tCbe TIim\>er0it tutorial Scried. ELEMENTARY . MATHEMATICAL ASTRONOMY, EXAMPLES AND EXAMINATION PAPERS. C. W. 0. ^ABLOW, M.A., B.Sc., GOLD MEDALLIST IN MATHEMATICS AT LONDON M.A., SIXTH WRANGLER, AND FIRST CLASS FIRST DIVISION PART II. MATHEMATICAL TRIPOS, CAMBRIDGE, AND GK H. BBYAN, D.So., M.A., F.E.S., SMITH'S PRIZEMAN, LATE FELLOW OK ST. PETER'S COLLEGE, CAMBRIDGE, JOINT AUTHOR OF " COORDINATE GEOMETRY, PART I.," " THE TUTORIAL ALGEBRA, ADVANCED COURSE," ETC. Third Impression (Second Edition). LONDON: W. B. OLIVE, (University Correspondence College Press], 13 BOOKSELLEB.S Row, STKAND, W.C. 1900. p PREFACE TO THE FIRST EDITION. FOR some time past it has been felt that a gap existed between the many excellent popular and non-mathematical works on As- tronomy, and the standard treatises on the subject, which involve high mathematics. The present volume has been compiled with the view of filling this gap, and of providing a suitable text-book for such examinations as those for the B.A. and the B.Sc. degrees of the University of London. It has not been assumed that the reader's knowledge of mathe- matics extends beyond the more rudimentary portions of Geometry, Algebra, and Trigonometry. A knowledge of elementary Dynamics will, however, be required in reading the last three chapters, but all dynamical investigations have been left till the end of the book, thus separating dynamical from descriptive Astronomy. The principal properties of the Sphere required in Astronomy have been collected in the Introductory Chapter ; and, as it is impossible to understand Kepler's Laws without a slight knowledge of the properties of the Ellipse, the more important of these have been collected in the Appendix for the benefit of students who have not read Conic Sections. All the more important theorems have been carefully illustrated by worked-out numerical examples, with the view of showing how the various principles can be put to practical application. The authors are of opinion that a far sounder knowledge of Astronomy can be acquired with the help of such examples than by learning the mere bookwork alone. Feb. 1st, 1892. PREFACE TO THE SECOND EDITION. THE first edition of Mathematical Astronomy having run out of print in less than eight months, we have hardly considered it advisable to make many radical changes in the present edition. We have, however, taken the opportunity of adding several notes at the end, besides answers to the examples, which latter will, we hope, prove of assistance, especially to private students ; our readers will also notice that the book has been brought up to date by the inclusion of the most recent discoveries. At the same time we hope we have corrected all the misprints that are inseparable from a first edition. Our best thanks are due to many of our readers for their kind assistance in sending us corrections and suggestions. Nov. 1st, 1892. CONTENTS. INTRODUCTORY CHAPTER. PAOB ON SPHERICAL GEOMETRY i Definitions ii Properties of Great and Small Circles iii On Spherical Triangles v CHAPTER I. THE CELESTIAL SPHERE. /Sect. I. Definitions Systems of Coordinates 1 II. The Diurnal Rotation of the Stars 13 III. The Sun's Annual Motion in the Ecliptic The Moon's Motion Practical Applications 20 CHAPTER II. THE OBSERVATORY. Sect. I. Instruments adapted for Meridian Observations 35 II. Instruments adapted for Observations off the Meridian 54 CHAPTER III. THE EARTH. Sect. I. Phenomena depending on Change of Position on the Earth 63 II. Dip of the Horizon 73 III. Geodetic Measurements Figure of the Earth 77 CHAPTER IV. THE SUN'S APPARENT MOTION IN THE ECLIPTIC. Sect. I. The Seasons 87 II. The Ecliptic 99 III. The Earth's Orbit round the Sun 105 CHAPTER V. ON TIME. ^/Sect. I. The Mean Sun and Equation of Time 115 II. The Sun-dial 125 III. Units of Time The Calendar 127 IV. Comparison of Mean and Sidereal Times 129 CONTENTS. CHAPTER VI. PACK ATMOSPHERICAL REFRACTION AND TWILIGHT 140 CHAPTER VII. THE DETERMINATION OF POSITION ON THE EARTH. Sect. I. Instruments used in Navigation 153 ^X, II. Finding the Latitude by Observation 102 ^ HI. To find the Local Time by Observation 171 IV. Determination of the Meridian Line 175 CXJ, V. Longitude by Observation 177 VI. Captain Sumner's Method 187 CHAPTER VIII. THK MOON. Sect. I. Parallax The Moon's Distance and Dimensions 191 II. Synodic and Sidereal Months Moon's Phases Mountains on the Moon 200 III. The Moon's Orbit and Rotation 209 CHAPTER IX. ECLIPSES. Sect. I. General Description of Eclipses 219 ,, II. Determination of the Frequency of Eclipses 224 III. Occultations Places at which a Solar Eclipse is visible 232 CHAPTER X. THE PLANETS. Sect. I. General Outline of the Solar System ... ... 238 II. Synodic and Sidereal Periods Description of the Motion in Elongation of Planets, as seen from the Earth Phases 244 III. Kepler's Laws of Planetary Motion 253 IV. Motion relative to Stars Stationary Points ... 258 V. Axial Rotations of Sun and Planets 264 CHAPTER XL THE DISTANCES OF THE SUN AND STARS. Sect. I. Introduction Determination of the Sun's Parallax by Observations of a Superior Planet at Opposition 267 II. Transits of Inferior Planets 271 ,, III. Annual Parallax, and Distances of the Fixed Stars 283 IV. The Aberration of Light ... 293 CONTENTS. DYNAMICAL ASTRONOMY. CHAPTER XII. PAOR THE ROTATION OF THE EARTH 315 CHAPTER XIII. THE LAW OP UNIVERSAL GRAVITATION. Sect. I. The Earth's Orbital Motion Kepler's Laws and their Consequences 337 II. Newton's Law of Gravitation Comparison of the Masses of the Sun and Planets 352 III. The Earth's Mass and Density 362 CHAPTER XIV. FURTHER APPLICATIONS OF THE LAW OF GRAVITATION. Sect. I. The Moon's Mass Concavity of Lunar Orbit... 371 II. The Tides 375 ,, III. Precession and Nutation 392 IV. Lunar and Planetary Perturbations 406 NOTES. Diagram for Southern Hemisphere 421 The Photochronograph 421 Effects of Dip, &c., on Rising and Setting 422 APPENDIX. Properties of the Ellipse 423 Table of Constants 426 ANSWERS TO EXAMPLES AND EXAMINATION QUESTIONS 428 INDEX 434 INTRODUCTORY CHAPTER, ON SPHERICAL GEOMETRY. Properties of the Sphere which will be referred to in the course of the Text. (1) A Sphere may be defined as a surface all points on which are at the same distance from a certain fixed point. This point is the Centre, and the constant distance is the Radius. (2) The surface formed by the revolution of a semicircle about its diameter is a sphere. For the centre of the semicircle is kept fixed, and its distance from any point on the surface generated will be equal to the radius of the semicircle. FIG. 1. (3) Let PqQP' be any position of the revolving semicircle whose diameter PP' is fixed. Let OQ be the radius perpendicular to PP', Cq any other line perpendicular to PP', meeting the semicircle in q. (We may suppose these lines to be marked on a semicircular disc of cardboard.) As the semicircle revolves, the lines OQ, Cgwill sweep out planes perpendicular to PP', and the points Q, q will trace out in these planes circles HQRK, hqrJc, of radii OQ, Cq respectively. From this it may readily be seen that Every plane section of a sphere is a circle, 4-STKON, 5 ii ASTRONOMY. Definitions. (4) A great Circle of a sphere is the circle in which it is cut by any plane passing through the centre (e.g., HQRK, PqQP' or PrRP ). A small circle is the circle in which the sphere is cut by any plane not passing through the centre (e.g., hqrk). (5) The axis of a great or small circle is the diameter of the sphere perpendicular to the plane of the circle. The poles of the circle are the extremities of this diameter. (Thus, the line PP is the axis, and P, P' are the poles of the circles HQK and hqJc). (6) Secondaries to a circle of the sphere are great circles passing through its poles. (Thus, PQP' and PRP" are secondaries of the circles HQK, hqk). FIG. 2. (7) The angular distance between two points on a sphere is measured by the arc of the great circle joining them, or by the angle which this arc subtends at the centre of the sphere. Thus, the dis- tance between Q and Bis measuredeither by the arc QE, or by the angle QOR. Since the circular measure of L QOR = (arc Qft) -f- (radius of sphere), it is usual to measure arcs of great circles by the angles which they subtend at the centre. This remark does not apply to small circles. (8) The angle between two great circles is the angle between their planes. Thus, the angle between the circles PQ, PR is the angle between the planes PQP', 7'EP'. It is called "the angle QPR." (9) A spherical triangle is a portion of the spherical surface bounded by three arcs of gr.eat_circles. Thus, in Fig. 2, PQR is a spherical triangle, but Pqr is not a spherical triangle, because qr is not an arc of a great circle. We may, however, draw a great circle passing through q and r, and thus form a spherical triangle Pqr. SPHERICAL GEOMETRY. ill Properties of Great and Small Circles. (10) All points on a small circle are at a constant (angular)) distance from the pole. For, as the generating semicircle revolves about PP 7 , carrying g along the small circle hk, to r, the arc Pq = arc Pr, and Z POq = L POr. The constant angular distance Pq is called the spherical, or angular radius of the small circle. The pole P is analogous to the centre of a circle in plane geometry. (11) The spherical radius of a great circle is a quadrant, or, All points on a great circle are distant 90 from its poles. For, as Q, by revolving about PP', traces out the great circle HQRK, we have L POQ = L POR = 90, and therefore, PQ, PE are quadrants. (12) Secondaries to any circle lie in planes perpendicular to the plane of the circle. For PP' is perpendicular to the planes of the circles HQK, liqk, therefore any plane through PP / , such as PQP' or PEP', is also per- pendicular to them. (13) Circles which have the same axis and poles lie in parallel, planes. For the planes HQK, hqk are parallel, both being perpen- dicular to the axis PP'. Such circles are often called parallels. (14) If any number of circles have a common diameter, their poles all lie on the great circle to which they are secondaries, and this great circle is a common secondary to the original circles. For if OA is the axis of the circle PQP', then OA is perpendicular- to POP'. Hence, if the circle PQP 7 revolves about PP', A traces out. the great circle HQRK, of which P, P 7 are poles. We likewise see that (15) If one great circle is a secondary to another, the latter is also a secondary to the former. This is otherwise evident, since their planes are perpendicular. (16) The angle between two great circles is equal to (i.) The angle between the tangents to them at their points of intersection ; (ii.) The arc which they intercept on a great circle to which they are both secondaries ; (iii.) The angular distance between their poles. Let Ft, Pu be the tangents at P to the circles PQ, PE, and let A, B bo the poles of the circles. If we suppose the semicircle PQP' to revolve about PP' into the position PEP', the tangent at P will revolve from Pt to Pu, the radius perpendicular to OP will revolve from OQ to 07?, and the axis will revolve from OA to OB. All these lines will revolve through an angle equal to the angle between the planes PQP', PRP / , and this is the angle QPE between the circles (Def. 8). BLenee, le between circles PQ, PR = L tPu = L QOR {y ASTEONO^TT. (17) The arc of a small circle subtending a given angle at the pole is proportional to the sine of the angular radius. Let qr be the arc of the small circle hqrJc, subtending L qPr at P, and let G be the centre of the circle. Evidently L qCr = L QOR (since Cq, Gr are parallel to OQ, OB). Hence, the arcs qr, QR are proportional to the radii Cq, OQ, . arc qr = G = Gq_ = ghl pQq = gin p^ arc QR OQ Oq But QR is the arc of a great circle subtending the same angle at the pole P hence the arc qr is proportional to sin Pq, as was to be shown. Since qQ = 90 - PQ, therefore sin Pq - cos gQ, so that the arc qr is proportional to the cosine of the angular distance of the small circle (jr from the parallel great circle QR. FIG. 3. FIG. 4. (18) Comparison of Plane and Spherical Geometry. It may be laid down as a general rule that great circles and small circles on a sphere are analogous in their respective properties to straight lines and circles in a plane. Thus, to join two points on a sphere means to draw the great circle passing through them. Secondaries to a great circle of the sphere are analogous to per- pendiculars on a straight line. The distance of a point from any great circle is the length of the arc of a secondary drawn from the point to the circle. Thus, rR is the distance of the point r from the great circle HQRK. SPHEEICAL GEOMETftf. V On Spherical Triangles* (19) Parts of a Spherical Triangle. A spherical triangle, like a plane triangle, has six parts, viz., its three sides and its three angles. The sides are generally measured by the angles they subtend at the centre of the sphere, so that the six parts are all expressed as angles. Any three given parts suffice to determine a spherical triangle, but there are certain " ambiguous cases " when the problem admits of more than one solution. The formulge required in solving spherical triangles form the subject of Spherical Trigonometry, and are in every case different from the analogous f ormulaj in Plane Trigonometry. There is this further difference, that a spherical triangle is completely determined if its three angles are given. Thus, two spherical triangles will, in general, be equal if they have the following parts equal : (i.) Three sides. (ii.) Two sides andincluded angle. (iii.) Two sides and one opposite angle. (iv.) Three angles, (v.) Twoanglesandadjacentside. (vi.) Two angles and one opposite side. Cases (iii.) and (vi.) may be ambiguous. (20) Right-angled Triangles. If one of the angles is a right angle, two of the remaining five parts will determine the triangle. (21) Triangle with two right angles. The properties of a spherical triangle, such as PQR, Fig. 3, in which one vertex P is the pole of the opposite side QR, are worthy of notice. Here two of the sides, PQ, PR, are quadrants, and two angles Q, R are right angles. The third side QR is equal to the opposite angle QPR, (22) Triangle with, three right angles. If, in addition, the angle QPR is a right angle (Fig. 4), QR will be a quadrant. The triangle PQR will, therefore, have all its angles right angles, and all its sides quadrants, and each vertex will be the pole of the opposite side. The planes of the great circles forming the sides, are three planes through the centre mutually at right angles, and they divide the surface of the sphere into eight of these triangles ; thus the area of each triangle is one-eighth of the surface of the sphere. (23) The three angles of a spherical triangle are together greater than two right angles. [For proof, see any text-book on Spherical Geometry.] (24) If the sides of a spherical triangle, when expressed as angles, are very small, so that its linear dimensions are very small com- pared with the radius of the sphere, the triangle is very approxi- mately a plane triangle. Thus, although the Earth's surface is spherical, a triangle whose sides are a few yards in length, if traced on the Earth, will not be distinguishable from a plane triangle. If the sides are several miles in length, the triangle will still be very nearly plane. vi AJSTKONOMY. (25) Any two sides 6f a spherical triangle are together greater than the third side. For if we consider the plane angles which the sides subtend at the centre of the sphere, any two of these are together greater than the third, by Euclid XL, 20. (26) The following application of (25) is of great use in astronomy, and is analogous to Euclid III., 7, 8. Let AHBK be any given great or small circle whose pole is P, Zany other given point on the sphere, and let the great circle ZP meet the given circle in the points A, B. Then A, B are the two points on the given circle whose distances from Z are greatest and least respectively. For let H be any other point on the circle. Join ZH, HP. Then, in spherical A ZPH, ZP + PH> ZH. But PH = PA ; /. ZP + PA > ZH, i.e., ZA>ZH. Also, if Z is on the opposite side of the circle to P, then ZII+PH>PZ', .:ZH + PB>PZ; .:ZH>PZ-PB, i.e., ZH>ZB. If Z' be a point on the same side of the circle as P, then PZ' + Z'H >PH. But PH - PB. .'. PZ'-t Z'H^PB. .-. Z'H>PB-PZ', i.e., Z'H>Z'B, as before. lie nee, A is further from Z, Z', and B is nearer to Z, Z', than any other point on the circle. (27) If H, K are the two points on the circle equidistant from Z, the spherical triangles ZPH, ZPK have ZP common, ZH = ZK (by hypothesis^), and PH = PK [by (10)], hence they are equal in all respects ; thus L ZPH = L ZPK, and L PZH = L PZK. Hence PH, PK are equally inclined to PB, as are also ZH, ZK. Similar properties hold in the case of the point Z'. These pro- perties are of frequent uw. ASTRONOMY. CHAPTEE I. THE CELESTIAL SPHERE. SECTION I. Definitions Systems of Co-ordinate*. 1 . Astronomy is the science which deals with the celestial bodies. These comprise all the various bodies distributed throughout the universe, such as the Earth (considered as a whole), the Sun, the Planets, the Moon, the comets, the fixed stars, and the nebulae. It is convenient to divide Astronomy into three different branches. The first may be called Descriptive Astronomy. It is concerned with observing and recording the motions of the various celestial bodies, and with applying the results of such observations to predict their positions at any subsequent time. It includes the determination of the distances, and the measurement of the dimensions of the celestial bodies. The second, or Gravitational Astronomy, is an appli- cation of the principles of dynamics to account for the motions of the celestial bodies. It includes the determination of their masses. The third, called Physical Astronomy, is concerned with determining the nature, physical condition, temperature, and chemical constitution of the celestial bodies. The first branch has occupied the attention of astronomers in all ages. The second owes its origin to the discoveries of Sir Isaac Newton in the seventeenth century ; while the third branch has been almost entirely built up in the present century. In this book we shall treat exclusively of Descriptive and Gravitational Astronomy. ASTRONOMY. : -;2: :The ;C.elesti.al Sphere. On observing the stars it is ' not^ 'difficult to imagine that they are bright points dotted about on the inside of a hollow spherical dome, whose centre is at the eye of the observer. It is impossible to form any direct conception of the distances of such remote bodies ; all we can see is their relative directions. Moreover, mof-t astronomical instruments are constructed to determine only the directions of the celestial bodies. Hence it is important to have a convenient mode of representing directions. FIG. 6. The way in which this is done is shown in Figure 6. Let be the position of any observer, A, , C, &c., any stars or other celestial bodies. About 0, as centre, describe a sphere with any convenient length as radius, and let the lines joining to the stars A, J3, C meet this sphere in a, ft, c respectively. Then the points a, I, c will represent, on the sphere, the directions of the stars A, H, C, for the lines joining these points to will pass through the stars themselves. In this manner we obtain, on the sphere, an exact representation of the appearance of the heavens as seen from 0. Such a sphere is called the Celestial Sphere. This sphere may be taken as the dome upon which the stars appear to lie. But it must be carefully borne in mind that the stars do not actually lie on a sphere at all, and that they are only so represented for the sake-of convenience. THE CELESTIAL SPHERE. 3. Use of the Globes. The representation of directions of stars by points on a sphere is well exemplified in the old- fashioned star globes. Such a globe may be used as the observer's celestial sphere ; but it must be remembered that the directions of the stars are the lines joining the centre to the corresponding points on the sphere ; for in every case the observer is supposed to be at the centre of the celestial sphere. The properties given in the Introduction on Spherical Geo- metry are applicable to the geometry of the celestial sphere. A knowledge of thorn will be assumed in what follows. 4. Angular Distances and Angular Magnitudes. Any plane through the observer will be represented on the celestial sphere by a great circle. The arc of the great circle a b (Fig. 6) represents the angle a 01 or A OB which the stars A, subtend at 0. This angle is generally measured in degrees, minutes, and seconds, and is called the angular distance between the stars. This angular distance must not be confused with their actual distance AB. In the same way, when we are dealing with a body pf perceptible dimen- sions, such as the Sun or Moon (DF, Fig. 6), we shall define its angular diametsr as the angle DOF, subtended by a diameter at the observer's eye. This angular diameter is measured by the arc df of the celestial sphere, that is, by the diameter of the projection of the body on the celestial sphere. From the figure it is evident that Od 01)' Since DF is the actual linear diameter of the body, mea- sured in units of length, the last relation shows us that the angular diameter (df) of a body varies directly as its linear diameter DF, and inversely as OD, the distance of the body from the observer's eye. As the eye can only judge of the dimensions of a body from its angular magnitude, this result is illustrated by the 1'act that the nearer an object is to the eye the larger it looks, and vice versd. Thus, if the distance of the object be doubled, it will only look half as large, as may be easily verified. 4 ASTRONOMY. 5. The Directions of the Stars are very approxi- mately independent of the Observer's Position on the Earth. This is simply a consequence of the enormously great dis- tances of all the stars from the Earth. Thus, let x (Fig. 7) denote any star or other celestial body, S, JZtwo different positions o^ the observer. If the distance SJ be only a very small fraction of the distance Sx, the angle Ex 8 will be very small, and this angle measures the difference be- tween the directions of x as seen from ^and from 8. In illustration, if we observe a group of objects a mile or two off, and then walk a few feet in any direction, we shall observe no perceptible change FIG. 7. in the apparent directions or relative positions of the objects. If Ex be drawn parallel to Sx, the angle xEx will be equal to ExS, and will therefore be very small indeed. Hence, Ex will very nearly coincide in direction with Ex'. Thus, considering the vast distances of the stars, we see that The lines joining a Star to different points of the Earth may be considered as parallel.* The stars will, therefore, always be represented by the same points on a star globe, or celestial sphere, no matter what be the position of the observer. The great use of the celestial sphere in astronomy depends on this fact. 6. Motion of Meteors. The projection of bodies on the celestial sphere is well illustrated by the apparent motion of a swarm of meteors. Where such a swarm is moving uniformly, all the meteors describe (approximately) parallel straight lines. II we draw planes through these lines and the observer, they will intersect in a common line, namely, the line through the observer parallel to the direction of the common motion of the meteors. The planes will, therefore, cut the celestial sphere in great circles, having this line as their common diameter. These great circles represent the apparent paths >i (he meteors on the celestial sphere. The paths appear, therefore, to radiate from a common point, namely, one of the extremities of this diameter. This point is called the Radiant, and by observing its position the direction of motion of the meteors is determined. * This is not true in the case of the Moon. tHE CELESTIAL StHE&E. 6 7. Zenith and Nadir. Horizon. If, through the observer, a line be drawn in the direction in which gravity acts (i.e., the direction indicated by a plumb-line), it will meet the celestial sphere in two points. One of these is vertically above the observer, and is called the Zenith; the other is vertically below the observer, and is called the Nadir. (Fig. 6, and Z, N, Fig. 8.) If the plane through the observer parallel to the surface of a liquid at rest be produced, it will cut the celestial sphere in a great circle. This great circle is called the Celestial Horizon. (Fig. 6, and sEnW, Fig. 8.) It is proved in Hydrostatics that the surface of a liquid at rest is a plane perpendicular to the direction of gravity. Hence, the celestial horizon is the great circle whose poles are the zenith and nadir. "We might have defined the horizon by this property. From the above definition, it is evident that, to an observer whose eye is close to the surface of the ocean, the celestial horizon forms the boundary of the visible portion of the celestial sphere. On land, however, the boundary, or visible horizon (as it is called), is always more or less irregular, owing to trees, mountains, and other objects. 8. Diurnal Motion of the Stars. If we observe the sky at different intervals during the night, we shall find that the stars always maintain the same configurations relative to one another, but that their actual situations in the sky, relative to the horizon, are continually changing. Some stars will set in the west, others will rise in the east. One star which is situated in the constellation called the l< Little Bear," remains almost FlG - 8 - fixed. This star is called Polaris, or the Pole Star. All the other stars describe on the celestial sphere small circles (Fig. 8) having a common pole P very near the Pole Star, and the revolutions are performed in the same period of time, namely, about 23 hours 56 minutes of our ordinary time. 6 ASTEONOMt. 9. Celestial Poles, Equator, and Meridian. The common motion of the stars may most easily be conceived by imagining them to be attached to the surface of a sphere which is made to revolve uniformly about the diameter PP'. The extremities of this diameter are called the Celestial Poles. That pole, P, which is above the horizon in northern latitudes is called the North Pole, the other, P\ is called the South Pole. The great circle, JEQR W, having these two points for its poles, is called the Celestial Equator. It is, therefore, the circle which would be traced out by the diurnal path of a star distant 90 from either pole. The Meridian is the great circle (PZP'N, Fig. 9) passing through the zenith and nadir and the celestial poles. It cuts both the horizon and equator at right angles [by Spher. Geom. (12), since it passes through their poles]. THE CELESTIAL SPHEKE. 7 10. The Cardinal Points. The East and West Points (J, W, Eig. 9) are the points of intersection of the equator and horizon. The North and South Points (&, S) are the intersections of the meridian with the horizon. Verticals. rSecondaries to the horizon, i.e., great circles through the zenith and nadir., are called Vertical Circles, or, briefly, Verticals. Thus, the meridian is a vertical. The Prime Vertical is the vertical circle (ZENTF) passing through the east and west points. Since P is the pole of the circle QERW, and ^is the pole of nEsWy therefore E, W are the poles of the meridian PZP'N. Hence the horizon, equator, and prime vertical which pass through E, W, are all secondaries to the meridian ; they therefore all cut the meridian at right angles. 11. Annual Motion of the Sun. The Ecliptic. The Sun, while participating in the general diurnal rotation of the heavens, possesses, in addition, an independent motion of its own relative to the stars. Imagine a star globe worked by clockwork so as to revolve about an axis pointing to the celestial pole in the same peri- odic time as the stars. On such a moving globe the directions of the stars will always be represented by the same points. During the daytime let the direction of the Sun be marked on the globe, and let this process be repeated every day for a year. We shall thus obtain on the globe a representation of the Sun's path relative to the stars, and it will be found that (i.) The Sun moves from west to east, and returns to the same position among the stars in the period called a year ; (ii.) The relative path on the celestial sphere is a great circle, inclined to the equator at an angle of about 23 27f. This great circle (CTL ===, Fig. 9) is called the Ecliptic. "We may, therefore, briefly define the ecliptic as the great circle which is the trace, on the celestial sphere, of the Sun's annual path relative to the stars. The intersections of the ecliptic and equator are called Equinoctial Points. One of them is called the First Point of Aries ; this is the point through which the Sun passes when crossing from south to north of the equator, and it is usually denoted by the symbol T The other is called the First Point of Libra, and is denoted by the symbol =0=, ASTKONOMY. 12. Coordinates. In Analytical Geometry, the position of a point in a plane is denned by two coordinates. In like manner, the position of a point on a sphere may be denned by means of two coordinates. Thus, the position of a place on the Earth is denned by the two coordinates, latitude and longitude. For fixing the positions of celestial bodies, the following different systems of coordinates are used. 13. Altitude or Zenith Distance and Azimuth. Let Fig. 10 represent the celestial sphere, seen from overhead, and lot x be any star. Draw the vertical circle ZxX. Then the position of x may be defined by either of the following pairs of coordinates, which are analogous to the Cartesian and polar coordinates of a point in a plane respectively : (a) The arc s X and the arc Xx ; (b) The arc Zx and the angle sZx. Practically, however, the two systems are equivalent ; for, since Z is the pole of sX, ZX = 90, therefore Zx = 90 xXj and angle sZx = arc sX, FIG. 10. The Altitude of a star (Xx} is its angular distance from the horizon, measured along a vertical. The Zenith Distance (abbreviation, Z.D.) is its angular distance from the zenith (Zx) , or the complement of the altitude. The Azimuth (sX or sZx) is the arc of the horizon inter- cepted between the south point and the vertical of the star, or the angle which the star's vertical makes with the meridian THE CELESTIAL SPHERE. 9 *14. Points Of the Compass. In practical applications of Astro- nomy to navigation, it is usual to measure the azimuth in "points" and " quarter points " of the compass. The dial plate of a mariner's compass is divided into 32 points, by repeatedly bisecting the right angles formed by the directions of the four cardinal points. Thus each point represents an angle of Hi degrees. The points are again subdivided into " quarter points " of 2\ degrees. Starting from the north and going round towards the east, the various points are denoted as follows : N., N. byB., N.N.E., N.E. by N., N.E., N.E. by E., E.N.E., E. by N. E., E. byS., E.S.E., S.E. by E., S.E., S.E. by S., S.S.E., S. by E. S., S. by W. S.S.W., S.W. by S., S.W., S.W: by W., W.S.W , W. by S. W., W. by N., W.N.W. N.W. by W., N.W., N.W. by N., N.N.W., N. by W. The quarter points are denoted thus : E.N.B. E. means one quarter point to the eastward of E.N.E., that is, 6 points, or 70 18' 45", from the north point, taken in an easterly direction. So, too, S.S.W. W. meafli 2J points, or 28 7' 30' , measured from the south point westwards. 15. Polar Distance, or Declination, and Hour Angle. From the pole P, draw through x the great circle PxM-, this circle is a secondary to the equator EQ, W. Then we may take for the coordinates of x the arc Px and the angle sPx. Or we may take the arc x3f, which is the complement of Px, and the arc QM, which = angle QPx. The North Polar Distance of a star (abbreviation, N.P.D.) is its angular distance (Pa;) from the celestial pole. The Declination (abbreviation, Decl.) is the angular distance from the equator (xM), measured along a secondary, and is, therefore, the complement of the N.P.D. The great circle PxM through the pole and the star is called the star's Declination Circle. The Hour Angle of the star (ZPx] is the angle which the star's declination circle makes with the meridian. The declination may be considered positive or negative, according as the star is to the north or south of the equator, but it is more usual to specify this by the letter N. or S., as the case may be, and this is called the name of the declination. The hour angle is generally measured from the meridian towards the west, and is reckoned from to 360. Either the declination and hour angle or the N.P.D. and hour angle may be taken as the two coordinates of a star. 10 ASTBONOHY. 16. Declination and Right Ascension. The position of a celestial body is, however, more frequently defined by its declination and right ascension. 'The declination has been already defined, in 15, as the angular distance of the star from the equator, measured along a secondary. (xM, Fig. 11.) The Right Ascension (E.A.) is the arc of the equator intercepted between the foot of this secondary and the First Point of Aries. Thus, T^, Fig. 11, is the E.A. of the star a:. The E.A. of a star is always measured from T eastwards reckoning from to 360. Thus the star w Piscium, whose declination circle cuts the equator 1 34' 18" west of T, has the E.A. 360 1 34' 18", or 358 25' 42". FIG. 11. 17. Celestial Latitude and Longitude. The position of a celestial body may also be referred to the ecliptic instead of the equator. The Celestial Latitude is the angular distance of the tody from the ecliptic, measured along a secondary to the ecliptic. (Hx, Pig. 11.) The Celestial Longitude is the arc of the ecliptic inter- cepted between this secondary and the first point of Aries, measured eastwards from T- (T#, Pig. 11.) tflE CELESTIAL SPHERE. ll 18. Latitude of the Observer. The celestial latitude and longitude, defined in the last paragraph, must not be confounded with the latitude and longitude of a place on the Earth, as there is no connection whatever between them. The Latitude of a place is the angular distance of its zenith from the equator, measured along the meridian. Thus, in Pig. 1 1 , ZQ, is the latitude of the observer. Since PQ nZ 90 ; .-. ZQ = nP, or in other words, The latitude of a place is the altitude of the Celestial Pole. The complement of the latitude is called the Colatitude. Hence, in Pig. 11, PZ is the colatitude of the observer, and is the angular distance of the zenith from the pole. In this book the latitude of an observer will generally be denoted by the symbol /, and the colatitude by c. The longitude of a place will be defined in Chapter III. 19. Obliquity of the Ecliptic. The inclination of the ecliptic to the equator is called the Obliquity. In Pig. 11, Q T C is the obliquity. As stated in 1 1 , this angle is about 23 27-'. We shall generally denote the obliquity by i. 20. Advantages of the Different Coordinate Systems. The altitude and azimuth of a celestial body indicate its position relative to objects on the Earth. Owing, however, to the diurnal motion, they are constantly changing. The N.P.D. and hour angle also serve to determine the star's position relative to the earth, and have this further advantage, that the N.P.D. is constant, while the hour angle increases at a uniform rate. Since the equator and first point of Aries partake of the common diurnal motion of the stars, the declination and right ascension of a star are constant. These coordinates are, there- fore, the most suitable for tabulating the relative positions of the various stars on the celestial sphere. The celestial latitude and longitude of a celestial body are also unaffected by the diurnal motion. They are most useful in defining the positions of the Sun, Moon, planets, and comets, for the first always moves in the ecliptic, while the paths described by the others are always very near the ecliptic. 21. Recapitulation. Por the sake of convenient refer- ence, we give on the next page a list of all the definitions of this chapter, with references to Pigs. 11, 12. ASTRON. c 12 ASTRONOMY. GREAT CIRCLES. Horizon, nEsW. Equator, EQWR. Meridian, ZsZ'n. Prime Vertical, ZEZ'W. THEIR POLES. Zenith, Z-, Nadir, Z '. North Pole, P ; South Pole, P. East Point, E\ West Point, W. NorthPoint, n ; South Point, s. Ecliptic, T i:Z ; Equinoctial Points, T, =2=, viz. : Eirst Point of Aries, T , and Eirst Point of Libra, b ; Yertical of Star, ZxX-, Declination Circle of Star, Pxlf. FIG. 12. COORDINATES. Altitude, Xx ; '") or Zenith Distance, Zx. ) North Polar Distance, Px. Declination, MX. Celestial Latitude, Hx. Azimuth, sX = sZx. Hour Angle, QM = ZPx. Bight Ascension, T ^ Celestial Longitude, OTHER ANGLES. Obliquity of Ecliptic (t) CT Q- Observer's Latitude (1) = ZQ = nP. Colatitude (c) = PZ. Notice that the circles on the remote side of the celestial sphere are dotted. CELESTIAL SPHEKE. 13 SECTION II. The Diurnal Rotation of the Stars. 22. Sidereal Day and Sidereal Time. A Sidereal Day is the period of a complete revolution of tlie stars about the pole relative to the meridian and horizon. Like the common day it is divided into 24 hours (h.), and these are subdivided into 60 minutes (m.) of 60 seconds (s.) each. The sidereal day commences at "Sidereal Noon," i.e., the instant when the first point of Aries crosses the meridian. The Astronomical Clock, which is the clock used in observatories, indicates sidereal time. The hands should indicate Oh. Om. Os. when the first point of Aries crosses the meridian, and the hours are reckoned from Oh. up to 24h., when T again comes to the meridian and a new day begins. From the facts stated in 8, it appears that the sidereal day is about 4 minutes shorter than the ordinary day. The stars are observed to revolve about the pole at a perfectly uniform rate, so that the sidereal day is of invariable length, and the angles described by any star about the pole are pro- portional to the times of describing them. Thus, the hour angle of a star (measured towards the west) is proportional to the interval of sidereal time that has elapsed since the star was on the meridian. Now, in 24 sidereal hours the star comes round again to the meridian, after a complete revolution, the hour angle having increased from to 360. Hence the hour angle in- creases at the rate of 15 per hour. Hence, also, it increases 15' per minute, or 15" per second. The hour angle of a star is, for this reason, generally measured by the number of hours, minutes, and seconds of sidereal time taken to describe it. It is then said to be expressed in time. Thus, The hour angle of a star, when expressed in time* is the interval of sidereal time that has elapsed since the star was on the meridian. In particular, since the instant when T is on the meridian is the commencement of the sidereal day, we see that The sidereal time is the hour angle of the first point of Aries when expressed in time. 14 ASTHONOMY. 23. To reduce to angular measure any angle ex- pressed in time. Multiply ~by 15. The hours, minutes, and seconds of time will thus be reduced to degrees, minutes, and seconds of angle. Conversely, to reduce to time from angular measure we must divide by 15, and for degrees, minutes, and seconds, write hours, minutes, and seconds. EXAMPLES. 1. To find, in angular measure, the hour angle of a star at 15h. 21m. 50s. of sidereal time after its transit. The process stands thus 15 21 50 230 27 30 /. the angular measure of the hour angle is 230 27' ?0" 2. To find the sidereal time required to describe 230 27' 30" (converse of Ex. 1). 15 ) 230 27 30 15 21 50 ; .-. required time = 15h. 21m. 50s. 24. Transits. The passage of the star across the meri- dian is called its Transit. Let x be the position of any star in transit (Fig. 13). The star's E.A. = T Q or rPQ = hour angle of T = sidereal time expressed in angle. Hence, the right ascension of a star, when ex- pressed in time, is equal to the sidereal time of its transit. In practice the R.A. of a star is always expressed in time. Thus, the R.A. of a Lyrse is given in the tables aa 18h. 33m. 14-8s., and not as 278 18' 42". THE CELESTIAL SPHEEE. 15 Again, let 2 be the meridian zenith distance Zx, considered positive if the -star transits north of the" zenith, d the star's north declination Qx, and I the north latitude QZ. Wo have evidently - Qx = QZ+Zx; d = i+*c or (star's N. decl.) = (lat. of observer) + (star's meridian Z.D.) This formula will hold universally if declination, latitude, and zenith distance are considered negative when south. Hence the R. A. and decl. of a star maybe found by observing its sidereal time of transit and its meridian Z.D., the latitude of the observatory being known. Conversely, if the R.A. and decl. of a star are known, we can, by observing its time of transit and meridian Z.D., deter- mine the sidereal time and the latitude of the observatory. By finding the sidereal time we may set the astronomical clock. A star whose E.A. and decl. have been tabulated, is called a known star. In Chapter II. we shall describe an instrument called the Transit Circle, which is adapted for observing the times of transit and meridian zenith distances of celestial bodies. 25. General Relation between R.A. and hour angle. Let x l (Fig. 13) be any star not on the meridian. Then z Qp Xl = L QPr- t rP^ = ^ QPr rM] hence, if angles are expressed in time, (star's hour angle) = (sidereal time) (star's H.A.). Hence, given the 11. A. and decl. of a star, we can find its hour angle and N.P.D. at any given sidereal time, and by this means determine the star's position on the 'observer's celestial sphere. Or we can construct the star's position thus On the equator, in the westward direction from Q, measure off Q T equal to the sidereal time (reckoning 15 to the hour). Prom T east- wards, measure f M equal to the star's It. A.; and from 3f, in the direction of the pole, measure off Mx l equal to the star's declinatiqn. We thus find the star x r 1 6 ASTRONOMY. *26. Transformations. If the R.A. and decl. of a star are given, its celestial latitude and longitude may be found, and vice versti ; but the calculations require spherical trigonometry. The process is analogous to changing the direction of the axes through an angle i, in plane coordinate geometry. Again, the Z.D. and azimuth may be calculated from the N.F.D. and hour angle, by solving the triangle ZPx^ We know the colatitude PZ, Px^ and L ZPx t , and we have to determine Zxi and L QZx } (= ISO PZxJ. In the last article we showed how to find the hour angle in terms of the R.A., or vice versA, the sidereal time being known. Hence we see that, given the coordinates of a star referred to one system, its coordinates referred to any other of the systems can bo calculated at any given instant of sidereal time. 27. Culmination and Southing of Stars. A celestial body is said to culminate when its altitude is greatest or least. Since the fixed stars describe circles about the pole, it readily follows, from Spherical Geometry (26), that a star attains its greatest or least zenith distance when on the meridian, and, therefore, that its culmination is the same as its transit. This is not strictly the case with the Sun, because, owing to its independent motion, its polar distance is not constant ; hence it does not describe strictly a small circle about the pole. When a star transits S. of the zenith it is said to south. 28. Circumpolar Stars. A Circumpolar Star at any place is a star whose polar distance is less than the latitude of the place. Its declination must, therefore, be greater than the colatitude. On the meridian let Px and Px' be measured, each equal to the KP.D. of such a star (Fig. 14). Then x and x' will be the positions of the star at its transits. Since Px < Pn, both x' and x will be above n. Hence, during a sidereal day a cir- cumpolar star will transit twice, once above the pole (at x) and once below the pole (at x'), and both transits will be visible. The two transits are distinguished as the upper and lower culminations respectively, and they succeed one another at intervals of 12 sidereal hours ( since xPx' = 180). The altitude of the star is greatest at upper, and least at lower culmination, as may easily be seen from Sph. Geom. (26) by considering the zenith distances. Hence the altitude is never less than nx, and the star is always above the horizon. Since THE CELESTIAL SPHEBE. nx-nP=Px = Px = nPnaf, 17 that is, The observer's latitude is half the sum of the altitudes of a circumpolar star at upper and lower culminations. Also, Px \ (nx nx) ; that is, The Star's N.P.D. is half the difference of its two meridian altitudes. These results will require modification if the upper culmi- nation takes place south of the zenith as at 8. The meridian altitude will then be measured by sS, and not nS. Here, nS = 180 sS, and we shall, therefore, have to replace the altitude at upper culmination by its supplement. South Circumpolar Stars. If the south polar dis- tance of a star is less than the north latitude of the observer, the star will always remain below the horizon, and will, therefore, be invisible. Such a star is called a South Cir- cumpolar Star. EXAMPLE. The constellation of the Southern Cross ( Crux) is invisible in Europe, for its declination is 62 30' S ; there- fore its south polar distance is 27 30', and it will, therefore, pot be visible in north latitudes higher than 27 30'. 18 ASTBONOMY. 29. Rising, Southing, and Setting of Stars. If the N. and S. polar distances of a star are both greater than the latitude, it will transit alternately above and below the horizon. This shows that the star will be invisible during a certain portion of its diurnal course. Astronomically, the star is said to rise and set when it crosses the celestial horizon. Let J, V be the positions of any star when rising and setting respectively. FIG. 15. In the spherical triangles Pnb, PI = Pb' (each being the star's KP.D.), right L Pnb = right L Pnb', and Pn is common. Hence the triangles are equal in all respects ; therefore Z nPb = Z nPb', and the supplements of these angles are also equal, that is, L sPb = L sPb'. But the angle sPb, when reduced to time, measures the interval of time taken by the star to get from b to the meri- dian, and sPV measures the time taken from the meridian to b'. Hence, The interval of time between rising and southing is equal to the interval between southing and setting. THE CELESTIAL SPHERE. 19 Thus, if , f are the times of rising and setting, and T the time of transit, we have T t tfT. The time of transit is the arithmetic mean between the times of rising and setting. In order to facilitate the calculations, tables have been constructed giving the values of T t for different latitudes and declinations. If the observer's latitude Pn and the star's polar distance Pb are known, it is possible (by Spherical Trigonometry) to solve the right- angled triangle PZm, and to calculate the angle nPb, and therefore also the angle &Ps. This angle, when divided by 15, gives the time T t. Moreover, the sidereal time of transit T is known, being equal to the star's R.A. Hence the sidereal times of rising and setting can be found. If the star is on the equator, it will rise at E and set at W. Since JSQWis a semicircle, exactly half the diurnal path will be above the horizon, and the interval between rising and setting will be 12 sidereal hours. If the star is to the north of the equator, it will rise at some point b between E and , so that L IPs > Z JEPs, i.e., / bPs > 90, and the star will he above the horizon for more than 12 hours. Similarly, if the star is south of the equator, it will rise at a point c between E and *, and will be above the horizon for less than 12 hours. Prom the equality of the triangles bPn, b'Pn (Pig. 15), we also see that nb = nb', and sb = sb'. Hence the diameter (ns) of the celestial sphere, joining the north and south points, bisects the arc (W) between the directions of a star at rising and setting. This gives us an easy method of roughly determining, by observation, the directions of the cardinal points ; but, owing to the usual irregularities in the visible horizon, the methoij is not very exac. 20 ASTRONOMY. SECTION III. The Sun's Annual Motion in the Ecliptic The Moon's Motion Practical Applications. 30. The Sun's Motion in Longitude, Bight Ascen- sion and Declination. In 11, we briefly described the Sun's apparent motion in the heavens relative to the fixed stars. "We defined a Year as the period of a complete revolution, starting from and returning to any fixed point on the celestial sphere. The Ecliptic was defined as the great circle traced out by the Sun's path, and its points of intersection with the Equator were termed the First Point of Aries and First Point of Libra, or together, the Equinoctial Points. We shall now trace, by the aid of Pig. 16, the variations in the Sun's coordinates during the course of a year, starting with March 21st, when the Sun is in the first point of Aries. We shall, as usual, denote the obliquity by i, so that i = 23 27' nearly. FIG. 16. On March 21st the Sun crosses the equator, passing through the first point of Aries (r). This is the Vernal Equinox, and it is evident from the figure that Sun's longitude = 0, B.A. = O, Decl. = 0. Prom March 21st to June 2 1st the Sun's declination is north, and is increasing. THE CELESTIAL SPHEEE. 21 On June 21st the Sun has described an arc of 90 from r on the ecliptic, and is at C (Fig. 16). This is called the Summer Solstice. If we draw the declination circle PCQ, the spherical triangle T OQ is of the kind described in Sph. Geom. (21), and CP is a secondary to the ecliptic. Hence (Sph. Geom. 26) the Sun's polar distance CP is a minimum and therefore its decl. a maximum. Also r Q = 90 and CQ = tCrQ = i. Hence Sun's longitude = 90, B.A. = 90 - 6h., N. Decl. = /, (a maximum). From June 21 to September 23 the Sun's declination is still north, but is decreasing. On September 23rd the Sun has described 180, and is at the first point of Libra (=), the other extremity of the common diameter of the ecliptic and equator. This is the Autumnal Equinox, and we have Sun's long. = 180, R.A. = 180 = 12h., Decl. = 0. From Sept. 23 to Dec. 22 the Sun is south of the equator, and its south declination is increasing. On December 22ud the Sun has described 270 from T, and is at L (Fig. 16). This is called the Winter Solstice. We have t L = 90, and the triangle . RL has two right angles at R, L (Sph. Geom. 21). The Sun's polar dis- tance LP is a maximum (Sph. Geom. 26), and *R = L = 90, LR = / L^R = i. Hence Sun's longitude = 270, R.A. = 270 = 18h., S. Decl. = i, (a maximum). From December 22 to March 21 the Sun's declination is still south, but is decreasing. Finally, on March 21, when the Sun has performed a com- plete circuit of the ecliptic, we have . Sun's long. = 360, B.A. = 360 = 24h., Decl. = 0. The longitude and R.A. are again reckoned as zero, and they, together with the declination, undergo the same cycle of changes in the following year. 22 ASTEONOMT. 31. Sun's Variable Motion in R.A. We observe that the Sun's right ascension is equal to its longitude four times in the year, viz., at the two equinoxes and the two solstices. At other times this is not the case. For example, between the vernal equinox and summer solstice we have T-3f< T$, .'. Sun's E.A. < longitude. Hence, even if the Sun's motion in longitude be supposed uniform, its R.A. will not increase quite uniformly. There is a further cause of the want of uniformity, namely, that the Sun's motion in longitude is not quite uniform ; but this need not be considered in the present chapter. 32. Direct and Retrograde Motions. The direction of the Sun's annual revolution relative to the stars, i.e., motion from west through south to east, is called direct. The opposite direction, that of the diurnal apparent motions of the stars or revolution from east to west, is called retrograde. The revolutions of all bodies forming the solar system, with the exception of some comets and one or two small satellites, are direct. We shall see in Chapter III. that the apparent retrograde diurnal motion may be accounted for by the direct rotation pf the Earth about its polar axis, THE CELESTIAL SPHERE. 23 33. Equinoctial and Solstitial Points Colures. From 30 it appears that the Summer and Winter Solstices may be defined as the times of the year when the Sun attains its greatest north and south declinations respectively. The corresponding positions of the Sun in the ecliptic ((7, Z, Fig. 17) are called the Solstitial Points. In the same way the Equinoctial Points (T, ) are the positions of the Sun at the Vernal and Autumnal Equinoxes when its declination is zero. The declination circle PTP'^j passing through the equi- noctial points, is called the Equinoctial Colure. The declination circle PCP'L, passing through the solstitial points, is called the Solstitial Colure. The latter passes through the poles of the ecliptic (7T, K'). 34. To find the Sun's Right Ascension and Decli- nation. In the "Nautical Almanack,"* the Sun's R.A. and declination at noon are tabulated for every day of the year. Their hourly variations are also given in an adjoining column. To find their values at any time of the day, we only have to multiply the hourly variation by the number of hours that have elapsed since the preceding noon, and add to the value at that noon. EXAMPLE. Tfl find the Sun's R.A. and decl. on September 4, 1891 at 5h. 18m. in^gjs^ afternoon. We find from the Almanack for 1891 under Septembers : Sun's R.A. a*oon = lOli. 52m. 15s., hourly variation 9'04s. N. Decl. at noon = 7 12' 12" 55'4" (1) RA. at noon = lOh. 52m. 15s. Increase in 5h. = 9'04s. x 5 = 45*2 18m. = 27 .-. R.A. at 5h. 18m. - lOh. 53m. 3s. (2) From the Almanack, decl. is less on September 5, and is therefore decreasing. N. Decl at noon = 7 12' 12" Decrease in 6h. = 55'4" x 5 = 4' 37" \ To be 18m. - 17") subtracted. N. Decl. at 6h. 18m. = 7 T 18 ' * Also in " Whitaker's Almanack," which may be consulted with advantage. 24 ASTRONOMY. 35. Rough Determination of the Sun's R.A. "We can, without the "Nautical Almanack," find to within a degree or two, the Sun's E. A. on any given date, as follow^ : A year contains 365 days. In this period the Sun's E.A. increases by 360. Hence its average rate of increase is very nearly 30 per month, or 1 per day. Knowing the Sun's E.A. at the nearest equinox or solstice, we add 1 for every day later, or subtract 1 for every day before that epoch. If the E.A. is required in time, we allow for the increase at the rate of 2h. per month, or 4m. per day. EXAMPLES. 1. To find the Sun's R.A. on January 1st. On December 22nd the R.A. = 18h. Hence on January 1st, which is ten days later, the Sun's R.A. = 18h. 40m. 2. To find on what date the Sun's R.A. is lOh. 36m. On Sep- tember 23rd the R.A. is 12h. Also 12h.-10h. 36m. = 84m., and the R.A. increases Sim. in 21 days. Hence the required date is 21 days before September 23, i.e., September 2nd, 36. Solar Time. Apparent Noon is the time of the Sun's upper transit across the meridian, that is, in north latitudes, the time when the Sun souths. Apparent Mid- night is the time of the Sun's transit across the meridian below the pole (and usually below the horizon). An Apparent Solar Day is the interval between two consecutive apparent noons, or two consecutive midnights. Like the sidereal day, the solar day is divided into 24 hours, which are again divided into 60 minutes of 60 seconds each. For ordinary purposes the day is divided into two portions : the morning, lasting from midnight to noon ; the evening, from noon till midnight ; and in each portion times are reckoned from Oh. (usually called 12h.) up to 12h. For astronomical purposes we shall find it more convenient to measure the solar time by the number of solar hours that have elapsed since the preceding noon. Thus, 6.30 A.M. on January 2nd will be reckoned, astronomically, as 18h. 30m. on January 1st. On the other hand, 12.53 P.M. will be reckoned as Oh. 53m., being 53 minutes past noon. During a solar day the Sun's hour angle increases from to 360. It therefore increases at the rate of 15 per hour. Hence The apparent solar time = the Sun's hour angle expressed in time. THE CELESTIAL SPHERE. 25 At noon the Sun is on the meridian. The sidereal time, being the hour angle of T, is the same as the Sun's H.A., i.e., Sidereal time of apparent noon Sun's R. A. at noon. At any other time, the difference between the sidereal and solar times, being the difference between the hour angles of T and the Sun, is equal to the Sun's E.A. Hence, as in 25, we have (Sidereal time) (apparent solar time) = Sun's R.A. If a and a + x are the right ascensions of the Sun at two consecutive noons, then, since a whole day has elapsed between the transits, the total sidereal interval is 24h. +#, and exceeds a sidereal day by the amount x. But the interval is a solar day. Hence, the solar day is longer than the sidereal day, and the difference is equal to the sun's daily motion in R.A.* 37. Morning and Evening Stars. Sunrise and Sunset. "When a star rises shortly before the Sun, and in the same part of the horizon, it is called a Morning Star. Such a star is then only visible for a short time before sunrise. When a star sets shortly after the Sun, and in the same part of the horizon, it is called an Evening Star. It is then only visible just after sunset. It will be readily seen from a figure, that a star will be a morning star if its decl. is nearly the same as the Sun's, while its E/.A. is rather less. Similarly, a star will be an evening star if its decl. is nearly the same as the Sun's, but its RA. somewhat greater. Thus, as the Sun's R.A. increases, the stars which are evening stars will become too near the Sun to to be visible, and will subsequently reappear as morning stars. The times of sunrise and sunset are calculated in the manner described in 29. The hour angles of the Sun, when crossing the eastern and western horizons, determine the intervals of solar time between sunrise, apparent noon, and sunset. The two intervals are equal, if the Sun's decl. be supposed constant from sunrise to sunset a result very approximately true, since the change of decl. is always very small. * Owing to the sun's variable motion in R. A., the apparent solar day is not quite of constant length. In the present chapter, however, it may be regarded as approximately constant. 26 ASTRONOMY. 38. The Gnomon. Determination of Obliquity of Ecliptic. The Greek astronomers observed the Sun's motion by means of the Gnomon, an instrument consisting essentially of a vertical rod standing in the centre of a hori- zontal floor. The direction of the shadow cast by the Sun determined the Sun's azimuth, while the length of the shadow, divided by the height of the rod, gave the tangent of the Sun's zenith distance. To find the meridian line, a circle was described about the rod as centre, and the directions of the shadow were noted when its extremity just touched the circle before and after noon. The sun's Z.D.'s at these two instants being equal, their azimuths were evidently (Sph. Geom. 27) equal and opposite, and the bisector of the angle between the two directions was therefore the meridian line. The Sun's meridian zenith distances were then observed both at the summer solstice, when the Sun's IS", decl. is i and meridian Z.D. least, and at the winter solstice, when the Sun's S. decl. is i and meridian Z.D. greatest. Let these Z.D.'s be z l and s 2 respectively, and let I be the latitude of the place of observation. From 24, we readily see that 2 t = l-i, 2 2 = Z+t, /: *=*(.+*,), * = i(v-i);. thus determining both the latitude and the obliquity. 39. The Zodiac. The position of the ecliptic was defined by the ancients by means of the constellations of the Zodiac, which are twelve groups of stars, distributed at about equal distances round a belt or zone, and extending about 8 on each side of the ecliptic. The Sun and planets were observed to remain always within this belt. The vernal and autumnal equinoctial points were formerly situated in the constellations of Aries and Libra, whence they were called the First Point of Aries and the First Point of Libra. Their positions are very slowly varying, but the old names are still retained. Thus, the " First Point of Aries" is now situated in the constel- lation Pisces. The early astronomers probably determined the Sun's annual path by observing the morning and evening stars. After a year the same morning and evening stars would be observed, and it would be concluded that the Sun performed a complete revolution in the year. THE CELESTIAL SPHEEE. 27 40. Motion of the Moon. The Moon describes among the stars a great circle of the celestial sphere, inclined to the ecliptic at an angle of about 5. The motion is direct, and the period of a complete " sidereal " revolution is about 27 days. In this time the Moon's celestial longitude increases by 360. "When the Moon has the same longitude as the Sun, it is said to be New Moon, and the period between consecutive new Moons is called a Lunation. AVhen the Moon has described 360 from new Moon, it will again be at the same point among the stars ; but the Sun will have moved forward, so that the Moon will have a little further to go before it catches up the Sun again. Hence the lunation will be rather longer than the period of a sidereal revolution, being about 29 \ days. The Age of the Moon is the number of days which have elapsed since the preceding new Moon. Since the Moon separates 360 from the Sun in 29j days, it will separate at the rate of about 12, or more accurately 12-|- , per day, or 30' per hour. This enables us to calculate roughly the Moon's angular distance from the Sun, when the age of the Moon is given, and conversely, to determine the Moon's age when its angular distance is given. EXAMPLE. On September 23, 1891, the Moon is 20 days old. To find roughly its angular distance from the Sun and its longitude on that day. (1) In one day the Moon separates 12^- from the Sun; therefore, in 20 days it will have separated 20 x 121, or 244, and this is the required angular distance from the Sun. (2) On September 23 the Sun's longitude is 180 ; therefore the Moon's longitude is 180 + 244 = 424 = 360 + 64, or 64. This method only gives very rough results; for the Moon's motion is far from uniform, and the variations seem very irregular. Moreover, the plane of the Moon's orbit is not fixed, but its intersections with the ecliptic (called the Nodes) have a retrograde motion of 19 per year. Hence, for rough pur- poses, it is better to neglect the small inclination of the Moon's orbit, and to consider the Moon in the ecliptic. If greater accuracy be required, the Moon's decl. and R.A. may be found from the Nautical Almanack. 28 ASTRONOMY. 41. Astronomical Diagrams and Practical Applica- tions. We can now solve many problems connected with the motion of the celestial bodies, such as determining the direc- tion in which a given star will be seen from a given place, at a given time, on a given date, or finding the time of day at which a given star souths at a given time of year. "We have, on the celestial sphere, certain circles, such as the meridian, horizon, and prime vertical, also certain points, such as the zenith and cardinal points, whose positions relative to terrestrial objects always remain the same. Besides these, we have the poles and equator, which remain fixed, with reference loth to terrestrial objects and to the fixed stars. "We have also certain points, such as the equinoctial points, and certain circles, such as the ecliptic, which partake of the diurnal motion of the stars, performing a retrograde revolution about the pole once in a sidereal day. Lastly, we have the Sun, which moves in the ecliptic, performing one retrograde revolution relative to the meridian in a solar day, or one direct revolution relative to the stars in a year, and whose hour angle measures solar time. In drawing a diagram of the celestial sphere, the positions of the meridian, horizon, zenith, and cardinal points should first be represented, usually in the positions shown in Pig. 18. Knowing the latitude nP of the place, we find the pole P. The points Q, ft, where the equator cuts the meri- dian, are found by making PQ = PR = 90 ; and the points Q, Ii, with E, W, enable us to draw the equator. We now have to find the equinoctial points. How to do this depends on the data of the problem. Thus we may have given (i.) The sidereal time ; (ii.) The hour angle of a star of known E.A. and decl ; (iii.) The time of (solar) day and time of year. In case (i.), the sidereal time multiplied by 15 gives, in degrees, the hour angle (Qf) of the first point of Aries. Measuring this angle from the meridian westwards, we find Aries, and take Libra opposite to it. Any star of known decl. and R.A. can be now found by taking on the equator = star's R.A., and taking on MP, MX = star's decl. THE CELESTIAL SPHERE. 29 The ecliptic may be drawn passing through Aries and Libra, and inclined to the equator at an angle of about 23 \ (just over right angle). As we go round from west to east, or in the direct sense, the ecliptic passes from south to north of the equator at Aries ; this shows on which side to represent the ecliptic. Knowing the time of year, we now find the Sun (roughly) by supposing it to travel to or from the nearest equinox or solstice about 1 per day from west to east. Finally, if the Moon's age be given, we find the Moon by measuring 12-i- per day, or 30' per hour eastwards from the Sun. P' FIG. 18. In case (ii.), we either know the hour angle, QMoi QPMof. a known star (#), or, what is the same thing, the sidereal interval since its transit ; or, in particular, it is given that the star is on the meridian. Each of these data determines J/~, the foot of the star's declination circle. From M we measure westwards equal to the star's R.A. This finds Aries. 80 ASTRONOMY. fn case (iii-)> the solar time multiplied by 15 gives the- Hun's hour angle QPS in degrees. From the time of year we can find the Sun's R.A., TJPS. From these we find Q,PT and obtain the position of Aries just as in case (ii.) It will be convenient to remember that azimuth and hour angle are measured from the meridian westwards, while right ascension and celestial longitude are measured from the first point of Aries eastwards. Thus, since the Sun's diurnal motion is retrograde, and its annual motion direct, the Sun's azimuth, hour angle, R.A., and longitude are all increasing. Most problems of this class depend for their solution chiefly on the consideration of arcs measured along the equator, or (what amounts to the same) angles measured at the pole. In another class of problems depending on the relation be- tween the latitude, a star's decl. and meridian altitude ( 24), we have to deal with arcs measured along the meridian. These two classes include nearly all problems on the celestial sphere which do not require spherical trigonometry. EXAMPLES. 1. To represent, in a diagram, the positions of the Sun and Moon, and the star Herculis as seen by an observer in London on Aug. 19, 1891, at 8 p.m., the following data being given : Latitude of London- = 51, Moon's age at noon on Aug. 19 = 14 days 19 hours, Moon's latitude = 2 S., K.A. of (Herculia = 16h. 37m., decl. = 31 48' N. The construction must be performed in the following order : (i.) Draw the observer's celestial sphere, putting in the meridian, horizon, zenith Z, and four cardinal points n, E, s, W. (ii.) Indicate the position of the pole and equator. The observer' s- latitude is 51. Make, therefore, nP = 51. P will be the pole. Take PQ = PR = 90, and thus draw the equator, QERW. (Hi.) Find the declination circle passing through the Sun. The- time of day is 8 p.m. Therefore the Sun's hour angle is 8 x 15, or 120. On the equator measure QK = 120 westwards from the- meridian. Then the Sun Q will lie on the declination circle PK. Since QW = 90, we may find K by taking WK = 30 = $ WR. (iv.) Find the first points of Aries and Libra. The date of obser- vation is August 19. Now, on September 23 the Sun is at =2=. Also- from August 19 to September 23 is 1 month 4 days. In this- interval the Sun travels about 34 from west to east. Hence the Sun is 34 west of rO=. And we must measure K* = 34 eastwards^ from 8, and thus find z. The first point of Aries ( T ) is the opposite point on the equator.. THE CELESTIAL SPHERE. 31 (v.) We may now draw the ecliptic Cri^= passing through the first points of Aries and Libra, and inclined to the equator at an angle of about 23 (i.e., slightly over of a right angle). The Sun is above the equator on August 19; hence the ecliptic cuts PK above K. This shows on which side of the equator the ecliptic is to be -drawn ; we might otherwise settle this point by remembering that the ecliptic rises above the equator to the east of T . The intersection of the ecliptic with PE determines Q, the position of the Sun. FIG. 19. ascenfion is 16h. 37m., in time, = 249 15' in angular measure. On the equator measure off T M = 249 15' in the direction west to east (i.e., the direction of direct motion) from T ; we must, therefore, take ^=M = 69 15'. On the declination circle HP, measure off MX = 31 48' towards P. Then x is the required position of Herculis. (vii.) Find the Moon. At 8 p.m. the Moon's age is 14d. 19h + 8h. = 15d. 3h. Hence, the Moon has separate/! from the Sun by about 185 in the direction west to east. Measure off }) = 185 from west to east, and put in }) about 2 below the ecliptic. The Moon's position is thus found. 32 ASTRONOMY. a/- 2. To find (roughly) at what time of year the Star o Cygni (R.A. = 20h. 38m., clecl. = 44 53' N.) souths at 7 p.m. Let o be the position of the star on the meridian (Fig 20). At 7 p.m. the Sun's western hour angle (QS or QPS) = 7h. = 105. Also TEQ, the Star's R.A. = 20h. 38m. Hence rRS, the Sun's R.A. = 20h. 38m. - 7h. = 13h. 38m. ; or, in angular measure, Sun's R.A. = 204 30'. Now, on September 23, Sun's R.A. = 180, and it increases at about 1 per day. Hence the Sun's R.A. will be 204 about 24 days later, i.e., about October 17th. 3. At noon on the longest day (June 21) a vertical rod casts on a horizontal plane a shadow whose length is equal p IG 20 to the height of the rod. To find the latitude of the place and the Sun's altitude at midnight. FIG. 21. From the data, the Sun's Z.D. at noon, Z, evidently = 45. Also, if QR be the equator, 0Q = Sun's decl. = i = 23 27' (approx.); .-. latitude of place = ZQ = 45 + 23 27' = 68 27'. If ' be the Sun's position at midnight, P0' = PQ = 90-2.327' = G6 33'. But Pn = lat. = 68 27'. ... Q' w = 68 27' -66 33' = 1 54'; and the Sun will be above the horizon at an alt. of 1 54' at midnight. THE CELESTIAL SPHERE. EXAMPLES. I. 1. Why are the following definitions alone insufficient? Tlie zenith and nadir are the poles of the horizon. The horizon is the great circle of the celestial sphere whose plane is perpendicular to the line joining the zenith and nadir. 2. The R.A. of an equatorial star is 270 ; determine approximately the times at which this star rises and sets on the 21st June. In what quarter of the heavens should we look for the star at mid- night ? 3. Explain how to determine the position of the ecliptic relatively - to an observer at a given hour on a given day. Indicate the position . of the ecliptic relatively to an observer at Cambridge at 10 p.m. at the autumnal equinox. (Lat. of Cambridge = 52 12' 51'6".) VV! i 4. Prove geometrically that the least of the angles subtended at an observer by a given star and different points of the horizon that which measures the star's altitude. 5. Show that in latitude 52 13' N. no circumpolar star when southing can be within 75 34' of the horizon. C. Represent in a figure the position of the ecliptic at sunrise on March 21st as seen by an observer in latitude 45. Also in lati- tude 67. , 7. If the ecliptic were visible in the first part of the preceding question, describe the variations which would take place during the day in the positions of its points of intersection with the horizon. 8. Determine when the star whose declination is 30" N. and whose . E.A. is 356 will cross the meridian at midnight. 9. The declination and R.A. of a given star are 22 N. and 6h. 20m. respectively. At what period of the year will it be (i.) a morning, (ii.) an evening star ? In what part of the sky would you then look for it ? 10. Find the Sun's R.A. (roughly) on January 25th, and thus de- termine about whatxtime Aldebaran (R.A. 4h. 29m.) will cross the meridian that night. 11. Where and at what time of the year would you look for Fomalhaut ? (R.A. 22h. 51m., decl. 30. 16' S.) 12. At the summer solstice the meridian altitude of the Sun is 75. What is the latitude of the place ? What will be the meridian altitude of the Sun at the equinoxes and at the winter solstice ? ~ 34 ASTRONOMY. EXAMINATION PAPER. I. 1. Explain how the directions of stars can be represented by means of points on a sphere. Explain why the configurations of the constellations do not depend on the position of the observer, and why the angular distance of two different bodies on the celestial sphere gives no idea of the actual distance between them. 2. Define the terms horizon, meridian, zenith, nadir, equator, ecliptic, vertical, prime vertical, and represent their positions in a figure. 3. Explain the use of coordinates in fixing the position of a body on the celestial sphere, and define the terms altitude, azimuth t polar distance, hour angle, right ascension, declination, longitude, latitude. Which of these coordinates alwa3 T s remain constant for the same star ? 4. Define the obliquity of the ecliptic and the latitude of the observer. Give (roughly) the value of the obliquity, and of the latitude of London. Indicate in a diagram of the celestial sphere twelve different arcs and angles which are equal to the latitude of the observer. 5. What is meant by a sidereal day and a sidereal hour ? How could you find the length of a sidereal day without using a tele- scope ? Why is sidereal time of such great use in connection with astronomical observations ? 6. Show that the declination and right ascension of a celestial body can be determined by meridian observations alone. 7. What is meant by a circumpolar star ? What is the limit of declination for stars which are circumpolar in latitude 60 N. ? Indicate in a diagram the belt of the celestial sphere containing all the stars which rise and set. 8. Define the terms year, equinoxes, solstices, equinoctial and solstitial points, equinoctial and solstitial colures. What are the dates of the equinoxes and solstices, and what are the corresponding values of the Sun's declination, longitude, and right ascension? Find the Sun's greatest and least meridian altitudes at London. 9. Why is it that the interval between two transits of the Sun or Moon is rather greater than a sidereal day ? Show how the Sun's R.A. may be found (roughly) on any given date, and find it on July 2nd, expressed in hours, minutes, and seconds. 10. Indicate (roughly) in a diagram the positions of the following stars as seen in latitude 51 on July 2nd at 10 p.m, : Capella (R.A. 5h. 8m. 38s., decl. 45 53' 10" N.), a Lyras (R.A. 18h. 33m. 14s., decl. 38 40' 57" N.), a Scorpii (R.A. 16h. 22m. 43s., decl. 26 11' 22" S.), a Ursse Majoris (R.A. lOh. 57m. Os., dec!. 62 20' 22" N.) CHAPTER II. THE OBSERYATOHY. SECTION I. Instruments adapted for Meridian Observations. 42. One of the most important problems of practical astro- nomy is to determine, by observation, the right ascension and declination of a celestial body. We have seen in Chapter I. that these coordinates not only suffice to fix the position of a star relative to neighbouring stars, but they also enable us to find the direction in which the star may be seen from a given place at a given time of day on a given date (41). More- over, it is evident that by determining every day the decli- nation and right ascension of the Sun, the Moon, or a planet, the paths of these bodies relative to the stars can be mapped out on the celestial sphere and their motions investigated. In Section II. of the preceding chapter we showed that the right ascension and declination of a star can be deter- mined by observations made when the star is on the meridian. We proved the following results : The star's R.A. measured in time is equal to the time of transit indicated by a sidereal clock ( 24). The star's north decl. d can be found from z its meridian zenith distance, and I the latitude of the observatory by the iormula d = l+z, where if the decl. is south d is negative, and if the star tran- sits south of the zenith z is negative (24). Lastly, I can be found by observing the altitudes of a circumpolar star at its two culminations, and is therefore known ( 28). Hence the most essential requisites of an observatory must include (i.) a clock to measure sidereal time, (ii.) a telescope so fitted as to be always pointed in the meridian, provided with graduated circles to measure its inclination to the ver- tical, and with certain marks to fix the position of a star in its field of view. 36 ASTRONOMY. 43. The Astronomical Clock is a clock regulated to indicate sidereal time. It should be set to mark Oh. Om. Os. at the time when the first point of Aries crosses the meridian. It will therefore gain about 4 minutes per day on an ordinary clock, or a whole day in the course of a year ( 22, 36). The clock is provided with a seconds hand, and the pendulum beats once every second, produc- ing audible "ticks"; hence an observer can estimate times by counting the ticks, whilst he is watching a star through a telescope. The pendulum is a compensating pendu- lum, or one whose period of oscillation is un- affected by changes of temperature. The form most commonly used is Graham's Mercurial Pendulum, in which the bob carries two glass cylinders containing mercury (Fig. 22). If the temperature be raised, the effect of. the increase in length of the pendulum rod is compensated for by the mercury expanding and rising in the cylinders. The same result is also effected in Harrison's Gridiron Pendulum, described in Wallace Stewart's Text-Boole of Heat, page 37. The clock is sometimes regulated by placing small shot in a cup attached to the pendulum. FIG. 23. THE OBSERVATORY. 37 44. The Astronomical Telescope (Fig. 23) consists essentially of two convex lenses, or systems of lenses, and 0', fixed at opposite ends of a metal tube, and called the object-glass and eye-piece respectively. The former lens receives the rays of light from the stars or other distant objects, and forms an inverted " image " (al) of the objects. The centre of the round object-glass is. called its " optical centre," and the image is produced as follows: Let AAA be a pencil of rays from a distant star. By traversing the object-glass these rays are refracted or bent towards the middle ray A 0, which alone is unchanged in direction. The rays all converge to a common point or "focus'' at a point a in A produced, and, if received by the eye after passing #, they would appear to emanate from a luminous point or " image " of the star at a. Similarly, the rays BBB, coming from another distant star, will converge to a focus at a point b in BO produced, and will give the effect of an " image" of the star at b. All these images (a, b) lie in a certain plane FN, called the focal plane of the object-glass, and they form a kind of picture or image of such stars as are in the field of view. The eye-piece 0' acts as a kind of magnifying glass, and enlarges the image ab just as if it were a small object placed in the focal plane FN. The figure shows how a second image A'B' is formed by the direction of the pencils of light after refraction through (/. This is the final image seen on looking through the telescope. The eye must be placed in the plane EE, so as to receive the pencils from A', B'. If, now, a framework of fine wires or spider's threads (Fig. 25) be stretched across the tube in the focal plane FNj these wires, together with the image (#J), will be equally magnified by the eye-piece. They will thus be seen in focus simultaneously with the stars, and the field of view will appear crossed by a series of perfectly distinct lines, which will enable us to fix any star's position, and thus determine its exact direction in space. Suppose, for example, that we have two wires crossing one another at the point F', and the telescope is so adjusted that the image of a star coincides with F', then we know that the star lies in the line joining F' to the optical centre of the object-glass. 00 ASTRONOMY. 45. The Transit Circle (Figs. 24, 26) is the instrument used for determining both right ascension and declination. It consists of a telescope, ST, attached perpendicularly to a light, rigid axis, WPPE, hollow in the interior. The ex- tremities of this axis are made in the form of cylindrical pivots, E, W, which are capable of revolving freely in two fixed forks, called Y's, from their shape. These Y's rest on piers of solid stone, built on the firmest possible foundations, and they are carefully fixed, so as always to keep the axis exactly hori- zontal and pointing due east and west. FIG. 24. In order to dimini?0i the effect of friction in wearing away the pivots, the axis is also partially supported at P, P upon friction rollers (not represented in the figure) attached to a THE OBSERVATORY. 3<> system of levers ( Q, Q) and counterpoises (R, R) placed within the piers. These support about four-fifths of the weight of the telescope, leaving sufficient pressure on the Y's to ensure- their keeping the axis fixed. Within the telescope tube, in the focal plane of the object- glass ( 44), is fixed a framework of cross wires, presenting^ the appearance shown in Fig. 25. Five, or sometimes seven, wires appear vertical, and two appear horizontal. Of the latter, one bisects the field of view ; the other is movable up and down by means of a screw, whose head is divided by graduation marks which indicate the position of the wire. The line joining the optical centre of the object-glass to the point of intersection of the middle vertical wire with the- fixed horizontal wire is called the Line of Colliinatiou. The wires should be so adjusted that the line of colliination is per- pendicular to the axis about which the telescope turns. For this purpose the framework carrying the wires can be moved horizontally, by means of a screw, into the right position. If the Y's have been accu- rately fixed, then, as the telescope turns, the line of collimation will always lie in the plane of the meridian. Hence, when a star transits we shall, on looking through the telescope, see it pass across the middle vertical, wire. Attached to the axis of the telescope, and turning with it, are two wheels, or graduated circles, GH, having their circumferences divided into degrees, and further subdivided by fine lines at (usually) intervals of 5'. By means of these graduations the inclination of the line of collimation to the vertical is read off by aid of sevi ral fixed compound micro- scopes, A, /, JB, pointed towards the circle. One of these microscopes (7), called the Pointer or Index, is of low magnifying power, and shows by inspection the number of degrees and subdivisions in the mark of the circle, which is opposite a wire bisecting its field of view. The pointer should read zero when the line of collimation points to the zenith, and the graduations increase as the telescope is. turned northwards. 40 FIG. 26, 46. Beading Microscopes. In addition to the pointer there are four (sometimes six) other microscopes, called Reading Microscopes, arranged symmetrically round each circle, as at ABCD (Fig. 26). These serve to determine the number of minutes and seconds in the inclination of the tele- scope, by means of the following arrangement. Inside the tube of each microscope in the focal plane of its object- glass* is fixed a graduated scale NL (Fig. 27) in the form of a strip of metal with fine teeth or notches. This scale, and the image of the telescope circle, formed by the object-glass of the microscope, are simultaneously viewed by the eye-glass, and present the appearance shown in Fig. 27. FIG. 27. A small hole O marks the middle notch, and 5 notches correspond to a division of the telescope circle, hence the number of notches from the hole to the next division of the circle gives the number of minutes to be added to the pointer reading. * A compound microscope, like a telescope, consists of an object- glass, which forms an image of an object, and an eye-piece which enlarges this image. A scale or wires fixed in the plane of the the image will, therefore, be seen in distinct focus, like the wires in the telescope. THE OBSERVATORY. 41 To read off the number of seconds, a pair of parallel wires, Sit, are attached to a framework, and can be moved across the field of view by means of a screw. One whole turn takes the wires from one notch of the metal scale to the next, i.e., over a space representing 1' on the telescope circle ; and the head of the screw is divided into 60 parts, each, therefore, representing V. The wires are adjusted so that the graduation on the telescope circle appears midway between them, and the reading of the screw-head then gives the number of seconds. With practice, tenths of a second can be estimated. The four microscopes of one of the circles are all read, and the best result is obtained by taking the mean of the readings. 47. Clamp and Tangent Screw. When it is required to rotate the telescope of the transit circle very slowly, this is done by means of the bar represented at LK in Fig. 24. The telescope axis may be firmly clamped to this bar by means of a clamp (not represented in the figure), which grips the rim of one of the circles as in a vice. When this has been done, the bar JTZ, and with it the telescope, may be slowly turned by means of a horizontal screw at Z, called the Tangent Screw, and provided with a long handle attached to it by a " universal joint." This handle is held by the observer, and he can thus turn the tangent screw without ceasing to watch the stars. 48. Arrangements for Illumination. As most obser- vations are conducted at night, the wires in the telescope and the graduations of the circles must be illuminated. This is done by a lamp placed exactly in front of one of the pivots, the light from which is concentrated by means of a bull's-eye lens in front and a mirror behind. Part of the rays are reflected, by a complicated arrangement of mirrors and prisms, so as to illuminate the parts of the graduated circle viewed by the microscopes. The rest of the light passes through a plate of red glass down the hollow axis to a ring- shaped mirror, whence it is reflected up to the wires ; thus the wires appear as dark lines on a dull red ground. There is also another arrangement for illuminating the wires from in front, if desired, so that they appear bright on a dark ground 42 ASTRONOMY. 49. Taking a Transit. Eye and Ear Method. If a star is to be observed with the transit circle, its R.A. and decl. must have been roughly estimated beforehand ; hence, its meridian Z.D. [= (star's decl.) (observer's lat.)} is known roughly. Before the star is expected to- cross the meridian, the telescope is turned by hand until the pointer indicates this roughly determined Z.D. ; this adjustment is sufficiently accurate to ensure the- star traversing the field of view. The telescope is then clamped ( 47). The observer now " takes a second" from the astronomical clock, i.e., he observes and writes down the- hour and minute, observes the second, and begins counting seconds by the clock's ticks. Thus, if he sees the time to be- llh. 23ni. 29s., he writes down "llh. 23m.," and at the- subsequent ticks he counts " 303132 33 " and so on ; in this way he knows, during the rest of the observation, t he- exact time at every clock -beat without looking at the clock. The star soon approaches the first vertical wire, and passes it, usually between two successive ticks. With practice, the observer is able to estimate fractions of a second as follows : Suppose the star crosses the wire between the 34th and 35th tick. The positions of the star are noticed at tick 34 and at tick 35, and by judging the ratio of their distances from the wire on the two sides, the observer estimates the time of crossing the wire by a simple proportion, and writes down, this time, say 34'6. The estimate is difficult to make,, because the two positions of the star are not visible simulta- neously, and the star does not stop at them, but moves continuously; hence to estimate tenths of seconds (as is usually done) requires much training and practice. Moreover, the observer must not lose count of the ticks of' the clock, for when he has written down the instant of transit. over the first wire the star will be nearing the second wire.* The time of transit over the second vertical wire is now estimated in the same way, and the process repeated at each wire. The average of the times of crossing the five or seven wires is taken as the time of transit ; in this way, * In most instruments the wires are placed at such a distance- that a star in the equator takes about 13 seconds from one wire- to the next. THE OBSERVATORY. 4$ the effect of small errors of observation will be much smaller than if the transit over one wire only were observed. This method of taking the time of transit is called the " Eye and Ear Method." While observing the transit, the observer turns the tele- scope by means of the tangent screw, until the horizontal wire bisects the image of the star ; during the rest of the observation the star will appear to run along the horizontal wire. After the observation, one of the circles is read by the pointer and the four microscopes. If the circle reads 0' 0", when the line of collirnation points to the zenith, the reading for the star will determine its meridian Z.D., in other cases we must subtract the zenith reading. Prom the meridian Z.D. the declination can be found. 50. The Chronograph. To obviate the difficulty of observing tr?nsits by the eye and ear method, an instrument called the Chronograph is now frequently used. A cylin- drical barrel, covered with prepared paper, is made to turn slowly and uniformly by clockwork about an axle, on which a screw is cut. In this way the barrel is made to move forward in the direction of its axis, about one-tenth of an inch in every revolution. The observer is furnished with a key or button, which is in electric communication with a pen or marker. At the instant when the star crosses one of the vertical wires, the observer depresses the key, and a mark is made upon the paper of the barrel. The astronomical clock, also, has electric communication with the marker, and marks the paper once every second, the beginning of a new minute being indicated, in some instruments, by the omission of the mark, in others, by a double mark. In this way, a record is made of the times of transit over the wires, the marks being arranged in a spiral, owing to the forward motion of the barrel. The distance of the beginning of any transit-mark from the previous second-mark can be measured at leisure with very great accuracy, and the time of transit may thus be readily calculated. Indeed, there is no difficulty in recording, by this method, the transits of two, or even more, near stars which are simultaneously in the field of view of the telescope, for the transit-marks of the different stars can be readily distinguished from one another afterwards. ASTRON. E 44 ASTRONOMY. 51. Corrections. After the transit of a star has been observed, certain corrections have to be allowed for in practice before its true B. A. and decl. are obtained. These corrections, which depend on errors of observation, may be conveniently classified as follows : (a) Corrections required for the Right Ascension : 1 . Error and rate of the astronomical clock. 2. Personal equation of the observer. 3. Errors of adjustment .of the transit circle, including (a) Collimation error. (5) Level error. (c) Deviation error. (d) Irregularities in the form of the pivots. (e) Corrections for the " vertically" and " wire intervals." (5) Corrections required in finding the Declination : 1. Beading for zenith point, or for the nadir, hori- zontal or polar point. 2. Errors of imperfect centering of the circles. 3. Errors of graduation. 4. Errors of " runs " in the reading microscopes. Besides these corrections, which we now proceed to de- scribe, there are others of a physical nature, such as refraction, parallax, aberration, the description of which will be given later. A correction is always regarded as positive when it must be added to the ol served value of a quantity in order to get the true value, negative if it has to be subtracted. (a) CORRECTIONS REQUIRED FOR THE RIGHT ASCENSION. . 52. Clock Error and Hate. A good astronomical clock can generally be regulated so as not to gain or lose more than about 2s. in a sidereal day. But to estimate times with greater accuracy, it is necessary to apply a correction to the time indicated, owing to the clock being either fast or slow. The Error of a clock is the amount by which the clock is sloiv when it indicates Oh. Om. Os. Thus, the error must be added to the indicated time in order to obtain the correct time. If the clock is fast, its error is negative. The Rate of the clock is the increase of error during 24 hours. It is, therefore, the amount which the clock loses in the 24 hours. If the clock gains, the rate is negative. THE OBSERVATORY. 45 The rate of a clock is said to be uniform or constant when the clock loses equal amounts in equal intervals of time. In a good astronomical clock, the rate should remain uniform for several weeks. 53. Correction for Error and Hate. If the error of a clock and its rate (supposed uniform) are known, the correct time can be readily found from the time shown by the clock. The method will be made clear by the following example : EXAMPLE. If the error of an astronomical clock be 2'52s., and its rate be O44s., to find to the nearest hundreth of a second the correct time of a transit, the observed time bythe clock being 19h.23m.25'44s. Here in 24h. the clock loses 0'44s. .-. in Ih. it loses -^ x 0'44s. = 0'0183s. Hence, loss in 19h. = 0'0183s. x 19 = 0'348s., and loss in 23m. = O'OOTs. At Oh. Om. Os. the clock error is = 2'52s. ; /. at 19h. 23m. 25'44s., clock is too slow by 2'52s. +0'355s. = 2'88s., /. the correct time = 19h. 23m. 25'44s. + 2'88s. = 19h. 23m. 28-32s. 54. Determination of Error and Rate of Clock. The clock error is found by observing the transit of a known star, i.e., a star whose R.A. and decl. are known. If the clock were correct, the time of transit (when cor- rected for all other errors) would be equal to the star's R.A. (see 24). If this is not the case, we have evidently (Clock error) = (Star's R.A.) (observed time of transit). This determines the clock error at the time of transit. To find the rate, the transits of the same star are observed on two consecutive nights. Let t and t x be the observed times of transit ; then x is the amount the clock has lost in 24 hours, i.e., the rate of the clock. Therefore (Bate of Clock) = (observed time of Isb transit) (observed time of 2nd transit). Having found the rate of the clock and its error at the time of transit, the error at Oh. Om. Os. may be found by subtracting the loss between Oh. Om. Os. and the transit. Stars used in finding clock error arc known as "Clock Stars." 46 ASTBONOMY. 55. Personal Equation is the error made by any par- ticular observer in estimating the time of a transit. Of two observers, one may habitually estimate the transit too soon, another may estimate it too late, but experience shows that the error made by each observer in taking times of transit by the same method is approximately constant. If all observations are made, by the same individual there will be no need to take account of personal equation, because the error made in taking a transit will be compensated by the error made in observing the clock stars to set the clock. If the two operations are performed by different observers, we must allow for the difference of their personal equations. Personal equation may be measured by an apparatus for observing the transit of a fictitious star, . 360 , 360 2. To find the number of feet in one fathom. By Ex. 1, 60 nautical miles = 69 ordinary miles j i.e., 60,000 fathoms = 69jt x 5280 feet ; /. 1 fathom = 69 * x 528 feet = 6'086 feet. 3. To express a metre in terms of a yard. By definition, 40,000,000 metres = Earth's circumference =24,900 miles ; .-. 1 metre = ^^S^Ur y ards = 1 '0956 yards. Tttfe EARTH. 69 95. Terrestrial Longitude. The Longitude of a place on the Earth is the angle between the terrestrial meridian through that place, and a certain meridian fixed on the Earth, and called the Prime Meridian. Thus, in Eig. 36, if PEP' represents the prime meridian, the longitude of any place q is measured by the angle RPq. The longitude of q is also measured by R Q, the arc of the equator intercepted between the meridian of the place and the prime meridian. FIG. 36. Since the latitude of q is measured by the arc Qq, we see that latitude and longitude are two coordinates denning the position of a place on the Earth just as decl. and 11. A., or celestial latitude and longitude define the position of a star.* The choice of a prime meridian is purely a matter of con- venience. The meridian of Greenwich Observatory is univer- sally adopted by English-speaking nations. The Erench use the meridian of Paris, and the University of Bolognahas recently proposed the meridian of Jerusalem as the universal prime me- ridian. Longitudes are measured both eastward and westward from the prime meridian, from to 180, not from to 360. *Note, however, that terrestrial latitude and longitude, being referred to the equator, correspond more nearly to declination and right ascension than to celestial latitude and longitude. ?0 ASTEOIfOMiT. 96. Phenomena depending on Change of Longitude. (i.) Let q, r (Fig. 37) be two stations in the same latitude, and let the longitude of q be L west of r, so that Z rPq = L. As the Earth revolves about its axis at the rate of 360 per sidereal day, or 15 per sidereal hour, the points q, r will be carried forward in the direction of the arrow. After an interval of -^ L sidereal hours, q will have revolved through Z and will arrive at the position originally occupied by r. Hence the appearance of the heavens to an observer at q will be same as it was, -^ L sidereal hours previously, to an observer at r. The stars will rise, south, and set -^ L hours earlier at r than at q. (ii.) If Aj B be two places in different latitudes, whose difference of longitude is Z, the transits of a star at A and B will take place when the meridian planes PAP' and PBP' (which are evidently also the planes of the celestial meridians of A, B respectively), pass through the direction of the star. Hence, in this case also, the transits will occur J-g- L hours earlier at B than at A. Now an observer at B will set his sidereal clock to indicate Oh. Om. Os. when T crosses the meridian of B. When T transits at A, the clock at B will mark -fa L h., but an observer at A will then set his clock at Oh. Om. Os. Hence, if the two clocks be brought together and com] ared, the clock from B will be -^ L h. faster than the clock from A. This fact may be expressed briefly by saying that the " local " sidereal time at B is T y h. faster than the local sidereal time at A. Since the Earth makes one revolution relative to the Sun in a solar day, in like manner the local solar time at B will be -jig-Z solar hours faster than the local solar time at A. Therefore, whether the local times be sidereal or solar, we have Longitude of A west of B = long, of B east of A = 15 {(local time at .B) (local time at A)}. In particular, Long, west of Greenwich = 15 {(Greenwich time) (local time)} = 15 (Greenwich time of local noon). THE EARTH. 71 97. To find the length of any arc of a given parallel of latitude, having given the difference of longitude of its extremities. [A small circle of the Earth parallel to the equator is called a Parallel of Latitude.] Let qr be the given arc of the parallel hqrk, I its latitude, and let qPr, the difference of longitudes of q and r, be = Z. Let a be the radius of the Earth. If the meridians of q, r meet the terrestrial equator in Q, R, we have, by Sph. Geom. (17), arc qr = arc QR X sin Pq = arc QR x cos I. But arc QR : circumference of Earth = Z : 360; .-. arc QR = 27T0Z/360 = 180 /. arc qr = iraL cos I 180 COROLLARY. Since V of arc of the equator measures a geographical mile, it follows that In latitude ?, the arc of a parallel corresponding to 1' difference of longitude is cos I geographical miles. 72 ASTRONOMY. 98. Changes of Latitude and Longitude due to a Ship's Motion. Suppose a ship, in latitude I, to sail m nautical miles in a direction A degrees west of north. If m is small, we may easily see (by drawing a diagram) that the ship would arrive at the same place by sailing m cos .4 nautical miles due north, and then sailing msinA nautical miles due west. Hence, The ship's latitude will increase by m cos^4 minutes ( 92). Its W. long, will increase by m sin^ sec I minutes ( 97, cor.). NOTE. The shortest distance between two points on a sphere is along a great circle. Hence, the shortest distance between two places in the same latitude is less than the arc of the parallel joining them (except at the equator). But the difference is imperceptible when the arc is small. 99. To explain the Gain or Loss of a Day in going round the World. If a traveller, starting from a place A, go round the world eastward, and if, during the voyage, the Earth revolves n times relative to the Sun, the traveller will have performed one more revolution relative to the Earth in the same direction, and therefore n + 1 revolutions relative to the Sun. Hence, to a person remaining at -df, the voyage will appear to have taken n days, while to the traveller, n + 1 days will appear to have elapsed in other words, the traveller will, apparently, have " gained a day." But, as he goes eastward, he will find the local time con- tinually getting faster, and he will have to move the hands of his watch forward Ih. for every 15, or 4m. for every 1 of longitude. Thus, by the end of the voyage he will have put his watch forward through 24h., and the day apparently gained will be made up of the times apparently lost every time the watch is put forward to local time. Similarly, a traveller going round the world westward, and starting and arriving back simultaneously with the first traveller, will have made n 1 revolutions relative to the Sun, instead of n. Hence, the journey will appear to have taken n 1 days, and he will apparently have lost a day. But, during the journey, he will have been continually moving the hands of his watch backwards, so that the 24h. apparently lost will be made up of the times apparently gained each time the watch is put back to local time. THE EAETH. 73 SECTION II. Dip of the Horizon 100. Definitions. Let be an observer situated above the surface of the land or sea. Draw OT, OT tangents to the surface. Then it is evident, from the figure, that only those portions of the Earth's surface will be visible whose distance from the observer is less than the length of the tangents OT, OT. FIG. 38. The boundary of the portion of the Earth's surface visible from any point is called the Offing or Visible Horizon. Hence, if OA CB be the Earth's diameter through 0, and the Earth be supposed spherical, the offing at is the small circle TtT, formed by the revolution of T about OB, and having for its pole the point A vertically underneath 0. If, however, the Earth be not supposed spherical, the form of the offing will, in general, be more or less oval, instead of circular. Conversely, since it is observed that the " offing " at sea is very approximately circular, whatever be the position of the observer, it may be inferred that the Earth is approxi- mately spherical. The Dip of the Horizon at is the inclination to the horizontal plane of a tangent from to the Earth's surface. Hence, if HOH' be drawn horizontally (i.e., perpendicular to OC\ the dip of the horizon will be the angle HOT. 74 ASTEONOMY. 101. To determine the Distance and Dip of the Visible Horizon at a given height above the Earth. Let h = A = given height of observer ; a = CA = Earth's radius; d OT = required distance of horizon ; D = L HOT = required dip expressed in circular D" the number of seconds in the dip D. (i.) By Euclid III. 36, OT 2 = OA . OB This determines d accurately. But in practical applications h is always very small compared with 2a ; therefore A 2 may be neglected in comparison with 2ah, and we have the approxi- mate formula, rf 2 = 2ah .*. d = */ (2a7i). (ii.) Since CTO is a right angle, .-. z OCT= complement of L COT '= L TOR= D. Therefore, D being expressed in circular measure, we hav = arc **'. (Sph. Geom., 17.) sin xP cos d But treating the small triangle x'xH&s plane (Sph. Geom., 24), and remembering that Z Pxx = 90, we have cos nxP ' .. t = If' sec d . sec nxP. lo Evidently the acceleration at rising = retardation at setting. COROLLARY 1. To an observer at the Equator, P coincides with w, .'. Z nxP = 0, .-. the time of rising is accelerated by -^D" sec d seconds. COROLLARY 2. If the star is on the equator, d = 0, x coincides with E, and z nEP = nP = I, .-. the acceleration = -&D" sec I seconds. THE EARTH. 77 SECTION III. Geodetic Measurements Figure of the Earth. 105. Geodesy is the science connected with the accurate measurement of arcs on the surface of the Earth. Such measurements may be performed with either of the two following objects : (i.) The construction of maps. (ii.) The determination of the Earth's form and magnitude. Only the second application falls within the scope of this book. 10G. Alfred Russell Wallace's Method of Finding the Earth's Radius. An approximate measure of the Earth's radius can be readily found by means of the following simple experiment, due to Mr. A. 11. Wallace. FIG. 41. Let Z, M, JV(Fig. 41) be the tops of three posts of the same height set up in a line along the side of a straight canal. Owing to the Earth's curvature the straight line LM will, if produced, pass a little above N. Hence, in order to see Z, M in a straight line, an observer at the post JV^will have to place his eye at a point JST, a little above JV, and the height -ZTJV may be measured. Let JL, .Of be also measured. Since the posts are of equal height, Z, Jf, N will lie on a circle concentric with, and almost coinciding with, the Earth's surface. Let the vertical KN meet this circle again in n. By Euclid III. 36, KL . EM = EN. Kn; .-. Kn = KL . EMI EN, and Radius of Earth = \ Kn (very approximately) _ EL . EM 1EN This method cannot be relied on where accuracy is required, for the small height EN is very dim cult to measure, and a very slight error in its measurement would affect the final result considerably. Moreover the observations are consider- ably affected by refraction. 78 ASTROXOMT. 107. Ordinary methods of Finding the Earth's Radius. "Where greater accuracy is required, the radius of the Earth is obtained by measuring the length of an arc of the meridian and determining the difference of latitude of its extremities; the radius may then be calculated as in 91. The instruments required for the observations include (i.) Measuring rods, such as the double bar ; (ii.) A theodolite, for measuring angles ; (iii.) A zenith sector. 108. Measurement of a Base Line. The first step is to measure, with extreme accuracy, the length of the arc joining two selected points, several miles apart, on a level tract of country ; this line is called a Base Line. A series of short upright posts are placed at equal distances apart along the base line, and they are adjusted till their tops are seen exactly in the same vertical plane, and are on the same level as shown by a spirit level. Across these posts are laid measuring rods of metal, whose length is very accurately known, and these are also adjusted in a line, and made level by the spirit level. These rods are not allowed to touch, but the small distances between their ends are measured with reading microscopes. In this way, a base line several miles long can be measured correctly to within a small fraction of an inch f *109. The Double Bar. If the measuring rods be made of a single metal, their length i>. ^ iron I \j' will vary with the tempera- } ture. This disadvantage is, c 'l however, sometimes obviated by the use of the double bar (Fig 42). It consists of two bars, al, cd, one of iron, the other of brass. These are joined together in the middle, and to their ends are hinged perpendicular pointers eac, fbd of such length that ea : ec = /& : fd = coefficient of linear expansion of iron : that of brass, = about 11 : 18.f If the temperature be raised, the rods will expand, say to a'b', c'd'. But aa' : cc' = ea : ec, therefore e, and similarly /, will remain fixed. Hence the distance ef will be unaffected by the changes of temperature. _ f Wallace Stewart's Heat, Table 22. Brass K THE EARTH. 79 110. Triangnlation. When once a base line has been measured, the distance between any two points on the Earth can be determined by the measurement of angles alone. For, calling the base line AB, let C be any object visible from both A and B. If the angles CAB, CBA be observed, we can solve the triangle H - G ABC and determine the lengths of the ,+''* sides CA, CB. Either of these sides, say ^S s CA, may now be taken as the base of a new triangle, whose vertex is another point, D. Thus, by observing the angles of the tri- angle A CD we can determine DA, DC in terms of the known length of AC. Pro- ceeding in this way, we may divide any country into a network of triangles connect- ing different places of observation A, B, (7, D, and the distance between any two of the places calculated, as well as the direction of ^C / the line joining them. Finally, two stations ^' (7, H are taken, which lie on the same meri- dian, and the distance CU is calculated ; in IG ' this way it is possible to measure an arc of the meridian. 111. The Theodolite. The measurement of the angles is far easier in practice than the measurement of a base line. The instrument used for measuring angles is called a Theo- dolite, and is really a portable form of altazimuth. It is provided with spirit-levels, by means of which the instrument fan be adjusted so that the horizontal circle is truly horizon- tal, and the vertical axis, therefore, truly vertical; the direction of the north point is usually found by means of a compass needle. Most theodolites are only furnished with a small arc of the vertical circle, sufficient for measuring the altitude of one terrestrial object as seen from another. By reading the horizontal circle of the theodolite, the azimuths of B, C, as seen from A, are found. By using the difference of azimuth instead of the angle ABC, it becomes unnecessary to take account of the height of the various stations above the Earth. For if A, B, C are replaced by any other points, A', B', C', at the sea level, and vertically above or below A, B, G t the vertical planes joining them will be unaltered in position, and therefore the azimuths will also be unaffected. 80 ASTRONOMY. 112. Having thus found, with great accuracy, the length of the arc joining two stations on the same meridian, it only remains now to observe their difference of latitude. The Zenith Sector is the most useful instrument for this purpose. It consists essentially of a long telescope ST (Eig. 44), mounted so as to turn about a horizontal axis, A, near its object-glass ; this axis is adjusted to point due east and west (as in the transit circle). Attached to the lower end near the eye piece is a graduated arc of a circle GH, whose centre is at A. The line of collimation of the telescope is indicated by cross-wires placed in the field of view. A fine plumb- line, AP, is attached to the axis A, and hangs freely in front of the graduated arc. The plumb-line should mark zero when the line of collimation points to the zenith. When the instrument is pointed to any star, the reading opposite the plumb-line will be the star's zenith distance This reading can be determined with great accuracy by means of a reading microscope. 113. A star is selected which transits near the zenith* and its meridian zenith distances are observed at the two stations. Let these be s and z' degrees. Then if /, and /. 2 are the latitudes of the stations, and d the declination, we have, by 24, l'-l= (d-z')-(d-z) = z-z'. Hence, if s is the measured length of the arc of the meri- dian joining the stations, and r the radius of the Earth, 91 gives 18Q * _ 13 _ FIG. 44. whence the Earth's radius is found. * This position is chosen because the effects of atmospheric refraction are least in the neighbourhood of the zenith, THE EARTH. 81 114. Exact Figure of the Earth. If the Earth were an exact sphere, the same value would be found for the radius r in whatever latitude the observations were made. But in reality the length of a degree of latitude, and therefore also r, is found to be larger when the observation is made near the poles than when made near the equator, and hence it is inferred that the meridian curve is somewhat oval. Let PQP'R represent the meridian curve, 00' two near places of observation on it. Then, if 0J5Tand O'K be drawn normal (i.e., perpendicular) to the Earth's surface at 0, 0', they will be the directions of the plumb lines of the zenith sectors at 0, 0'. Hence the observed difference of latitudes or meridian altitudes at 0, 0' will give the angle OKO'. Eegarding the small arc 00' as an arc of a circle whose centre is JT, we shall have approximately, Circular measure of OKO' = arc 00' -f- OJT, arc 00' 180 s _ circ. measure of OKO' TT I' V and hence r, calculated as in 113, is the length OK. The length OK is called the radius of curvature of the arc, and K is called the centre of curvature ; they are respec- tively the radius and centre of the circle whose form most nearly coincides with the meridian along the arc 00'. This radius of curvature OK is not, in general, equal to C, the distance from the centre of the Earth, owing to the Earth FlG - 45 - not being quite spherical. As the result of numerous observations, the meridian curve is found to be an ellipse (see Appendix), whose greatest and least diameters, called the major and minor axes, are the Earth's equatorial and polar diameters respectively. The Earth's surface is the figure formed by making the ellipse revolve about its minor axis POP'. This figure is called an oblate spheroid. ASTRONOMY. 115. To find the Equatorial and Polar Radii of Cur- vature of the meridian curve, supposing 1 it to be an ellipse. Let PQP'R be the ellipse. Let 2, 2i be the lengths of its equatorial and polar diameters QCR, PCP'. Let r v r z be the required radii of curvature at Q and P respectively. Take any point on the ellipse, and let the normal at meet the two axes in G and g respectively. It is proved in treatises on Conic Sections* that OG : Og = CP* : C& = i 2 : a\ First take very near to Q. Then OG will become equal to the radius of curvature r^ ; also Og will evidently become ulti- mately equal to CQ or a. Therefore, ^ : a = b* : a? ; Next take very near to P. to I and Og to r%. Therefore, I : r 2 = W : 2 ; Thus r x , r 2 are found in terms of a, r = Then G will become equal r = and I may be found ; I r*r. Conversely, if r, and r 2 are known, for, by solving, we find a = %/(rfr ~We notice that since a > J, .*. r^ < r r That the equatorial radius of curvature is less than the polar is also evident from the shape of the curve. This, as the figure shows, is most rounded at Q, It, and flattest or least rounded at P, P'. Hence it will require a smaller circle to fit the shape of the curve at the equator than at the poles. 116. Exact Dimensions of the Earth. The lengths of the Earth's equatorial and polar semi-diameters, , i, are a = 3963-296 miles, I = 3949'791 miles. Thus, the Earth's equatorial semi-diameter exceeds its polar semi-diameter by 13-505 miles. * Appendix, Ellipse (9). THE EAETH. 83 The mean radius of an oblate spheroid is the radius of a sphere of equal volume, and is equal to ^/(a-1}. Thus, the Earth's mean radius is approximately 3958-8 miles. The ellipticity or compression (0) is the fraction For the Earth, c = - nearly. 293 The eccentricity (e) is given by the relation a~ Hence L l = s (I-* 2 ) = 8 (1 e)*; .-. !-* = (! -- 90- i. Thus, if the Sun be visible all day long during a certain period of the year, the latitude must be greater than 66 32^' K. or S. These circumstances have led to the following definitions. The Tropics are the two parallels to the Earth's equator in north and south latitude , or 23 27|-'. The northern tropic is called the Tropic of Cancer, the southern the Tropic of Capricorn. The Arctic and Antarctic Circles are respectively the parallels of north and south latitude 90 *, or 66 32f. These four parallels divide the Earth's surface into five regions or zones. The portion between the tropics is called the Torrid Zone. The portion between the tropic of Cancer and the arctic circle is called the North Temperate Zone. The portion between the tropic of Capricorn and the antarctic circle is called the South Temperate Zone. The portions north of the arctic circle, and south of the antarctic circle are called the Frigid Zones, and are distin- guished as the Arctic and Antarctic Zones. 121. Sun's Diurnal Path at Different Seasons and Places. "We shall now describe the various appearances presented by the Sun's diurnal motion at different times of the year, beginning in each case with the vernal equinox. We shall first suppose the observer at the Earth's equator, and shall then, describe how the phenomena are modified as he travels northward towards the pole. SUN'S APPARENT MOTION IN THE ECLIPTIC. 89 122. At the Earth's equator, I = 0, and the poles of of the celestial sphere are on the horizon (P, P', Fig. 47). Hence, between sunrise and sunset, the Sun has always to revolve about the poles through an angle 180, and the days and nights are always equal, each being 12 hours long. On March 21 the Sun is on the celestial equator, and it describes the circle EZW, rising at the east point, passing through the zenith at noon, and setting at the west point. Between March 21 and Sept. 23, the Sun is north of the celestial equator; it therefore rises north of E., transits north of the zenith Z, and sets north of W. Its IS", meridian zenith distance 2 is always equal to its !N". declination d (since by 24, 2 d I and I = 0) . Hence, from March 21 to June 21, z increases from to i N. On June 21, z has its greatest JN". value f, and the Sun describes the circle E'QW, where ZQ' = i. From June 21 to Sept. 23, z decreases from i to 0. On Sept. 23, the Sun again describes the great circle EQ W. Between Sept. 23 and March 21, the Sun is south of the equator, and therefore it transits south of the zenith. "We now have z = d, both being S. From Sept. 23 to Dec. 22, the Sun's south Z.D. at noon, 2, increases from to i. On Dec. 22, 2 has its greatest value i (south) and the Sun describes the circle E 'Q," W" where ZQ, " = i. From Dec. 22 to March 21, 2 diminishes again from to 0. On March 21, the Sun again describes the circle EQW, and the same cycle of changes is repeated the following year. 90 ASlRONOM*. 123. In the Torrid Zone North of the Equator". On March 21, the Sun describes the equator KQW (Fig, 48), rising at ^and setting at W. Here L ZPE L ZPW 90, and the day and night are each 12h. long. The Sun transits S. of the zenith at Q, where ZQ = z =7. From March 21 to June 21, d increases from to t, and the Sun's diurnal path changes from EQVto E'QW. The hour angles at rising and setting increase from ZPE and ZPWiQ ZPE' and ZPW, respectively ; hence the days increase and the nights decrease in length. The day is longest on June 21, when the hour angle ZPE' is greatest. The increase in the day is proportional to the angle EPE', and is greater the greater the latitude I. At first the Sun transits S. of the zenith, and z = ld. "When d = , z = 0, and the Sun is directly overhead at noon. After this, the Sun transits N. of the zenith, and z = d L On June 21,2 attains its maximum N. value ZQ' = il. From June 21 to Sept. 23, the phenomena occur in the reverse order. The diurnal path changes gradually back to EQW. The day diminishes to 12h. The Sun, which at first continues to transit N". of the zenith, becomes once more ver- tical at noon when d again = I, and then transits S. of the zenith. From Sept. 23 to Dec. 22, the Sun's path changes from EQWto E"Q'W". The eastern hour angle at sunrise decreases to ZPE"; thus the days shorten and the nights lengthen. The day is shortest on Dec. 22. Also z increases from I to 1 -f i. On Dec. 22, s attains the maximum value ZQ" = -f-, and the Sun is then furthest from the zenith at noon. From Dec. 22 to March 21, the length of the day increases again to 12 hours, and the Sun's meridian zenith distance decreases to z = L 124. On the Tropic of Cancer, I = i. The variations in the lengths of day and night partake of the same general character as in tbe Torrid Zone. But the Sun only just reaches the zenith at noon once a year, namely, on the longest day, June 21. At other times the Sun is south of the zenith at noon, and z attains the maximum value 2* on December 22. TIIE SUN'S APPABENT MOTION IN THE ECLIPTIC. 9l Z Q' 2 P' FIG. 49. 125. In the North Temperate Zone I > i but < 90 - i. Here the variations in the lengths of day and night are similar, hut more marked, owing to the greater latitude. On March 21, the Sun describes the equator EQWR (Fig. 49), which is bisected by the horizon ; hence the day is 1 2h . long. The length of the day increases from March 21 to June 21. The day is longest on June 21, when the jSun describes E'Q'WR', and the hour angles ZPE', ZPW are greatest. The days diminish to 12h. on Sept. 23, when the Sun again describes EQ, WE. The day is shortest on Dec. 22, when the Sun describes E"Q!'W"R". From Dec. 22 to March 21, the days increase in length, and on March 21 the day is again 12 hours long. The difference between the longest and shortest days is the time taken by the Sun to describe the angles E'PE", W"PTP', and is therefore = iV ( ^ E'PE" + L W'PW} = A . / E'PE". It will be seen that L E'PE" is greater in Fig. 49 than in Fig. 48, thus the variations are more marked in the tem- perate zone than in the torrid zone. The variations increase as the latitude increases. The Sun never readies the zenith' in the temperate zone, but always transits south of the zenith. The Sun's zenith distance at noon is least on June 21, when z = ZQ ' = li, and is greatest on Dec. 22, when % = ZQ" = l+i. At the equinoxes (March 21 and Sept. 23), z = ZQ = /. ASTEON. H 92 ASTRONOMY. 126. On the Arctic Circle, I = 90 t. Hence on June 21, when the Sun's KP.D. = 90-*', the Sun at midnight will only just graze the horizon at the north point without actually setting. On Dec. 22 at noon, the Sun's Z.D. = 90, and the Sun will just graze the horizon without actually rising. As in the preceding case, the days increase from Dec. 22 to June 21, and decrease from June 21 to Dec. 22; on March 21 and Sept. 23, the day and night are each 12h. long. 127. In the Arctic Zone we have l> 90- 1, and the variations are somewhat different (Fig. 50). On March 21, the Sun describes the circle EQW, and the day is 12h. long. As d increases, the days increase and the nights decrease, and this continues until d = 90 I. When this happens, the Sun at midnight only grazes the horizon at n. Subsequently, while ^>90 I, the Sun remains above the horizon during the whole of the day, circling about the pole like a circumpolar star. This period is called the Per- petual Day. During the perpetual day, the Sun's path continues to rise higher in the heavens every twenty -four hours until June 21, when the Sun traces out the circle R' Q'. The Sun's least and . greatest zenith distances will then be ZQ! = I i , and ZR' 180 tZ respectively. After June 21, the Sun's path will sink lower and lower. When d is again 90 I the perpetual day will end. Subsequently, the Sun will be below the horizon during part of each day. The days will then gradually shorten and the nights lengthen. On Sept. 23, the Sun will again describe the circle EQ, W, and the day and night will each be 12 hours long. The days will continue to diminish till the Sun's south declination d' 90 L When this happens the Sun at noon will only just graze the horizon at s. While d' >90 Z, the Sun remains continually below the horizon. This period is called the Perpetual Night. On Dec. 22 the Sun traces out the circle R"Q" below the horizon. When d' is again = 90 /, the perpetual night will end. Subsequently, the day will gradually lengthen until March 21, when it will again be 12 hours long. THE SUN'S APPARENT MOTION IN THE ECLIPTIC. 98 Z P FIG. 50. Sun's altitude mil attain its greatest val ,' on June 21 when the Sun will trace out the circle QK ' 21 ther night. . of the equator. In fact, if we consider two antipodal or places at opposite ends of a diameter of the Earth at one place will coincide with the night at the other l , equat r and antarctic c Me, the longest day, and June 21 the shortest. Within the antarctic circle there will be perpetual day for j certam penod before and after Dec. 22, and perpetual for a certain period before and after June 21. _ in _ _ _ _ r V OF THK UNIVERSITY 94 ASTRONOMY. The variations in the Sun's north zenith distance at noon will be the same as the variations in the south zenith distance in the corresponding north latitude six months earlier.* 130. The Seasons. Having thus described the variations in the Sun's daily path at different times and places, we shall now show how these variations account for the alternations of heat and cold on the Earth. Astronomically, the four seasons are denned as the portions into which the year is divided by the equinoxes and the solstices. Thus, in northern latitudes, Spring commences at the Yernal Equinox (March 21), Summer ,, ,, Summer Solstice (June 21), Autumn ,, ,, Autumnal Equinox (Sept. 23), Winter ,, Winter Solstice (Dec. 22). It is obvious that the temperature at any place will depend in a great measure upon the length of the day. While the Sun is above the horizon, the Earth is receiving a considerable portion of the heat of his rays, the remaining portion being absorbed by the Earth's atmosphere through which the rays have to pass. When the Sun is below the horizon, the Earth's heat is radiating away into space, although the heated atmosphere retards this radiation to a considerable extent. Thus, on the whole, the Earth is most heated when the days are longest, and conversely. The variations in the Sun's meridian altitude have a still greater influence on the temperature. When the Sun's rays strike the surface of the Earth nearly perpendicularly, the same pencil of rays will be spread over a smaller portion of the surface than when the rays strike the surface at a considerable angle ; hence the quantity of heat received on a square foot of the surface will be greatest when the Sun is most nearly vertical. By this mode of reasoning it is shown in Wallace Stewart's Text-Boole of Light, 10, that the intensity oi illumination of a surface is proportional to the cosine of the angle of incidence, and the same argument holds good with * The student will find it instructive to trace out fully the varia- tions in S. latitudes corresponding to those described in 122-128. See diagram, p. 421. IN THE ECLIPTIC. 95 regard to radiant heat as well as light. Hence the Sun's heat- ing power when ahove the horizon is always proportional to the cosine of the Sun's zenith distance or the sine of its altitnde. In this proof, however, the absorption of heat by the Earth's atmosphere has been neglected. But when the Sun's rays reach the Earth obliquely, they will have to pass through a greater extent of the Earth's atmosphere, and will, therefore, lose more heat than when they are nearly vertical. This cause will still further increase the effect of variations in the Sun's altitude in producing variations in the temperature. 131. Between the Tropics the combination of the two causes above described tends to produce high temperatures, subject only to small variations during the year. The Sun's meridian altitude is always very great, and the variations in the lengths of day and night are small. If the latitude be north, the Sun's heating power is greatest while the Sun transits north of the zenith. During this period the Sun's meridian altitude is least when the days are longest. Thus the effects of the two causes in producing variations in the Sun's heat counteract one another, to a certain extent, and give rise to a period of nearly uniform but intense heat. In the North Temperate Zone, the Sun is highest at noon when the days are longest, and therefore both causes combine to make the spring and summer seasons warmer than autumn and winter. But the highest average tempera- tures occur some time after the summer solstice, and the lowest temperatures occur after the winter solstice ; for the Earth is gaining heat most rapidly about the summer solstice, and it continues to gain heat, but less rapidly, for some time afterwards. Similarly, the Earth is losing heat most rapidly at the winter solstice, and it continues to lose heat, but less rapidly, for some time afterwards. Por this reason, summer is warmer than spring, and winter is colder than autumn. ' As we go northwards, the Sun's altitude at noon becomes generally lower throughout the year, and the climate therefore becomes colder. At the same time, the variations in the length of the day become more marked, causing a greater fluctua- tion of temperature between summer and winter. 96 ASTRONOMY. Within the Arctic Circle there is a warm period during the perpetual day, but the Sun's altitude is never sufficiently great to cause very intense heat. During the perpetual night the cold is extreme ; and the low altitude of the Sun, when above the horizon at intermediate times, gives rise to a very low average temperature during the year. In the Southern Hemisphere the seasons are reversed ; for, in south latitude I, when the Sun's south declination is d, the same amount of heat will be received from the Sun as in north latitude I, when his north declination is d. Hence, the seasons corresponding to our spring, summer, autumn and winter will begin respectively on September 23, December 22, March 21, and June 21, and will be separated from the corre- sponding seasons in north latitude by six months. 132. Other Causes affecting the Seasons and Climate. It is found (as will be explained in the next section) that the Sun's distance from the Earth is not quite constant during the year. The Sun is nearest the Earth about December 3 1 , and furthest away on July 1 (these are the dates of perigee and apogee respectively) . As shown in Wallace Stewart's Text-Book oj Light, 9, the intensity of illumination, and therefore also of heating, due to the Sun's rays, varies inversely as the square of the Sun's distance. Hence the Earth receives, on the whole, more heat from the Sun after the winter solstice than after the summer solstice. This cause tends to make the winter milder and the summer cooler in the northern hemisphere, and to make the summer hotter, and the winter colder in the southern hemisphere. The variations in the Sun's distance are, however, small, and their effect on the seasons is more than counter- acted by purely terrestrial causes arising from the unequal distribution of land and water on the Earth. The sea has a much greater capacity for heat than the rocks forming the land ; it is not so readily heated or cooled. In the southern hemisphere the sea greatly preponderates, the largest land- surfaces being in the northern hemisphere. Hence, the climate of the southern hemisphere is generally more equable, and the seasons are not so marked as in the northern hemi- sphere, quite in contradiction to what we should expect from the astronomical causes. THE SUN'S APPAKENT MOTION IN THE ECLIPTIC. 97 133. Times of Sunrise and Sunset. The times of sunrise and sunset at Greenwich are given for every day of the year in Whitaker > & and other almanacks. For any other latitude, the Sun's declination must be found from the almanack, the times of sunrise and sunset can then be found by means of tables of double entry constructed for the pur- pose (29). These are called ''Tables of Semidiurnal and Seminocturnal Arcs." . They give, for different latitudes and declinations, the interval between apparent noon and sunset, .#., the apparent time of sunset, or half the length of the day. Subtracting this from 12 hours, the apparent time of sunrise is found, and is half the length of the night. If, as in 129, we consider two antipodal places A and S, the planes of their horizons will be parallel, and the Sun will be above the horizon at A when he is below the horizon at J3, and vice versd. Hence, the apparent time of sunrise (measured from noon) in N. latitude I will be the apparent time of sunset (measured from midnight) in S. latitude I on the same date. For this reason the tables are usually constructed only for N. latitudes. For S. latitudes they give the time of sunrise instead of sunset. The times found in this manner will be the local solar times. To reduce to Greenwich solar time we must add or sub- tract 4m. for each degree of longitude, according as the place is W. or E. of Greenwich. 134. To find the length of the perpetual day and night at places within the Arctic or Antarctic Circles. The perpetual day lasts while the Sun's declination at local midnight is greater than the colatitude (or complement of the latitude), during spring and summer. The perpetual night lasts while the Sun's S. decl. at local noon is greater than the colat. during autumn and winter. The Sun's decl. at Green- wich noon being given for every day of the year, in the Nautical Almanack, it is easy to find, to within a day, the durations of the perpetual day and night in any given latitude greater than 66 32|'. 98 ASTRONOMY. 135. To find the time the Sun takes to rise or set. Let D" be the Sun's angular diameter, measured in seconds. When the Sun begins to rise, his upper limb just touches the horizon, and his centre is at a depth \D" below the horizon. When the Sun has just finished rising, his lower limb touches the horizon, and his centre is at an altitude |_D" above the horizon. During the sunrise, the centre rises through a vertical height D". The problem is closely similar to that of 104, where the effect of dip is considered. Hence if t seconds be the time taken in rising, d the declination of the Sun's centre, and x the inclination to the vertical of the Sun's path at rising (Hx'x or nxP, Tig. 40) we have t = -jV D" sec d sec #, = 4 sec d sec x x (O's angular diameter in minutes). As in 104, this gives, for a place on the equator, t -^D n sec d, and at an equinox in latitude ?, t = T V D" sec I. EXAMPLE. At an equinox in latitude 60, the O's angular diameter being 32', the time taken to rise will be = 4 x 32 x sec 60 seconds = 256s. = 4m. 16s. 136. Note. It may be mentioned that, owing to atmos- pheric refraction, the Sun really appears to rise earlier and set later than the times calculated by theory. As the pheno- mena of refraction will be discussed more fully in Chapter VI., it will be sufficient to mention here that the rays of light from the Sun are bent to such an extent by the Earth's atmosphere that the whole of the Sun's disc is visible when it would just be entirely below the horizon if there were no atmosphere. Moreover, there is daylight, or rather twilight, for some time after the Sun has vanished, so that what is commonly called night does not begin for some time after sunset. For the same reasons, the perpetual day at a place in the arctic circle is lengthened, and the perpetual night shortened, by several days. The time taken in rising and setting is, however, prac- tically UTI affected. 99 SECTION II. The Ecliptic. 137. The First Point of Aries. In determining the right ascensions of stars, the first step must necessarily be to find accurately the position of the first point of Aries, since this point is taken as the origin from which R.A. is measured. In other words, we must first find the R.A. of one star. When this is known we can use that star as a " clock star," to determine the sidereal time and clock error ; and, these being known, we can then find the R.A. of any other star, as explained in Chapter II. But until the position of T has been found, the methods of Chapter II. will only enable us to find the difference of R.A. of two stars by observing the difference of their times of transit, as indicated by the astro- nomical clock, and will determine neither the sidereal time nor the clock error, nor the R.A.'s of the stars. 138. First Method. The position of T may be found thus : At the vernal equinox the Sun's declination changes from south to north, or from negative to positive. Let the Sun's declination be observed by the Transit Circle at the pre- ceding and following noons, and let the observed values be ^and -f^ 2 (.*., ^ S., and d t IT.). Let t v 2 be the corre- sponding times of transit of the Sun's centre, as observed by the astronomical clock, and let T^ the time of transit of any star, be also observed. Then, T tfj = difference of R. A. of star and Sun at first noon, Tt z = ,, ,, ,, at second noon. Let T rfj = ^ and T t z = 2 . "We have Increase in Sun's decl. in the day = d% ( d l ) = d z + d ly ,, ,, R.A. ,, = t t t l a l a. 2 , and both coordinates increase at an approximately uniform rate during the day. Therefore the Q's decl. will have increased from d l to in a time d l /(d l + d^ of a day, and the corresponding increase in R.A. will be fa-oa) x dj(d l + d,\ The Sun is now at T, .' O's R.A. is now = 0. Hence, The star's R.A. = a, - ^"^ *** + A 100 A.STEONOMT. *139. Flamsteed's Method for finding the First Point of Aries. The principle of the method now to be described is as follows : Let 8 lt $ be two positions of the Sun shortly after the vernal and before the autumnal equinox respectively, and such that the declinations S l J/j and SM are equal. Then the right-angled triangles r^/"A and ^MS will be equal in all respects, and we shall therefore have FIG. 52. At noon, some day shortly after March 21, the Sun is observed with the Transit Circle, say when at 8 V We thus determine its meridian zenith distance z 15 and also the dif- ference between the times of transit of the Sun and some fixed star x, whose R.A. is required. This difference, which is the difference of E.A. of the Sun and star, we shall call a r If d l be the Sun's declination, and I the observer's latitude, we shall have = a We now have to determine J/7V, the difference of R.A. of the Sun and star shortly before September 23, when the Sun'g declination SMis again equal to d r But the Sun can only be observed with the Transit Circle at noon, and it is highly improbable that the Sun's declination will again be exactly equal to d 1 at noon on any day. We shall, however, find two consecutive days in September on which the declinations at noon, S 2 M 2 and $ 3 Jf 3 , are respectively greater and less than d^ THE SUN'S APPARENT MOTION^ Iff T3l2 Let 2 2 and 2 8 be the observed meridian zenith distances at 3 and S & ; d% and $s> we ma y assume that the Sun's decl. and R.A. both vary at a uniform rate, so that the change in the decl. is always proportional to the corresponding change in R.A.* Therefore, and MN= M^N-M,M= a,- Now we have shown that -M 1 N= HN- ^ -12 hours; = 6h. + This determines T-ZV, the star's R. A., in terms of 15 a v 3 , the observed differences between the times of transit of the Sun and star, and d lt d^ d^ the Sun's declinations at the three observations. But we need not even find the declinations, for d l = l-z v d, = l-z v d s = l-% ; therefore, substituting, we have The star's R.A., r^= 6h.+f j ^-f^-^^ (a^-a,) } . 2 3~ 2 2 In applying either of the above methods to the numerical calcula- tion of the right ascension of any star, it is advisable to follow the various steps as we have described them, instead of merely sub- stituting the numerical values of the data in the final formulas. * In other words, we assume, as in Trigonometry, that tho " principle of proportional parts " holds for the small variations in decl. and E.A. during the day. 10 V 2 ASTRONOMY. *140. The Advantages of Flamsteed's Method. Among these the following may be mentioned. 1st. The method does not require a knowledge of the latitude, for we do not require to find the Sun's declination. Hence, errors arising from inaccurate determination of the latitude are avoided. 2nd. One great source of error in determining Z.D.'s is the refrac- tion of the Earth's atmosphere. Since the Sun is observed each time in the same part of the sky, z lt z 2 , 3 will be nearly equally affected by refraction. Hence, the " principle of proportional parts" will hold, so that the small differences in the true Z.D.'s are proportional to the differences in the observed Z.D.'s. Hence we may use the observed Z.D.'s uncorrected for refraction. EXAMPLE. To find the Right Ascension of Sirius and the clock errors in March and Sept., 1891, from the following data, the rate of the clock being supposed correct. (Decl. of Sirius = 16 34' 2" S.) Mar. 25, 1891. Sept. 18. Sept. 19 Decl. of Sun at noon... Time of transit of Sun Time of transit oiSirius 1 48' 56" Oh. 15m. 36s. 6h. 39m. 10s. 1 53' 0" llh. 42m. 42s. 6h. 40m. 25s. 1 29' 43" llh. 46m. 17s. 6h. 40m. 25s. OnMar.25,(R.A.ofSmws)-(Sun'sR.A.)=6h 39m. 10s. -Oh. 15m. 36s. =6h.23m.34s. Hence, in angular measure, the difference of R.A. is about 96. Draw the diagram as in Fig. 52, but make the angle SiPN = 96; iV will therefore lie between M l and J5f 2 , instead of where represented. Also, since Sirius is south of the equator, it should be represented at a point x on PN produced through N. In this figure we shall have 8^ = 148'56"; MiN = 6h.39m.10s. -Oh.15m.36s. = 6h.23m.34a. S 2 3f 2 = 153' 0"; NM. 2 = llh.42m.42s.-6h.40m.25s. = 5h. 2m. 17s. = 129'43"; NM 3 = Ilh.46m.l7s.-6h.40m.25s. = 5h. 5m.52s. Also, SM is by construction equal to S^M } . Hence, applying the principle of proportional parts, we have SoMg-giJf! = 4' 4" = 244 S. 2 M 2 -S 3 M 3 23' 17" 1397' and M%M 3 = 3m. 35s. = 215s. ; .-. M*M = 215 x 244/1397 = 37'5 seconds ; .-. NM = 5h. 2m. 17s. + 37s. = 5h. 2m. 54s. Now, NMt-NAI = NT -N~ = 2Nr -12h. bonce, TN = 6}i. + %(NMi-NM) = 6h. + i(6h.23m.34s.-5h.2m.54s.) = 6h. + (lh. 20m. 40s.) = 6h. 40m. 20s. Thus the right ascension of Sirius = 6h, 40m. 20s. Also, clock error in March = 6h.40m.20s.-6h.39m.10s. = + 1m. 10s. Sept. = 6h.40m.20s.-6h.40m.25s. = 5s. 103 141. Precession of the Equinoxes. Thus far we have treated the first point of Aries as being fixed, and this will evidently be the case if the equator and ecliptic are fixed in direction. But if the right ascensions of various stars are observed over an interval of several years, it will be found that the position of the first point of Aries is slowly changing, and that it moves along the ecliptic in the retrograde direc- tion at the rate of about 50-2" in a year. This motion is called Precession of the Equinoxes, or, briefly, Precession. Precession is found to be due almost entirely to gradual changes in the direction of the plane of the equator, the ecliptic remaining almost fixed among the stars. Its effect is to produce a yearly increase of 50-2" in the celestial longi- tudes of all stars, their latitudes being constant. In a large number of years the effect of precession will be considerable. Thus, T will perform a complete revolution in the period 360x60x60 years, i.e., about 25,800 years. o(j' 2i At the present time the vernal equinoctial point has moved right out of the constellation Aries into the adjoining con- stellation Pisces. It still, however, retains the old name of lt First Point of Aries." Similarly, the autumnal equinoctial point is in the constellation Virgo, but it is still called the " First Point of Libra." The rate of precession can be found very accurately by observations of the first point of Aries separated by a con- siderable number of years. The larger the interval, the larger is the change to be observed, and the less is the result affected by instrumental errors. *142. Correction for Precession in using Flamsteed's Method. During the interval that elapses between the two observations in Flamsteed's method, the right ascension of the observed star will have increased slightly, owing to precession, and the E.A. given by the formula will be the arithmetic mean of the E.A.'s at the times of the two observations.f As the change in E.A. is very approximately uniform, this mean will be the star's E.A. at a time exactly half way between the two observations, i.e., at the summer solstice. t This may be most readily seen by imagining the equator and ecliptic to be at rest, and the change in E.A. to be due to motion of the star. 104 ASTROJDMT. 143. Determination of Obliquity of Ecliptic. The method now used for finding the obliquity of the ecliptic is similar in principle to that of 38, hut the Sun's meridian zenith distance is observed by means of the transit circle instead of the gnomon. The obliquity is equal to the Sun's greatest declination at one of the solstices. Since observations with the Transit Circle can only be performed at noon, while the maximum declination will probably occur at some intermediate hour of the day, it will be necessary, in exact determinations, to make observations of the Sun's decl. for several days before and after the solstice. Prom these it is possible to determine the maximum decl. ; the method is, however, too complicated to be described here. For rough purposes the Sun's greatest noon decl. may be taken as the measure of the obliquity. 144. When the position of T has been determined, the obliquity can also be found by a single observation of the Sun's E-.A. and decl. For we thus find the two sides T-3/i MS of the spherical triangle T^S, and these data are sufficient to determine both the obliquity flfTS, and the Sun's longitude T S. THE SUN'S APPARENT MOTION IN THE ECLIPTIC. 105 SECTION III. The Earth's Orbit about the Sun. 145. Observations of the Sun's Relative Orbit. By daily observations with the Transit Circle, the decl. and R. A. of the Sun's centre at noon are found for every day of the year. From these data the Sun's long, is calculated, as in 144, by solving the spherical triangle T SM (Fig. 53). If the obliquity of the ecliptic is also known, we have three data, any two of which suffice to determine the long., T$. Thus the accuracy of the observations can be tested, and the Sun's motion at various times of the year can be accurately determined. Although the determination of the Sun's actual distance from the Earth in miles is an operation of great difficulty, it is easy to compare the Sun's distance from the Earth at dif- ferent times of the year, for this distance is always inversely proportional to the Sun's angular diameter. This property is proved in 4, but numerous simple illustrations may also be used to show that the angular diameter of any object varies inversely with its distance (see 4). The Sun's angular diameter may be readily observed by means of the HeHometer ; or, if preferred, any other form of micrometer may be used. The Sun's distances at two different observations will be in the reciprocal ratio of the corresponding angular diameters. Thus, by daily observation, the changes in the Sun's distance during the year may be investigated. If the circular measure of the Sun's angular diameter is 2r, then Trr 2 is called the Sun's apparent area. In fact, this is the area of a disc which would look the same size as the Sun if placed at unit distance from the eye. EXAMPLE. The Sun's angular diameter is 31' 32" at midsummer, and 32' 36" at midwinter. To find the ratio of its distances from the Earth at these times. The distances being inversely proportional to the angular dia- meters, we have Dist. at midsummer = 82' 36" = 1956 = 489 _ , 1 , Dist. at midwinter 31' 32" 1892 473 "** Hence the Sun is further at midsummer than at midwinter, in the proportion of very nearly 81 to 30. 106 ASTRONOMY. 146. Kepler's First and Second Laws. We may no\\ construct a diagram of the Sun's relative orbit. Let E repre- sent the position of the Earth, ET the direction of the first point of Aries. Then, by making the angle TES equal to the Sun's longitude at noon, and ES proportional to the Sun's distance, we obtain a series of points S, S'... , 8 r .. , representing the Sun's position in the plane of the ecliptic, as seen from the Earth at noon on different days of the year. Draw the curve passing through the points S, S'... , S r .. ; this curve will represent the Sun's orbit relative to the Earth, and it will be found that I. The Sun's annual path is an ellipse, of which the Earth is one focus. II. The rate of motion is everywhere such that the radius vector (i.e., the line joining the Earth to the Sun) sweeps out equal areas in equal intervals of time. These laws were discovered by Kepler for the motion of Mars about the Sun, and he subsequently generalized them by showing that the orbits of all the other planets, including the Earth, obeyed the same laws. In their general form they are known as Kepler's First and Second Laws. [See p. 253.] FIG. 54. 147. Perigee and Apogee. When the Sun's distance from the Earth is least, the Sun is said to be in perigee. When the distance is greatest, the Sun is said to be in apogee The positions of perigee and apogee are called the two Apses of the orbit ; they are indicated at p, a in Fig. 54. The line pEa joining them is the major axis of the ellipse (Ellipse, 4), and is sometimes also called the apse line. 107 148. Verification of Kepler's First Law. The Sun's angular diameter is observed to be greatest on Dec. 31, and least on July 1 ; we therefore conclude that these are the days on which the Sun passes through perigee and apogee respectively. The positions of perigee and apogee being thus found, the angle TEp is known, which is the long, of perigee. From the winter solstice to perigee is about 10 days. Hence, during this interval the Sun will have moved through an angle of about 10 ; .-. longitude of perigee = 270 + 10 = 280 roughly. To verify that the orbit is an ellipse, it is now only neces- sary to show that the relation connecting ES and the angle pES is the same as that which holds in the case of the ellipse. If the orbit is an ellipse of eccentricity r8 l i .:rK>r8 t ; .-.Dispositive. At the autumnal equinox, since fC = TQ==z 180, S lt &J will both coincide with b; .. JEJ 8 = 0. In a similar manner we may show that : From the autumnal equinox to the winter solstice, U 3 is negative. At the winter solstice, 1 2 as O. From the winter solstice to the vernal equinox, E% is positive. Collecting these results, we see that (i.) From equinox to solstice /^ is negative. (ii.) From solstice to equinox J 3 is positive. (iii.) Es vanishes four times a year, viz., at the equinoxes and solstices. ON TIME. 121 163. Graphic Representation of Equation of Time. The values of the equation of time at different seasons may now be represented graphically by means of a curved line, in which the abscissa of any point represents the time of year, and the ordinate represents the corresponding value of the equation of time. In the accompanying figure (Fig. 59) the horizontal line or axis from JE^ to E^ represents a year, the twelve divisions representing the different months as indicated. The thin curve represents the values of E^ the portion of the equation of time due to the unequal motion ; this curve is obtained by drawing ordinates perpendicular to the horizontal axis and proportional to E r Where the curve is below the horizontal line E i is negative. FIG. 59. The thick curved line is drawn in a similar manner, and represents, on the same scale, the values of E v the equation of time due to the obliquity. In drawing the diagrams to scale, it is necessary to know the maximum values of JEJ, E v These can be calculated, but the calculations do not depend on elementary methods alone. We shall therefore have to assume the following facts : The greatest value of E l is about 7 minutes. Hence the greatest distances of the thin and thick curves from the horizontal axis should be taken to be about 7 and 1 units of length respectively. 122 ASTRONOMY. We may now draw the diagram representing JE, the total equation of time. We have Hence, at every point of the horizontal line we must erect an prdinate whose length is equal to the algebraic sum of the ordinates (taken with their proper sign) of the two curves which represent E^ and E The extremities of these ordi- nates will determine a new curve which represents E. FlG. 60. This curve is drawn separately in the annexed diagram (Fig. 60). It cuts the horizontal axis in four points. At these points the ordinate vanishes, and E is zero. Hence, The Equation of Time vanishes four times a year. 164. Alternative Proof. But without representing the values of the equation of time graphically, it can he readily proved that E vanishes four times a year. The proof depends on the fact stated in the last paragraph, that The greatest equation of time due to the obliquity is greater than the greatest equation due to the eccentricity. Off TIME. 123 From 162 it is evident that JS t must attain its greatest positive value some time between a solstice and the following equinox, and its greatest negative value between an equinox and the following solstice. These maxima occur, in fact, in the months : February, May, August, November. Their values, with the proper signs, are respectively about -flOm., 10m., 4- 10m., 10m. Now, E l is never greater than the maximum value of 7m. ; hence, whether E^ is positive or negative, the total equation, E^ -f EX corresponding to either of these maxima, must have the same sign as E Hence, in the year beginning and ending with the date of the maximum value of E^ in February, E will have the following signs alternately : + + + Thus, 73 changes sign, and thereJx^e vanishes, four times in the year. 165. Miscellaneous Remarks. "From Pig. 59 it will be seen that the largest fluctuations in the equation of time occur in the autumn and winter months ; during spring and summer they are much smaller. The days on which the equation of time vanishes are about April 16, June 15, September 1, and December 25. Between these days E increases numerically, and then decreases, attaining a positive or negative value at some inter- mediate time. These maxima are : + 14m. 28s. on February 11 ; 3m. 49s. on May 14 ; -f-6m. 17s. on July 26 ; 16m. 21s. on November 3. 166. Inequality in the Lengths of Morning and Afternoon. If we neglect the small changes in the Sun's declination during the day, the interval from sunrise to apparent noon is equal to the interval from apparent noon to sunset ( 37). But by morning and afternoon are meant the intervals between sunrise and mean noon, and between mean noon and sunset respectively. Hence, unless mean and appa- rent noon coincide, i.e., unless the equation of time vanishes, the morning and afternoon will not be equal in length. 4-STKOF, K 124 ASTBONOMT. Let r, * be the mean times of sunrise and sunset, E the equation of time. Then 12h. r = interval from sunrise to mean noon. But apparent noon occurs later than mean noon by E\ .'. 12h. r-\-E interval from sunrise to apparent noon. Similarly, sJE= interval from apparent noon to sunset; .-. 12h.-r+.# =*-.#, or r + s= 12h. + 2^, so that the sum of the times of sunrise and sunset exceeds 12 hours by twice the equation of time. The length of the morning is 12h. r, and that of the afternoon is . Now the last relation gives 2J0 = -(12-r); .-. 2 (equation of time) = (length of afternoon) (length of morning). About the shortest day (December 22) the curve represent- ing the equation of time is going upwards, hence E is increasing. But the length of day is changing very slowly (because it is a minimum), hence, for a few days, the half length, jE", may be regarded as constant. Hence, must increase, and, therefore, the mean time of sunset is later each day. Similarly, it may be shown that sunrise is also later. The afternoons, therefore, begin to lengthen, while the mornings continue to shorten. Similarly, about June 21, the afternoons continue to lengthen after the longest day, although the mornings are already shortening. EXAMPLE. On Nov. 1, the sun-dial is 16m. 20s. before the clock. Given that the Sun rose at 6h. 54m., find the time of sunset. Time from sunrise to mean noon = 12h. 6h. 54m. = 5h. 6m. apparent noon to mean noon = Oh. 16m. 20s. sunrise to apparent noon = 4h. 49m. 40s. apparent noon to sunset = 4h. 49m. 40s. mean noon to sunset = 4h. 49m. 40s. - 16m. 20s. = 4h. 33m. 20s. Hence, the time of sunset was 4h. 33m., correct to the nearest minute. ON TIME. J 25 SECTION II. The Sun-dial essentially of a rod or flat the direction of the celestial pole The shadow Pia. 61. The plane through OA, the edge of the style, and throueh the edge of the shadow, evidently passes through the Sun also it passes through the celestial pole, therefore it will meet ' "Declination ce shln t plane, which is the plane of the known t,? p P arent f 00 ?' and whose position is supposed 123' oli e A 1 Y- rder , to graduate the p late for *~ 1, ^, do clock, it is only necessary to determine the posi- esof no,-' " - c "' w l ' 1 ' ' , 45 . ' &0 '' Wlth the meridian plane. Since be thn - S P erour ' te e P^e* wffl the planes bounding the shadow at 1, 2 3 o'clock rcspecfavely. If we join the points 0,, On. dm *'"&* , hese , lines of shadow in the plane of the m * * he cirram ^ of the dial-plate 126 ASTRONOMY. 168. Geometrical Method of Graduating the Dial- plate. To find the planes OA i., OAn., &c., suppose a plane AKR drawn through A perpendicular to OA, meeting the plane of the dial-plate in KR and the meridian plane in A~s.ii. If, in this plane, we take the angles xn.-4i., i.^n., n.^tm., &c., each = 15, the points i., n., m...., &c., will evidently determine the directions of the shadow at 1, 2, 3,... o'clock respectively. FIG. 62 But in practice it is much more convenient to perform the construction in the plane of the dial itself. Imagine the plane AKR of Fig. 62 turned about the line KR till it is brought into the plane of the dial, the point A of the plane being brought to U (Fig. 62). Then, by making the angles xii. 7i., i. ^n., n..Z7"m., &c., each = 15, we shall obtain the same series of points i., n., m. as before. If the dial -plate is horizontal, and I is the latitude of the place (xn. OA), we have evidently therefore the following construction : On the meridian line, measure xn. = OA sec I, and xn. U = xn. A = Oxn. sin I. Draw -ZTxn. R perpendicular to OU. Make the angles xn.ZTi., i.Z7n., n.JTni., &c., each = 15, taking i., n., m., &c., on KR. Join 0i., On., 0m., &c., and let the joining lines meet the circumference of the dial in 1, 2, 3, &c. These will be the required points of graduation for 1, 2, 3,... o'clock respectively. 127 SECTION III. Units of Time The Calendar. 169. Tropical, Sidereal, and Anomalistic Years. Hitherto we have defined a year as the period of a complete revolution of the Sun in the ecliptic. In order to give a more accurate definition, however, it is necessary to specify the starting point from which the revolution is measured. "We are thus led to three different kinds of years. A Tropical Year is the period between two successive vernal equinoxes, or the time taken by the Sun to perform a complete revolution relative to the first point of Aries. The length of the tropical year in mean solar time is very approximately 365d. 5h. 48m. 45 -5 Is. at the present time. For many purposes it may be taken as 365 days. A Sidereal Year is the period of a complete revolution of the Sun, starting from and returning to the secondary to the ecliptic through some fixed star. Thus, after a sidereal year the Sun will have returned to exactly the same position among the constellations. If T were a fixed point among the stars, the sidereal and tropical year would be exactly of the same length. But T has an annual retrograde motion of 50-22" among the stars ( 141). Consequently, the tropical year is rather shorter than the sidereal. An Anomalistic Year is the period of the Sun's revo- lution relative to the apse line in other words, the interval between successive passages through perigee. Owing to the progressive motion of the apse line, the positions of perigee and apogee move forward in the ecliptic at the rate of 11-25" per annum ( 153). Hence the anomalistic year is rather longer than the sidereal. It is easy to compare the lengths of the sidereal, tropical, and anomalistic years. For, relative to the stars, In the sidereal year the Sun describes 360, In the tropical year it describes 360 - 50-22", In the anomalistic year it describes 360 -f 11-25" ; .'. (Sidereal year) : (tropical year) : (anomalistic year) = 36O : 360 50-22": 360+11'25". From this proportion it will be found that the sidereal year is about 20 m. longer than the tropical, and 4 Jin. shorter than the anomalistic. 128 AST&ONOMf. 170. The Civil Tear. For ordinary purposes, it is important that the year shall possess the following qualifications : 1 st. It must contain an exact (not a fractional) number of days. 2nd. It must mark the recurrence of the seasons. Now the tropical year marks the recurrence of the seasons, but its length is not an exact number of days, being, as we have seen, about 365d. 5h. 48m. 45 -5 Is. To obviate this disadvantage, the civil year has been introduced. Its length is sometime? 365, and sometimes 366 days, but its average length is almost exactly equal to that of the tropical year. Taking an ordinary civil year as 365d., four such years will be less than four tropical years by 23h. 15m. 2'04s. ? or nearly a day. To compensate foi this dilierence, every fourth civil year is made to contain 366 days, instead of 365, and is called a leap year. For convenience, the leap years are chosen to be those years the number of which is divisible by 4, such as 1892, 1896. The introduction of a leap year once in every four years is due to Julius Caesar, and the calendar constructed on this principle is called the Julian Calendar. Now three ordinary years and one leap year exceed four tropical years by 24h. 23h. 15m. 2'04s., t.g.!44m. 57j96a^ Thus, 400 years of the Julian Calendar will exceed 400 tropical years by (44m. 57'96s.) x 100, i.e., by 3d._2h^56m.36s. To compensate for this difference, Pope ~Gregory~XlII. arranged that three days should be omitted in every 400 years. This correction is called the Gregorian correction and is made as follows : Every year whose number is a multiple 0/100 is taken to be an ordinary year of 365 days, instead of being a leap year of 366, unless the number of the century is divisible by 4; in that case the year is a leap year. EXAMPLES. (i.) 1892 is divisible by 4, .*. the year 1892 is a leap year, (ii.) 1900 is a multiple of 100, and 19 is not divisible by 4, .'. 190O is not a leap year. (Hi.) 2000 : the number of tho century is 20, and is divisible by 4, .'. 2OOO is a leap year. The Gregorian correction still leaves a small difference between the tropical year and the average length of the civil year, amounting to only Id. 5h. 26m. in 4,000 years. 171. A Synodic Year is a period of 12 lunar months, being nearly 355 days. The name is, however, rarely used. OK TIME. 129 SECTION IV. Comparison of Mean and Sidereal Times. 172. Relation between Units. One of the most important problems in practical astronomy is to find the sidereal time at any given instant of mean solar time, and conversely, to find the mean time at any given instant of sidereal time. Before doing this it is necessary to compare the lengths of the mean and sidereal days. We have seen ( 169) that a tropical year contains abont 365| mean solar days. In this period both the true and mean Sun describe one complete revolution, or 360 from west to east relative to T ; or, what is the same thing, T describes one revolution from east to west relative to the mean Sun. But the mean Sun performs 365 revolutions from east to west relative to the meridian at any place. Therefore T performs one more revolution, i.e., 366 revo- lutions, relative to the meridian. Now, a sidereal day and a mean solar day have been defined ( 22, 159) as the periods of revolution of the mean Sun and of T relative to the meridian ; .-. SOS} mean solar days = 366| sidereal days. Prom this relation we have, One mean solar day = ( 1 + ~ ) sidereal days \ ODO^-/ = (1 + '002738) sidereal days = 24h. 3m. 56'5s. sidereal time = 1 sidereal day + 4m. 4s. nearly; .. one mean solar hour = Ih. + 10s. s. sidereal time, and 6m. of mean solar time = Gin. + Is. sidereal time nearly. In like manner we have One sidereal day = ( 1 ) mean solar days \ o66f/ -- (1 -002730) mean days = 23h. 56m. 4' Is. mean time = 1 mean day 4m. -f 4s. nearly ; .*. one sidereal hour = Ih. 10s. +-J-S. of mean time, and 6m. sidereal time = 6m. Is.meansolartimenearly. L 130 tf __ |4 173. From the results of the last paragraph we have the following approximate rules : (i.) To reduce a given interval of mean time to sidereal time, add 10s. for every hour, and Is. for every 6m. in the given interval. For every minute so added, sub- tract Is. (ii.) To reduce a given interval of sidereal time to mean time, subtract 10s. for every hour, and Is. for every 6m. in the given interval. Then add Is. for everr minute so subtracted. EXAMPLE 1. Express in sidereal time an interval of 13h. 23m. 25s. mean time. The calculation stands as follows : H. M. s.^ Mean solar interval =13 23 25 Add 10s. per hour on 13h.... ' 2 10 Is. per 6m. on 23m. ... ... ... 4 13 25 39 Subtract Is. per 1m. on 2m. 13'8s. ... 2 .-. Required sidereal interval = 13 25 37 EXAMPLE 2. Find the mean solar interval corresponding to 14h. 45m. 53s. of sidereal time. The calculation stands as follows : H. M. s. Given sidereal interval ... ... ... ... = 14 45 53 Subtract 10s. per hour on 14h. = 2m. 20s. \ o 28 Is. per 6m. on 46m. (nearly )= 8s. / 14 43 25 Add Is. per 1m. on 2m. 28s. .... ... 3 .. Required interval of mean time =14 43 28 If accuracy to within a few seconds is not required, the second correction of Is. per 1m. may be omitted. On the other hand, if the interval consists of a considerable number of days, or if accuracy to the decimal of a second is needed, the results found by the rules will no longer be correct. "We must, instead, add 1/365^- of the given mean solar interval to get the sidereal interval, or subtract 1/366 J of the given sidereal to get the mean solar interval. In order to still further simplify the calculations, tables have been constructed ; in most cases, these give the quantity to be added or subtracted according as we are changing from mean to sidereal, or from sidereal to mean time. Off TIME. 131 174. To find the sidereal time at a given instant of mean solar time on a given date at Greenwich. The Nautical Almanack* gives the sidereal time of mean noon at Greenwich on every day of the year. Now the given mean time represents the number of hours, minutes, and seconds which have elapsed since mean noon, expressed in mean time. Convert this interval into sidereal time ; we then have the sidereal interval which has elapsed since mean noon. Add this to the sidereal time of mean noon ; the result is the sidereal time required. Thus, let m he the mean time at the given instant, mea- sured from the preceding mean noon, * the sidereal time of mean noon from the Nautical Almanack, and let k = l/365 ; so that l+ is the ratio of a mean solar unit to the corresponding sidereal unit. Then, from mean noon to given instant, Interval in mean time = m .*. interval in sidereal time = m+lan But, at mean noon, sidereal time = s .*. at given instant, required sidereal time, s-=s Q +m+km. If the result he greater than 24h., we must subtract 24h., for times are always measured from Oh. up to 24h. EXAMPLE. Find the sidereal time corresponding to 8h. 15m. 40s. P.M. on Dec. 20, given that the sidereal time of mean noon was I7h. 55m. 8s. From mean noon to the given instant, the interval in mean time is 8h. 15m. 40s. Converting this interval to sidereal time, by the method of 173, we have Mean solar interval = 8h. 15m. 40s. Add 10s. per hour on 8h. 1m. 20s. Add Is. per 6m. on 15m. 40s. 3s. 8h. 17m. 3s. Subtract Is. per 1m. on 1m. 23s. ls._ .*. Sidereal interval since mean noon = 8h. 17m. 2s. But sidereal time of mean noon = I7h. 55m. 8s. .*. Sidereal time at instant required = 26h. 12m. 10s. Or, deducting 24h., sidereal time is = 2h. 12m. 10s. * Or Whitaker's Almanack, which may be used if the Nautical is not at hand. 132 ASTEONOMf. 175. To find the mean solar time corresponding to a given instant of sidereal time at Greenwich. Subtract the sidereal time of mean noon from the given sidereal time ; this gives the interval which has elapsed since mean noon, expressed in sidereal time. Convert this interval into mean time ; the result is the mean time required. Let It = 1/366 J ; so that 1 k' is the ratio of a sidereal to a mean solar unit. Let the given sidereal time = , and let the sidereal time of the preceding mean noon = ; Then, from mean noon to given instant, Interval in sidereal time = s S Q .'. interval in mean time = (s s ) '(s s ). .'. required mean time m = (s * ) M> ). If s be less than , we must add 24h. to s in order that the times s, s may be reckoned from the same transit of T EXAMPLE. Find the solar time corresponding to 16h. 3m. 42s. sidereal time on May 6, 1891, sidereal time at mean noon being 2h. 52m. 17s. Sidereal interval since mean noon = 16h. 3m. 42s. -2h. 52m. 17s. = 13h. llm. 25s. /. Mean solar interval ( 173) = 13h. llm. 25s. -2m. 10s. 2s. + 2s. = 13h. 9m. 15s. Hence, 13h. 9m. 15s. is the mean time; which, in our usual reckoning, would be called Ih. 9m. 15s., on the morning of May 6 ( 36). The sidereal timo was also 16h. 3m. 42s. a sidereal day or 23h. 56m. 4s. previously, i.e., Ih. 13m. 11s. a.m. on the morning of May 5. % 176. To find the mean time corresponding to a given instant of sidereal time at Greenwich (alterna- tive method). The Nautical Almanack also contains the mean time of tl Sidereal Noon," i.e., the mean time when T is on the meridian, and when the sidereal clock marks Oh. Om. Os. Let this be m , and let s be the given sidereal time, k' the factor l/366 as before. Then From sidereal noon to given instant, sidereal interval = s ; .-. ,, ,, mean solar,, = s-k's. But, at sidereal noon, mean time = m ; /. at given instant, The required mean time = m Q +s Jc's. Oft TIME. 133 177. To find the sidereal time from the mean solar, or the mean time from the sidereal, in any given longitude. If the longitude is not that of Greenwich, the ahove methods will require a slight modification, because the sidereal time of mean noon and mean time of sidereal noon are tabulated for Greenwich. In such cases, the safest plan is as follows : Find the Greenwich time corresponding to the given local time ( 96). Convert this Greenwich time from mean to sidereal, or sidereal to mean, as the case may be, and then find the corresponding local time again. Let the longitude be L west of Greenwich (Z being nega- tive if the longitude is east), lot m l be the mean and s l the sidereal local time, m, s the corresponding times at Greenwich, and let /c, &', w> , have the same meanings as in 172-4. By 96 we have, whether the times be local or sidereal, (Greenwich time) (local time in long. Z W.) = T 1 - F Zh. = 4Z m. Therefore, s s l ^L = m m r (i.) If m l is given and s l is required, we have (in hours), By 174, 8 = 8H + s l = srfa = (ii.) If s 1 is given and m l is required, we have By 175, 176, m = (-* ) - #(-) or = i +-#, i.e. m *-*~^ s -* > + L EXAMPLE. Find the solar time when the local sidereal time is 5h. 17m. 32s. on March 21, the place of observation being Moscow (long. 37 34' 15" B.) ; given that sidereal time of mean noon was 23h. 54m. 52s. at Greenwich. Eeduced to time ( 23), 37 34' 15" is 2h. 30m. 17s. /. Greenwich sidereal time at instant required = 5h. 17m. 32s. -2h. 30m. 17s. = 2h. 47m. 15a. Sidereal interval since Greenwich noon = 2h. 47m. 15s. + 24h -23h. 54m. 52s. = 2h. 52m. 23s. .'. Greenwich mean time = 2h. 52m. 23s. -20s. -9s. = 2h. 51m. 54s. .'. Moscow mean time = 2h. 51m. 54s. + 2h. 30m. 17s. - 5h. 22m. 11s 134 AStKONOiTT. 178. Equinoctial Time. For the purpose of comparing the times of observations made at different places on the Earth, another kind of time has been introduced. The Equinoctial Time at any instant is the interval of time that has elapsed since the preceding vernal equinox, measured in mean solar units. The advantage of equinoctial time is that it is independent of the observer's position on the Earth, since the instant when the Sun passes through T is a perfectly definite instant of time, and is independent of the place of observation. On the other hand, mean time and sidereal time, being measured from the transits of the mean Sun and of T across the meridian, depend on the position of the meridian that is, on the longitude of the observer. The chief disadvantage of equinoctial time is that since the tropical year contains 365d. 5h. 48m. 46s., and not exactly 365 days, the vernal equinox will occur 5h. 48m. 46s. later in the day every year, so that at the end of each tropical year the equinoctial clock will have to be put back 5h. 48m. 46s. Hence also the same equinoctial time will represent a different time of day on the same date in different years. The disadvantages of using local time are obviated in Great Britain by the universal use of " Greenwich Mean Time." 179. Practical Applications. In 41 we showed how to determine roughly the time of night at which a given star would transit on a given day of the year. "With the intro- duction of mean time, in the present chapter, we are in a position to obtain a more accurate solution of the problem. Por the R.A. of any star (expressed in time) is its sidereal time of transit. If this be given, we only have to find the corresponding mean time ; this will be the required time of transit, as indicated by an ordinary clock. In the calculations required in converting the time from one measure to the other, it is advisable not to quote the formula? of 174-177, but to go through the various steps one by one. If neither the sidereal time of mean noon nor the mean time of sidereal noon is given, we must fall back on the rough method of 35. ON TIME. 135 EXAMPLES. 1. Find the solar time at 5h. 29m. 28s. sidereal time on July 1, 1891 ; mean time of sidereal noon being 17h. 20m. 8s. Sidereal interval from sidereal noon to the given instant = 5h.29m.28s. .-. Mean solar interval = 5h. 29m. 28s. 50s. 5s. + Is. = 5h.28m.34s. i.e., Mean solar time = 5h. 28m. 34s. + I7h. 20m. 83. =22h. 48m. 42s. ; or, lOh. 48m. 42s. A.M., July 2. It was also 5h. 29m. 28s., a sidereal day or 23h. 56m. 4s. pre- viously, i.e., lOh. 52m. 38s. a.m. July 1. ^ 2. To find the mean time of transit of Aldelaran at Greenwich on December 12, 1891. Given H M s R.A. of Aldelaran = 4 29 40 ; Sidereal time of noon, December 12, 1891 = 17 23 56. Since the star's R.A. is less than the sidereal time of noon, we must increase the former by 24h., in order that both may be mea- sured from the same " sidereal noon." H. M. s. Sidereal time of transit + 24h. = 28 29 40 Subtract noon = 17 23 56 /. Sidereal interval from noon to transit =11 5 44 To convert into mean solar units, subtract 1 49 .'. Mean Solar interval from noon to transit =11 3 55 .'. Aldelaran transits at llh. 3m. 55s. mean time. 7~&~*ii- 3. To find the (local) sidereal time at New York at 9h. 25m. 31g. (local mean time) on the morning of September 1, 1891. Longitude of New York = 74 W. Sidereal time of mean noon at Greenwich, Sept. 1 = lOh. 42m. 24s. The given local mean time is measured from midnight, therefore we must take the time measured from noon as H. M. s. August 31, 1891. = 21 25 31 Add for 74 west longitude reduced to time = 4 56 .*. Greenwich mean time is, August 31, 26 21 31 or, September 1, 2 21 31 To convert this interval to sidereal units, add 24 /. Sidereal time elapsed since Greenwich noon = 2 21 55 But at Greenwich noon, sidereal time (by data) = 10 42 24 .-. Sidereal time at Greenwich is 13 4 19 Subtract for 74 west longitude, 4 56 .-. Sidereal Time at New York = 8 18 9 136 ASTRONOMY. 4. To find the Paris mean time of transit of Regulus at Nice on December 26, 1891. H. M. s. Longitude of Paris = 2 21' E.E.A. of Regulus =10 2 34 Nice = 7 18' E. Sidereal time at Greenwich noon = 18 18 48 Here local sidereal time of transit at Nice 10 2 34 Subtract east longitude of Nice, T 18', in time 29 12 V. Greenwich sid. time of transit at Nice -I- 24h. C3 33 22 Subtract Greenwich sidereal time at noon, 18 18 48 /. Sidereal interval since Greenwich noon 15 14 34 To convert to mean solar units, subtract 2 30 .'. Greenwich mean time = 15 12 4 Add east longitude of Paris, expressed in time = 9 24 .-. Paris mean time of transit = 15 21 28 That is, 3h. 21m. 28s. in the morning on December 27. 5. Find the E.A. of the Sun at true noon on October 8, 1891, given that the equation of time for that day is 12m. 24s., and that the sidereal time of mean noon on March 21 was 23h. 54m. 52s. Mean solar interval from mean noon March 21 to mean noon Oct. 8 = 201 days. Mean solar interval from mean noon to apparent noon on Oct. 8 = -12m. 24s. /. interval from mean noon on March 21 to apparent noon on Oct. 8 = 201d.-12m. 24s. Now, in 365 days the mean Sun's E.A. increases 24h., and the increase takes place quite uniformly. .'. increase in mean Sun's E.A. in 201 days H. M. s. = 24h.x 201 -=-365| = 13 12 27 Add mean Sun's E.A. on March 21 ( = sidereal time of mean noon) = 23 54 52 .'. mean Sun's E.A. at mean noon Oct. 8 = 37 7 19 or, subtracting 24h., =13 7 19 Subtract change of E.A. in 12m. 24s. = 2 .'. mean Sun's E.A. at apparent noon Oct. 8 =13 7 17 But true Sun's E.A. mean Sun's E.A = equation of time = 12 24 /. True Sun's E.A. at apparent noon Oct. 8 = J2h. 54m. 53s, ON TIME. 137 EXAMPLES. -V. 1; To what angles do Sidereal Time, Solar Time, and Mean Time correspond on the celestial sphere ? Are these angles measured direct or retrograde ? 2. Draw a diagram of the Equation of Time, on the supposition that perihelion coincides with the vernal equinox. 3. On May 14 the morning is 7'8 minutes longer than the after- noon : find the equation of time on that day. 4. On a sun-dial placed on a vertical wall facing south, the position of the end of the shadow of a gnomon at mean noon is marked on every day of the year. Show that the curve passing through these points is something like an inverted figure of eight. 5. Why are not the graduations of a level dial uniform ? Show that they will be so if the dial be fixed perpendicular to the index. 6. Show that if every 5th year were to contain 366 days, every 25th year 367 days, and every 450th year 368 days, the average length of the civil year would be almost exactly equal to that of the tropical year. How many centuries would have to elapse before the difference would amount to a day ? 7. Give explicit directions for pointing an equatorial telescope to a star of R.A. 22h., declination 37 N., in latitude 50 N., longitude 25 E., at lOh. Greenwich mean time, when the true Sun's E.A. is 14h. 47m. 17s., and the equation of time is 16m. 14s. 8. If the mean time of transit of the first point of Aries be 9h. 41m. 24*4s., find the time of the year, and the sidereal time of an observation on the same day at Ih. 22m. 13'5s. 9. At Greenwich, the equation of time at apparent noon to-day is - 3m. 39'42s., and at apparent noon to-morrow it will be 3m. 35'39s. Prove that the mean solar time at New York corresponding to ap- parent time 9 A.M. there this morning is 8h. 56m. 2O9s., having given that the longitude of New York is 74 I' W. 10. Find the sidereal time at apparent noon on Sept. 30, 1878, at Louisville ( long. 85 30' W.) having given the following from the Nautical Almanack : At mean noon. Sun's apparent right ascension. Sept. 30. 12h. 26m. 23'16s. Oct. 1. 12h. 30m. 0'51s. Equation of time to be added to mean time. 10m. 0'77s. 10m. 19-98s, 138 ASTBONOMT. MISCELLANEOUS QUESTIONS. 1. Explain how to determine the position of the ecliptic relatively to an observer in S. latitude at a given time on a given day. 2. Indicate the position of the ecliptic relatively to an observer at Cape Town (lat. 33 56' 3'5" S.) at noon on August 3. 3. Explain why a day seems to be gained or lost by sailing round the world. State which way round a day seems to be lost, and give the reason why. 4. If the inclination of the ecliptic to the equator were 60, instead of 23 27^', describe what would be the variations in the seasons to an observer in latitude 45, illustrating your description with a diagram. 5. Describe the changes of position in the point of the Sun's rising at different times of the year, and at different points on the Earth's surface. 6. If the equator and ecliptic were coincident, what kind of curve would be described in space by a point on the Earth's surface, say at the equator, daring the course of the year ? 7. Examine when that part of the equation of time due to the eccentricity of the Earth's orbit is positive. 8. On September 22, 1861, the times of transit of a Lyrse and of the Sun's centre over the meridian of Greenwich were observed to be 18h. 32m. 51'3s. and 12h. Om. 23'3s. by a sidereal clock whose rate was correct. Given that the R. A. of a Lyrae was 18h. 31m. 43'9s., find the Sun's B..A. and the error of the clock. 9. Define mean time and sidereal time, and compare the lengths of the mean second and the sidereal second. 10. If a, a' are the hour angles in degrees of the Sun at Greenwich, at t and t' hours mean time, show that the equations of time at the preceding and following mean noons, expressed in fractions of an hour, are respectively a't-at' 21 .X(24-Q-a(24 f) oir TIME. EXAMINATION PAPER. v. 1. Define tiie dynamical mean Sun and the mean Sun, stating at what points they have the same R.A., and when the former coin- cides with the true Sun. Show that the mean Sun has a uniform diurnal motion, and state how it measures mean time. 2. Define the equation of time. Of what two parts is it generally taken to consist? State when each of these parts vanishes, is positive, or negative. Give roughly their maximum values, and sketch curves showing their variations graphically. 3. Show that the equation of time vanishes four times a year. 4. If, on a certain day, the sun-dial be 10 minutes before the clock, what is the value of the equation of time on that day ? Will the forenoon of that day or the afternoon be longer, and by how much ? 5. Define the terms solar day, mean solar day, sidereal day. What is the approximate difference and the exact ratio of the second and third ? 6. Define the terms civil year, anomalistic year, equinoctial time. Why was this last introduced ? 7. Show how to express mean solar time in terms of sidereal time, and vice versd. 8. If the mean Sun's R.A. at mean noon at Greenwich on June 1 be 4h. 36m. 54s., find the sidereal time corresponding to 2h. 35m. 45s. mean time (1) at Greenwich, (2) at a place in longitude 25 E. 9. On what day of the year will a sidereal clock indicate lOh. 20m. at 4 P.M. ? .10. In what years during the present century have there been five Sundays in February ? When will it next happen ? ASTBON. L CHAPTER VI. ATMOSPHERICAL KEFBACTION AND TWILIGHT. 180. Laws of Refraction. It is a fundamental prin- ciple of Optics that a ray of light travels in a straight line, so long as its course lies in the same homogeneous medium ; but when a ray passes from one medium into another, or from one stratum of a medium into another stratum of dif- ferent density, it, in general, undergoes a change of direction at their surface of separation. This change of direction is called Refraction.* Letarayof light S0(Fig. 64) pass at from one medium into another, the two media being separated by the plane surface AB, and let OT be the direc- tion of the ray after refraction in the second medium. Draw ZOZ' the normal or perpendicular to the plane AB at 0. Then the three laws of refraction may be stated as follows : I. Thfrancident and refracted rays SO, OTand the normal ZOZ 1 all lie in one Diane. . . is a constant quantity, being tlie same for all directions of the rays, so long as the two media are the same.] This constant ratio is called the relative index of refraction of the two media, and is usually denoted by the Greek letter fi. * For a fuller description, see Stewart's Light, Chap. YI. f The value of the ratio varies slightly for rays of different colours, but with this we are not concerned in the present chapter. ATMOSPHERICAL REFRACTION AND TWILIGHT. 141 Thus, if TO be produced backwards to S', sin ZOS = /i sin Z' OT = p sin ZOS r , The angles ZOS and Z' OT are usually called the angle of incidence and the angle of refraction respectively. III. When light passes from a rarer fo a denser medium, the angle of incidence is greater than the angle of refraction. Since Z ZOS> L Z'OT, sin ZOS > sinZ'OT and /. /i > 1. 181. General Description of Atmospherical Refrac- tion. If the Earth had no , atmosphere, the rays of light proceeding from a celestial body would travel in straight Lines right up to the obser- ver's eye or telescope, and we should see the body in its actual direction. But when a ray Sa (Fig. 65) meets the uppermost layer AA' of the Earth's atmo- sphere, it is refracted or bent out of its course, and its direc- FlG - tion changed to aft. On passing into a denser stratum of aiv at BB', it is further bent into the direction be, and so on ; thus, on reaching the observer, the ray is travelling in a direction OT, different from its original direction, but (by Law I.) in the same vertical plane. The body is, therefore, seen in the direction OS', although its real direction is aS or OS. Also, since the successive horizontal layers of air A A', BB', CC', ... are of increasing density, the effect of refraction is to bend the ray towards the perpendicular to the surfaces of separation, that is, towards the vertical. Hence : The apparent altitudes of the stars are increased by refraction. In reality, the density of the atmosphere increases gradually as we approach the Earth, instead of changing abruptly at the planes A A', BB', .... Consequently, the ray, instead of describing the polygonal path Sabc 0, describes a curved path, but the general effect is the same. 142 ASTRONOMY. 182. Law of Successive Refractions. Let there be any number of different media, separated by parallel planes JLA', ', CO', HH' (Fig. 66), and let Sale OT represent the path of a ray as refracted at the various surfaces. Then it is a result of experiment that the final direction S'T of the ray is parallel to what it would have been if the ray had been refracted directly from the first into the last medium without traversing the intervening media. Thus, if a ray SO, drawn parallel to Sa, were to pass directly from the first medium to the last by a single refrac- tion at 0, its refracted direction would be the same as that actually taken by the ray Sa, and would coincide with OT. S' FIG. 66. FIQ. 67. 183. The Formula for Astronomical Refraction. "We shall now apply the above laws to determine the change in the apparent direction of a star produced by refraction. Since the height of the atmosphere is only a small fraction of the Earth's radius, it is sufficient for most purposes of approximation to regard the Earth as flat, and the surfaces of equal density in the atmosphere as parallel planes. "With this assumption, the effect of refraction is exactly the same ( 182) as if the rays were refracted directly into the lowest stratum of the atmosphere, without traversing the intervening strata. ATMOSPHERICAL EEFEACTION AND TWILIGHT. 143 Let OS (Fig. 67) be the true direction of a star or other celestial body. Then, before reaching the atmosphere, the rays from the star travel in the direction SO. Let their direction after refraction be S'OT, then OS' is the apparent direction in which the star will be seen, and the angle SOS' is the apparent change in direction due to refraction. The normal OZ points towards the zenith. Hence ZOS is the star's true zenith distance, and ZOS 1 or Z'OT is its apparent zenith distance, and the first and third laws of refraction show that the star's apparent direction is displaced towards the zenith. Let L ZOS' = 8, tS'OS = u, and .-. L ZOS = z + u ; and let /j be the index of refraction. By the second law of refraction, sin (a -f u) = p sin s. sin 2 cos w-j-cos s sin u = yu sin z. Now the refraction u is in general very small. Hence,' if u be measured in circular measure, we know by Trigonometry that sin u = , and cos u = 1 very approximately. Therefore we have sin 2 -f- u cos 8 = (j. sin 8 ; Let U be the amount of refraction in circular measure when the zenith distance is 45. Putting s 45, we have .-. u = Thus the amount of refraction is proportional to the tangent of the apparent zenith distance. The last result does not depend on the fact that the refrac- tion is measured in circular measure. Hence, if w", U" be the numbers of seconds in u, U, we have u" = U" tan a. The quantity U" is called the coefficient of refraction. Since U is the circular measure of Z7", we have V" = 180X60X6 . V= 206265 (,,-1), 7T whence, if U" is known, p cm be found, and conversely, 144 ASTEONOMT. 184. Observations on the preceding Formula. In the last formula u" represents the correction which must be added to the apparent or observed zenith distance in order to obtain the true zenith distance. By the first law, the azimuth of a celestial body is unaltered by refraction. Thus the time of transit of a star across the meridian, or across any other vertical circle, is unaltered by refraction. In using the transit circle, there will, therefore, be no cor- rection for observations of right ascension, but in finding the declination the observed meridian Z.D. will require to be increased by U" tan z. A star in the zenith is unaffected by refraction, and the correction increases as the zenith distance increases. When a star is near the horizon, the formula u" = U" tan z fails, since it makes u" = co, when 2 = 90. In this case u is no longer a small angle, so that we are not justified in putting sin u = u and cos u = 1. But there is a more important reason why the formula fails at low altitudes, namely, that the rays of light have to traverse such a length of the Earth's atmo- sphere that we can no longer regard the strata of equal density as bounded by parallel planes. In this case, it is necessary to take into account the roundness of the Earth in order to obtain any approach to accurate results. For zenith distances less than 75, the formula is found to give fairly satisfactory results ; for greater zenith distances it makes the correction too large-. The coefficient of refraction U" is found to be about 57", when the height of the barometer is 29 -6 inches and the temperature is 50. But the index of refraction depends on the density of the air, and this again depends on the pressure and temperature. Hence, where accurate corrections for refraction are required, the height of the barometer and thermometer must be read. Any want of uniformity in the strata of equal density, or any uncertainty in determining the temperature, will introduce a source of error ; hence it is desirable that the corrections shall be as small as possible. For this reason observations made near the zenith are always the most reliable, ATMOSPHERICAL EEFEACTION AND TWILIGHT. 145 *185. Cassini's Formula. The law of refraction was also investi- gated by Dominique Cassini on the hypothesis that the atmosphere is spherical but homogeneous throughout ; in this way he obtained the approximate formula u = (ju1) tans (1 n sec 2 s), where n is the ratio of the height of the homogeneous atmosphere to the radius of the Earth. Cassini's formula may be proved as follows : Let SO'O be the path of a ray of light from a star 8. By hypothesis this ray undergoes a single refraction on entering the homo- geneous atmosphere at 0'. Let be the position of the observer, G the centre of the Earth. Produce 00' to 8', CO to Z, and GO' to Z'. Let u = L SOS' (in circular measure), Then, by 183, if u is small, we have u = (jti 1) tans'; but here z' is not the apparent zenith distance, so that we must express tan z' in terms of tan z. Draw CT perpendicular to O'O pro- duced, and O'N perpendicular to COZ. Then O'T tans' = TG = OTtansj tan_s_ = OT = 1 + ^0 tans'" Or OT OIV sec z __ , O# gec 2 00 = 1 + ' FIG. 68. 00 cos z But ON is very approximately the height of the homogeneous atmosphere OH, and is therefore = n . OG ; tan 2 . , tans tans' = 1 + n sec 2 z ; tans' 1 + Ti sec 2 z whence, by substituting in the formula, we have , .. , tan z 1 4 n sec 2 z = (fj. 1) tans {l wsec 2 s + 'n- 2 sec 4 s n 3 sec 6 s, &c.} Now n is very small ; we may therefore neglect its square and higher powers; hence we obtain approximately u= (/* !) tan z (1Ti sec- s), which is Cassini's formula. If the value of n be properly chosen, Cassini's formula is found to give very good results for all zenith distances up to 80. 146 . A8TBONOMY. 186. To determine the Coefficient of Refraction from Meridian Observations. Assuming the " tangent law," u= 7"tanz, the coefficient of refraction U may be found from observations of circumpolar stars as follows. Let Z D z 2 , the apparent zenith distances of a circumpolar itar, be observed at upper and lower culminations respectively. Then the true zenith distances will be %i -f 7~tan z l and z 2 + 7"tan z 2 . Now, the observer's latitude is half the sum of the meri- dian altitudes at the two culminations ( 28), hence if I be the latitude, we have or 90-? = i(2 1 + 2 2 ) + J^(tanz 1 +tanE 2 ) ...... (i.). Now let a second circumpolar star be observed. Let its apparent zenith distances at upper and lower culminations be z' and % '. Then we obtain in like manner 90-Z = i (a'+a") + i T(tan z' + tan 2") ...... (ii.). Eliminating I from (i.) and (ii.) by subtraction, we have (tan Zj + tan z 2 ) (tan z' + tan z")* If the two stars have the same declination, we shall have Zj = z and z 2 = z", and the above formula will fail. Hence it is important that the two observed stars should differ con- siderably in declination ; the best results are obtained by selecting one star veiy near the pole (e.g-, the Pole Star) and the other about 30 from the pole. 187. Alternative Method (Bradley's). Instead of using a second circumpolar star, Bradley observed the Sun's apparent Z.D.'sat noon at the two solstices. Let these be Z v Z y By 38, since the true Z.D.'s are Z^ + 7tan Z t and Z^ + U tan Z v Z^ + Z7tan Z l = l- i, Z^ + Vtan Z z = l + i; (i = obliquity.) .-. 2Z = ' 1 + 4+7(tan;+tan 2 ) ......... (iii.). Eliminating ?from (i.), (iii.), we have ^hence 7" is found. UNIVERSITY TWILIGHT,. 147 188. Other Methods of finding the Refraction. Suppose that at a station on the Earth's equator, either a star on the celestial equator, or the Sun at an equinox, is observed during the day. Its diurnal path from east to west passes through the zenith, and during the course of the day its true zenith distance will change uniformly at the rate of 15 per hour. Thus the true Z.D. at anytime is known. Let the apparent Z.D. be observed with an altazi- muth. The difference between the observed and the calcu- lated Z.D. is the displacement of the body due to refraction. By this method we find the corrections for refraction at different zenith distances without making any assumptions regarding the law of refraction. Except at stations on the Earth's equator, it is not possible to observe the refraction at different zenith distances in such a simple manner. Nevertheless, methods more or less similar can be employed. For this purpose the zenith distances of a known star are observed at different times. The true zenith distance at the time of each observation can be calculated from the known R.A. and declination ( 26). Hence the refraction for different zenith distances of the star can be determined. This method is very useful for verifying the law of refraction after the star's declination and the observer's latitude have been found with tolerable accuracy. Moreover, it can be employed to find the corrections for refraction at low altitudes when the " tangent law " ceases to give approximate results. 189. Tables of Mean Refraction. From the results of such observations tables of mean refraction have been con- structed by Bessel,* and are now used universally. These are calculated for temperature 50 and height of barometer 29*6 inches ; they give the refraction for every 5' of altitude up to 10, for larger intervals at altitudes between 10 and 54, and for every 1 at altitudes varying from 54 to 90. Other tables give the "Correction for Mean Refraction," which must be added to or subtracted from the mean refrac- tion given in the first table in allowing for differences in the temperature and barometric pressure. The corrections for temperature and pressure are applied separately. * See any book of Mathematical Tables, such as Chambers'^. 148 ASTBOffOMY. 190. Effects of Refraction on Rising and Setting. At the horizon the mean refraction is about 33' ; con- sequently a celestial body appears to rise or set when it it is 33' below the horizon. Thus, the effect of refraction is to accelerate the time of rising, and to retard, by an equal amount, the time of setting of a celestial body. In particular, the Sun, whose angular diameter is 32', appears to be just above the horizon when it is really just below. The acceleration in the time of rising due to refraction can be investigated in exactly the same way as the acceleration due to dip ( 104). If u" denotes the refraction at the hori- zon in seconds, d the declination, x the inclination to the vertical of the direction in which the body rises, the accelera- tion in the time of rising in seconds = u" sec x sec d. lo Taking the horizontal refraction as 33', or 1980", and putting x = 0, d = 0, we see that at the Earth's equator at an equinox, the time of sunrise is accelerated by about 2m. 12s. owing to refraction, "When the Sun or Moon is near the horizon, it appears distorted into a somewhat oval shape. This effect is due to refraction. The whole disc is raised by refraction, but the refraction increases as the altitude diminishes ; so that the lower limb is raised more than the upper limb, and the vertical diameter appears contracted. The horizontal dia- meter is unaffected by refraction, since its two extremities are simply raised. Hence, the disc appears somewhat flat- tened or elliptical, instead of truly circular. According to the tables of mean refraction, the refraction on the horizon is 33', while at an altitude 30', the refraction is only 28' 23", and at 35' it is 27' 41". Hence, taking the Sun's or Moon's diameter as 32', the lower limb when on the horizon is raised about 5' more than the upper. The con- traction of the vertical diameter, therefore, amounts to 5', i.e., about one-sixth of the diameter itself, so that the appa- rent vertical and horizontal angular diameters are approxi- mately in the ratio of 5 to 6. ATMOSPHERICAL REFRACTION AND TWILIGHT. 149 191. Illusory Variations in Size of Sun and Moon. The Sun and Moon generally seem to look larger when low down than when high up in the sky. This is, however, merely a false impression formed by the observer, and is not in accordance with measurements of the angular diameter made with a micrometer. When near the horizon, tho eye is apt to estimate the size and distance of the Sun and Moon by comparing them with the neighbouring terres- trial objects (trees, hills, &c.). When the bodies are at a considerable altitude no such comparison is possible, and a different estimate of their size is instinctively formed. 192. Effect of Refraction on Dip, and Distance of the Horizon. Since refraction increases as we approach the Earth, its effect is always to bend the path of a ray of light into a curve which is concave downwards (Fig. 69). FIG. 69. Let be any point above the Earth's surface, and let T' be the curved path of the ray of light which touches the Earth at T' and passes through 0. Then OT' is the distance of the visible horizon. Draw the straight tangent OT, then OT would be the distance of the visible horizon if there V^ere no refraction; hence, it is evident from the figure that The Distance of the horizon is increased by refraction. Draw OT", the tangent at to the curved path OT', then OT" is the apparent direction of the horizon. Hence, from the figure we see that The Dip of the horizon is diminished by refraction. Both dip and distance are still approximately proportional to the square root of the height of the observer. 150 ASTBONOMY. 193. Effect of Refraction on Lunar Eclipses and on Lunar Occupations. In a total eclipse the Moon's disc is never perfectly dark, but appears of a dull red colour. This effect is due to refraction. The Earth coming between the Sun and Moon prevents the Sun's direct rays from reach- ing the Moon, but those rays which nearly graze the Earth's surface are bent round by the refraction of the Earth's atmosphere, and thus reach the Moon's disc. From observing the " occultations " of stars when the unilluminated portion of the Moon passes in front of them, we are enabled to infer that the Moon does not possess an atmosphere similar to that of our Earth. For the directions of stars would be displaced by the refraction of such an atmosphere just before disappearing behind the disc, and just after the occultation ; and no such effect has been observed. 194. Twilight. The phenomenon of twilight is also due to the Earth's atmosphere, and is explained as follows : After the Sun has set, its rays still continue to fall on the atmosphere above the Earth, and of the light thus received a considerable portion is reflected or scattered in various directions. This scattered light is what we call twilight, and it illuminates the Earth for a considerable time after sunset. Moreover, some of the scattered light is transmitted to other particles of the atmosphere further away from the Sun, and these reflect the rays a second time ; the result of these second reflections is to further increase the duration of twilight. Twilight is said to end when this scattered light has entirely disappeared, or has, at least, become imperceptible. From numerous observations, twilight is found to end when the Sun is at a depth of about 18 below the horizon. If the Sun does not descend more than 18 below the horizon, there will be twilight all night. Let I = latitude, d = Sun's declination, then it is easily seen by a figure that the Sun's depth below the horizon at midnight = 90 dl. This depth is less than 18, if I > 72 an arc of 25 x 10'. .*. from 26 20' on limb to of vernier, represents an arc of 25 x 10' -25 x (10' -10") ; i.e., 25 x 10", or 4' 10". Hence the zero mark of the vernier scale is at a distance 26 20' + 4' 10" from the zero on the limb, and the reading is 26 24' 10". f 199. The Errors of the Sextant need not be described in detail. If the sextant does not read zero when the two mirrors are parallel, it is said to have an Index Error, and a constant correction for index error must be added to all readings made with the instrument. There are also errors due to eccentricity or want of coincidence between the centre about which the index bar turns, and the centre of the limb, errors of graduation, &c. 200. To determine the Index Error of the Sextant, In all goou sextants the graduated limb is continued backwards for about 5 behind the zero point. This portion of the limb is called the "arc of excess," and is used for finding the index error, as follows. The Sun or full Moon is observed ; the two images of its disc are brought into contact. Let e be the index-error, r the sextant reading, D the angular diameter of the disc, then we have evidently D = r + e. Now let the index bar be moved along the arc of excess until the images again touch, the image which was before uppermost being undermost. If the reading on the arc of excess be r', we have now D = r' + e, or D = / e. Hence, 2e = r'r. f The simpler forms of mercurial barometer are provided with a vernier by means of which the height of the mercury is read off to the nearest hundredth of an inch. The student will find it of great assistance to carefully examine the vernier in such an instrument 158 ASTEONOMY. 201. To take altitudes at Sea by the Sextant. The principal use of the sextant is for finding altitudes. Now the altitude of a star is its distance from the nearest point of the celestial horizon. To find this, the sextant is so adjusted that the reflected image of the star appears to lie on the ofiing or visible horizon ; when the plane of the sextant is slightly turned, the image of the star should just graze the horizon without going below it. The sextant reading then gives the star's angular distance from the nearest point of the "offing." Subtract the dip of the horizon and the correc- tion for refraction, both of which are given in books of mathematical tables. The star's true altitude is thus obtained. 202. To take the Altitude of the Sun or Moon. In observing the Sun's altitude, the " index " shades must be turned into position between the two mirrors, and the instru- ment adjusted so that the Sun's lower limb appears just to graze the horizon. The reading of the sextant, when corrected for dip and refraction, gives the altitude of the Sun's lower limb. Add the Sun's angular semi-diameter, which is given in the Nautical Almanack ; the altitude of the Sun's centre is then obtained. Both the Sun's altitude and its angular diameter may be obtained by observing the altitudes of the upper and lower limbs. The difference of the two corrected readings gives the Sun's angular diameter, and half the sum of the readings gives the altitude of the Sun's centre. If this method is used, allowance must be made for the change in the Sun's altitude between the observations. For this purpose, three observations must be made. First take the altitude of the Sun's lower limb, then of the upper limb, and lastly, again of the lower limb. Also note the time of each observation. The difference between the first and third readings determines the Sun's motion in altitude ; from this, by a simple proportion, the change in altitude between the first and second observations is found, and thus the alti- tude of the lower limb at the second observation is known. We can now find the Sun's angular diameter, and the altitude of its centre at the second observation. THE DETERMINATION OP POSITION ON THE EAETH. 159 Let , = time of 1st observation, when a = alt. of lower limb ; 3 = time of 2nd observation, when I = alt. of upper limb ; # 3 = time of 3rd observation, when a' = alt. of lower limb ; Then in time t s t v the alt. of lower limb increases a' a. .'. in time 2 ^ it increases (a a) x f ~. h~h Hence if 2 denote the alt. of lower limb at second observation, tn - t j 69 - TI This finds # 2 , and we then have Sun's angular diameter = l a. r Alt. of Sun's centre at second observation = In taking the altitude of the Moon, the altitude of the illuminated limb must be observed, an'l the angular semi- diameter, as given in the " Nautical Almanac," must be added or subtracted, according as the lower or upper limb is illuminated. 203. Artificial Horizon for Land Observations. Owing to the absence of a well-defined offing on land, an artificial horizon must be used. This is simply a shallow dish of mercury, protected in some manner from the disturbing effect of the wind. The sextant is used to observe the angular distance between a star and its image as reflected in the mercury. Half this angular distance is the star's apparent altitude ; correcting this for refraction, the true altitude is obtained (S = p-P-s. . . z >S^i, = p -P s -i- w- 226 [As an examplo, the student may show that the greatest latitudes the Moon can have, in ordei that it may bo partially or wholly within the penumbra at opposition arc p + s + P-f m and p + s + P m respectively.] *291. Greatesb Latitudes of the Moon at Syzygy. Since S and V are in the ecliptic, it follows that when the Moon is in conjunction or opposition, the plane of the paper in !Fig. 89 is perpendicular to the ecliptic. Therefore the angles VEM^ VEM. 2 measure the Moon's latitude at con- junction, and SJEm v SEm. 2 , SEm 3 measure its latitude at opposition in the positions represented. The above expres- sions are, therefore, the greatest possible latitudes at syzygy consistent with eclipses of the kinds named. Now, taking the mean values we have, roughly, s=l6'- m=l5'; p = 57' ; P = 0' 8". Substituting these values, and collecting the results, we have, roughly, the following limits for the Moon's geocentric lati- tude, or angular distance from the line of centres : (1) For a lunar eclipse, VEM^ =p+ P-s+m = 56'; (2) Fora total lunar eclipse, VEM^p + Psm = 26'; (3) For a solar eclipse, SEm^ =zp P+s + m 88'; (4) For an annular eclipse, SEm z = pP+s m = 58'. Lastly, taking the Sun at apogee, and the Moon at perigee, we have, m = 17' and s = 16' nearly, whence we have, in the most favourable case, (40) For a total solar eclipse, SEm^ -=p P 8+ m = 58'. 292. Ecliptic Limits. From the last results it appears that a lunar eclipse cannot occur unless at the time of oppo- sition the Moon's latitude is less than about 56', and that a solar eclipse cannot occur unless at conjunction the Moon's latitude is less than about 88'. Now the Moon's latitude depends on its position in its orbit relatively to the line of nodes ; hence there will be corresponding limits to the Moon's distance from the node consistent with the occurrence of eclipses. These limits are called the Ecliptic Limits. *The ecliptic limits may be computed as follows : Let the geocentric direction of the Moon's centre be represented on the celestial sphere by Jf. Let JV represent the node, ECLIPSES. 227 secondary to the ecliptic. [The ecliptic limit, strictly speak- ing, means the limit of NH measured along the ecliptic, and not that of NMJ\ Now the limit of latitude MH lias been calculated in the last paragraph for the different cases. Let this be denoted by I. Also let I be the inclination of the Moon's orbit to the ecliptic. Then in the spherical triangle NHM, right-angled at H t we have HM = I, and z JINM= /; both of these are known, hence NIL can be calculated. FIG. 90. For rough purposes it will be sufficient either to treat the small triangle ITNMas a plane triangle (Sph. Geom. 24), or to regard Mil as approximately the arc of a small circle, whose pole is JV. The first method gives .-. NH= I cot /. Or, adopting the second method, we have (Sph. Geom. 17) I = MH= z MNITx sin NH = /sin Nil-, .-. sin NH =1/1, whence the ecliptic limit Nil is found. EXAMPLES. 1. To find the Lunar Ecliptic Limit. For a lunar eclipse we have, by 291, I = 56'. Also, 1-5 roughly. Ilence = sin 11 (from table of natural sines) and the lunar ecliptic limit is about 11. 2. To find the Solar Ecliptic Limit. For a solar eclipse we have I = 88'. Hence, taking I = 5 as before, we have = sin 17, roughly, and the solar ecliptic limit is about 17. 228 ASTRONOMY. 293. Major and Minor Ecliptic Limits. Owing to the variations in the distances of the Sun and Moon their parallaxes and angular semi-diameters are not quite constant. Hence the exact limits of the Moon's latitude /, as calculated by the method of 291, are subject to small variations. This alone would render the ecliptic limits variable. But there is another cause of variation in the ecliptic limits, arising from the fact that Z the inclination of the Moon's orbit, is also variable, its greatest and least values being about 5 19' and 4 57'. The greatest and least values of the limits for each kind of eclipse are called the Major and Minor Ecliptic Limits. For an eclipse of the Moon the major and minor ecliptic limits have been calculated to be about 12 5' and 9 30' re- spectively at the present time. For an eclipse of the Sun the limits are 18 31' and 15 21' respectively. Thus a lunar eclipse may take place if the Moon, when full, is within 12 5' of a node; and a lunar eclipse must take place if the full Moon is within 9 30' of a node. Similarly, a solar eclipse may take place if the Moon, when new, is within 18 31', and a solar eclipse must take place if the new Moon is within 15 21' of a node. The mean values of the lunar and solar ecliptic limits are now 10 47' and 16 56'. But the eccentricity of the Earth's orbit is very slowly decreasing ; consequently the major limits are smaller and the minor limits larger than they were, say, a thousand years ago. 294. Synodic Revolution of the Moon's Nodes. An eclipse is thus only possible at a time when the Sun is within a certain angular distance of the Moon's nodes. Hence the period of revolution of the Moon's nodes, relative to the Sun, marks the recurrence of the intervals of time during which eclipses are possible. This period is called the period of a synodic revolution of the nodes. In 273 it was stated that the Moon's nodes have a retro- grade motion of about 19 per annum, more exactly 19 21'. In one year (365d.) the Sun, therefore, separates from a node by 360+ 19 21' or 379*35, hence it separates 360 in (360 x 365)/379-35 days, or about 346'62d. This, then, is the period of a synodic revolution of the node. EClttSES. 229 In a synodic lunar month (29| days), the Sun separates from the line of nodes by an angle 379jx29j-h365, or 30 36', a result which will be required in the next paragraph. 295. To find the Greatest and Least number of Eclipses possible in a Year. Let the circle in Fig. 9* represent the ecliptic, and let JVJ n be the Moon's nodes. Take the arcs NL, NL', til, nl' each equal to the lunar eclip- tic limit, and NS, JVb", ns, ns' each equal to the solar ecliptic limit. Then the least value of S3' or ss' is twice the minor solar ecliptic limit, and is 30 42', and this is greater than 30 36', the distance traversed by the Sun relative to the nodes between two new Moons. Hence, at least one new Moon must occur while the Sun is travel- ling over the arc SS', and two may occur. Therefore there must be one, and there may le two eclipses of the Sun, while the Sun is in the neighbourhood of a node. Again, the greatest value of LL', IV is double the major lunar ecliptic limit, and is, therefore, 24 10'. This is consider- ably less than the space passed over by the Sun relative to the nodes between Fio.TJi. two full Moons. Hence, there cannot be more than one full Moon while the Sun is in the arc LL\ and there may be none. Therefore there cannot le more than one eclipse of the Moon while the Sun is in the neighbourhood of a node, and there may be none at all. 296. The case most favourable to the occurrence of eclipses is that in which the Moon is new just after the Sun has come within the solar ecliptic limits, i.e., near S. There will then be an eclipse of the Sun. When the Moon is full (about 14| clays later) the Sun will be near N, at a point within the lunar ecliptic limits ; there will therefore be an eclipse of the Moon. At the following new Moon the Sun will not have reached S'-, and there will be a second eclipse of the Sun. In six lunations from the first eclipse the Sun will have travelled through just over 180, and will be within the space ss', near s ; there will therefore be a third eclipse of the Sun. 2,30 ASTROttOMt. At the next full Moon the Sun will be near , and there will be a second eclipse of the Moon. The Sun may just fall within the space 88 near *' at the next new Moon ; there will then be & fourth eclipse of the Sun. In twelve lunations from the first eclipse, the Sun will have described about 368, and will, therefore, be about 8 beyond its first position, and well within the limits ss' ; there will, therefore, be & fifth eclipse of the Sun. About 14f days later, at full Moon, the Sun will be well within the lunar ecliptic limits LL, and there will be a third eclipse of the Moon. All these eclipses occur in 12J lunations, i.e., 369 days, or a year and four days. "We cannot, therefore, have all the eight eclipses in one year, but There may be as many as seven eclipses in a year, namely, either five solar and tivo lunar, or four solar and three lunar. 297. The most unfavourable case is that in which the Moon is full just before the Sun reachesthe ecliptic limits at L. At new Moon the Sun will be near N, and there will be one solar eclipse. At the next full Moon the Sun will have passed L', so that there will be no lunar eclipse. After six .lunations the Sun will not have arrived at I. At the next new Moon the Sun will be within the ecliptic limits, and there will be a second solar i-flijM-e. At the next full Moon the Sun will be again just beyond I', and at 12 lunations from first full Moon, the Sun may again not have quite reached L. At 12 lunations there will be a third solar eclipse. The interval between the first and third eclipses will be 12 lunations, or about 354 days. If, therefore, the first eclipse occurs after the llth day of the year, i.e., January 11, the third will not occur till the following year. Therefore, The least possible number of eclipses in a year is two. These must both le solar eclipses. ECLIPSES. 231 298. The Saros of the Chaldeans. The period of a synodic revolution of the nodes is ( 294) approximately 346-62 days. Hence, 19 synodic revolutions of the node take 6585-78 days. Also 223 lunar months = 6585*32 days. It follows that after 6585 days, or 18 years 11 days, the Moon's nodes will have performed 19 revolutions relative to the Sun, and the Moon will have performed 223 revolutions almost exactly. Hence the Sun and Moon will occupy almost exactly the same position relative to the nodes at the end of this period as at the beginning, and eclipses will therefore recur after this interval. The period was discovered by observation by the Chaldean astronomers, who called it the Saros. By a knowledge of it they were usually able to predict eclipses. Indeed, in the records of eclipses handed down to us in the form of cuneiform inscriptions, they invariably stated whether the circumstances accorded with prediction by the Saros or not. A "synodic revolution of the Moon's apsides," or the period in which the Sun performs a complete revolution relative to the Moon's apse line, occupies 411-74 days. Hence sixteen such revolutions occupy 6587*87 days, or about two days longer than the Saros. Therefore the Moon's line of apsides also returns to very nearly the same position relative to the Sun and Moon. Hence, the solar eclipses, as they recur, will be nearly of the same kind (total or annular) in each Saros. The whole number of eclipses in a Saros is about 70. The average of all eclipses from B.C. 1207 to A.D. 2162 shows that there are 20 solar eclipses to 13 lunar. The present values of the mean solar and lunar ecliptic limits, 16 56', and 10 47', are in the ratio of 31 : 18 very nearly. This ratio gives, on the whole, a higher average proportion of solar eclipses to lunar than that given above. It must, however, be remembered that all the angles used in calcu- lating the limits are subject to gradual changes. Con- sequently the numbers of eclipses in that period aro subject to very gradual variation ; after a large number of Saroses have recurred, the order of eclipses in each will have changed, 232 ASTRONOMY. *SECTION III. Occultations Places at which a Solar Eclipse is visible. 299. Occultations. When the Moon's disc passes in front of a star or planet, the Moon is said to occult it. An occultation evidently takes place whenever the ap- parent angular distance of the Moon's centre from the star becomes less than the Moon's angular semi- diameter. As the apparent position of the Moon is affected by parallax, the cir- cumstances of an occultation are different at different places on the Earth's surface. FIG. Let m denote Moon's angular semi-diameter, p its horizontal parallax. In the figure, let E and M be the centres of the Earth and Moon, and let s C, sC' represent the parallel rays coming from a star, and grazing the Moon's disc. These rays cut the Earth's surface along a curve 00*, and it is evident that only to observers at points within this curve is the star hidden by the Moon's disc. Let EC, Es, EM, EC' cut the Earth's surface in c, x, m, c' ; the rays EC, EC' cut the Earth's surface in a small circle cc, whose angular radius mEc = MEC = m. Let d be the geocentric angular distance SEM between the Moon's centre and the star. Then the angle ECO = angle subtended by the Earth's radius EO at C ; = parallax of C when viewed from ; = ^sin COZ(\ 249); = p sin OEx (by parallels). But ECO = CEs ; = angle subtended by ex ; .% sin OEx = aD " Ic **'. P ECLIPSES. 233 Hence we have the following construction for the curve separating those points on the Earth's surface at which the occultation is visible at a given instant from those at which the star is not occulted. Taking the sublunar point m as pole, describe a circle cc on the terrestrial globe, with the Moon's angular semi-diameter (m) as radius. Through the sub- stellar point x draw any great circle, cutting this small circle in any point c. Measure along it an arc c such that sin c is always the same multiple f ) of me. The locus of the points 0, thus determined, is the curve required. Half of the circle cc' consists of points under the advancing limb of the Moon; hence, over the portion of the curve 00' corresponding to this half -circle, the occultation is just beginning. At points on the other half of cc the Moon's limb is receding ; hence over the other portion of 0' the star is reappearing from behind the Moon's disc. Since the greatest and least values of ex in any position are d+m and d m, it is evident that the greatest value of d for which an occultation can take is when d m=p; d=m+p. 300. Occnltation of a Planet. If s be a planet, the lines Es, Os can no longer be regarded as rigorously parallel; but the angle between them, Es 0, = angle subtended at s by the Earth's radius EO = parallactic correction at ( 248) = P sin ZOs ( 249) = P sin OEx very nearly. As before, EGO - p sin OEx. But ECO = EsO + CEs; .-. p sin OEx = P sin OEx + ex ; sin OEx = -^. pP With this exception, the construction is the same as for a star. If the planet be so large that we must take account of its angular diameter, the method of the next paragraph must be used. 234 ASTRONOMY. 301. Eclipse of the Sun. There is a total eclipse of the Sun, provided the Moon's disc completely covers the Sun's; this occurs if the Moon's angular semi-diameter (m) is larger than the Sun's (s), and the apparent angular distance between the Sun's and Moon's centres (as seen from any point at which the eclipse is visible) is less than m s. ilcnce, if the Moon's angular semi-diameter were reduced to m s, the Sun's centre would then be occulted. Hence the points 0, whose locus encloses the places from which the eclipse is visible, can be found as follows : With centre m the sublunar point, and angular radius m s, describe a circle. Through the subsolar point x draw any arc of a great circle xc, cutting the circle in 0, and take 0, on xc produced, such that xo p-p For an annular eclipse m < *, and the apparent angular distance between the centres is s m\ hence the same con- struction is followed, save that s m is the angular radius of the small circle first described. For a partial solar eclipse, the angular radius is s + m. When a planet has a sensible disc, the beginning of its occupation may be compared to a partial eclipse of the Sun ; and the planet is entirely occulted when the conditions are satisfied corresponding to those for a total eclipse. EXAMPLE. Supposing the centres of the Earth, Moon, and Sun to be in a straight line and the Moon's and Sun's semi-diameters to be exactly 17' and 16', to find the angular radii of the circles on the Earth over which the eclipse is total and partial respectively, taking the relative horizontal parallax as 57'. At those points at which the eclipse is total, the apparent angular distance between the centres, as displaced by parallax, must be not greater than 17' - 16', or 1'. Hence, since the centres are in a line, with the Earth's centre, the parallactic displacement must be not greater than 1'. Hence, if z be the Sun's zenith distance at the boundary, then 57' sin z = 1' ; .*. sin z = ?, or approximately cir- cular measure of z = -^j-. But a radian contains about 57 ; .*. -gV of a radian = 1 approx. Hence the eclipse is total over a circle of angular radius 1 about the sub-solar point. Similarly, the eclipse is partial if 57' sin z < 16' + 17', or 33', or sin < ff, or '58. From a table of natural sines, we find that sin- 1 '58 = 35^ roughly ; therefore the angular radius is 35 3 . ECLIPSES. 235 EXAMPLES ON ECLIPSES GENERALLY. 1. To find (roughly) the maximum duration of an eclipse of tho Moon, and the maximum duration of totality. From 291 we see that a lunar eclipse will continue as long as the Moon's angular distance from the line of centres of the Earth and Sun is less than 58', and the eclipse will continue total while the angular distance is less than 26'. Hence, the maximum duration of the eclipse is the time taken by the Moon to describe 2 x 58', or 116', and the maximum duration of totality is the time taken to describe 2 x 26', or 52'. Now the Moon describes 360 (relative to the direction of the Sun) in the synodic month, 29 -r days. Therefore, the times taken to describe 116' and 52'' respectively are 29^x116 , 29jj<52 d 360x60 360x60 ay8> ,. i.e. 3h. 48m. and Ih. 42m., and these are the maximum durations of the eclipse and of totality. The eclipse of Nov. 15, 1891, lasted 3h. 28m., and was total for Ih. 23m. 2. To calculate roughly the velocity with which the Moon's shadow travels over the Earth. (Sun's distance = 93,000,000 miles.) The radius of the Moon's orbit being about 240,000 miles, its cir- cumference is about 1,508,000 miles. Relative to tho line of centres, the Moon describes the circumference in a synodic month, i.e., about 29 days. Hence its relative velocity is about 1,508,000 -f- 29, or 51,000 miles per day, i.e., 2,100 miles per hour. If q 'denote the point where the middle of the shadow reaches the Earth (Fig. 88), and if the Earth's surface at q is perpendicular to Sq, wo have velocity of q : vel. of M = Sq : 8M = 93,000,000 : 93,000,000-240,000 - 1-0026 nearly. Hence the velocity of the shadow at q = vel. of M very nearly = 2,100 miles an hour. To find the velocity of the shadow relative to places on the Earth, we must subtract the velocity of the Earth's diurnal motion. This, at the Earth's equator, is about 1,040 miles an hour. Hence, if the Earth's surface and the shadow are moving in the same direction, the relative velocity is about 1,060 miles an hour. 3. To find the maximum duration of totality of the eclipse of the example on page 234, neglecting the obliquity of the ecliptic. The angular radius of the shadow being 1, or about 69 miles, its diameter is 139 miles. The obliquity of the ecliptic being neglected, the eclipse is central at a point on the equator, and the shadow and the Earth are therefore moving in the same direction with relative velocity 1,060 miles an hour (by Question 2). The greatest duration of totality is the time taken by the shadow to travel over a distance equal to its diameter, i.e., 139 miles, and is therefore 139 x 60/1000 minutes, i.e., 7'9 minutes (roughly). ASTKON. E 236 ASTEONOMT. EXAMPLES. IX. 1. If a total lunar eclipse occur at the summer solstice, and at the middle of the eclipse the Moon is seen in the zenith, find the latitude of the place of observation. 2. If there is a total eclipse of the Moon on March. 21, will the year b.e favourable for observing the phenomenon of the Harvest Moon? 3. Having given the dimensions and distances of the Sun and Moon, show how to find the diameter of the umbra where it meets the Earth's surface. 4. Calculate (roughly) the totality of a solar eclipse, viewed from the Equator at the Equinox, supposing Moon's diameter 2,1GO miles, Sun's diameter 400 times Moon's ; Distance of Moon from Earth 222,000 miles ; Distance of Sun from Earth 92,000,000 miles. 6. If 8 is the semi-diameter of the Sun, and p, P the horizontal parallaxes of the Sun and the Moon at the time of a lunar eclipse, show that to an observer on the Earth the angular radius of the Earth's shadow at the distance of the Moon is P + p S, and that of the penumbra P + p + S. Determine, also, the length of the shadow. 6. If the distance of the Moon from the centre of the Earth is taken to be 60 times the Earth's radius, the angular diameter of the Sun to be half a degree, and the synodic period of the Sun and Moon to be 30 days, show that the greatest time which can be occupied by the centre of the Moon in passing through the umbra of the Earth's shadow is about three hours, and explain how this method might be employed to find th ^ s the planet nearest the Sun, its dis- tance on the above scale being represented by 4. It is characterized by its small size, the great eccentricity of its elliptical orbit, amounting to about 3-, and the great inclina- tion of the orbit to the ecliptic, namely, about 7. The sidereal period of revolution round the Sun is about 88 of our days. Thus, Mercury's greatest and least distances from the Sun ore in the ratio of 1+i : 1 i (cf. 149), or 3:2. Professor Schiaparelli, of Milan, has found that Mercury rotates on its axis once in a sidereal period of revolution ; consequently it always turns nearly the same face to the Sun, like the Moon does to the Earth ( 276). Owing, however, to the great 'eccentricity of the orbit, the " libration in longitude " is much greater than that of the Moon, amounting to 47. Consequently, rather over one quarter of the whole surface is turned alternately towards and away from the Sun, three-eighths is always illuminated, and three-eighths is always dork. 240 ASTKONOMY. 306. Venus, ? , is the next planet, its mean distance from the Sun being represented by about 7 (really 7-2). Its orbit is very nearly circular, arid is inclined to the ecliptic at an angle of about 3 23'. Yenus revolves about the Sun in a period of 224 days. 307. The Earth, , comesnext, its mean distance being re- presented by 10, audits orbit very nearly circular (eccentricity ^i_). Its period of revolution in the ecliptic is 365|- days, and its period of rotation is a sidereal day, or 23h. 56m. mean time. It is the nearest planet to the Sun having a satellite (the Moon, ([ ), which revolves about it in 27|- days. 308. Mars, n - This method is, however, much modified by the fact that the real orbits are not circles, but ellipses. EXAMPLE 1. Given that the greatest elongation of Yenus is 45, find its distance from the Sun, that of the Earth being 93,000,000 miles. Here distance of Venus = 93,000,000 sin 45 = 93,000,000 x V^ = 93,000,000 x -70711 = 65,760,000 miles. EXAMPLE 2. Taking the Earth's distance as unity, to find the distance of Mercury, having given that Mercury's greatest elonga- tion is 22i. The distance of Mercury = Ixsin 22 = ^{1(1 -cos 45)} = ^(2-^/2) = -38268. THE PLAN 318. Changes in Elongation of a Superior Planet. Let us now compare the apparent motion of the superior planet 7with that of Sun. Since it revolves about the Sun in the same direction as the Earth does, but more slowly, the line SJwill move, relative to SE, in the opposite or retrograde direction. Hence, in considering the changes in the position of the planet relative to the Sun, we may regard SE as a fixed line, and J must then revolve about 8 in the circle ARBTwifh a retrograde motion, i.e., in the same direction as the hands of a watch.* At A the planet is in opposition with the Sun, and its elongation is 180. At B it is in conjunction, and its elongation is 0. If, however, we were to refer the directions of the Earth and planet to the Sun, the planet would be in heliocentric conjunction with the Earth at A, and in helio- centric opposition at B. The planet is nearest the Earth at A, and since its orbital Telocity is constant, its relative angular velocity is then greatest, and the elongation SEJ is decreasing at its most rapid rate. As the planet moves round from opposition A to FIG. 96. conjunction B, the elongation SEJ decreases continuously from 180 to 0. At R the elongation is 90, and the planet is said to be in quadrature. * As a simple illustration, both the hour and minute hands of a watoh revolve in the same directions, but the minute hand goe* faster and leaves the hour hand behind. Hence the hour hand separates from the minute hand in the opposite direction to that in which both tare moving. 248 ASTRONOMY". At conjunction, J?, the elongation is 0; and we may also consider it to be 360. As the planet revolves from B to A, the elongation (measured round in the direction BRA} de- creases from 360 to 180. At T the elongation is 270, and the planet is again seid to be in quadrature. At A the elongation is again 180, the planet being once more in opposition. After this the elongation decreases from 180 to as before, as the planet's relative position changes from A through R to B. The cycle of changes recurs in the synodic period, i.e., the period between two successive conjunctions or oppositions. "We see that the elongation decreases continually from 360 to as the planet revolves from conjunction round to con- junction, and there is no greatest elongation. FIG. 97. 319. To compare (roughly) the Distance of a Superior Planet with that of the Earth. Here there is no greatest elongation, and therefore we must resort to another method. Let the planet's elongation SEJ (Fig. 97) be observed at any instant, the interval of time which has elapsed since the planet was in opposition being also observed. Let this interval be , and let 8 denote the length of the planet's synodic period. Then, in time S the angle JSE increases from to 360 j therefore, if we assume the change to take place uniformly, the angle JSE at time t after conjunction is = 860 x tlS THE PLANETS. 249 Hence, JSE is known. Also JJES has been observed, and SJE (= 18QJJSSJSE) is therefore also known. Therefore we have, by plane trigonometry, Distance of Planet _ SJ _ sin SEJ Distance of Earth " 8E sin SJE which determines the ratio of the distances required. This method is also applicable to the inferior planets. It is, however, not exact, owing to the fact that the planetary motions are not really uniform (see 327). *320. It U not necessary to observe the instant of conjunction or opposition. If 8 is known, t\ro observations of the elongation and the elapsed time are sufficient to determine the ratio of the distances. The requisite formulea are more complicated, but they only involve plane trigonometry. We, therefore, leave their investigation as an exercise to the more advanced student. EXAMPLE. To calculate the distance of Saturn in terms of that of the Earth, having given that 94 days after opposition the elonga- tion of Saturn was 84 17', and that the synodic period is 376 days. Given also tan 5 43' = !. Let the Sun, Earth, and Saturn be denoted by fif, E, J. In 376 days / J8E increases from to 360. .*. in 94 days after opposition L JSE = 90 j also, by hypothesis, L JES = 84 17'. Distance of Saturn = SJ = ^ flj?/ = ^ 840 1? ,, Distance of Earth SE -r cot 6 43' = ~ = 10. Therefore the distance of Saturn, as calculated from the given data is 10 times that of the Earth. 321. The synodic period of an inferior planet may be found very readily by determining the time between two transits of the planet across the Sun's disc and counting the number of revolutions in the interval. For a superior planet this is not possible, and we must, instead, find the interval between two epochs at which the planet has the same elongation. 250 ASTRONOMY. 322. Relations between the Synodic and Sidereal Periods. The relation between the synodic and sidereal periods is almost exactly the same as in the case of the Moon, the only difference being that the planets revolve about the Sun and not about the Earth. The sidereal period of a planet is the time of the planet's revolution in its orbit about the Sun relative to the stars. The synodic period is the interval between two conjunc- tions with the Earth relative to the Sun. It is the time in which the planet makes one whole revolution as compared with the line joining the Earth to the Sun. Let S be the planet's synodic period, P its sidereal period, Yihc length of a year, that is, the Earth's sidereal period, all the periods being supposed measured in days. Then, in one clay, the angle described by the planet about the Sun 360/P, the angle described by the Earth = 360/F, and the angle through which their heliocentric directions have separated = 860//S. If the planet be inferior, it revolves more rapidly than the Earth, and 360/ represents the angle gained by the planet in one day. 360 360 360 If the planet be superior, it revolves more slowly than the Earth, and 360/ is the angle gained by iheJZarth i-i one day. 360 _ 360 360 ~^~ ~Y~ ~ ; or _i=i-JL. 3 1 P From these relations, the sidereal period can be found if the synodic period is known, and vice vtrsd. THE PLANETS. 323. Phases of the Planets. As the planets derive their light from the Sun, they must, like the Moon, pass through different phases depending on the proportion of their illuminated surface which is turned towards the Earth. Phases of an Inferior Planet. An inferior planet V will evidently be new at inferior conjunction A, dichotomized like the Moon at its third "quarter at greatest elongation 7", full at superior conjunction B, dichotomised like the Moon at first quarter when it again comes to greatest elongation at IT. Thus, like the Moon, it will undergo all the possible different phases in the course of a synodic revolution. There is, however, one important difference. As the planet revolves from A to B its distance from the Earth increases, and its angular diameter therefore decreases. Thus the planet appears largest when new and smallest when full, and the variations in the planet's brightness due to the differ- ences of phase arc, to a great extent, counterbalanced by the changes in the planet's distance. For this reason, Venus alters very little in its brightness (as seen by the naked eye) during the course of its synodical revolution. FIG. 98. The phase is determined by the angle 8 VE, and this is the angle of elongation of the Earth as it would appear from the planet. The illuminated portion of the visible surface of the planet at V is proportional to 180-F.#, and the proportion of the apparent area of the disc which is illumi- nated varies as 1 + cos S VE or 2 cos 2 \ S VE. ( Cf. 263). The phases of Venus are easily seen through a telescope. ASTROX. 252 ASTRONOMY. 324. Phases of a Superior Planet. For a superior planet J the angle SJE never exceeds a certain value. It is greatest when SJEJ = 90, being then the greatest elongation of the Earth as it would appear from the planet. Hence tho planet is always nearly full, being only slightly gibbous, and the phase is most marked at quadrature. FIG. 90. The gibbosity of Mars, though small, is readily visible at quadrature, about one-eighth of the planet's disc being obscured. The other superior planets are, however, at a distance from the Sun so much greater than that of the Earth that they always appear very approximately full. 325. The "Phases" of Saturn's Kings are due to an entirely different cause. The plane of the rings, like the plane of the Earth's equator, is fixed indirection, and inclined to the ecliptic at an angle of about 28. Hence, during the course of the planet's sidereal revolution, the Sun passes alternately to the north and south side? of the rings (just as in the phenomena of the seasons on our Earth, the Sun is alternately N. and S. of the equator). The Earth, which, relatively to Saturn, is a small distance from the Sun, also passes alterrately to the north and south sides of the rings, and we see the rings first on one side and then on the other. At the instant of transition the rings are seen edgewise, and are almost invisible. Unless Saturn is in opposition at this instant, the Sun and Earth, do not cross the plane of the rings simultaneously, and between their passages there is a B'.iort interval during which the Sun and Earth are on opposite sides of the plane; and the unilluminaled side of the rings is turned towards the Earth. The last " dis- appearances" of the rings occurred in Sept., 1891 May, 1892, but they occur twice in each sidereal period, or once about every 15 years. Other interesting appearances are presented by the shadows thrown by the planet on the rings and by the rings o^ *he planet. THE PLANETS. 253 SECTION III. Kepler's Laws of Planetary Motion. 326. Kepler's Three Laws. We have already seen that the orbits of most of the planets are nearly circular, their -distances from the Sun being nearly constant and their motions being nearly uniform. A far closer approximation to the truth is the hypothesis held for a long time by Tycho Brahe and other astronomers, namely, that each planet re- solved in a circle whose centre was at a small distance from the Sun, and described equal angles in equal intervals of time about a point found by drawing a straight line from the Sun's centre to the centre of the circle and producing it for nn equal distance beyond the latter point. The true laws which govern the motion of the planets were discovered by the Danish astronomer Kepler, in connection with his great work on the planet Mars (De Motibus Stellae Jfartis). After nine years' incessant labour the first and second of the following laws were discovered, and shortly afterwards the third. I. Every planet moves in an ellipse, with the Sun in one of the foci. II. The straight line drawn from the centre of the Sun to the centre of the planet (the planet's "radius vector") sweeps out equal areas in equal times. III. The squares of the periodic times of the several planets are proportional to the cubes of their mean distances from the Sun. These laws are known as Kepler's Three Laws. We Tiave already proved that the first two laws hold in the case of the Earth. The third law is also found to hold good for the Earth as well as the other planets, and this fact alone .affords strong evidence that the Earth is a planet 254 ASTRONOMT. By the mean distance of a planet is meant the arith- metic mean between the planet's greatest and least dis- tances from the Sun. If p, a (Fig. 100) be the planet' im- positions at perihelion and aphelion (i.e., when nearest and furthest from the Sun respectively), the planet's mean distance = (Sp + Sa) = \pa = \ (major axis of ellipse described) (147). The periodic times are, of course, the sidereal periods. Hence the third law is a relation between the sidereal periods- and the major axes of the orbits. FIG. 100. 327. Verification of Kepler's First and Second Laws. We will now roughly sketch the principle of the- methods by which Kepler determined the orbit of Mars, and thus proved his Eirst and Second Laws. A verification of the laws in the case of the Earth. has already been given, and we have shown ( 145) how to determine exactly the position of the Earth at any given time ; we may regard this, there- fore, as known. We may also suppose the length of the sidereal period of Mars to be known, for the average length of the synodic period may be found, as in 261, and the sidereal period may be deduced by the formula of 322. Let the direction of the planet be observed when it is at any point J/ in its orbit, the Earth's position being E. When the planet has returned again to Jf after a sidereal revolution, the Earth will not have returned to the same place in it* 'J HE PLAXETS. 255 orbit but will be in a different position, say F Let no the planet's new direction FM be observed * the "rie k ^T g F hG ^\ m0ti< 2' We knmv ^> ^ nd tne angle -^A Prom th the ieF 0< 2' We nmv ^> ^ nd tne angle -^A Prom the observations of the two directions of J/ we know the angles SEM ^\ SFM. These "* G sufficient to enable us to solve the quadrilateral Via. 101. . . We can thus determine SM and the angle .2SJ/; whence the dis ance and d lre ction of M from the Suu am found? Similarly, any other position of Mars in its orbit can be found by two observations o the planet's sidereal period separated by the interval of the planet's sidereal revolution. In s way, by a senes of observations of Mars, extending ovei dS Clirccti011 d daity ' * For simplicity we suppose Mars to move in thelcliptic plane The methods require some modification when the inclination of the orbits 1S taken into account, but the general principle is the same. r A 256 ASTEOXOMY. b28. Verification of Kepler's Third Law. Kepler '& Third Law can le verified much more easily, especially if we make the approximate assumption that the planets revolve uniformly in circles about the Sun as centre. The sidereal periods of the different planets can be found by observing the average length of the synodic period (the actual length of any synodic period is not quite constant, owing to the planet not revolving with exactly uniform velocity) and applying the equations of 322. The distance of the planet may be compared with that of the Earth, either by observing the greatest elongation (317) in the case of an inferior planet, or by the method of 319. It is then easy to verify the relation between the mean distances and periodic times of the several planets. In the table of 314, the student will have little difficulty in verifying (especially if a table of logarithms- be employed) that the square of the ratio of the periodic time of the planet to the year (or periodic time of the Earth) is in every case equal to the cube of the ratio of the planet's mean distance to that of the Earth.* The data being only approximate, however, the law can only be veri- fied as approximately true, although it is in reality accurate. Owing to the importance of Kepler's Third Law, we append the following examples as illustrations. EXAMPLES. 1. Given that the mean distance of Mars is 1'52 times that of the Earth, to find the sidereal period of Mars. Let T be the sidereal period of Mars in days. Then, by Kepler'a Third Law, /. T - 305^ x A/(3-511S) - 305 L x T874 = 684'5. Hence, from the given data, the period of Mars is T874 of a year, or 684-5 days. Had we taken the more accurate value of the relative distance, viz., l - 5237, we should have found for the period the correct value, namely, 687 days. * In other words, 2 log (period in years) = 3 log (distance in terms- of Earth's distance). THE PLANETS. 257 2. The synodic period of Jupiter being 399 days, to find its distance from the Sun, having given that the Earth's mean distance is 92 million miles. Let T be the sidereal period of Jupiter. Then, by 322, JL = _1 1 33f T 365^ 399 36o x 399' . = 11-82, or nearly 12 years. Let a be the distance of Jupiter in millions of miles. Then by Kepler's Third Law, I SL V = ( Y - 144 \ 92 ; IT I /. a =92 x3/(l44) = 92x5-24 = 482; that is, Jupiter's distance is 482 millions of miles. By taking T - 11'82 and the Earth's distance as 92'04, we should have found the more accurate value 477'6 for Jupiter's distance in millions of miles. 329. Satellites. The motions of the satellites about any planet are found to obey the same laws as those which Kepler investigated for the orbits of the planets. For example, the Moon's orbit about the Earth is an ellipse, and (except so far as affected by perturbations) satisfies both of Kepler's First and Second Laws. When a number of satellites are revolv- ing round a common primary (i.e., planet) as is the case with Jupiter, the squares of their periodic times are found, in every case, to be proportional to the cubes of their mean distances from the planet.* EXAMPLE. To compare (roughly) the mean distances of its two satellites from Mars. The periodic times are 30^ h. and7|h. respec- tively, and these are in the ratio (nearly) of 4 to 1. Hence the mean distances are as 4^ : 1, or %/W : 1. Now, 2-yi6 = s/128 = 5 very nearly (since 5 :J - 125). Hence, the mean distances are very nearly in the ratio of 5 to 2. * Of course the relation docs not hold between the periodic times and mean distances of satellites revolving round different planets, nor between those of a satellite and those of a planet. 258 ASTRONOMY. SECTION IV. Motions Relative to Stars Stationary Points. 330. Direct and Retrograde Motion. We have described ( 316-318) the motion of a planet relative to the Sun. In considering- its motion relative to the stars we must take account of the Earth's motion. FIG. 103. An inferior planet moves more swiftly than the Earth. Hence at inferior conjunction the line ^^(Fig. 102) joining them is moving in the direction of the hands of a watch. The planet therefore appears to move retrograde. At greatest elonga- tion ( 7, U') the planet's own motion is in the line joining it to the Earth, and hence produces no change in its direction ; but the Earth's direct motion causes the line EU or EU' ta turn about U QT U' with a rotation contrary to that of the hands of a watch; and therefore the apparent motion is direct. Over the whole portion UBU' of the relative orbit both the Earth's motion and the planet's combine to make the planet's apparent motion direct. There must, therefore, be two positions, M between A and U and N between U' and A, at which the motion is checked and reversed. At these two positions the planet is said to be stationary. A superior planet moves slower than the Earth ; hence at opposition the line EA (Fig. 103) joining them is turning in the direction of the hands of a watch. The planet therefore appears to move retrograde. At quadrature (2t, T) the Earth is moving along RET] hence its motion produces no change in the planet's direction. Hence the planet's direct motion about THE PLANETS. 259 the Sun makes its apparent motion also direct. In all parts of the arc RBT the orbital velocities of Earth and planet conspire to produce direct motion. Hence the planet is stationary at If, between A and H, and at JV between In both cases the longitude increases from J/ to JV and decreases from Nto J/; hence it is a maximum at JV and a minimum at M. After a complete synodic revolution the planet's elongation is the same as at the beginning, and the Sun's longitude has been increased ; therefore the planet's longitude has also increased. Hence the direct preponderates over the retrograde motion. FIG. 105. FIG. 104. 331. Alternative explanation. We may also proceed ;as follows. Let E, J represent two planets at heliocentric conjunction. Let E^ E^ E z , ..., J l} J^ 7 8 , ..., be their successive positions after a series of equal intervals. To find the apparent motion of /among the stars, as seen from J2, take any point E, and let E\, E'l, JS3, ... (Fig. 105) be parallel respectively to E^'E^ E./^ .... Then the points 1, 2, 3, ... represent 7's direction as seen from ^at a 'series of equal intervals, starting from opposition. 260 ASTKONOMY. Again, if Jl, <72, <73 be taken parallel to (Pig. 108), the points 1, 2 now represent j's direction as seen from J. We observe from Figs. 107, 108 that the relative motion is retrograde from 1 to 2, and becomes direct near 3. At the instant at which this takes place, either planet must be stationary, relative to the other. Since J 4 E 4 is nearly a tan- gent to JS's orbit, E is near its greatest elongation, and J is near quadrature at the positions 4 ; hence, E appears stationary from /between inferior conjunction and greatest elongation ; and J appears stationary between opposition and quadrature. FIG. 107. We notice that <71, J2, . . . are parallel to E 1, E<2, but measured in opposite directions, showing that the motion of E relative to J is the same (direct, stationary, or retrograde) as that of / relative to E. THE PLANETS. 261 332. Effects of Motion in Latitude. Hitherto we have supposed the planet to move in the ecliptic. When, however, the small inclination of the orbit to the ecliptic is taken into account, it is evident that the planet's latitude is subject to periodic fluctuations. The points of intersection of the planet's orbit with the ecliptic are (as in the case of the Moon) called the Nodes. Whenever the planet is at a node its latitude is zero; and this happens twice in every sidereal period of revolution. A planet is stationary when its longitude is a maximum or mininium, but unless its latitude should happen to be a maximum at the same time, the planet does not remain actually at rest. When the change from direct to retrograde motion, and vice versa, is combined with the variations in lati- tude, the effect is to make the planet describe a zigzag curve, sometimes containing one or two loops, called " loops of retrogression." This is readily verified by observation. t Ecliptic FIG. 109. Fig. 109 is an example of the path of Venus in the neigh- bourhood of its stationary points, the numbers representing its positions at a series of intervals of ten days. Here,, the planet is stationary close to the node JV, between 4 and 5, and it describes a loop in the neighbourhood of the- stationary point near 9, where its motion changes from re- trograde to direct. The student will find it an instructive exercise to trace out the path of any planet in the neighbourhood of its retrograde motion, using the values of its decl. and R.A., at intervals of a few days, as tabulated in the Nautical or Almanack. 262 ASTRONOMY. 333. To find the condition that two planets may be stationary as seen from one another, assuming the orbits circular and in one plane. Let P, Q be the positions of the planets at any instant ; P', Q' their position? after a very short interval of time. Then, if PQ and P'Q' are parallel, the direction of either planet, as seen from the other, is the same at the beginning and end of the interval ; that is, P is stationary as seen from Q, and Q is stationary as seen from P. Let u, v represent the orbital velocities of the planets P, Q ; a, b the radii $P, SQ respectively. FIG. 110. Draw P'J/, Q'N perpendicular to PQ. Then, in the stationary position, we must have P'M = Q'N. But PP', QQ', being the arcs described by the two planets in the same interval, are proportional to the velocities u, v. Therefore P'M, Q'N are proportional to the component velocities of the planets perpendicular to PQ. These com- ponent velocities must, therefore, be equal, and we have u sin P'PM= v sin Q'QN. "Whence, since P'P is perpendicular to SP and Q'Qto SQ, u cos SPQ = v cos SQN = vcos SQP (i.), and this is the condition that the planets may be stationary relative to one another. THE PLANETS. 263 *334. To find the angle between the radii vectores in the station- ary position, and the period during which a planet's motion is retrograde. By projecting SQ, QP on SP, we have a = 6 cos PSQ + PQ cos SPQ. Similarly b = a cos PSQ + PQ cos SQP. .-. cos SPQ : cos SQP = a - b cos PSQ : b - a cos PSQ. Whence, by (i.), u (a-b cos PSQ) +v (b-acosPSQ) = 0; = "" '" . (ii.). ac -\-lni By means of Kepler's Third Law, we can express the ratio of u to v in terms of a and b. For if T l} To denote the periodic times, then. evidently uT v = 2-, vT 2 = 2.T& ; .-. u : v = aT 2 : Z-7',. But T, I T, = ^ ; bl ; Substituting in (ii.), we have cos PSQ = ^ [From this result it may be easily deduced that tan t PSQ = ( 1=? /SQ ) * \l + cosPSQ/ In the above investigation PSQ is .the angle through which SQ 1 separates from SP between heliocentric conjunction and the station- ary point. Hence, since L PSQ increases from to 360 in the synodic period S, the time taken from conjunction to the stationary - 4. /P-SO point =Sx ^T' If L PSQi = L PSQ, there is another stationary point before con- junction, when the planets are in the relative positions P, Q. Hence, the interval between the two stationary positions is twice the time taken by the planets to separate through /PSQ, and is therefore This represents the interval during which the motion of either planet, as seen from the other, is retrograde. During the remainder of the synodic period the motion is direct, and the time of direct motion is therefore 264 ASTROXOMY. SECTION Y. Axial Rotations of Sun and Planets. 335. The Period of Eotation of the Sun can be found by observing the passage of sunspots across the disc. These spots, by the way, are very easily exhibited with any small telescope by focussing an image of the Sun on to a piece of white paper placed .a few inches in front of the eye-glass for to look straight at the Sun would cause blindness. As the Sun's axis of rotation is nearly perpendicular to the ecliptic, the rotation of the spots is seen in perspective, and makes them appear to move nearly in straight lines across the disc. From this observed apparent motion (as projected on the celestial sphere in a manner similar to that explained in 263) their actual motion in circles about the Sun's axis is readily determined. For example, if a spot move^ from the -centre of the disc to the middle point of its radius, we may readily see that the angle turned through = sin" 1 -| = 3C. The spots are observed to return to the same position in about .27| days, and this is their synodic period of rotation relative to the Earth. Call it 8, and let T be the time of a sidereal rotation, T the length of the year. Then, as in the case of an inferior planet ( 322), we may show that J_ = 1 j_ m 1 = 1 t 1 8 ~* T - Y ' T ~ 27i "*" 365^ ; whence the true period of rotation T = 25| days (roughly). It has been observed that spots near the Sun's equator rotate rather faster than those near the poles. This proves the Sun's surface -to be in a fluid condition, for no rigid body could rotate in this way. 336. Periods of Rotation of Planets. The rotation period of a .superior planet is easily found by observing the motions of the markings across its di-c near opposition, allowance being made for the motions of the Earth and planet. The surface of Mars has well- -defined markings, which give the period 24h. 37m. The principal mark on Jupiter is a great red spot amid his southern belts, which rotates in the period of 9h. 56m. Saturn rotates in lOh. 14m. For an inferior planet, the period is more difficult to observe. There is still some uncertainty as to whether Venus rotates in about ;23h. 21m., or whether, like Mercury, it always turns the same face to the Sun. There are no well-defined markings, and, as the greatest elongation is only 45, Venus can only be seen for part of the night as an evening or morning star, and in the most favourable positions only a portion of the disc is illuminated. Moreover, refraction, modified by air-currents, prevents the planet from being seen distinctly when near the horizon. If the same markings are :Been on the disc of a planet on consecutive nights, they may either hare remained turned towards the Earth, or they may have rotated through 360 during the day ; hence the difficulty of deciding between the two alternative hypotheses. Before the researches of Schiapa- j-elli ( 305), it was believed that Mercury also rotated in about 24h. THE PLANETS. 265 EXAMPLES. X. 1. The Earth revolves round the Sun in 365'25 days, and Venus in 224'7 days. Find the time between two successive conjunctions of Venus. 2. If Venus and the Sun rise in succession at the same point of the horizon on the 1st of June, determine roughly Venus' elongation. 3. Find tbe ratio of the apparent areas of the illuminated portions of the disc of Venus when dichotomized and when full, taking Venus' distance from the Sun to be T 8 T of that of the Earth. 4. Mars rotates on his axis once in 24 hours, and the periods of the sidereal revolutions of his two satellites are 1\ hours and 3O hours respectively. Find the time between consecutive transits over the meridian of any place on Mars of the two satellites respectively. 5. A small satellite is eclipsed at every opposition. Find an . expression for the greatest inclination which its orbit can have to the plane of the ecliptic. 6. If the periodic time of Saturn be 30 years, and the mean dis- tance of Neptune 2,760 millions of miles, find (roughly) the mean distance of Saturn and the periodic time of Neptune. (Earth's mean distance is 92 millions of miles.) 7. If the synodic period of revolution of an inferior planet were a. year, what would be its sidereal period, and what would be its mean distance from the Sun according to Kepler's Third Law ? 8. Jupiter's solar distance is 5'2 times the Earth's solar distance ' t find the length of time between two conjunctions of the Earth and. Jupiter. 9. Saturn's mean distance from the Sun is nine times the Earth's mean distance. Find how long the motion is retrograde, having given cos" 1 \ = 65. 10. Show that if the planets further from the Sun were to move -with greater velocity in their orbits than the nearer ones, there would be no stationary points, the relative motion among the stars "being always direct. AVhat would be the corresponding phenomenon, if the velocities of two planets were equal ? 266 ASTRONOMY. EXAMINATION PAPER. X. 1. Explain the apparent motion of a superior planet. Illustrate* by figures. 2. Describe the apparent course among the stars of an inferior planet as seen from the Earth, and the changes in appearance which the planet undergoes. 3. Define the sidereal and synodic period of a superior or inferior planet, and find the relation between them. Calculate the synodic period of a superior planet whose period of revolution is thirty years. 4. How is it that Venus alters so little in apparent magnitude (as- seen by the naked eye) in her journey round the Sun ? Why does- not Jupiter exhibit any perceptible phases ? 5. State Bode's Law connecting the mean distances of the various; planets from the Sun. 6. Prove that the time of most rapid approach of an inferior planet to the Earth is when its elongation is greatest, and that the-' velocity of approach is then that under which it would describe its- orbit in the synodic period of the Earth and the planet. Give the- corresponding results for a superior planet. (The orbits are to be- taken circular and in the same plane.) 7. AVhat is meant by stationary points in the apparent motion of a planet ? Prove that, if a planet Q is stationary as seen from P> then P will be stationary as seen from Q. 8. State Kepler's Three Laws, and, assuming the orbits of the- Earth and Venus to be circular, show how the Third Law might be verified by observations of the greatest elongation and synodic period of Venus. 9. Find the periods during which Venus is an evening star and a morning star respectively, being given that the mean distance of Venus from the Sun is '72 of that of the Earth. 10. Having given that there will be a full Moon on the 5th of June, that Mercury and Venus are both evening stars near their greatest elongations, that Mars changed from an evening to a morning star- about the vernal equinox, and that Jupiter was in opposition to the Sun on April 21st, draw a figure of the configuration of these heavenly bodies on May 1st. (All these bodies may be supposed to move in one plane.) CHAPTER XI. THE DISTANCES OF THE SUN AND STARS. SECTION I. Introduction Determination of the Surfs Parallax by Observations of a Superior Planet at Opposition* 337. In Chapter VIII. , Section I., we explained the nature of the correction known as parallax, and showed how to find the distance of a celestial body from the Earth in terms of its parallax. We also described two methods of finding the parallax of the Moon or of a planet in opposition the first by meridian observations at two stations, one in the northern and the other in the southern hemisphere ( 252) ; the second by micrometric observations made at a single observatory shortly after the time of rising and shortly before the time of setting of the planet or observed body ( 254). In both methods the position of the body is compared with that of neighbouring stars. This is impossible in the case of the Sun, for the intensity of the Sun's rays necessitates the use of darkened glasses in observations of the Sun, and these render all near stars invisible. Of course the star could theoretically be dispensed with in the method of 252, but only (as there explained) at a great sacrifice of accuracy ; and if a star is used which crosses the meridian at night, the temperature of the air has changed considerably, and the corrections for refraction are therefore quite different, besides which other errors are introduced by the change of temperature of the instrument. * The student will find it of great advantag3 to revise Section I. of Chapter VIII. before commencing the present Section. A.STKON". T 268 ASTRONOMY. In 264 we described a method, due to Aristarchus, in which the ratio of the Sun's to the Moon's distance was determined by observing the Moon's elongation when dicho- tomized, but this method was rejected, owing to the irregular boundary of the illuminated part of the disc, and the con- sequent impossibility of observing the instant of dichotomy. 338. Classification of Methods. The principal prac- ticable methods of finding the Sun's distance may be con- veniently classified as follows : A. Geometrical Methods. (1) By observations of the parallax of a superior planet at opposition (Section I.). (2) By observations of a transit of the inferior planet Venus (Section II.). B. Optical Methods (Section IV.). (3) By the eclipses of Jupiter's satellites (Roemer's Method). (4) By the aberration of light. C. Gravitational Methods (Chapter XIV., Section IV.). (5) By perturbations of Venus or Mars. (6) By lunar and solar inequalities. 339. To find the Sun's Parallax by Observation of the Parallax of Mars. By observing the parallax of Mars when in opposition, the Sun's parallax can readily be found. For the observed parallax determines the distance of Mars from the Earth, and this is the difference of the dis- tances of the Sun from the Earth and Mars respectively. The ratio of their mean distances may be found, if we assume Kepler's Third Law ( 326), by comparing the sidereal period of Mars with the sidereal year, and is therefore known. Hence the distance of either planet from the Sun may readily be found, and the Sun's parallax thus determined. The parallax of Mars in opposition may be observed by cither of the methods described in Chapter VIII., Section I. The method of 252 (by meridian observations at two stations) was employed by E. J. Stone in 1865. The observa- tions were made at Greenwich and at the Cape, and the Sun's parallax was computed as 8 -943". The method of 254 (by observations at a single observatory) was employed by Gill at Ascension Island in 1879, and the result was 8-783", THE DISTANCES OF THE SUN AND STAES. 269 EXAMPLE. If the parallax of Mars when in opposition be 14", to find the Sun's parallax, assuming the distances of the Sun from the Earth and Mars to be in the ratio of 10 : 16. The distance of the Earth from Mars in opposition is the difference of the Sun's distances from the two planets. Hence Distance of Earth from Mars I Distance of Earth from Sun = 16 - 10 : 10 = 3 : 5. But the parallax of a body is inversely proportional to its dis- tance ( 250). .'. Parallax of Sun : Parallax of Mars = 3:5; .'. Sun's parallax = 8 * 14// = 8'4". 5 *340. Effect of Eccentricities of Orbits. Owing to the eccen- tricities of the orbits of the Earth and Mars, their distances from the Sun when in opposition will not in general be equal to their mean distances, and therefore their ratio will differ from that given by Kepler's Third Law. But, by the method of 145, the Earth's dis- tance at any time may be compared with its mean distance, and similarly, since the eccentricity of the orbit of Mars and the position of its apse line are known, it is easy to determine the ratio of Mars' distance at opposition to its mean distance, and thus to compare its distance with that of the Earth. 341. Sun's Parallax by Observations on the Aste- roids and on Venus. The Sun's parallax may also be found by observing the parallax of one of the asteroids when in opposition, the method being identical with that employed in the case of Mars. In this way Galle, by meridian obser- vations of the parallax of Flora at opposition in 1873, com- puted the Sun's parallax at 8'873", and Lindsay and Gill, by observing the parallax of Juno in 1877, found the value 8-765". The next planet, Jupiter, is too distant to be utilized in this way. Its parallax at opposition is less than a quarter of the Sun's parallax, and is too small to be observed with sufficient accuracy. The Sun's parallax might also be found by an observation of Venus near its greatest elongation. The ratio of its distance to the Sun's might be calculated and its parallax found by the method of 252, and that of the Sun deduced. The method of 254 could not be employed, because one of the observa- tions would have to be made in full sunshine. 270 ASTLONOMT. EXAMPLES. 1. Having given that the greatest possible parallax of Mars when in opposition is 21'OS", to find the Sun's mean parallax, the eccentri- cities of the orbits of the Earth and Mars being ^ and T y respec- tively, and the periodic time of Mars being 1'88 of a year. The parallax of Mars is greatest when Mars is nearest the Earth ; hence the greatest possible value occurs when, at opposition, Mars is in perihelion and the Earth is at aphelion. Let r, r' denote the mean distances of the Earth and Mars from the Sun respectively. By Kepler's Third Law we have r' 3 _ (1-88) 2 . . r' n .J , ~ 9q ;. 3 - -p > - = (The calculation is most easily performed with a table of logarithms.) But since the Earth is in aphelion, its distance from the Sun at the time of observation is greater than its mean distance by Jj, and is therefore = r (1 + e^) = 1-017 r. Also the distance of Mars from the Sun at perihelion = r'(l-TV) = (l-iV)x 1-523 r = (1-523- -090) r- l'433r. Hence the least distance of Mars from the Earth at opposition = -416 r. Therefore, since r is the Sun's mean distance from the Earth, we have Observed parallax of Mars : mean parallax of Sun = 1 : '416; /. Sun's mean parallax = 21-08" x -416 = 8'77". 2. To find the Earth's moan distance from the Sun, and its dis- tances at perihelion and aphelion, taking the Sun's parallax as 8"79". If a denote the Earth's equatorial radius, we have, approximately, r = __ a __ = <* - ^ a x 2Q6 ' 265 sin 8*79" circ. meas. of 8'79" 879 * Taking a = 3963'3, this gives r (Earth's mean solar distance) = 93,002,000 miles, correct to the nearest thousand miles. Also, perihelion distance from Sun = 93,002,000 x (1-^) = 93,002,000-1,550,000 - 91,452,000 miles, and aphelion distance = 93,002.000 x (1 + -L) = 93,002,000 + 3,550,000 = 94,552,000 miles. DISTANCES OF THE SUN AND STARS. 271 SECTION II. Transits of Inferior Planets. 342. When Yenus is very near the ecliptic at inferior con- junction, it passes in front of the Sun's disc, appearing like a black dot on the Sun. Now the circumstances of such a transit are different at different places, for although both the Sun and planet are displaced by parallax, their displace- ments arc different, and their relative directions are therefore not the same. Now the ratio of the parallaxes of the Sun and planet at conjunction can be calculated from comparing their periodic times, or from the ratio of their distances, as determined by observations of the planet's greatest elonga- tion or otherwise. Hence, by comparing the circumstances of the transit at different places, it becomes possible to deter- mine the parallaxes of both the Sun and planet. The various methods of finding the Sun's parallax from observing transits of Venus may be classified as follows : (i.) By simultaneous observations of the relative position of the planet at different stations, either by micrometric mea- surements, or from photographs. (ii.) Delislis method, by comparing the times of the begin- ning or end of the transit at stations in different longitudes. (iii.) Halley's method, by comparing the durations of the transit at stations in different latitudes. Of these methods Halley's is the earliest, Delisle's the next. 343. First Method. Let P and p be the horizontal parallaxes of the Sun and of Yenus respectively at the time of transit. Then, at a place where the planet's zenith distance is z, its direction is depressed by parallax through an angle p sin z ( 249) ; also the Sun is depressed through P sin z* Hence the planet appears to be brought nearer to the Sun's lower limb by an angle (pP) sin %. If, now, the positions of the planet relative to the Sun's disc be simultaneously observed at any two or more different places, and the Sun's zenith distances be also determined, the difference of parallaxes p P can be readily found. Thus, if one of the stations be chosen where the Sun is * Strictly speaking, this should be P sin z,, where z\ is the Z.D. of the Sun's centre, but z l is very nearly equal to 2, and no sensible error is introduced by taking z instead of K\. 272 ASTBONOMt. vertical, and another where the Sun is on the horizon, the relative displacement will be zero at the former station, and p -P at the latter. Hence, the two directions of the planet relative to the Sun will he inclined at an angle p P. If two stations are at opposite ends of a diameter of the Earth, the angular distance between the relative positions will be '2 (pP). Hence, in either case, 2? P can be readily found. Let now / and r denote the distances of Yenus and the Earth from the Sun respectively. Then, if The the ratio of the sidereal period of Yenus to a year, we have, by Kepler's Third Law (assuming the orbits circular), r'/r = T\ whence the ratio of r' to r is found. Also, since Yenus is in conjunction, its distance from the Earth is = rr'. There- fore p : P= r : r r' t and ^_=n' = JL-l. p-P r V Whence, since the ratio of r to r is known, and P p has been observed, the Sun's horizontal parallax Pmay be found. We have roughly (by Bode's Law) r' =T*O r i an ^ therefore Hence the displacement of Yenus on the Sun's disc at a place where its zenith distance is 2, is about | P sin 2. The apparent position of Yenus on the Sun's disc may be observed either by measuring the planet's distance from the edge of the disc with a micrometer or heliometer, or by taking a photograph of the Sun. But the photographic method, though easier, does not give such accurate results. For, to obtain P correct to O'Ol", it would be necessary to find 2(p-P) correct to ^-xO'Ol", or about 0'05". Since the Sun's dia- meter is 32', the greatest possible difference of positions would be 20 x 32 x 60 ' r 37400' of the Sun's diameter. It is difficult to obtain a good photograph of the Sun more than 4J inches in diameter, and it would, therefore, be necessary to measure the planet's position correct to - & zoo f an inch, a degree of accuracy unattainable in practice. The slightest distortion or imperfection in the photographic plate would render the observations worthless. THE DISTANCES OF THE SUN AND STARS. 273 344. Delisle's Method. In this method, the Sun's parallax is determined by observing the difference between the times at which the transit begins or ends at different places. Let A, B be two stations near the Earth's equator in widely different longitudes, say at the ends of the diameter of the Earth, and in the plane containing UV, the path of Yenus' relative motion. Draw AUL and BVL, touching the Sun in L and cutting the path of Yenus in &, V. Then, when Yenus reaches U the transit begins at A, the planet appearing to enter the Sun's disc at L t and when Yenus is at V the transit begins at B. In the interval between the times of commencement of the transit as seen from A and B, the planet moves through the angle ULVor ALB about the Sun relative to the Earth, and this angle, being the angle sub- tended at the Sun by the Earth's diameter AB, is twice the Sun 1 9 parallax. FIG. 111. But the rate of relative angular motion of Yenus is known, being 360 in a synodic period. Hence the angle TJLV, described in the observed interval, is known, and the Sun's parallax is thus found. In a similar way, the Sun's parallax may be determined by < observing the interval between the times at which the transit ends at two stations A, B. We should have to draw two tangents from A, B to the opposite side of the Sun (M ). As before, the angle described by Yenus in the observed interval is twice the Sun's parallax. 274 ASTRONOMY. In employing Delisle's method, the observed times ol ingress or egress must be the Greenwich times, or must be reckoned from an epoch common to both observers. For this reason the difference of longitudes of the two stations must be accurately known. In the following example the ob- served interval 690s. corresponds to 8-86" of parallax, and it follows that an error of Is. in the estimated interval would give rise to an error of just over 0-01" in the computed parallax. Hence if the interval of time be estimated correct to the nearest second, the parallax will be correct to two decimals of a second. In practice it would be dim cult to make observations from the extremity of a diameter of the Earth, but the method is readily modified so as to be applicable when the stations are not so favourably situated. EXAMPLE. Given that the synodic period is 584 days, and that the difference between the times of ending of a transit, as seen from opposite ends of a diameter of the Earth, is llm. 30s., to find the Sun's parallax. In 584 days Venus revolves through 360 about the Sun relative to the Earth ; therefore its angular motion per minute 360 x 60 x 60 . , = - seconds = T541". 584 x 24 x 60 Therefore in 11 Jm. Venus describes an angle T541" x 11^ = 1772". This angle is twice the Sun's parallax ; .'. Sun's parallax = 8'86". 345. Halley's Method. The method now to be de- scribed was invented by Dr. Halley in 1716, and was first put into use at the transits in 1761 and 1769. In Halley's method the times of duration of the transits are observed from two stations A, B, one in north and the other in south latitude, in a plane as nearly as possible perpendicular to the ecliptic, or, more strictly, to the relative path of Venus. Take this plane as the plane of the paper in Fig. 112, and suppose also (for the purpose of simplifying the explanation) that -4, B are at the ends of a diameter of the Earth. Let LM be the diameter of the Sun's disc perpendicular to the line of centres, and let the directions of Yenus A V, BV, when pro- duced, meet the disc in a, I. Then a, b are the relative posi- tions of Yenus as seen at conjunction from A and 7>. THE DISTANCES OP TRK SUN A\0 STARS. 275 In Fig. 113 the Sun's disc is represented as seen from the Earth ; a, I are the positions of Venus as seen on the disc from A, B, projected on L1I, in Fig. 112, and PQR, PQR' are the apparent paths of Venus as it appears to cross the disc at B and A respectively. As in 343, the angular measure of the arc db or QQ! measures the sum of the displacement of Venus due to relative parallax at A and B, and this, in the circumstances here considered, is twice the difference of the parallaxes of the Sun and Venus. Now the observed times of duration of the transit ut A and B are the times taken to describe the chords P' Q-R' and PQR respectively. Knowing the synodic period of Venus and the ratio of its distances from the Sun and Earth, the rate at which Venus travels across the Sun's face can be found. Hence, the angular lengths of the chords PQR, P'Q'R' can be found. Also the Sun's angular diameter ZJ/ is known. Hence the angular distances OQ, OQ', QQ f can be calculated, for we have (very approximately) whence the and QQ' = OQ'-OQ.. Hence QQ' is known, and therefore the difference of parallaxes of Venus and the Sun is found ; Sun's parallax may bd found as in 343. 276 ASTRONOMY. *34C. Or if AB be known in miles, the length cf ab in miles can be found from the proportion ab : AB = Va : VA, and then, the angle aAb being known (being the angular measure of QQ')> we can find the Sun's distance in miles, for we have circular measure of L aAb = t ; whence aA Sun's distance Aa (in miles) = lep S th ab (in mi]es > . circular measure of L aAb The working of Halley's method will be made much clearer by a careful study of the following numerical examples. The student should copy Pigs. 112 and 113. EXAMPLES. 1. To find the angular rate at which Venus moves across the Sun's disc. Let 8, E, V denote the Sun, Earth, and Venus respectively (Fig. 112). From the example of 344, 8V separates from 8E with relative angular velocity, about 8, of T54" per minute, or 1' 32'4" per hour. But Venus is nearer the Earth than the Sun in the ratio 28 : 72 (roughly). And we have angular velocity of EV I ang. vel. of 8V Therefore EV separates from E8 with angular velocity = ^ x 1' 32-4" per hour = 3' 57'6" per hour = V per minute very nearly. 2. Neglecting the motion of the observatory due to the Earth's rotation, find the position on the Sun's disc of the chord PR, tra- versed by the planet, in order that the trsviisit may take four hours. Draw the figures as in 345. In four hours Venus moves 4x3' 58' , or very nearly 16' relative to the Sun (by Ex. 1) ; /. the chord PR must measure 10'. Hence PR is equal to the Sun's angular semi-diameter OP. Therefore, PR is a side of a regular inscribed hexagon in the Sun, and L MOP = 30. 3. If, at A, B. at opposite ends of a diameter of the Earth perpen- dicular to the piano of the ecliptic, the durations of transit are 3h. 21in. and 4h. respectively, to find the Sun's parallax. tHB DISTANCES OP THE SUN AND STARS. 2^7 Here tlie arc PR takes 39m. longer to describe than P*R'. Hence it is longer by 39 x 4", or 156". Draw R'K perpendicular to PR. Then, KR = ^PR-Pit) = x 156" * 78". Now, by Example 2, Z Jf OR = 60? And JBE', being very small, is approximately a straight line perpen- dicular to OR ; .'. R'RK = 30 approximately. Hence Q'Q = R'K = RKtan 30 = RK^/% = ^V3" = 45" nearly. But angular measure of Q'Q : twice Sun's parallax = 8V:EV= 18:7; .*. twice Sun's parallax = 45'' x T ^ = 17'50"; .'. Sun's parallax = 8*75". 4. A transit of Venus was observed from two stations selected as favourably as possible, one in N. the other in S. latitude, the zenith distances of the planet being 53 8' (sin 53 8' = '8) and 30 respectively. Given that the times occupied by the planet in pass- ing across the disc were 4h. 52m. and 4h. 30m., to find the Sun's parallax, assuming the distances of Venus and the Earth from the Sun to be in the ratio of 18 : 25 and neglecting the rotation of the Earth. Venus moves nearly 4" per minute relative to the Sun; hence in 4h. 30m. it moves through 18'. In 4h. 52m. it moves through 19 7 28" j /. in Fig. 113, P'Q' = 18' x = 9', PQ = 19' 28" x * = 9-73', and the Sun's semi-diameter SP ~ 16' nearly; .'. SQ =Vsp-'-pQ2 = v/256- 94-67 - 12W; SQ' = v/SP' 2 -P'Q 2 = v/256 -81 = 13-23 7 ; /. QQ' = -53 7 = 31-8". Now, if A and B be well chosen, QQ' is the sum of the relative displacements of Venus at the two stations. Let P be the Sun's parallax, p that of Venus ; then we have QQ' = (2>-P)( s in 2 + sin z'} = (p - P) x (sin 30 + sin 53 8') 1*3 Again, P:p= : /. P = 24-5" x T 7 ? = 9'5". Hence, with the given data, the Sun's parallax is 9'5". 278 \STRONOMY. 347. Difficulties of Observing the Duration of a Transit. In Examples 3, 4, above, the observed differences of duration were 39m. and 20m. respectively. An error of one second in the estimated durations of transit would give rise to an error of less than O'l per cent., and if we could be sure of observing the durations to within a second, the Sun's parallax could be found correct to two decimal places. But in practice it is extremely difficult to estimate the times of beginning and ending of a transit, even to the nearest second. For in the first place, Venus, when seen through the telescope, is not a mere point, but a disc of finite dimensions, its angular diameter at conjunction being about 67", or one-thirtieth of the diameter of the Sun. Hence its passage across the edge of the disc from external to internal contact occupies an interval which is never less than about 17s. (See Example on page 279.) est FIG. 114. Now, it is impossible to observe the first external contact ( U} of Venus with the Sun, because the planet is invisible until it has cut off a perceptible portion from the edge of the Sun's disc, and by that time it has advanced considerably beyond the point of contact. The last external contact ( F') at the end of the transit is also difficult (though rather less so) to observe, for a similar reason. For this reason, the internal contacts U', V. are alone observed, and a correction is applied for the angular semi- diameter of Venus. But in observing the first internal contact U\ when the planet's disc separates from the edge of the Sun, another difficulty, in the form of an optical illusion, makes itself manifest. THE DISTANCES OF THE SUN AND STARS. 279 Instead of remaining truly circular, the planet's disc appears to become elongated towards the edge of the Sun, and remains for some time connected with the edge by a narrow neck called the " black drop." This breaks suddenly at last, but not until the planet has separated some distance from the Sun's edge.* Even if the "black drop" be remedied, the atmosphere surrounding the planet Venus renders the con- tacts uncertain and ill-defined. It is worthy of notire that in Dclisle's method the times of ingress and egress at both stations are equally affected by the "black drop" appearance, and therefore it has no effect on the computation, provided that both observers take the same stage of the phenomenon for the observed time of ingress. EXAMPLE. Having given that the angular diameter of Venus at conjunction is 67", to find the interval between external and internal contact (i.) when Venus passes across the centre of the Sun's disc, (ii.) in the circumstances of Example 2, 346. (i.) Between external and internal contacts the planet moves through a distance equal to its angular diameter; therefore, since its rate of motion is 4" per second, the time occupied = 67 -f 4s. = 17s. very nearly. (ii.) Here the planet is 67" nearer the centre at internal than at external contact. Now the planet's direction of motion UV is inclined at angle 60 to the radius through the centre of the disc (Fig. 114). Hence the planet's .component relative velocity along the radius is 4" cos 60 per second, and therefore the interval required, in seconds, 67 = 67 4 cos 60 2 = 33-5s. 348. Recent Determinations of the Parallax of the Sun. Professor Arthur Auwers, the well-known Berlin astronomer, has recently (December 11, 1891) completed the calculations based on the observations in Germany of the transit of Venus in 1882. He finds that the parallax of the Sun is 8 800 seconds, with an error of 0*03 of a second at most. From the old observations of the transits of 1761 and 1709, Prof. Newcomb has lately computed the parallax at 8'79". * The " black drop " may be illustrated by holding two globes in the sunshine, at different distances from a white screen, and moving them until their shadows nearly touch. 280 ASTRONOMY. 349. Advantages and Disadvantages of H alley's and Delisle's Methods. In Halley's method the observed data are the intervals of time occupied by Yenus in crossing the Sun's disc at the two stations. It is not necessary to know the actual times of the transit ; hence neither the Greenwich time nor the longitude of the observatories need be known. In Delisle's method it is essential that the Greenwich times of the observations should be known with great accuracy, but it is not necessary to observe both the beginning and end of the transit at the two stations. Still, if these be both observed, we have two independent data for calculating the parallax, which afford some test of the accuracy of the computations. On the other hand, Delisle's method possesses the advan- tage that the places of observation mut be near the Earth's equator, and it may therefore be possible to select the stations nearly at opposite ends of a diameter of the Earth, and thus to get the greatest effect of parallax, while in Halley's method it is necessary that the stations shall be in as high latitudes as possible, and, owing to the practical difficulties of taking observations near the poles, the greatest effect of parallax cannot be utilized. Delisle's method is most easily employed if the transit is nearly central, i.e., if Venus passes nearly across the centre of the Sun's disc. This condition is fatal to the success of Halley's method ; here the best results are obtained when Yenus transits near the edje of the disc. For in Fig. 113 (page 27ij) we have OQ' 2 -OQ 2 = QP 2 -Q'P' 2 , PR-P'R' QP+Q^ 2 OQ + OQ' Hence the effect on QQ' of a small error in the computed length of PR or PR' will be least when QP + Q'P' is smallest and OQ+O'Q' is largest, a condition satisfied when the transit takes place near the edge M of the disc. On the other hand, for a nearly central transit, OQ, O'Q' would be email, and very slight errors in the estimated lengths of PR, P'R' would produce such large errors in the computed displacement QQ' as to render the method practically worthless. The transits of 1874 and 1882 were both favourable to the use of ll'illey's method. THE DISTANCES OP THE SUN AND Sf AKS. 2S1 *350. To determine the frequency of Transits of Venus. Since the Sun's angular semi-diameter is about 16', a transit of Yenus only occurs when the angular distance between the centres of the Sun and Yenus, as seen from some place on the Earth, is 1 6'. Hence, neglecting the effects of the relative parallax (P-p =. 23" by Ex. 3. 346, and this is small compared with 16'), Yenus must be at an angular dis- tance (SEV] < 16' from the ecliptic at the time of conjunc- tion. Hence the planet's heliocentric latitude JSSVmust be less than 16' xEVjSV, that is l6'x T 7 ? , or about 6'. Now the orbit of Yenus is inclined to the ecliptic at about 3 23', or 203'. Hence, by a method similar to that of 292, we see that the planet must be at a distance from the node of not more than about sin'^f-g = sin' 1 -^ (roughly) = 142', in order that a transit may take place. The smallncss of this limit alone shows that transits of Yenus are of rare occurrence. Now, a synodic period of Yenus contains about 584 days, that is, 1-599, or, more accurately, 1-598662 of a year. Hence five synodic revolutions occupy almost exactly eight years, the difference only amounting to T -^ of a year. This difference corresponds to an arc of f~f , or 2 24' on the ecliptic. This arc is much less than the doulle arc 3 24' within which transits take place. Hence it frequently happens that, eight years after one transit has taken place, the Sun and Yenus arc again at conjunction within the necessary limits, and another transit occurs near the same node. But after sixteen years, conjunction will occur at 4 48' from its first position ; this is greater than 3 24' ; hence there cannot be more than t\vo transits near the same node at intervals of eight years. And if a transit should be central, occurring almost exactly at the node, the conjunctions occurring eight years before and after would fall outside the required limits, and no second transit would then take place in eight years. Again, it maybe shown that 1-598662x147 = 235-003. Hence 147 synodic periods of Yenus occupy almost exactly 235 years, the difference being only '003 of a year. Thus a transit of Yenus may recur at the same node at an interval of 235 years. And it is possible to prove that thore is no 282 ASTRONOMY. intermediate interval between. 8 and 235 years at "which transits recur at the same node. If the orbits of the Earth and Venus were circular, a transit at one node would be followed by one at the opposite node in 11 3^ or 121* years. For 1-598662x71 = 113+'005; 1-598662x76 = 121*--002. But this result is modified by the eccentricities of the orbits (which now cause a difference of nearly a day in the times taken by the Earth to describe the two halves into which its orbit is divided by the line of nodes). In reality it is found that the intervals between transits of Venus occur at present in the following order : 8, 105*; 8, 121*; 8, 105*; 8, 121*. Transits have occurred, and are about to occur, in 1761, 1769, 1874, 1882, 2004, 2012 (the thick and thin type being used to distinguish the two different nodes). . Transits of Mercury occur much more frequently than transits of Venus. For although the orbit of Mercury is inclined to the ecliptic at about twice as great an angle as that of Venus, this cause is more than compensated for by the greater proximity of the planet to the Sun ; and since the synodic period of Mercury is only about % of that of Venus, conjunctions occur five times as often, so that we should ceteris paribus expect five times as many transits. By a method similar to that employed for Venus it is found that transits occur at the same node at intervals of 7, 13, 33, or -46 years. The next transit will occur in 1894. Although transits of Mercury thus occur far more often than transits of Venus, they cannot be used to determine the Sun's parallax with such accuracy, for Mercury is so near the Sun that the parallaxes of the two bodies are more nearly equal, and the planet's relative displacement is therefore much smaller than that of Venus. Moreover, Mercury moves much more rapidly across the Sun's disc, giving less time for accurate observations ; besides which, owing to the great eccentricity of the orbit, the ratio of Mercury's to the Earth's distance from the Sun cannot be so exactly computed. THE DISTANCES OF THE SUN AND STAES. 283 SECTION III. Annual Parallax, and Distances of the. Fixed Stars. 352. Annual Parallax, Definition. By Annual Parallax is meant the angle between the directions of a star as seen from different positions of the Earth in its annual orbit round the Sun. We haye several times ( 5, 247) mentioned that the fixed stars have no appreciable geocentric parallax. Their distances from the Earth are so great that the angle subtended at one of them by a diameter of the Earth is far too small to be observable even with the most accurately constructed instru- ments. But the diameter of the Earth's annual orbit is about 23,400 times as great as the Earth' R diameter, or about 186 million miles (twice the Sun's distance), and this diameter subtends, at certain of the nearest fixed stars, an angle sufficiently great to be measurable, sometimes amounting to between \" and 2". Now, the Earth, by its annual motion, passes in six months from one end to the other of a diameter of its orbit ; hence, by observing the same star at an interval of six months, its displacement due to annual parallax can be measured. Since the Sun is fixed, the position of a star on the celestial sphere is correctedfor annual parallax by referring its direction to the centre of the Sun; this is called the star's heliocentric direction, as in 156. The correction for annual parallax is the angle between the geocentric and heliocen- tric directions of a star. Let S be the Sun, E the Earth, x the star (Fig. 115). Then Ex is the apparent or geocentric direction of the star, Sx its heliocentric direction, and z ExS is the correction for annual parallax. This angle is also equal to xEx ! where Ex 1 is parallel to Sx. We notice that the correction for annual parallax (ExS) is the angular distance of the Earth from the Sun as they would appear if seen ly an observer on the star. ASTRON. u 284 ASTBONOMY. 353. To find the Correction for Annual Parallax. Let r = JES = radius of Earth's orbit. = Sx = distance of star. E = L SEx = angular distance of star from Sun. p = z ExS = annual parallax of star. By trigonometry we have in the triangle SEx smSJEx Sx ' whence* sinp= sin.E ..................... (i.). Hence the parallactic correction p is greatest when E = 90. This happens twice a year, and the corresponding positions of the Earth in its orbit are evidently the inter- sections of the ecliptic with a plane drawn through S per- pendicular to Sx. Let this greatest value of p be denoted by P, then P is called the star's annual parallax, or simply the star's parallax. j Putting E= 90 in (i.), we have and therefore sin^? = sin P . sin E. * Notice the close similarity between the present investigation and that of 249. f There is no risk of confusion in the use of the term parallax alone, because a star has no geocentric parallax. The " parallax " of a body means its equatorial horizontal parallax if the body belongs to the solar system. If not, its " parallax " is its annual parallax. THE DISTANCES OF THE SUN AND STAES. 285 But the angles P, p are always very small ; therefore their sines are very approximately equal to their circular measures. Thus we have approximately _P (in circular measure) = ~ , d # = JPsin E; and, if P", p" denote the numbers of seconds in P, p, P" = 180x60x60 -L = 206,265 r (approximately)) v d> d and p" = P" sin E. 354. Relation between the Parallax and Distance of a Star. If a star's parallax be known, its distance from the Sun is given by the formula fll = 180X60X60 JL =ao6266 r TT d d whence d = S r = 206,265 , where r is the Sun's distance from the Earth. For most purposes r may be taken as 93 million miles. EXAMPLES. 1. The parallax of Castor is 0'2" ; to find its distance. We have d = 206265 *- = 206,265x93,000,000 P" 0-2 = 5 x 206,265 x 93,000,000 = 95,900,000,000,000, or 959 x 10 11 miles approximately. It would be useless to attempt to calculate more figures of the result with the given data, which are only approximate. It is most convenient (besides being shorter) to write the result in the second form. 2. To find the distance of a Centauri (i.) in terms of the Sun's distance, (ii.) in miles, taking its parallax to be 0750". Here d = . r = 275,000r "75 275 x 10 3 x 93 x 10 fi = 25,575 x 10 9 256 x 10 11 miles approximately. 286 ASTRONOMY. 355. General Effects of Parallax. Since Ex is parallel to Sz, it is in the same plane as US and JSx. Hence the lines ES, Ex, Ex' cut the celestial sphere of E at points 8, x, a? , lying in one great circle, and we have the two following laws : (i.) Parallax displaces the apparent position of a star from its heliocentric position in the direction of the Sun. (ii.) The parallactic displacement of any star at different times varies as the sine of its angular distance from the Sun. FIG. 117. FIG. 118. Let Fig. 118 represent the observer's celestial sphere, S the Sun. Let x be the apparent or geocentric position of the star, whose parallax is P. Draw the great circle Sx and produce it to # , making XX Q = P sin Sx. Then # represents the star's heliocentric position, and this is its position as corrected for annual parallax. Conversely, if the star's heliocentric position # is given, we may obtain its geocentric or apparent position x by join- ing tf $, and on it taking xp = P sin 8x = P sin Sx Q very approximately (for the difference between P sin Sx and P sin Sx is exceedingly small, and may be neglected). The terms Parallax in Latitude and Parallax in Longitude are used to designate the corrections for parallax which must be applied to the celestial latitude and longitude of a star respectively. Similarly, the parallax in decl. and parallax in R. A. denote the corresponding corrections for the decl. and E-.A. THE DISTANCES OP THE SUN AND STARS. 28? 356. To show that any star, owing to parallax, appears to describe an ellipse. In Fig. 117, Ex' is parallel to the star's heliocentric direc- tion; therefore, x is fixed, relative to the Earth. Moreover, x'x = ES. Hence, as the Sun 8 appears to revolve about the Earth in a year, the star x will appear as though it revolved in an equal orbit about its heliocentric position x ', in a plane parallel to the ecliptic. FIG. 119. Let the circle MN(Fig. 119) represent this path, which the star x appearsto describe in consequence of parallax. This circle is viewed obliquely, owing to its plane not being in general perpendicular to Ex'\ hence, if mn denote its projection on the celestial sphere, the laws of perspective show that mn is an ellipse. (Appendix, 12.) This small ellipse is the curve described by the star on the celestial sphere during the year. Particular Cases. A star in the ecliptic moves as if it revolved about its mean position in a circle in the ecliptic plane, hence its projection on the celestial sphere oscillates to and fro in a straight line (more accurately a small arc of a great circle) of length 2P. For a star in the poie of the ecliptic the circle MN is per- pendicular to Ed) hence Ex describes a right cone, and the projection x describes on the celestial sphere a circle, of angular radius P, about the pole K. If the eccentricity of the Earth's orbit be taken into account, the curve MN will be an ellipse instead of a circle, but its projection mn will still be an ellipse. 288 ASTRONOMY. 357. Major and Minor Axes of the Ellipse. We shall now prove the following properties of the small ellipse described during the course of the year by a star whose parallax is P, and celestial latitude I. (i.) (A) The length of the semi-axis major is P. (B) The major axis is parallel to the ecliptic. (c) When the star is displaced along the major axis it has no parallax in latitude. (D) At these times the Sun's longitude differs from the star's by 90. (ii.) (A) The length of the semi-axis minor is Psinl. (B) The minor axis is perpendicular to the ecliptic. (c) When the star is displaced along the minor axis it has no parallax in longitude. (D) At these times the Sun's longitude is either equal to the star's, or differs from it ly 180. On the celestial sphere let a? denote the heliocentric position of the star, ABAB' the ecliptic, JTits pole, the secondary to the ecliptic through the star. Then, if S is the Sun, the star X Q is displaced to #, where x = P sin xS. THE DISTANCES OF THE SUN AND STAES. 289 (i.) The displacement is greatest when sin x Q S is greatest, and this happens when Bin# /S=: 1, # =90. If, therefore, we take A, A' on the ecliptic so that A) A' are the corresponding positions of the Sun. JTow A, A are the poles of BKB* (Sph. Geom., 11, 14, 15), and therefore the great circle Ax^A is a secondary to BKB'. Hence, if a, a' denote the displaced positions of the star, aa' is perpendicular to JK7?, and is therefore, parallel to the ecliptic. Also, x Q a = x$ = P sin 90 = P ; therefore the semi-axis major of the ellipse is P. Since AB = AB = 90, the star's longitude ( r B) differs from the Sun's longitude at A or A by 90. And since the star is displaced parallel to the ecliptic, its latitude, or angular distance from the ecliptic, is unaltered, and therefore the parallax in latitude is zero. (ii.) The parallactic displacement is least when sin xJ3 is least, and this happens when S is at B. For (Sph. Geom., 26) B is the point on the ecliptic nearest to # . Also, since sin xj = sin (180 # #) = sin x^B, it follows that the parallactic displacement is also least when S is at B'. If, therefore, b, V be the extremities of the minor axis, the arc lit is along JO?, and is therefore perpendicular to the ecliptic. Also, xjb = x$ = P sin x B = P sin / ; therefore the semi-axis minor is P sin I. When the Sun is at B, it has the same longitude as the star ; when at ^, the longitudes differ by 180. And since the star is displaced in a direction perpendicular to the ecliptic, its longitude TB is unaltered; therefore the parallax in longitude is zero, 290 ASTRONOMY. The parallax in latitude is evidently equal to the apparent angular displacement of the star resolved parallel to x Q K, and its maximum value is xjb, or xj)'. The parallax in longitude is not equal to the star's angular displacement perpendicular to Jx Q , but to the change of longitude thence resulting, and this is measured by the angle xKx^. Hence, in Tig. 120, (i.) The maximum parallax in latitude = x Q b = I* sin . (ii.) The maximum parallax in longitude = L x^Ka = x^Kd = x Q a/sin Kx Q (Sph. Geom. l7)=P/eos x Q B = P sec L 358. To determine the Annual Parallax of any Star, the following methods have been employed : (i.) The absolute method, by the Transit Circle ; (ii.) Bessel's, or the differential method, by the micrometer or heliometer j (iii.) The photographic method. The absolute method consists simply in observing with the Transit Circle the apparent decl. and R.A. of a star at different times in the year. From the small variations in these coordinates it is possible to find the star's parallax. Although this method has been successfully employed, it possesses many disadvantages. For the observations are con- siderably affected by errors of adjustment of the Transit Circle and by refraction. Moreover, several other causes give rise to variations in the star's apparent decl. and R.A. during the year. These include aberration (vide Section I V.) , precession ( 141), and nutation, all of which produce dis- placements much larger than those due to parallax. In 372 we shall see that when either the latitude or longi- tude is most affected by parallax it is unaffected by aberra- tion. Hence the best plan is to find the changes in these coordinates when they are respectively most affected by parallax. These changes are P sin I and Psec I ( 357) and from them P may be found. 359. Bessel's Method consists in observing with a micro- meter (^79) or heliometer ( 80) the variations in the angular distance and relative position of two optically near stars during the course of a year. THE DISTANCES OF THE STJN AND STARS. 291 The stars, being nearly in the same direction, are very nearly equally affected by refraction, and we may also men- tion that the same is true of aberration, precession and nutation. These corrections do not therefore sensibly affect the relative angular distance and positions of the stars. On the other hand, the two stars may be at very different dis- tances from the Earth ; if so, they are differently displaced by parallax, and their angular distance and position undergo variations depending on their relative parallax or difference of parallax. Hence, by observing these variations during the year the difference of parallax can be found. This method does not determine the actual parallax of either star. But if one of the observed stars is very bright and the other is very faint, it is reasonable to assume that the former is comparatively near the Earth, while the latter is at such a great distance away that its parallax is insensible. Under such circumstances the observed relative parallax is the parallax of the bright star alone. By making compari- sons between the bright star and several different faint stars in its neighbourhood, this point may be settled. If a considerable discrepancy .is found in the observed relative parallaxes, one or more of the comparison stars must themselves have appreciable parallaxes, but since the vast majority of stars in any neighbourhood are too distant to have a parallax, we shall be able to find the parallax not only of the star originally observed, but of that with which we had first compared it. The parallax of a star can never be negative ; if the relative parallax should be found to be negative, we should infer that the comparison star has the greater parallax, and is therefore nearer the Earth. 360. The Photographic Method is identical in prin- ciple with the last, but instead of observing the relative distances of different stars with a micrometer, portions of the heavens are photographed at different seasons, and the dis- placements due to parallax are measured at leisure by comparing the positions of any star on the different plates. This method has been used by Dr. Pritchard, of Oxford, and possesses the advantages of great accuracy, combined with convenience. 292 ASTRONOMY. 361. Parallaxes of certain Fixed Stars. The nearest stars are a Centauri, with a parallax of 0-75", and 61 Cygni, with parallax 0-54". Among others, the following may he mentioned: a Lyra, 0-18", Sirim, 0-2", Arcturm, 0-1 3", Polaris, 0-07", a Aquilce, 0-19". Of these, 61 Cygni is hy no means bright ; and a companion star to Sirius is invisible in all but two or three of the best telescopes. So it is not an invariable rule that faint stars are most distant, and have no appreciable parallax ; it is, however, true in the great majority of cases.* 362. Proper Motions. Binary Stars. Many stars, instead of being fixed in space, are gradually changing their positions. They are then said to have a proper motion. This motion may partly belong to the star, but is also partly an apparent motion, due to the fact that the solar system is itself moving through space in the direction of a point in the constellation Hercules. The displacement due to this cause can be allowed for approximately. Many of these motions, like that of our own Sun, are apparently progressive ; i.e., the star moves with constant velocity and in the same direction. Others are orbital, i.e., the star revolves about some other star, or (more accurately) two stars revolve about their common centre of mass. Such a system of stars is called a Binary Star. It is usually seen by the naked eye as a single heavenly body, its components being too near to be distinguished. Frequently a system of stars has itself a progressive motion; and sometimes an apparently progressive motion may really be an orbital one, with a period so long that the path has not sensibly diverged from a straight line during the short period for which stellar motions have been watched. A progressive or orbital motion cannot be confounded with the displacement due to annual parallax, for the former is always in the same direction, and the latter has a period dif- fering from a year, while parallax always produces an annual variation. * These figures can only be regarded as very rough approxima- tions, for considerable discrepancies exist between the values foun4 by different methods. THE DISTANCES OP THE SUN AND STAES. 293 SECTION IV. The Aberration of Light. 363. Velocity of Light. "We now come to certain methods of finding the Sun's distance which depend on the fact that light is propagated through space with a large but measurable velocity. The velocity of light has been measured by laboratory experiments in two different ways, invented by two French physicists, Fizeau and Foucault. For the description of these the reader is referred to "Wallace Stewart's Text Book of Light, Chapter IX.* The experiments give the velocity of light in air ; the velocity in vacuo can be obtained by multiplying this by the index of refraction of air.f The latter quantity may be found either by direct experiment or from the coeffi- cient of astronomical refraction (see 183). In 1876, Cornu, by employing Fizeau's method, found the velocity of light in vacuo to be 300,400,000 metres per second. Still more recently, Michelson, by a modification of Foucault's method, has found the velocity to be 299,860,000 metres, or 186,330 miles per second ; this may be taken as the most probable value. 364. Roemer's Method. The Equation of Light. In the last chapter we stated that Jupiter has four satellites, which revolve very nearly in the plane of the planet's orbit. Consequently a satellite passes through the shadow cast by Jupiter once in nearly every revolution, and is then eclipsed, as is our Moon in a lunar eclipse. Since the orbits and periods of the satellites have been accurately observed, it is possible to predict the recurrence of the eclipses, so that when one eclipse has been observed the times at which subsequent eclipses will begin and end can be computed. ]S"ow, the Danish astronomer Roemer in 1675 observed a remarkable discrepancy between the predicted and the observed times of eclipses. If of two eclipses one happens when Jupiter is near opposition, and the other happens near the planet's superior conjunction, the observed interval * The student will find it useful to read this chapter before com- mencing the present section, t Stewart's Light, 41. 294 ASTROXOMT. between the former and the latter is always greater than the computed interval ; similarly the observed interval between an eclipse near superior conjunction and the next eclipse near opposition is always less than the computed interval. The eclipses at conjunction arc thus always retarded, relatively to those at opposition, by an interval of time which is observed to be about 16m. 40s. As explained by Roemer, this apparent retardation is due to the fact that light travels from Jupiter to the Earth with finite velocity, and therefore takes 1 6m. 40s. longer to reach the Earth when the planet is furthest away at superior conjunction (B} than when the planet is nearest the Earth at opposition (A). The relative retardation is the difference between the times taken by the light to travel over the distances AE and BE. But BE- AE = 2S. Therefore the retardation is twice the time taken by the light to travel from the Sun to the Earth. Taking the retardation as 16m. 40s., we see that light takes 8m. 20s. to travel from the Sun to the Earth. This interval is sometimes called the " equation of light?' If we know the equation of light and the velocity of light, \ve may calculate the Sun's distance. Conversely, if the Sun's distance and the equation of light are known, the velocity of light can be determined. Knowing the Sun's distance, the Sun's parallax can be computed, as in Chapter VIII., Section I. The present method differs from those described in Sections I., II., in that it gives the distance instead of the parallax of the Sun. THE DISTANCES OF TIFE SUN AND STARS. 295 EXAMPLE 1. To find the Sun's distance, having given that the velocity of light is 186,330 miles per second, and that eclipses of Jupiter's satellites which occur when the planet is furthest from the Earth, are retarded 16m. 40s. relatively to those which occur when the planet is nearest. Here the time taken by light to pass over a diameter of the Earth's orbit is 16m. 40s. ; therefore light travels from the Sun to the Earth in 8m. 20s., or 500 seconds. /. the Sun's distance = 186.330 x 500 miles = 93,165,000 miles. EXAMPLE 2. Taking the value of the Sun's distance calculated in the preceding example, the Sun's parallax will be found to be about 8-78". 365. The Aberration of Light is a displacement of the apparent directions of stars, due to the effect of the Earth's motion on the direction of the relative velocity with which their light approaches the earth. The rays of light emanating from a star travel in straight lines through space* with a velocity of about 186,330 miles per second. We see the star when the rays reach our eye, and the appearance presented to us depends solely on how the rays are travelling at that instant. If the Earth were at rest, and there were no refraction, we should see the star in its true direction, "because the light would be travelling towards our eyes in a straight line from the star. But in every case the direction in which a star is seen is the direction of approach of the light-rays from the star at the instant of their reaching the eye. Now the velocity of approach is the relative velocity of the light with respect to the observer. If the observer is in motion, this relative velocity is partly due to the motion of the light and partly due to the motion of the ob- server. If the observer happens to be travelling towards or away from the source of light, the only effect of his motion will be to increase or decrease the velocity of approach of the light, without altering its direction, but if he be moving in any other direction, his own motion will alter the direction of the relative velocity of approach, and will therefore alter the direction in which the star is seen. * Of course the rays are refracted when they reach the Earth's atmosphere, but the effects of refraction can be allowed for separately. 296 A8TEONOMY. Suppose the light to be travelling from a distant star x in the direction xO. Let T"be the velocity of light, and let it be represented by the length M 0. Suppose also that an observer is travelling along the direction NO with velocity w, represented by the straight line NO. Then, if we regard as a fixed point, the light is approaching with velocity re- presented by MO. Also since the observer is approaching with velocity represented by NO, the point is approaching the observer jVwith an equal and opposite velocity repre- sented therefore by ON. Hence the whole relative velocity with which the light is travelling towards the observer is the resultant of the velocities represented by M and ON. By the Triangle of Velocities this resultant velocity is repre- sented in magnitude and direction by MN. Hence MN represents the direction of approach of the light towards the observer's eye. Therefore when the observer has reached the star is seen in the direction Ox' drawn parallel to NM, although its real direction is Ox, In consequence, the star appears to be displaced from its true position x to the position x'. This displacement is called the aberration of the star, and its amount is, of course, measured by the angle xOx. This angle is sometimes called the angle of aberration or the aberration error. 366. Illustrations of Eelative Velocity and Aberration. The following simple illustrations may possibly assist the reader in understanding more thoroughly how aberration is produced. (1) Suppose a shower of rain-drops to be falling perfectly vertically, with a velocity, say, of 40 feet per second. Then, if a man walk through the shower, say with a velocity of 4 feet THE DISTANCES OF THE SUN AND STABS. 297 per second, the drops will appear to be coining towards him, and therefore to be falling in a direction inclined to the vertical. Here the man is moving towards the drops with a horizontal velocity of 4 feet per second, and therefore the drops appear to be coming towards the man with an equal and opposite horizontal velocity of 4 feet per second. Their whole relative velocity is the resultant of this horizontal velocity and the vertical velocity of 40 feet per second with which the drops are approaching the ground. By the rule for the compo- sition of velocities, this lelative velocity makes an angle tan~ l -$ or tan* 1 '1 with the vertical. Hence the man's own motion causes an") apparent displacement of the direction of the rain from the vertical Y" through an angle tan' 1 g l. This angle corresponds to the angle ofJ aberration in the case of light. (2) Suppose a ship is sailing due south, and that the wind is blow- ing from due west with an equal velocity. Then to a person on the ship the wind will appear to be blowing from the south-west, its southerly component being due to the motion of the ship, which is approaching the south. In this case the ship's velocity causes the wind to apparently change from west to south-west, i.e., to turn through 45. We might, therefore, consistently say that the " angle of aberration " of the wind was 45. 367. Annual and Diurnal Aberration. A point on the Earth's surface is moving through space with a velocity compounded of (i.) The orbital velocity of the Earth in the ecliptic about the Sun ; (ii.) The velocity due to Earth's rotation about the poles. These give rise to two different kinds of aberration, known respectively as annual and diurnal aberration. Now the Earth's orbital velocity is about 2?r x 93,000,000 miles per annum, or rather over 18 miles per second, while the velocity due to the Earth's rotation at the equator is roughly 2^x4000 miles per day, or 0*3 miles per second. The former velocity is about T oi^o ^ * ne velocity of light, and therefore the annual aberration is a small though measurable angle. The latter velocity is only -fa as great ; hence the diurnal aberration is much smaller and less important. For this reason the term " aberration" always signifies annual aberration, unless the word "diurnal" is also used. We shall now consider the effects of annual aberration, leaving diurnal aberration till the end of this section. 298 * ASTRONOMY. 368. To determine the correction for aberration on the position of a Star. Let Ox be the actual direction of a star x seen from the Earth at ; U the direction of the Earth's orbital motion at the time of observation. On Ox take OM representing on any scale the velocity of light, and draw MY parallel to 017, and representing on the same scale the velocity of the Earth. Then YO represents the relative velocity of the light in magnitude and direction, so that OYx is the direction in which the star x is seen (Mg. 123). [For if ON be drawn parallel and equal to YM, the parallelogram of velocities MNOT shows that 21/0, the actual velocity of the light- rays in space is the resultant of the two velocities TO and NO, or YO and MY, and therefore YO is the required relative velocity.] FIG. 124. Since Ox, Ox, and U all lie in one plane, it follows, by representing their directions on the celestial sphere, that a star is displaced by aberration along the great circle joining its true place to the point on the celestial sphere towards which the Earth is moving. The displacement xOx is called the star's aberration error. Let it be denoted by y, and let u = NO = velocity of Earth, V MO = velocity of light. Then the triangle OM Y gives sinMOY_ MY _ u t sin M YO ~~ MO ~~ V* or sin y = -^ sin HYO = ~ sin THE DISTANCES OF THE SUN AND STARS. 299 The aberration error y is, therefore, greatest when UOx' 90. Let its value, then, be k. Putting UOx = 90, we have sin H =s and .*. sin y = sin k sin UOoc. The angle UOx is called the Earth's Way of the star, and k is called the Coefficient of Aberration. Since a and k are both small, we have, approximately y = k sin (Earth's way), k (in circular measure) = u/ V\ and, therefore, if y", k" denote the number of seconds in y, k respectively ?/' = k" sin (Earth's way), , 180x60x60 u ~iT ' v = 206,265 x. velocity of light 369. General effect of Aberration on the Celestial Sphere. Neglecting the eccentricity of the Earth's orbit, the direction of motion of the Earth, in the ecliptic plane, is always perpendicular to the radius vector drawn to the Sun. Hence, on the celestial sphere, the point 7", towards which the Earth is moving, is on the ecliptic, at an angular distance 90 behind the Sun. This point is sometimes called the apex of the Earth's Way. Let x' denote the observed position of the star. Draw the great circle x' 7, and produce it to a point x , such that xx' = k sin x' U. Then x represents the star's true position, corrected fV aberration. Conversely, if we are given the true position x, we can find the apparent position x' by joining #7 and taking xx' = k sinxV, for it is quite sufficiently approximate to use k sin xU instead of k sin x'U. A3TROX. X 300 ASTRONOMT. "We thus have the following laws : (i.) Aberration produces displacement in the apparent position of a star towards a point U on the ecliptic, distant 90 behind the Sun. (ii.) The amount of the displacement varies as the sine of the Earth's Way of the star, i.e., the star's angular distance from the point U. FIG. 125. FIG. 123. 370. Comparison between Aberration and Annual Parallax. The student will not fail to notice the close analogy between the corrections for aberration and annual parallax. The point 7 for the former corresponds to the point S for the latter, in determining the direction and magnitude of the displacement. In fact, the aberration error of a star is exact/// the same as its parallactic correction would be three months earlier (when the Sun was at U) if the star's annual parallax ice re k. There is, however, this important difference that the annual parallax depends on a star's distance, whilst the constant of aberration k is the same for all stars. For k depends only on the ratio of the Earth's velocity to the velocity of light, and not on the star's distance. The value of k in seconds is about 20'492" ; for rough purposes it mnv he taken as 20'5". THE DISTANCES OF THE SUN AND STARS. 301 371. To show that the aberration curve of a star is an ellipse. This result, which follows immediately from the analogy between aberration and parallax, may be proved inde- pendently as follows: On Ox (Fig. 125), the true direction of a star #, take Ox to represent the velocity of light, and xM to represent the Earth's velocity. Then M meets the celestial sphere in m, the star's apparent position. As the Earth's direction of motion in the ecliptic varies, while its velocity remains constant, Jfdescribes a circle, about x as centre in a plane parallel to the ecliptic plane. The projection of this circle on the celestial sphere is an ellipse (cf. 356), and this is the curve traced out by a star during the year in consequence of aberration. Particular Cases. A star in the ecliptic oscillates to and fro in a straight line, or more accurately an arc of a great circle of length 2/j. A star at the pole of the ecliptic revolves in a small circle of radius k (cf. 356). 372. Major and Minor Axes of the Aberration Ullipse. By writing Z7"for S and k for P in the investiga- tion of 357, we obtain the analogous results relating to the ellipse described by a star in consequence of aberration, namely : (i.) (A) The length of the semi-axis major is k. (B) The major axis of the ellipse is parallel to the ecliptic. (c) When the star is displaced along the major axis it has no aberration in latitude. (D) At these times the Surfs longitude is either equal to the .star's, or differs from it by 180.* (ii.) (A) The length, of the semi-axis minor is k sin I. (B) The minor axis is perpendicular to the ecliptic. (c) When the star is displaced along the minor axis, it has no aberration in longitude. (D) At these times the Sun's longitude differs from the star's .by 90. COROLLARY. The maximum aberration in longitude = k sec? (cf. 357, ii.). * Note that (i., r) and (ii., D) are the reverse of the corresponding -properties in 357. 302 ASTRONOMY. *373. Effect of Eccentricity of Earth's Orbit. Owing to the elliptic form of the Earth's orbit the Earth's velocity is not quite uniform, and therefore the coefficient of aberration is subject to small variations during the year. The earth's velocity is greatest at perihelion and least at aphelion. The angular velocities at those times are inversely proportional to the squares of the corresponding distances from the Sun, but the actual (linear) velocities are in- versely proportional to the distances themselves, and these are in the ratio of l-e : 1 + e, or 1 --fa : 1 + ^ ( 149). Since the coeffi- cient of aberration is proportional to the Earth's velocity, its greatest and least values are therefore in the ratio of 61 : 59, and are respectively and of its mean value. Moreover, the direction of the Earth's motion is not always exactly perpendicular to the line joining it to the Sun, hence the " apex of the Earth's way," towards which a star is displaced, may be distant a little more or a little less than 90 from the Suu at different seasons. The aberration curve is still an ellipse. The student who has read the more advanced parts of particle dynamics may know that the curve MN, tracecfout by M, is in this case the "hodograph " of the Earth's orbital motion. It is also known, in the case of elliptic motion, such as the Earth's, that this hodograph is a circle, whose centre does not, however, quite coincide with x. Hence the aberration-curve hTc is an ellipse. 374. Discovery of Aberration. Aberration was dis- covered by Bradley, in 1725, in the course of a series of observations made with a zenith sector on the star y Draconis for the purpose of discovering its annual parallax. The star's latitude was observed to undergo small periodic variations during the course of the year, and these differed from the variations due to annual parallax in the fact that the dis- placement in latitude was greatest when the Sun's longitude differed from that of the stars ly 90 ; that is, at the time when the parallax in latitude should be zero ( 357, i., c.). The fact that the phenomenon recurred annually led Bradley to suppose that it was intimately connected with the Earth's motion about the Sun, and he was thus led to adopt the explanation which we have given above, It will be seen that the pecu- liarity which led Bradley to discard annual parallax as an explanation is quite in harmony with the results of 372. 375. To Determine the Constant of Aberration by Observation. The constant k can best be found by observ- ing different stars with a zenith sector or transit circle, as in., the direct method of finding a star's parallax ( 358). THE DISTANCES OF THE SUN AND STAUS. 303 The differential method of 359 cannot be used, because the coefficient of aberration is the same for all stars. But aberration is much larger than parallax (the coefficient of aberration being 20-49", while the greatest stellar parallax is < I"), and can therefore be found directly with greater accuracy. Of course it is necessary to make corrections for refraction and precession. The former correction is the most liable to uncertainty, as it varies slightly according to atmo- spheric conditions. But, as all stars have the same constant of aberration, a star may be selected which transits near the zenith, and is therefore but little affected by refraction. This condition was secured by Bradley when he observed the star y Draconis. The star is very favourable in another respect, for its longitude is very nearly 270. It therefore lies very nearly in the " solstitial colure," its declination circle passing nearly through the pole^f the ecliptic. J, At the vernal equinox, the star's longitude is less than the Sim's by 90, and it is therefore displaced away from the poles of the ecliptic and equator through a distance k" sin ?, its decimation being therefore decreased by k" sin I. At the autumnal equinox its declination is increased by k" sin L Hence the difference of the apparent declinations = 2k' 1 sin ?, and this is also the difference of the star's apparent meridian zenith distances. By observing these, k" may be found, 7 being of course known. The value of k" is very approximately 20 - 493". 304 A.STBONOMY. 376. Relation between the Coefficient of Aberration and the Equation of Light. "We have seen (368) that JU 180x60x60 u r x T~ 'T '. ................. W ' where V is the coefficient of aberration in seconds, u the velocity of the Earth, V that of light, hoth of which we will suppose measured in miles per second. Now let r represent the radius of the Earth's orbit (sup- posed circular) in miles. Then in one sidereal year, or 365 J days, the Earth travels round its orbit through a distance 2irr miles. Hence the Earth's velocity in miles per second is 365ix 24x60x60 Substituting in (i.), we have jfc" - 15 JL 365 V But r/ Fis the time taken by the light to travel from the Sun to the Earth, measured in seconds, or the " equation of light." Hence, The coefficient of aberration in seconds = --- x number of seconds taken by Sun's light to 365 4 reach Earth. Thus, by observing the retardation of the eclipses of Jupiter's satellites at superior conjunction, the coefficient of aberration can be found independently of the methods of 375, the number of days (365^) in the sidereal year being of course known. The close agreement between the values found thus and by direct observation affords the strongest evidence in support of Bradley 's explanation of aberration. EXAMPLE. To find the coefficient of aberration in seconds, having given that light takes 8m. 20s. to travel from the Sun to the Earth. Here the required coefficient of aberration ,// 15 x 500 7oOO THE DISTANCES OF THE SUN AKD STARS. 305 377. To find the time taken by the light from a star to reach the Earth. It is sometimes convenient to estimate the distance of a star by the number of years which the light from it takes to reach the Earth. This may he determined from a knowledge of the star's parallax, and of the coefficient of aberration, without knowing either the Snn's distance or the velocity of light. Let the parallax of a star he = P" in seconds = P radians, and let the coefficient of aberration = k" seconds = k radians. Then, if r, d be the Earth's and star's distances from the Sun, we have p _ r 7 _ velocity of Earth d ' " velocity of light " Now, in one year, the Earth travels over a distance 2?rr ; 27Tf .-. in one year light travels a distance ; /J .-. the number of years taken by light to travel from the star (distance d) to the Earth ' \ k ~ 27TT ~~ 27TP 27TP"' The distance travelled by light in a year is sometimes called a " light-year." Hence, The product of a star's parallax and its distance in light- years is equal to the coefficient of aberration divided by 2?r. EXAMPLES. 1. To find how long the light would take to reach U8 from a star having a parallax Ol". The required time, in years, 1 fc" 10x20-49x7 == = --- approximately STT 0-1 2 x 22 = 32-6. 2. To find the time taken by the light from the nearest star, a Centauri, taking its parallax as 075". The parallax is 7'5 times that of the star in the last question, therefore its distance is 10/75 as great, and the time taken by the light = ^ = 4-35 years. 7'5 306 ASTRONOMY. 378. Relation between the Coefficient of Aber- ration, the Sun's Parallax, and the Velocity of Light. -It follows from 376 that if the coefficient of aberration k" be determined by observation, the fraction rjV is also known, independently of observations of the eclipses of Jupiter's satellites. And if F, the velocity of light, be deter- mined experimentally by the method of Foucault or Fizeau, the Sun's distance r can be found. Thus the Sun's parallax can be calculated from the coefficient of aberration and the velocity of light. And generally, if, of the four quantities, Sun's parallax, coefficient of aberration, velocity of light, and length of sidereal year in days, any three are observed, the value of the fourth may be deduced from them. In this manner Foucault, by his determination of the velocity of light, in 1862, found the Sun's parallax to be 8'86". Cornu, by experiments in 1874 and 1877, combined with the values for k" determined by Struve, obtained the values 8-83" and 8*80" respectively. Hichelson's experiments make the parallax 8- 793". EXAMPLE. If the velocity of light = 186,000 miles per second and the Earth's radius (a) = 3,960 miles, to prove that the product of the Sun's parallax and the coefficient of aberration, both measured in seconds, is 180'35. The Sun's parallax P" = 18 * G0 * 60 n- r 15 r GO P " fc" = 18 x G0 x ( > x 60 a 200205 x GO 3000 14G1* V 14G1 186000 = 180-35. 379. Planetary Aberration. The direction of any planet is affected by aberration, which is due partly to the motion of the Earth, and partly to that of the planet itself. For, during the time occupied by the light in travelling from a planet to the Earth, the planet itself will have moved from the position which it occupied when the light left it. We shall, however, show that the direction in which a planet is seen at any instant was the actual direction of the planet relative to the Earth at the instant previously when the lujlt left the planet. THE DISTANCES OF THE SUN AND STARS. 307 Let t be the time required by the light to travel from the planet to the Earth. Let P, Q be the positions of the planet and Earth at any instant ; P', Q' their positions after an interval t. The light which leaves the planet when at P reaches the Earth when it has arrived at Q' ; the direction of the actual motion of the light is, therefore, along PQ. Eut PQ' and Q Q' are the spaces passed over by the light and the Earth* FIG. 128. respectively in the time t (and QQ' is so small an arc that it may be regarded as a straight line). Therefore QQ' : PQ' = velocity of Earth : velocity of light. Hence it follows from 368 that the line PQ represents the direction of relative velocity of the light with respect to the Earth. Therefore, when the Earth is at Q' the planet is seen in a direction parallel to PQ, and its apparent direction is exactly what its real direction was at a time t previously. The same is true in the case of the Sun or a comet, or any other body, provided that the time taken by the light from the body to reach the Earth is so small that the Earth's motion doe. not change sensibly in direction in the interval. The aberration of the planet at any instant is the angle between the apparent direction PQ and the actual direction P'Q. 308 ASTRONOMY. EXAMPLE. To find the effect of aberration on the positions of (i.) the Sun, (ii.) Saturn in opposition, taking its distance from the Sun to be 9^ times the Earth's. (i.) The light takes 8m. 20s. to travel from the Sun to the Earth therefore the Sun's apparent coordinates at any instant are its actual coordinates 8m. 20s. previously. Thus, its apparent decl. and R.A. at noon are its true decl. and R.A. at 23h. 51m. 40s., or llh. 51m. ,40s. A.M. Now the Sun describes 360 in longitude in 365 days. Hence, in, 500 seconds it describes 20'492", and the Sun's aberration in longi- tude is 20'492". This is otherwise evident from the fact that the- Earth's way of the Sun is 90 ; and it is at rest, consequently its aberration = fc. (ii.) The distance of Saturn from the Earth at opposition is. = 9| 1, or 8 times the Sun's distance. Light travels over this distance in 8m. 20s. x 8 = 500x8|s. = Ih. 10m. 50s. Therefore, the apparent coordinates are the actual coordinates Ih. 10m. 50s. previously. Thus the observed decl. and R.A. at midnight (12h. 0*m. Os.) are the- true decl. and R.A. at lOh. 49m. 10s. 380. Diurnal Aberration is due to the effect of the Earth's diurnal rotation about the poles on the relative velo- city of light. As the Earth revolves from west to east, the portion of the- motion of an ohserver due to this diurnal rotation is in the- direction of the east point E of the horizon. The effect of diurnal aberration can thus be investigated by methods precisely similar to those of 368, Staking the- place of U.* Hence, every star x is displaced by diurnal aberration, towards the east point E. And if x' be its displaced position,, then the displacement xx' = A sin x E, where ,-, i n . volocitv of observer Circular measure of A = ^-- . velocity ot light * The student will find it useful to go through the various steps- of 368-371, considering the diurnal motion. THE DISTANCES OF THE STTN AND STARS. 309" Taking a for the Earth's radius, V for the velocity of light, let the observer's latitude be I. In a sidereal day (86164-1 mean seconds) the Earth's rotation carries the observer round a small circle, whose dis- tance from the Earth's axis is a cos , and whose circumference is, therefore, 'lira cos I. Hence, the observer's velocity = miles per second : 86164-1 ( F 1-irtt COS / .*. circular measure of A = 86164-1 x V 1 .'. A" (number of seconds in A] __ 180x60x60 x 2Tra cos/ TT 86164-1 7' Ida cos I , t = approximately. Thus, the coefficient of diurnal aberration varies as the cosine of the latitude. If K" denote the coefficient of diurnal aberration at the equator in seconds, we therefore,, have K" = - l5 * 39fi3 = n-32" V 186,000 A" = K" cos I = O'32' cos L * Effect of Diurnal Aberration on Meridian Observations. The correction for diurnal aberration is greatest when the star is 90 from the east point, i.e., is on the meridian. In this case, the displacement is perpendicular to the meridian, and is equal to A". The star's meridian altitude is thus unaffected, but its time of transit is somewhat retarded at upper culmination, and (for a cir- cumpolar star) accelerated at lower culmination, since the star appears on the meridian, when it is really A" west of the meridian- The effect of diurnal aberration on the time of transit is thus equi- valent to that of a small collimation error A" in the Transit Circle. For a star on the equator, seen from the Earth's equator, the- retardation of the time of transit would be -^ K " seconds, = -g^ of a second nearly, and it would be difficult to observe such a small interval. 310 ASTRONOMY. 381. To determine the Coefficient of Diurnal Aber- ration "by Observations of the Azimuths of Stars when on the Horizon. When a star is rising or setting it is evidently displaced by diurnal aberration along the horizon towards the east point. Consider two stars, one of which rises S. of E., and the other "N. of E. It is evident that their rising points are drawn towards one another. But the stars set S. of W. and "N. of W., and their displacements are still towards the E. point ; hence, their setting points are separated away from one another. And, if the stars, at rising and setting, be carefully observed with an altazimuth, the difference between their azimuths at setting will exceed that between their azimuths at rising by an amount proportional to the diurnal aberration. From this, the coefficient of diurnal aberration may be found. The azimuths are unaltered by refraction ( 184), but the times of rising and setting are slightly altered by refraction. If the co- efficient of refraction be the same at both observations, however, the acceleration in rising will be equal to the retardation at setting, .and the refraction will increase the azimuths at rising and setting by the same amount ; thus the data will be unaffected. If the tem- perature of the air has changed considerably between rising and setting, it is only necessary to make the observations at equal intervals before and after the stars transit. .382. Relation between the Coefficients of Aberra- tion and the Sun's Parallax. We have evidently 7T" _ velocity of diurnal motion at equator k" velocity of Earth's orbital motion But the velocities in miles, per sidereal day, are 2-n-a and This gives the coefficient of diurnal aberration at the equa- tor in terms of the coefficient of annual aberration and the Sun's parallax. Conversely, if it were possible to observe the coefficient of diurnal aberration accurately, we should thus have another way of finding the Sun's parallax. But the smallness of the diurnal aberration renders it im- possible to obtain good results by this method. THE DISTANCES OF THE SUN AND STARS. 311 EXAMPLES. XI. 1. Prove that cosec 876" = 23546 approximately, and thence that the distance of the Sun is nearly 81 million geographical miles, the angle 8' 76" being the Sun's parallax, and a geographical mile sub- tending 1' at the Earth's centre. 2. Find the Sun's diameter in miles, taking the Sun's parallax as 8'8", its angular diameter as 32', and the Earth's radius as 3,960 miles. 3. A spot at the centre of the Sun's disc is observed to subtend an angle of 5". What is its absolute diameter? 4. Show, by means of a diagram, that the general effect of the Earth's diurnal rotation is to shorten the duration of a transit of Venus, and that this circumstance might be used to find the Sun's parallax. 5. Supposing the equator, ecliptic, and orbit of Venus all to lie in one plane, and that a transit of Venus would last eight hours, at a point on the Earth's equator, if the Earth were without rotation ; show that, if the Sun is vertically overhead at the middle of the transit, the duration is diminished by about 9m. 55?s. owing to the Earth's rotation, taking the Sun's parallax to be 8'8", and the syn- odic period of Venus to be 586 clays. 6. If the annual parallax be 2", determine the distance of the star, taking the Sun's distance to be 90,000,000 miles. Hence, deduce the distance of a star whose pamllax is 0'2". 7. Find, roughly, the distance of a star whose parallax is 0'5", given that the Sun's parallax is 9", and the Earth's radius is 4000 miles. 8. The parallax of 61 Cyyni is O'o", and its proper motion, per- pendicular to the line of sight, is 5" a year; compare its velocity in that direction with that of the Earth in its orbit round the Sun. 9. Account for the following phenomena : (i.) all stars in the ecliptic oscillate in a straight line about their mean places in the course of the year ; (ii.) two very near stars in the ecliptic appear to Approach and recede from one another in the course of the year. 10. Suppose the velocity of light to be the same as the velocity of the Earth round the Sun. Discuss the effect on the Pole Star as- seen by an observer at the North Pole throughout the year. 312 ASTRONOMY. 11. Sound travels with a velocity 1,100 feet per second. Deter- mine the aberration produced in the apparent direction of sound to a person in a railway train travelling at sixty miles an hour, if the source of sound be exactly in front of one of the windows of the carriage. 12. Show that, in consequence of aberration, the fixed stars whose latitude is I appear to describe ellipses whose eccentricity is cos I. 13. How must a star be situated so as to have no displacement -due to (i.) aberration, (ii.) parallax? "Where must a star be so that the effect may be the greatest ? 14. On what stars is the effect of aberration or parallax to make them appear to describe (i.) circles, (ii.) straight lines? 15. Show that the effect of annual parallax on the position of a fitar may be represented by imagining the star to move in an orbit equal and parallel to the Earth's orbit, and that the effect of aber- ration may be represented by imagining it to revolve in a circle whose radius is equal to the distance traversed by the Earth while the light is travelling from the star. 16. Supposing the star 17 Virginia to be situated (as it nearly is) at the first point of Libra, find the direction and magnitude of its displacement due to aberration about the 21st day of every month of the year, taking the coefficient of aberration to be 20*5". When is its aberration greatest ? 17. At the solstices show that a star on the equator has no aber- ration in declination. If its R.A. be 22h., show that its time of transit is retarded at the summer and accelerated at the winter solstice by "68 of a second. 18. If the coefficient of aberration be 20", and an error of 2,000 miles a second be made in determining the-velocity of light, find, in miles, the consequent error in the value of the Sun's mean distance as computed from these data. 19. Show that when a planet is stationary its position is unaffected by aberration. 20. Taking the Earth's radius as 4,000, velocity of light 186,000 miles per second, show that the coefficient of diurnal aberration at the equator is about one-third of a second. THE DISTANCES OF THE SILX AND STARS. 313 MISCELLANEOUS QUESTIONS. 1. Explain the following terms: asteroid, libration, lunation parallax, perihelion, planet's elongation, right ascension, synodical period, gyxygies. ztn : .th. 2. Given that the R.A. of Orion's belt is 80, show by a figure its position at different hours of the night about March 21 and September 23. 3. Prove that the number of minutes in the dip is equal to the number of nautical miles in the distance of the visible horizon. 4. Show how to determine the latitude of a place by meridional observations on a circumpolar star, taking into account the refraction 5. Show how to find longitude from lunar distances. The cleared lunar distance of a star at 8h. 30m. local mean time is 150'45", and the tabular distances are 150'0" at 6h. and 151'30' / at 9h. of Green- wich mean time. Find the longitude. 6. At what time of the year can the waning moon best be seen ? 7. On July 21 at 2 A.M. the Moon is on the meridian. What is the age of the Moon ? Indicate the position on the celestial sphere of a star whose declination is and whose R.A. is 30. 8. Taking the distance of Venus from the Sun to be f of that of the Earth, find the ratio of the planet's angular diameters at superior and inferior conjunction and greatest elongation, and draw a series of diagrams showing the changes in the planet's appearance during a synodic period, as seen through a telescope under the same magnifying power. 9. Defining a lunar day as the interval between two consecutive transits of the Moon across the meridian, find its mean length in (i.) mean solar, and (ii.) sidereal units. 10. At what season is the aberration of a star least whose R.A. is $0 and whose declination is 60 ? 11. Show that the constant of aberration can be determined by observation of Jupiter's satellites, without a knowledge of the radius of the Earth's orbit. 12. How is it possible to calculate separately the aberration the constant of aberration being supposed unknown annual parallax, and proper motion of a star, from a long series of observations of the apparent place of a star ? 3H ASTRONOMY. EXAMINATION PAPEK. XL 1. Why is the method for finding the Moon's parallax not available- in the case of the Sun? Show how the determination of the parallax of Mars leads to the determination of the Sun's parallax. 2. Show how the Sun's parallax can be found by comparing the times of commencement or of termination of a transit of Venus at two- stations not far from the Earth's equator. 3. Show how the Sun's parallax can be found by comparing the- durations of a transit of Venus at two stations in high N. and S. latitudes. Why is this method not available when the transit is- central ? 4. Distinguish between solar and stellar parallax. Towards what point does a star seem to be displaced by heliocentric parallax ? Find an expression for the displacement. 5. Describe Bessel's method of determining the annual parallax of a fixed star. 6. How might the Sun's parallax be determined by observations of the eclipses of Jupiter's satellites? 7. Explain the aberration of light, and investigate the direction- and magnitude of the displacement which it produces on the- apparent position of a star. 8. Show that owing to aberration a star in the pole of the ecliptic appears to describe a circle, and that a star in the ecliptic appears-, to oscillate to and fro in a straight line during the course of the year^ 9. Show how the velocity of light may be determined from the aberration of.a star when the Sun's mean distance is known. 10. Investigate the general effects of diurnal aberration due to- the Earth's rotation about its axis. In what direction nre stars- displaced by diurnal aberration ? Show that the coefficient of diurnal aberration at a place in latitude I is K cos I, where K is the- coefficient at the equator. DYNAMICAL ASTRONOMY, CHAPTER XII. THE ROTATION OF THE EARTH. 383. Introductory. In the preceding chapters we have shown how the motions of the celestial bodies can be determined by actual ob crvation, and we have also explained certain resulting phenomena. But no use has yet been made of the principles of dynamics ; consequently we have been unable to investigate the causes of the various motions. In par- ticular, while we have assumed that the diurnal rotation of the stars is an appearance due to the Earth's rotation, we have not as yet given any definite proof that this is the only pos- sible explanation. The ancient Greeks accounted for the motions of the solar system by means of the Theory of Epicycles, according to which each planet moved as if it were at the end of a system of jointed rods rotating with uniform but different angular velocities. Suppose AB, BO, CD to be three rods jointed together at II, C. Let A be fixed ; let AB revolve uniformly about A ; let BC revolve with a different angular velocity about B ; and let CD revolve with another different angular velocity about C. Then, by properly choosing the lengths and angular velocities of the rods, the motion of J), relative to A, may be made nearly to represent the motion, relative to the Earth, of a planet. Copernicus (A.D. 1500 eirc.) was the first astronomer who explained the motions of the solar system on tho theory that the diurnal motion is due to the Earth's rotation, and that the Earth is one of the planets which revolve round the Sun. This theory was adopted by Kepler (A.D. 1609 circ.) whose laws of planetary motion have already been mentioned ( S26). A.STRON. Y 316 ASTRONOMY. These laws were, however, unexplained until their true cause was found by Newton (A.B. 1687) by his discovery of the law of gravitation. 384. Arguments in Favour of the Earth's Rota- tion. Without appealing to dynamical principles, the pro- bability of the Earth's rotation about its axis (87) may be inferred from the following considerations : (i.) If the Earth were at rest, we should have to imagine the Sun and stars to be revolving about it with inconceivably great velocities. If the Earth rotates, the velocity of a point on its equator is somewhere about 1,050 miles an hour. But since the Sun's distance is about 24,000 times the Earth's radius, the alternative hypothesis would require the Sun a body whose diameter is nearly 110 times as great as that of the Earth to be moving with a velocity 24,000 times as great, or about 25,000,000 miles an hour; while most of the fixed stars are at such distances from the Earth that they would have to move with velocities vastly greater than the velocity of light. It is inconceivable that such should be the case. (ii.) The diurnal rotations all take place about the pole, and are all performed in the same period a sidereal day. This uniformity is a natural consequence of the Earth's rota- tion, but it' the Earth were at rest, it could only be explained by supposing the stars to be rigidly connected in some manner or other. Were such a connection to exist it would be difficult to explain the proper motions of certain fixed stars, and the independent motions of the Sun, Moon, and planets. (iii.) By observing the motion of the spots on the Sun at different intervals, it is found that the Sun rotates on its axis. Moreover, similar rotations may be observed in the planets ; thus, Mars is known to rotate in a period of nearly 24 hours. There is, therefore, nothing unreasonable in suppos- ing that the Earth also rotates once in a sidereal day. (iv.) The phenomenon of diurnal aberration affords a proof of the Earth's rotation. Were it not for the difficulty of its observation, this proof alone would be conclusive. We may mention that diurnal parallax ^ould be equally well accounted for if the celestial bodies revolved round the Earth; not so, however, diurnal aberration, THE B.OTATIOTT OF THE EARTH. 317 385. Dynamical Proofs of the Earth's Rotation. The following is a list of the methods by which the Earth's rotation is proved from dynamical considerations : (1) The eastward deviation of falling bodies. (2) Eoucault's pendulum experiment. (3) Foucault's experiments with a gyroscope. (4) Experiments on the deviation of projectiles. (5) Observations of ocean currents and trade winds. (6) Experiments on the differences in the acceleration of gravity in different latitudes, due to the Earth's centrifugal force, as observed by counting the oscillations of a pendulum ; combined with (7) Observations of the figure of the Earth. 386. The Eastward Deviation of Falling Bodies. If the Earth is rotating p.Jbout its polar axis, those points which are furthest from the Earth's axis move with greater velocity than those which are nearer the axis. Hence the top of a high tower moves with slightly greater velocity than the base. If, then, a stone be dropped from the top of the tower, its eastward horizontal velocity, due to the Earth's rotation, is greater than that of the Earth below, and it falls to the east of the vertical through its point of projection. The same is true when a body is dropped down a mine. This eastward deviation, though small, has been observed, and affords a proof of the Earth's rotation. Consider, for example, a tower of height h at the equator. If a be the Earth's equatorial radius, the base travels over a distance 2ira in a sidereal clay, owing to the Earth's rotation, while the top of the tower describes 2ir(a + h) per sidereal day. Thus, the velocity at the top exceeds that at the bottom by 2irh per sidereal day. If h be measured in feet, the difference of velocities is irh/'SQOO indies per sidereal second, and is sufficiently great to cause a small but perceptible deviation when a body is let fall from a high tower. The earliest experiments were too rough to show this deviation, and were, therefore, used as evidence against, instead of for, the Earth's rotation. The deviation can only be observed in experi- ments conducted with very great care, and it is very difficult to measure. Its amount is largely modified by the resistance of the air and other causes, and therefore differs considerably from that by theory. 318 ASTRONOMY. 387. Foucault's Pendulum Experiment. In 1 85 1 , M . Foucault invented an experiment by which the Earth's rota- tion is very clearly shown. A pendulum is formed of a large metal ball suspended by a fine wire from the roof of a high building, and is set in motion by being drawn on one side and suddenly released ; it then oscillates to and fro in a vertical plane. If now the pendulum be sufficiently long and heavy to continue vibrating for a considerable length of time, the plane of oscillation is observed to very gradually change its direction relative to the surrounding objects, by turning slowly round from left to right at a place in the northern hemisphere, or in the reverse direction in the southern. If the experiment is performed in latitude ?, the plane of oscillation appears to rotate through 15 x sin I in a sidereal hour, 360 sin lin a sidereal day, or 360 in cosec I sidereal days. This apparent rotation is accounted for by the Earth's rotation, as follows. (i.) Let us first imagine the experiment to be performed at the north pole of the Earth. Let the pendulum AB be vibrating about A in the arc BB' in the plane of the paper. The only forces acting on the bob are the tension of the string BA and the weight of the bob acting vertically downwards ; both are in the plane of the paper. The Earth's rotation about its axis CA pro- duces no forces on the bob. Hence there is nothing whatever to alter the direction of the plane of oscillation ; this plane therefore remains fixed in space. But the Earth is not fixed in space ; it turns from west to east, making a complete direct revolution in a sidereal day. Hence the plane of the pendulum's oscillation appears, to an observer not conscious of his own motion, as though it rotated once in a sidereal day, in the reverse or retrograde direction (east to west). If, however, he were to compare the plane of oscillation not with the Earth but with the stars, whose directions are actually fixed in space, he would FIG. 129. THE ROTATION OF THE EARTH. 319 see that it always retained the same position relatively to them. Since, then, the pendulum at the pole of the Earth appears to follow the stars, it evidently appears to rotate in the same direction as the hands of a watch at the north pole, and in the direction opposite to the hands of a watch at the south pole. 7 (ii.) Next suppose the experiment performed at the Earth's equator. If the bob be set swinging in the plane of the equator, take this as the plane of the paper (Fig. 130). The direction of the vertical AQC is now rotating about an axis through C per- pendicular to the plane of the paper ; hence it always remains in that plane. Hence there is nothing whatever to turn the plane of oscillation of the pen- dulum out of the plane of the Earth's equator. It therefore continues always to pass through the east and west points, and there is no apparent rotation of the plane of oscillation. FIG. 130. If the pendulum do not swing in the plane of the equator, the explanation is much more complicated. As the Earth rotates, the direction of gravity performs a direct revolution in a sidereal day. Hence, relative to the point of support, gravity is gradually and continuously turning the bob west- wards, in such a way as to keep its mean position always pointed towards the centre of the Earth. When the bob is south of its position of equilibrium, this westward bias tends to turn the plane of oscillation in the clockwise direction, but when the bob is north of the mean position, the west- ward bias has an equal tendency to turn the plane in the reverse direction. Consequently the two effects counter- act one another, and therefore produce no apparent rotation of the plane of oscillation relative to surrounding objects. 320 ASTRONOMY. Qu (in.} Lastly, consider the case of an observer in latitude I (Fig. 131). Let w denote the angular velocity with which the Earth is rotating about its polar axis CP. It is a well-known theorem in Rigid Dynamics that an angular velocity of rotation about any line maybe resolved into components about any two other lines, by the parallelo- gram law, in just the same way as a linear velocity or a force along that line; this theorem is called the Parallelogram of Angular Velocities. Applying it to the angular velocity n about CP, we may resolve it into two components FIG 131. and n cosPCO or n sin I about CO, n sin PGO or n cos I about a line CO' perpendicular to CO, and we may consider the effects of the two angular velocities separately. As in case (i.), the component nsin I causes the Earth to turn about CO, without altering the direction in space of the plane of oscillation ; this plane, therefore, appears to rotate relatively in the reverse or retrograde direction, with angular velocity n sin I. As in case (ii.)> the angular velocity n cos I about CO' produces no apparent rotation of the plane of oscillation relative to the Earth. Hence the plane of oscilla- tion appears to revolve, relative to the Earth, with retrograde angular velocity n sin I. But the angular velocity n = 15 per sidereal hour = 360 per sidereal day. Therefore the plane of oscillation turns through 15 sin I per sidereal hour = 360 sin I per sidereal day, 360 fcnd its period of rotation = , - n sin I = cosec I sidereal days. THE IIOTATI02? OF THE EARTH. 321 388. The Gyroscope or Gyrostat is another apparatus used by Foucault to prove the Earth's rotation. It is simply a large spinning-top, or, more correctly, a heavy revolving wheel IT (Fig. 132), whose axis of rotation AB is supported by a framework, so that it can turn about its centre of gravity in any manner. Thus, by turning the wheel and the inner frame A CBD about the bearings CD, and then turning the outer frame DECF about the bearings EF, the axis AB (like the telescope in an altazimuth or equatorial) can be pointed in any desired direction. The three axes A B, CD, EF all pass through the centre of gravity of the top ; hence its weight is entirely supported, and does not tend to turn it in any way; and the bearings A, B, C, D, E, JPare very light, and so constructed that their friction may be as small as possible. The top may be spun by a string in the usual way, and it continues to spin for a long time. FIG. 132. When a symmetrical body, such as the wheel H, is revolv- ing rapidly about its axis of figure, and is not acted on by any force or couple, it is evident that no change of motion can take place, and therefore the axis of rotation AB must remain fixed in direction. This is the case with the gyro- scope, for, from the mode in which the weight of the wheel is supported, there is no force tending to turn it round. When the experiment is performed it is observed that the axis AB follows the stars in their diurnal motion ; if pointed to any star, it always continues to point to that star, its posi- tion relative to the Earth changing with that of the star. Hence it is inferred that the directions of the stars arc fixed in space, and that the diurnal motion is not due to them, but to the rotation of the Earth. 322 ASTllOXOMY. 389. If while the gyroscope is spinning rapidly any attempt be made to alter the direction of the axis of rotation AB by pushing it in any direction, a very great resistance will be experienced, and the axis will only move with great difficulty. This shows that the small friction at the pivots CD, EF can have but little effect in turning the axis of the top, and therefore the gyroscope spins as if it were practically free, as long as its angular velocity remains considerable. The following additional experiments with the gyroscope can be also used to prove the Earth's rotation. Experiment 1. Let the hoop CEDF be steadily rotated about the line EF. The line AB is no longer free to take up any position, for the pivots and D obviously force it always to be in a plane through EF and perpendicular to plane CEDF. Hence the axis of rotation is no longer able to maintain always the same position, unless that position coincides with EF. The result is that the axis gradually turns about CD till it does coincide with EF, the di- rection of rotation of the wheel being the same as that in which frame is forced to revolve. It will then have no further tendency to change its place. Of course we suppose the hoop turned so quickly that the effect of the slow motion of the Earth is imperceptible. Experiment 2. We may now repeat Experiment 1, using the Earth's rota- tion. Let the framework CEDFbe fixed in a horizontal position, the line CD being held pointed due east and west. The axis AB is then free to turn in the plane of the meridian. Now, owing to the Earth's rotation, the framework carrying CD is turning about the Earth's polar axis, and this causes the top to turn till its axis points to il\e celestial poles. The result of experiment agrees with theory, thus affording a further proof of the Earth's rotation about the poles. Experiment 3. Let the framework CEDF be clamped in a vertical plane. The axis AB can then turn in a horizontal plane, but it cannot point to the pole. It will, however, try to point in a direction differing as little as possible from the direction of the Earth's axis, and will therefore turn till it points due north and south. This has also been verified by actual observation. Experiments 3 and 2, if performed with a sufficiently perfect gyroscope, would enable us to find the north point, and then to find the celestial pole, and thus determine the latitude without observing any stars. By means of Foucault's pendulum experiment we could also (theoretically) determine the latitude,. THE ROTATION OF THE EARTH. 323 390. The Deviation of Projectiles. If we suppose a cannon ball to be fired in any direction, say from the Earth's North Pole, the ball will travel with uniform horizontal velocity in a vertical plane. But, as the Earth rotates from right to left, the object at which the ball was aimed will be carried round to the left of the plane of projection, and therefore the ball will appear to deviate to the right of its mark. At the South Pole the reverse would be the case, because in consequence of the direction of the vertical being reversed, the Earth would revolve from left to right ; hence the ball would deviate to the left of its mark. At the equator no such effect would occur. The deviation, like that in Foucault's pendulum, depends on the Earth's component angular velocity about a vertical axis at the place of observation, and this component, in latitude I, is n sin ( 387, iii.). Now the Earth rotates about the poles through 15" per sidereal second. Hence, if t be the time of flight measured in sidereal seconds, the deviation is = nt sin I = 15". t sin , and it is necessary to aim at an angle 15". t sin I to the left of the target in N. lat. /, or 15". t sin/ to the right in S. lat. I. The formula is sufficiently approximate even if t be measured in solar seconds. It is necessary to allow for this deviation in gunnery thus affording another proof of the Earth's rotation. 391. The Trade Winds are due to a similar cause. The currents of air travelling towards the hotter parts of the Earth at the equator, like the projectiles, undergo a deviation towards the right in the northern hemisphere, and towards the left in the southern. This deviation changes their directions from north and south to north-east and south-east respectively. In a similar manner the Earth's rotation causes a deviation in the ocean currents, making them revolve in a direction opposite to that of the Earth's rotation, which is "counter clockwise " in the N. and " clockwise " in the S. hemisphere. The rotatory motion of the wind in cyclones is also due to the Earth's rotation. 324 ASTEONOMY. 392. Centrifugal Force. If a body of mass m is revolving in a circle of radius r with uniform velocity v under the action of any forces, it is known that the body has an acceleration v*/r towards the centre of the circle.* Hence the forces must have a resultant mv*/r acting towards the centre, and they would be balanced by a force mv 2 /r acting in the reverse direction, i.e., outwards from the centre. This force is called the centrifugal force. Thus, in consequence of its acceleration, the body appears to exert a centrifugal force outwards. If it be attached to the centre of the circle by a string, the pull in the string is mv*/r. If m be measured in pounds, r in feet, and v in feet per second, then mv^/r represents the centrifugal force in poundals. Similarly, in the centimetre-gramme-second system of units, mv*/r is the centrifugal force in dynes. If n represent the body's angular velocity in radians per second, v = nr, and the centrifugal force is therefore mn*r. 393. General Effects of the Earth's Centrifugal Force. If the Earth were at rest the weight of a body would be entirely due to the Earth's attraction. But in con- sequence of the diurnal rotation the apparent weight is the resultant of the Earth's attraction and the centrifugal force. Let QOR represent a meridian section of the Earth (Fig. 134). Consider a body of mass m supported at any point on the Earth's surface. Since the Earth is nearly, but not quite, spherical, the force ^ of the Earth's attraction on a unit mass is not directed exactly to the Earth's centre, but along a line OK. But, owing to the body's central acceleration along OM, the force which it exerts on the support is not quite equal to the Earth's attraction mg^ but is compounded of mg Q acting along OJT, and the centri- fugal force m . ri* . MO acting along M 0. On KQ, take a point G such that KG * See any book on Dynamics. THE ROTATION OF THE EARTH. 325 then, by the triangle of force?, OG is the direction of the re- sultant force exerted by the body on its support, and this force is the apparent weight of the body. Hence, also OG represents the apparent direction of gravity, or the verti- cal as indicated by a plumb-line. Producing GO, KO to Z, Z", we see that the effect of centrifugal force is to displace the vertical from Z" towards the nearest pole (P). The angle ZGQ measures the (geographical) latitude of the place, and is greater than Z ' KQ, which would measure the latitude if the Earth were at rest. Hence the apparent latitude of any place is increased ly centrifugal force. I FIG. 134. Again, if the apparent weight be denoted by mg, we have, by the triangle of forces, g :y = GO: KQ-, now from the figure it is evident that G < IL(), and there- fore g < # . Hence the apparent iveight of a body is diminished by centrifugal force. 394. Effect on the Acceleration of a Falling Body. If a body is falling freely towards the Earth near 0, the whole acceleration of its motion in space is due to the Earth's attraction, and is # , along OK. But the Earth at has itself an acceleration ri*OM to wards 31. Hence the accelera- tion of the body relative to the Earth is the resultant of Now we have roughly ff = 32-18 feet per second per second, a = 3963 miles = 3903 x 5280 feet, and n = 2?r radians per sidereal day = radians per mean solar second. Hence A = 3963x5280x4.' = . 86164 x 86164 and therefore - = = nearly. ff 32-18 28J Hence = - 1 sin 2? ~ 289 2 Since d is small, this gives approximately 1 sin 21 circular measure ot d = -^- - FIG. 135. " (number of seconds in 180x60x60 289X27T 206265 sin 21 578 sin 21 = 357" sin Hence the deviation D = 5' 57". sin 21, and this is the in- crease of latitude due to centrifugal force. COROLLAEY. The deviation of the vertical due to centri- fugal force is greatest in latitude 45 (v sin 2? = 1), and is there 5' 57". * Since the Earth is not quite spherical, g is not the same at as at the equator. The difference may be neglected, however, when multiplied by the small constant jy. 328 ASTRONOMY. 397. Figure of the Earth. In 114 we stated that the form of the Earth has been observed to be an oblate spheroid. Now it has been proved mathematically that a mass of gravitating liquid- when rotating takes the form of an oblate spheroid whose least diameter is along its axis of rotation. Thus the Earth's form may be accounted for on the theory that the Earth's surface was formerly in a fluid or molten state, and that it then assumed its present form, owing to its diurnal rotation. We thus have another argument in favour of the Earth's rotation ; but it is only fair to say that this theory of the Earth's origin has not been satisfactorily demonstrated. It accounts satisfactorily, however, for the form of the surface of the ocean. This theory may be illustrated by the following general considera- tions. When a mass of liquid is acted on by no forces beyond the attractions of its particles, it is easy to realize that the whole is in equilibrium in a spherical form, being then perfectly symmetrical. If, however, the fluid be rotating about the axis PGP', the centri- fugal force tends to pull the liquid away from this axis and towards the equatorial plane. The liquid would, therefore, fly right off, but its attraction is always trying to pull it back to the spherical form. Hence, the only effect of centrifugal force (which, for the Earth, is small compared with gravity) is to distort the liquid from its spheri- cal form by pulling it out towards the equator ; and it is therefore reasonable to suppose that the fluid will assume a more or less oblate figure, whose equatorial is greater than its polar diameter. It may also be remarked that the form assumed by the liquid would be such that the effective force of gravity (i.e., the resultant of the attraction and centrifugal force) on the surface would every- where be perpendicular (i.e., normal) to the surface. *398. Gravitational Observations. If the Earth were a sphere, its attraction g would everywhere tend to its centre, and would be of the same intensity at all points on its surface, while the variations in g, the apparent intensity of gravity, would be entirely due to the Earth's centrifugal force, its value in latitude I being proportional to 1 -^-g cos- 1 ( 396). By comparing the values of g at different places, we should then be able to demonstrate the Earth's centri- fugal force, and hence prove its rotation. But, owing to the Earth's ellipticity, its attraction gr does not pass through the centre, except at the poles and equator, and its intensity is not everywhere con- stant. It is, therefore, important to determine experimentally the values of g at different stations. By allowing for centrifugal force, the corresponding values of the Earth's attraction g can be found, and the variations in its intensity at different places afford a measure of THE ROTATION OF THE EARTH. 329 tlie amount by which the Earth differs from a sphere. We thus have a gravitational method of finding the Earth's ellipticity. But the Earth's ellipticity can also be determined by direct obser- vation, as explained in Chapter III., Section III. The agreement between the results thus independently obtained furnishes another proof of the Earth's rotation. In consequence of the EarthVellipticity it is found (by observa- tion) that the difference in the intensity of gravity between the polo and equator is increased from ^-g- to -3^-5 of the whole. 399. To compare the Intensity of Gravity at different places. The intensity of gravity may be measured by the force with which a body of unit mass is drawn towards the Earth. This cannot be measured by weighing a body with a common balance, because the weights of the body and of the counterpoise, by means of which it is weighed, are equally affected by variations in the intensity of gravity, and two bodies of equal mass will, therefore, balance one another when placed in the scale pans, no matter what be the intensity of gravity. In fact, by weighing a body with weights in the ordinary way, we determine only its mass, and not the absolute force with which it is drawn to the Earth. We might determine the intensity of gravity by means of a " spring balance," for the elasticity of the spring does not depend on the intensity of gravity, and therefore the extension of the spring gives an absolute measure of the force with which the body is drawn towards the Earth. If the apparatus were to support a mass of one pound, first at the equator and then at the pole, the force on it would be greater at the latter place by about - l ^, and this spring would thcro be extended about -j--^ more. It would be very difficult to construct a spring balance sufficiently sensitive to show such a small relative difference of weight, but it has been done. Aticood's machine might be used to find g, but this method is not capable of giving very accurate results. The most accurate method of finding g is by timing the oscillations of a pendulum of known length. [* A theoretical simple pendulum, consisting of a mere heavy par- ticle of no dimensions, suspended by a thread without weight, is of course impossible to realize in practice, but the difficulty is over- come by the use of a pendulum called Captain Rater's Reversible Pendulum. This pendulum is a bar which can be made to swing ab ut either of two knife-blades fixed ?.t opposite sides of, but un- equal distances from, its centre of gravity, and it is so loaded that the periods of oscillation, when suspended from either knife-edge, are equal. It is then known that the pendulum will swing about either knife-edge in just the same manner as if it were a simple pendulum whose whole mass was concentrated at the other knife- edge. The distance between the knife-edges is, therefore, to be regarded as the length of the pendulum.'] 330 400. Oscillation? of a Simple Pendulum. In a simple pendulum, formed of a small heavy particle suspended by a fine light thread of length I, the period of a complete oscillation to and fro is the time of a single swing or " leat " being of course half of this. Hence by observing the time of oscillation t and measuring the length I, the intensity of gravity g can be found. By the " seconds pendulum " is meant a pendulum in which one beat occupies one second, hence a complete oscillation occupies two seconds. EXAMPLK. Having given that the length of the seconds pendulum is 99'39 centimetres, to find g in centimetres per second per second. t = 2nVZ/7 = 2 seconds, and I = 99'39 cm., .-. g =^i = 99-39 x (3-1416) 2 = 981. It is often necessary to compare the lengths of two pendulums whose periods of oscillation are very nearly equal, to find the effect of small changes in the length of a pendulum due to variations in temperature, or, in comparing the intensity of gravity at different places, to find the effect of a small alteration in the value of g on the period of oscillation and on the number of oscillations in a given interval. If the differ- ences are small, the calculations may be much simplified by means of the following methods of approximation.* 401. To find the change in the time of oscillation of a pendulum, and in the number of oscillations in a given interval, due to a small variation in its length or in the intensity of gravity. If t be the time of a complete oscillation of a pendulum of length J, we have, by 400, ? = 47T 2 - (i). * The same results can of course be obtained by means of the differential calculus. THE HOTATIOX OF THE EARTH. 331 (i.) Suppose the length increased to 1' 9 and let t' be the new period of oscillation. We have t* = 47T 3 -. g Therefore, by division, *"_r *T*T and therefore also I'*-?-,, ..t + t I'-l ~1T ~ t} ~T These formulae are exact. But if I' is very nearly equal to , t' is very nearly equal to t, and therefore, putting t + t'= 2t, we have approximately Q t'-t_l'-l T T' whence, if t, I be known, the change V t, consequent on the increase of length I' I, may be readily found approximately without the labour of extracting any square roots. (ii.) Suppose the intensity of gravity increased to g', the length I being unaltered, and let t' be the new period. Since we have, by division, and therefore also But, if , g are very nearly equal to t', /, this gives approximately 2 """ = - . * f/ ASTEON. Z 332 ASTEONOMf. (iii.) If I and g both vary, becoming V and g', we have, in like manner Therefore also I 9 or approximately, if the variations are small, 2 *-* = r ~ ? - ff '- ff t I the Phenomena at so^^mf 80 deSCnbeth P ondin g phenomena in the . 3. If a railway is laid along a meridian, and a train is travelling anTast^To " ^^^ P ole > "yeBtigate whether it wHIexert an eastward or a westward thrust on the rails, and why 4. A bullet is fired in N. latitude 45, with a velocity of 1 600 frpf 5 Ihn ft ' f f?i I how many feet it will deviate to the right. a bodv 7ti * 6 Ear S W6re t0 r tate ^venteen times as fist a body at the equator would have no weight. 6. If the Earth were a homogeneous sphere rotating so fast thaf ^ f * he last q uest! . show that the Earth's 336 ASTRONOMY. EXAMINATION PAPEB. XII. 1. Give reasons for supposing that the diurnal rotation of th( heavens is only an appearance caused by a real rotation of the Earth. Name methods by which it has been claimed that this ii proved. 2. Describe the gyroscope experiment, and the gyroscope. 3. Give any theoretical methods of determining latitude withou observing a heavenly body. 4. Describe Foucault's experiment for exhibiting the Earth' rotation ; and find the time of the complete rotation of the plane c vibration of a simple pendulum fieely suspended in latitude 60. 5. Having given that the Earth's circumference is 40,000 kilc metres, find the acceleration of a body at the equator due to th Earth's rotation in centime bres per second per second, and takinj <7 , the acceleration of gravity, to be 981 of these units, deduce t ratio of centrifugal force to gravity at the equator. 6. What is meant by the vertical at any point of the Earth surface ? Supposing the Earth to be a uniform sphere revolvir round a diameter, calculate the deflection of the vertical from t normal to the surface. 7. State what argument is drawn from the Earth's form to suppo the hypothesis of its rotation. 8. Why is it that the intensity of gravity is less at the equal than in higher latitudes ? Show that the alteration in the appare weight of a body due to centrifugal force varies nearly as cos where I is the latitude, and state the ratio of centrifugal force gravity at the equator. 9. If a body is weighed by a spring balance in London and Quito, a difference of weight is observed. Why is this not observed" an ordinary pair of scales be used ? 10. Show that an increase in the intensity of gravity will cat a pendulum to swing more rapidly, and vice vers&. If the accele: tion of gravity be increased by the small fraction l/r of its vali show that a pendulum will gain one complete oscillation in every CHAPTER XIII* THE LAW OF UNIVERSAL GRAVITATION. SECTION I.^T/ie Earttis Orlital Motion Kepler's Laws and their Consequences. 405. Evidence in favour of the Earth's Annual Motion round the Sun. The theory that the Earth is a planet, and revolves round the Sun, was propounded by Copernicus (circ. 1530) and received its most convincing proof, over 150 years later from Newton (A.D. 1687), who accounted for the motions of the Earth and planets as a consequence of the law of universal gravitation. This proof is based on dynamical principles ; but the following arguments, based on other considerations, afford independent evidence in favour of the theory that the Earth revolves round the Sun rather than the Sun round the Earth. (i.) The Sun's diameter is 110 times that of the Earth's, and it is much easier to believe that the smaller body revolves round the larger, than that the larger body revolves round the smaller. If the dynamical laws of motion be assumed, it is impossible to gee how the larger body could revolve round the smaller, unless either its mass and. therefore its density were very small indeed, or the smaller one were rigidly fixed iu some way. (ii.) The stationary points, and alternately direct and retro- grade motions of the planets, are easily accounted for on the theory that the Earth and planets revolve round the Sun (Chap. X.) in orbits very nearly circular, and it would be impossible to give such a simple explanation of these motions on any other theory. It is true that we might suppose, with Tycho Erahe (circ. 1600), that the planets revolve round the Sun as a centre, while that body has an orbital motion round the Earth, but this explanation would be more complicated than that which assumes the Sun to be at rest. And it would be hard to explain how such huge bodies as Jupiter and Saturn could be brought to describe such complex paths. 33$ ASTRONOMY. (iii.) As seen through a telescope, Venus and Mars are found to be very similar to the Earth in their physical charac- teristics, and their phases show that, like the Earth and Moon, they are not self-luminous. It is, therefore, only natural to suppose that their property of revolving round the Sun is shared by the Earth. Moreover, the Earth's relative distance from the Sun agrees fairly closely with that given by Bode's law ; hence there is a strong analogy between the Earth and the planets. (iv.) The orbital motion of the Earth is in strict accordance with Kepler's Laws of Planetary Motion. In particular, the relation between the mean distances and periodic times given by Kepler's Third Law ( 326) is satisfied in the case of the Earth's orbit. Moreover, a similar relation is observed to hold between the periodic times of Jupiter's satellites and their mean distances from Jupiter. Hence it is probable that the Earth and planets form, like Jupiter's satellites, one system revolving about a common centre. But it is improbable that the Sun and Moon should both revolve about the Earth, for their distances from it and their periods are not connected by this relation. (v.) The changes in the relative positions of two stars during the year in consequence of annual parallax can only be accounted for on the hypothesis either of the Earth's orbital motion, or of a highly improbable rigid connection between all the nearer stars and the Sun, compelling them all to execute an annual orbit of the same size and position. (vi.) The aberration of light affords the most convincing proof of all. In particular, the relation between the coefficient of aberration and the retardation of the eclipses of Jupiter's satellites has been fully verified by actual observations, and affords incontestible evidence that the phenomenon is actually due to the finite velocity of light, as explained in Chapter XI. And the alternative hypothesis which would account for annual parallax would not give rise to aberration, but would produce an entirely different phenomenon. Hence the evi- dence derived from the aberration of light, unlike the previous evidence, furnishes a conclusive proof, and not merely an argument, in favour of the Earth's orbital motion. THE LAW OF UNIVERSAL GRAVITATION. 339 406. NEWTON'S THEORETICAL DEDUCTIONS FROM KEPLER'S LAWS. Kepler's Three Laws of planetary motion naturally suggest the following questions : (1) What makes the planets move in ellipses ? (2) Why does the radius vector from the Sun to any planet trace out equal areas in equal times ? (3) Why are the squares of the periodic times proportional to the cubes of the mean distances from the Sun ? These questions were first answered by Newton about 1687, or nearly sixty years after the death of Kepler. The theore- tical interpretation of the Second Law necessarily precedes that of the first; accordingly we now repeat the laws in their new order, together with Newton's interpretations of them. Kepler's Second Law. The radius vector joining each planet to the Sun moves in a plane describing equal areas in equal times. NEWTON'S DEDUCTION. The force under which a planet describes its orbit always acts along the radius vector in the direction of the Sun's centre. Kepler's First Law. The planets move in ellipses, having the Sun in one focus. NEWTON'S DEDUCTION. The force on any planet varies inversely as the square of its distance from the Sun. Kepler's Third Law. The squares of the periodic times of the several planets are proportional to the cubes of their mean distances from the San. NEWTON'S DEDUCTION. The forces on different planets vary directly as their masses, and inversely as the squares of their distances from the Sun, or, in other words, the accelerations of different planets, due to the Sun's attraction, vary inversely as the squares of their distances from the Sun. 340 ASTRONOMY. .If, as we have every reason for believing, the planets are material bodies, Newton's laws of motion show that they cannot move as they do unless they are acted on by some force, otherwise they would either be at rest or move uni- formly in a straight line. Kepler's Second Law then enables us to determine the direction of this force, his First Law enables us to compare the force at different parts of the same orbit, and his Third Law enables us to compare the forces on different planets. 407. We have seen that the orbits of most of the planets are nearly circular, the eccentricities being small, except in the case of Mercury. Before proceeding to the general discussion of the dynamical interpretation of Kepler's Laws, it will be convenient therefore to consider the case where the orbits are supposed circular, having the Sun for centre. Kepler's Second Law shows that under such circumstances the planets will describe their orbits uniformly, and it hence follows that the acceleration of a planet has no component in the direction of motion, but is directed exactly towards the centre of the Sun. The law of force can now be deduced very simply, as follows : KEPLER'S THIRD LAW FOE, CIRCULAR ORBITS. 408. To compare the Sun's attractions on different Planets, assuming that the orbits are circular and that the squares of the periodic times are propor- tional to the cubes of the radii. Suppose a planet of mass J/is moving with velocity v in a circle of radius r. Let T be the periodic time, P the force to the centre. . .', . Since the normal acceleration in a circular orbit is 2 /r, therefore * In the period T 7 the planet describes the circumference lira ; .-. vT= 2vr. Substituting for v, we have P iii^: _ ^L ***** jf ~- r 2 -yT TfLE LAW OF UNIVERSAL GRAVITATION. $4l Let M 1 be the mass of another planet revolving in a cir- cular orbit of radius r', T its periodic time, P' the force of the Sun's attraction ; then we have in like manner p ,_ JT 4*V - ^ x r * ' By Kepler's Third Law, r- r Therefore the forces on different planets vary directly as their masses and inversely as the squares of their distances from the Sun. COEOLLAEY 1. Let P = CM/r* ; then C is called the abso- lute intensity of the Sun's attraction, and we see that The absolute intensity of the Sun's attraction is the same for all planets. For c The constant C evidently represents the force with which the Sun would attract a unit mass at unit distance, or the acceleration which the Sun would produce at unit distance. COROLLAS r 2. If another body be revolving in an orbit of radius / in a period T\ under a different central force, which produces an acceleration C"// 2 at distance r', we have tT=l=! and (7= .-. C'T : CT 2 = i : r 8 , a formula which enables us to compare the absolute intensities of two different centres of force, which attract inversely as the squares of the distances, when the periodic times and distances of two bodies revolving about them are known. 342 ASTRONOMY. 409. To compare the velocities and angular velo- cities of two planets moving in circular orbits. If v, v are the velocities, n, ri the angular velocities (in radians per unit time), we have Also v = rw, v = rn' ; .*. v:v'= r~*: r'~*. EXAMPLES. 1. If the Earth's period were doubled, to find what would be its new distance from the Sun. If r, r' be the old and new distances, Kepler's Third Law gives r' 3 : r 3 = 2 2 : 12; /. S = r x */4 = 92,000,000 x T587 = 146,000,000 miles. 2. If the Earth's velocity were doubled, its orbit remaining cir- cular, to find its new distance. Here r' : r = v 2 : t/ 2 = 1 : 4 ; ... r'=ir = 23,000,000 miles. 3. If the Earth's angular velocity were doubled, to find its new distance. The new angular velocity being double the old, the new period would be half the old, and therefore r' 3 :*- 3 =()': I'; /. r' = r x */i = r/ V4 = 92,000,000 -f- T587 = 92,000,000 x -63 = 58,000,000 miles. 4. To find what would be the coefficient of aberration to an observer situated on Venus. The coefficient of aberration (in circular measure) is the ratio of the observer's velocity to the velocity of light ; hence, if fc, k' are the coefficients on the Earth and Venus, _ = t/ r^ = \r_ /100 k v r-* V r' V 72 ; .. k' = 20-493" x A/(l-38*) = 20'493" x M785 = 24-151". THE LAW OF UNIVERSAL GRAVITATION. 343 We shall now prove Newton's deductions from Kepler's Laws, for the general case of elliptic orbits, employing, how- ever, different and simpler proofs to those used by Newton. 410. Areal Velocity. definition. If a point P is moving in any path MPK about a centre S, the rate of increase of the area of the sector MSP, bounded by the fixed line SM and the radius vector SP, is called the areal velocity of P about the point S. If the radius vector SP describes equal areas in equal times, in accordance with Kepler's Second Law, the areal velocity of P about S is of course constant, and is then measured by the area of the sector described in a unit of time. If the rate of description of areas is not constant, we must, in measuring the areal velocity at any point, pursue a similar course to that adopted in measuring variable velocity at any instant, as follows : FIG. 136. If the radius vector describes the sector PSP' in the inter- val of time t, then the average areal velocity over the arc PP' is measured by the ratio area PSP' time t (Thus the average areal velocity is the areal velocity with which a radius vector, sweeping out equal areas in equal times, would describe the sector PSP' in the same time t.) The areal velocity at a point P is the limiting value of the average areal velocity over the arc PP when this arc is infinitisimally small. 344 ASTRONOMY. 411. Relation between the Areal Velocity and the Actual (linear) Velocity. Let PP' be the small arc described by a body in any small interval of time t. Let be the actual or linear velocity of the body, h its areal velocity. Since the arc PP is supposed small, we have PP'=vt, area PSP'=M. Draw S Y perpendicular on the chord PP' produced. Then &PSP'= | (base) x (altitude) or FIG. 137. But when the arc PP' is infinitesimally small, PFis the tangent at P, and SYis therefore the perpendicular from S on the tangent at P. If this perpendicular be denoted we have therefore or (areal vel. about S) = J (velocity) x (perp. from S on tangent). COROLLARY. Por planets moving in circular orbits of radii r, r, h = |IT, and h'= \v'r r . But v I v = r"* : r x ~ J ; A:A'=r:r'; hence the areal velocity of a planet moving in a circular orbit is proportional to the square root of the radius, . . THE LAW OP UNIVERSAL GRAVITATION. 345 412. PROPOSITION I. If a particle moves in such a manner that its areal velocity about a fixed point is constant, to prove that the resultant force on the particle is always directed towards the fixed point. [Newton's Deduction from Kepler's Second Law.] Let a body be moving in the curve PQ in such a way that its areal velocity about S remains constant. Let v, v' be the velocities at P, Q, and let PF, QY, the corresponding directions of motion, intersect in R. Drop SY, S Y perpen- dicular on PF, QY. Since the areal velocities at P and Q are equal, .-. v.SY=v'. SY. But SY= Rsin&KF, SY = SItsmSltY. .-. v$wSRY=v sin FIG. 138. i.e.) Component velocity at P perpendicular to BR = component vel. at Q pcrp. to SB. Therefore, as the particle moves from P to Q, its velocity perpendicular to JRS is unaltered, and therefore the total change of velocity is parallel to ItS. This is true whether the arc PQ be large or small. But if the arc PQ be taken infinitesimally small, the average rate of change of velocity over PQ, measures the acceleration at P, and P coincides with P. Therefore the direction of the acceleration of the particle nt any point of its path always passes through S, and there- fore the force acting on the particle also always passes through S. 346 ASTEONOMT. 413. Conversely, if the force on the particle always passes through 8, the areal velocity about 8 remains constant. For in passing from P to Q, the direction of motion is changed from PR to EQ, and the same change of velocity could therefore be produced by a suitable single blow or instantaneous impulse acting at R. And since the force at every point of PQ always passes through 8, this equivalent impulse must evidently also pass through 8 ; it must therefore act along RS. Hence the velocity perpendicular to R8 is unaltered by the whole impulse, and is the same at P as at Q j therefore FIG. 139. v sin 8RT = v' sin 8RT j therefore v .SY = v' .SY 1 ; therefore areal vel. at P = areal vel. at Q. 414. PROPOSITION II. A particle describes an ellipse under a force directed to wards the focus ; to show that the force varies inversely as the square of the dis- tance. [Newton's Deduction from Kepler's First Law.] If h is the constant areal velocity, we have, by (i.), We will now express the kinetic energy of the particle in terms of r, its distance from the focus. Let its mass be M. In the Appendix (Ellipse 11) it is proved that for the ellipse whose major and minor axes are 20, 2J, m, j. 2 4# 4tfa / 2 1 \ Therefore t? 8 = - = ^- ( --- J. jt? 2 i 2 \ r a r and kinetic energy at distance r THE LAW OP T7NIVEESAI, GBAVITATION. 347 If v is the velocity at distance r', we have, similarly, and therefore, for the increase of kinetic energy, (in.). FIG. 140. Now the increase of kinetic energy is equal to the work done by the impressed force in bringing the particle from distance r to distance r. The resolved part of the displace- ment in the direction of the force is rr'. Hence if P' denote the average value of the force between the distances r and r', we have Work done = P' (r-r'} = JJf^-^W = ~^(~ - -M b* rr rr Put / = r ; then the average force P' becomes the actual force P at distance r. Therefore A. j- j. \ orce at distance r) = 2 This is proportional to 1/r 2 . Therefore the force varies inversely as the square of the dis- tance. \STRON. 2 A 348 ASTRONOMY. 415. PROPOSITION III. Having given that the squares of the periodic times of the planets are proportional to the cubes of the semi-axes major of their orbits, to compare the forces acting on different planets. [Newton's Deduction from Kepler's Third Law.] Let T be the periodic time of any planet; then, by hypothesis, the ratio is the same for all planets. In the last proposition (vi.) we showed that the force at distance r is given by -p _ Let this be put = Jf(7/r 2 , where C is some constant ; then 4h~a , .. N = ........................ ( yu -)- Now in the period T the radius vector sweeps out the area of the ellipse, and this area is nab (Appendix, Ellipse 13). Hence, since the areal velocity is h, we have hT= irab. Substituting the value of h from this equation in (vii.), we have But a*/T 2 is the same for all the planets ; therefore C is con- stant for all the planets, and since the force it follows that The forces on different planets are proportional to tlieir masses divided by the squares of their distances from the Sun. Or, as in 408, Cor. 1, 27)e absolute intensity of the Sun's attraction ( C) is the same for all the planets. CoROLLAjtY. Let accented letters refer to the orbit of another particle revolving round a different centre of force of intensity C'. Then, by (viii.), FC: T'*C r = a s : a'\ THE LAW OP UNIVERSAL GRAVITATION. 349 416. Other Consequences of Kepler's Laws. (i.) In 150 we showed that, in consequence of Kepler's Second Law being satisfied by the Earth in its annual orbit, the Sun's apparent motion in longitude is inversely propor- tional to the square of the Earth's distance from it. Since the areal velocity of any planet about the Sun always remains constant^ it may be shown in like manner that its angular velocity is inversely proportional to the square of its distance from the Sun. FIG. 141. For, if the planet's radius vector revolves from SP to SP in the time t, and if the arc PP' is very small, we have area SPP' = SP* x Z PSP' ( 150), the angle being measured in radians ; area SPP' _ i o p2 v tPSP ___ = f * __, i.e., (areal velocity of P) = %SP* x (angular velocity of P), provided that the angular velocity is measured in radians per unit of time. If n denote the angular velocity, h the areal velocity, and r the distance SP, we have therefore And since h is constant, n is inversely proportional to r. * (ii.) If the mass of the planet is M, its momentum is Mv along PY, and the moment of this momentum about 8 is = Mv x 8T = Mvp = 2hM. ( 411.) This is the planet's angular momentum, and is constant, since 7i in- constant. 350 ASTRONOMY. *417. Having given, in magnitude and direction, the velocity of a planet at any point of its orbit, to construct the ellipse described under the Sun's attraction. Let the attraction at distance r be 0/r 2 per unit mass, where C is given. Suppose that at the point P of the orbit the planet is moving with velocity v in the direction PT. We have v x ST = 2h, which determines h. Also, from (vii.), G = 47i 2 a/6 2 . Substituting in (ii.), *-c(- 1)...(,). Hence, by considering the planet at P, we have SP Now v, G, and SP are known ; hence the last equation determines the semi-axis major a. If r = SP, we have 2a 2C-ru 2 ' Let H be the other focus of the ellipse. Then it is known (Ellipse 8) that HP, SP make equal angles with PT. Also SP + HP = 2a. Hence, we can construct the position of If by making / TPI = / TPS, and producing IP to a point H such that PH = 2a-SP. The ellipse can now be constructed as in Appendix (Ellipse 2). COROLLARY 1. Since SP + HP = 2a, equation (x.) gives SP.a COROLLARY 2. Substituting for h in terms of G, we see from equation (iv.) that the work done when the body moves from dis- tance r to distance / is *jfafJL-4. * This result is also proved independently in many treatises on dynamics, but a fuller investigation would be out of place here. THE LAW OF UNIVERSAL GBAVITATION. 351 Hence the work done by a mass M in falling from distance 2a to distance r is = MQ (--- = iMu 2 byfxi. \ r 2a/ kinetic energy of the planet when at distance r. Therefore, if a circle be described about the centre of force 8, with radius equal to the major axis 2a, the velocity at any point of the orbit is that which the planet would acquire by falling freely from the circle to that point under the action of the attracting force. COROLLARY 3. If the planet be revolving in a circle, r a, and therefore v 2 = C/r = C/a, as in 408. COROLLARY 4. If v 3 = 2C/r, (x.) gives I/a = 0; /. a = oo. Hence the velocity is that acquired by falling from an infinite distance. In this case, the orbit is not an ellipse, but a parabola, a conic section satisfying the " focus and directrix " definition of Appendix (1), but having its eccentricity equal to unity. If v z > 2C/r, the velocity is greater than that due to falling from infinity, a comes out negative, and the orbit is a hyperbola, a conic section satisfying the focus and directrix definition, but having its eccentricity e greater than unity. A few cornel- have been observed to describe parabolas and hyper- bolas .-ibout the Sun. In such a case the motion is not periodic; the comet gradually moves away TO an infinite distance, and is lost for ever, unless the attraction of some other heavenly body should happen to divert its course, and send it back into the solar system. EXAMPLE. To find how long the Earth would take to fall into the Sun if its velocity were suddenly destroyed. If the Earth's velocity were very nearly, but not quite destroyed, it would describe a very narrow ellipse, very nearly coinciding with the straight line joining the point of projection to the Sun. The major axis of this ellipse would be very nearly equal to the Earth's initial distance from the Sun, and therefore the Earth would have very nearly gone half round the narrow ellipse when it would collide with the surface of the Sun. Hence, if r denote the Earth's distance from the Sun, the semi- rnajor axis of the narrow ellipse is \r, and the periodic time in this ellipse would be ()* years. The Earth would therefore collide with the Sun in 2 x (1)^ years = years = - years = x 1-4142 days = 64 days nearly. 8 352 ASTRONOMY. SECTION II. Newton's Law of Gravitation Comparison of the Masses of the Sun and Planets. 418. In the last section we showed that the Sun attracts any planet of mass M at distance r with a force CM/r 1 , where C is a constant. If we assume the truth of Newton's Third Law of Motion (i.e., that action and reaction are equal and opposite), the planet must also attract the Sun with an equal and opposite force CM/r 2 . Since in the former case the force is proportional to the mass of the attracted body, and in the latter to the mass of the attracting body, it is reasonable to suppose that the attraction between two bodies is propor- tional to the mass of each. Moreover, the motions of the various satellites, such as the Moon, confirm the theory that they revolve in their orbits under the attraction of their respective primary planets. From evidence of this character Newton, after many years of careful investigation, enunciated his Law of Universal Gravitation, which he stated thus : Every particle in the universe attracts every other particle with a force proportional to the quantities of matter in each, and inversely proportional to the square of the distance between them. By quantity of matter is, of course, meant mass, and the word attracts implies that the force between two particles acts in the straight line joining them and tends to bring them together. If M, M' be the masses of two particles, and r the distance between them, the law asserts that either particle is acted on by a force, directed towards the other, of magnitude where k has the same value for all bodies in the universe. The constant is called the constant of gravitation. *419. Astronomical Unit of Mass. Taking any fundamental units of length and time, it is possible to choose the unit of mass such that fc = 1. This unit is called the astronomical unit of mass. Hence, if M, M ' are expressed in astronomical units, the force between the particles is equal to MH'jr". It is, however, usually more convenient to keep the unit of mass arbitrary, and to retain the constant fc. THE LAW OF UNIVERSAL GRAVITATION. 353 420. Remarks on the Law of Gravitation. Newton's Law states that not only do the Sun, 'the planets and their satellites, and the stars, mutually attract one another, but every pound of matter on one celestial body attracts every other pound of matter, on either the same or another body. But it is well-known that two spheres attract one another in just the Fame way as if the whole of the mass of either were concentrated at its centre, provided that the spheres are either homogeneous or made up of concentric spherical layers, each of uniform density. Since the Sun and planets are very nearly spherical, and their dimensions are very small compared with their distances, we see that their attractions may be very approximately found by regarding them as mere particles, instead of taking separate account of the individual particles forming them. Moreover, every planet is attracted by every other planet, as well as by the Sun. But the mass of the Sun, and con- sequently its attraction, is so much greater than that of any other member of the solar system, that the planetary motions are only very slightly influenced by the mutual attractions. Kepler's Laws, therefore, still hold approximately, but the orbits are subject to small and slow changes or perturbations. The Moon , on the other hand, is far nearer to the Earth than to the Sun ; hence the Moon's orbital motion is mainly due to the Earth's attraction. The chief effect of the Sun's attraction on the Earth and Moon is to cause them together to describe the annual orbit ; but it also produces pertur- bations or disturbances in the Moon's relative orbit ( 272) with which we are not here concerned. The fixed stars also attract one another and attract the solar system, which in its turn attracts the stars. The proper motions of stars are probably due to this cause ; but when we consider the vast distances of the stars, and remember that the attraction varies inversely as the square of the distance, it is evident that the relative accelerations are mostly too feeble to have produced any sensible changes of motion within historic times, and that countless ages must elapse before such changes can be discerned. 354 ASffcONOMf* 42 1 . Correction of Kepler's Third Law. Prom the fact that a planet attracts the Sun with a force equal to that with the Sun attracts the planets, it may he shown that Kepler's Third Law cannot he strictly true, as a consequence of the law of gravitation. Not only will the planet move under the Sun's attraction, but the Sun will also move under the planet's attraction. Eut since the forces on the two hodics are equal, while the mass of the Sun is very great compared with the mass of any planet, it follows that the acceleration of the Sun is very small compared with that of the planet, and hence the Sun remains very nearly at rest. We may, however, obtain a modification of Kepler's Third Law, in which the planet's reciprocal attraction is allowed for as follows : ,, Let S, M be the masses of the Sim and planet; then the attraction betweeen them is This attraction, acting on the mass JJ/of the planet, produces an acceleration of the planet towards the Sun equal to The corresponding attraction on the mass 8 of the Sun pro- duces an acceleration, in the reverse direction, of Hence the whole acceleration of the planet relative to the Sun is iM, yd instead of kS/r z , as it would be if the Sun were at rest. Hence the absolute intensity of the planet's acceleration towards the Sun is k (S + M), and this depends on the values of both M and 8. Let now T be the periodic time, r the planet's mean distance from the Sun, or the semi-axis major of the relative orbit ; then, by 408 (for a circular orbit), or 415 (for an elliptical orbit), 1HE LAW OF UNIVERSAL GRAVITATION. 355 If M 1 be the mass of another planet, we have in like manner for its orbit Jc (8 +11') T l = 4^ r'\ Therefore T 2 (8+3T) : T l (8 + M 1 ) = r 3 : r*, the correct relation between the periods and mean distances. It is known that different planets have different masses. Hence, the fact that Kepler's Third Law is approximately true shows that the masses of the planets are small compared with that of the Sun. 422. Motion relative to Centre of Mass. The mutual attractions of the Sun and planet have no influence on the position of the centre of mass (commonly called the " centre of gravity ") of the solar system ; hence, in consider- ing the relative motions, that point may be treated as fixed. It is known from general dynamical principles that when a system of bodies are under the influence of their mutual reactions or attractions alone, the centre of mass of the whole system is not accelerated. But it may be interesting to prove independently that when two bodies, such as the Sun and a planet, attract one another, they both revolve about their centre of mass. Let us suppose (to take a simple case) the relative orbit circular and of radius (P=) r, the angular velocity being n. Then, if G be the point about which the planet (P) and Sun (S) revolve, individually, we have n*xGP = acccl. of planet = kS/r* ; w 2 x GS = acccl. of Sun = Hence MxGP= Sx GS ; which relation shows that G is the common centre of mass, as was to be proved. In the case of three or more bodies, such as the Sun and pLinets, the centre of mass is still the common centre about which they revolve, but the corresponding investigation is more difficult, owing to the effect of the mutual attractions of the planets in producing perturbations. It may be mentioned that the mass of the Sun is so large, compared with those of the planets, that, although the further planets arc so very distant, the centre of mass of the whole solar system always lies very near the Sun. 356 ASTRONOMY. 423. Verification of the Theory of Gravitation for the Earth and Moon. Before considering the motions of the planets about the Sun, Newton investigated the orbital motion of the Moon about the Earth, with the view of dis- covering whether the Earth's attractive force, which retains the Moon in its orbit, is the same force as that which pro- duces the phenomenon of gravity at the Earth's surface. If we assume that the force varies inversely at the square of the distance, and that the Moon's distance is 60 times the Earth's radius, the acceleration of gravity at the Moon should be (-frV) 2 g, where g is the acceleration of gravity on the Earth's surface. But the acceleration g = 32-2 feet per sec. per sec. ; .. accel. at Moon's distance = 32-2/3600 feet per sec. per sec. = 32-2 feet per min. per min. From the length of the lunar month and the Moon's dis- tance in miles, Newton calculated what must be the normal acceleration -of the Moon in its orbit. At the time of his first investigation (1666) the Earth's radius and the Moon's dis- tance were but imperfectly known, and the Moon's normal acceleration, as thus computed, came out only about 27 feet per minute per minute. Some fifteen years later, the Earth's radius, and consequently the Moon's distance, had been measured with much greater accuracy, and, working with the new values, Newton found that the Moon's normal accelera- tion to the Earth agreed with that given by his theory. Taking the lunar sidereal month as 27 -3 days, the Earth's radius as 3960 miles, and the radius of the Moon's orbit as 60 times the Earth's radius, the angular velocity (n) of the Moon, in radians, per minute is 27T 27-3x24x60' The Moon's distance in feet (d) = 3960 X 60 x 5280. Hence the Moon's normal acceleration (tfd) in feet per minute per minute = 3 150 XJL XJ280 x 47T 2 _ 2xll0 2 X7r 2 (27'3) 2 x 24 2 x 60 2 = (27-3) 2 x 10 = 32 approximately, thus agreeing with that given by the law of gravitation. THE LAW OF UNIVERSAL GRAVITATION. 357 EXAMPLE. Having given that a body at the Earth's equator loses 1/289 of its weight in consequence of centrifugal force, (i.) To calculate the period in which a projectile could revolve in a circular orbit, close to, but without touching the Earth, and (ii.) To deduce the Moon's distance. (i.) The centrifugal force on the body would have to be equal to its weight, and would therefore have to be 289 times as great as that at the Earth's equator. Hence the projectile would have to move -v/289, or 17 times as fast as a point on the Earth's equator, and would therefore have to perform 17 revolutions per day.* Therefore the period of revolution = j\- of a day. (ii.) Assuming the law of gravitation, the periodic times and dis- tances of the projectile and Moon must be connected by Kepler's Third Law. Hence, taking the Moon's sidereal period as 27| days, we have, if a = Earth's rad., d = Moon's dist., .-. d z = a? x (17 x 27^) 2 = a 3 {^^} 2 = a 3 . 215915'i ; .-. d = a x 3/215915-1 = 59'99; .'. distance of Moon = 60 x Earth's radius almost exactly. 424. Effect of Moon's Attraction. Moon's Mass. If we take account of the Moon's attraction on the Earth we must introduce a correction analogous to that made in Kepler's Third Law (421). If J/, m are the masses of the Earth and Moon, the whole relative acceleration is k(lf+m)l$, instead of kM/d*. But, if g n is the acceleration of gravity on the Earth's surface, ^ = and, if I 7 is the length of the sidereal month, then, by 421, jJf-fW/nj w-jr * 1H "^ :=r ^^' This formula might be used (and has been used by Airy) to find m/M, the ratio of the Moon's to the Earth's mass, in terms of the observed values of a, d, g^ T. It is not, how- ever, a very accurate method, owing to the smallness of M/Jf. * Relative to the Earth it would perform 16 or 18 revolutions per day, according to whether it was revolving in the same or the opposite direction to the Earth. 358 ASTRONOMY. 425. To find the ratio of the Sun's Mass to that of the Earth. Let /S, M, m be the masses of the Sun, Earth, and Moon, d, r the distances of the Moon and Sun from the Earth, T, Y the lengths of the sidereal lunar month and year respectively. Then, if k be the gravitation constant, the Earth's attraction on the Moon is = klfm/d', and its intensity is kM. The Sun's attraction on the Earth is = kSM/r 1 , and its intensity is kS. Therefore, by 415, Corollary, kM . T 2 = 47rd 3 , kS . F 2 = 47rV ; whence the ratio of the Sun's to the Earth's mass may be found. If we take account of the attraction of the smaller body on the larger, the whole acceleration of the Earth, relative to the Sun, is k (S + M+m)/^ (since the Sun is attracted by the Moon as well as the Earth), and that of the Moon, relative to the Earth, is k (M+m)/eP. Hence the corrected or more exact formula is Since the Moon's mass is about -fa of that of the Earth, the first or approximate formula can only be used if the cal- culations arc not carried beyond two significant figures. In this manner it is found that the Sun's mass is about 331,100 times that of the Earth. EXAMPLES. 1. To compare, roughly, the masses of the Earth and Sun, taking the Sun's distance to be 390 times the Moon's, and the number of sidereal months in the year to be 13. We have 8:M = ~:l^ t . mass of Sun 390? _ 2 _ mass of Earth 18* " l51 ' 000 - To the degree of accuracy possible by this method, the Sun's mass is therefore 350,000 times that of the Earth. THE LAW OP UNIVERSAL GRAVITATION. 359 2. To find the ratio of the masses, taking the Moon's mass as Jj of the Earth's, and the number of sidereal months in the year as 13^. 390 3 _ 390 3 x 3 2 _ 5338710 --,.._ oooooy ; .'. 8 = 333668 (M + m) = 333668 (1 + ff T ) 31 = 337,787 M. 426. To determine the mass of a planet which has one or more satellites. The method of the last paragraph is obviously applicable to the case of any planet which has a satellite. We require to know the mean distance and the periodic time of the satellite. The former may be easily found by observing the maximum angular distance of the satellite from its primary, the distance of the planet from the Earth at the time of observation having been previously computed. The periodic time of the satellite may also be easily observed. Let M' , ml be the masses of the planet and satellite, d' their distance apart, r' their distance from the Sun, T' the period of revolution of the satellite, Y' the planet's period of revolution round the Sun. Using unaccented letters to re- present the corresponding quantities for the Earth and Moon we have, roughly, tE __ M'T" 2 __ SY'' 2 = SY 2 _ 31T* k " d'* ' r* r & d* ' or, more accurately, (S+M+m'^Y* k ' ~~d'~ _ (S+M+m) Y' whence the mass of the planet, or, more correctly, the sum of the masses of the planet and satellite, may be determined in terms of the mass of the Sun, or the sum of the masses of the Earth and Moon. We do not require to know the periodic time and mean distance of the planet from the Sun, since the above expressions enable us to express the required mass, M' + m' 9 in terms of the year and mean distance of the Earth, or in terms of the lunar month and the mean distance of the Moon. 360 EXAMPLE. To find the mass of Uranus in terms of that of the Sun, having given that its satellite Titania revolves in a period of 8 days 17 hours at a distance from the planet = '003 times the distance of tho Earth from the Sun. Let M be the mass of Uranus, then we have d 3 . r 3 and, by Kepler's Third Law, r*/Y' 2 is the same for Uranus as for the Earth. Hence M'8= C' 003 ) 3 . I 3 (8d. 17h.) a ' (365d.6h.) 2? *L = I 3 Y /365d.6h. 8 UOOO/ \8d.l7h. = ?7_ x /8766\ 2 ~ 10 9 \ 209 / Thus, the mass of Uranus is to that of the Sun in the ratio of 1 to 21,Q53. *427. The Masses of Mercury and Venus (which have no satellites) could theoretically be found by determining their mean distances from the Sun by direct observation, and comparing them with those calculated from their periodic times by Kepler's Third Law. For, if 3T is the mass of such a planet, we have (S+_J)_r 2 = (8 + M + m) T- ~ This enables ns to find the stun of the masses of the Sun and planet, and, the Sun's mass being known, the planet's mass could be found. This method is, however, worthless, because the masses of Mercury nd Venus are only about TOO^WO ^ *injW of that of the Sun > and in order to calculate one significant figure of the fraction M'/S it would be necessary to know all the data correct to about seven significant figures, a degree of ^Qewracy unattainable in practice. For this reason it is necesgar,y ?to Calculate the masses of these planets by means of the pefltwrbatiqns t they produce on one another and on the Earthy these p*L r All r ) ia ^9 I n j' vW^ ^ e discussed in the next chapter. T1TR LAW OP TTNTVERSAL GRAVITATION". 36 1 428. Centre of Mass of the Solar System. When the masses of the various planets have been found in terms of the Sun's mass, the position of the centre of mass of the system can be found for any given configuration, and can thus be shown to always lie very near the Sun. EXAMPLES. 1. To find the distance of the centre of mass of the Earth and Sun from the centre of the Sun. Here the mass of the Sun is 331,100 times the Earth's mass, and the distance between their centres is about 92,000,000 miles. Hence, the centre of mass of the two is at a distance from the Sun's centre of about 92,000,000 =278mil 331,100 + 1 2. To find the centre of mass of Uranus and the Sun, and to show that it lies within the Sun. The distance of Uranus from the Sun is 19'2 times the Earth's distance, and its mass is 1/21053 of the Sun's. Hence the C,M- ia at a distance from the Sun's centre of 92,000,000 x 19-2 21053 + 1 The Sun's semi-diameter is 433,200 miles ; hence the centre of mass of the Sun and Uranus is at a distance from the Sun's centre of rather less than the radius. 3. In the case of Jupiter, the mean distance is 5'2 times that of the Earth, and the mass is 1/1050 of that of the Sun j hence the C.M. is at a distance 5-2 x 92,000,000 1050 + 1 This is just greater than the Sun's radius (433,200), showing that the centre of mass lies just without the Sun's surface. 362 ASTBONOMY. SECTION III. The Earth's Mass and Density. 429. The so-called " Weight of the Earth " really means the Earth's mass, and the operation called " weighing the Earth," in some of the older text-books, means finding the mass of the Earth. In the last section we explained how to compare the masses of the Sun and certain planets with that of the Earth, and in the next chapter we shall give methods applicable to a planet having no satellites. But before the masses can be expressed in pounds or tons it is necessary to determine the Earth's mass in these units. The methods of doing this all depend on comparing the Earth's attraction with that of a body of known mass and distance ; and the only difficulty lies in determining the \atter attraction, since the force between two bodies of ordinary dimensions is always extremely small. The following methods have been used. The first two are by far the best. (1) By the "Cavendish Experiment," or the balance. (2) By observations of the influence of tides in estuaries. (3) By the "Mountain" method. (4) By pendulum experiments in mines. 430. The " Cavendish Experiment " owes its name to its having been first used to determine the Earth's mass by Cavendish, about the year 1798. The essential principle of the method consists in comparing the attractions of two heavy balls of known size and weight with the Earth's attraction. Since the attraction of a sphere at any point is proportional directly to the mass of the sphere and inversely to the square of the distance from its centre, it is evident that by comparing the attractions of different spheres such as the Earth and the experimental ball of metal we can find the ratio of their masses. The comparison is effected by means of a torsion "balance. Two equal small balls A, B are fixed to the ends of a light beam suspended from its middle point by means of a slender vertical thread or "torsion fibre" (in his recent experiments, Professor C. V. Boys has used a fine fibre of spun quartz), so as to be capable of twisting about in a horizontal plane (the plane of the paper in Fig. 143). Two heavy metal balls C, D, are brought near the small balls A, (as shown in the THE LAW OP UNIVERSAL GRAVITATION. 363 figure), and their attraction causes the beam to turn about 0, say from its original position of rest XX' to the position AB. As the beam turns the fibre twists ; this twisting is resisted by the elasticity of the fibre, which produces a couple, propor- tional to the angle of twist XOA, tending to untwist it again. Let us call this couple /x L XOA, where / is a constant depending on the fibre, called its " torsional rigidity" The beam AB assumes a position of equilibrium when the moments about of the attractions of the large spheres (7, D on the balls A, B, just balance the "untwisting couple" /x Z XOA. The angle XOA being measured, and the dimensions of the apparatus being supposed known, the attractions of the spheres can now be determined in terms of the torsional rigidity. FIG. 143. The value of /is found in terms of absolute units of couple by observing the time of a small oscillation of the beam when the balls A, B have been removed. [The beam will then swing backwards and forwards like the balance wheel of a chronometer (204). The greater the torsional rigidity, the more frequently will it reverse the motion of the beam, and the more frequent will be the oscillations.*] Hence finally the attractions between the known masses C, D and A, B are found in terms of known units of force, and by comparing these attractions with that of gravity the Earth's mass is found. * The student who has read Kigid Dynamics should work out the formula. ASTRON. 2 B 364 ASTRONOMY. In practice, instead of measuring the angle XOA, the masses C, I) are subsequently placed on the reverse side of the beam, say with their centres at c, fl, and they now deflect the beam in the reverse direction, say to ab. The angle measured is the whole angle aOA, and this angle is ticice the angle XOA, if the positions CD and cd are symmetrically arranged with respect to the line XOX'. In the earlier experiments the beam AB was six feet long, and the masses C, D were balls of lead a foot in diameter. Quite recently, however, Professor C V. Boys, by the use of a quartz fibre for the suspending thread, has performed the experiment on a much smaller scale, the whole apparatus being only a few inches in size and being highly sensitive. He uses cylinders instead of spheres for the attracting bodies, and this introduces extra complications in the calculations. Although the above description shows the general principle of the method, many further precautions are required to ensure accuracy. A full description of these would be out of place here. 431. The common balance has also been used to deter- mine the Earth's mass. In this case the differences of weight of a body are observed when a large attracting mass is placed successively above and below the scale-pan containing it. EXAMPLE. To find the Earth's mass in tons, having given that the attraction of a leaden ball, weighing 3 cwt., on a body placed at a distance of 6 inches from its centre is '0000000432 of the weight of the body. Let M be the mass of the Earth in tons. The mass of the ball in tons is = /. The Earth's radius in feet = 3960 x 5280 = 20,900,000 roughly ; and the distance of the body from the ball in feet = ^. Hence, since the attractions of the Earth and ball are proportional directly to the masses and inversely to the squares of the distances from their centres, -0000000432 ; 1 = _ ' ' 2 (20,900,000) (20,900,000^ x J_ = 3 x 209- x 10*0 x -0000000432 5 x 432 x 10 - 10 5 432 2160 = 6067 x 10 1S . Hence the mass of the Earth is (roughly) 6067 million billion tons. THE LAW OF UNIVERSAL GRAVITATION. 365 432. To determine the Earth's Mass by observa- tions of the Attraction of Tides in Estuaries. A method which admits of very great accuracy is that in which the mass of the Earth is found hy comparing it with that of the water brought by the tide into an estuary. Con- sider an observatory situated (like Edinburgh Observatory) due south of an arm of the sea, whose general direction is east and west. The direction of its zenith, as shown either by a plummet or by the normal to the surface of a "bowl of mercury, is not the same at high tide as at low, because the additional mass of water at high tide produces an attraction which deflects the plummet and the nadir point northward, and hence displaces the zenith towards the south. Hence the latitude of the observatory is less at high tide than at low ; and the difference is a measurable quantity. The great advantage of this method is that the mass which deflects the plumb-line can be measured with great certainty ; for the density of the sea-water is exactly known (and, unlike that of the rocks in the next methods, is uniform throughout) and the shape and height of the layer of water brought in are known from the ordnance maps, and the tide measurements at the port. *433. In the Pendulum Method the values of g r the acceleration of gravity, are compared by comparing the oscil- lations of two pendulums at the top and bottom of a deep mine. The difference of the two values is due to the attrac- tion of that portion of the Earth which is above the bottom of the mine ; this exerts a downward pull on the upper pen- dulum, and an upward pull on the lower one. If the Earth were homogeneous throughout, the values of g at the top and bottom would be directly proportional to the corresponding distances from the Earth's centre. If this is not observed to be the case, the discrepancy enables us to find the ratio of the density of the Earth to that of the rocks in the neighbourhood of the mine. If the latter density is known, the Earth's density can be found, and knowing its volume, its mass can be computed. But this method is very liable to considerable errors, arising from imperfect knowledge of the density of the rocks overlying the mine. 366 ASTEONOMT. *434. In the Mountain Method the Earth's attraction is com* pared with that of a mountain projecting above its surface. Suppose a mountain range, such as Schiehallien in Scotland, running due E. and W. ; then at a place at its foot on the S. side the attraction of the mountain will pull the plummet of a plumb line towards the N., and at a place on the N. side the mountain will pull the plummet to the S. Hence the Z.D. of a star, as observed by means of zenith sectors, will be different at the two sides, and from this difference the ratio of the Earth's to the mountain's attraction may be found. In order to deduce the Earth's density it is then necessary to determine accurately the dimensions and density of the mountain. This renders the method very inexact, for it is impossible to find with certainty the density of the rocks throughout every part of the mountain. 435. Determination of Densities. Gravity on the Surface of the Sun and Planets. When the mass and volume of a celestial body have been computed, its average density can, of course, be readily found. By dividing the mass in pounds by the volume in cubic feet, we find the average mass per cubic foot, and since we know that the mass of a cubic foot of water is about 62 J Ibs., it is easy to compare the average density with that of water. The deter- mination of densities is particularly interesting, on account of the evidence it furnishes regarding the physical condition of the members of the solar system. The Earth's density is about 5*58. Prom knowing the ratios of the mass and diameter of the Sun or a planet to that of the Earth, we can compare the intensity of its attraction at a point on its surface with the intensity of gravity on the Earth. It may be noticed that attraction of a sphere at its surface is pro- portional to the product of the density and the radius. For the attraction is proportional to mass -*- (radius) 2 , and the mass is proportional to the density x (radius) 3 ; .*. the attraction at the surface is proportional to the density x radius. EXAMPLES. 1. To find the Earth's average density and mass, having given that the attraction of a ball of lead a foot in diameter, on a particle placed close to its surface, is less than the Earth's attraction in the proportion of 1 : 20,500,000, and that the density of lead is 11'4 times that of water. THE LAW OF UNIVERSAL GRAVITATION. 367 Let D be the average density of the Earth. Then, since the radii of the Earth and the leaden ball are | and 20,900,000 feet respectively, and the attractions at their surfaces are proportional to thei* densities multiplied by their radii, /. 1 : 20,500,000 = ll'4xi : -Dx 20,900,000; /. average density of Earth. D = 5'7x|^ = 5-6. Hence the average mass of a cubic foot of the material forming the Earth is 5 - 6 x 62'5 pounds. But the Earth is a sphere of volume |TT (20,900,000) 3 cubic feet. Hence the mass of the Earth, with these data, = -I* x 209 3 x 10 15 x 5-6 x 62-5 pounds = 1H38 x 10 22 pounds = 597 x 10 19 tons. 2. To calculate the mean density of the Sun from the following data: Mass of O = 330,000 . (mass of $) ; Density of = 5 '58 ; Q's parallax = 8'8"j Q's angular semi-diameter = 16'. The radii of the Sun and Earth being in the ratio of the Sun's angular semi-diameter to its parallax ( 258), we have Q's radius 16' 960 inQ-1 - = - = - = iuy j. j (p's radius 8'S 8'8 /. volume of Sun = (109'1) 3 . (vol. of Earth) = 1,298,000 . (vol. of Earth) roughly. But mass of Sun = 330,000. (mass of Earth) ; . density of Sun ^ 330 = _1_ nearl density of Earth 1298 3'9 /. density of Sun = 1'4. 3. To find the number of poundals in the weight of a pound at the surface of Jupiter, taking the planet's radius as 43,200 miles and density 1^ times that of water. Taking the Earth's radius as 3960 miles and density as 5'58, we have (gravity at surface of Jupiter) : (gravity on Earth) = 1-33 x 43,200 : 5-58 x 3960. But at the Earth's surface the weight of a pound = 32-2 poundals ; therefore on the surface of Jupiter the weight of a pound = 83-7 poundals. 368 ASTRONOMY. EXAMPLES. XIII. 1. Taking Neptune's period as SO years, and the Earth's velocity as 91 miles per second, find the orbital velocity of Neptune. 2. If we suppose the Moon to be 61 times as far from the Earth's centre as we are, find how far the Earth's attraction can pull the Moon from rest in a minute. 3. If the Earth possessed a satellite revolving at a distance of only 6,000 miles from the Earth's surface, what would be approximately its periodic time, assuming the Earth to be a sphere of 4,000 miles radius ? 4. Assuming the distance between the Earth's centre and the Moon's to be 240,000 miles, and the period of the Moon's revolution 28 days, find how long the month would be if the distance were only 80,000 miles. 5. Calculate the mass of the Sun in terms of that of Mars, given that the Earth's mean distance and period are 92 x 10 6 miles and 365i days, and the mean distance and period of the outer satellite of Mars are 14,650 miles and Id. 6h. 18m. 6. Show that the periodic time of an asteroid is 3| years, having given that its mean distance is 2'305 times that of the Earth. 7. Show that we could find the Sun's mass in terms of the Earth's, from exact observation of the periods and mean distances of the Earth and an asteroid, by the error produced in Kepler's Third Law in consequence of the Earth's mass. 8. Show that an increase of 10 per cent, in the Earth's distance from the Sun would increase the length of the year by 56' 14 days. 9. The masses of the Earth and Jupiter are approximately "iuroVoT) an d ToW respectively of the Sun's mass, and their distances from the Sun are as 1 : 5. Show that Kepler's Laws would give the periodic time of Jupiter too great by more than 2 days. 10. Prove that the mass of the Sun is 2 x 10 27 tons, given that the mean acceleration of gravity on the Earth's surface is 9'81 metres per second per second, the mean density of the Earth is 5*67, the Sun's mean distance T5 x 10 s kilometres, a quadrant of the Earth's circumference 10,000 kilometres, and taking a metre cube of water to be a ton. THE LAW OF UNIVERSAL GRAVITATION. 369 11. Having given that the constant of aberration for the Earth is 20'49", and that the distance of Jupiter from the Sun is 5'2 times the distance of the Earth from the Sun, calculate the constant of aberration for Jupiter. 12. If the mass of Jupiter is T oVo f t ne m ass of the Sun, show that the change in the constant of aberration caused by taking into account the mass of Jupiter is 004" nearly (see Question 11). 13. Find the centre of mass of Jupiter and the Sun. Hence find the centre of mass of Jupiter, the Sun, and Earth, (1) when Jupiter is in conjunction, (2) when in opposition. (Sun's mass = 1,048 times Jupiter's = 332,000 times Earth's. Jupiter's mean distance = 480,000,000 miles ; Earth's = 93,000,000 miles.) 14. If the intensity of gravity at the Earth's surface be 32 - 185 feet per second per second, what will be its value when we ascend in a balloon to a height of 10,000 feet ? (Take Earth's radius = 4,000 miles and neglect centrifugal force.) Would the intensity be the same on the top of a mountain 10,000 feet high ? If not, why not ? 15. Show how by comparing the number of oscillations of a pendulum at the top and bottom of a mountain of known density, the Earth's mass could be found. 16. How would the tides in the Thames affect the determination of meridian altitudes at Greenwich observatory theoretically ? . 17. If the mean diameter of Jupiter be 86,000 miles, and his mass 315 times that of the Earth, find the average density of Jupiter. 18. If the Sun's diameter be 109 times that of the Earth, his mass 330,000 times greater, and if an article weighing one pound on the Earth were removed to the Sun's surface, find in poundals what its weight would be there. 19. Taking the Moon's mass as ^ that of the Earth, show that the attraction which the Moon exerts upon bodies at its surface is only about l-5th that of gravity at the Earth's surface. 20. If the Earth were suddenly arrested in its course at an eclipse of the Sun, what kind of orbit would the Moon begin ^0 describe ? 370 ASTRONOMY. EXAMINATION PAPER. XIII. 1. State reasons for supposing that the Earth moves round the Sun, and not the Sun round the Earth. 2. State Kepler's Laws, and give Newton's deductions therefrom. .3. If the Sun attracts the Earth, why does not the Earth fall into the Sun ? 4. Show that the angular velocitiesof two planets are as the cubes of their linear velocities. 5. State Newton's Law of Gravitation, and prove Kepler's Third Law from it for the case of circular orbits, taking the planets small. 6. Explain clearly (and illustrate by figures or otherwise) what is meant by a force varying inversely as the square of the distance. 7. Are Kepler's Laws perfectly correct ? Give the reason for your answer. What is the correct form of the Third Law if the masses of the planets are supposed appreciable as compared with the mass of the Sun ? 8. How can the mass of Jupiter be found ? 9. Show that if a body describes equal areas in equal times about a point, it must be acted on by a force to that point. 10. Find the law of force to the focus under which a body will describe an ellipse ; and if C be the acceleration produced by the force at unit distance, T the periodic time, and 2a the major axis of the ellipse, find the relation between 0, a, T. CHAPTEE XIV, FURTHER APPLICATIONS OF THE LAW OF GRAVITATION. SECTION 1. The Hoori 's Mass Concavity of Lunar Orbit. 436. The Earth's Displacement due to the Moon. In Section II. of the last chapter we saw that when two bodies are under their mutual attraction they revolve about their common centre of mass. Thus, instead of the Moon revolving about the Earth in a period of 27-^ days, both bodies revolve about their centre of mass in this period, although from the Moon's smaller size its motion is more marked. In this case both the Earth and Moon are under the attraction of a third body the Sun which causes them together to describe the annual orbit. But the Sun's dis- tance is so great compared with the distance apart of the Earth and Moon, that its attraction is very nearly the same, both in intensity and direction, on both bodies. To a first approximation, therefore, the resultant attraction of the Sun is the same as if the masses of both the Earth and Moon were collected at their common centre of mass. Hence it is strictly the centre of mass of the Earth and Moon, and not the centre of the Earth, which revolves in an ellipse about the Sun with uniform areal velocity, in accordance with the laws stated in 155. And, owing to the revolution of the Moon, the Earth's centre revolves round this point once in a sidereal month, threading its way alternately in and out of the ellipse described, and being alternately before and behind its mean position. 372 ASTRONOMY. This displacement of the Earth has been used for finding the Moon's mass in terms of tho Earth's, by determining the common centre of mass of the Earth and Moon, as follows. FIG. H4. Let EV J/,, 6 r 1 (Fig. 144) be the positions of the centres of the Earth and Moon, and their centre of mass, at the Moon's last quarter, JE* 2 , M v 2 and E# M# G 3 their positions at new Moon and at first quarter respectively, S the Sun's centre. Then, at last quarter, E^ is behind # and the Sun's longi- tude, as seen from E^ is less than it would be as seen from G l by the angle E^SG^. At first quarter, JE 9 is in front of G 3 , and therefore the Sun's longitude is greater at E% than at G s by the angle G^SE y If, then, the observed coordinates of the Sun be compared with those calculated on the supposition that the Earth moves uniformly (i.e., with uniform areal velocity), its longitude will be found to be decreased at last quarter and increased at first quarter. From observing these displacements the Moon's mass may be found. For, knowing the angle of displacement E l SG l and the Sun's distance, the length E l G l may be found. Also the Moon's distance E^ is known. And, since G l is the FURTHER APPLICATIONS OF THE LAW OF GRAVITATION. 373 centre of mass of the Earth and Moon, mass of Moon : mass of Earth = E l G l : G^J/j; whence the mass of the Moon can be found. The Sun's displacement at the quarters could he found by meridian observations of the Sun's R.A. with a transit circle. The displacement of the Earth will also give rise to an apparent displacement, having a period of about one month, in the position of any near planet ; this could be detected by observations on Mars, when in opposition, similar to those used in finding solar parallax ( 339). From this and other methods it is found that the mass of the Moon is about 1/81 of that of the Earth. The Moon's density, as thus deduced, is about 3 '44, or of that of the Earth. EXAMPLE. To compare the masses of the Moon and Earth, having given that the Sun's displacement in longitude at the Moon's quadratures is equal in f of the Sun's parallax. Since L E i SG i = f the angle subtended by Earth's radius at S, therefore E& = | (Earth's radius). But E^Mi = 60 (Earth's radius) ; .'. EjJf, = 80. #!(?,; .'. GjMi = 79..E? 1 G 1 , and mass of Moon : mass of Earth =E^G^ : O i M l = 1 : 79; /. the Moon's mass = 1/79 of the Earth's mass. 437. Application to Determination of Solar Paral- lax. If the Moon's mass be found by any other method, the above phenomena give us a means of finding the Sim's parallax and distance. For we then know E^ G l : G^^ and therefore E& and the angle E.SG, is found by observation. But the exact ratio of IS 1 SG 1 to the parallax is known, for it is equal to that of ^ G l to tbe Earth's radius ; hence the Sun's parallax and distance can be found. Since the Moon's mass can be found with extreme accuracy by many different methods, this method is quite as accurate as many that have been used for finding the solar parallax. *438 Concavity of the Moon's Path about the Sun. - The Moon, by its monthly orbital motion about the Earth, threads its way alter- nately inside and outside of the ellipse which the centre of mass of the Earth and Moon describes in its annual orbit about the Sun, ,374 ASTRONOMY. Hence the path described by the Moon in the course of the year is a wavy curve, forming a series of about thirteen undulations about the ellipse. It might be thought that these undulations turned alternately their concave and convex side towards the Sun, but the Moon's path is really always concave ; that is, it always bends towards the Sun, as shown in Fig. 145, which shows how the path passes to the inside of the ellipse without becoming convex. To show this it is necessary to prove that the Moon is always being accelerated towards the Sun. Let n, n' be the angular velo- cities of the Moon about the Earth and the Earth about, the Sun respectively. Then, when the Moon is new, as at M 2 (Fig. 145), its acceleration towards G 2 , relative to G 2 , is n 2 . MG 2 . But (? 2 has a normal acceleration n'' 2 G^S 'towards 8. Hetice the resultant accelera- tion of the Moon Jf 2 towards 8 is n'-G. 2 S- FIG. 145. Now, there are about 13 sidereal months in the year ; therefore TO = 13X- Also E^S is nearly 400 times E 2 M 2 , and therefore G 2 S is slightly over 400 times GM- 2 . Therefore roughly n'"G z S : n*M 2 Gz = 400 : 182 ; .'. ri-G^S > n 2 (? 2 3f 2 . Thus, the resultant acceleration of M is directed towards, not away from 8, even at Jf 2 , where the acceleration, relative to (? 2 . is directly opposed to that of G 2 . Therefore the Moon's path is constantly being bent (or deflected from the tangent at M 2 ) in the direction of the Sun, and is concave towards the Sun. *439. Alternate Concavity and Convexity of the Path, of a Point on the Earth.. In consequence of the Earth's diurnal rota- tion, combined with its annual motion, a point on the Earth's equator describes a wavy curve forming 365 undulations about the path described by the Earth's centre. In this case, however, it may be easily shown in the same way that the acceleration of the point towards the Earth's centre is greater than the acceleration of the Earth's centre towards the Sun. The path is, therefore, not always concave to the Sun, being bent away from the Sun in the neighbourhood of the points where the two component accelera- tions act in opposite directions. FURTHER APPLICATIONS OF THE LAW OF GRAVITATION. 375 SECTION II. The Tides. In the last section we investigated the displacements due to the -Moon's attraction on the Earth as a whole. We shall now consider the effects arising from the fact that the Moon's attractive force is not quite the same either in magnitude or direction at different parts of the Earth, and shall show how the small differences in the attraction give rise to the tides. 440. The Moon's or Sun's Disturbing Force. Let C, Jf be the centres of the Earth and Moon ; AC A' the Earth's diameter through M ; B, B' points on the Earth such that M C = MB = MB'. Let 3/, m denote the masses of the Earth and Moon, a the Earth's radius, d the Moon's distance. The resultant attraction of the Moon on the Earth as a whole is &Mm/ CM*, and the Earth is therefore moving with acceleration km/ CM* towards the common centre of mass of the Earth and Moon, as shown in 422, 424. FIG. 146. (i.) N"ow at the sublunar point A the Moon's attraction on unit mass is km/AM 2 and is greater than that at C (since AM < CM). Hence the Moon tends to accelerate A more than C and thus to draw a body at A away from the Earth, with relative acceleration F, where l CA d*(d-aj 2 d* (1-a/dy Since a/d is a small fraction, we have, to a first approximation, 376 ASTRONOMY. (ii.) At A the Moon's attraction per unit mass is km/A'M' 2 , and is less than that at (7, since AM > CM. Hence the Moon tends to accelerate C more than A, and thus to draw the Earth away from A with relative acceleration F', where = Jan To a first approximation, therefore, (l+a/d Thus a body either at A or A tends to separate from the Earth, as if acted on by a force away from C, of magnitude approximately = 2kina/d* per unit mass. FIG. 147. (iii.) Consider now the effect of the Moon's attraction on a body at B. This produces a force per unit mass of which may be resolved into components and Since we have taken 7>3f = CM, the first component is equal to km/ CM 3 ; that is, to the force at C. This component therefore tends to make a body at B move with the rest of the Earth, and produces no relative acceleration. Therefore the Moon tends to draw a body at B towards the Earth with relative acceleration /, represented by the second component ; thus FUKTHER APPLICATIONS OF THE LAW OP GRAVITATION. 377 The point B is approximately the end of the diameter BCB perpendicular to AC (since BM, CM, B'M are nearly parallel in the neighbourhood of the Earth). Hence the relative acceleration at B is approximately per- pendicular to CM, and its magnitude f=km=km . d it' Similarly at B' the Moon tends to draw a body towards C, with relative acceleration /= kma/d*. At either of these points, B, B', therefore, a body tends to approach the Earth, as if acted on by a force towards the Earth's centre, of magnitude kma/d 3 per unit mass. Generally, the Moon's attraction at any point tends to accelerate a body, relatively to the Earth, as if it were acted on by a force depend- ing on the difference in magnitude and direction between the Moon's attractions at that point and at the Earth's centre. This apparent force is called the Moon's disturbing force or tide-generating force. AVc sec that the dis- turbingforce produces a pull along^L4' and a squeeze along////. A similar consequence arises from the attraction of the Sun. The Sun's actual attraction on the Earth as a whole keeps the Earth in its annual orbit, but the variations in the attraction at different points give rise to an apparent distribution of force on the Earth which is the Sun's disturbing force or tide-generating force. 441. To find approximately the Moon's or Sun's Disturbing Force at any point. Let be any point of the Earth. Draw ON perpen- dicular on CM. " Then the difference of the Moon's attractions at and N tends to accelerate towards JV, with a relative acceleration 1cm . NO/d* [by 440 (iii.)]- Also, the difference of the attractions at N, C tends to accelerate TV away from C with a relative acceleration 2km. CN/d 3 [by 440 (i.)]. The whole acceleration of 0, relative to C, is compounded of these two relative accelerations. Therefore, if X. Foe the components of the disturbing force at in the directions CN, NO, ON v ^ .- , Y= K 378 ASTRONOMY. 442. Hence the following geometrical construction ; On CN produced take a point IT such that Tlicn the line OH represents the disturbing force at in direction, and its magnitude is v 7 on F = 1cm. -. d* The Sun's tide-raising force may be found exactly in the same way. The force is everywhere directed towards a point on the diameter of the Earth through the Sun, found by a similar construction to the above. And if r, S denote the Sun's distance and mass, the force is proportional to S/r s instead of mj&. In all these investigations we see that the tide-raising force due to an attracting body is proportional directly to its mass and inversely to the cube (not the square) of its distance. From this it is easy to compare the tide-raising forces due to different bodies acting at different distances. EXAMPLES. 1. To compare the tide-raising forces due to the Sun and Moon. The masses of the Sun and Moon are respectively 331,000 and gL times the Earth's mass. Also, the Sun's distance is about 390 times the Moon's. .*. Sun's tide-raising force : Moon's tide-raising force = 33 : 73 nearly = 3:7 nearly. Thus the Sun's tide-raising force is about'three-sevenths of that of the Moon. 2. To find what would be the change in the Moon's tide-raising force if the Moon's distance were doubled and its mass were in- creased sixfold. If /,/'betheold and new tide-raising forces at corresponding points, /'/= f = 4 f J ' J 2 3 ' l a> 4 Therefore the tide-raising force would have three-quarters of its present value. 3. To compare the Moon's tide-raising forces at perigee and apogee. The greatest and least distances of the Moon being in the ratio of FURTHER APPLICATIONS OF THE LAW OF GRAVITATION. 379 l + TT to ! Te> or 19 to 17 ( 270), the tide-raising power at perigee is greater than at apogee in the ratio of 19 3 : 17 3 or 6859 : 4913, or roughly 7 : 5. 4. To compare the maximum and minimum values of the Sun's tide-raising force. The eccentricity of the Earth's orbit being ^ these are in the ratio of (1 + sV) 3 : (1 g^) 3 , or approximately 1 + ^ : I--&, or 21 : 19. As before, the force is greatest at perigee and least at apogee. Moon 443. The Equilibrium Theory of the Tides. Let us imagine the Earth to be a solid sphere covered with an ocean of uniform depth. If we plot out the disturbing forces at different points of the Earth by the construction of 442, we shall find the distribution represented in Eig. 148, the lines representing the forces both in magnitude and direction. Here the disturbing force tends to raise the ocean at the sub-lunar point A and at the opposite point A, and to de- press it at the points B, B'. At intermediate points it tends to draw the water away from B and B\ towards A and A'. Hence the surface of the ocean will assume an oval form, as represented by the thick line in Eig. 148, and there will be high water at the sublunar point A and the opposite point A, low water along the circle of the Earth BB', distant 90 from the sublunar point. Thus we have the same tides occurring simultaneously at opposite sides of the Earth. It may be shown that the oval curve aba'b' is an ellipse whose major axis is aa'. The surface of the ocean, therefore, assumes the form of the figure produced by revolving this ellipse about its major axis. This figure is called a prolate spheroid, and is thus distinguished from an oblate spheroid, which is formed by revolution about the minor axis. 2c ASTRON. 380 ASTEONOMT, But though this is the form which the ocean would assume if it were at rest, a stricter mathematical investigation shows that the Earth's rotation would cause the surface of the sea to assume a very different form. In fact, if the Earth were covered over with a sufficiently shallow ocean of uniform depth, and rotating, we should really have low tide very near the sublunar point A and its anti- podal point A', and high tide at the two points on the Earth's equator distant 90 from the Moon (Fig. 149). If the Moon were to move in the equator, the equilibrium theory would always give low ^ water at the poles. This phenomenon is uninnue'nced by the Earth's rotation, and since the Moon is never more than about 28 from the equator, we see that the Moon's tide-raising force has the general effect of drawing some of the ocean from the poles towards the equator. *444. A few other consequences of the equilibrium theory may also be enumerated. (1) According to it the height of the tides, or the difference of height between high and low water at any place, is directly proportional to the tide-generating force, and consequently, with the results of Example 1 of 442, the heights of the solar and lunar tides are in the proportion of 3 to 7. (2) Since the distortion of the mass of liquid is resisted by gravity, the height of the tide depends on the ratio of the tide-producing force to gravity, and therefore is inversely proportional to the intensity of gravity, and therefore to the density of the Earth ; if the density were halved, the height of the tides would be doubled. (3) If the diameter of the Earth were doubled, its density remaining the same, the inten- sity of gravity and the tide-producing force would both be doubled, since both are proportional to the Earth's radius. This would cause the ocean to assume the same shape as before, only all its dimensions would be doubled. f Consequently the height of the tide would also be doubled, and it thus appears that the height of the tide is pro- portional to the Earth's radius. We thus have the means of comparing the tides which would be produced on different celestial bodies, for the above properties show that the height of tide is proportional to ma/Dd?, where a and D are the radius and density of the body under consideration, TO, d the mass and distance of the disturbing body. *445. Canal Theory of the Tides. As an illustration, let us consider what would happen in a circular canal, not extremely deep, supposed to extend round the equator of a re- f Of course this is not a very strict proof. FTTRTHEB APPLICATIONS OF THE LAW OF GBAYITATION. 381 volving globe. Then, in Fig. 149, it is clear that the direction of the disturbing force would, if it acted alone, cause the water in the quadrants AB and AB' to flow towards A ; and, in the quadrants A'B and AB', towards A. Hence this force acts in the same direction as the Earth's rotation in the quadrants B'A and BA, and in the opposite direction in AB and AB'. Hence, as the water is carried from A to B, it is constantly being retarded, from B to A it is accelerated, from A to B' it is retarded, and from B' to A it is again accelerated, the average velocity being, of course, that of the Earth's rotation. Hence the velocity is least at B and ', and greatest at A and A'. Moon Now, it is easy to see that when water moves steadily in a uniform canal it must be shallow where it is swift and deep where it is slow. For, if we consider any portion of the canal, say AB, the quantity that flows in at one end A is equal to the quantity that flows out at the other end B. But it is evident that if the depth of the canal were doubled at any point without altering the velocity of the liquid, twice as much liquid would flow through the canal ; consequently, in order that the amount which flows through might be the same as before, we should have to halve the velocity of the liquid. This shows that where the canal is deepest the water must be travelling most slowly. Con- versely, where the velocity is least the depth must be greatest, and where the velocity is greatest the depth must be least. Hence the depth is least at A and A', and greatest at B and J?, just the opposite to what we should have expected from the equilibrium theory. 882 ASTRONOMY. In a canal constructed round any parallel of latitude the same would be the case ; and hence, if we could imagine a uniform ocean replaced by a series of such parallel canals, low tide would occur at every place when the Moon was in the meridian. This theory (due to Newton), though sounder than Laplace's equilibrium theory, is still not quite mathematically correct. The true explanation of the tides, even in an ocean of uniform depth, is far more complicated, and quite beyond the scope of this book. 446. Lunar Day and Lunar Time. According to either hypothesis, the recurrence of high and low water depends on the Moon's motion relative to the meridian ; hence, in investigating this, it is convenient to introduce another kind of time, depending on the Moon's diurnal motion. The lunar day is the interval between two consecutive upper transits of the Moon across the meridian. In a lunation, or 29 mean solar days, the Moon performs one direct revolution relative to the Sun, and therefore per- forms one retrograde revolution less relative to the meridian. Thus 29J mean days = 28J lunar days ; whence the mean length of a lunar day = O + BT) mean 8 lar days = 24h. 50m. 32s. nearly. The lunar time is measured by the Moon's hour an^le, converted into hours, minutes, and seconds, at the rate of ^15 to the hour. ttjItTHEB, APPLICATIONS OP THE LAW OF GRAVITATION. 383 *447. Semi-diurnal, Diurnal and Fortnightly Tides. It has been found convenient to regard the tides produced by the Moon's disturbing force as divided into three parts, whose periods .are half a day, a day and a fortnight, the " day " being the lunar day of the last paragraph. If we adopt the equilibrium theory as a working hypothesis, the lunar tide must be highest when the Moon is nearest to the zenith or nadir. Hence high tide takes place at the Moon's upper and lower transits, when its zenith distance and nadir distance are least respectively. But, for a place in N. lat. (Fig. 150) when the Moon's declination is K, it describes a small circle Q,'R\ and its least zenith distance ZQ 'is less than its least nadir distance NR ; hence the two tides are unequal in height. This phenomenon can be represented by supposing a diurnal tide, high only once in a lunar clay, combined with a semi-diurnal tide, high twice in this period. Again, the Moon's meridian Z.D. and N.D. go through a complete cycle of changes, owing to the change of the Moon's declination, whose period is a month. But after half a month, the Moon's declination will have the same value but opposite sign, and hence the diurnal circles Q[R' , Q,"R" ', equidistant from the equator Q,R, are described at intervals of a fortnight. But NJR"= ZQ', ZQ"=: NR' ; hence the two tides have the same heights. This can be represented by supposing a fort- nightly tide of the proper height combined with the diurnal and semi-diurnal ones. In just the same way the smaller tides caused by the Sun may be artificially represented by combining a diurnal and semi-diurnal tide (the solar day being used) and a six-monthly tide. 448. Spring and Neap Tides. Priming andLagging We have hitherto considered chiefly the tides due to the action of the Moon. In reality, however, the tides are due to the combined action of the Sun and Moon, the tide-raising forces due to these bodies being in the proportion of about 3 to 7 (Ex. 1, 442). We shall make the assumption that the height of the tide at any place is the algebraic sum of the heights of the tides which would be produced at that place by the Sun and Moon separately. 384 ASTRONOMY. / At new or full Moon the Sun is nearly in the line AA\ (, and the tide-raising powers of the Sun and Moon both act in )the same direction, and tend to draw the water from B^B' to i A^^AL^ hence the whole tide is that due to the sum of the ) separate disturbing forces of the Sun and Moon. The tides / are then most marked, the height of high water and.depth_pf I low water being at their maximum. Such tides are called Spring Tides. We notice that the height of the spring tide = 1 +f or \- ^ * na * f tnc lunar tide alone. Moon At the Moon's first or last quarter the Sun is in a line BB' perpendicular to A A'. Hence the Sun tends to draw the water away from A, A' to B, B >', while the Moon tends to draw the water in the opposite direction. The Moon's action being the greater, preponderates, but the Sun's action diminishes the tides as much as possible. The variations are therefore at their minimum, although high water still occurs at the same time as it would if the Sun were absent. These tides are called Neap Tides. The^ndght of the neap, _ tide is the difference of the heights of the lunar and solar tides, and is therefore f of that of the lunar tide. Hence spring tides and neap tides are in the ratio of (roughly) 10 to 4. For any intermediate phase of the Moon, the Sun's action is somewhat different. Between new Moon and first quarter, the Sun is over a point S l behind A. Here the Moon tends to draw the water towards A, A', and the Sun tends to draw the water towards S l and the antipodal point $ s . Therefore the com- bined action tends to draw the water towards two points Q, Q' APPLICATIONS OF THE LAW OF GRAVITATION". 385 between A and S l and between A and S. 6 respectively, whose longitudes are rather less than those of A and A respectively. The resulting position of high water is therefore displaced to the west, and the high water occurs earlier than it would if due to the Moon's influence alone. The tides are then said to prime. Between first quarter and full Moon the Sun is over a point $ 2 between ' and A, and the combined action of the Sun and Moon tends to draw the water towards two points jR, R', whose longitudes are slightly greater than those of A, A. The resulting high tides are therefore displaced east- wards, and occur later than they would if the Sun were absent. The tides are then said to lag. Between full Moon and last quarter the Sun is over some point $ 8 between and A', but the antipodal point S l is between A and B' ; hence the tide primes. Between last quarter and new Moon, when the Sun is at a point S between B and A, it is evident in like manner that the tide lags. Hence Spring Tides occur at the syzygies (conjunction and opposition). Neap Tides occur at the quadratures. From syzygy to quadrature, the tide primes. From quadrature to syzygy, the tide lags. The heights of the spring and neap tides vary with the varying distances of the Sun and Moon from the Earth. Spring tides are the highest possible when both the Sun and Moon are in perigee, while neap tides are the most marked when the Moon is in apogee but the Sun is in perigee (because the Sun then pulls against the Moon with the greatest power, as far as the Sun's action is con- cerned). Both the spring and neap tides, and also the priming and lagging, are on the whole most marked when the Sun is near per ^ee, i.e.. about January. It may be here stated, without proof, that, taking the Sun s am Moon's tide-raising forces to be in the proportion ot o 7, ti maximum interval of priming or lagging is found \ 61 minutes. 386 ASTRONOMY. 449. Establishment of the Port. Both the equilibrium and canal theories completely fail to represent the actual tides on the sea, owing to the irregular distribution of land and water on the Earth, combined with the varying depth of the ocean. These circumstances render the prediction of tides by calculation one of the most complicated problems of prac- tical astronomy, and the computations have to be based largely on previous observations. In consequence of the barriers offered to the passage of tidal waves by large continents, lunar high tide does not occur either when the Me on crosses the meridian, as it would on the equilibrium theory, or when the Moon's hour angle is 90, as it would on the canal theory. But this continental retardation causes the high tide to occur later than it would on the equilibrium theory, by an interval which is constant for any given place. This interval, reckoned inlunar hours, is called the Establishment of the Port for the place considered. Thus the establish- ment of the port at London Bridge is Ih. 58m., so that lunar high water occurs Ih. 58m. after the Moon's transit, i.e., when the Moon's hour angle, reckoned in time, is Ih. 58m. The same causes affect the solar tide as the lunar, hence the Sun's hour angle (or the local apparent time) at the solar high tide is also equal to the establishment of the port. The actual high tide, being due to the Sun and Moon con- jointly, is earlier or later than the lunar tide by the amount of priming or lagging. By adding a correction' for this to the establishment of the port, the lunar time of high water may be found for any phase of the Moon ; and we notice in particular that at the Moon's four quarters (syzygies and quadratures), the lunar time of high water is equal to the establishment of the port. And, knowing the lunar time of high water, the corresponding mean time can be found, for (mean solar time) (lunar time) = (mean 0's hour angle) ( ([ 's hour angle) = ( d 's R.A.) -(mean Q's R.A.) [since R.A. and hour angle are measured in opposite directions]. Now the Moon's R.A. is given in the Nautical Almanack for every hour of every day in the year. Also the mean Sun's R.A. at noon is the sidereal time of mean noon, and is given FURTHER APPLICATIONS OF THE LAW OF GRAVITATION. 387 in the Nautical Almanack. Hence the mean Sun's R.A. [which = (sidereal time) - (mean time)] is easily found for any intermediate time. Hence the mean time of high water can be readily found. The establishments of different ports, and the times of high water at London Bridge, are given in the Nautical Almanack. *450. If only a very rough calculation is required, we may proceed as in 35, 40. We assume the Moon's R.A. to increase uniformly ; we shall then have ( <[ 's R.A.) - ( 0's R.A.) = ( O elongation) ; .'. (solar time) = (lunar time) + ( MN^ since the sides of the triangle J/iJVyV 2 are each less than 90. But when the Moon comes to My let another impulse act towards H. This will deflect the direction of motion from M^N^ to M^, and the Moon will now begin to describe the great circle NJil^N^ whose nodes Ny N^ are still further behind their initial positions. The inclination of the orbit to the ecliptic will, however, be increased this time. It is easy to see that the same general effect takes place when the Moon is acted on by a continuous force, always FURTHER AT-PLICATIOKS OP THE LAW OP GRAVITATION. 409 tending towards the ecliptic, instead of a series of impulses. Such a force continuously deflects the Moon's direction of motion, and draws the Moon down so that it returns to the ecliptic more quickly than it would otherwise. Hence the Moon, after leaving one node, arrives at the next hefore is has quite described 180, and the result is an apparent retrograde (never direct] motion of the nodes, combined with periodic, hut small, fluctuations in the inclination of the orbit. *474. The retrograde motion of the Moon's nodes is, in some respects, analogous to the precession of the equinoxes, and, although the analogy is somewhat imperfect, the former phenomenon gives an illustration of the way in which the latter is produced. If the Earth had a string of satellites, like Saturn's rings, chosely packed together in a circle in the plane of the equator, the Sun's disturbing force, ever ac derating them towards the ecliptic, would, as in the case of the Moon, cause a retrograde motion of the points of intersection of all of their paths with the ecliptic, and this would give the appearance of a kind of retrograde precession of the plane of the rings. If the particles, instead of being separate, were united into a solid ring, the general phenomena would be the same. And it is not unnatural to expect that what occurs in a simple ring should also occur, to a greater or less degree, in the case of other bodies that are somewhat flattened out perpendicularly to their axis of rotation, such as the Earth, thus accounting for the precession of the equinoxes. (Of course this is only an illustration, not a rigorous proof; in fact, if the Earth were qiiite spherical it would behave very differently.) FIG. 163. *475 Perturbations due to Average Value of Eadial Disturbing Force -Let d be the Moon's distance. Then, when the Moon is m conjunction or opposition, the Sun's disturbing force acts away from the Earth, and is of magnitude 2k8d/* (Fig 163) When the Moon is in quadrature the disturbing force acts towards the B. but is only half as great. Hence, on the average, the disturbing force tends to pull the Moon away from the Earth. In consequence, the Moon's average centrifugal force must be rather less than it would be at the same distance from the Eartl ,f there were no disturbing force, and the effect of this is^ to make the month a little longer than it would be otherwise for the same c tance of the Moon. 410 ASTRONOMY. Moreover, the disturbing force increases as the Moon's distance increases, but the Earth's attraction diminishes, being proportional to the inverse square of the distance ; this has the effect of making the whole average acceleration along the radius vector decrease more rapidly as the distance increases than it would according to the law of inverse squares. The result of this cause is the progressive motion of the apse line. It is difficult to explain this in a simple manner, but the following arguments may give some idea of how the effect takes place. At apogee the Moon's average acceleration is less, and at perigee it is greater than if it followed the law of inverse squares and had the same mean value. Hence, when the Moon's distance is greatest, as at apogee, the Earth does not pull the Moon back so quickly, and it takes longer to come back to its least distance, so that it does not reach perigee till it has revolved through a little more than 180. Similarly, at perigee the greater average acceleration to the Earth does not allow the Moon to fly out again quite so quickly, and it does not reach apogee till it has described rather more than 180. Hence, in each case, the line of apsides moves forward on the whole. *476. Variation, Evection, Annual Equation, Parallactic Inequality. When the Moon is nearer than the Earth to the Sun (M } , Fig. 162), the Moon is more attracted than the Earth, and therefore the disturbing force is towards the Sun ( 472). Its effect is, therefore, to accelerate the Moon from last quarter to con- junction, and to retard it from conjunction to first quarter. When the Moon is more distant than the Earth from the Sun (A/ 3 , Fig. 163), it is less attracted than the Earth, and therefore the disturbing force is away from the Sun. Thus the Moon is accelerated from first quarter to full Moon, and retarded from full Moon to last quarter.f Hence we see that the Moon's motion in each case must be swiftest at conjunction and opposition, and slowest at the quadratures. This phenomenon is known as the Variation. The force towards the Earth is greatest at the quadratures, and least at the conjunction and opposition, since at the former the Sun pulls the Moon towards, and at the latter away from the Earth. Either cause tends to make the orbit more curved at the quadratures and less curved at the syzygies. For, if v is the velocity, R the radius of curvature, then v*]R = normal acceleration. Hence R is greatest, and the orbit therefore least curved, when v is greatest, and the normal acceleration is least. The effect of this cause would be to distort the orbit, if it were a circle, into a slightly oval curve, which would be most flattened, and therefore narrowest (compare f These retardations and accelerations are closely analogous to those of the water in an equatorial canal ( 445). -PUHTHEB APPLICATIONS OF THE LAW OF GRAVITATION. 411 arguments of 114, 115), at the points towards and opposite the Sun ; most rounded, and therefore broadest, at the points distant 90 from the Sun. Of course the Moon's undisturbed orbit is not really circular, but elliptic, and far more elliptic than the oval into which a circular orbit would be thus distorted. But a distortion still takes place, and gives rise to periodical changes in the eccentricity, depending on the position of the apse line, and known as evection. The Sun's disturbing force is greatest when the Sun is nearest, and least when the Sun is furthest. These fluctuations, between perihelion and aphelion, give rise to another perturbation, called the whose most noticeable effect consists in the con- sequent variations in the length of the month ( 475). If, instead of resorting to a first approximation, we employ more accurate expressions for the Sun's disturbing force on the Moon, it is evident that this force is greater when the Moon is near con- junction than at the corresponding position near opposition ; just as the disturbing force which produces the tides is really greater under the Moon than at the opposite point. Hence the Moon is mere disturbed from last quarter through new Moon to first quarter than from first quarter through full Moon to last quarter. Hence the time of first quarter is slightly accelerated, and that of last quarter retarded. This is called the Moon's Parallactic Inequality. Its amount is proportional to TcScP/r*, instead of fcSeF/r^like the other perturbations). For many reasons this perturbation is of considerable use in determinations of the Sun's mass and distance. 477. Planetary Perturbations. The Sun's mass is so great, compared with the masses of the planets, that the orbital motion of one planet about the Sun is but slightly affected by the attraction of any other planet. The mutual attractions of the planets, and their actions on the Sun, give rise to small planetary perturbations, which cause each planet to diverge slowly from its elliptical orbit, besides accelerating or retarding its motion. Since the orbital motions of the planets are all usually referred to the Sun as their common centre or " origin," nnd not to the centre of mass of the solar system, the perturba- tions of one planet, due to a second, depend, not on the actual acceleration produced by the latter, but on the differences of the accelerations which it produces on the former planet and on the Sun. As in the case of the Moon, the force which produces this difference of accelerations is called the disturbing force. ASTRON. 2 E 412 ASTRONOMt. *478. Geometrical Construction for the Disturbing Force. The approximate expressions, investigated in 472, for _the Sun s dis- turbing force on the Moon, are inapplicable to the disturbing force of one planet on another, because the distance of the distorting body from the Sun is no longer very large, compared with that of the disturbed body. We must, therefore, adopt the following con- struction (Fig. 164) : Let P, Q be two planets, of masses Jf, If ; 8 the Sun. Then the planet P produces an acceleration fcM/PQ 2 on Q along QP, and an acceleration IcM/PS* on 8 along SP. To find the acceleration of Q, relative to 8, due to this cause, take a point T on PQ such that PT : P3 = PS 2 : PQ 2 . Then the accelerations of 8, Q, due to P, are fcjf . SP//SP 3 and TcM . TP/SP 3 respectively. Hence, by the triangle of accelerations, the acceleration of Q, relative to 8, is represented^ in magnitude and direction by fclf .TS/SP 3 . Therefore the disturbing force per unit mass on Q, due to P, is parallel to T/8, and of magni- tude TeM. TS/SP*. FIG. 164. Similarly, if we take a point 2* on QP such that QT : Q8 = QS 2 : QP 2 , the disturbing force per unit mass on P, due to Q, is parallel to T'8, and is of magnitude JcM' . T'S/SQ 3 . The disturbing force on Q, due to P, and that on P, due to Q, are not equal and opposite, because they depend on the planets' attrac- tions on 8, as well as on their mutual attractions. When PQ = P/8, the points Q, T evidently coincide, and the dis- turbing force on Q is along the radius vector QS. When PQ < P8, PT>PQ, so that the disturbing force on Q tends to pull Q about 8 (as in Fig. 164) towards P, and when PQ>PS, the disturbing force tends to push Q about 8 away from P. JTTETHER APPLICATIONS OF THE LAW OP GBAVITATION. 413 f 7 f ' * h i? distlu * bin S f on P is along PS. 1*2 PUU P about 5 toward Q. an * when push P about fl owa^/ from Q. *479. Periodic Perturbations on an Interior Planet.-Let us con- sider, m the first place, the perturbations produced by one planet E on another planet F, whose orbit is nearer the Sun- as for example, the perturbations produced by the Earth on Venus by Jupiter or Mars on the Earth, or by Neptune on Uranus. Let A, B be the positions of the planet, relative to E, when in heliocentri^ conjunction and opposition respectively; U, IT points on the relative orbit such that EU = EU' = E8. (These points are near but not quite coincident with the positions of greatest elonga- tion. ) Then, if we only consider the component relative acceleration of F perpendicular to the radius vector Fflf, this vanishes when the planet is at U or 17, as shown in the last paragraph. FIG. 165. The tangential acceleration also vanishes at A and B. Over the arc U'AU the relative acceleration is towards E, therefore the planet's orbital velocity is accelerated from U' to A j similarly it is retarded from A to U. Again, at a point F 2 on the arc UBU', the relative acceleration is away from the Earth, and this accelerates the planet's orbital velocity between U and B, and retards it between B and U'. It follows that V is moving most swiftly at A and B, and most slowly at U and U'. Hence, if we neglect the eccentricity of the orbit, we see that the planet, after passing A, will shoot ahead of the position it would occupy if moving uniformly; thus the disturbing force displaces the planet forwards during its path from A to near U. Somewhere near U, when the planet is moving with its least velocity, it begins to lag behind the position it would occupy if moving uniformly; thus from near Uto B the disturbing force dis- places the planet backwards. Similarly, it may be seen that from B to near U' the planet is displaced forwards, and from near U ' to A it is displaced backwards. 414 ASTRONOMY. The principal effect of the component of the disturbing force along the radius vector, is to cause rotation of the planet's apsides, as in the case of the Moon. The direction of their rotation depends on the direction of the force, and is not always direct. The eccen- tricity of the orbit is also affected by this cause, as in the phenome- non of lunar evection, and the periodic time is slightly changed. Owing to the inclination of the planes of the orbits of E, V, the attraction of E, in general, gives rise to a small component perpen- dicular to the plane of F's orbit, which is always directed towards the plane of E's orbit. This component produces rotation of the line of nodes, or line of intersection of the planes of the two orbits. This rotation is always in the retrograde direction, and is to be explained in exactly the same way as the rotation of the Moon's nodes. It is thus a remarkable fact that since all the bodies in the solar system (except the satellites of Uranus and Neptune) rotate in the direct direction, all the planes of rotation and revolution, and all their lines of intersection (i.e., the lines of nodes, and the lines of equinoxes) in the whole solar system, with the above exceptions, have a retrograde motion. A FIG. 166. *480. Periodic Perturbations of an Exterior Planet. The accele- rations and retardations produced by a planet E on one J, whose orbit is more remote from the Sun, during the course of a synodic period, may be investigated in a similar manner to the corre- sponding perturbations of an interior planet, assuming the orbits to be nearly circular. If SJ is less than 2SE there are two points if, N on the relative orbit at which EM = EN = E8. At these points the disturbing torce is purely radial, and it appears, as before, that the planet J is lerated from heliocentric conjunction A to M, and from helio- Ce ^ r i C T PPOSition B to N ' retarded from N to A, and from M to B. c, ' then E8 hence the attraction of E is greater on the hun than on J, and the disturbing force therefore always accele- rates the planet J towards B. Thus the planet's orbital velocity increases from A to B, and decreases from B to A, and it is greatest at a and least at A. Therefore from B to A the planet is displaced in FURTHER APPLICATIONS OF THE LAW OF GRAVITATION. 415 advance of its mean position, and from A to B falls behind its mean position. The effects of the radial and orthogonal components of the dis- turbing force in altering the period and causing rotation of the apse line, and regression of the nodes, can be investigated in the same way for a superior as for an inferior planet. *481. Inequalities of Long Period. If the orbits of the planets were circular (except for the effects of perturbations), and in the same plane, their mutual perturbations would be strictly periodic, and would recur once in every synodic period. Owing, however, to the inclinations and eccentricities of the orbits, this is not the case. The mutual attractions of the planets produce small changes in the eccentricities and inclinations, and even in their periodic times, which depend on the positions of conjunction and opposition relative to the lines of nodes and apses. Neglecting the motion of these latter lines, the perturbations would only be strictly periodic if the periodic times of two planets were commensurable ; the period of recurrence being the least common multiple of the periods of the two planets. But when the periodic times of two planets are nearly but not quite in the proportion of two small whole numbers, inequali- ties of long period are produced, whose effects may, in the course of time, become considerable. Thus, for example, the periodic times of Jupiter and Saturn are very nearly but not quite in the proportion of 2 to 5. If the propor- tionality were exact, then 5 revolutions of Jupiter would take the same time as 2 revolutions of Saturn 5 and, since Jupiter would thus gain three revolutions on Saturn, the interval would contain 3 synodic periods. Thus, after 3 synodic periods had elapsed from conjunction, another conjunction would occur at exactly the same place in the two orbits, and the perturbations would be strictly periodical. But, in reality, the proportionality of periods is not exact; the positions of every third conjunction are very slowly revolving in the direct direction. They perform a complete revolution in 2,640 years. But there are three points on the orbits at^which con- junctions occur, and these are distant very nearly 120 from one another. It follows that when the positions of conjunction have revolved through 120, they will again occur at the same points on the orbits, and the perturbations will again be of the same kind as initially. The time required for this is one-third the above period, or 880 years, and consequently Jupiter and Saturn are subject to lono-- period inequalities which recur only once in 880 years. Again, the periodic times of Venus and the Earth are nearly in the proportion of 8 to 13 ; consequently 5 conjunctions of Venus occur in almost exactly 8 years, thus giving rise to perturbations havinj a period of 8 years. But the proportion is not exact, and, consequently, there are other mutual perturbations having a very long period. 416 ASTBOtfOMY. One of the most important secular perturbations is the alternate increase and decrease in the eccentricity of the Earth's orbit. This, at the present time, is becoming gradually more and more circular, but in about 24,000 years the eccentricity will be a minimum, and will then once more begin to increase. The effects of this cause on the climate of the Earth's two hemispheres have already been considered ( 463). 482. Gravitational Methods of Finding the Sun's Distance. The Earth's perturbations on Mars and Venus furnish a good method of finding the Sun's distance. For the magnitude of these perturbations depends on the ratio of the Earth's mass, or rather the sum of the masses of the Earth and Moon (since both are instrumental in producing the perturbations), to the Sun's mass. Hence, if S, M, m denote the masses of the Sun, Earth, and Moon, it is possible, from observations of these perturbations, to find the ratio of (M+m) : S. But, if r, d be the distances of the Sun and Moon from the Earth, T and Fthe length of the sidereal lunar month and year, we have, by Kepler's corrected Third Law, (M+m) T* : (S+M+m) Y z = d 3 : r 8 ; whence the ratio of r to d is known. If, now, the Moon's distance d be determined by observation in any of the ways described in Chapter VIII., or by the gravitational method of 423, the Sun's distance r may be immediately found. This method was used by Leverrier in 1872. From obser- vations of certain perturbations of Venus he found the values 8-853" and 8-859" for the Sun's parallax, while the rotation of the apse line of Mars gave the value 8'866". The perturbations of Encke's comet were used in a similar way by Von Asten, in 1876, to find the Sun's parallax, the value thus obtained being rather greater, viz., 9-009". The lunar perturbations also furnish data for determining the Sun's distance, the principal of these being the parallactic inequality of the Moon ( 476). Several computations of the Sun's parallax have thus been made, the results being 8 -6" by Laplace in 1804, 8-95" by Leverrier in 1858, 8-838" by Newcomb in 1867. See also 437 for the determination of the parallax from the apparent monthly displacement of the Sun. FUETHEB APPLICATIONS OF THE LAW OF GRAVITATION. 417 483. Determination of Masses. The mass of any planet which is not furnished with a satellite can be deter mined in terms of the Sun's mass by means of the perturba- tions it produces on the orbits of other planets. The amount of these perturbations is always proportional to the disturbing force, and this again is proportional to the mass of the disturbing planet. In this manner the mass of Yenus has been found to be about 1/400,000 of the Sun's mass, and that of Mercury about 1/5,000,000. 484. The Discovery of Neptune. The narrative of the discovery of Neptune is one of the most striking and remark- able in the annals of theoretical astronomy, and forms a fitting conclusion to this chapter. In 1795, or about 14 years after its discovery, the planet Uranus was observed to deviate slightly from its predicted position, the_ observed longitude becoming slightly greater than that given by theory. The_ discrepancy increased till 1822, when Uranus appeared to undergo a retardation, and to again approach its predicted position. About 1830 the observed and computed longitudes of the planet were equal, but the retardation still continued, and by 1845 Uranus had fallen behind its computed position by nearly 2'. As early as 1821, Alexis Bouvard pointed out that these discrepancies indicated the existence of a planet exterior to Uranus, but the matter remained in abeyance until 1846, when the late Mr. (afterwards Prof.) Adams, in Cambridge, and M. Leverrier, in Paris, independently and almost simul- taneously, undertook the problem of determining the position, orfcit, and mass oi_an unknown planet which would give rise to the observed perturbations. Adams was undoubtedly the first by a few months in performing the computations, but the actual search for the planet at the observatory of Cam- bridge was delayed from pressure of other work. Meanwhile Leverrier sent the results of his calculations to Dr. Galle, of Berlin, who, within a few hours of receiving them, turned his telescope towards the place predicted for the planet, and found it within about 52' of that place. Subsequent exami- nation of star charts showed that the planet had been pre- viously observed on several occasions, but had always been mistaken for a fixed star. 418 ASTRONOMY. It will be seen from 479 that the acceleration of Uranus up to 1822, and its subsequent retardation, are ^at once accounted for by supposing an exterior planet to be in helio- centric conjunction with the Sun about the year 1822. But Adams and Leverricr sought for far more accurate details concerning the planet. At the same time the data afforded by the observed perturbations of Uranus were insufficient to determine all the unknown elements of the new planet's orbit, and therefore the problem admitted of any number of possible solutions. In other words, any number of different planets could have produced the observed perturbations. To render the problem less indeterminate, however, both astronomers assumed that the disturbing body moved nearly in the plane of the ecliptic and in a nearly circular orbit, that its distance and period were connected by Kepler's Third Law, and that its distance from the Sun followed Bode's Law. The latter assumption led to considerable errors, including an erroneous estimation of the planet's period by Kepler's Third Law. For when Neptune was observed, its distance was found to be only 30 '04 times the Earth's distance, instead of 38-8 times, as it would have been according to Bode's Law. Nevertheless, the actual planet was subsequently found to fully account for all the observed perturbations of Uranus. The discovery of Neptune affords most powerful evidence of the truth of the Law of Gravitation, and so indeed does the theory of perturbations generally. The fact that the planetary motions are observed to agree closely with theory, that computations of astronomical constants (such as the Sun's and Moon's distances), based upon gravitational methods, agree so closely with those obtained by other methods, when possible errors of observation are taken into account, affords an indisputable proof that the resultant acceleration of any body in the solar system can always be resolved into com- ponents directed to the various other bodies, each__cpmponent boingju:oportional directly to the mass and inversely to the sijuare of the distance of the corresponding body. Such a truth cannot be regarded as a fortuitous coincidence ; it can only be explained by supposing every body in the universe to attract every other body in accordance with Newton's Law of Universal Gravitation. FUKTHEK APPLICATIONS OF THE LAW Qlf OKAV1TATION. EXAMPLES. XIY. 1. If the Sun's parallax be 8'80", and the Sun's displacement at first quarter of Moon 6'52", calculate the mass of the Moon, the Earth's radius being taken as 3,963 miles. 2. Supposing the Moon's distance to be 60 of the Earth's radii, and the Sun's distance to be 400 times that of the Moon, while his mass is 25,600,000 times the Moon's mass, compare the effects of the Sun and Moon in creating a tide at the equator, in the event of a total eclipse occurring at the equinox. 3. If the Earth and Moon were only half their present distance from the Sun, what difference would this make to the tides ? Cal- culate roughly what the proportion between the Sun's tide-raisng power and the Moon's would then be, assuming the Moon's distance from -the Earth remained the same as at present. 4. Taking the Moon's mass as -^ of the Earth's, and its distance as 60 times the Earth's radius, show that the Moon's tide-raising force increases the intensity of gravity by 1/17,280,000 when the Moon is on the horizon, and that it decreases the intensity of gravity by 1/8,640,000 when the Moon is in the zenith. 5. Compare the heights of the solar tides on the Earth and on Mercury, taking the density of Mercury to be twice that of the Earth, its diameter "38 of the Earth's diameter, and its solar distance 38 of the Earth's solar distance. 6. Explain how the pushing forward of the Moon by the tidal wave enlarges the Moon's orbit. 7. Show that, owing to precession, the right ascension of a star at a greater distance than 23| from the pole of the ecliptic will undergo all possible changes, but that a star at a less distance than 23^ will always have a right ascension greater than twelve hours. 8. Prove that for a short time precession does not alter the decli- nations of stars whose right ascensions are 6h., or 18h. 9. Exhibit in a diagram the position of the pole star (R.A. = Ih. 20m., decl. = 88 40 7 ) relative to the poles of the equator and ecliptic, and hence show that owing to precession its R.A. is increas- ing rapidly, but that its polar distance is decreasing. 10. Describe the disturbing effects of Neptune on Uranus for a short time before and after heliocentric conjunction, pointing out when Uranus is displaced in the direct, and when in the retrograde direction. 420 A.STRONOMr. EXAMINATION PAPER. XIV. 1. Show that the Moon's orbit is everywhere concave to the Sun. 2. Show that the tide-raising force of a heavenly body is nearly proportional to its (mass) -f- (distance) 3 . 3. How is it that we have tides on opposite sides of the Earth at once ? 4. Explain the production of the tides on the equilibrium theory. 5. Define the terms spring tide, neap tide, priming and lagging , establishment of the port, lunar time. 6. What is meant by the expression " Luni-solar Precession" ? Describe the action of the Sun and of the Moon in causing the Precession. 7. Give a general description of Precession. Does precession change the position of (a) the equator, (6) the ecliptic among the stars ? 8. Describe nutation. What is the cause of Lunar Nutation? What is meant by the equation of the equinoxes ? 9. Give a brief account of the discovery of Neptune. 10. Explain how the retrograde motion of the Moon's nodes is caused bv the Sun's attraction on the Earth and Moon. NOTE I. DIAGRAM FOE SOUTH LATITUDES. In order to familiarize the student with astronomical diagrams drawn under different conditions, we subjoina/ftgure showing the principal circles of the celestial splndre of an observer in South latitude 45 at about 19h. of JSereal time (QWRr = 270+15 = 19h.). The figure Jnows also the Sun's daily paths at t^e" solstices ; also the arcs T^R^= QM, and MX, which measure the E.A. and N. decl. of the star x. R" fi NADIR FIG. 169. N.POLE NOTE II. THE PHOTOCHEONOGRAPH. Quite recently photography has been applied to recording transits, as an alternative for the methods explained in Chap. II., 49, 50. The image of the observed star is 422 ASTRONOMY. jjf ejected on a sensitized plate placed in front of the transit circle, and, owing to the diurnal motion, it moves horizontally across the plate. The plate is made to oscillate slightly in a vertical direction, by means of clockwork, say once in a second, and this motion, combined with the horizontal motion of the image, causes it to describe a zigzag or wavy streak on the plate. The star's position at each second is indicated by the undulations, and the position of these is capable of being measured with great exactness. NOTE III. NOTE ON 104. It may be proved, by Spherical Trigonometry, that sin nP = sin xP sin nxP, or sin I = cos d sin nxP , cos 2 d cos 2 nxP = cos 2 d cos 2 d sin 2 nxP = cos 2 d sin 2 1 = cos (d + 1} cos (d I) ; acceleration t = - 15 U" 15 1 ; but they are of little astronomical importance, except as representing the paths described by non-periodic comets. 424 ASTRONOMY. 2. An ellipse has two foci (each focus having a corresponding directrix), and the sum of the distances of any point from the two foci is constant. Thus in Fig 169, 8, Hare the two foci, and the sum 8P + PH is the same for all positions of P on the curve. From this property an ellipse may easily be drawn. For, let two small pins be fixed at 8 and H, and let a loop of string SPH be passed over them and round a pencil-point P ; then, if the pencil be moved so as to keep the string tight, its point P will trace out an ellipse. For SP + PH +H8 = constant, and /. SP + PR = constant. 3. For all positions of P on the ellipse, SP is inversely propor- tional to 1 -t e cos ASP, so that SP(l + ecos^4SP) = I = constant, e being the eccentricity and 8A the line through 8 perpendicular to the directrix. >r 4. The line joining the two foci is perpendicular to the directrices. The portion of this line (AA'), bounded by the curve, is called the major axis or axis major. Its middle point C is called the centre, and the^ curve is symmetrical about this point. The line BCB', drawn through the centre perpendicular to AC A' and terminated by the curve, is called the minor axis or axis minor. The lengths of the major and minor axes are usually denoted by 2a and 26 respectively. 5. The extremities A, A' of the major axis are called the apses or apsides. Since, by (2), 8P + HP is constant, therefore, taking P at A or A', SP + HP = SA + HA = 8 A' + HA' = $(SA + HA + SA' + HA') evidently = AA' = 2a. Taking P at B, SB + HB = 2a ; .'. SB (evidently) = HB = a = CA. PBOPERTIES OF THE ELLIPSE. 425 6. The eccentricity e = 08/CA ; /. CS = e . CA, and 52 = 02 = SB 2 - CiS 2 (Euc. I. 47) = a*-a 2 e 2 = a 2 (1-e 2 ) j Hence also S4= (L4-CS = a(l-e) and 8 A' = CA' 7. The latus rectum is the chord LSL' drawn through the focus perpendicular to the major axis AA'. Its length is 21, where I = a (Ift-e 2 ). Also I is the constant of (3), for when P coincides with L, ASP = 90; .'. cos ASL = 0, and 8L = I. [Fig. 168.] 8. The tangent fPT and normal PGg, at P, bisect respectively the exterior and interior angles (SPI, 8PH) formed by the lines JSP, HP. 9. If the normal meets the major and minor axes in G, g, PO : Pg = OB 2 : CA 2 (= 6 2 : a 2 ). 10. If ST, drawn perpendicular on the tangent at P, meets HP produced in I, then evidently SP IP ; .'. HI = SP + HP = 2a [by (2)]. If HT' is the other focal perpendicular on the tangent, it is known that rectangle 8T . HT' = constant = 6 2 . 11. Relation between the focal radius SP and the focal perpen- dicular on the tangent ST. Let SP = r, 8T = p. Then cos TIP = cos T8P = pfr. By Trigonometry, = I8* + IH--2. 18. IH. cos SIHi 4aV = 4j> 2 + 4a 2 - 8pa x p/r ; 2 (1-e 2 ) 2a This may also be proved from the similarity of the triangles 8PT } HPT', which gives 8T : HT' = SP : HP ; /. ST 2 : ST.HT = )S(P : .HP and ST.HF = b 2 (10) j .\ p 2 : b 2 =r : 2a-r. 12. If a circular cone (i.e., either a right or oblique cone on a circular base) is cut in two by a plane not intersecting its base, the curve of section is an ellipse. More generally, the form of a circle represented in perspective, or the oval shadow cast by a spherical globe or a circular disc on any plane, are ellipses. A circle is a particular form of ellipse for the case where 6 = a and /. e = 0. 13. The area of the ellipse is -nab. 426 ASTRONOMY. TABLE OF ASTRONOMICAL CONSTANTS. (Approximate values, calculated, when variable, for the Spring Equinox, A.D. 1900.) THE CELESTIAL SPHERE. Latitude of London (Greenwich Observatory), 51 28' 31", Cambridge Observatory, 52 12' 51". Obliquity of Ecliptic, 23 27' 8", OPTICAL CONSTANTS. Coefficient of Astronomical Eefraction, 57". Horizontal Eefraction, 33'. Coefficient of Aberration, 20'493". Velocity of Light in miles per second, 186,330. metres 299,860,000. Equation of Light, 8m. 18s TIME CONSTANTS. Sidereal Day in mean solar units = 1 l/366days = 23h. 56m.4'ls. Mean Solar Day in sidereal units = 1 + 1/365| days = 24h. 3m. 56'5s. Year, Tropical, in mean time, 365d. 5h. 48m. 45'51s. Sidereal, 365d. 6h. 9m. 8'97s. Anomalistic, 365d. 6h. 13m. 48'09s. Civil, if the number of the year is not divisible by 4, or if it be divisible by 100, bnt not by 400, 365 days. In other cases, 366 Month, Sidereal, 27'32166d. = 27d. 7h. 43m. 11 '4s. Synodic, 29'53059d. = 29d. 12h. 44m. 30s. Metonic Cycle, 235 Synodic Months = 6939'69d = 19 tropical years (all but 2 hours). Period of Botatiou of Moon's Nodes (Sidereal), 6793'391d. = 18'60yr. (Synodic), 346'644d. = 346d. 14Jh. Apsides (Sidereal), 3232'575d. = 8'85yr, (Synodic), 411'74d. Saros 223 Synodic Months = 6585'29d. = 18*0906 yr, = 18 yr. 10 or 11 days. = 19 Synodic periods of Moon's Nodes (very nearly/ = 16 Apsides (nearly). Equation of Time, Maximum due to Eccentricity, 7m. Obliquity, 10m, TABLE OF ASTRONOMICAL CONSTANTS. 427 Equatorial Circumference, THE EARTH. Equatorial Eadius, 3963-296 miles. Polar 3949-791 Mean 3959"! 22,902 360 x 60 = 21,600 geographical miles. 4 x 1Q7 = 40,000,000 metres. Ellipticity or Compression, l-f-293. Eccentricity, '0826. Density (Water = 1), 5'58. Mass, 6067 x 10 18 tons. Mean Acceleration of Gravity in ft. per sec. per sec., 32-18. Eatio of C entrif ugal Force to Gravity at Equator, 1 -. 289. Eccentricity of its Orbit, l-s-60. Annual Progressive Motion of Apse Line, H'25". Eetrograde Motion of Equinoxes (Precession), 50-22". Period of Precession, 25,695 years. Nutation, 18'6 Greatest change in Obliquity due to Nutation, 9'23". Equation of Equinoxes, 15' 37". THE SUN. Mean Parallax, 8'80". Angular Semi-diameter, 16' 1". Distance in miles, 92,800,000. Diameter in miles, 866,400. in Earth's radii, 109. Density in terms of Earth's, ^. (taking water a& lj, 1 J 4. Mass in terms' of Earth's, 324,439. Period of Axial Eotation, 25d. 5h. 37m. THE MOON. Mean Parallax, 57' 2707". Angular Semi-diameter, 15' 34". Distance in miles, 238,840, in Earth's radii, 60'27. in terms of Sun's distance, 1/389. Diameter in miles, 2,162. in terms of Earth's, 3/11. Density in terms of Earth's, '61. (taking water as 1), 3'4. Mass, in terms of Earth's, 1/81. Eccentricity of Orbit, 1/18. Inclination of Orbit to Ecliptic, 5 8'. Ecliptic Limits, Lunar, 12 5' and 9 30'. Solar, 18 31' and 15 21'. Tide-raising force in terms of Sun's, 7/3. ASTRON. 2 F ANSWERS. NOTE. Where only rough values of the astronomical data are given in the questions, the answers can only be regarded as rough approximations, not as highly accurate results. It is impossible to calculate results correctly to a greater number of significant figures than are given in the data employed, and any extra figures so calculated will necessarily be incorrect. As the use of working examples is to learn astronomy rather than arithmetic, it is ad- visable to supply from memory the rough values of such astronomi- cal constants as are not given in the questions. These values will thus be remembered more easily than if the more accurate values were taken from the tables on pages 426, 427, though reference to the latter should be made until the student is familiar with them. I. EXAMPLES (p. 33). 1. Only their relative positions are stated; these do not completely fix them. 2. 6 P.M., 6 A.M.; on the meridian. 8. On September 19. 9. (i.) Early in July ; (ii.) middle of June the Sun passes it about June 26. 10. 304 = 20h. 16m.; at 8h. 13m. P.M. 11. Near the S. horizon about 10 P.M. early in October. 12. 38 27', 51 33', 28 5', or if Sun transits N. of zenith 8 27', 81 33', 58 5'. I. EXAMINATION PAPER (p. 34). 7. 30. 8. 61 58' 37", 15 4' 21". 9. 6h. 43m. 16s. (roughly). 10. The figure should make Capella slightly W. of N., altitude about 15; o Lyras a little S.E. of zenith, altitude about 75; a Scorpii slightly W. of S., altitude about 12; o Ursse Hajoris N.W., altitude about 60. ANSWERS. 429 II. EXAMPLES (p. 61). . Direct. 7. Interval = 12 sidereal hours. 9. 2 3 29' 58'5". 11. 12 39' 9". 12. I7h. 29m. 52'42s. II. EXAMINATION PAPER (p. 62). 6. Positive. 1O. lrn.2'52s., + 0718. III. EXAMPLES (p. 84). 2. 4,267ft. 3. aN., L-90W. and o S., L + 90W., if L = W. longitude given place. 5. 13m. 6. 39-8 miles. 7. 3960. 8. 6084ft. 10. 49' 6" per hour. MISCELLANEOUS QUESTIONS (p. 85). 2. N.P.D. = 85, hour angle = 30 W. 3. Because declination circle has not been defined. 5. 22h. 40m., 9h. 20m., 14h. Om., 19h. 36m.. 10. 52". ' V III. 1 EXAMINATION PAPER (p. 86). 1. 24,840 miles, 3,953 miles. 2. 3-285 ft., 6,084 ft., T69ft. per second. 3. 507 ft. 5. 3,437,700 fathoms, 6,366,200 metres (roughly), 1,851-851 metres. 9. See 97, cor. IV. EXAMPLES (p. 113). 5. 45. 7. Star, 6h. 15m. 26'35s. ; Sun, Oh. 13m. 51'90s. 10. 3481 : 3721, or 29 : 31 nearly. IV. EXAMINATION PAPER (p. 114). 3. See 130, 151. 3. Oh. 36m. 21'26s. (Note that the clock has a losing rate of 3m. 22'05s. on sidereal time ; it gives solar time approxi- mately.) V. EXAMPLES (p. 137). 1. Retrograde. 3. 3'9m. 6. 347 centuries exactly. 7. Star's hour angle = 4h. llm. 3s., N.P.D. = 53. 8. October 28, 15h. 39m. 27'32s. 1O. 12h. 27m. 13'26s. at Louisville = 18h. 9m. 13'26 at Greenwich. 430 ASTRONOMY. MISCELLANEOUS QUESTIONS (p. 138). 3. Eastward. 5. Use Figs. 47, 50. 6. See 439. 7. See 161. 8. llh. 59m. 15'9s. ; - 1m. 7'4s. 9. 366-25 : 365'25 or 1465 : 1461. V. EXAMINATION PAPER (p. 139). 4. - 10m. ; morning 20m. longer. 5. See 172. 8. (i.) 7h. 13m. 5s. ; (ii.) 7h. 12m. 48s. 9. June 26. 10. 1824, 1852, 1880, 1920. VI. EXAMPLES (p. 151). 3. 3,963 miles. 4. From 50 9' 47" to 49 59' 55" (refraction at altitude 5 - 9' 47" by tables). 5. 44 53' 28". 8. 84 33' ; 377 miles or 327 nautical miles. VI. EXAMINATION PAPER (p. 152). 4. 462". 7. 44 58' 54". 1O. Ih. 12m. VII. EXAMPLES (p. 188). 1. 37 49'. 2. 51 44' 26-09*. 4. 50 54' 58'6" or 60 43' 23'6" according aa star transits N. or S. of zenith. 5. 44 55', or, if corrected for refraction (cf. Ex. 2, p. 168), 44 53' 54". 6. 51 33', 38 27', 61 54'. 8. - 10m. } i.e., 10m. fast. 9. 12 30'. 1O. Ih.Om. 11. 2 32'. 12. 27'. 13. See 237. 18. Lat. = cos- l -fr = 87 54' nearly. VIII. EXAMPLES (p. 217). 2. 92,819,000 (see Ex. 2, p. 195). 3. At 6 p.m. ; about same length as Midsummer Sun, i.e., 16|h. 4. See 261. 5. 8' 48". 6. Use 266. 7. lOd. 4h. at noon. 8. Gibbous, bright limb turned slightly below direction of W. Hour angle = 30, decl. = 0. 10. (i.) No harvest moon ; (ii.) Phenomena practically unaltered. ANSWERS, 431 VIIL EXAMINATION PAI^ER (p. 218). 4. See 260. 7. 71 33". 9. When we have a solar eclipse. IX. EXAMPLES (p. 236). 1. 23|S. 2. Favourable if moon passes from N. to S. at ecliptic on March 21. 4. 4m. 38s. 5. Length = (Earth's radius) -~ sin (8 P). 7. 6h. 32m. if month unaltered; or, by 329, a lunation = about 10 days, and then time = 2h. 10m. 8. 40 Earth's radii = 158,000 miles (roughly). 9. Total Solar. 1O. 128' (cf. 291). IX. EXAMINATION PAPER (p. 237). 6. 850,000, 230,000, and 5,800 miles (roughly). 7. See 292, 295-297. 9. No. 10. In Fig. 93 take M on xm produced, such that sin xM = #m/(p - P). X. EXAMPLES (p. 265). 1. 291'96 days, or, if conjunctions are of the same kind, 583'92 days. 2. 40. 3. 19 : 6, or nearly 3:1. 4. IQi^h., 120h. 5. p + P s with notation of 290. 6. 888 million miles, 164 yrs. 7. 6 months ; Vi or '63 of Earth's mean distance. 8. 398 days. 9. f of a year = 137 days. 10. Stationary at heliocentric conjunction only, never retrograde. X. EXAMINATION PAPER (p. 266). 3. i_i_ years = 378 days. 4. See 323, 324. The alterations in Venus's brightening are really not inconsiderable (see Ex. 3, p. 205). 6. Most rapid approach at quadrature j velocity that with which the Earth would describe its orbit in synodic period. 9. 287 days. 10. Draw the circular orbits about , radii 4, 7, 10, 16, 52 ( 304). The heliocentric longitudes (measured from Q T ) are roughly as follows: $153, ? 175, 0220, 192; correction for, 192. Geodesy, 77. Geographical latitude, 83. mile, 67. Gibbosity of Mars, 252. Gibbous Moon, 203. Globes : their use, 3. Gnomon, 25, 125. Golden Number, 215. Gravitation : Newton's law of> 352 ; remarks, 353 ; verification for the Earth and Moon, 356. Gravity: to compare its intensity at different places, 329, 334; to find its value, 334. GREGORY, Pope : his correction of the Julian Calendar, 128. Gyroscope or Gyrostat, 321, 395. INDEX. 43? HALLET : his method of deter- mining the Sun's parallax by ob- serving a transit of Venus, 271. Harvest Moon, 216. Heliocentric latitude, 112. longitude, 112. Heliometer, 59. Horizon, celestial, 5; artificial,159, visible, 5, 73-76. - dip of, 73. Horizontal parallax, 191. point, 50. Hour angle, 9 ; expressed in time, 13; its connection with right ascension, 15. circle, 56. Instruments for meridian obser- vations, 35 ; for ex-meridian observations, 54; for geodesy, 78-80 ; for navigation, 153. Introductory Chapter on Spheri- cal Geometry, i.-vi. JULIUS C^SAR : his calendar, 128. Juno, 241,269. Jupiter, 241 ; its satellites, 241. KATER'S reversible pendulum, 329. KEPLER : his laws of planetary motion, 106, 111, 253 ; verifica- tion of his first law, 107, 254 ; verification of the second law, 108, 254 ; deductions from the second law, 109; verification of the third law, 256 ; Newton's deductions from his laws, 339, 345, 346, 348 ; his third law for circular orbits, 340 ; correction of the third law, 354. Knot, 68. Known star, 15, 45. Lagging of the tides, 383-5. Latitude of a place defined, 10; phenomena depending on change of latitude, 65 ; change due to ship's motion, 72. Latitude (continued) : determi- nation by meridian observa- tions, 162 ; determination by ex-meridian observations, 169. celestial, 10. geocentric, 83, 112. geographical, 83. heliocentric, 112. parallel of, 71 ; length of any arc of a given parallel, 71- Leap year, 128. Libra, first point of, 7. Light, refraction of, 140 ; its velo- city, 293; aberration of, 295 to find the time taken by the light from a star to the Earth,. 305. Light-year, 305. Local time: its determination,. 171. Log-line : its use in navigation,. 68. Longitude, celestial, 10. geocentric, 112. heliocentric, 112. terrestrial, 69 ; phenomena depending on change of terres- trial longitude, 70 ; change due- to ship's motion, 72; its deter- mination at sea, 177; the method of lunar distances, 179;. clearing the distance, 179 ; its determination by celestial signals, 181 ; its determination, on land, 182 ; its determination by transmission of chronome- ters, 182; by chronograph, 184^ by terrestrial signals, 185; by Moon-culminating stars, 186;. bv Captain Sumner's method,. 187, Loop of retrogression, 261. Lunar distances, determination of longitude by, 179. geocentric, 180. mountains: determination of their height, 207. Lunation. 27. 488 INDEX. Mars, 240 ; Kepler's observations on Mars, 254 ; its parallax used to determine that of the Sun, 268. Mass, astronomical unit of, 352. Mean noon, 117. solar day, 117. solar time, 117 ; its deter- mination at a given instant of sidereal time, 132. Sun, 116, 117. time, 116. Mercury, 239 ; its period of rota- tion, 264 ; frequency of its transits, 282; its mass, 360, 417. Meridian, celestial, 6. line : its determination, 175. prime, 69. terrestrial, 64. it Meteors : their motion, 4. Metonic cycle, 215. Metre, 67. Micrometers, 58. Midnight, apparent, 24. Mile, geographical, 67. nautical, 67. Moon : its motion, 27 ; its age, 27 ; itsposition denned by its centre, 53; illusory variations in its size, 149 ; method of taking its altitude by the sextant, 158; determination of its parallax, 196 ; its distance, 197 ; its dia- meter determined, 199; its elongation, 200 ; determination of its synodic period, 201 ; its phases, 202; relation between phase and elongation, 204 ; its use in finding the Sun's dis- tance, 205; its appearance i-elative to the horizon, 206; determination of the height of lunar mountains, 207 ; its orbit about the Earth, 209; eccen- tricity of its orbit, 210; its nodes, 210; its perturbations, 210, 407 ; retrograde motion of its nodes, 211, 408, 409. Moon (continued) : progressive motion of its apse line, 211, 410 ; its rotation, 212 ; its librations, 213; general effects of libra- tion, 214 ; its eclipses, 219-221 ; determination of its geocentric distance consistent with an eclipse, 224; its greatest lati- tude at syzygy consistent with an eclipse, 226; synodic revo- lution of its nodes, 228; its occupations, 232 ; verification of the law of gravitation, 356; effect of its attraction, 357 ; its mass, 357 ; concavity of its path about the Sun, 374 ; its disturb- ing or tide-generating force, 375, 377 ; its orbital motion accelerated by tidal friction, 388 ; its form and rotation, 391 ; its disturbing couple on the Earth, 392 ; the rotation of its nodes, 408; its other in- equalities, 410, 411. Nadir, 5. point, determination of, 49. Nautical mile, 67. Neptune, 243 ; its discovery, 417. New Moon, 27. NEWTON, Sir ISAAC : his deduc- tions from Kepler's laws, 339, 345, 346, 348 ; his law of uni- versal gravitation, 352. Nodes, 27, 210 ; their retrograde motion, 211. North polar distance of a circuin- polar star, 17. Number of eclipses in year, 229. Nutation, lunar, 401 ; its general effects, 402 ; its discovery, 403 ; to correct for, 403 ; its physical causes, 404. monthly, 406, solar, 405. Obliquity of ecliptic, 11 j its de termination, 26. INDEX. 439 Observatory, 35. Occultafcions, 232. Offing, 73. Opposition, 200. Parallactic inequality, 411. Parallax, 179, 191; geocentric parallax, 191; horizontal paral- lax, 191 ; general effects of and correction for geocentric parallax, 192 ; relation between horizontal parallax and dis- tance of celestial body, 194; compared with refraction, 195; parallax of Moon determined, 196 ; parallax of planet deter- mined, 198 ; relation between parallax and angular diameter, 199 ; determination of the Sun's parallax, 268 et seqq.; annual parallax denned, 283; to find the correction for annual parallax, 284; relation between the parallax and dis- tance of a star, 285 ; its general effects on the position of a star, 286; determination of the an- nual parallax of a star, 290. Pendulum, Foucault's, 318 ; Cap- tain Kater's reversible, 329 j oscillations of a simple pen- dulum, 330; to find the change in the time of oscillation due to a variation in its length or in the intensity of gravity, 330; to compare the times of oscillation of two pendulums of nearly equal periods, 333 ; pen- dulum method of finding the Earth's mass, 365. Perigee, 106, 210. Perihelion, 111. Perpetual day : determination of its length, 97. Personal equation, 46. Phases of Moon, 202 ; of planet, 251, 252. Perturbations, lunar, 210, 407; rotation of nodes, 408 ; due to average value of radial disturb- ing force, 409; variation, evec- tion, annual equation and parallactic inequality, 410, 411. planetary, 411 ; periodical, 413, 414 ; inequalities of long period, 415 ; secular, 416. Photography, stellar, 60, 421. Planet : its position defined by centre, 53; determination of its parallax, 198 ; its occulta- tion, 235 ; definition, 238 ; in- ferior and superior planets, 244 ; changes in elongation of a inferior planet, 244 ; to find the ratio of the distance from the Sun of an inferior planet to that of the Earth, 246; changes in elongation of a superior planet, 247; to com- pare the distance from the Sun of a superior planet with that of the Earth, 248 ; determina- tion of the synodic period of an inferior planet, 249; relation between the synodic and side- real periods of a planet, 250 ; phases of the planets, 251, 252 ; motions relative to stars, 258 ; transits of inferior planets, 271; its aberration, 306, 307; to compare the velocities and angular velocities of two planets moving in circular orbits, 342 ; having given the velocity of a planet at any point of its orbit, ^ to construct the ellipse de- scribed under the Sun's attrac- tion, 350 ; to find the mass of a planet which has one or more satellites, 359; its perturba- tions, 411 j masses determined, 417. Points of the compass, 9* Polar distance, 9- point : its determination, 51. 440 INDEX. Pole, celestial, 6. terrestrial, 64. Port, establishment of the, 386. Precession of the equinoxes, 103, 392. Earth's axis, 396. a spinning-top, 395. luni-solar, 393 ; to apply the corrections for, 397 ; various effects of, 398; its effects on the climate of the Earth's hemispheres, 400. Prime vertical, 7. Prime vertical instrument : deter- mination of latitude by its use, 170. Priming of the tides, 383-5. -Quadrature, 200. Hadiant, 4. Eeading microscope, 40. Refraction, 140; laws of R., 140; relative index of E., 140; general description of atmo- spherical R., 141 ; its effect on the apparent altitude of a star, 141 ; law of successive R., 142; formula for astro- nomical R., 142 ; Cassini's for- mula, 145; coefficient found by meridian observations, 146 ; other methods of determination, 147 ; its effects on rising and setting, 148; effects on dip and distance of horizon, 149 ; effects on lunar eclipses and occulta- tions, 150, 221 ; comparison of R. with parallax, 195. Retrograde motion, 22, 258. Right ascension, 10 ; expressed in time, 14 ; connection with hour angle, 15. ROEMER : his method of finding the velocity of light, 293. dotation of "Earth, 64, 315 ; of Moon, 212; of Moon'snodes,211, 408 ; of Sun and planets, 264. Saros of the Chaldeans, 231. Satellite, defined, 238 ; their obe- dience to Kepler's laws, 257. Saturn, 242 ; phases of its rings, 252. Seasons, 94 ; effect of the length of day on temperature, 94 ; other causes affecting tempera- ture, 94; unequal length of, 109. Secondary, iii., 238. Sextant, 154 ; its errors, 157 ; determination of theindex error, 157 ; method of taking altitudes at sea, 158 ; method of taking altitudes of Sun or Moon, 158. Sidereal day, 13. month, 200 ; its relation to the synodic month, 200. noon, 13. period, 200, 250. time, 13, 25 ; its disadvan- tages, 115 ; its determination at a given instant of mean solar time, 131 ; its determination at Greenwich or in any longitude, 133. year, 127. Solar day, apparent, 24. system, tabular view of, 243 its centre of mass, 361. time, 24 ; its disadvantages, 115. Solstices, 21, 23. Solstitial colure, 23. points, 23. Southing of stars, 16. Spectrum analysis, 60.' Stars : independence of their di- rections relative to observer's position on the Earth, 4 ; their diurnal motion, 5, 13; culmi- nation, 16 ; southing, 16 ; cir- cumpolar stars, 16 ; rising and setting, 18 ; time of transit, 19 ; to show that a star appears to describe an ellipse, owing to parallax, 287 ; owing to aber- ration, 301. INDEX. 441 Stars, morning and evening, 25. Stationary points, 258 ; their de- termination, 262, 263. Sub-solar point, 187. Summer solstice, 21. and winter, causes of, 94. SUMNER, Captain : his method of finding longitude, 187. Sundial, 125 ; geometrical method of graduation, 126. Sun : its annual motion, 7 ; its annual motion in the ecliptic, 20 ; its motion in longitude, right ascension and declination, 20, 21 ; its variable motion in right ascension, 22 ; determi- nation of its right ascension and declination, 23, 24 ; its position defined by its centre, 53 ; its diurnal path at different sea- sons and places, 88 ; to find length of time of sunrise or sunset, 98 ; observations of its relative orbit, 105 ; its apparent area, 105, 109 ; its apparent annual motion accounted for, 110; illusory variations in size, 149 ; method of finding its alti- tude by the sextant, 158 ; diffi- culty of finding its parallax, 197; its distance determined by Aristarchus, 205 ; solar eclipses, 219, 222, 234; description, 238 ; its period of rotation, 264 ; de- termination of its distance from the Earth, 268 et seqq. ; its paral- lax determined by observation of the parallax of Mars, 268; parallax by observations on the asteroids and Venus, 269; paral- lax determined by observations -of the transit of Venus, 271 et $eqq.; advantages and disadvan- tages of Halley's and Delisle's methods, 280 ; relation between coefficient of aberration, Sun's parallax, and velocity of light, 306. Sun (continued) : to find the ratio of its mass to the Earth's, 358 : gravity on its surface, 366 ; its parallax determined by observa- tions of lunar and solar displace- ments of the Earth, 373; its disturbing or tide-generating force, 375, 377 ; its mass com- pared with that of the Moon, from observations of the rela- tive heights of the solar and lunar tides, 388 ; its disturbing couple on the Earth, 392; gravi- tational methods of finding its distance, 416. Synodic month, 200. period, 200, 250. Syzygy, 200. Telescope, astronomical, 37. Terrestrial equator, 64. longitude, 69. meridian, 64. pole, 64. Theodolite, 79. Tidal constants, 387. friction, 388 ; application to the solar system, 392. Tides, 375 ; equilibrium theory of their formation, 379; canal theory, 380; semi - diurnal, diurnal, and fortnightly tides due to the Moon, 383; semi-diur- nal, diurnal, and six-monthly tides due to the Sun, 383; spring and neap tides, 383_j their priming and lagging, 383- 385 ; establishments of ports 386. Time: its reduction to circular measure, 14 ; relation between the different units, 129, 134. equinoctial, 134. local : its determination by method of equal altitudes, 171, 172. lunar, 382. Trade winds, 323. 442 LtfDEX. Transit, 14 ; eye and ear method of taking transits, 42; of Venus, 271-282"; of Mercury, 282. circle, 38 ; corrections re- quired for right ascension, 44 ; corrections required for decli- nation, 49. Triangulation, 79. Tropics, 88. Tropical year, 127. True Sun, 117. Uranus, 242. Variation, 410. Venus, 240; its period of rota- tion, 264; observations of its transit used to determine the Sun's parallax, 271 ; determi- nation of the frequency of its transits, 281; its mass, 360, 417. Velocity, angular, 342. area!, 343. of light, 293. Velocities of planets compared, 342. Vernier, 157. Vernal equinox, 20, Vertical, 7. circle, 7. prime, 7. Vesta, 2-40. WALLACE, ALFRED RUSSELL : hi method of finding the Earth's- radius, 77. Waning and waxing Moons, 203. Winter solstice, 21. Year, 20. anomalistic, 127. civil, 128. ' leap, 128. sidereal, 127. synodic, 128. tropical, 127. Zenith, 5. distance, 8. point, 51. sector, 8& Zodiac, 25. PRINTED AT THE BURLINGTON PRESS, CAMBRIDOB. THIS BOOK IS DUE ON THE LAST DATE STAMPED BELOW AN INITIAL FINE OF 25 CENTS WILL BE ASSESSED FOR FAILURE TO RETURN THIS BOOK ON THE DATE DUE. THE PENALTY WILL INCREASE TO SO CENTS ON THE FOURTH DAY AND TO $1.OO ON THE SEVENTH DAY OVERDUE. UNIVERSITY .QF.CALIFORNIA LIBRARY