THE LIBRARY 
 
 OF 
 
 THE UNIVERSITY 
 
 OF CALIFORNIA 
 
 DAVIS 
 
 GIFT OF 
 
 PROFESSOR H.B. WALKER 
 
PHYSICAL 
 MEASUREMENTS 
 
 DUFF and EWELL 
 
BLAKISTON'S SCIENCE SERIES 
 
 PHYSICAL 
 MEASUREMENTS 
 
 BY 
 
 A. WILMER DUFF 
 
 PROFESSOR OF PHYSICS IN THE WORCESTER POLYTECHNIC INSTITUTE 
 
 AND 
 
 ARTHUR W. EWELL 
 
 PROFESSOR OF PHYSICS IN THE WORCESTER POLYTECHNIC INSTITUTE 
 
 SECOND EDITION, REVISED AND ENLARGED 
 WITH 78 ILLUSTRATIONS 
 
 PHILADELPHIA 
 
 P. BLAKISTON'S SON & CO 
 
 1012 WALNUT STREET 
 1910 
 
 LIBRARY 
 
 (JNfVFRSITY nr rmtr/\nutn 
 
COPYRIGHT, 1910, BY P. BLAKISTON'S SON & Co. 
 
 Printed by 
 
 The Maple Press 
 
 York, Pa. 
 
PREFACE. 
 
 Our intention in writing this book was not to give an 
 account of physical laboratory methods in general, but to 
 describe a limited number of carefully chosen exercises such 
 as we have found in our experience to be suitable for the 
 laboratory work of students who have had a fair course 
 in General College Physics. 
 
 The descriptions of the exercises will usually fit apparatus 
 and conditions of considerable diversity, but many practical 
 details have been included where experience has shown that 
 they are necessary. Other instructors who may adopt the 
 book will probably find some of the exercises unsuited to 
 their classes, but the list is sufficiently extensive to afford 
 a considerable variety of selection. 
 
 The descriptions of apparatus are intended to be read by 
 the student with the apparatus before him. Hence elaborate 
 illustrations have been thought unnecessary. For an ex- 
 tended account of certain special topics, such as the theory 
 of the balance and the construction of galvanometers, 
 references to other works have been given. 
 
 Usually several text-books and special treatises have 
 been referred to at the beginning of the account of an experi- 
 ment. It is assumed that each student will have one of the 
 text-books and that some of the special works will be found in 
 the reference room of the laboratory. While the reference 
 is generally to the latest edition (at the present date, 1910), 
 those who have different editions will have no difficulty in 
 finding the passages referred to. Each instructor who uses 
 
 v 
 
VI PREFACE 
 
 the book will exercise his discretion as to what preliminary 
 reading will be required and will issue the necessary instruc- 
 tions to his class. 
 
 We are indebted to Dr. Albert W. Hull for assistance in 
 reading the page proof. Many of the tables have been taken 
 from Ewell's Physical Chemistry. 
 
CONTENTS. 
 
 PAGE 
 
 GENERAL INTRODUCTION i 
 
 i. Purpose of Course. 2. General Directions. 3. Reports. 
 4. Errors. 5. Errors of Observation. 6. Possible Error 
 of a Calculated Result. 7. General Method for the Possible 
 Error of a Result. 8. Some General Notes on Errors. 9. 
 Probable Error of a Mean. 10. Limits to Calculations, n. 
 Notation of Large and Small quantities. 12. Plotting of 
 Curves. 
 
 MECHANICS 13 
 
 13. The Use of a Vernier. 14. Vernier Caliper. 15. Microm- 
 eter Caliper. 16. Micrometer Microscope. 17. Com- 
 parator. 1 8. Spherometer. 19. Dividing Engine. 20. 
 Cathetometer. 21. Barometer. 22. The Balance. 23. Ad- 
 justment of Telescope and Scale. 24. Time Determination. 
 
 I. To Make and Calibrate a Scale. 
 II. Errors of Weights. 
 
 III. Volume, Mass, and Density of a Regular Solid. 
 
 IV. Mohr-Westphal Specific Gravity Balance. 
 V. Density by the Volumenometer. 
 
 VI. Density of Gases. 
 
 VII. Acceleration of Gravity by Pendulum. 
 
 VIII. Coefficient of Friction. 
 
 IX. Hooke's Law and Young's Modulus. 
 
 X. Rigidity (or Shear Modulus). 
 
 XL Viscosity. 
 
 XII. Surface Tension. 
 
 HEAT 63 
 
 25. Radiation Correction in Calorimetry. 26. The Beck- 
 mann Thermometer. 
 XIII. Thermometer Testing. 
 XIV. Temperature Coefficient of Expansion. 
 
 vii 
 
VI11 CONTENTS. 
 
 PAGE 
 
 XV. Coefficient of Apparent Expansion of a Liquid. 
 
 XVI. Coefficient of Increase of Pressure of Air. 
 
 XVII. Pressure of Saturated Water Vapor. 
 
 XVIII. Hygrometry. 
 
 XIX. Specific Heat by Method of Mixture. 
 
 XX. Ratio of Specific Heats of Gases. 
 
 XXI. Latent Heat of Fusion. 
 
 XXII. Latent Heat of Vaporization. 
 
 XXIII. Latent Heat of Vaporization. Continuous-flow 
 
 method. 
 
 XXIV. Thermal Conductivity. 
 
 XXV. The Mechanical Equivalent of Heat. 
 XXVI. The Melting-point of an Alloy. 
 XXVII. Heat Value of a Solid. 
 XXVIII. Heat Value of a Gas or Liquid. 
 XXIX. Pyrometry. 
 
 SOUND 119 
 
 XXX. The Velocity of Sound. 
 XXXI. The Velocity of Sound by Kundt's Method. 
 
 LIGHT 124 
 
 27. Monochromatic Light. 28. Rule of Signs for Spherical 
 Mirrors and Lenses. 
 XXXII. Photometry. 
 
 XXXIII. Spectrometer Measurements. 
 
 XXXIV. Radius of Curvature. 
 XXXV. Focal Length of a Lens. 
 
 XXXVI. Lens Combinations. 
 
 XXXVII. Magnifying Power of a Telescope. 
 
 XXXVIII. Resolving Power of Optical Instruments. 
 
 XXXIX. Wave-length of Light by Diffraction Grating. 
 
 XL. Interferometer. 
 
 XLI. Rotation of Plane of Polarization. 
 
 ELECTRICITY AND MAGNETISM 153 
 
 29. Resistance Boxes. 30. Forms of Wheatstone's Bridge. 
 31. Galvanometers. 32. Correction for Damping of a Bal- 
 listic Galvanometer. 33. Galvanometer Shunts. 34. 
 Standard Cells. 35. Device for Getting a Small E. M. F. 
 36. Double Commutator. 37. Relations between Electrical 
 Units . 
 
CONTENTS. 
 
 IX 
 
 XLII. Horizontal Component of the Earth's Magnetic 
 
 Field. 
 
 XLIII. Magnetic Inclination or Dip. 
 XLIV. Measurement of Resistance by Wheatstone's 
 
 Bridge. 
 
 XLV. Galvanometer Resistance by Shunt Method. 
 XLVI. Galvanometer Resistance by Thomson's Method 
 XLVII. Measurement of High Resistance (i). 
 XLVIII. Measurement of High Resistance (2). 
 XLIX. Measurement of Low Resistance (i). 
 L. Measurement of Low Resistance (2). 
 LI. Measurement of Low Resistance (3). 
 LI I. Comparison of Resistances by the Carey-Foster 
 
 Method. 
 
 LIII. Battery Resistance by Mance's Method. 
 LIV. Temperature Coefficient of Resistance. 
 LV. Specific Resistance of an Electrolyte. 
 LVI. Comparison of E. M. F.'s by High Resistance 
 
 Method. 
 LVII. Comparison of E. M. F.'s and Measurement of 
 
 Battery Resistance by Condenser Method. 
 LVIII. Calibration of Voltmeter. 
 LIX. Calibration of Ammeter. 
 LX. Comparison of Capacities of Condensers. 
 LXI. Absolute Determination of Capacity. 
 LXII. Coefficients of Self-induction and of Mutual Induc- 
 tion. 
 
 LXIII. Strength of a Magnetic Field by a Bismuth Spiral. 
 LXIV. Study of a Ballistic Galvanometer. 
 
 LXV. Magnetic Permeability. 
 LXVI. Magnetic Hysteresis, 
 j VXTTT / (a) Mechanical Equivalent of Heat. 
 
 \ (b) Horizontal Intensity of Earth's Magnetism. 
 LXVIII. Thermoelectric Currents. 
 LXIX. Elementary Study of Resistance, Self-induction, 
 
 and Capacity. 
 LXX. Self-induction, Mutual Induction, and Capacity, 
 
 Alternating Currents. 
 LXXI. Dielectric Constant of Liquids. 
 LXXII. Electric Waves on Wires. 
 
 TABLES 
 
 I. Four-Place Logarithms. 
 
 2 35 
 
CONTENTS. 
 
 II. Trigmometrical Functions. 
 
 III. Reduction to Infinitely Small Arc. 
 
 IV. Barometer Corrections. 
 
 V. Density and Specific Volume of Water. 
 VI. Density of Gases. 
 VII. Density, Specific Heat, and Coefficient of Expansion of 
 
 Metals. 
 VIII. Density, Specific Heat, and Coefficient of Expansion of 
 
 Miscellaneous Substances. 
 XL Elastic Moduli. 
 
 X. Surface Tension. 
 XI. Coefficient of Viscosity. 
 XII. Specific Heats of Gases. 
 
 XIII. Vapor Pressure of Water. 
 
 XIV. Boiling-point of Water. 
 
 XV. Wet and Dry Bulb Hygrometer. 
 XVI. Vapor Pressure of Mercury. 
 XVII. Melting-points of Metals. 
 XVIII. Wave-lengths of Light. 
 XIX. Refractive Indices. 
 
 XX. Specific Rotatory Power. 
 XXL Photometric Table. 
 
 XXII. Specific Resistance and Temperature Coefficient of 
 Metals. 
 
 XXIII. Specific Resistance and Temperature Coefficient of 
 
 Solutions. 
 
 XXIV. Dielectric Constants. 
 
 INDEXT 255 
 
PHYSICAL MEASUREMENTS. 
 
 INTRODUCTION. 
 
 1. Purpose of Course. 
 
 Intelligent work requires a clear perception of the end 
 in view. It is important, therefore, to remember that the 
 purpose of a course in Laboratory Physics is not only the 
 attainment, by personal experimentation, of a more definite 
 knowledge of the facts and principles of physics and an 
 acquaintance with the use of measuring instruments and 
 methods, but also the acquisition of a scientific habit of 
 accuracy and carefulness in observing and examining phe- 
 nomena and drawing conclusions therefrom. 
 
 2. General Directions. 
 
 Much time in the laboratory will be wasted unless some 
 preparation be made before coming to the laboratory. The 
 purpose and general method of the measurement to be made 
 should be examined with the aid of the text-book and some 
 of the references preceding the directions. This may 
 usually be done in a few minutes at home, whereas it might 
 require an hour or more in a laboratory where a number 
 of people are moving around. 
 
 The readings made in the laboratory should always be 
 recorded in a firmly bound book reserved for this purpose 
 only, and never on loose slips of paper or in a book that may 
 become dog-eared and untidy. When, for convenience or 
 of necessity, two work together at an experiment, each 
 should keep his own notes of the measurements made, and, 
 
 i 
 
2 INTRODUCTION. 
 
 whenever possible, each should make a separate set of read- 
 ings for himself, and these should be as independent as 
 possible. 
 
 No operation should be performed or measurement 
 made unless the purpose and meaning of it are understood ; 
 otherwise it may be made imperfectly or some essential 
 part of it may be overlooked. 
 
 3. Reports. 
 
 An essential part of the work is a written report on each 
 experiment completed. This should be handed in within a 
 week after the work is finished. In preparing the report 
 the writer has to make clear to himself the purpose and 
 bearing of each part of the work and examine critically the 
 value and accuracy of the final result. This exercise is as 
 valuable as the experimental work itself. The report 
 should be as brief as possible, consistently with giving the 
 following information : 
 
 The purpose of the experiment (including the definition 
 of the leading terms, such as coefficient of friction, mechani- 
 cal equivalent of heat, etc.) ; 
 
 A brief statement of the method used ; 
 
 A statement (tabulated if possible) of the observations 
 and readings made : 
 
 An outline of the calculation of the final result (omit- 
 ting the details of the numerical work) ; 
 
 A criticism of the reliability of the result ; 
 
 Brief answers to the questions appended to the directions. 
 
 4. Errors. 
 
 A perfectly accurate experimental result is impossible; 
 but some estimate can usually be formed as to the magni- 
 tude of the possible error and this is frequently of the 
 greatest value. An experimental result of unknown reli- 
 ability is often of very little value. Hence an estimate of 
 
ERRORS OF OBSERVATION. 3 
 
 the accuracy of a measurement is very desirable in an 
 account of the work. 
 
 Inaccuracy may arise from several different causes 
 (i) errors of observation, due to the inherent limitations 
 of the observer's powers of observing and judging; (2) 
 instrumental errors, arising from imperfections in the work 
 of the instrument maker in constructing and subdividing 
 the scale used by the observer; (3) mistakes, such as the 
 mistaking of an 8 for a 3 on a scale; (4) systematic errors 
 due to faultiness in the general method employed. 
 
 Instrumental errors may be decreased by using more 
 accurate instruments or by calibrating the scales of the instru- 
 ments used, that is, ascertaining and allowing for the errors 
 in their graduation. This is frequently a difficult operation 
 and unsuited for an elementary course. We shall, therefore, 
 usually assume that the accuracy of the instruments is 
 such that the instrumental errors are less than the errors of 
 observation. 
 
 Mistakes in reading can be eliminated by care and repeti- 
 tion. Systematic errors are apt to arise when some indirect 
 method of arriving at a result is adopted, a direct method 
 being difficult or impossible. For example, the length of 
 a wave of light cannot be measured directly and a method 
 depending on diffraction or interference is usually employed 
 (Exp. XXXIX). A careful study of the method used will 
 often enable us to eliminate such errors by improving 
 the details of the method, or, where this cannot be done, 
 some estimate of the uneliminated errors can often be formed. 
 
 5. Errors of Observation. 
 
 Different methods of estimating the magnitude of errors of 
 observation may be employed, the choice depending on the 
 nature of the measurements. In many cases the quantity 
 can be measured several times and the mean taken, it 
 being probably more accurate than a single observation. 
 In other cases circumstances do not permit repetition and a 
 
4 INTRODUCTION. 
 
 single observation must suffice. In either case the observer 
 can, from the circumstances of the case, say with a high de- 
 gree of probability that the error cannot be greater than a 
 certain magnitude. This we shall call the "possible error" 
 of the measurement. It does not strictly mean the greatest 
 possible error, since a greater error might be theoretically 
 possible but very improbable. 
 
 (a) When Only a Single Observation is Made. For example, 
 a liquid, the temperature of which is varying slowly, is kept 
 well stirred and the temperature is observed by means of a 
 thermometer graduated to degrees. The temperatur at a 
 certain time is noted as being between 36 and 37 and the 
 observer, estimating to o . i of a division, records the tem- 
 perature as 36.3; but he does not trust his estimate closer 
 than o . i ; that is, he considers that the real temperature may 
 be as high as 36 . 4 or as low as 36 . 2. He therefore states 
 the temperature as 36.3 with a possible error of 0.1, or 
 36.3o.i. The actual error may, of course, be less than 
 o. i; the latter is only a reasonable estimate of the limit of 
 error of observation. 
 
 (b) When Several Different Observations of a Quantity 
 are Made. The mean of a number of observations of a 
 quantity is more trustworthy than a single reading, for 
 observations that are too large are likely to counterbalance 
 others that are too small. Greater confidence can be 
 placed in the mean when the separate readings differ but 
 little from the mean than when they differ greatly. The 
 average of the differences between the mean and the 
 separate readings is called the mean deviation. It can be 
 shown (as indicated in 9) that when ten observations are 
 made, the probability that the actual error is greater than 
 the mean deviation is very small, about i in 100, while if 
 15 observations are made it is reduced to i in 1000. Even 
 if only 5 observations are made (which is rather too small a 
 number) the probability is only i in 15. Hence, when a 
 quantity is measured several times, the average deviation 
 may be taken as a measure of the possible error. 
 
POSSIBLE ERROR OF A CALCULATED RESULT. 5 
 
 6. Possible Error of a Calculated Result. 
 
 A piece of laboratory work usually calls for the measure- 
 ment of several different quantities and the calculation of a 
 result by some formula. Knowing the possible errors of 
 the separate quantities we can deduce the possible error 
 of the result, but the method will vary with the nature of 
 the arithmetical operations. 
 
 (a) Possible Error of a Sum or Difference. The possible 
 error of a sum or difference is the sum of the possible errors 
 of the separate quantities, for each possible error may be 
 either positive or negative. 
 
 Example. A bulb containing air (Exp. VI) weighs 
 2o.i425g. 0.0002 g. and after the air has been pumped 
 out it weighs 20.0105 g. o.ooo2g. Hence the weight of the 
 air is o . 1 3 2 o g. o . 0004 g. Since it is sometimes erroneously 
 assumed that a derived result must be accurate to as high a 
 percentage as the measurements from which it is deduced, 
 it should be noticed in the above that, while the separate 
 weights are found to 0.001%, the weight of the air is only 
 ascertained to 0.3%. 
 
 (b) Possible Error of a Power. If a measured quantity 
 x is in doubt by p per cent (p being small), the nth power 
 of x is in doubt by up per cent. For 
 
 x\ i- 
 
 IOO/ I \ IOO 
 
 squares and higher powers of p/ioo being neglected. 
 
 Example of (a) and (b). 
 
 T = 3. 506 .005 and/ = 2.018 -003. (Exp. VII). What 
 is the possible error of T 2 t 2 ? T 2 = 12. 2 9 and since T may be in 
 error by 1/7 %, T 2 maybe in error by 2/7 % or .04. Hence 
 T 2 = 12 . 29 .04. Similarly 2 = 4.07 .01. Hence T 2 t 2 = 
 
 8. 22 .05. 
 
 (c) Possible Error of a Product or Quotient. The percent- 
 age by which a product or quotient is in doubt is the sum of 
 
6 INTRODUCTION. 
 
 the percentages by which the separate quantities are in 
 doubt. For if the quantities be 
 
 / P \ / <1 \ 
 
 x( i ) and y{ i I 
 
 V ioo/ \ ioo/ 
 
 their product is 
 
 x ( I JL\J 1 A\^ xy L P^} 
 
 \ IOO/ \ IOO/ \ IOO/ 
 
 and their quotient is 
 
 p/ioo and q/ioo being assumed small. It is evident that 
 a similar statement applies to any number of products and 
 quotients. 
 
 Example of (b) and (c) . 
 
 The diameter of a sphere (Exp. Ill) is measured by 
 a vernier caliper and found to be i . 586 cm., but the vernier 
 only reads to 1/50 mm.; so the possible error is .002 cm. or 
 1/8 of i%. The sphere is weighed in a balance such that i 
 mg. added to one pan does not cause an observable 
 change of the pointer, while 2 mg. does, and the weight is, 
 therefore, 16.344 g. with a possible error of .002 g. or 
 1/80 %. The calculated value of the density is 7 . 827 ; but the 
 volume may be in error by 3/8% and the mass by 1/80%. 
 Hence the density may be in error by 3 / 8 + 1 / 80 % , or practi- 
 cally 3/8%. Hence the proper statement of the density 
 is 7 . 83 with a possible error of .03 or 7 . 83 . 03. 
 
 7. General Method for the Possible Error of a Result. 
 
 The above rules for sums, differences, powers, products, 
 and quotients will usually suffice for finding the possible 
 error of a result calculated from the measurements of 
 several quantities. But when several of these operations 
 are combined, or when the formula for calculation contains 
 
GENERAL METHOD FOR THE POSSIBLE ERROR. 7 
 
 one of the quantities more than once, the effects of the 
 several errors may be difficult to trace by these means. The 
 following general method is always applicable. It may be 
 carried out by simple arithmetic, but is simplified by an 
 elementary use of the calculus. 
 
 To find to what extent the possible error in one of the 
 quantities affects the result, we may calculate the result 
 assuming all the quantities to be quite accurate and then 
 repeat the calculation after changing one of the quantities 
 by its possible error. The difference in the result will be the 
 effect sought. If we do the same for each of the other 
 quantities, the final possible error of the result will be the 
 sum (without regard to sign) of the parts due to the separate 
 quantities. 
 
 This, however, is equivalent to differentiating the whole 
 expression, first with regard to one quantity, then with 
 regard to a second and so on and finally adding the partial 
 differentials. It will be seen from the following examples 
 that the process is much simplified by taking the logarithm 
 of the whole formula before differentiating. 
 
 (i) Time of Vibration of a Pendulum (Exp. VII). If in 
 time T a pendulum makes n fewer vibrations than the 
 pendulum of a clock that beats seconds and if t is the time 
 of a single vibration, 
 
 T 
 
 t = T-n 
 Taking logarithms, 
 
 log t = log T-log(T-n) 
 Hence by differentiating, 
 
 &_sr sr 
 
 t T T-n 
 
 This means that if T be changed by a small quantity, ST 
 the consequent change, Bt, in t is given by the formula. If 
 
8 INTRODUCTION. 
 
 the possible error of T be 2 seconds, by putting 8T = 2 
 the value of $t will be the possible error of t. If T be 862 
 seconds and n be 17, 
 
 & 17X2 
 
 This indicates one of the advantages of taking logarithms. 
 It gives us at once the ratio of $t to t, or (multiplied by 100) 
 the percentage by which t is in doubt. 
 
 (2) Specific Heat by the Method of Mixture (Exp. XIX). 
 Let r = 95 be the initial temperature of the specimen, 
 2 = 25 that of the water, and let t = 45 be the final tempera- 
 ture of the mixture, and let the possible error of each ther- 
 mometer reading be o. 2. The formula for calculation is 
 
 ^ 
 
 M(T-t) 
 
 We shall consider how far the possible errors in the ther- 
 mometer readings affect x, leaving the consideration of the 
 other terms (the errors of which are likely to be much smaller) 
 to the reader. 
 
 log x=log(m+m l s) +log(tt ) logM log(T-t) 
 
 Proceeding as in (i) above, we find the effects of the possible 
 errors of T, t , and / respectively as follows : 
 
 $x SjT o. 2 
 
 = 0.4% 
 
 T t 50 
 
 & 0.2 
 
 /-/ 20 
 
 = 1 .O 
 
 70 
 
 
 Total = 2. 8% 
 
 This example will show a second advantage in the method of 
 taking logarithms. It separates the various terms and so 
 simplifies the process. 
 
OF A MEAN. 9 
 
 8. Some General Notes on Errors. 
 
 The statement of a possible error should contain only one 
 significant figure. (A zero that serves only to fix the decimal 
 point, such as the zeros in 0.0026, is not a significant 
 figure). Thus in the last example in 6, 3/8% of 7.83 is 
 0.0293, which shows that the second decimal place in 7.83 
 is in doubt by 3. Hence it would be superfluous to add 
 figures to show that the third and fourth decimal places are 
 also in doubt. 
 
 Measurements sometimes seem so accurate that one is 
 tempted so say that "the possible error is practically zero 
 and need not be considered." This is never literally true. 
 One factor may be so accurately determined, compared with 
 other factors, that the effect of its possible error on the result 
 might seem to be negligible ; but only a calculation can show 
 this and the calculation will frequently show the opposite. 
 An illustration occurs in the first example of 6, considered 
 in connection with the other measurements required to 
 determine the density of air in Exp. III. 
 
 The consideration of possible errors is of great importance 
 in deciding what care need be expended in determining the 
 various factors in a complex measurement and what are the 
 best conditions for obtaining an accurate result. This 
 applies more especially to advanced and difficult measure- 
 ments, but illustrations will occur in this book. (Exp. XIX.) 
 But, as one of the purposes of this course is to teach the most 
 exact use of the measuring instruments, measurements 
 should usually be made as accurately as the instruments 
 will permit. 
 
 9. "Probable Error" of a Mean. 
 
 There is another method of indicating the reliability of 
 measurements which possesses some advantages over the 
 one that we have explained, though it is not so generally 
 applicable. When a large number of observations of a 
 quantity have been made, we can, by means of formulas 
 
10 INTRODUCTION. 
 
 deduced from the mathematical Theory of Probability, cal- 
 culate the probability that the mean is not in error by more 
 than a given amount. When a coin is tossed up it is an even 
 chance whether it will come down a head or a tail; the 
 chance or probability of its being a head is, therefore, i in 2 
 or 1/2. Now the "probable error" of the mean of a num- 
 ber of readings is defined as a magnitude such that it is an 
 even chance whether the error is greater or whether it is 
 less than this magnitude. In other words, the probabil- 
 ity of the error exceeding the "probable error" is 1/2. 
 One formula for calculating the "probable error" is the 
 following : 
 
 average deviation 
 #=0.84^ 
 
 A/number of observations 
 
 This method is useful as a method of indicating the reli- 
 ability of measurements when each of all the quantities that 
 occur in the experiment can be measured several times. 
 For when each mean and its "probable error" has been 
 found we can calculate the "probable error" of the final 
 result. For further details we shall refer the reader to other 
 works (e. g., Merriman's "Least Squares"). We shall not 
 have frequent occasion to refer to " probable errors," since in 
 most cases some of the quantities that have to be determined 
 cannot be measured more than once. 
 
 In justification of the use of the mean deviation as a 
 measure of the possible error, we may note that by the above 
 formula for e, when 10 observations have been made, the 
 mean deviation equals 3 . 6 <?. Now a reference to tables 
 of the probability of errors (e. g., Smithsonian Tables) shows 
 that the probability of an error greater than 3.8 e is about i 
 in 100. 
 
 10. Limits to Calculations. 
 
 By the above methods the possible error in any calcula- 
 tion from experimental quantities may be deduced. The 
 magnitude of the possible error in any calculation indicates 
 

 PLOTTING OF CURVES. II 
 
 how far it is useful and desirable to carry the calculation. 
 A calculation should be carried as far as, but not farther than, 
 the first doubtful figure. This rule must be applied not only 
 to the calculation of the final result, but also to each inter- 
 mediate step. When a calculation is carried too far, useless 
 and very unscientific labor is expended, and when it is not 
 carried far enough very absurd results are often obtained. 
 
 In addition and subtraction a place of decimals that is 
 doubtful in any one of the quantities is doubtful in the 
 result. 
 
 In multiplication and division (performed in the ordinary 
 way) decimal places that have not been determined are 
 usually filled up by zeros. 'Any figure in the result that 
 would be altered by changing one of these zeros to 5 is 
 doubtful. 
 
 ii. Notation of Very Large and Very Small Numbers. 
 
 Partly to save space and partly to indicate at once the 
 magnitude of very large' or very small numbers, the following 
 notation is used. The digits are written down and a decimal 
 point placed after the first and its position in the scale 
 indicated by multiplying by some power of 10. Thus 
 42140000 is written 4.2i4Xio 7 and .00000588 is written 
 5.88X10-. This also enables us to abbreviate the multi- 
 plication and division of such numbers. Thus 42 140000 X 
 .00000588 is the same as 4. 214 X 5-88X 10 and 42140000-7- 
 .00000588 is the same as (4. 214-=- 5. 88) X io 13 . 
 
 12. Plotting of Curves. 
 
 It is assumed that the general method of the represen- 
 tation of the connection between two related quantities by 
 means of a curve is familiar to the reader from his work in 
 Graphical Algebra or elsewhere. Attention may be called 
 to the following points: 
 
 i. Mark experimental points clearly by crosses or circles 
 surrounding the points. 
 
12 INTRODUCTION. 
 
 2. The curve should be drawn so as to strike an average 
 path among the points; it does not have to pass through even 
 one point. 
 
 3. Abscissas may be drawn to any scale and ordinates to 
 any scale. Record the main division of the scales along 
 each axis. (Do not record the individual observations.) 
 
 4. For convenience, all the abscissas may be diminished 
 by the same amount before plotting, and the same is true of 
 ordinates. 
 
 5. Ordinates and abscissas should be drawn to such scales 
 that the curve occupies a large part of the paper. 
 
 6. Curves should be drawn carefully and neatly by means 
 of curve forms. 
 

 MECHANICS. 
 
 13. The Use of a Vernier. 
 
 The vernier is a contrivance for reading to fractions of the 
 units in which scales are graduated. It is a second scale 
 parallel to the main scale of the instrument and so divided 
 that n of its units equal n i or n -f i of the units of the scale. 
 If 5 is the length of a scale unit and v that of a vernier unit, 
 in the first case 
 
 nv=(n i)s or s v s-^-n; 
 in the second case 
 
 nv=(n + i)s or v s = s + n. 
 
 Hence the unit of the vernier is less or greater than that of 
 the scale by one-wth of a scale unit. 
 
 If we did not have a vernier there would be something in 
 the nature of an index to indicate what division of the scale 
 should be read in making a certain measurement and 
 fractions would be estimated by eye. The zero of the vernier 
 is taken as such an index, the whole number of scale divisions 
 being the number just below 
 the zero of the vernier, while 
 the fraction of a scale division 
 
 is determined with the vernier. A |J I I IJ I I I I ,| 
 
 If the wth division of the FlG T 
 
 vernier coincides with the scale 
 
 division, the zero of the vernier must be m wths of a scale 
 
 unit from the scale division just below it. Thus the use of 
 
 a vernier divided into n parts is equivalent to subdividing 
 
 the scale unit into n parts. 
 
 If no vernier division exactly coincides with a scale 
 division there will be two vernier divisions nearly coincident 
 
MECHANICS. 
 
 
 
 FIG. 2. 
 
 with scale divisions, and one can often estimate fractions 
 of the fraction given by the vernier. In figure i the 
 whole number of scale divisions is 5.2, and evidently the 
 third and fourth vernier divisions most nearly coincide with 
 scale divisions. Since the third division is 
 somewhat nearer a coincidence, we may call 
 the fraction 3 . 3 tenths, or, the full reading 
 will be 5 . 233. 
 
 14. Vernier Caliper. 
 
 The vernier caliper consists of a straight 
 graduated bar, and two jaws at right angles 
 to it, one of which is fixed while the other 
 is movable. The position of the movable 
 jaw can be accurately determined by means 
 of the scale and a vernier which should read 
 zero when the jaws are in contact. (If this be 
 not the case, allowance must be made for the zero reading) . 
 
 15. Micrometer Caliper. 
 
 The micrometer caliper is a U-shaped piece of metal in 
 one arm of which is a steel plug with a carefully planed face 
 and through the other arm of which passes a screw with a 
 plane end parallel and opposite to that of the 
 screw. A linear scale on the frame reads zero 
 approximately when the plug and screw are in 
 contact, and its reading in any other position 
 indicates the whole number of turns of the 
 screw and consequently the number of mm. 
 (or 1/2 mm. or 1/40 in. as the pitch of the 
 screw may be) between the screw and the plug. 
 Fractions of a turn are read on the divided 
 head. As contact approaches, the screw should be turned 
 with a very light touch and the same force used for 
 different contacts. Some micrometers are provided with a 
 ratchet head which permits only a definite, moderate 
 pressure. 
 
MICROMETER MICROSCOPE. 15 
 
 16. Micrometer Microscope. 
 
 The micrometer microscope is a microscope with cross- 
 hairs at the focus. In one type of instrument these cross- 
 hairs are movable by a micrometer screw. In the other 
 and more common type the whole microscope is moved by 
 a micrometer screw (see Fig. 4). The most elaborate 
 instruments have both movements. The rotations of the 
 screw are read on a fixed linear scale while the fraction of a 
 rotation is read by a circular scale attached to the screw, 
 and thus the amount of movement is ascertained if the 
 pitch of the screw is known. The pitch is best determined 
 by reading a length on a reliable scale placed in the field of 
 view. 
 
 17. Comparator. 
 
 The comparator consists essentially of a pair of micro- 
 scopes movable along a horizontal bar to which they are at 
 right angles. The length to be measured is placed under the 
 
 FIG. 4. 
 
 microscopes. The eye-piece of each microscope is first focused 
 clearly on the cross-hairs and the whole microscope focused 
 without parallax on the point to be observed, so that the 
 image of the point coincides with the intersection of the 
 cross-hairs. The object is then removed and a good scale 
 put in its place, and a reading of the scale gives the required 
 length, this reading being facilitated by the use of microm- 
 eter screws. 
 
1 6 MECHANICS. 
 
 18. Spherometer. 
 
 The Spherometer is an instrument with four legs, three 
 of which form the vertices of an equilateral triangle, while 
 the fourth is at the center of the triangle. The fourth leg 
 can be screwed up and down and the distance of its extrem- 
 ity from the plane of the extremities of the other three legs 
 can be accurately measured by means of a linear scale 
 attached to the fixed legs and a circular scale attached to 
 the movable leg. The linear scale gives the number of 
 complete turns of the screw and the circular scale the frac- 
 tion of a turn. These scales are read when the screw makes 
 contact with an object placed beneath. 
 
 The position of contact may be determined by noticing 
 that the screw turns very easily for a fraction of a turn just 
 after contact begins. This is due to reduced friction in the 
 bearing, owing to the weight of the screw falling on the 
 body in contact and to the back-lash of the screw, the 
 frame not yet being raised. The screw should 
 be lowered until it thus begins to turn very 
 easily; it should then be turned back again 
 until it just again begins to turn hard and a 
 reading then made. A less sensitive method 
 is to turn the screw down until the instrument 
 is felt to rock or wobble and make a reading; 
 then raise the screw until rocking just ceases, 
 make another reading and take the mean of 
 FIG. 5. "the two as the contact reading. In some 
 spherometers the end of the screw on making 
 contact raises two levers arranged to greatly magnify the 
 motion. The screw is lowered until the top lever comes to 
 some definite position, for instance, with the end opposite 
 to a stud in the frame of the instrument. 
 
 The zero reading is the reading of the scales when the end 
 of the screw is in the plane of the ends of the legs, and is 
 so called because the instrument is most frequently used to 
 determine distances above or below this plane. It may be 
 

 THE DIVIDING ENGINE. 17 
 
 obtained by placing the instrument on a very plane plate of 
 glass. Several zero readings should be made before and 
 after making readings with the object in position under the 
 screw, for the zero reading is likely to change from slight 
 disturbances of the adjustment of the instrument and ther- 
 mal expansion due to the heat of the hand. If the mean of 
 the zero readings made after reading on an object should be 
 decidedly different from the mean of those made before, 
 readings should be again made on the object and the first 
 zero readings discarded. After each reading, the screw 
 should be turned up through at least a quarter revolution, 
 that the readings may be entirely independent. 
 
 The unit of the linear scale may be obtained by com- 
 parison with a standard steel scale. If the plane of the 
 circular scale is not exactly perpendicular to the axis, the 
 linear scale may not give the correct number of turns. It 
 is therefore best, as a check upon the linear scale readings, 
 to count the number of rotations. 
 
 19. The Dividing Engine. 
 
 The dividing engine received its name from its being 
 originally made to subdivide scales, diffraction-gratings, 
 etc. An equally frequent use of the instrument is for the 
 accurate measurement and calibration of scales, gratings, 
 etc. 
 
 It consists essentially of (i) a very carefully made hori- 
 zontal screw, the ends of which are so supported that the 
 screw is free to rotate, but not to advance or recede; (2) a 
 nut movable on the screw and bearing against (3) a plat- 
 form movable along a track which is parallel to the screw; 
 (4) a micrometer microscope which in some instruments is 
 held in a support movable along a rail on the same bed 
 plate as the track, but in other instruments is carried on the 
 platform; (5) dividing gear for making scales, etc.; (6) a 
 divided circular scale attached to the screw with a vernier 
 attached to the bed plate. 
 
1 8 MECHANICS. 
 
 If, by means of the circular scale and vernier, the rota- 
 tion of the screw can be read to i in a very large number, 
 say in i in 1,000, then, since the nut moves a distance equal to 
 the "pitch" of the screw (measured parallel to the axis) 
 when the screw is given one complete rotation, it follows 
 that the movement of the nut and platform can be 
 read to a correspondingly small fraction of the pitch of 
 the screw. 
 
 The object whose length is to be measured is placed on 
 the movable platform (in the case of instruments of the first 
 type mentioned under (4) above) . The microscope is focused 
 on one end of the length to be measured, so that the inter- 
 section of the cross-hair coincides with that end. By turn- 
 ing the screw until the movement of the platform brings the 
 other end of the length to be measured into coincidence 
 with the intersection of the cross-hairs and observing the 
 number of turns and parts of a turn, the length of the object 
 in terms of the pitch of the screw as unit is obtained. The 
 true pitch of the screw must itself be obtained by com- 
 paring it by the same method with some accurately known 
 length, such as a length on a standard meter. 
 
 The adjustment of the microscope consists of two steps: 
 (i) the eye-piece must be focused on the cross-hairs (but 
 the eye-piece must not be taken out lest the cross-hairs be 
 injured) ; (2) the whole microscope must be moved toward 
 or away from the object until it is seen without parallax, i. e., 
 until the relative position of the cross-hairs and the image 
 of the object is not changed by shifting the eye sidewise. 
 The length to be measured must then be placed parallel 
 to the screw. This is attained when, by rotation of the screw, 
 the image of each end can be brought to coincidence with 
 the intersection of the cross-hairs. One of the most fre- 
 quent sources of error, in using a measuring instrument 
 on the screw principle, is back-lash or lost motion. To 
 avoid this the screw should always be turned in the same 
 direction during a measurement. In many dividing 
 engines back-lash is impossible because the motion of the 
 
CATHETOMETER. 19 
 
 screw cannot be reversed, the platform can only be moved 
 in the reverse direction by unclasping the nut (which, for 
 this purpose, is a split nut held together by a clasp and 
 spring) . 
 
 Usually the handle for turning the screw is not attached 
 to the screw or circular scale, but to a separate disk rotating 
 co-axially with the screw. The motion of the handle in 
 one direction is communicated to the screw by a ratchet; 
 when the handle is reversed the ratchet slips freely. As an 
 aid to counting the number of turns of the screw in measuring 
 a considerable length, two detents are sometimes geared to 
 the screw in such a way that only a definite number of turns 
 can be given to the screw at a time, after which the handle 
 must be turned back for beginning a new number of turns. 
 
 A more complete description of the dividing engine will 
 be found in Stewart and Gee, I, 16. 
 
 20. Cathetometer. 
 
 The cathetometer is a vertical pillar supported on a tripod 
 and leveling screws, and capable of rotation about its axis; 
 the pillar is graduated and a horizontal telescope with cross- 
 hairs is borne by a carriage that travels on the pillar and 
 can be clamped at any desired position. A slow-motion 
 screw serves for accurate adjustment of the position of the 
 telescope. 
 
 Adjustments.* (i) The intersection of the cross-hairs, 
 X, must be in the optical axis of the telescope. To secure 
 this, focus X on some mark, rotate the telescope about its 
 own axis and see whether X remains on the mark. If not, 
 the adjusting screws of the cross-hairs must be changed 
 until this is attained. 
 
 (2) The level must be properly adjusted. Level the 
 telescope until the bubble comes to the center of the scale. 
 Turn the level end for end. If the bubble does not come to 
 
 * Adjustments (i) and (2) are not usually required and should not be made 
 without the advice of the instructor. 
 
20 MECHANICS. 
 
 the same position, the level must be adjusted until it will 
 stand this test. 
 
 (3) The scale must be vertical. If there are separate 
 levels for the shaft, this is readily attained. If there is but 
 one level for telescope and shaft, this and the next adjust- 
 ment must be made simultaneously. 
 
 (4) The telescope must be perpendicular to the scale. 
 The top of the scale, T, may be regarded as having two 
 degrees of freedom first, parallel to the line of two level- 
 ing screws of the base, A and B; second, in a line through 
 the third leveling screw, C, perpendicular to AB. If A 
 and B be screwed equal amounts in opposite directions, T 
 will move parallel to A B. If C only be turned, T will 
 move perpendicular to AB. 
 
 First make the telescope horizontal and parallel to AB. 
 Turn the shaft through 180. It is easily seen that if the 
 telescope makes an angle a with the normal to the scale, 
 turning the scale through 180 will cause the telescope to 
 make an angle 2a with its former direction. Hence, with 
 the leveling screw of the telescope, correct half the error 
 in the level, and, by turning A and B equally in opposite 
 directions, correct the remainder. Turn the telescope to 
 the first position and repeat the above adjustments, then to 
 the second and continue as often as is necessary. Then turn 
 the telescope normal to AB and adjust by C. When the 
 adjustment is complete, turning the shaft through any 
 angle will not alter the position of the bubble. 
 
 Unless the cathetometer is on a perfectly immovable sup- 
 port, perfect adjustment is not possible and too much time 
 should not be spent in adjusting, provided the telescope is 
 accurately level at each reading. 
 
 The eye-piece of the telescope is focused (but not removed) 
 until the cross-hairs seem perfectly distinct and the focus 
 of the objective changed until the object is seen ' very 
 distinctly and without parallax, i. e., with no relative 
 motion with respect to the cross-hairs when the eye is 
 moved about. 
 
BAROMETER. 2 1 
 
 21. Barometer. 
 
 Text-book of Physics (Duff}, pp. 157-158; Watson's Physics, pp. 
 150-154; Ames' General Physics, pp. 176-177; Crew's Physics, 
 pp. 165-166. 
 
 If the barometer is of Fortin's cistern form, the cistern is 
 raised or lowered by means of the screw at the bottom 
 until the mercury just meets an ivory stud near the side of 
 the cistern. A collar to which is attached a vernier is so 
 placed that the top of the meniscus of the mercury column 
 is tangent to the plane of the two lower edges. The height 
 of the barometer should be reduced to zero by the formula 
 
 h = h(i . 00016 2t) 
 
 where h is the observed height, h the height at o and /the 
 temperature Centigrade. For the expansion of the mercury 
 will increase the height in the ratio (i . + .000181*) and the 
 expansion of the brass scale will reduce the apparent height 
 in the ratio (i - .000019*) (Table VII). 
 
 The siphon barometer has two scales, graduated on the 
 glass tube, in opposite directions, from a common zero. 
 The length of the mercury column is obviously the sum of 
 the readings of the mercury levels in the two tubes. Since 
 the coefficient of linear expansion of glass is only about 
 0.000008, the correction formula becomes 
 h = h(i- .000173*). 
 
 Since the mercury may adhere to the glass to some extent, 
 barometer tubes should be tapped gently before reading. 
 
 22. The Balance. 
 
 Kohlrausch, 7-11; Watson's Practical Physics, 25, 26. 
 Weighing by a Sensitive Balance. On first using a sensitive 
 balance note the position, purpose, and structure of the 
 following parts: 
 
 The beam, The knife-edges and planes, 
 
 The pointer, The arrestment, 
 
 The pillar, The rider-arms, 
 
 The pan-supports. 
 
22 MECHANICS. 
 
 By the sensibility of a balance is meant the amount of de- 
 flection of the beam produced by a given small weight. 
 Consider how the sensibility depends on (i) the length of 
 the beam, (2) the weight of the beam and pans, (3) the 
 distance of the point of suspension of the beam from its 
 center of gravity. Will the sensibility of a certain balance 
 be different with different loads on the pans and why? 
 How can the sensibility be varied with a given load? (See 
 references.) 
 
 Precautions in Use of Balance. 
 
 1. Note the maximum load that may be placed on the 
 balance and take care not to exceed it. 
 
 2. Always lift the beam from the knife e<Jges before in 
 any way altering the load on the pans. 
 
 3. Do not stop the swinging of the balance with a jerk. 
 It is best to stop it when the pointer is vertical. 
 
 4. To set the beam in vibration, do not touch it with the 
 hand, but taise and lower the arrestment. 
 
 5. Place the large weights in the center of the pan. 
 
 6. Make final weighings with the case closed. 
 
 7. Replace all weights in their proper place in the box 
 when they are not actually in use. Do not use weights 
 from different boxes. 
 
 8. Do not place anything in contact with a pan that is 
 liable to injure it. 
 
 9. Avoid, if possible, weighing a hot body. 
 
 10. Never handle the weights with the fingers, as this 
 may change some of the weights appreciably. Always use 
 the pincers. 
 
 Notice the dimensions of the weights in the box, e. g., 
 50 g., 20 g., 10 g., 10 g., 5 g., 2 g., i g., i g., etc. Instead of 
 weights of 0.005 > - 2 g-> o.ooi g., o.ooi g., it is customary 
 to use a rider of o . 01 g., which can be placed on the beam at 
 various distances from the center. The beam is for this pur- 
 pose graduated into 10 divisions, which may be still further 
 subdivided. Thus the o.oio g. rider placed at the division 
 4 of the beam is equivalent to o . 004 g. placed on the pan. 
 
THE BALANCE. 23 
 
 The zero-point of the balance is the position on the scale 
 behind the pointer at which, the pans being empty, the 
 pointer would ultimately come to rest; it must not be con- 
 fused with the zero of the scale. As much time would be 
 wasted in always waiting for the pointer to come to rest, 
 the zero of the balance is best obtained from the swings of 
 the pointer. For this purpose, readings of the successive 
 "turning-points" are made as follows three successive 
 turning-points on the right and the two intermediate ones 
 on the left, or vice versa; e. g., 
 
 Turning points. 
 L. R. 
 
 -1-3 + 2-1 
 
 I.I -f 2. O 
 
 I .O 
 
 Mean, 1.13 +2.05 
 
 -1.13 
 
 Zero-point = +0.92^-2 = +0.46. 
 
 By taking an odd number of successive turning-points 
 on one side and the intermediate even number on the other 
 side and then averaging each set, we eliminate the effect of 
 the gf adual decrease of amplitude of the swing. 
 
 The resting-point of the balance with any loads on the 
 pans is the point at which the pointer would ultimately 
 come to rest, and is found in the same way as the zero-point. 
 If the resting-point should happen to be the same as the 
 zero-point, the weight of the body on 
 one pan is immediately found by the ~ 
 weights on the other pan and the posi- 
 tion of the rider. Usually, however, this 
 
 will not be so. With the rider at a suitable division, find the 
 resting-point on one side of the zero-point, and then, after 
 .altering the rider one place, find the resting-point on the 
 other side of the zero. By interpolation the change of the 
 position of the rider necessary to make the resting-point 
 coincide with the zero-point is deduced. For example, the 
 
24 MECHANICS. 
 
 zero is +o . 46 ; with the rider at 4 the resting-point is +0.51; 
 with the rider at 5 the resting-point is + o . 10. By changing 
 the rider from 4 to 5, o.ooi g. was added. To bring the 
 resting-point to the zero we should have added 0.05-;- 
 (0.510.10) of o.ooi g. or o.oooi g. approximately. 
 Hence the weight of the body is the weight on the pan plus 
 0.0041 g. 
 
 The arms of the balance may be unequal. If this be so, 
 the weight obtained above will not be the true weight. To 
 eliminate this error the body must be changed to the other 
 pan and another weighing made. If / be the length of the 
 left arm and r that of the right and if u be the counterbal- 
 ancing weight when the body is in the left pan and v when 
 it is in the right, while w is the true weight of the body then, 
 lw=ru, lv=rw 
 
 (The geometric mean of two very nearly equal quantities 
 is nearly equal to their arithmetic mean.) The ratio of the 
 arms of the balance may also be calculated, since 
 
 The buoyancy of the air on the weights and on the body 
 must be allowed for in accurate work. To the apparent 
 weight of the body must be added a correction equal to the 
 weight of the air displaced by the body and from the apparent 
 weight must be subtracted the weight of the air displaced 
 by the weights. In each case the weight of the air displaced 
 can be calculated if its volume and density are known. 
 This correction in any case is very small. A small per- 
 centage error in the correction will not appreciably affect 
 the calculated true weight. Hence approximate values of 
 the volumes of the body and weights may be used. In 
 finding the volume of the weights the density of brass 
 weights may be taken as 8.4. The density of air at o and 
 760 mm. may be taken as .0013, and its density at the 
 temperature of the laboratory and the pressure indicated 
 
ADJUSTMENT OF TELESCOPE AND SCALE. 25 
 
 by the barometer may be calculated by the laws of gases. 
 Hence the temperature and barometric pressure should be 
 obtained. 
 
 23. Adjustment of Telescope and Scale. 
 
 To adjust a telescope and scale, determine approximately 
 the location of the normal to the mirror, either by finding 
 the image of one eye or the image of an incandescent lamp 
 held near the eye. Move the stand supporting the telescope 
 and scale until the center of the scale is about in line with 
 the normal. Look along the outside of the telescope at the 
 mirror and move the scale up and down, or, if this is not 
 possible, raise or lower the stand until you see the reflection 
 of the scale in the mirror. It may be a help to illuminate 
 the scale with an incandescent lamp. Look through the 
 telescope pointed at the mirror, and change the focus until 
 the scale is seen distinctly. Remember that the more 
 distant the object, the more the eye-piece must 'be pushed in, 
 and that the image of the scale is at about twice the distance 
 of the mirror. 
 
 24. Time Signals. 
 
 A convenient source of time signals for a laboratory is a 
 chronometer which either opens or closes a circuit containing 
 batteries, sounders, etc., every second with an omission 
 at the end of each minute. 
 
 The individual second intervals indicated by a chro- 
 nometer, so arranged, are likely to be somewhat inaccurate, 
 and therefore, when an accurate interval of one second is 
 required, a second's pendulum should be used with a plat- 
 inum point making contact with a drop of mercury, and thus, 
 if desired, closing an electric circuit. Since it is difficult 
 to set the mercury drop exactly in the center of the path, 
 alternate seconds are likely to be too long. Therefore, if 
 possible, a two seconds' interval should be employed, 
 alternate contacts being disregarded. If these contacts 
 cause confusion, a pendulum omitting alternate contacts 
 may be used (see Ames and Bliss, p. 486). 
 
I. TO MAKE AND CALIBRATE A SCALE. 
 
 To illustrate the use of the dividing engine (described on 
 page 17) a short scale is to be engraved in millimeters on a 
 strip of nickel-plated steel and then calibrated by compari- 
 son with the average millimeter of a standard scale. 
 
 Arrange the cogs of the dividing gear so that each fifth 
 mm. division shall be longer than the intermediate divisions 
 and each tenth division still longer. Test this adjustment 
 on a rough test strip. Next clamp the strip to be divided 
 on the platform of the engine so that it is parallel to the 
 screw; this can be tested by observing the edge of the 
 strip in the microscope as the platform is advanced by the 
 screw. Care should be taken to clamp the pillar that sup- 
 ports the divider so that the point of the divider moves 
 perpendicular to the length of the scale. A scale of 2 
 or 3 cms. should then be marked out on the steel strip 
 and the temperature of the platform ascertained by a 
 thermometer. 
 
 This scale is next to be calibrated. The exact pitch of the 
 screw is first obtained in terms of the mm. of the standard. 
 For this purpose a considerable length, e. g., a decimeter, 
 of the standard should be measured on the engine. This 
 should be done for three different parts of the screw. The 
 agreement of the three determinations will afford some 
 indication of the uniformity of the screw. The scale should 
 then be measured mm. by mm. For the first reading the 
 circular scale of the screw may be set to zero when the 
 cross-hairs coincide with the zero division of the scale to 
 be measured, and thereafter the screw should be turned 
 always in the same direction and only arrested for a reading 
 of the circular scale and vernier (the total number of turns 
 being also noted) when the microscope shows that the middle 
 of a division has come to coincide with the intersection of 
 
 26 
 
ERRORS OF WEIGHTS. 27 
 
 the cross-hairs. As this coincidence approaches, the handle 
 should be turned slowly, and if turned too far the reading 
 at that point must be omitted altogether. The handle 
 should also be turned slowly when contact with the detent 
 approaches so that the screw may not be arrested with a 
 jerk. 
 
 As a check on the work, the whole length of the scale 
 should be measured. 
 
 In calculating the true length of the divisions, allowance 
 must be made for the temperature of the standard which 
 may be taken as the temperature of the platform of the 
 engine. The standard is correct at the temperature marked. 
 From its coefficient of expansion calculate the length of its 
 mm. at the temperature of observation and then deduce the 
 pitch of the screw at the same temperature. Then from the 
 readings made, calculate the length of each millimeter of the 
 scale and, by addition, draw up a table showing the true 
 distance of each division from the zero division. 
 
 Questions. 
 
 1 . Enumerate the possible sources of error in the use of the divid- 
 ing engine for the manufacture of scales. 
 
 2. At what temperature would the whole length of your scale be 
 an exact number of centimeters? (Table VII.) 
 
 II. ERRORS OF WEIGHTS. 
 
 Kohlrausch, 12; Watson's Practical Physics, 27. 
 
 Weights by good makers are usually so accurate that 
 errors in them may for most purposes be neglected. But 
 when less perfect weights are to be used or when weighings 
 are to be made with the highest possible degree of accuracy, 
 the errors in the weights must be carefully ascertained. 
 
 We shall suppose that a 100 g. box of weights is to be 
 tested, and that a reliable 100 g. weight is supplied as a 
 standard, and that an accurate 10 mg. rider is supplied for 
 making the weighings. The weights of the box will be 
 denoted by 100', 50', 20', 20", 10', and so on, and the 
 
28 MECHANICS. 
 
 sum 5' + 2' + 2" + 1' by 10". To find the six unknown quan- 
 tities, 100', 50', 20', 20", io r , 10", we must make six weigh- 
 ings and obtain six relations between these quantities. 
 Such a set of weighings are indicated in the following table. 
 Each should be performed by the method of double-weigh- 
 ing described on page 24. 
 
 10' =10" +a 
 
 20' =10' + 10" +b 
 
 20" =20' +c 
 
 50' =20' +20" +10' +d 
 ioc/ =50' + 20' +20" + 10' +e 
 100 = 100' +/ 
 
 To solve these equations, substitute the value of 10' given 
 by the first in the second; then substitute the value of 20' 
 given by the second in the third, and so on to the last, when 
 the value of 10" in terms of the standard 100 and a, b, c, d, 
 e, f will be obtained. The calculation of the other quantities 
 will then present no difficulty. To standardize the box 
 completely the same process must be applied to 10', 5', 2', 2" ', 
 i', i", and similarly to the smaller weights. 
 
 III. VOLUME, MASS, AND DENSITY OF A REGULAR 
 
 SOLID. 
 
 The mass of the specimen (a sphere or cylinder) is found 
 by weighing on a sensitive balance (see p. 21). To eliminate 
 the inequality of the arms of the balance, the body should be 
 weighed in both pans (p. 24). The zero-point and resting- 
 points of the balance should be found by the method of 
 vibrations and the various precautions in the use of the 
 balance must be carefully observed. Allowance should be 
 made for air buoyancy (p. 24) and corrections should be 
 applied to the weights, if the weights have been corrected in 
 the preceding experiment, or if a table of corrections is 
 supplied. 
 
MOHR-WESTPHAL SPECIFIC GRAVITY BALANCE. 29 
 
 The dimensions of the specimen are measured by a 
 micrometer caliper (p. 14) or a vernier caliper (p. 14). 
 If the body is spherical, ten measurements of the diameter 
 should be made and the average taken; if it is cylindrical 
 ten measurements of the diameter and ten of the length 
 should be made. 
 
 From the mass and the volume, the density (or mass per 
 c.c.) is deduced. 
 
 The ratio of the arms of the balance should also be derived 
 from the results of the double weighing (p. 24). 
 
 The possible error of the density determination should be 
 calculated as illustrated on p. 6. 
 
 Questions. 
 
 1. If the object aimed at were merely the density of the body, 
 which of the above measurements should be improved in precision 
 and to what extent would it need to be improved? 
 
 2. If the above improvement were not possible, how much of the 
 refinement of measurement of the other quantity might be dis- 
 carded ? 
 
 IV. MOHR-WESTPHAL SPECIFIC GRAVITY 
 BALANCE. 
 
 Kohlrausch, p. 45; Stewart and Gee, I, 92, III. 
 
 This is a convenient form of hydrostatic balance for 
 rinding the density of a liquid by determining the buoyancy 
 of the liquid on a float hung from an arm of the balance 
 and immersed in the liquid. Instead of weights riders are 
 used, the arm of the balance from which the float hangs 
 being graduated into ten divisions. The float is made of 
 such a size that when hanging in air from the graduated 
 arm of the balance (which is less massive than the other 
 arm) it will just produce equilibrium. Four riders of differ- 
 ent mass are employed, each one being ten times as heavy 
 as the next smaller. The largest rider is of such a size 
 that if the float hanging from the balance be immersed in 
 water at 15 C. the addition of the rider to the hook at the 
 
30 MECHANICS. 
 
 end of the beam will restore equilibrium. Hence it counter- 
 balances the buoyancy of the water on the float. Thus it is 
 evident that if the water be replaced by a liquid of unknown 
 density at the same temperature (so that the volume of the 
 float is the same) and if the largest rider under the circum- 
 stances produces equilibrium when placed at the sixth 
 division, then for equal volumes, this liquid can weigh 
 only six-tenths as much as water, or its density is 0.6. A 
 second rider, one-tenth as heavy as the first, would evidently 
 
 FIG. 7. 
 
 enable us to carry the process one decimal place farther, 
 etc. For liquids of a density exceeding unity, another 
 rider equal to the largest must be hung from the end of the 
 beam, and still a third may be necessary for liquids of den- 
 sity above 2. 
 
 From the above it will be seen that (i) the balance must 
 be adjusted by the leveling screw on the base until the end 
 .of the beam is opposite the stud in the framework when the 
 float is suspended in the air; (2) the beaker must always 
 be filled to the same level, that level being such that when 
 the liquid is water at 15 C. the balance is in equilibrium 
 with the largest rider hanging above the float, and (3) the 
 liquid tested must be at 15 C. 
 
MOHR-WESTPHAL SPECIFIC GRAVITY BALANCE. 31 
 
 As an exercise in the use of this balance, find what shrink- 
 age of volume there is in the solution of some salt (e. g., 
 common salt, ammonium chloride or copper sulphate) 
 in water and find how the shrinkage varies with the con- 
 centration. Solutions may be made up by weighing out 
 very carefully on a sensitive balance (see p. 21), 0.5 gm., 
 i gm., 4 gm., 10 gm., etc., of the salt and dissolving each in 
 a deciliter of water. When the density of a solution has 
 been found, the percentage contraction is calculated from 
 the sum of the volumes of the constituents before mixture 
 and the volume of the solution after mixture; the volume 
 in each case equals the mass divided by the density. The 
 densities of various salts are given in Table VII. 
 
 The densities found and the percentages of contraction 
 should be represented by curves with percentages of salt 
 as abscissae. If any determination of density be largely in 
 error it will be shown by the curve. 
 
 If time permit, determine the density at 1 5 of equivalent 
 solutions* of several salts having the same base, e. g., 
 NaCl; 1/2 Na 2 SO 4 ; NaNO 3 ; etc., and compare with the densi- 
 ties of similar solutions with a different base, e. g., NH 4 C1; 
 1/2 (NH 4 ) 2 SO 4 , NH 4 NO 3 , etc. The difference in density 
 between corresponding salts should be approximately 
 constant (Valson's Law of Moduli). Find similarly the 
 difference in densities contributed by the acid radicals, 
 e. g., NaN0 3 and NaCl; NH 4 N0 3 and NH 4 C1, etc. 
 
 Questions. 
 
 1 . What sources of error may there be in a determination of density 
 by this method ? 
 
 2. How might the accuracy of the riders be tested? 
 
 3 . How might the accuracy of graduation of the beam be tested ? 
 
 4. What effect has capillarity? 
 
 5. Explain the Law of Moduli. f 
 
 * The chemical equivalent of a substance is the atomic or molecular weight 
 divided by the valency. Two solutions are equivalent if the number of grams 
 of each dissolved in one liter (or that proportion) is the same fraction of the 
 respective chemical equivalent. 
 
 t Phy. Chem., Ewell, p. 159. 
 
3 2 
 
 MECHANICS. 
 
 V. DENSITY BY VOLUMENOMETER. 
 
 Gray's Treatise on Physics, I, 426. 
 
 When the density of such substances as gunpowder, 
 sugar, starch, etc., is to be determined, neither the method 
 of immersion in a liquid nor that of the direct measure- 
 ment of mass and volume can be employed. The method 
 then usually employed is that of the volumenometer. This 
 is a method of immersion in air instead of immersion in 
 water, with an application of Boyle's Law 
 instead of Archimedes' principle. The volume 
 of the body is found by placing it in a glass 
 vessel and noting how much the pressure in 
 the vessel changes when the air is allowed to 
 expand. 
 
 A gas-washing bottle of about 150 c.c. capa- 
 city, A, into which the body is to be introduced, 
 is connected, by heavy pressure tubing, with 
 an open, U-shaped, mercury manometer (see 
 Fig. 8). The bottle, A, is closed by the stopper 
 6, which should be lubricated with rubber 
 grease,* and forced into A to a definite mark. 
 DE is raised until the mercury in the burette BC is at a 
 division B which is carefully observed. The pressure, 
 P, in A is carefully determined from the difference in 
 mercury levels and the barometer. By the use of a 
 rear mirror, parallax may be avoided and a small square 
 will assist in reading a scale between the two arms of the 
 manometer. The accuracy of the readings may be increased 
 by using a cathetometer (p. 19). Lower DE until the mer- 
 cury is at a division K and again determine the pressure, p. 
 Let the volume between B and K be v. Let V be the 
 volume of A, and connecting tubing, to B. By Boyle's 
 Law: 
 
 FIG. 8. 
 
 * Equal parts pure rubber gum, vaseline, and paraffin. The two latter are 
 melted together and the rubber is cut into small pieces and dissolved in the 
 heated liquid, 
 
DENSITY OF AIR. 33 
 
 Make at least six determinations of P and p, bringing 
 the mercury each time to the same points B and K which 
 should be as far apart as is convenient. Calculate V from 
 the mean values. 
 
 Now introduce a carefully weighed amount of the assigned 
 powder into the bottle, A (which may be disconnected at 
 e), and insert the stopper, b, to its former depth. Again 
 determine the pressures, P' and p r when the mercury level 
 is at B and K respectively. Repeat as before. If x is the 
 unknown volume of the powder, the previous equation 
 becomes 
 
 from which x may be calculated. From the volume and 
 mass of the powder its density is determined. 
 
 If time permit, determine the density also with a specific 
 gravity flask (pyknometer) . Weighings should be made 
 of (i) bottle empty; (2) bottle filled with a liquid of known 
 density which is inert toward the body, and (3) with a 
 known mass of the body in it, the rest of the bottle being 
 filled with the liquid. An equation for density can be 
 worked out. 
 
 The possible error in the determination of the density 
 is found by methods explained on pages 3,8. 
 
 Questions. 
 
 1. What sources of error remain uneliminated? 
 
 2. With a view to greater accuracy what suggestions would you 
 make as to the most suitable magnitudes for x and v? 
 
 VI. DENSITY OF AIR. 
 
 The density of air at atmospheric pressure, or its mass 
 per cubic centimeter, might be obtained by weighing a flask 
 containing air at atmospheric pressure and then re-weigh- 
 ing it after all the air has been removed by an air-pump. 
 The difference of weight, together with the volume of the 
 flask, would give the density of the air. In practice the 
 3 
 
34 MECHANICS. 
 
 procedure has to be modified, because it is impossible to 
 completely exhaust the flask of air. The modification con- 
 sists in finding the pressure of the air remaining in the flask 
 and taking account of it. 
 
 Let D be the required density at the room temperature 
 and pressure, P. Let d be the density of the remaining air 
 when the pressure has been reduced to p. Let the weight 
 of the flask when filled with air be W and let w represent 
 its weight when exhausted to the pressure p. 
 
 W-w = V(D-d) 
 By Boyle's Law 
 
 D _P D-d_P-p 
 
 ~d = ~p " ~D~ P 
 Therefore 
 
 'P-p V P-p 
 
 A convenient form of flask is a round-bottom flask from 
 which part of the neck has been cut off and which is closed 
 by a rubber stopper containing a glass tube with a glass 
 stop-cock. The rubber stopper will hold tighter if lubricated 
 with rubber grease * before insertion. 
 
 If the flask, as found, is dry, it will be better to postpone 
 finding its volume until the end of the experiment, as the 
 operation requires it to be filled with water. Moreover, 
 of the two weighings for finding the mass of air removed, 
 it is better to make the one with the flask partly exhausted 
 first, for the weighing with the air admitted can be made 
 immediately after, without handling the flask or removing it 
 from the balance, a point of some importance where the 
 difference of weight to be measured is so small. To save 
 delay in weighing the flask after it has been exhausted, the 
 zero reading of the fine balance used should be obtained 
 before the flask is exhausted. For the method of accurate 
 weighing, by oscillations, see page 23. 
 
 *See note, p. 32. 
 
DENSITY OF AIR. 35 
 
 A Bunsen's aspirator or a Geryk pump is satisfactory 
 for exhausting the flask. The flask should be connected 
 to the aspirator or pump, through a bottle for catching any 
 water or mercury. An open-tube manometer connected 
 to the tube that joins the aspirator or pump and flask will 
 give the pressure. 
 
 There should be a stop-cock or a rubber pinch-cock in 
 the connection between the manometer and the pump or 
 aspirator. When a sufficiently high exhaustion has been 
 secured this cock should be closed for several minutes to 
 ascertain if there is any leakage. If not, both ends of the 
 manometer should be read and the stop-cock of the flask 
 closed. Before removal of the flask, the other cock should 
 be opened that the. rest of the apparatus may fill with air. 
 If by any chance a small quantity of water should pass into 
 the manometer, allowance should be made for it, the density 
 of mercury being taken as 13.6. 
 
 The flask is then weighed as quickly as possible on a 
 fine balance, the method of vibration being used. It may 
 be necessary to hang the flask by a fine wire to the hook 
 which carries the pan. This weighing is repeated with the 
 stop-cock open, but with the flask otherwise undisturbed. 
 The atmospheric pressure is obtained from a reading of the 
 barometer (see p. 21). 
 
 The volume of the flask may be obtained by filling it 
 with distilled water and weighing it on an open balance. 
 To get the flask just filled to the stop-cock, the stopper 
 (removed for filling the flask) should be thrust in with the 
 stop-cock open, the stop-cock should then be closed, and 
 any water above the stop-cock should be removed. Of 
 course, the stop-cock should be replaced at its original 
 depth, which should be marked. The density of water at 
 different temperatures will be found in Table V. 
 
 When the experiment is completed, place the open flask 
 inverted on a frame to dry, so that it may be ready for the 
 next person who uses it. 
 
 The density of dry air may be found in the same way, 
 
36 MECHANICS. 
 
 the flask being several times exhausted and refilled through 
 a drying-tube. Similarly the density of any other gas, e. g., 
 carbon dioxide, may be found by filling the flask from a 
 generator. The gas must be admitted to the exhausted 
 flask very slowly and the exhaustion and filling must be 
 repeated to insure the (almost) complete removal of 
 the air. 
 
 In reporting, deduce from your measurement of the 
 density of air or gas, its density at o C. and 760 mm. by 
 using Boyle's and Charles' Laws. Find also the possible 
 error of the measurement of density (p. 5). 
 
 Questions. 
 
 1 . Would the first results be affected by the presence of water in 
 the flask? Explain. 
 
 2 . Should the flask weigh more filled with dry air or filled with 
 moist air, both at atmospheric pressure? Why? 
 
 VII. ACCELERATION OF GRAVITY BY PENDULUM. 
 
 Text-book of Physics (Duff}, 117; Watson's Physics, 112114; 
 Watson's Practical Physics, 46-49; Ames' General Physics, 
 pp. 74, 91, 135; Crew's Physics, 85, 86. 
 
 The acceleration of gravity, g, is most readily obtained 
 from the length and time of vibration of a pendulum. The 
 time of vibration of an ideal simple pendulum, i. e., a heavy 
 particle vibrating at the end of a massless cord would be 
 
 ,- T 
 
 / being the length of the pendulum. If the bob is a ball so 
 large that the mass of the suspending wire is negligible, 
 the above formula will apply provided the radius of the 
 ball is negligible compared with the length of the pendulum. 
 If these assumptions may not be made, the pendulum must 
 be regarded as a physical pendulum and its moment of 
 
ACCELERATION OF GRAVITY BY PENDULUM. 37 
 
 inertia about the suspension considered. Under these 
 circumstances the formula 
 
 t = 
 
 Mgh 
 
 must be used, where / is the moment of inertia of the entire 
 pendulum about the knife-edge, M is the total mass and h 
 is the distance from the knife-edge to the center of gravity 
 of the whole. If the mass of the suspension is negligible 
 it is only necessary to consider the moment of inertia 
 of the ball about the knife-edge. It is easily shown that 
 the latter formula then reduces to the formula for the 
 simple pendulum, provided the length of the pendulum is 
 taken as the distance from the knife-edge to the center of 
 the ball plus 2r 2 / $1 where r is the radius of the ball. Hence 
 to find g there are three quantities, t, I, and r, to be 
 measured. 
 
 A convenient form of pendulum consists of a spherical 
 bob into which screws a nipple through which a fine wire 
 is passed and secured. To the upper end of the wire is 
 soldered a stirrup of brass which rests on a knife-edge of 
 steel. A short platinum wire should be soldered to the 
 lower side of the bob. 
 
 For accurately measuring the length of the pendulum a 
 cathetometer (see p. 19), which should be carefully adjusted, 
 may be used. (If necessary, the measurement of length may 
 be postponed until the time has been observed). The 
 horizontal cross-hair of the cathetometer is first focused 
 on the knife-edge, the fine screw being used for the final 
 adjustment of the telescope, and the scale and vernier are 
 then read. The telescope is then lowered and set on either 
 the top or bottom of the bob, whichever is the more definite. 
 These readings should be repeated several times, beginning 
 each time with the knife-edge. If the adjustments are 
 imperfect, the telescope should at least be made exactly 
 level before each reading. The diameter of the bob may be 
 
38 MECHANICS. 
 
 measured by means of a micrometer or a vernier caliper 
 (see p. 14). 
 
 For fixing the vertical position of the pendulum, two 
 vertical pointers may be so placed that, when the pendulum 
 is at rest, the pendulum suspension and two pointers are 
 in one plane. The eye of the observer should always be 
 kept in this plane in using the first two methods. The pen- 
 dulum is set vibrating in an arc of 3 or 4 cms. Several 
 attempts may be necessary to get the pendulum vibrating 
 exactly perpendicular to the knife-edge with the bob free 
 from rotation. 
 
 The time of vibration is most readily obtained with 
 precision when the pendulum is very nearly a second's 
 pendulum, i.e., when the period of a complete vibration is 
 very nearly two seconds. For the determination of the 
 period several methods are available. The first and roughest 
 method given below will serve for adjusting the pendulum 
 to the required length. 
 
 (A) In the first method for determining the period, 
 time is found by the relay (p. 25) and the number of 
 vibrations in three minutes is counted, fractions of a vibra- 
 tion being roughly estimated. This is repeated several 
 times. Or a stop-watch or stop-clock may be used, but it 
 should be rated by comparison with a chronometer or 
 standard clock. The stop-watch is started as the pendulum 
 crosses the plane of observation and "one" is counted the 
 next time the pendulum crosses the plane in the same 
 direction. The watch is stopped on the 5oth vibration; 
 and the whole repeated five times. The mean time divided 
 by 50 will give a fair value for the period. 
 
 (B) A second and much more accurate method of obtain- 
 ing the time of vibration is the method of coincidences. This 
 consists in finding the rate at which the pendulum gains or 
 loses as compared with a standard clock or chronometer. 
 It is applicable only when the periods of pendulum and clock 
 or chronometer are nearly the same or when one is nearly 
 an exact multiple of the other. The method receives its 
 
ACCELERATION OF GRAVITY BY PENDULUM. 39 
 
 name from the fact that what is observed is the "coincidence 
 interval" or the interval between the moment when a 
 passage of the pendulum through the vertical coincides 
 with some signal from the clock to the next time when such a 
 coincidence occurs. 
 
 In a coincidence interval, the pendulum must gain or 
 lose one vibration as compared with the chronometer or 
 other time standard. If n such coincidence intervals occur 
 in T sec., the number of vibrations of the pendulum during 
 this time is (Tri). Hence if / is the time of one vibration, 
 
 Tn 
 and the period of a complete vibration is 
 
 27 
 
 Tn 
 
 A convenient form of signal is given by the chronometer 
 and relay described on page 25. It is advisable to have the 
 coincidence interval something between 30 seconds and 3 
 minutes, and, if necessary, the length of the pendulum should 
 be changed for the purpose. 
 
 After the coincidence interval has been roughly deter- 
 mined by a few observations, the following modification of 
 the method will give it much more accurately. Calling the 
 time of the first coincidence zero seconds, observe the sec- 
 ond on which the next coincidence occurs and then the next, 
 until four have been observed. Then, after allowing a 
 considerable number of coincidences to pass unnoted, but 
 keeping note of the time, observe the number of the seconds, 
 counted from the original coincidence, upon which four more 
 successive coincidences occur. 
 
 From the first set of coincidences, three estimates of the 
 coincidence interval will be obtained and three others from 
 the second set, the mean of all giving an approximate esti- 
 mate. Then let the time of the first coincidence of the first 
 set be subtracted from the time of the first of the second 
 
40 MECHANICS. 
 
 set, also the time of the second coincidence of the first set 
 from that of the second of the second set, etc. These dif- 
 ferences give four estimates of the time, T; of some unknown 
 integral number, n, of coincidence intervals. If the mean 
 of these four estimates be divided by the mean time of a 
 single coincidence interval as already found, the quotient 
 will be n plus or minus a small fraction. This fraction is 
 due to inaccuracy in the estimates of the coincidence intervals 
 and should be dropped. The period / of the pendulum may 
 now be calculated. The plus sign in the denominator is 
 used if the pendulum is the faster. 
 
 The following aid to the observation of coincidences is 
 suggested. Keeping the eye constantly in the proper plane 
 for observation, make a dot on a piece of paper at each click 
 of the- relay. When there appears to be coincidence, pro- 
 long the dot into a stroke. To avoid recording every click, 
 a cross may be used instead of a dot for marking a minute, 
 and the clicks may be passed unrecorded until the next 
 minute, or coincidence. There may be several successive 
 clicks during which there appear to be coincidences, in 
 which case several successive strokes should be made and 
 the mean taken. From these dots, strokes, and crosses, the 
 times of coincidence may be deduced. Or, a dial indicating 
 seconds may be employed, the second when there first 
 appears to be a coincidence being observed and the second 
 when there first appears to be no coincidence. Since the 
 clock cannot be observed immediately, the ticks are counted 
 until the clock is observed and then subtracted ; minutes must 
 be noted and recorded if they are not recorded on the clock. 
 
 (Q A third method consists in modifying the second 
 method so that coincidences of two sounds are observed. 
 The pendulum is made to actuate a sounder or telephone 
 each time it passes through the vertical and a coincidence is 
 observed when the sounder and relay strike together. A 
 block of wood with a narrow trough filled full of mercury 
 is placed in a mercury tray and is adjusted beneath the 
 pendulum so that the platinum wire on the under side of the 
 
COEFFICIENT OF FRICTION. 41 
 
 bob just touches the mercury when the pendulum is at rest, 
 and crosses the narrow trough at right angles when the 
 pendulum is in motion,. A wire soldered to the knife-edge 
 is connected in series with several batteries, a sounder or 
 telephone, and the mercury trough. The final adjustment 
 of the mercury trough is made with the leveling screws of 
 the mercury tray. Care should be taken not to spill the 
 mercury. 
 
 From the possible errors in the measurements of / and t 
 deduce the possible errors in the value found for g 
 (see p. 7). 
 
 Questions. 
 
 1. Does the friction of the knife-edges and of the air increase or 
 decrease the value of g? 
 
 2. Why should coincidence be observed exactly, for the plane 
 containing the position of rest? 
 
 3. What would be the result of increasing the arc of vibration to 
 10 cm.? (Table III.) 
 
 4. Why should the top reading of the cathetometer always precede 
 the bottom reading? 
 
 5. Design, if possible, a scheme of electrical connections such 
 that the sounder will only operate when there is a coincidence. 
 
 VIII. COEFFICIENT OF FRICTION. 
 
 Text-book of Physics (Duff), 126130; Watson's Physics, 96-100; 
 Ames 1 General Physics, p. 118; Crew's Physics, 117; DanielVs 
 Physics, pp. 176-184. 
 
 The coefficient of friction of two surfaces is the ratio 
 of the force of friction opposing the incipient or actual 
 relative motion to the force pressing the two surfaces to- 
 gether. The force requisite to start the motion is greater 
 than that required to sustain the motion, i. e., the "coeffi- 
 cient of static friction" is greater than that of "kinetic fric- 
 tion." Moreover, the coefficient of kinetic friction is not 
 quite constant, but varies somewhat with the speed. 
 
 (A) The coefficient of static friction of one surface on 
 another may be found by means of a block of the former 
 resting on a slide of the latter. One end of the slide is gently 
 
42 MECHANICS. 
 
 elevated by a screw until the block just fails to stand sta- 
 tionary on the slide. The tangent of the angle which the 
 slide then makes with the horizontal equals the coefficient 
 of static friction (see references). The tangent may be 
 measured by some simple method, using meter-stick, plumb- 
 line and level or square. Several entirely independent 
 adjustments for this angle and measurements of the tan- 
 gent should be made, the adjusting screw being each time 
 turned some distance down so that the influence of the pre- 
 vious setting may be avoided. The friction may vary some- 
 what from point to point, and if so, different points should 
 be chosen for the separate trials. 
 
 The accuracy of the determination of the tangent should 
 be calculated to see whether the possible errors will account 
 for the variations of the coefficient. Such, however, will 
 probably not be found to be the case. 
 
 (B) The coefficient of kinetic friction may be determined 
 by the same apparatus if we can find the acceleration 
 with which the block moves down the slide when the latter 
 is tilted beyond the angle of repose. For, if the acceleration 
 of the block is a and its mass m and the angle of inclination 
 of the slide i, then the component of gravity down the slide 
 is mg sin i and the pressure on the slide is mg cos i. Hence, 
 if JJL is the coefficient of friction, by Newton's second law, 
 
 m a m g sin i t a m g cos i 
 
 a 
 
 and, = tan ^ ;. 
 
 g cos i 
 
 This process will give the mean coefficient of friction for the 
 range of speeds through which the block passes, but for the 
 low speeds in question the coefficient does not vary much. 
 
 The acceleration, a, is found by a method frequently 
 employed in physical measurements. A tuning-fork (fre- 
 quency of 50 or less) is fastened in a clamp attached to a 
 support above the slide. A stylus of spring brass with a 
 steel needle point is attached to one prong and just behind 
 this stylus is a second stationary stylus which is attached to 
 
COEFFICIENT OF FRICTION. 43 
 
 the support. A long and narrow glass plate is covered with 
 the washing compound called "Bon Ami" by transferring, 
 with a wet cloth, a little of the paste from the cake to the glass, 
 and then spreading it out in a thin layer. The block is 
 then raised to the top of the slide and secured by a trigger. 
 The support that carries the fork is raised and lowered and 
 the fork is adjusted in the clamp until each stylus touches 
 the coated glass, making with it an angle of about 45, the 
 stylus on the fork being exactly in front of the other stylus. 
 
 The frame-work is then lifted until neither stylus touches 
 the glass. The fork is set in vibration by drawing the prongs 
 together with the fingers and releasing them, or by with- 
 drawing a wooden wedge, and is then adjusted until its 
 stylus vibrates an equal amount on each side of the other 
 (stationary) stylus. The frame-work is then lowered until 
 the styli touch the glass and the block is immediately re- 
 leased by the trigger. 
 
 A wave line should be obtained with a straight line ex- 
 actly in the center, the amplitude of the wave line on each 
 side of the straight line being several millimeters. Since 
 in any measurement the effect of inaccuracies at the ends is 
 less important the greater the quantity measured, we meas- 
 ure the distance passed over during several vibrations of 
 the fork. This distance divided by the time in which it 
 was traversed, i. e., by the period of the fork multiplied by 
 the number of vibrations, gives the average velocity of the 
 block during this time. If T be the period of the fork and 
 x the distance passed over in n complete vibrations of the 
 fork, the average velocity is oc/nT. Similarly we find 
 the average velocity for the next n complete vibrations. 
 The average acceleration will be the difference between 
 these average velocities divided by their separation in time 
 or nT; for since each velocity is the average we may con- 
 sider it as belonging to the middle of the time for which 
 it is the average. From several successive groups of n 
 vibrations several values of the acceleration are obtained 
 and the mean taken. 
 
44 MECHANICS. 
 
 It remains to determine the period of the fork. Two 
 methods will be described, (a) The fork is clamped be- 
 side a small electro-magnet connected through a battery 
 with a pendulum which closes the. circuit every second (see 
 p. 25). To the armature of the electro-magnet a stylus is 
 also attached. A plate of glass covered with " Bon Ami" is 
 clamped on a movable block so that each stylus rests upon 
 it. The electro-magnet and fork may have any relative 
 position which may be convenient, but the styli should not 
 be far apart. The fork is set vibrating and the block with 
 the glass is drawn along, the fork making a wave line and 
 the other stylus a straight line broken (or notched) every 
 second. With a square, lines are drawn at right angles to 
 the glass through the beginnings of alternate second sig- 
 nals and the number of complete vibrations, estimated to 
 tenths, is counted between the lines. 
 
 (6) This is known as a stroboscopic method and depends 
 upon the persistence of vision. The fork is watched through 
 
 holes in a disk revolving at a con- 
 stant speed. The holes are equally 
 spaced in concentric circles, the 
 number per circle increasing with 
 the radius. The speed of the disk 
 is varied until the fork appears 
 stationary when viewed through 
 FIG. 9. the holes of a particular circle. If 
 
 there are m holes in the circle, and 
 
 if the disk revolves n times per second, the frequency of the 
 fork is obviously mn. By varying the speed and using other 
 holes, additional determinations may be made. The speed 
 of the disk is obtained by determining, with a counter, the 
 number of revolutions in a given time. 
 
 (C) Another method of finding the coefficient of kinetic 
 friction is to make the slide horizontal and find the force 
 required to keep the block in uniform motion after it has 
 been started. For this purpose a braided cord is attached 
 to one end of the block, passed over a pulley at the end of 
 
COEFFICIENT OF FRICTION. 45 
 
 the slide, and attached to a scale pan, to which weights are 
 added. In this case the weight of the pan and weights 
 must not be taken as the force acting on the block, for some 
 force is required to overcome the friction of the pulley. 
 The amount required must be found by a separate experi- 
 ment. Two pans are attached to the ends of the cord 
 hanging over the pulley and sufficient equal weights are 
 placed on the pans to make the pressure on the pulley the 
 same as in the main experiment where the parts of the cord 
 were at right angles. The additional weight on one scale pan 
 requisite to keep the whole in constant motion when started 
 is the force needed to overcome the friction of the pulley, 
 and is, therefore, the correction required. 
 
 With this apparatus we may also test whether the coeffi- 
 cient of friction varies when weights are added to the block. 
 The correction for friction of the pulley does not need to 
 be re-determined experimentally, but may be calculated 
 from the former determination, on the assumption that the 
 friction of the pulley is proportional to the pressure on it. 
 
 The possible error in the results of the first and last 
 methods is easily determined. The most, accurate way of 
 finding the possible error in method (B) is by means of 
 formulae deduced by the differential calculus (see p. 6), 
 but a much simpler and a sufficiently accurate method is 
 the following: Note that an overestimate of i will increase 
 the value of fj. and the same will be the effect of an under- 
 estimate of a. Hence the coefficient should be recalculated 
 with tan i and cos i increased and a decreased by their 
 possible errors and the change found in the, coefficient may 
 be taken as the final possible error. The possible error of a 
 may be taken as its mean deviation and the possible errors 
 of tan i and cos i may be deduced from the measurements 
 from which they were obtained. 
 
 Questions. 
 
 1 . In the second method, why is it desirable that the straight 
 line be exactly in the middle of the wave line? 
 
 2. In the third method, what error would be introduced if the cord 
 from the block was not exactly horizontal ? 
 
46 MECHANICS. 
 
 IX. HOOKE'S LAW AND YOUNG'S MODULUS. 
 
 Text-book of Physics (Duff), 168, 171, 173; Watson's Physics, 
 172, 173; Ames 1 General: Physics, pp. 144, 145, 153, 154; 
 Crew's Physics, 126-129. 
 
 H coke's Law states that, for small strains, stress and 
 strain are proportional. Young's Modulus, E, is the con- 
 stant ratio of stress to strain for a stretching strain, the 
 stress being taken as the force per unit cross section and 
 the strain as the stretch per unit of length, or, if F is the 
 whole force, A the area of cross section, Lthe whole length, 
 and / the increment of length, 
 
 (A) The quantity most difficult to measure is /, the small 
 increase of length. If a wire be supported at one end and 
 force applied to the other end, there is danger that the 
 support may yield slightly, and a slight amount of 
 yielding will cause a proportionally large error in 
 the estimate of the small increase in length. The 
 peculiarity of the first method described below is 
 the means adopted to eliminate the yield of the 
 support. The increase of the length of the wire 
 under experiment is found by comparison with 
 another wire under constant stretch attached to 
 the same support as the former wire. One wire 
 
 6 carries a scale and the other a vernier opposite 
 the scale. If there be any doubt which is vernier 
 (see p. 13) and which is scale, comparison should 
 be made with an ordinary steel scale. The screws 
 FIG. 10. by means of which the wires are clamped to scale 
 and vernier should be adjusted until scale and 
 vernier tend to lie in one plane. A light rubber band may 
 then be slipped over scale and vernier to keep them together. 
 The stretch may be produced by means of lead weights. 
 The value of these weights should be determined by a 
 
HOOKE'S LAW AND YOUNG'S MODULUS. 47 
 
 platform balance. To produce a suitable stretch it may 
 be advisable to add two or more weights at a time. We shall 
 suppose that two are added, but the description can readily 
 be modified to suit any number. The greatest weight 
 should not be more than half that required to break the 
 wire. (A copper wire o.oi sq. cm. section will break at 40 
 kgs. ; brass, 60 kgs. ; iron, 60 kgs.) Suppose, then, two 
 weights are added at a time and each stretch observed. 
 When the maximum number has been added the weights 
 should be removed in the same order, readings being again 
 taken as they are removed. The whole series of observa- 
 tions should be repeated at least three times. Such readings 
 should always be arranged in tables having in a line or 
 column all the readings for a particular pair of weights. 
 The length of the wire may be measured by means of a 
 long beam compass and the diameter should be measured 
 at least a dozen times at different places and in different 
 directions by means of a micrometer caliper (see p. 14). 
 
 Before calculating, the dimensions should be expressed in 
 centimeters and the weights in dynes. First find the mean 
 value of / for each pair of weights when added and when 
 removed and then the value of F -=- / for each of these values 
 of / and the respective F's. Find the mean value of F + l 
 and the greatest percentage deviation from the mean. This 
 will give the percentage deviation from Hooke's Law since 
 F-r-l should be a constant, A and L being practically con- 
 stant. The final value of Young's Modulus should be 
 stated in the notation explained on page n. 
 
 The possible errors of the different quantities measured 
 may be taken as the mean deviation in each case. The per- 
 centage error of the final value of E will be, as is readily seen 
 from the formula, the sum of the percentage errors of F, L, I, 
 and twice the percentage error of the radius (see p. 6). 
 
 (B) Young's Modulus may also be found by means of 
 the flexure of a bar. For, in bending (within limits) one 
 side of a bar is stretched and the other compressed (nega- 
 tively stretched) , and so Young's Modulus is the only constant 
 
4 8 
 
 MECHANICS. 
 
 that need be considered. The amount of bending might be 
 deduced from the sag of the center or end of the bar, but a 
 much more delicate method is the following optical one: 
 
 A bar of rectangular cross section is laid on two knife- 
 edges and at each end is attached an approximately vertical 
 mirror in mountings that admit of considerable adjustment. 
 A vertical scale, nearly in line with the bar, is reflected 
 from the farther mirror into the nearer and thence into a 
 telescope also nearly in line with the bar. When a weight 
 is attached to the center of the bar, the bar is bent and 
 another part of the scale is reflected into the telescope. 
 This arrangement serves to determine the angle of bending. 
 For suppose the difference of the scale-readings on the 
 horizontal cross-hairs of the telescope be D cms. (Fig. n) 
 and let the distance between the two mirrors be p and the 
 distance of the scale from the farther mirror q, then, if the 
 change of inclination of each mirror be i, 
 
 D 
 
 "f Q n * _ 
 Lctll l> 
 
 For a consideration of figure 11 will show that d l = p tan zi; 
 But since i is a small angle tan 22 = 2 tan i and tan 
 
 ., = q tan 42. 
 ^ = 4 tan i. 
 
 tan 
 
 FIG. ii. 
 
 From tan i, the weight R in dynes applied at the center, the 
 length of the bar between the knife edges, /, the breadth, b, 
 and the thickness, a, Young's Modulus, E, is obtained, by 
 the equation 
 
 a s b tan i 
 
HOOKE'S LAW AND YOUNG'S MODULUS. 
 
 49 
 
 PROOF. 
 
 Let the y axis coincide with the radius from the center of curvature 
 of the bar to the center of the bar, and let the x axis be the tangent 
 to the central axis at this point. Designate distances from the elas- 
 tic central axis, LOM (Fig. 12), along other radii by z. The elastic 
 central axis remains unchanged in length. The curvature at any 
 point P on LOM is the rate of change of the directions of the tangent. 
 The angle the tangent line at P makes with the #-axis is a small one 
 and may be taken as dy/dx (which is really the tangent of that 
 angle) . The rate of change of the direction of the curve at the point 
 x, y, is therefore d 2 y/dx' 2 , which therefore equals the curvature. But 
 the curvature also equals i/r, r being the radius of curvature. Hence 
 
 Now consider two strips of the beam distant z from LOM. By 
 the bending these strips are changed in length in the proportion z/r or 
 2 d 2 y/dx 2 . (For, consider figure 13; the proportional change of 
 length is 
 
 G'H'-GH 
 GH 
 
 FIG. 12. 
 
 FIG. 13. 
 
 By definition of Young's Modulus, if a force F applied to a rod of 
 cross section A and length L produce an extension /, 
 
 EAl 
 
 F = 
 
 L ' 
 
 where E is Young's Modulus. The stress in a strip of width b and 
 thickness dz is obtained by putting bdz for A and z d 2 y/dx 2 for l-i-L, 
 which gives 
 
 Hence the moment about P of the restoring force in the strips z is 
 
50 MECHANICS. 
 
 The moment about P of the stress in the whole cross section is the 
 integral with reference to z of the above expression for values of z 
 from o to \a or 
 
 For equilibrium this must equal the moment of %R about P or 
 
 d 2 y^ 6R n 
 dy 6R \lx x 2 
 
 ' dx Ea 3 b( 2 2 j 
 At a point of support 
 
 Hence by substitution 
 
 Rl 2 
 
 4 a 3 6 tan i' 
 
 (By integrating again, the value of y at a point of support or the 
 deflection of the beam is obtained. This is left as an exercise for 
 the student.) 
 
 The adjustment of the apparatus is most readily made 
 as follows. Place the telescope and scale nearly in the line 
 of the mirrors and, glancing above the telescope, set the 
 farther mirror so that the nearer mirror is seen by reflection 
 and then the latter so that the scale is seen. Then adjust 
 the eye-piece of the telescope so that the cross-hairs are as 
 distinct as possible and finally focus the telescope until the 
 scale is seen. The bar must not be strained beyond the 
 limits of elasticity. For adjustment of the telescope and 
 scale, see p. 25. Equal weights, perhaps 100 grams at a 
 time, should be added, but this process should be stopped 
 when it is found -that the scale-reading no longer changes in 
 the same proportion as the weights. Determine carefully 
 by several readings, with and without this maximum weight 
 attached, the change of scale-reading. The width and 
 thickness of the bar may be measured by a micrometer 
 caliper (see p. 14), a number of readings of each at different 
 points being made. 
 
 In calculating, use this weight for R and the average of 
 the changes of deflection for D and take the mean deviation 
 
RIGIDITY. 51 
 
 as the measure of the possible error of D, a, and b. The 
 percentage possible error of tan i is deduced from the possible 
 errors of D and 2p+4q. The possible error in the latter 
 term is twice the possible error in p plus four times that in q. 
 
 (C) A simple optical method may also be employed for 
 finding the extension of a wire. In this method, one side 
 of a small bench carrying a vertical mirror is supported by 
 the end of the wire and the other by a fixed bracket. The 
 deflection of the mirror when weights are added to the wire 
 is read by a scale and telescope. The details of the method 
 may readily be worked out by anyone who has followed the 
 preceding methods. 
 
 (D) (Searle's Method.) The extension may also be deter- 
 mined from the change of position of a level supported by 
 the two wires. The lowering of the stretched wire is com- 
 pensated by a micrometer screw which therefore reads the 
 extension. For details, see Watson's Practical Physics, 45. 
 
 Questions. 
 
 1. How closely is it worth while to measure the length of the wire 
 in the first method ? 
 
 2. Which of the first two methods is the more accurate and what 
 is the chief weakness of the other ? 
 
 3. In the second method why is nothing said as to the distance 
 of the mirrors beyond the knife-edges? Might they be placed inside? 
 
 4. Reduce your results to tons and inches. 
 
 X. THE RIGIDITY (OR SHEAR -MODULUS). 
 
 Text-book of Physics (Duff}, 119, 170; Watson's Physics, 171, 174, 
 175; Ames' General Physics, pp. 151153; Crew's Physics, 131, 
 132; Duff's Mechanics, 117, 130, 131. 
 
 The rigidity of any material is the resistance it offers 
 to change of shape without change of volume. It is meas- 
 ured by the ratio of the shearing stress to the shear pro- 
 duced. In the twisting of a wire or rod, within moderate 
 limits, there is no change of volume. Hence this affords 
 a means of finding the rigidity of the material. The con- 
 stant or modulus of torsion of a particular wire is the couple 
 
52 MECHANICS. 
 
 required to twist one end of unit length of the wire through 
 unit angle, the other end being kept fixed. If it be denoted 
 by r and if the length of the wire be L the couple required to 
 twist the wire through unit angle is r/L. If now to the wire 
 be attached a mass of moment of inertia, /, and the wire 
 and the mass be set into torsional vibrations, the time of a 
 semi-vibration is, by the principles of Simple Harmonic 
 Motion (see references) , 
 
 If t, L and / be found r can be deduced. From the modulus 
 of torsion of the particular wire the rigidity n of the material 
 of which the wire consists can be deduced ; for 
 
 2T 
 
 Proof. 
 
 Suppose unit length of the wire to be twisted through unit angle. 
 The vibrations are due to the restoring couple at the lower end pro- 
 duced by the twist. Let the cross section of the end be divided into 
 concentric rings and let the radius of one ring be x and its width dx; 
 its area is z-xxdx. Relatively to the fixed end it is displaced through 
 unit angle. Hence the linear displacement (supposed small) of the 
 ring whose radius is x is x times unit angle or simply x. This is by 
 definition the shear and hence the shearing stress is nx. This is 
 the restoring force per unit area of the cross section. Hence the 
 restoring force of the ring whose radius is x is 2xnx 2 dx. The effect 
 of this force in producing rotation depends on its moment about the 
 axis or 2nnx*dx. The moment of the restoring force of the whole end 
 section is the sum of expressions like 2xnx 3 dx for values of x between 
 o and r or ^nnr*. This is by definition the modulus of torsion, r and 
 gives us the above equation. It should be noticed that it is not a 
 constant for the material of the wire, but depends on the dimensions 
 of the particular wire. 
 
 The length, L, may be measured by means of a long 
 beam compass which is afterward compared with a fixed 
 brass scale. The radius, R, may be measured by a microm- 
 eter caliper (see p. 14), measurements being made at a great 
 many different places and the mean taken. 
 
 The moment of inertia, /, of the wire and attached mass 
 
RIGIDITY. 53 
 
 might be roughly obtained by calculation, but it is better 
 to apply an experimental method that is used in other cases. 
 This consists in adding to the vibrating mass, of unknown 
 moment of inertia, another mass of such regular form that 
 its moment of inertia can be accurately calculated, and 
 finding the times of vibration before (t) and after (T) adding 
 this mass. If the original moment of inertia be / and the 
 added moment of inertia i: 
 
 whence 
 
 One of the simplest forms of added inertia is that of a solid 
 cylinder of circular cross section vibrating about an axis 
 through the center of the axis of the cylinder and at right 
 angles to it. The vibrating mass may then be in the form of 
 a hollow cylinder in which the solid cylinder may be placed. 
 If / be the length and r the radius of the solid cylinder of 
 mass m: 
 
 /I 2 r*\ 
 =m ( +- . 
 \ia 47 
 
 The quantities / and r can be obtained with sufficient pre- 
 cision by measurement with a steel scale divided to mm.'s, 
 and m may be found by a platform balance. In the above 
 formula for i, it is assumed that the axis of rotation is per- 
 pendicular to the axis of the cylinder. That this may be so 
 the carrier must be carefully leveled. This may be done 
 by supporting close under it a rod that is carefully leveled 
 by a spirit-level and comparing the carrier as it swings 
 with the leveled rod. The end of the vibrating mass should 
 be provided with an index, such as a vertical needle. A 
 stationary vertical wire is placed in front of this index 
 when the latter is in the position of rest. The body is set 
 
54 MECHANICS. 
 
 vibrating through an angle of between 60 and 90, all 
 pendulum vibrations being carefully suppressed. 
 
 The time of vibration may be found by much more accurate 
 methods than simply timing a certain number of vibrations. 
 The most common methods for accurately timing vibrations 
 are the "method of coincidences" and the "method of 
 passages." The former is especially useful for finding the 
 time of vibration of a pendulum whose half period is approx- 
 imately one second (Exp. VII). The method of passages 
 will be found suitable for the present experiment. It con- 
 sists in noting as accurately as possible the time of every 
 nth passage of the vibrating system through its mean posi- 
 tion or position of rest. The value to be chosen for n is a 
 matter of convenience when two observers work together, one 
 counting the seconds and the other noting the passages, or 
 when a single observer has a chronometer in front of him. 
 But a single observer noting time by a clock circuit and 
 sounder should choose for n an odd number such that n 
 semi- vibrations occupy a little more than a minute. (It is 
 supposed that there is a minute signal, such as the omission 
 of a tick; see page 25.) 
 
 The passages are observed as follows: After a minute 
 signal, the seconds are counted until a passage occurs, for 
 example, from left to right. The second and fraction of a 
 second of this passage is recorded. The succeeding passages 
 in each direction are counted until the minute signal, 
 after which the seconds are again counted until the passage 
 occurs from right to left, for which the second and fraction 
 of a second is recorded. Obviously, if n has been properly 
 chosen, the passage just recorded is the nth. The succeeding 
 passages are counted until the next minute ignal, after 
 which the second and fraction of a second of the 2wth 
 vibration (from left to right) is recorded. 
 
 The following suggestion may aid in counting the seconds 
 and estimating fractions of a second. The observer should 
 keep counting seconds (not necessarily out loud) along 
 with the clock; when the number of the second is of two or 
 
RIGIDITY. 55 
 
 more syllables, the accent should be thrown on one syllable 
 whose sound should coincide with the tick; thus, eleven, 
 thirteen, fourteen, etc., twenty-one, twenty-to?, etc. The 
 passage will usually occur somewhere between two ticks. 
 To estimate at what point of time between the two seconds 
 the passage takes place, the indications of the eye may be 
 used to reinforce those of the ear. Suppose A (in Fig. 14) 
 to be the mean position of the index on the vibrating body, 
 then if B and C be its positions at the fifth and sixth 
 ticks, respectively, and if BA be six- 
 '* > tenths of the distance BC, it is evident 
 < tf'Vv' that the true time of passage is 5.6 
 seconds. With practice the eye can 
 become very expert in making such 
 
 judgments, and, for the purpose of attaining such skill, the 
 method should be used from the beginning, although at first 
 not much reliance can be placed on the judgment. 
 
 For simplicity of description we shall suppose that n is 5, 
 but the proper substitutions must be made if n has any 
 other value. When the approximate time of 5 vibrations 
 has been obtained by observing a few passages, all of the 
 passages need not thereafter be observed in order to ascertain 
 when 'each fifth passage is due, for this can readily be foreseen 
 by adding to the time of the last observed passage the 
 known approximate time of 5 vibrations. Further assistance 
 is obtained by recording the time of the o, loth, 2oth, etc., 
 in a second column, the first column being headed "left 
 to right" and the second "right to left." In this way such 
 a record as the following is obtained: 
 
 M. S. 
 
 l/U JX. 
 
 M. S. 
 
 JX. LU 
 
 M. S. 
 
 M. 
 
 S 
 
 (0) . . . . 
 
 (SO) ...- 
 
 ($)'- 
 
 (55).. 
 
 
 (10) 
 
 (60) 
 
 (15) 
 
 (65) 
 
 
 (20) 
 
 (70) 
 
 (25) 
 
 (75) 
 
 
 (30) 
 
 (80) 
 
 (35) 
 
 (85) 
 
 
 (40) 
 
 (90) 
 
 (45) 
 
 (95) 
 
 
56 MECHANICS. 
 
 from which the time of vibration is calculated thus : 
 
 M. s. M. s. 
 
 (50)- (o) .... (55)- (5) .... 
 (60) (10) (65) (15) 
 
 (70) (20) (75) (25) 
 
 (80) (30) (85) (35) 
 
 (90) (40) (95) (45) 
 
 Mean of 50 vibrations . ... 
 
 Final mean of 50 vibrations = . . . .possible error. . . . 
 Final mean of one vibration = . . . . possible error .... 
 To find the possible error in the value found for n, first 
 eliminate r and i from the equation given above and express 
 n in terms of the quantities observed L, /, T, t R, m. (r 2 / 4 
 is so small compared with I 2 / 12 that the effect of the possible 
 error in the former may be neglected.) T and t come in 
 only in the form (T 2 t 2 ) and the possible error in this may 
 be found by methods stated on page 5. 
 
 Questions. 
 
 1. To increase the accuracy of the result, which quantity would 
 have to be measured more closely ? 
 
 2. What sources of error are there other than those referred to in 
 the text? 
 
 XL VISCOSITY. 
 
 Text-book of Physics (Duff), 196-198; Watson's Physics, 161; 
 Ames' General Physics, pp. 139, 168. 
 
 A solid has rigidity; that is, it offers a continued resist- 
 ance to forces tending to change its shape. A liquid has no 
 rigidity and offers no continued resistance to forces tend- 
 ing to change its shape; that is, the smallest force if given 
 time will produce an unlimited change in the shape of the 
 liquid. But the rate at which a liquid changes its shape 
 under a given force is not the same for all liquids. Some 
 liquids change very slowly and are called viscous liquids, 
 others change rapidly and are called mobile liquids. The 
 
VISCOSITY. 57 
 
 action of both can be stated in terms of a property called 
 viscosity. 
 
 The viscosity of a fluid may be defined as the ratio of the 
 shearing stress in the fluid to the rate of shear. From 
 this general definition a simpler definition can be readily 
 deduced. A shear consists essentially in the sliding of layer 
 over layer and the shearing is the force per unit area re- 
 quired to produce the shear. Hence we have the following 
 equivalent definition: "The coefficient of viscosity is the 
 tangential force per unit of area of either of two horizontal 
 planes at unit distance apart, one of which is fixed while 
 the other moves with unit velocity, the space between the 
 two being filled with the liquid." (Maxwell.) 
 
 (A) The flow of liquid through a capillary tube is essen- 
 tially of the nature of sliding of layer over layer. The 
 cylindrical layer in immediate contact with the tube remains 
 fixed or at least has no motion parallel to the axis of the 
 tube, and the immediately adjacent layer slides over it, the 
 next layer slides over the second, and so on up to the center 
 of the tube. (In a tube of greater than capillary bore this 
 is not so, for there are eddies in the motion. This distinc- 
 tion is in fact the best definition of the term capillary.) 
 
 Thus, if we measure the force causing flow through the 
 tube and the rate of flow, we shall be in a position to deduce 
 the coefficient of viscosity of the fluid. In fact, if M be the 
 mass of liquid of density d that flows in time t, through a 
 vertical tube of length / and radius of bore r, and if h be the 
 vertical distance from the level of the liquid in the reser- 
 voir above the tube to the lower end of the tube, the coeffi- 
 cient of viscosit is 
 
 Proof. 
 
 Suppose all the liquid in a capillary tube of length / and radius r 
 to be solidified except a tubular layer of mean radius x and thickness 
 dx. If there be a difference of pressure p (per unit of area) between 
 the two ends, the solid will attain a steady velocity such that the 
 
58 MECHANICS. 
 
 viscous i 
 ends. I 
 ity that 
 
 viscous resistance just equals the whole difference of pressure on its 
 ends. Hence it follows from the definition of the coefficient of viscos- 
 
 } xdx 
 
 pxx 
 Hence, v = r , 
 
 2/73 
 
 If q be the volume of the core that flows out per second, 
 
 . . 
 
 2/7} 
 
 Suppose now another layer liquefied. There will follow a further 
 flow represented by the same expression but with a different value 
 for x. Let the process be continued until the whole is liquid, then 
 the whole flow per second, Q, will be the sum of all the values of q 
 for values of x between o and r. Hence 
 
 If the tube be vertical and the flow be due to gravity only, instead of 
 p we must put gdh. If M be the mass of density d that flows out in 
 time t, 
 
 In the above it was tacitly assumed that the liquid adheres to the 
 tube without any slip. If there were any slip the outflow would be 
 increased by it and the above expression would not hold. Poiseuille 
 and others verified the above formula in all cases, thus showing that 
 no slip occurs. (A more formal proof of the above equation is given 
 in Tait's "Properties of Matter," 317). 
 
 A piece of capillary tubing should be chosen whose bore 
 is as nearly as possible circular in section. This can be 
 tested by examining the ends under a micrometer micro- 
 scope (see p. 15). If the section is found to be nearly cir- 
 cular the principal diameters of the bore should be measured. 
 This should, however, only be regarded as a preliminary 
 measurement, serving as a test of the circularity of the bore 
 and a check on the following more satisfactory method. 
 
 The mean radius of the bore can be best determined by 
 weighing the amount of mercury that fills a measured length 
 of the tube. For this purpose the tube should be first 
 cleaned by attaching it to the end of a rubber tube, at 
 the other end of which is a hollow rubber ball, and thus 
 drawing through it and forcing out a number of times (i) 
 chromic acid; (2) distilled water; (3) alcohol, and finally 
 drying it by sucking air through it. Then draw into the tube 
 
VISCOSITY. 
 
 59 
 
 T 
 
 a column of clean mercury and measure its length as accu- 
 rately as possible by a comparator (see p. 15). 
 
 The mass of the mercury should next be ascertained by 
 weighing it with great care in a sensitive balance (for full 
 directions see pp. 21-24). The mercury should not be 
 dropped directly on the scale pan, but into 
 a watch-glass or paper box placed on the 
 scale pan. From these measurements and 
 the density of mercury at the temperature 
 of observation (see Table VII) the diameter 
 of the bore is obtained. It may be noted 
 that since it is r 4 that is used in the formula 
 for viscosity and r 2 that is obtained directly 
 from the mercury measurements of the bore 
 the value of r need not be deduced. The 
 length of the tube may be measured by the' 
 comparator as already described. 
 
 The tube is then attached vertically by a 
 rubber connection to a funnel and the mass 
 of water that flows through the tube in a 
 given time found by weighing a beaker (i) 
 empty and (2) containing the water that 
 has passed. The time is obtained by ob- 
 serving a clock ticking seconds or a chronometer. It is 
 evident that the greater the whole time the less the per-, 
 centage error in time due to errors in observing the time 
 of starting and stopping, and so, too, the greater the whole 
 mass the less the percentage error in weight. Hence the 
 time and the mass should be sufficiently great to make the 
 percentage errors in them less than those in / and r 4 . To 
 prevent evaporation from the beaker it should be covered 
 by a sheet of paper pierced by a hole through which the 
 tube passes. While the liquid is flowing the temperature of 
 the water in the funnel should be noted. 
 
 The value of h is the mean of its values at the beginning 
 and end of the flow. These values are best obtained by a 
 cathetometer (p. 19). For this purpose a very simple form 
 
 FIG. 15. 
 
60 MECHANICS. 
 
 of instrument may be used. A vertical scale is placed near 
 the apparatus for viscosity and the cathetometer (a telescope 
 that may be leveled, movable along a vertical column that 
 .may be made truly vertical) placed so that its telescope 
 (leveled to horizontality) may be turned, so that the inter- 
 section of its cross-hairs coincides alternately with the image 
 of the water surface and that of the scale. This gives the 
 level of the surface of the liquid on the vertical scale. The 
 level of the lower end of the vertical tube is obtained in the 
 same way, whence h is obtained. 
 
 The viscosity of alcohol may be measured by the same 
 means, particular care being taken to prevent evaporation. 
 The possible error of the result is readily calculated from 
 the possible errors of the separate measurements. The 
 possible error of r is not needed, but that of r 4 should be 
 deduced directly from the determination of r 2 . 
 
 Questions. 
 
 i. Could the radius be found satisfactorily by measurements 
 with a micrometer microscope? Explain. 
 
 -2. What mass of this liquid would flow through a tube i mm. in 
 diameter and i meter long, under a constant head of 2 meters ? 
 
 3. Two square flat plates of 20 cm. edge are separated by i mm. of 
 this liquid. What force would be required to move one with a 
 velocity of 30 cm. per second, the other being at rest? 
 
 XII. SURFACE TENSION. 
 
 Text-book of Physics (Duff), 206214; Watson's Physics, 155160; 
 Ames' General Physics, pp. 182-190; Crew's Physics, 149-160; 
 Poynting and Thomson, Properties of Matter, Chap. XIV. 
 
 The height to which liquid rises or is depressed in a 
 capillary tube depends on the surface tension of the liquid, 
 the angle of capillarity, and the radius of the tube. From 
 measurements of the height, h, and radius, r, the surface 
 tension is deduced if the angle of capillarity is known, for 
 (see references) 
 
 rdgh 
 
 T = 
 
 2 cos a 
 
SURFACE TENSION. 6 1 
 
 d being the density of the liquid and g the acceleration of 
 gravity. In the case of perfectly pure distilled water the 
 angle of capillarity a or the angle at which the surface of 
 the liquid meets the glass, is zero and so cos a= i. 
 
 It is important that the capillary tube be quite clean. 
 The cleaning should be performed with chromic acid and 
 distilled water. The height of the water in the tube can be 
 measured in two ways. One method is to place a scale 
 etched on mirror glass behind the tube. The mean level 
 of the meniscus-shaped surface of the liquid in the tube 
 and the ordinary plane surface of the liquid in the vessel 
 should be read. A preferable method is to measure the 
 distance between the two surfaces with a cathetometer 
 (see p. 19). 
 
 To make certain that the inner surface of the tube is 
 wet by the water and that the angle of capillarity is zero, 
 the tube should be thrust deeper into the liquid and then 
 withdrawn before the levels of the surfaces are read. This 
 should be repeated and the height read several times, 
 different parts of the scale being used, but the part of the tube 
 in which the liquid rises remaining the same. If the motion 
 of the liquid in the tube is sluggish or uncertain, the tube 
 should be more carefully cleaned. Finally, the point to 
 which the liquid rises in the tube should be marked on the 
 tube by a sharp file. 
 
 The tube should then be carefully broken at the point 
 marked and its diameter should be carefully measured by 
 means of a micrometer microscope (see p. 15). If the 
 section of the bore is not circular, the greatest and least 
 diameters should be carefully measured and the mean 
 taken, but if they differ very much the result will not be 
 satisfactory. 
 
 The whole should be repeated with as many tubes of 
 different sizes as time will permit. The temperature at 
 which the work is performed should be stated. 
 
 If time permit, determine also the surface tension of an 
 assigned solution. 
 
62 MECHANICS. 
 
 Make an estimate of the possible error for the results 
 obtained by one of the tubes. 
 
 (Apparatus for determining the surface tension at different 
 temperatures is described in Findlay's Practical Physical 
 Chemistry, p. 78; and E well's Physical Chemistry, p. 117.) 
 
 Questions. 
 
 1. Are the errors of measurement sufficient to explain the dif- 
 ferences between results with different tubes? 
 
 2. What other sources of error may there be? 
 
 3. How could the surface tension of mercury be obtained in an 
 analogous way? 
 
 4. How high would this liquid rise in a tube o. i mm. in diameter? 
 
HEAT. 
 
 25. Radiation Correction in Calorimetry. 
 
 Watson's Practical Physics, 82; Ostwald's Phys. Chem. Meas., p. 
 124-127; Poynting and Thomson, Heat, Chap. XVI. 
 
 A body which is above the temperature of surrounding 
 bodies falls in temperature at a rate that is proportional to 
 the excess of its temperature above that of its surround- 
 ings. This is Newton's Law of Cooling* If the mean 
 excess of the body's temperature in any time be known and 
 also its rate of loss of temperature at some particular ex- 
 
 FIG. 16. 
 
 cess, its mean rate of loss of temperature is readily deduced, 
 and this multiplied by the time for which the mean is taken 
 will give the whole loss of temperature. 
 
 Consider the case of the heating of a vessel containing 
 water by the passage of steam into the water. If a curve 
 (Fig. T 6 from o upwards) showing the rise of temperature of 
 
 63 
 
64 HEAT. 
 
 the water be drawn and the same continued after the water 
 has reached its highest temperature, the latter, or straight 
 line part of the curve, will give the rate of loss of temperature 
 at the highest temperature attained. Let us denote this 
 rate by r (degrees per minute) . If the excess of temperature 
 when the temperature is highest is t and if the mean excess 
 during the whole time of rise of temperature is i' then the 
 mean rate of cooling was by Newton's Law rt f \t and this 
 multiplied by the whole time of rise of temperature, T (min- 
 utes), gives the whole loss of temperature. Hence the final 
 (highest) temperature must be corrected by addition of rTt'/.t. 
 If the curve of rffee of temperature is a straight line, t' is half 
 of t and the correction is rT/2. When the curve is not 
 approximately a straight line the whole time T must be 
 divided into a number of intervals (each perhaps of 30 sec.) 
 and t f must be obtained by averaging the excesses in these 
 intervals. 
 
 When the calorimeter is cooled below the temperature 
 of the room (e. g., by adding ice to the water) the calor- 
 imeter gains temperature by radiation from the surround- 
 ings; but the above method will still apply except that we 
 shall have to do with rates of warming instead of rates of 
 cooling and the correction of the final temperature will be 
 subtractive. 
 
 If the calorimeter starts below the temperature of the 
 room and is heated above it, the correction must be made in 
 two parts as above. In this case we must find the initial 
 rate of warming (before the hot body is placed in the calor- 
 imeter) and also the final rate of cooling (after the highest 
 temperature was attained). The correction will also be in 
 two parts when the calorimeter starts above the room tem- 
 perature and ends below it. 
 
 If the main rise of temperature is closely represented by 
 a straight line, it is easily shown* that the correction amounts 
 to the algebraic average of the initial and final rates multi- 
 plied by the whole time that the calorimeter is heating or 
 
 * Ewell's Physical Chemistry, p. 84. 
 
THE BECKMANN THERMOMETER. 65 
 
 cooling. In fact, if the water is T l minutes below the 
 temperature of the room and T 2 minutes above the room 
 temperature, the radiation correction is (T 2 r 2 T 1 r 1 )/2 and 
 this differs from (T i +T 2 )(r 1 +r 2 )/2 by (T l r 2 T 2 r l )/ 2, which 
 is zero, since under these circumstances the rate is propor- 
 tional to the time that the water is above or below the room 
 temperature. This fact is particularly useful in cases where 
 the surrounding temperature is indefinite (Exp. XXVII, for 
 example) . 
 
 In every calorimetry experiment where the temperature 
 changes, this radiation correction must be applied, and there- 
 fore the initial and final rates of change of temperature must 
 be determined. The rate may usually be found with sufficient 
 accuracy by reading the temperature every minute for five 
 minutes. In very accurate work, more careful methods 
 must be applied. 
 
 26. The Beckmann Thermometer. 
 
 Watson's Practical Physics, 102; Findlay, Practical Physical Chem- 
 istry, pp. 114117; Ostwald, Phys, Chem. Meas., pp. 119120. 
 
 The Beckmann thermometer is used for determining 
 changes in temperature. The bulb is large and the stem is 
 small so that a small change of temperature is shown by a 
 large change in reading. The amount of mercury may be 
 varied, and the temperature corresponding to a particular 
 reading will vary with the amount of mercury in the bulb 
 and stem. There is a reservoir at the end of the stem into 
 which surplus mercury may be driven by warming the bulb. 
 A gentle jar will detach the mercury in this reservoir when 
 sufficient has been expelled. If one desires to study high 
 temperature changes, the bulb is warmed until the thread 
 of mercury extends to the reservoir, when the mercury in 
 the reservoir is joined to it. The bulb is then allowed to 
 cool until sufficient mercury has been drawn over, when the 
 thread is detached from the mercury in the reservoir by a 
 gentle jar. Several trials are often necessary before the 
 5 
 
66 HEAT. 
 
 proper amount of mercury is secured. In an improved type 
 of Beckmann thermometer, two reservoirs are provided, and 
 the first has a scale which tells the amount of mercury 
 required in that reservoir for different ranges of temperature. 
 Beckmann thermometers are delicate and expensive and 
 must be handled with the greatest care. 
 
XIII. THERMOMETER TESTING. 
 
 Watson's Practical Physics, 59-69; Edser, Heat, pp. 23-36; Text- 
 book of Physics (Duff), pp. 189, 190; Watson* s Physics, 177- 
 182; Ames' General Physics, pp. 220-224; Crew's General Physics, 
 249-252. 
 
 The readings of a thermometer gradually change for a 
 long time after the thermometer has been filled. The cause 
 of this is the gradual recovery of the bulb from the effect 
 of the very great heating to which the glass was subjected 
 when the thermometer was made. The shrinkage is rapid 
 at first and slower afterward, but may continue for years. 
 Hence the necessity for re-determining, from time to time, 
 the so-called "fixed points" of a thermometer, namely, the 
 reading in melting ice, and that in steam at standard, pres- 
 sure. When the thermometer is first graduated it is usually 
 done by determining the fixed points and dividing the 
 distance between them into 100 equal parts laid off on the 
 stem. This assumes that the bore is uniform or that, by 
 calibration of the bore, the variations of the bore are deter- 
 mined and allowed for in a table of corrections to be applied 
 to the readings of the thermometer in order to obtain the 
 true temperature. Usually the variations of the bore are 
 too small to have any appreciable effect except in cases 
 where extreme accuracy is aimed at. Nevertheless, every 
 thermometer needs to be carefully examined in this regard. 
 Let us suppose that on the scale laid off on the stem the 
 true readings in ice and steam have been obtained and for 
 the moment let us suppose that the bore is quite uniform. 
 To see how to make corrections for other points on 
 the scale we must consider the elementary definition of 
 temperature. 
 
 Temperature on the mercury scale is defined by the 
 expansion of mercury (relatively to glass). Let f ]00 be 
 
68 HEAT. 
 
 the volume of a mass of mercury at the temperature of 
 steam under a pressure of 76 cm. and let V Q be its volume 
 at the temperature of melting ice. The degree is defined 
 as the rise of temperature that would produce an expansion 
 of (^ 100 ^o)/ IOO > an d T above zero is, therefore, the rise of 
 temperature that will produce an expansion of T(v 1Q<) v ) / 1 oo. 
 Hence if at T the volume of the mercury be v, 
 
 v v f 
 
 IOO 
 
 .'. T ' = 100. 
 
 This definition depends only on the expansion of mercury 
 and the expansion of the particular glass used and is other- 
 wise independent of the size and shape of the thermometer. 
 Now regard the thermometer tube under test as simply a 
 graduated cylinder of constant cross section containing 
 mercury. Let the height of the mercury as read on the 
 scale when the thermometer is in melting ice be a; when it is a 
 steam at 76 cm. let it be b, and when it is at the temperature 
 T, let it be t. Then 
 
 v v ta 
 Hence 
 
 J. \V \MJ , 
 
 b a 
 
 where T is the true temperature when the reading of the 
 thermometer is t. By this equation values of T for values 
 of t for every five degrees should be calculated. Having 
 thus drawn up a table of true temperatures we subtract the 
 scale-reading from the true temperature and thus get a 
 correction (positive or negative), which added to the scale- 
 reading gives the true temperature. 
 
 This is on the assumption that the bore is sensibly uni- 
 form. The only quite satisfactory method of testing this 
 
THERMOMETER TESTING. 69 
 
 is to calibrate the bore by measuring the length of a short 
 thread of mercury at different positions in the tube. This 
 process requires considerable time and the following will 
 usually suffice: Two thermometers for which tables of 
 true readings have been drawn up as above, are compared 
 at regular intervals (say every five degrees) between zero 
 and 100 by being used simultaneously to measure the tem- 
 perature of a body. If, after corrections, the readings of 
 the thermometers are not sensibly different, this shows that 
 the bores of both must be practically uniform. If they do 
 differ appreciably, then the bore of one or both must be 
 variable. If they be compared with a third thermometer, 
 the one with the variable bore will be detected and it must 
 be then calibrated. 
 
 Testing Zero-point. A calorimeter consisting of a small 
 copper vessel inside of a larger is suitable for holding the ice. 
 Both vessels should be washed in ordinary tap water. 
 The space between the two vessels should be filled with 
 cracked ice, and the inner vessel filled with cracked ice and 
 then distilled water poured in until the vessel is filled to the 
 brim. The thermometer having been washed clean, is 
 inserted in the inner vessel, just sufficient of the stem being 
 exposed to admit of the zero being observed. When the 
 reading has fallen to i the reading should be observed 
 every minute until it is stationary for five minutes. This 
 stationary temperature, read to o . i degree, is the true zero 
 point, or a in the above equation. 
 
 Sources of Error. 
 
 (1) Impurity in the ice or water. 
 
 (2) The presence of water above o near the bulb of the 
 thermometer. 
 
 Testing Boiling-point. The form of boiler used for this 
 test consists of a vessel for boiling water surmounted by a 
 tube up which the steam passes, this tube being enclosed in 
 another down which the steam passes to an exit tube and a 
 pressure gauge (see Fig. 17). Half fill the lower part of the 
 vessel with water. Push the thermometer to be tested 
 
HEAT. 
 
 ] 
 
 through a cork in the top until the boiling-point is only a 
 degree or two above the cork, but take care that the 
 bulb of the thermometer does not reach down to the 
 water. Apply heat, adjusting it carefully as boiling begins, 
 so that the pressure inside, as indicated by the pressure 
 
 gauge, shall not materially ex- 
 ceed atmospheric pressure. 
 Some excess is, of course neces- 
 sary, if there is to be a free 
 flow of steam. What excess is 
 permissible may be deduced from 
 the consideration that a rise of 
 pressure of i cm. (of mercury 
 column) corresponds to a rise of 
 boiling-point of 0.373 (see Table 
 XIII). If water is used in the 
 pressure gauge, a pressure of i 
 cm. of water column would cor- 
 respond to only 0.03 rise of 
 steam temperature. If the ther- 
 mometer be graduated to degrees 
 only, an error of 0.03 in finding 
 the boiling-point is negligible. 
 
 Read the barometer and reduce 
 the height to zero degrees (p. 21). 
 
 To the boiling-point thus 
 found a correction must be ap- 
 plied, for the difference between the atmospheric pressure 
 at the time and that of a standard atmosphere (76 cm. of 
 mercury). Find from Table XIV the true temperature of 
 the steam at this pressure, and the difference between the 
 boiling-point observed on the thermometer and this tem- 
 perature. Since this temperature is always within a few 
 degrees of 100, the thermometer will have practically the 
 same error at 100. Therefore b in the above equation 
 may be taken as 100 plus or minus the difference between 
 the observed boiling-point and the true boiling temperature. 
 
 FIG. 
 
TEMPERATURE COEFFICIENT OF EXPANSION. 71 
 
 Comparison of Two Thermometers. The most satisfac- 
 tory method is to immerse the thermometers in steam above 
 water boiling under a pressure that can be regulated. A 
 simple means that is sufficient if the thermometers are of 
 the same length and graduated to degrees only, is to use 
 the thermometers simultaneously to find the temperature of 
 a block of good conducting material (copper or brass) im- 
 mersed in a vessel of water the temperature of which can 
 be gradnally raised by a burner. The thermometers should 
 be thrust in holes close together in the block and before 
 each reading the burner should be removed and the water 
 well stirred for a minute so that the temperature of the 
 block shall become uniform. 
 
 Questions. 
 
 1. Which should be determined first, boiling-point or freezing- 
 point, and why? 
 
 2 . How much error might there be in determining the boiling-point 
 if only the bulb were immersed in the steam? 
 
 3. Why is there no need to take account of barometric pressure 
 in finding the zero-point? 
 
 XIV. TEMPERATURE COEFFICIENT OF EXPANSION. 
 
 Edser, Heat, pp. 39-61; Text-book of Physics (Duff}, pp. 194-197; 
 Watson's Physics, 184, 185; Ames' General Physics, pp. 229 
 233; Crew's General Physics, 263-265. 
 
 For measuring the thermal expansion of a body, choice 
 may usually be made from a variety of methods. The par- 
 ticular method chosen will depend on the form of the speci- 
 men. The expansion of a metal rod may be measured by 
 means of a spherometer or by means of two reading micro- 
 scopes focused on definite marks near the ends of the speci- 
 men. The expansion of a wire is best measured by an opti- 
 cal lever method. The expansion of a solid of irregular 
 form can be found by a hydrostatic method, namely, by 
 weighing it in a liquid at different temperatures, it being 
 supposed that the density of the liquid at different tempera- 
 tures is known. 
 
HEAT. 
 
 (A) Expansion of a Metal Rod.- The rod is supported 
 at the lower end on a firm point and is heated by being en- 
 closed in a tube through which steam is passed from a 
 boiler. A spherometer (see p. 16) is so supported that the 
 end of the screw can be brought down on the flat end of the 
 rod. The spherometer is supported in the 
 hole, slot, and plane method, so that its 
 position is definite and not liable to be dis- 
 turbed by thermal expansion of the sup- 
 porting surface. 
 
 The rod is first measured by means of an 
 ordinary meter scale divided to mms. It is 
 then accurately placed in position in the 
 heating tube, the end of the rod projecting 
 through corks. Through the cork at the 
 upper end should also pass a glass tube for 
 the entry of the steam, while a similar tube 
 at the lower end serves to drain off the water. 
 At least six readings of the spherometer 
 scales should be made at the room tempera- 
 ture. Then pass steam into the jacket 
 about the rod. Every few minutes read the 
 temperature of the interior as given by two 
 thermometers at different heights and read 
 the spherometer. When the temperature 
 has become constant, make at least six read- 
 ings of the spherometer and several readings 
 of the thermometer. Always estimate tenths 
 of the smallest division. From the differ- 
 ence in spherometer readings, the length, and the change 
 in temperature, calculate the coefficient of expansion. 
 
 (B) Expansion of a Wire. For this an optical lever 
 method is most suitable. A mechanical lever or system of 
 levers is sometimes employed for magnifying small mo- 
 tions. A ray of light reflected from a mirror that is tilted 
 by the expansion serves' the purpose of a long index arm 
 much better, inasmuch as it has no weight itself and may 
 
 FIG. 18. 
 
TEMPERATURE COEFFICIENT OF EXPANSION. 73 
 
 be taken as long as we wish. The wire is hung vertically, 
 the lower end being solidly clamped, and the upper end 
 carrying a sleeve on which rests one leg of a small three- 
 legged bench, on which a mirror is mounted. The other 
 two legs rest on a fixed bracket. The wire is enclosed by a 
 tube through which a current of steam is passed from a 
 boiler and into which two thermometers are thrust to read 
 the temperature. A drainage tube at the lower end allows 
 the escape of water. The wire is prolonged above the mir- 
 ror and is attached to a spring by which the wire is kept 
 stretched. The image in the 
 mirror of a vertical scale is _ __ _____ ---- T 
 
 observed by a reading telescope r"~~ ---- - ------ 
 
 (see p. 25 for adjustments), ^fr-lL^ 
 and the change of reading, d, 
 on the horizontal cross-hairs 
 
 of the telescope, produced by the expansion is noted. Let 
 the length of the wire between the clamp and the support 
 of the bench be /, and let the length of the bench between the 
 point of the movable leg and the line of the other legs be a. 
 Let the distance of the scale from the mirror be L, and the 
 change of temperatures be (t 2 /J . Then, remembering that 
 a ray of light reflected from a mirror turns through twice the 
 angle that the mirror turns through, it is easily seen from 
 the figure that the expansion is ad/2L and the coefficient 
 of expansion is 
 
 ad 
 
 The most difficult quantity to determine with a high degree 
 of precision is a. It may be measured by means of a mi- 
 crometer microscope or a dividing engine (see p. 17). A 
 simpler and more accurate method is to place the optical 
 lever so that the movable leg is on the (vertical) screw of 
 a micrometer caliper (p. 14), while the other legs are on 
 a fixed support and then focus the telescope and scale on 
 the mirror. When the screw is turned the movable leg 
 
74 HEAT. 
 
 is raised a known amount. From this, the distance between 
 the mirror and the scale, and the scale-readings, a is deduced. 
 This calibration may be avoided by placing two legs 
 of the mirror bench upon the collar attached to the wire 
 and resting the third leg directly upon the micrometer screw. 
 The extension may also be measured by Searle's combina- 
 tion of level and micrometer screw (see end of Exp. IX). 
 
 XV. COEFFICIENT OF APPARENT EXPANSION OF A 
 
 LIQUID. 
 
 Edser, Heat, pp. 64-71; Text-book of Physics (Duff), pp. 198-201; 
 Watson's Physics, pp. 211-213; -Awes' General Physics, pp. 233- 
 235; Crew's General Physics, 266, 267. 
 
 The object of this experiment is to determine the coeffi- 
 cient of apparent expansion of some salt solution with refer- 
 ence to glass. A vessel holds M grams of liquid at t and 
 m grams at a higher temperature, t'. Let V be the volume 
 of the vessel at the lower temperature. Since we are 
 considering the apparent expansion, i. e., the expansion 
 with reference to the vessel, we may consider V to be also 
 the volume of the vessel at the higher temperature. The 
 volume of i gram at t is therefore V/M and at *', V/m. 
 The increase in volume is 
 
 m M \ Mm 
 The coefficient of apparent expansion, e, is this apparent 
 expansion divided by the original volume V/M and the 
 range of temperature (t f t) or . 
 
 M-m 
 
 m(t'-t)' 
 
 A glass bulb with a re-curved capillary stem is used. 
 To fill the bulb with a liquid, warm it with the hand or by 
 playing a flame about some distance beneath it. Remove it 
 from the source of heat and plunge the end of the stem into 
 the liquid. As the air in the bulb cools liquid will be drawn in. 
 
COEFFICIENT OF APPARENT EXPANSION OF A LIQUID. 75 
 
 To expel liquid, warm the bulb gently, keeping it so turned 
 that the stem is filled with the liquid ; when the liquid ceases 
 to come out, invert it so that the stem is highest, and allow 
 it to partially cool. Repeat until all the liquid is expelled. 
 
 Clean the bulb by drawing in a little distilled water, or, 
 if the interior be foul, first use chromic acid. Finally rinse 
 the interior with alcohol. Remove the alcohol and dry the 
 interior, if necessary playing a flame about some distance 
 beneath. 
 
 To determine the density of the (cold) salt solution, 
 thoroughly cleanse a tall measuring glass and a suitable 
 hydrometer (variable immersion). Pour enough of the 
 salt solution into the measuring glass to float the hydrometer, 
 read the density, and pour the solution back into the bottle. 
 
 Weigh the bulb very carefully on a sensitive balance 
 (see pp. 21-25). Support the bulb in a clamp stand, clasping 
 the stem between half corks. Fill a small beaker with the 
 salt solution and support it so that the end of the stem dips 
 into the solution. Warm the bulb, playing a Bunsen flame 
 beneath. Never allow the flame for an instant to remain 
 stationary beneath the bulb, and until the bulb contains con- 
 siderable warm liquid, do not allow the flame to touch the bulb, 
 and then only where there is liquid. Alternately warm the 
 bulb and allow it to cool a little until the bulb is filled. 
 When it is partly full 1 it may be best to gently boil the liquid 
 in the bulb. When the bulb is almost full the liquid can be 
 made to expand to fill the entire stem. Then allow it to 
 cool completely while it draws over liquid from the beaker. 
 
 When the bulb is cooled to the temperature of the room, 
 support it in a copper vessel in which water is kept at a 
 constant temperature, a few degrees warmer than the room. 
 When the temperature has been kept constant for five 
 minutes (by the addition of small amounts of hot or cold 
 water, if necessary) and has been frequently stirred, read 
 the temperature (as always estimating tenths). Remove 
 any liquid adhering to the end of the stem, remove the bulb 
 from the bath, dry the exterior, and weigh. Handle the 
 
76 HEAT. 
 
 bulb carefully with a cloth about it so that no liquid may 
 be expelled. Weigh a small, clean, dry beaker. Support 
 the bulb again in the copper bath with the beaker beneath 
 the end of the stem, to catch any liquid expelled. Heat the 
 water in the bath to boiling. When the temperature has 
 been constant for live minutes, read the temperature, catch 
 on the side of the small beaker any liquid adhering to the 
 end of the stem, remove the bulb from the bath, dry the 
 exterior, and weigh. Weigh the small beaker with the 
 liquid contained. Carefully remove the liquid from the 
 bulb and stem as described above. 
 
 The difference between the two weights of the bulb when 
 filled with liquid gives the weight M m of liquid expelled. 
 The difference between the weight of the flask dry and after 
 being in the second bath gives the final weight of liquid in the 
 bulb. The expelled liquid is saved simply as a check and is 
 not used at all if the above difference be slightly greater. 1 
 
 A specific-gravity bottle may be substituted for the bulb, 
 but is not as satisfactory. 
 
 Questions. 
 
 1. Why is double weighing unnecessary? 
 
 2. Why is M m determined more accurately from the difference 
 of the two weighings than from the weight of the liquid expelled. 
 
 3. How might the coefficient of expansion of a solid, attainable 
 only in the form of small lumps, be found by an extension of this 
 method ? 
 
 4. How might the absolute expansion of a liquid be found by the 
 above apparatus? 
 
 XVI. COEFFICIENT OF INCREASE OF PRESSURE OF 
 
 AIR. 
 
 Edser, Heat, pp. 106 in; Poynting and Thomson, Heat, pp. 45-49; 
 Text-book of Physics (Duff), pp. 188, 203-205; Watson's Pracr 
 tical Physics, 78, 79; Watson's Physics, 195198; Ames' 
 General Physics, p. 240; Crew's General Physics, 269. 
 
 If the volume of a mass of gas remains constant while 
 its temperature is raised, its pressure increases according 
 to the law 
 
COEFFICIENT OF INCREASE OF PRESSURE OF AIR. 77 
 
 in which P is the pressure at o C., P the pressure at the 
 temperature t, and a is a constant called the coefficient of 
 increase of pressure. If the pressure were kept constant 
 and the volume allowed to increase, the law of increase of 
 volume would be similar, and it is found that the constant 
 a is practically the same in both cases. 
 
 It is, however, difficult to keep the volume exactly con- 
 stant, for the containing vessel will expand when heated 
 (the volume of the vessel would also increase because of the 
 increase of pressure to which it is subjected, but this may 
 be neglected since it is extremely small). If p is the ob- 
 served pressure at temperature / and P the observed pres- 
 sure at o 
 
 /> = P (i+a'*), 
 
 where a' is the coefficient of apparent increase of pressure 
 (seeExp. XV). 
 
 To correct for the expansion of the vessel, we must 
 suppose the final volume of the gas compressed in the pro- 
 portion in which the capacity of the vessel expanded. The 
 law of expansion of the vessel is 
 
 V = V Q (l +b t), 
 
 where b is the coefficient of cubical expansion of the vessel. 
 
 To get the pressure P that would keep the volume of the 
 
 gas absolutely constant, we must multiply p by (i +6 t), 
 
 .'. P=P (i+a'/)(i+6/) 
 
 And so the true coefficient of increase of pressure a is ob- 
 tained from the apparent coefficient of increase of pres- 
 sure a', by adding the coefficient of cubical expansion of the 
 vessel, or, 
 
 a = a f + b. 
 
 The air (or gas) is enclosed in a bulb to which is con- 
 nected a mercury manometer. The pressure indicated by 
 the manometer is obtained from readings of the mercury 
 levels on a scale between the two columns, or, preferably, 
 with a cathetometer (p. 19). 
 
HEAT. 
 
 If the true increase of pressure of dry air is desired the 
 air must first be carefully dried. To fill the bulb with dry 
 air it may be connected through a drying tube (containing 
 chloride of calcium) with an air-pump and the bulb several 
 times exhausted and refilled with air sucked through the 
 drying tube. (If the bulb be already filled with dry air 
 the process will be unnecessary.) 
 
 The bulb is then connected to the manometer. The 
 bulb is first immersed in a bath of ice and water as nearly 
 as possible at o, and the movable column 
 of the manometer is adjusted until the 
 mercury in the other column is at a definite 
 point, as high as possible without entering 
 the contraction where connection is made 
 with the bulb. The temperature and pres- 
 sure are read as carefully as possible, at 
 least six times, when both have become 
 quite steady, the manometer being read- 
 justed before each reading. 
 
 The bath of ice and water is now replaced 
 by one of water at about 10 and the movable column is 
 readjusted until the mercury in the stationary column is at 
 the former point, that the volume of the gas may remain 
 constant. The temperature and pressure are read when 
 they have become steady. The water is then heated to 
 about 20 and the observations are repeated. Readings are 
 thus made at intervals of about i o until the water boils. 
 
 The pressure and temperature when the water is boiling 
 should be read at least six times, the mercury level in the 
 stationary column being adjusted to the constant point before 
 each reading. It is at the initial temperature (near o) and 
 the final temperature (near 100) that the most reliable 
 observations are obtained, and it is upon these that the 
 most reliable estimate of the coefficient of expansion is 
 founded. (The readings at intermediate temperatures are 
 made in order to test the law of expansion.) If the two arms 
 of the manometer are of different radii, there will be a con- 
 
 FlG. 20. 
 
COEFFICIENT OF INCREASE OF PRESSURE OF AIR. 79 
 
 stant difference of level due to capillarity. This should be 
 read when the bulb is disconnected and allowance should be 
 made for it at other times. Read the barometer (p. 21) and 
 the temperature of the barometer and of the mercury in the 
 manometer. 
 
 Tabulate from your observations (a) the temperatures; 
 (b) the differences in level of the mercury columns; (c) these 
 differences reduced to zero degrees; (d) the pressures as 
 calculated from (c) and the barometer heights (reduced to 
 zero degrees). 
 
 The test of the law of expansion is made by plotting 
 the curve of pressure and temperature, the former as ordi- 
 nates, the latter as abscissas. This should be nearly a 
 straight line. The averages for the first point (o) and the 
 last (about 100) are to be taken as fixing the straight line. 
 The divergence of intermediate points from the straight line, 
 while not sufficient to invalidate the conclusion that the in- 
 crease of pressure is linear, will illustrate the difficulty of 
 keeping the temperature at intermediate points constant 
 for a sufficient length of time for the air in the bulb to come 
 wholly to the temperature of the water. 
 
 Calculate from these two average pressures and tem- 
 peratures, the coefficient of apparent increase of pressure (a') , 
 and, obtaining the coefficient of cubical expansion of the 
 glass (6) from Table VIII, find the true coefficient of increase 
 of pressure (a) . (Remember that the coefficient of cubical 
 expansion is three times the coefficient of linear expansion.) 
 
 If time permit, increase the range of temperature by 
 observations below o in a freezing mixture and above 100 
 in heated oil. 
 
 Questions. 
 
 1 . Why must (a) the air be dry ? (b) a capillary connect the bulb 
 and the manometer ? 
 
 2. What would be the percentage error if the expansion of the 
 bulb was neglected? 
 
8o 
 
 HEAT. 
 
 XVII. PRESSURE OF SATURATED WATER VAPOR. 
 
 Poynting and Thomson, Heat, Chap. X; Edser, Heat, pp. 220-228; 
 Text-book of Physics (Duff), pp. 226-230; Watson's Physics, 
 216-218; Ames' General Physics, pp. 264-269; Crew' s General 
 Physics, 279-281. 
 
 The object of this experiment is to find the pressure of 
 saturated water vapor at different temperatures. By pres- 
 sure of saturated water vapor at a given temperature, or, 
 as it is often called, maximum pressure of water vapor, or, 
 equilibrium pressure, is denoted the pressure of 
 the vapor above water in a closed vessel at the 
 given temperature after a steady state has been 
 reached. A liquid continues to give off vapor 
 from the surface, or, "evaporate," as long as the 
 pressure of the vapor above the liquid is less 
 than the saturated vapor pressure, independent 
 of the total atmospheric pressure above the 
 liquid. After the pressure of the vapor reaches 
 the saturated vapor pressure for that tempera- 
 ture, the total quantity of vapor in the atmos- 
 phere above the liquid remains constant, since 
 for any vapor given off from the surface an equal 
 quantity is condensed. 
 
 There are two chief methods of finding the 
 saturated vapor pressure, the static method and 
 the kinetic method. 
 
 (A) In the static method some water (or other liquid) is 
 introduced into the space at the top of a barometric column 
 which is surrounded by a bath, the temperature of which 
 can be varied. The pressure of the vapor is found by 
 measuring with a cathetometer (p. 19) the height of the 
 mercury column and subtracting this from the barometric 
 reading, each being reduced to zero (p. 21). By varying 
 the temperature of the bath, the vapor pressure at various 
 temperatures is obtained. 
 
 (B) In the kinetic method the quantity measured is the 
 
 FIG. 21. 
 
PRESSURE OF SATURATED WATER VAPOR. 
 
 8l 
 
 temperature of the steam above water boiling under different 
 measured pressures. When a liquid boils, bubbles of vapor 
 are formed throughout the interior of the liquid. In forming 
 these bubbles, the vapor overcomes the pressure of the 
 atmosphere above the liquid, therefore the pressure of the 
 vapor must equal the atmospheric pressure, and obviously 
 the vapor in the bubbles is saturated. Hence, in measuring 
 the atmospheric pressure above a liquid boiling at a known 
 temperature, we find the saturated vapor pressure of the 
 liquid at this temperature, and this is Regnault's method, 
 which method is followed in this experiment. 
 
 FIG. 22. 
 
 In Regnault's apparatus the total pressure above the 
 surface of the liquid can be kept very constant. As the 
 liquid is heated, the vapor is condensed in a Liebig con- 
 denser, and as the pressure of vapor distributed through 
 several conducting vessels is the vapor pressure correspond- 
 ing to the vessel at lowest temperature, the pressure exerted 
 by the vapor cannot exceed the maximum pressure corre- 
 sponding to the temperature of the tap water, and is there- 
 fore very small. As the temperature of the boiler changes, 
 the temperature of the air in the boiler varies, but a large 
 air reservoir surrounded by water is connected between the 
 condenser and the manometer and air-pump or aspirator, 
 which makes the volume of the air in the boiler small com- 
 6 
 
82 HEAT. 
 
 pared with the total volume of air in the system, and thus 
 the increase of pressure due to the heating of the air in the 
 boiler is small. 
 
 The boiler should be about two-thirds full of water. Fill 
 with water the small tube running down into the boiler 
 (which tube is closed at the bottom), and insert in this 
 tube through a cork one of the thermometers tested by the 
 observer. Draw out any water which may be in the air 
 reservoir by means of the stopper underneath. Fill the 
 surrounding vessel with water. (Rubber stoppers should 
 be lubricated with rubber grease (note p. 32) before inser- 
 tion.) Exhaust the air from the system to the highest 
 vacuum attainable by means of a Geryk pump or aspirator. 
 Close all the cocks through which connection is made 
 to the aspirator and let the system stand a few minutes to 
 see if there is any leakage. If not, start a gentle stream of 
 water through the condenser, and place a Bunsen flame 
 under the boiler. Read the barometer and its temperature 
 (see p. 21). 
 
 When the temperature as registered by the thermometer 
 in the boiler becomes very steady, read it, and at once 
 record the two extremities of the mercury column of the 
 manometer. Let in a little air by first opening and then 
 closing a cock near the air-pump, and then opening and 
 closing a cock nearer the apparatus. Increase the pressure 
 at first by about 15 mm., gradually increasing the steps 
 and when near atmospheric pressure change the pressure 
 by about 1 2 cm. The reason for the difference in pressure 
 in the steps is that it is better to have the steps represent 
 about equal changes of temperature, for instance, about 5. 
 
 From the corrected barometer reading and the differences- 
 in height of the mercury columns, calculate the pressures. 
 Tabulate pressures and temperatures and also plot them, 
 making temperatures abscissas and pressures ordinates. 
 
 Ramsay and Young's method for measuring the vapor 
 pressure of a small quantity of liquid is described in Wat- 
 son's Practical Physics, 94. 
 
HYGROMETRY. 83 
 
 Questions. 
 
 1. State precisely what two quantities you have observed in the 
 second method and what relation they bear to the pressure and temp- 
 erature of saturated vapor. 
 
 2. What condition determines whether a liquid will boil or evapo- 
 rate at a given temperature? 
 
 3. What was the actual vapor pressure above the boiling liquid? 
 (Table XIII.) 
 
 4. What determines (a) the lowest temperature, (b) the highest 
 temperature for which this apparatus is applicable? 
 
 XVIII. HYGROMETRY. 
 
 Poynting and Thomson, Heat, pp. 209-215; Davis, Elementary Me- 
 teorology, Chap. VIII; Robson, Heat, 7375; Watson's Prac- 
 tical Physics, 9597; Text-book of Physics (Duff), pp. 239242; 
 Watson's Physics, 220, 221; Ames 1 General Physics, pp. 265- 
 268. 
 
 Three methods will be used for studying the hygrometric 
 state of the atmosphere. The first method (A) determines 
 the dew-point, the second (B) determines, indirectly, the 
 actual vapor pressure, and the third (C) determines the 
 relative humidity. 
 
 (A) Regnanlt's Hygrometer. A thin silvered-glass test- 
 tube is half-filled with ether. The test-tube is tightly 
 closed by a cork through which passes a sensitive ther- 
 mometer which gives the temperature of the ether. Two 
 glass tubes also pass through the cork, one extending to the 
 bottom, the other ending below the cork. An aspirator 
 gently draws air from the shorter tube. The ether is 
 evaporated by the air bubbles and the entire vessel cools. 
 The silvered surface and the thermometer are watched 
 through a telescope and the temperature is read the moment 
 moisture appears on the metal. The air current is stopped 
 and the temperature of disappearance of the moisture is 
 observed. This is repeated several times and the mean 
 is taken as the dew-point. The detection of moisture is 
 facilitated by observing at the same time a similar piece 
 of silvered glass which covers a part of the test-tube, but 
 which is insulated from it. The temperature of the air 
 
84 HEAT. 
 
 should also be carefully determined, preferably with a ther- 
 mometer in a similar apparatus where there is no evapora - 
 tion. 
 
 An arrangement of two small mirrors at right angles 
 so placed as to reflect light from the two tubes into the 
 telescope will facilitate the comparison. 
 
 (B) Wet and Dry Bulb Hygrometer. Two thermometers 
 are mounted a few inches apart. About the bulb of one is 
 wrapped muslin cloth to which is attached a muslin wick 
 dipping in water. The other is bare. The temperatures 
 of both are read when they have become steady. The 
 temperature of the first thermometer will be lower than that 
 of the bare thermometer, on account of the evaporation of 
 the water. From the difference of temperature of the two 
 thermometers and the temperature of the bare thermometer 
 the actual vapor pressure may be determined with the aid 
 of empirical tables (see Table XV). For more accurate 
 apparatus, see references. 
 
 (C) Chemical Hygrometer. Fill three ordinary balance 
 drying vessels with pumice. Saturate two with strong 
 sulphuric acid and the third with distilled water. Weigh 
 very carefully the two which have the acid and then connect 
 them to an aspirator, with the water absorption vessel 
 between them. After a gentle stream of air has passed 
 through for a considerable time, disconnect and weigh the 
 sulphuric acid vessels. The ratio of the gains in weight 
 will obviously be the relative humidity. Observe also the 
 temperature of the air. 
 
 If not directly determined, calculate from your observa- 
 tion, by each of the three methods, the dew-point, the 
 actual vapor pressure, the relative humidity, and the amount 
 of moisture in the atmosphere per cubic meter. Tabulate 
 your results. Table XIII gives the vapor pressures of water 
 at different temperatures. 
 
SPECIFIC HEAT BY THE METHOD OF MIXTURE. 85 
 
 XIX. SPECIFIC HEAT BY THE METHOD OF MIXTURE. 
 
 Edser, Heat, pp. 122-136; Text-book of Physics (Duff), pp. 208211; 
 Watson's Physics, 200-201; Watson's Practical Physics, 82- 
 84; Ames' General Physics, pp. 250-252; Crew's General Physics, 
 252. 
 
 The specfic heat of a substance is the number of calories 
 required to raise the temperature of one gram of the 
 substance one degree centigrade, or the number of calories 
 given up by one gram in cooling one degree centigrade. 
 In the method of mixture a known mass (M) of the sub- 
 stance, heated to a known temperature (T), is immersed in a 
 known mass of liquid (m) of known specific heat (for water 
 = i), at a known temperature (/ ), and the unknown mean 
 specific heat (x) of the substance is deduced from these data 
 and the temperature (t) to which the mixture rises. Water 
 is the liquid employed unless there would be a chemical 
 reaction on immersion. 
 
 The liquid must be contained in a vessel which is also 
 heated by the immersion of the hot body. The heating of 
 the vessel is equivalent to the heating of a certain addi- 
 tional quantity of water. This equivalent quantity of water 
 (e) is called the water equivalent of the vessel. It is practi- 
 cally equal to the mass of the vessel (mj multiplied by the 
 specific heat (5) of the material of the vessel. Theoretically 
 it may be obtained by noting the temperature of the vessel 
 and pouring into it a known mass of water at a known 
 temperature and noting the final temperature. This is an 
 inverted form of the method of mixture applied to finding 
 the specific heat of the vessel. But as we shall presently 
 see, it is the method of mixture applied under very unfavor- 
 able conditions and will not usually give a very satisfactory 
 result. Another method will be recommended below. 
 
 The equation for finding the specific heat is obtained by 
 equating the heat given up by the hot body to that taken up 
 by the water and containing vessel. Hence 
 
 Mx(T-t)=(m+e)(t-t ). (i) 
 
86 HEAT. 
 
 Sources of Error. 
 
 (1) Loss of heat while the hot body is being transferred 
 to the water. 
 
 (2) Loss of heat by radiation, conduction, or evapora- 
 tion while the mixture is assuming a uniform temperature. 
 
 (3) Errors in ascertaining the true temperature includ- 
 ing errors in the thermometers. 
 
 Choice of Best Conditions. As the accuracy of this de- 
 termination depends largely on the selection of suitable 
 conditions, we shall consider how these may be chosen so 
 that unavoidable errors in the separate measurements may 
 affect the result as little as possible. 
 
 By taking the logarithms of both sides of (i) and dif- 
 ferentiating partially, we obtain (see pp. 7, 8.) 
 
 (2) (3) (4) 
 
 f&cl _8M [~S*~| Jm_ |~8*1 Be 
 
 [He \M~~W [Hc\m~ m +e [x\e~m+e 
 
 (5) (6) " (7) 
 
 M * T [H * ry (T-w 
 
 [x\T = T-t' UJ'o *-* ' Up (T-t)(t-t Y 
 
 The left-hand side of (2) stands for "the ratio that the 
 possible error (8#) in x, due to the possible error (8M) in 
 M, bears to x," and so for the other equations. 
 
 M and m can be measured with great precision; hence 
 (2) and (3) are negligible. From (4) it is seen that the 
 water equivalent of the calorimeter must be found with 
 some care. From (5) and (6) it is seen that* the ranges 
 T t and t t Q should be as great as possible (see, how- 
 ever, "sources of error" above). This is also consistent 
 with the indications of (7), for although (T t ) enters 
 the numerator, the product of T t and t t Q is in the 
 denominator. Moreover, it is seen from (5), (6), and (7) 
 that if equal errors are made in observing T, t, and t , the 
 effect of the error in / may equal the sum of the effects of 
 the errors in T and t . Hence the necessity of determining 
 t with special care. But, allowing an unavoidable error in 
 
SPECIFIC HEAT BY THE METHOD OF MIXTURE. 87 
 
 t, how can its effect be made as small as possible by prop- 
 erly choosing the quantities, M, m, t , Tf Let us suppose 
 T / is taken as great as possible under the circumstances. 
 How can (T t)X (* * ) be made as great as possible? The 
 sum of T t and / 1 is T t Q , a fixed quantity. Hence their 
 product is, by algebra, a maximum when they are equal or 
 t is midway between T and t Q . But it is seen from (i) that 
 this also requires MX and m + e to be equal. Hence we see 
 that for the best results, T and t should be as far apart as 
 possible and the heat capacity of the specimens should be as 
 nearly as possible equal to the heat capacity of the water and 
 the vessel that contains it. 
 
 The logical procedure, then, would be to roughly de- 
 termine x by the method of mixture, using any convenient 
 values of M and m, and with this rough value for x, calcu- 
 late what ratio of M to m would best satisfy the above con- 
 dition. Moreover, it is seen from (4) above that m should 
 be as large as is consistent with other conditions. Then we 
 should proceed to arrange a new experiment to be per- 
 formed under the more favorable conditions for precision. 
 
 We now see why it is not easy to determine the water 
 equivalent of the vessel directly. Its heat capacity is small 
 compared with that of the water that would fill it, and so 
 the change of the temperature of the water would be small 
 and difficult to determine accurately. If a much smaller 
 quantity of water were used, a large part of the surface of 
 the vessel would be left uncovered, and its temperature 
 could not be determined. Hence it is better to determine 
 the specific heat of the material of the vessel, using the ordi- 
 nary method of mixture and a mass of the same material as 
 the vessel. Then multiplying the mass of the vessel by its 
 specific heat, we have its water equivalent. 
 
 It is desirable that the body should have such a form 
 that it and the water in which it is immersed should rapidly 
 come to a common temperature. Filings, shot, thin strips, 
 wire or small pieces would best satisfy this condition. 
 Larger solid masses are more rapidly (and therefore with 
 
88 HEAT. 
 
 less loss of temperature) transferred from the heater to the 
 water, and to give the water ready access to them they may 
 be perforated with holes, through which, by moving the 
 mass up and down in the water, the water may be made to 
 circulate. The following directions apply primarily to this 
 latter form, but may be readily adapted to the other forms. 
 
 Two forms of heater will be here described, (i) The 
 steam heater. A copper tube large enough to admit the 
 specimen is enclosed, except at the ends, by an outer copper 
 vessel which is to act as a steam jacket to the inner vessel. 
 Steam from a simple form of boiler is admitted to the 
 jacket through a tube near the top of the jacket and escapes 
 through an outlet near the bottom. If the body to be 
 heated is a solid mass, it is suspended in the heater by a long 
 string that passes through a cork that closes the upper 
 end of the heater. (If the specimen is in the form of shot 
 or clippings they are placed in a dipper that fits into the 
 heater.) A thermometer passed through the cork or cover 
 of the dipper is pushed down until it comes into contact with 
 the body tested. The lower end of the heater is also closed 
 with a cork. 
 
 Such a steam heater ultimately brings the specimen to a 
 very steady temperature, but it has the disadvantage of 
 heating very gradually. If the boiler which supplies steam 
 to the jacket has a closed tube extending from the top into 
 the interior of the boiler, of slightly larger diameter than 
 the specimen or dipper, either of the latter may be placed 
 therein and rapidly heated to about the steam temperature, 
 when they are transferred to the steam heater. 
 
 (2) The Electric Heater. A metallic tube is heated by 
 a strong current of electricity passing through a coil of wire 
 of high resistance that surrounds the tube. The current 
 can be varied by changing a variable resistance in circuit 
 with the heating coil. With an alternating current the 
 resistance may be an inductive resistance or choking coil 
 consisting of wire surrounding a soft-iron wire core. A low 
 resistance allowing a high current is used until the tern- 
 
SPECIFIC HEAT BY THE METHOD OF MIXTURE. 89 
 
 perature rises to the desired point (perhaps near 100) and 
 then the current is reduced to the strength that will keep 
 the temperature constant, as indicated by a thermometer 
 hung in the heater. The body is introduced into the heater 
 exactly as in the case of the steam heater. The proper 
 method of varying the resistance can only be learned by 
 some practice. 
 
 The two thermometers used should be those for which 
 tables of corrections have been obtained earlier. 
 
 The calorimeter may be prepared while the specimen is 
 being heated. It consists of a smaller copper or aluminum 
 can highly polished on the outside and enclosed in a larger 
 one brightly polished on the inside, but well insulated from 
 it by corks or cotton-wool. A wooden cover fits over both 
 vessels and has holes for thermometer and stirrer and an 
 opening giving access to the interior vessel. A convenient 
 form has a trap-door which slides open in two halves, 
 exposing the entire inner vessel. A screen with sliding or, 
 preferably, double swinging doors, should separate the 
 calorimeter from the boiler, heater, etc. 
 
 The "water equivalent" of the receiving vessel means 
 the water equivalent of the inner vessel together with the 
 stirrer, if one be used. It is advisable that the stirrer 
 should also be of copper or aluminum. (A stirrer is, how- 
 ever, not necessary when the specimen is in the form of one 
 large block) . 
 
 At certain times, in the manipulation of this experiment 
 the co-operation of two persons is desirable, and, for 
 economy of time, two determinations should be made 
 simultaneously, two heaters, two specimens, and two 
 calorimeters being used. One specimen should peferably 
 be of the same material as the calorimeter, so that the water 
 equivalent may be determined. 
 
 The body whose specific heat is to be determined is 
 weighed to o . i gm. and placed in the heater along with a 
 thermometer. The inner vessel of the calorimeter (including 
 the stirrer) is weighed to o.i gm. Water near the tern- 
 
90 HEAT. 
 
 perature of the room is poured into it until it is judged that 
 when the hot body is immersed it will be completely covered 
 and the water will rise to within a couple of centimeters of 
 the top of the vessel. The vessel and water are then 
 weighed to o . i gm. The inner vessel is now replaced in the 
 outer and the cover adjusted and closed. 
 
 When the temperature of the specimen has remained 
 constant for ten minutes, it may be assumed that the hot 
 body is throughout at the temperature of the heater. The 
 next steps require two persons, and as it is important that it 
 should be carried out promptly and neatly, it should be 
 carefully considered before being performed. One person 
 constantly stirs the water in the calorimeter and reads the 
 temperature, to tenths of a degree, every minute for five 
 minutes. He then opens the cover and slides the calorimeter 
 beneath the heater. The other observer has meanwhile 
 made a careful final observation of the temperature of the 
 heater and removed the lower stopper. As soon as the 
 calorimeter is in position, he lowers the hot body, without 
 splash, into the water. The calorimeter must then be 
 immediately removed and the cover closed. 
 
 One observer should then keep the mixture stirred by 
 moving the body up and down with the aid of the string 
 and note the temperature at as short equal intervals as 
 possible (perhaps every 15 sees.) while the other records 
 the readings. After the highest temperature has been 
 reached, the readings are continued every minute for at 
 least five minutes. 
 
 Simple and obvious modifications of the above procedure 
 are required if the specimen is in the form of shot or clippings. 
 
 After rough calculations of water equivalents and specific 
 heats, the observers should exchange duties and repeat 
 the whole, using masses that most nearly accord with the 
 conditions laid down in the considerations stated above. 
 Before calculating final results, make corrections of the 
 temperatures, T, t, t , according to the tables of corrections 
 already obtained for the thermometers used. 
 
RATIO OF SPECIFIC HEATS OF GASES. 91 
 
 Plot all the temperature observations of the final deter- 
 mination, and correct for radiation according to the direc- 
 tions given on pages 63, 64. The possible error of the result 
 should be calculated as explained on pages 7, 8.) 
 
 Questions. 
 
 1. How might the present method be adapted to find the specific 
 heat of a liquid? 
 
 2. Considering evaporation, loss of heat when transferring the 
 hot body, and any other sources of error that may occur to you, is 
 your result more probably too high or too low ? 
 
 XX. RATIO OF SPECIFIC HEATS OF GASES. 
 (Clement and Desormes* Method.) 
 
 Edser, Heat, pp. 3 1 7-3 2 5 ; Text-book of Physics (Duff) , pp. 211-212, 
 267, 268; Watson's Physics, 259-260; Watson's Practical 
 Physics, 105; Ames' General Physics, pp. 252-256. 
 
 The gas is compressed into a vessel until the pressure has a 
 value which we will designate by p v The vessel is then 
 opened for an instant, and the gas rushes out until the 
 pressure inside falls to the atmospheric pressure, p Q . This 
 expansion may be made so sudden that it is practically 
 adiabatic and the temperature of the gas will therefore 
 fall. After the vessel has been closed for a few minutes, the 
 gas will have warmed to the room temperature, /, and the 
 pressure, p 2 , will be above that of the atmosphere. Consider 
 one gram of the gas. During the adiabatic expansion, its 
 volume changed from v l to v 2 , according to the adiabatic 
 equation for pressure and volume (see references) 
 
 W Po 
 
 Since the initial and final temperatures are the same, and 
 since the volume remains v 2 while the gas is warming and 
 the pressure is rising from p to p 2 , by Boyle's Law 
 
92 HEAT.. 
 
 Hence f, the ratio of specific heats, is given by the equation 
 
 A large carboy is mounted in a wooden case and may be 
 surrounded with cotton batting. The neck is closed with 
 a rubber stopper through which passes a T-tube connected 
 on one side with a compression pump (e. g., a bicycle pump), 
 
 FIG. 23. 
 
 and on the other side with^a manometer containing castor 
 oil.* A large glass tube, which may be closed by a rubber 
 stopper, also passes through this large stopper. A little 
 sulphuric acid in the bottom of the carboy keeps the air 
 dry. A very fine copper wire and a very fine constantin wire 
 
 * The density of castor oil is about .97, but it should properly be deter- 
 mined (Exp. VIII). 
 
RATIO OF SPECIFIC HEATS OF GASES. 93 
 
 pass tightly through minute holes in the stopper and meet at 
 the center of the carboy, in a minute drop of solder. 
 
 The air in the carboy is compressed until the difference in 
 pressure is about 40 cm. of oil (=Pi po). The tube connect- 
 ing with the pump is closed, and, after waiting about 15 
 minutes to allow the air inside to regain its initial tempera- 
 ture (as shown by the pressure becoming constant), the ends 
 of the oil column are carefully read. The carboy is now 
 carefully surrounded with cotton batting, which may have 
 been removed to facilitate cooling. The air inside is momen- 
 tarily allowed to return to atmospheric pressure by removing, 
 for about one second, the rubber stopper from the glass tube. 
 After waiting until the air inside has assumed the room 
 temperature (shown by the pressure becoming constant), 
 the final pressure p 2 is determined. The cotton-wool had 
 better be removed during this stage. 
 
 Connect the wires to a calibrated galvanometer (Exp. 
 LVIII), apply the initial compression p lt and observe the 
 reading of the galvanometer when it has become steady. 
 Remove the stopper as before (for not over one second) , re- 
 place the stopper, and observe the galvanometer reading. 
 The proper reading to record is the fairly steady deflection 
 which is attained immediately after the stopper is removed. 
 There are liable to be rapid fluctuations which should be 
 disregarded, and of course the temperature does not long 
 remain steady, owing to heating or cooling from the outside. 
 Record as before the final pressure p 2 . Repeat several times, 
 starting with the same initial pressure p v Record the 
 temperature of the room, t, and p , the height of the baro- 
 meter (p. 21). 
 
 Calculate y, the ratio of specific heats by the above equation . 
 Calculate the change of temperature from the mean of the 
 galvanometer deflections and the constants of the thermo- 
 couple and galvanometer. 
 
 Compare the result with T l T where T l is + 273 and 
 r o is calculated from the adiabatic equation for temperature 
 and pressure (see references). 
 
94 HEAT. 
 
 Unless exceedingly fine wire is employed (preferably 
 No. 40, B. & S.), the heat capacity of the wire is relatively 
 so great that the thermocouple will not show the full change 
 of temperature. 
 
 Draw a curve with volumes as abscissae, and pressures 
 as ordinates, which will represent the changes in this 
 experiment. 
 
 (Let specific volumes, i. e., volumes of one gram, be abscissae. 
 Calculate from Table VI and the laws of gases the specific volumes 
 corresponding to the room temperature and p , p lt and p 2 , and draw 
 the corresponding isothermal. Draw the horizontal line correspond- 
 ing to p . Draw a vertical through the point corresponding to p ? on 
 the above isothermal. The intersection of these two straight lines 
 will evidently be p , v 2 ). 
 
 Questions. 
 
 1. Do you see any objection to an initial exhaustion of the gas 
 in place of the compression? 
 
 2 . What are the advantages and disadvantages of a large opening ? 
 Short time of opening? Castor oil manometer? 
 
 3. How would an aneroid manometer be preferable in this experi- 
 ment to a liquid manometer? 
 
 XXI. LATENT HEAT OF FUSION. 
 
 Edser, Heat, pp. 145-149; Text-book of Physics (Duff), p. 225; 
 Watson's Physics, 211; Watson's Practical Physics, 88; 
 Ames' General Physics, pp. 260, 261; Crew's General Physics, 
 286. 
 
 The latent heat of fusion of a substance is the number of 
 calories required to melt one gram of the substance. The 
 most common method of measuring it is a method of mixture 
 similar to that used in finding the specific heat of a solid. 
 A known mass of the solid at its melting-point is placed in a 
 known mass of the liquid at a known temperature, and the 
 temperature of the liquid observed after the solid has com- 
 pletely melted. Allowance must be made for the water 
 equivalent of the calorimeter and correction must be made 
 
LATENT HEAT OF FUSION. 95 
 
 for the effect of radiation to or from the calorimeter while 
 melting is taking place. The error due to radiation may be 
 made small by having the liquid initially as much above 
 the temperature of its surroundings as finally it falls below. 
 Thus loss and gain by radiation will approximately balance. 
 Nevertheless, since the calorimeter will probably not be the 
 same length of time above the temperature of the sur- 
 roundings as below, there will be a residual error for which 
 correction must be made. 
 
 The calorimeter consists as usual of an inner can polished 
 on the outside to diminish radiation, and enclosed in an 
 outer can polished on the inside. The space between the 
 two cans may be filled by cotton-wool to prevent air currents, 
 and still further prevent communication of heat. The inner 
 can is weighed, first empty and then half-filled with warm 
 water about 1 5 above the room temperature. It is then 
 placed in the outer can as described above and covered by a 
 wooden cover having holes for thermometer and stirrer 
 and a hinged cover giving access to the inner vessel. 
 
 The temperature is carefully noted each minute until it 
 has fallen to about 10 above the room temperature. In the 
 meantime, ice is broken to pieces of about a cubic centimeter 
 in volume. These pieces are carefully dried in filter paper. 
 A careful observation of the temperature of the water in 
 the calorimeter having been made and the time noted, a 
 piece of ice is dropped in without splashing and kept under 
 water by a piece of wire gauze attached to the stirrer. 
 The temperature is noted every half-minute as the ice melts, 
 the water meantime being kept stirred. The rate at which 
 ice is dropped in is regulated simply by the rate at which 
 it can be dried and the temperature and time noted. 
 
 The process is continued until the temperature has fallen 
 to about 10 below the room temperature. Then the 
 addition of ice is discontinued and the temperature of the 
 water further noted every minute for four or five minutes. 
 Finally, the weight of the inner can and its contents is 
 obtained in order that the mass of the ice may be deduced. 
 
96 HEAT. 
 
 After the proper weight of ice has been ascertained, the 
 experiment should be repeated with a single piece of approxi- 
 mately this weight. As there are considerable sources of 
 error that cannot be eliminated, the whole determination 
 should be repeated as often as time will permit. 
 
 All the temperature observations should be plotted 
 against the time and the radiation correction determined 
 as described on pages 63 and 64. In reporting, consider 
 the possible error of your result so far as it depends on the 
 possible error of your weighings and observations of tem- 
 perature. State also any other sources of error that may 
 have affected your result. 
 
 Questions. 
 
 1 . What advantages are there in the use of one large lump over an 
 equal mass of small ones? 
 
 2. Why must the water in the inner vessel be pure? 
 
 3. Is it preferable to have the air about the calorimeter moist or 
 dry? Explain. 
 
 XXII. LATENT HEAT OF VAPORIZATION. 
 
 Ames' General Physics, p. 269; Watson's General Physics, 214; Crew's 
 Physics, 287; Watson, Practical Physics, 89-91; Edser, Heat, 
 pp. 150-9; Text-book of Physics (Duff), p. 231. 
 
 The latent heat of vaporization of a substance is the number 
 of calories required to change one gram of the substance 
 from liquid to vapor. The usual method of measuring 
 it is a method of mixture. A known mass of vapor, at a 
 known temperature, is discharged into a known mass of 
 liquid, at a known initial temperature, and the final tem- 
 perature is noted. The same precautions are necessary 
 as in finding the latent heat of fusion. The arrangement 
 of the calorimeter is also the same. To minimize radiation 
 the water should be initially as much below room tempera- 
 ture as it finally rises above, say 15. The initial rate of 
 warming should also be obtained, and also the final rate of 
 cooling. 
 
LATENT HEAT OF VAPORIZATION. 
 
 97 
 
 Several different forms of boiler have been devised for 
 the purposes of this determination. Two will be briefly 
 described. 
 
 In Berthelot's boiler the delivery tube passes out through 
 the bottom of the boiler, which is heated by a ring burner 
 that surrounds the tube. Thus the tube is so far as possible 
 jacketed by the boiling water. The usual form of this 
 
 boiler is somewhat fragile, but a 
 good substitute may be made 
 from a round-bottomed boiling 
 flask the neck of which has been 
 shortened (Fig. 24). 
 
 FIG. 24. 
 
 FIG. 25. 
 
 In the electrically heated boiler the heating of the water 
 is produced by a coil of wire that is immersed in the water 
 and is heated by a current of electricity. The current must 
 be kept regulated by a rheostat, so that boiling proceeds at 
 a moderate rate. 
 
 The chief difficulty is in delivering the steam dry. Con- 
 densation is apt to take place in the delivery tube. This 
 can be reduced by inserting a trap in the delivery tube be- 
 tween the boiler and the calorimeter. The trap should, 
 from time to time, be cautiously heated by a Bunsen burner 
 to prevent condensation, but in general, it is better to dis- 
 pense with the trap and make the exposed part of the 
 7 
 
98 HEAT. 
 
 delivery tube as short as possible and carefully cover it with 
 cotton-wool. 
 
 If the delivery tube simply passed to a sufficient depth 
 beneath the water, the steam would be delivered at greater 
 than atmospheric pressure, as the pressure of a certain depth 
 of water would have to be overcome. Hence it is better 
 to let the delivery tube pass into a condensing-box im- 
 mersed in the water. The latter must also be open to the 
 atmosphere by another tube. To prevent any escape of steam 
 by this tube it may be closed by a little cotton-wool. The 
 amount of steam that has been condensed is obtained by 
 weighing the condensing-box (well dried) before it is placed 
 in the calorimeter, and again with the contained water 
 at the end of the experiment. The temperature of the steam 
 is deduced from the barometric pressure. A pressure gauge 
 attached to the boiler affords a means of estimating how far 
 the pressure differs from atmospheric pressure. 
 
 For the best results, certain precautions must be ob- 
 served. The delivery tube must not be connected to the 
 coridensing-box until steam has begun to pass freely, and 
 as dry as possible, from the tube. Connection should not 
 be attempted until the temperature of the water has been 
 carefully ascertained and care has been taken that every- 
 thing is ready for making a deft and prompt connection. 
 After the temperature of the well stirred water in the inner 
 calorimeter has been read every minute for five minutes, the 
 connection is made. The temperature is read every half- 
 minute, the water meantime being kept well stirred by a 
 stirrer (which should be of the same material as the calor- 
 imeter and condenser in order to simplify the calculation of 
 the water equivalent). The flame of the ring-burner must 
 be regulated so that the steam does not pass too rapidly. This 
 may be gauged by the rate of the rise of the temperature 
 of the calorimeter, which should not exceed 4 or 5 per 
 minute. In finding the subsequent rate of cooling, the 
 boiler should be disconnected from the condenser and the 
 tube leading to the condenser should be closed by plugs of 
 
LATENT HEAT OF VAPORIZATION. 99 
 
 cotton-wool to prevent evaporation; but in subsequently 
 weighing the condenser the wool should not be included. 
 The whole determination should be repeated as many times 
 as possible. 
 
 A formula for the calculation of the latent heat may 
 be readily worked out. Account must be taken of the 
 water equivalent of calorimeter, condenser, and stirrer. The 
 correction for radiation is made by the method stated 
 on pages 63, 64. 
 
 The possible error of the result, so far as it depends on 
 the readings made, should be calculated, and other possible 
 sources of error should be mentioned. 
 
 Questions. 
 
 1 . State the advantages and disadvantages of a rapid flow of steam 
 
 2. Explain why the latent heat should vary with the atmospheric 
 pressure. 
 
 3. Must the boiling water be pure? Explain. 
 
 XXIII. LATENT HEAT OF VAPORIZATION. 
 
 Continuous -flow Method. 
 
 Ames' General Physics, p. 269; Watson s Physics, 214; Crew 1 s General 
 Physics, 287; Watson, Practical Physics, 89-91; Edser, Heat, 
 pp. 150-9; Text-book of Physics (Duff}, p. 231. 
 
 The apparatus for this method may be readily constructed 
 from a Liebig's condenser. Water enters at D and leaves at 
 C through T-tubes connected to the condenser by short 
 rubber tubes. Superheated steam enters at A through a 
 T-tube and the condensed water drops into a covered beaker 
 E. The steam is superheated as it flows through a glass 
 tube FE. This is first covered with asbestos over which a 
 heating coil of wire* is wrapped, the coil being covered by 
 a second layer of asbestos. AB and FE are mounted on 
 a wooden frame and A B is thickly covered with cotton- 
 wool to prevent radiation. Thermometers 7\, T 3 , T 2 , T 4 , 
 
 * "Nichrome" wire (supplied by the Driver-Harris Co., New York) is very 
 suitable. 
 
IOO 
 
 HEAT. 
 
 give the respective temperatures of the superheated steam, 
 the outflowing water, the inflowing water, and the water of 
 condensation. The supply of water may come from the 
 water mains, if this is sufficiently constant in temperature. 
 
 FIG. 26. 
 
 It is, however, much better to have a supply of from 5 to 
 lo gallons in an elevated tank and keep the flow constant by 
 an overflow regulator as indicated in figure 29 (Exp. 
 XXVIII). 
 
 The boiler to supply the steam should be large enough to 
 allow of a flow for two hours without refilling (one to two 
 
LATENT HEAT OF VAPORIZATION. IOI 
 
 liters will suffice). The current in the superheating coil 
 should be regulated by a rheostat so that the superheated 
 steam is at about 105. Some time should be spent in test- 
 ing adjustments to obtain a suitable current and a rate of 
 flow of water that will give a rise of temperature of about 
 20. The tank should be connected with the water service 
 so that it can be readily filled. The water as it comes 
 from the mains will probably be below room temperature 
 and this is an advantage, since with a suitable rate of flow 
 of the steam the water that drops into E will differ but little 
 from room temperature and will suffer little loss of heat by 
 radiation. This will require a proper regulation of the 
 burner that heats the boiler. The burner should be sur- 
 rounded by a shield of sheet-iron or asbestos to prevent 
 fluctuations caused by air-currents. 
 
 The thermometers 7\, T 2 , and T 3 should be read once a 
 minute (e. g., T 2 20 sec. after T l and T 3 20 sec. after T 2 ) . From 
 the mean of each of these readings, the temperature of E, 
 the mass of water that flows out at C, and the mass of the 
 water that drops into E, the latent heat can be calculated. 
 The specific heat of the superheated steam may be taken as 
 0.5. A formula can be readily constructed to express the 
 fact that the heat given up by the steam and condensed water 
 equals the heat carried off by the current of water. 
 
 Questions. 
 
 1 . Why does not the water equivalent of the condenser need to be 
 considered ? 
 
 2. How could you find the amount of error due to conduction of 
 heat from the superheater to the water in the condenser? 
 
 3. How could you find the amount of error due to radiation from 
 the condenser? 
 
 4. What other sources of possible error are there in this method? 
 
IO2 
 
 HEAT. 
 
 XXIV. THERMAL CONDUCTIVITY. 
 
 Edser, Heat, pp. 416-430; Watson, Practical Physics, 106, 107; 
 Text-book of Physics (Duff), pp. 216-220; Watson's Physics, 
 236238; Ames' General Physics, p. 288; Crew's General Physics, 
 254, 256. 
 
 The thermal conductivity of a substance is the amount 
 of heat transmitted per second per unit of area through a 
 plate of the substance of unit thickness, the temperature of 
 the two sides differing by i and the flow having become 
 steady. If K be the thermal conductivity, and if a plate 
 of thick-ness / and area A be kept with one side at a tem- 
 perature t, and the other at a lower temperature, t', the 
 number of calories that will flow through the plate in time 
 7, after the flow has become steady, will be 
 
 KA(t-t')T 
 
 I 
 
 whence K can be derived if the other quantities are observed 
 or measured. 
 
 Thermal conductivity is in general difficult to measure 
 satisfactorily. The following very simple method cannot 
 
 be relied on to closer than a 
 few per cent., but it only re- 
 quires a small portion of the 
 time that the more accurate 
 methods call for. 
 
 A rod or wire of the substance 
 to be tested is inserted at one 
 end into a heavy block of metal, 
 which is heated to a constant 
 high temperature in a bath, 
 through the bottom of which 
 the rod passes. At its lower 
 end the rod is screwed into a heavy block of brass or 
 copper of mass M and specific heat s, which is initially at a 
 very low temperature. Heat is thus conducted by the rod 
 from the bath to the lower block. If the latter neither lost 
 
 FIG. 
 
THERMAL CONDUCTIVITY. 103 
 
 nor gained heat by convection or radiation, and if there 
 were no losses from the sides of the rod, we could calculate 
 the conductivity of the rod from its dimensions and the 
 mass, specific heat, and rise of temperature of the lower 
 block. The loss of heat from the surface of the rod is 
 almost wholly prevented by enclosing it by a glass tube, 
 which does not come into direct contact with the rod, and 
 wrapping the glass tube with cotton wool and paper. 
 
 To allow for radiation or convection to or from the lower 
 block the experiment is modified as follows: The block 
 is enclosed in a vessel surrounded by a water-jacket, through 
 which water at a constant temperature, t', circulates. Now, 
 the rate at which the lower block receives heat through the 
 rod, when the former is at the temperature t f , is the mean 
 of the rates at which it receives heat when it is n degrees 
 below t', and when it is n degrees above t' '. For let R, R lt 
 and R 2 represent the rates of conduction of heat (flow of 
 heat in one second) to the lower block at temperatures t', 
 t' n, and t f +n, the upper end being at temperature /. 
 Then 
 
 KA(t-t') 
 
 R 
 
 I 
 
 KA[t-(t'-n)] 
 I 
 
 KA[t-(t'+7i)] 
 
 whence 
 
 Again, when the lower block is at the same temperature 
 as the jacket, it neither receives heat from nor gives heat 
 to the jacket. And when it is n degrees below it gains 
 heat as rapidly as it loses heat when it is n degrees above. 
 Thus by taking the mean rate as above, the effects of radia- 
 tion to or from the block are eliminated. In fact, adding 
 a to R lt to allow for the gain by radiation, and subtracting 
 
104 HEAT. 
 
 a from R 2 , to allow for loss by radiation, would leave 
 unchanged. 
 
 The same would hold true for any other pair of temper- 
 atures equidistant from the temperature of the jacket. 
 If the rates of rise at two temperatures equidistant from 
 the temperature of the jacket be r t and r 2 , by what has been 
 said the rate at the temperature of the jacket would be 
 i( r i+ r 2)- Hence, the rate at which the body must be 
 gaining heat is Ms^(r 1 +r 2 ). Hence, by the definition of 
 thermal conductivity, 
 
 or, 
 
 A(t-t') 
 
 The lower block should be cooled initially to about 12 
 below the temperature of the water that circulates through 
 the jacket by being placed in a bath of ice and water (or 
 snow). When taken out, it must be carefully dried. The 
 jacket may be kept at a constant temperature by water 
 passing and repassing through it between two large vessels, 
 which are alternately raised and lowered about every five 
 or ten minutes. The temperature of the water should be 
 frequently read by a thermometer (which may conveniently 
 pass through a large cork that floats on the surface of the 
 water). If the temperature of the water should show a 
 tendency to rise or fall, a small quantity of cooler or warmer 
 water, respectively, may be added. If the vessels be large 
 and the temperature of the room does not vary widely, 
 there should be no difficulty in keeping the water constant to 
 within . 2 for a sufficient length of time. 
 
 The hot bath is in the form of a trough, which is heated 
 at one end, while the conducting rod passes into the tank at 
 the other end. To prevent direct radiation from the burner 
 to the rod, thick screens of wood and asbestos are interposed. 
 The temperature of the lower block should be read at least 
 
THE MECHANICAL EQUIVALENT OF HEAT. 105 
 
 every minute by means of a thermometer passing through 
 the cork or fiber cover and inserted into a hole in the block, 
 the unoccupied space in the hole being filled with mercury. 
 The readings of temperature will, from various causes, be 
 slightly irregular. They should therefore be plotted in a 
 curve, and the irregularities eliminated by taking the more 
 correct values from the curve. The temperature of the 
 upper block should be read frequently by a thermometer 
 thrust into it. Small quantities of boiling water should be 
 added frequently to the bath to compensate for evaporation. 
 
 To gain some idea of the amount of reliance to be placed 
 on the result, the mean rate of rise for each pair of degrees 
 equidistant from the temperature of the jacket should be 
 obtained, and the final mean of all taken in calculating. 
 It is, however, to be noted that the temperatures nearest 
 the jacket temperature should give the best results, since 
 there the radiation is a minimum, and therefore any defect 
 in the method of correcting for radiation a minimum. 
 
 (For comparing the conductivities of poorly conducting 
 substances Lees and Chorlton's apparatus is quite satis- 
 factory. Directions for its construction and manipulation 
 are given in Robson's Heat, page 135.) 
 
 XXV. THE MECHANICAL EQUIVALENT OF HEAT. 
 
 Griffith, Thermal Measurement of Energy, Chap. Ill; Edser, Heat, 
 Chap. XII; Text-book of Physics (Duff), pp. 259-262; Watson's 
 Physics, 250-251; Ames 1 General Physics, pp. 203-205; 
 Crew's General Physics, 289; Rowland, Physical Papers, pp. 
 343-476. 
 
 The mechanical equivalent of heat is the number of units 
 of mechanical energy that, completely turned into heat, 
 will produce one unit of heat, or, in the c. g. s. system, the 
 number of ergs in a calorie. The apparatus here described 
 is a copy of that used in the University of Cambridge, 
 England, and the following description and introduction 
 is partly taken from that issued to students in that university. 
 
io6 
 
 HEAT. 
 
 In this apparatus mechanical energy is expended in working 
 against friction, thus producing heat, which is measured 
 by the rise in temperature of a known mass of water. 
 
 A vertical spindle carries at its upper end a brass cup. 
 Into an ebonite ring concentric with the cup there fits 
 tightly one of a pair of hollow truncated cones. The second 
 cone fits into the first, and is provided with a pair of steel 
 pins which correspond to two holes in a grooved wooden 
 
 FIG. 28. 
 
 disk, which prevents the inner from revolving when the 
 spindle and the outer cone revolve. A cast-iron ring, 
 resting on the disk and fixed by two pins, serves to give a 
 suitable pressure between the cones. A brass wheel is 
 fixed to the spindle, and, by a string passing round this 
 wheel and also round a hand-wheel, motion is imparted to 
 the spindle. A pair of guide pulleys prevents the string 
 from running off the wheel. Above the wheel is a screw 
 cut upon the spindle. This screw actuates a cog-wheel of 
 100 teeth, which makes one revolution for every 100 revo- 
 lutions of the spindle. 
 
THE MECHANICAL EQUIVALENT OF HEAT. 107 
 
 To the base of the apparatus one end of a bent steel rod 
 is attached; the rod can be fixed in any position by a nut 
 beneath the base. The other end of the rod carries a cradle, 
 in which runs a small guide pulley on the same level with the 
 disk. The cradle turns freely about a vertical axis. A fine 
 string is fastened to the disk and passes along the groove in 
 its edge; it then passes over the pulley and is fastened to a 
 mass of 200 or 300 grams. On turning the hand- wheel it is 
 easy to regulate the speed so that the friction between the 
 cones just causes the mass to be supported at a nearly 
 constant level. To prevent the string from running off 
 the guide pulley, a stiff wire with an eye is fixed to the 
 cradle and the string is passed through this eye. It also 
 passes through an eye fixed to the steel rod, to prevent 
 the weight from being wound up over the pulley. 
 
 The rubbing surfaces of the cones must be carefully 
 cleaned and then four or five drops of oil must be put be- 
 tween them; the bearings of the spindle and guide pulley 
 should also be oiled. The cones are then weighed together 
 with the stirrer. The inner cone is then filled to about i 
 cm. from its edge with water 2 or 3 below the temperature 
 of the room and the system is again weighed. The cones 
 are then placed in position in the machine and a thermometer 
 is hung from a support so that it passes through the central 
 aperture in the disk and almost touches the bottom of the 
 inner cone. 
 
 One observer, X, takes his place at the hand-wheel, and 
 the other, Y, at the friction machine. By working the 
 machine the water is now warmed up until its temperature is 
 nearly equal to that of the room. The index of the counting 
 wheel is read and the temperature of the water is carefully 
 observed every minute for five minutes. Immediately after 
 the last reading, X turns the wheel fast enough to raise the 
 mass until the string is tangential to the edge of the disk. 
 If the string be not tangential the moment of its tension 
 about the axis of revolution is seriously diminished. Y 
 stirs the water and notes the temperature at each passage 
 
108 HEAT. 
 
 of the zero of the counting-wheel past the index; each 
 passage of the zero after the first corresponds to 100 revolu- 
 tions of the spindle. Y gives a signal at each passage of the 
 zero and X notes the time by aid of a watch. After Y has 
 recorded the temperature upon a sheet of paper previously 
 ruled for the purpose, he also records the time observed by 
 X. After about 1000 revolutions the motion is stopped 
 and the readings of the index of the counting wheel and of the 
 thermometer are recorded. Observations of the temper- 
 ature are continued every minute for five minutes, the stirring 
 of the water being continued. 
 
 The temperature observations are plotted against the 
 time, and the radiation correction is determined as explained 
 on pages 63 and 64. The heat produced is readily cal- 
 culated from the mass of water, the water equivalent of 
 the cones and stirrer, and the corrected rise of temperature. 
 
 From the initial and final readings of the counting wheel 
 and the number of complete revolutions the exact number 
 (n) of revolutions made by the spindle is deduced. The 
 work done is calculated as follows: When the spindle 
 has made n turns the work spent in overcoming the friction 
 between the cones is the same as would have been spent 
 if the outer cone had been fixed and the inner one had been 
 made to revolve by the descent of the mass of M grams. 
 In the latter case M would have fallen through innr cm. 
 where r is the radius of the groove of the wooden disk, which 
 must be measured. Hence the total work spent against 
 friction and turned into heat is 2nnrMg ergs. In the report, 
 estimate the possible error of the result as far as it depends 
 upon the errors of observations and measurements. 
 
 Questions. 
 
 1 . What amount of error is due to neglect of the work spent against 
 friction of the bearing of the outer cone ? 
 
 2. Why must the wheel be turned faster as the experiment pro- 
 ceeds? 
 
 3. What effect on the result has the variation of the viscosity of 
 the oil? 
 
THE MELTING-POINT OF AN ALLOY. IOQ 
 
 XXVI. THE MELTING-POINT OF AN ALLOY. 
 
 Robson, Heat, pp. 77-79; Findlay, Phase Rule, pp. 220-223; Ewell, 
 Physical Chemistry, pp. 271-272. 
 
 If an alloy is melted and is allowed to cool while its tem- 
 perature is continuously observed, and a curve be then drawn 
 with times as abscissae and temperature as ordinates, it 
 will be found that at certain points the curvature abruptly 
 changes, the fall of temperature being decreased or even 
 ceasing. At the moment corresponding to such a point, the 
 alloy is radiating heat to the room, and the fact that its 
 temperature does not fall as rapidly indicates that heat is 
 being produced internally by some change of state of the 
 material. Such a point is therefore a solidifying-point of 
 some constituent of the alloy or of the eutectic alloy. 
 
 The assigned* metals are carefully weighed and melted in 
 an iron cup. A copper-constantinf thermocouple is plunged 
 into the liquid metal and kept there until the entire mass is 
 solid. A porcelain tube should cover one wire for some 
 distance from the junction. The terminals are connected 
 to a calibrated galvanometer through a resistance such that 
 the maximum deflection will keep on the scale. The galvan- 
 ometer is read every half-minute and the time of each read- 
 ing is noted. When the readings are commenced, the metal 
 should be considerably above the melting-point and the 
 readings should be continued for some time after the metal 
 is apparently solid. For calibration of the galvanometer, see 
 Exp. LVIII. 
 
 Plot the galvanometer deflections against the time. De- 
 termine the electromotive force corresponding to the galvan- 
 ometer deflections where the curvature changed, and from 
 the constants of the thermocouple, or a chart giving the 
 
 * Tin and lead are suitable metals. The changes of curvature are more 
 distinct if the former is in excess. The eutectic of tin and lead is composed 
 f 37% lead, 63% tin, and melts at 182.5 (Rosenhain and Tucker, Roy. Soc. 
 Phil. Trans., 1908, A. 209, p. 89). 
 
 f "Advance" Wire, Driver-Harris Co., Harrison, N. J. 
 
1 10 HEAT. 
 
 temperature for different electromotive forces, determine 
 the temperature of these points. 
 
 Tabulate the observed temperatures of these transition 
 points and your opinion of what they represent. 
 
 Questions. 
 
 1. Explain why the second transition point is represented by 
 a horizontal portion of the cooling curve while the first transition 
 point is merely represented by a change of curvature. 
 
 2. Will the temperature of the first point vary with the initial 
 concentration ? 
 
 3. Will the temperature of the second transition point vary with 
 the initial concentration ? 
 
 XXVII. HEAT VALUE OF A SOLID. 
 
 Ferry and Jones, pp. 237-242. 
 
 (A) HEMPEL BOMB CALORIMETER (Constant-volume 
 Calorimeter) 
 
 A pellet of the fuel to be tested is formed in a press, a 
 cotton cord being imbedded with a loose end. After being 
 pared down to about i gm. and brushed, it is carefully 
 weighed. It is then suspended in a Hempel combustion 
 bomb, and the thread is wrapped around a platinum wire 
 connecting the platinum supports of the basket. The ter- 
 minals attached to these supports are connected with several 
 Edison or storage cells sufficient to just bring the wire 
 to a brilliant incandescence (as ascertained by a preliminary 
 trial) . 
 
 The bomb is charged with oxygen under at least fifteen 
 atmospheres' pressure, either from a charged cylinder or 
 produced by a retort. Bomb and pressure gauge should be 
 immersed in water while the oxygen is being supplied. As- 
 certain that the bomb valve is open and that all connections 
 are screwed tight. Open the cylinder valve (if a cylinder is 
 used) until the pressure becomes high, and then close. Lift 
 the bomb out of the water, loosen one of the connections, 
 and allow the mixture of air and oxygen to escape; then 
 
HEAT VALUE OF A SOLID. Ill 
 
 tighten, replace in water, open the cylinder valve again until 
 the pressure becomes high (at least fifteen atmospheres); 
 close both the cylinder valve and that of the bomb, and 
 finally disconnect and dry the bomb. (If the oxygen is 
 produced in a retort, partly fill the latter with a five to 
 one mixture of potassium chlorate and manganese dioxide, 
 connect to the bomb and pressure gauge, and heat the upper 
 part slowly with a Bunsen burner.) 
 
 Attach the electrical terminals, place the bomb in the 
 special vessel containing about a liter of water, adjust the 
 Beckmann thermometer to read about i (see p. 65), stir 
 the water continually, and read its temperature every half- 
 minute for five minutes, estimating to tenths of the smallest 
 graduation. Close the electric switch; after a few seconds 
 open it and read the temperature of the continually stirred 
 water for ten minutes. 
 
 Let H be the heat value of the fuel and m the mass of 
 the specimen, M the mass of the water, e the water equivalent 
 of the bomb, t l the initial temperature, and t 2 the final tem- 
 perature (corrected for radiation, see p. 63.) Then 
 
 eWi-tJ 
 
 cal. pergm. 
 wi 
 
 The best method in practice to determine e is to repeat 
 the determination, using 'salicylic acid as fuel, and assum- 
 ing its heat value to be 5300 calories per gram. 
 
 (B) ROSENHAIN'S CALORIMETER (Constant-pressure 
 Calorimeter) 
 
 Phil. Mag., VI, 4, p. 451. 
 
 Instead of burning the fuel in a fixed volume of highly 
 compressed oxygen, the oxygen is supplied continuously 
 at only slightly above atmospheric pressure. 
 
 The coal is pulverized and a sample is compressed, in a 
 special screw press, into a pellet weighing about one gram. 
 This is placed on a porcelain dish which rests on the bottom 
 
112 
 
 HEAT. 
 
 of the inside chamber. The ignition wire should be about 
 3 cm. of No. 30 platinum wire and the external terminals 
 should be connected to storage-battery terminals through a 
 key and a resistance such that the wire will glow^brightly. 
 A gasometer is charged with oxygen from a cylinder or gen- 
 erated from "oxone" and water. The action of the dif- 
 ferent valves having been studied, the apparatus should 
 be assembled, the upper side valve (see Fig. 29) being closed 
 and the ball valve lowered. Connect with 
 the oxygen supply through a wash-bottle, 
 turn on a very gentle stream of oxygen, 
 and pour into the outer vessel a measured 
 volume of water, at the room tempera- 
 ture, so that the combustion chamber is 
 just covered. 
 
 If a Beckmann thermometer (see p. 65) 
 is used, adjust to read between o and i 
 in this water. The bulb should be sup- 
 "~\CT_-Jj/ ported on a level with the center of the 
 combustion chamber. Read the tempera- 
 ture every half-minute for five minutes, 
 then increase the oxygen current, and care- 
 fully read the temperature and the time, close the key and 
 ignite the pellet with the hot platinum wire, and immediately 
 remove the wire. During such operations it is best to hold 
 the inner vessel steady by grasping the oxygen inlet tube. 
 Keep the water pressure in the gasometer constant, and as 
 combustion proceeds increase the flow of oxygen. If pos- 
 sible, read the thermometer every half-minute. 
 
 When combustion has ceased, move the hot wire about 
 to igni-te any unconsumed particles. Keep the wire hot as 
 short a time as possible and remove it immediately from any 
 combustion, otherwise it is liable to be melted. 
 
 Finally, turn off the oxygen supply, open the upper valve, 
 and raise the ball valve, allowing the water to enter the 
 inner chamber. Then force out the water by closing both 
 valves and turning on the oxygen. Record the highest 
 
 FIG. 29. 
 
HEAT VALUE OF A GAS OR LIQUID. 113 
 
 temperature and the time and the temperature every 
 half-minute for five minutes. For radiation correction, 
 see p. 63, and for formulae, see (A) preceding. 
 
 To determine e, assemble the apparatus, including the 
 Beckmann thermometer, and pour in 1000 c.c. of water. 
 Determine very carefully the temperature with a 0.1 
 thermometer and then add 500 c.c. of water at about 50, the 
 temperature of which has also been very carefully deter- 
 mined. Determine also very carefully the final steady 
 temperature and from these data determine e. 
 
 For anthracite coal add sugar in the proportion 3:1. 
 (Heat of combustion of sugar = 3 900 calories per gram.) 
 
 Questions. 
 
 1. Calculate the heat value of (a) one kilo of this substance, (b) 
 one short ton in B. T. U. per lb., (c) the mechanical energy equivalent 
 to the latter. 
 
 2. What error would be caused by (a) an error of 20 in the water 
 equivalent? (b) allowing a current of 5 amperes to flow through a 
 platinum wire of 2 ohms' resistance for 5 seconds? (c) Neglecting 
 the radiation correction ? 
 
 XXVIII. HEAT VALUE OF A GAS OR LIQUID. 
 
 Ferry and Jones, pp. 243-246. 
 
 The heat value will be determined with Junker's cal- 
 orimeter. 
 
 (A) A measured volume (v liters) of gas under an ob- 
 served pressure, p, is burned in the calorimeter, and the 
 rise of temperature, from tj to t 2 , of a mass of M gr. of 
 water, is determined. The flow of water and gas is so regu- 
 lated that the burned gas leaves the calorimeter at approxi- 
 mately the temperature of the entering gas, and there 
 should be a difference of at least 6 in the temperature of the 
 in- and outflowing water. Also, the flow of water must be 
 sufficient to furnish a constant small overflow at the supply 
 reservoir (see Fig. 30). The burner should be lighted out- 
 side the calorimeter. When the temperatures indicatedjDn 
 the various thermometers have become constant, note the 
 gasometer reading, and immediately collect in graduates the 
 8 
 
HEAT. 
 
 heated overflowing water, and also the water condensed 
 by the combustion of the gas. Let the mass of the latter be 
 m gr., and its temperature /'. Note the temperatures of the 
 inflowing and outflowing water every 15 sec. until two or 
 
 three liters have passed through. 
 Then immediately note the gaso- 
 meter reading and remove the 
 graduates. 
 
 Assuming that condensation of 
 the gas occurs at 100, the heat 
 liberated is m [536+100 2']. If 
 H represents the heat value of 
 the gas in gram-calories per liter 
 and v the volume, reduced to o, 
 and 760 mm., 
 
 M(t 2 -t 1 )-m(6 3 6-t f ) 
 
 FIG. 30. (B) To determine the heat 
 
 value of a liquid fuel, the gas 
 
 burner is replaced by a suitable lamp which is attached to 
 one arm of a balance. The rate at which the liquid is 
 consumed is determined from the weights in the pan, on the 
 other side, at different times. It is best to make the weight 
 in the pan slightly deficient and note the exact time when 
 the balance pointer passes zero, as the liquid is consumed. 
 Practically complete combustion is obtained with a "Primus" 
 burner, supplied from a reservoir where the liquid is under 
 considerable pressure. With very volatile liquids, the 
 opening of the burner must be large and the pre-heating 
 tubes must be in the cooler part of the flame. 
 
 Express your results in (a) calories per liter (b) B. T. U. 
 per gallon or cubic foot. 
 
 Questions. 
 
 1. Is the heat value of a gas, in calories per gram, definite? Per liter? 
 
 2. What difference would there be in the result if all the water 
 vapor escaped without condensing? 
 
 3. Why is no radiation correction necessary? 
 
PYROMETRY. 115 
 
 XXIX. PYROMETRY. 
 
 Edser, Heat, pp. 339411; Bulletin, Bureau of Standards, I, 2, pp. 
 189-255; Watson, Practical Physics, 208-210; Le Chatelier, 
 High Temperature Measurements, Chaps. Ill, VI, IX. 
 
 This exercise is a study of three of the methods used in 
 the measurement of very high temperatures. 
 
 A hollow black body is to be heated electrically and the 
 temperature of its interior is to be determined by means 
 of a calibrated thermocouple. A platinum resistance ther- 
 mometer is to be calibrated, and also an incandescent lamp 
 is to be calibrated for use as an optical pyrometer. 
 
 Electric Furnace. The electric furnace (see Fig. 31) 
 consists of a thin porcelain cylinder about 15 cm. long and 
 
 FIG. 31. 
 
 10 cm. in diameter upon which is wound about 5 m. of 
 No. 22 " Nichrome" wire,* if a 220-volt supply is to be used. 
 Whatever the voltage, the winding must be such as to 
 consume about half a kilowatt. The ends of the cylinder 
 are closed by porcelain caps with proper apertures and the 
 whole is surrounded by many layers of asbestos. 
 
 Heating and cooling must be very gradual so that the 
 thermocouple and platinum thermometer may acquire 
 the temperature of the furnace. The highest temperature 
 should not exceed 1000. 
 *$ee note bottom of page 99, 
 
Il6 HEAT. 
 
 Thermocouple. A platinum and platinum +10% rhodium 
 thermocouple should be connected to a galvanometer through 
 a key and such a resistance as will keep the deflection on the 
 scale at the highest temperature. The galvanometer with 
 resistance should be calibrated as a voltmeter (Exp. LVIII). 
 The chart or table accompanying the couple gives the 
 temperature of the hot junction when the electromotive 
 force is known. (The cool junction should be in ice and 
 water.) 
 
 Platinum Resistance Thermometer. The platinum resist- 
 ance thermometer consists of a coil of fine platinum wire 
 (for example, 50 cm. of No. 30) wound on a porcelain frame 
 and surrounded by a glazed porcelain tube. The coil 
 constitutes one arm of a Wheatstone's bridge. A pair of 
 dummy leads are connected to an adjoining arm (see figure) 
 and compensate for the heating of the lead wires. A 
 suitable switch connects the galvanometer to either the 
 bridge or the thermocouple. 
 
 Optical Pyrometer. The optical pyrometer consists 
 essentially of a lens and a miniature incandescent lamp. The 
 lens focuses the interior of the enclosed furnace (an ideal black 
 body) on the filament of the lamp. The current through 
 the filament is adjusted until the tip of the filament is 
 invisible against the image of the furnace. When this is 
 true, both must be emitting similar light, and therefore 
 they must be at approximately the same temperature. A 
 small eye-piece aids in observing the tip of the filament. 
 The incandescent lamp filament is to be calibrated; i.e., the 
 current necessary to heat the tip of the filament to different 
 temperatures is to be determined. The temperature of the 
 filament is determined by finding the temperature of the 
 furnace, by the thermocouple, when the two have the same 
 temperature. (The incandescent lamp circuit contains an 
 ammeter for measuring the current, which is omitted from 
 
 Fig. 31.) 
 
 Observations. While the furnace is slowly heating, and 
 also while it is slowly cooling, observe at frequent intervals, 
 
PYROMETRY. 117 
 
 (A) the galvanometer deflection with the thermocouple, 
 and the resistance in the galvanometer circuit; (B) the 
 resistance of the platinum thermometer; (C) the current 
 through the filament. The three observations should be 
 made in succession and the times of each recorded. (D) 
 Calibrate the galvanometer as described in Exp. LVIII, if 
 the constant is not furnished. (E) If time permits, use the 
 optical thermometer to determine the temperature of 
 various distant, brightly heated bodies; e. g., melted silver, 
 iron, or copper. An image of the hot body is formed upon 
 the tip of the filament, and the current through the latter 
 is adjusted until the two are indistinguishable. The tem- 
 perature corresponding to this current is obtained from the 
 calibration (see (e) below). 
 
 Report, (a) Tabulate readings, (b) Plot the three sets 
 of readings against the time, (c) Make a second plot with 
 resistance as abscissae and, for ordinates, the platinum 
 temperatures, as given by Callender's equation, 
 
 RIRQ 
 
 pi IOO - . 
 
 where R is the extrapolated resistance at o and R 1 is the 
 resistance at 100. Transfer to this plot also the readings 
 of the true temperature, /, and determine the mean value of 
 Callender's difference constant, 8, by applying at several 
 points the equation 
 
 For pure platinum 8 is 1.50. The platinum resistance 
 thermometer is the most accurate, convenient method of 
 measuring temperature below 1000. 
 
 (d) Construct a curve which gives the temperature of the 
 tip of the incandescent lamp filament plotted against the 
 current, (e) Finally, determine by the latter curve the 
 temperatures of any bodies tested with the optical pyrometer 
 and record the results. (For a discussion of the error in 
 
Ilg HEAT. 
 
 assuming for different bodies that the radiation is similar to 
 that from a black body, see the above reference to the 
 Bulletin of the Bureau of Standards and Haber, "Thermo- 
 dynamics of Gas Reactions" pp. 281-291.) 
 
 Questions. 
 
 1. Would you expect the platinum resistance thermometer to 
 attain a slightly higher or a slightly lower temperature than the 
 thermocouple? Explain. 
 
 2. Is any correction required for the absorption of the lenses 
 in such an optical pyrometer? Explain. 
 
SOUND. 
 
 /^s\ 
 
 XXX. THE VELOCITY OF SOUND. 
 
 Ttxt-book of Physics, (Duff), pp. 319, 320, 334-337; Watson's Physics, 
 287, 288, 309; Ames' General Physics, pp. 337, 363; Crew's 
 General Physics, 213; Poynting and Thomson, Sound, Chap. VII- 
 
 The velocity of sound in a medium can be found if the 
 wave-length, L, in the medium of a note of frequency n can 
 be determined; for v = n L. If 
 the medium be contained in a 
 tube, one end of which is closed, 
 the closed end must be a node 
 and the open end a loop. Hence 
 the length of the tube must be 
 an odd number of quarter wave- 
 lengths. Such a tube will reso- 
 nate to a fork, if the wave-length 
 of a natural vibration of the pipe 
 be the same as the wave-length 
 to which the fork gives rise. 
 Thus, if the length of pipe that 
 resonates to a fork of known 
 pitch be measured, we have the 
 means of finding the velocity of 
 sound. 
 
 A long glass tube is mounted 
 on a stand. Water is introduced 
 from the bottom, where is at- 
 tached a rubber tube provided 
 with a pinch-cock and connected 
 to a glass bottle. By raising and 
 lowering the bottle, water may be 
 brought to any height in the tube. The additional connec- 
 tions represented in figure 32, permit raising or lowering the 
 water level by opening pinch-cocks. When A is opened 
 
 119 
 
 
 FIG. 32. 
 
120 SOUND. 
 
 the tube fills and the tube empties upon opening D. C 
 and B are pinch-cocks which regulate the rate of flow. 
 The vessels are so large, compared with the capacity of the 
 tube that only rarely is it necessary to transfer water from 
 the lower to the upper vessel. 
 
 If a tuning-fork be vibrated above the tube, resonance will 
 first take place when the air column is approximately one- 
 quarter wave-length of the fork, next when three-quarter 
 wave-lengths, etc. In reality, the first loop is not exactly at 
 the open end of the pipe, but a short distance beyond the 
 open end. The distance between two nodes is accurately 
 half a wave-length, and it is from this distance that the 
 wave-length is best determined. 
 
 The tuning-fork may be sounded by gently striking the 
 end of the fork against the knee or a block of soft wood. 
 The fork should be held above the end of the tube, so that 
 the plane of the prongs includes the axis of the tube. Each 
 node should be located very carefully, at least four times, 
 each location being tested both with the water rising 
 and with the water falling, and the distance of each position 
 from the open end noted. The mean is taken as the true 
 distance. The whole should then be repeated with a fork 
 of different pitch. Observe the temperature and barometric 
 pressure, and measure the diameter of the tube. With a 
 little practice one can often locate nodes corresponding to 
 the higher modes of vibration of the fork. This should 
 be tried and, from their wave-lengths and the velocity of 
 sound as already determined, the pitch of these higher sounds 
 can be calculated. 
 
 The pitch of the forks used may be determined by com- 
 parison with a standard fork by the method of beats. The 
 standard is mounted on a resonance box and is set in vibra- 
 tion by pulling the prongs together with the fingers and 
 then releasing them. The fork of unknown pitch is sounded 
 in the usual way, and the end of the shank is set upon the 
 resonance box of the standard. If the two forks are of 
 nearly the same pitch, beats will be heard. With a stop- 
 
THE VELOCITY OF SOUND. 121 
 
 watch the time of ten, fifteen, or twenty beats, as may be 
 convenient, is several times determined. Dividing by the 
 time, we have the number of beats per second, and this 
 is the difference of pitch of the two forks. To determine 
 which fork is the higher, add a little wax to one prong of 
 the fork used in the experiment. Since this increases the 
 inertia of the fork, it decreases its pitch. If originally the 
 two forks have very nearly the same pitch, so that there 
 is only a fraction of a beat per second, very small amounts 
 of wax should be added. A large piece of wax might 
 change the pitch of the fork from above that of the standard 
 to below. (Time may often be saved by comparing the 
 fork with the standard during the necessary delays of the 
 work described below.) 
 
 The velocity of sound in carbon dioxide may be deter- 
 mined in a similar manner. The water surface is first 
 lowered to the bottom of the tube and the tube is filled with 
 the gas from a generator through a small tube lowered just 
 to the water surface (not below) . The generator consists of 
 a tube filled with marble, surrounded by dilute hydrochloric 
 acid. In filling the generator with marble, use only whole 
 pieces, carefully excluding any dust or pieces small enough 
 to drop through into the outer vessel. If the gas is not 
 evolved in sufficient quantity, add hydrochloric acid to the 
 outer vessel. When the air in the tube has been entirely 
 displaced by the gas, a lighted match introduced into the top 
 of the tube will be extinguished. The delivery tube is now 
 withdrawn from the resonance tube, while the gas still flows 
 out to fill the volume occupied by the delivery tube. 
 
 The water surface is now slowly raised and the nodes 
 located for one of the forks. Observations can only be made 
 with the water rising, for, when the water surface is lowered, 
 air enters the tube. Hence each node can be located but 
 once. The tube is again filled and the nodes redetermined 
 with the same fork. 
 
 From the distance between the nodes and the pitch of the 
 fork the velocity of sound is determined. 
 
1.2 SOUND. 
 
 For each gas, find the correction for the open end; that 
 is, the displacement of the loop beyond the end of the tube. 
 Use the mean position of the highest node and one-fourth 
 of the mean wave-length. Find what fraction this displace- 
 ment is of the radius of the tube. 
 
 Calculate, from the mean values of the velocities, the 
 velocity of sound in each gas at o C. (see references). 
 Find also the ratio of the specific heats from the velocity 
 at o C., the standard barometric pressure (both in absolute 
 units) and the density as given in Table VI. 
 
 Questions. 
 
 1. Explain why (a) readings are made with both rising and falling 
 water (b) the plane of the prongs of the fork must contain the axis 
 of the tube. 
 
 2. What is the influence of atmospheric moisture upon the velocity 
 of sound? 
 
 XXXI. VELOCITY OF SOUND BY KUNDT'S METHOD. 
 
 Text-book of Physics (Duff), p. 338; Watson's Physics, 317; Ames' 
 General Physics, p. 364; Crew's General Physics, 215; Poynting 
 and Thomson, Sound, pp. 115-117; Watson's Practical Physics, 
 113- 
 
 A glass tube, A G, about a meter long and about 3 cm. 
 internal diameter is closed at one end by a tight-fitting 
 piston, C, and at the other end by a cork through which 
 passes a glass tube having at one end a loosely fitting card- 
 board disk, D (Fig. 33). The glass tube should be about a 
 meter long. A little dry powdered cork is sprinkled in 
 the tube, the stopper at G is loosened, and a current of air, 
 dried by passage through several drying tubes, is slowly 
 forced through the hollow rod of the piston, C. The stopper 
 at G is then replaced and the glass tube, F, is held at the 
 center and stroked longitudinally with a damp cloth. The 
 piston, C, is adjusted until the powder collects in the sharp- 
 est attainable ridges. These ridges will appear where the 
 pressure changes are least; that is, at the loops. Measure 
 carefully the distance between two extreme ridges and 
 
VELOCITY OF SOUND BY KUNDT's METHOD. 123 
 
 divide by the number of segments into which the tube is 
 divided. This distance (between two loops) is a half wave- 
 length of the waves in the tube. Disturb the powder and 
 make a new adjustment of the piston, C, and a new measure- 
 ment of the half wave-length. Make a third repetition of 
 the adjustments and readings. 
 
 
 
 7^1 
 
 
 1 1 
 
 
 c r~ 
 
 j 
 
 A C 
 
 D 
 
 FIG. 33. 
 
 6 
 
 Fill the tube with another dried gas, for example, carbon 
 dioxide, illuminating gas, hydrogen, oxygen, or hydrogen 
 sulphide, and determine the half wave-length. If n is the 
 constant pitch of the note emitted by the glass tube and / is 
 the wave-length in the gas 
 
 v i l i 
 v=nl .'. =~i. 
 
 ^2 k 
 
 Since the velocity changes at the same rate with change of 
 temperature in all gases (see references), the velocity of 
 sound or compressional waves at zero degrees in any other 
 gas than air can be calculated from the ratio of the wave- 
 lengths at a common temperature, and the velocity in air 
 at zero degrees (33,200 cm. per second). 
 
 From the velocity of sound at zero degrees in the gases 
 other than air and the standard atmospheric pressure cal- 
 culate the ratio of specific heats, 7- (see references). Table 
 VI gives the densities of the more common gases and vapors 
 at zero degrees and a pressure of 76 cm. of mercury = 1013200 
 dynes per square centimeter. 
 
 Questions. 
 
 1. Calculate (a) the velocity of compressional waves in glass, 
 (b) the elasticity E. (Notice that each end of the glass rod must be a 
 loop, and the center a node. The density of glass can be obtained 
 from Table VIII.) 
 
 2. Why must the glass rod be set in longitudinal vibration? 
 
 3. Why does the powder collect at the loops? 
 
LIGHT. 
 
 27. Monochromatic Light. 
 
 The simplest and most useful monochromatic light is the 
 sodium flame. Sodium may be introduced into a Bunsen 
 flame by surrounding the tube of the burner with a tightly 
 fitting cylinder of asbestos which has been saturated with a 
 strong solution of common salt and formed into cylindrical 
 shape by wrapping around the burner while still damp. 
 As the top of the cylinder is exhausted, it should be torn off 
 and the rest of the tube pushed up into the lower part of 
 the flame. A piece of hard-glass tubing held in the flame 
 will also give a good sodium light. 
 
 Elements giving red, green, blue, and violet light will be 
 found in Table XVIII. Salts of these elements (e. g., 
 KN0 3 , SrCl 2 , CaCl 2 , LiCl) may be introduced into the outer 
 edge of a bunsen flame, either in a thin platinum spoon, on 
 copper gauze, or by a piece of wood charcoal which has 
 absorbed a solution. If a very intense light is not required, 
 a vacuum tube is a very satisfactory source (Table XVIII). 
 Intense light of one general color may be obtained by filtering 
 sun light or the light from an arc light through colored glass 
 or gelatine. The solutions given in the accompanying table 
 give much purer monochromatic light. 
 
 Light Filters (Landolt).* 
 
 Color 
 
 Thickness 
 of layer 
 (mm.) 
 
 Aqueous 
 solution of 
 
 j Grams per 
 
 IOO C.C. 
 
 Average 
 wave-length 
 (Angstrom units) 
 
 Red.. 
 
 20 
 
 20 
 
 , Crystal violet 560 
 Potassium chromate 
 
 .005 
 10. 
 
 6560 
 
 Green 
 
 2O 
 20 
 
 ; Copper chloride 
 Potassium chromate 
 
 60. 
 10. 
 
 5330 
 
 Blue.. 
 
 20 
 2O 
 
 l Crystal violet 
 1 Copper sulphate 
 
 .005 
 15- 
 
 4480 
 
 * Mann, Manual of Advanced Optics, p. 185. 
 
 124 
 
SPHERICAL MIRRORS AND LENSES. 125 
 
 28. Rule of Signs for Spherical Mirrors and Lenses. 
 
 Mirrors. Consider the side upon which the incident light 
 falls as the positive side of the mirror. If the object, the 
 image, or the principal focus is on this side, their respective 
 distances, (u, v, f=rj 2) will be positive; if on the other side, 
 negative. Therefore, the focal length (and hence radius) is 
 positive for concave mirrors and negative for convex. The 
 object distance, u, will obviously, in most cases, be positive. 
 
 The formula for all spherical mirrors is : 
 
 I I I 2 
 
 U V f Y 
 
 if the signs of the numerical quantities which are substituted 
 for u, v, f, and r are determined by the above rule. 
 
 Lenses. Let all the distances, u, v, f, r lt r 2 be positive 
 for the double convex lens, when the object is outside the 
 principal focus; that is, in the most common case. The 
 formula for all lenses is then 
 
 U V f 
 
 As an illustration of the application of this rule, consider the 
 signs of these distances when an image of a real object is 
 formed by a double concave lens. The distance, u, of the 
 object is obviously measured on the same side as it would 
 be in the standard case of the double convex lens and is, 
 therefore, positive. The distances / and v are, however 
 measured on the same side of the lens as the object, or 
 opposite to the standard case with the double convex lens, 
 and are, therefore, negative. r lt the radius of the front face, 
 is on the same side as the object, while in the case of the 
 double convex lens this radius is on the other side, therefore 
 r lt and similarly r 2 , is negative for a double concave lens. 
 
XXXII. PHOTOMETRY. 
 
 Text-book of Physics (Duff), p. 353; Watson's Physics, 361-364; 
 Ames' General Physics, pp. 437, 442; Watson's Practical Physics, 
 pp. 382-387; Edser, Light, pp. 9-20; Stine's Photometrical 
 Measurements; Palaz' Photometry. 
 
 The intensity of illumination of a surface by a source of 
 light of small area varies inversely as the square of the 
 distance. Hence it follows that, if two lights produce equal 
 intensities of illumination at a point, P, their illuminating 
 powers, or the intensities of illumination they can produce at 
 unit distances, are directly as the squares of their distances 
 from P. This is the basis of all practical methods of com- 
 paring illuminating powers. 
 
 As a means of testing when two different sources of light 
 produce equal illumination at a point, various so-called 
 screens have been used. The one that has been most 
 extensively employed is Bunsen's grease-spot screen. It is 
 based on the fact that a grease-spot on paper is invisible 
 when the paper is equally illuminated on both sides, since 
 viewed from one side as much light is gained by transmission 
 from the farther side as is lost by transmission to the farther 
 side. 
 
 Another screen more perfect in some respects is that of 
 Lummer and Brodhun. A white opaque disk (see figure 34) 
 is illuminated on opposite sides by the two sources of light. 
 An arrangement of mirrors and lenses enables one eye to 
 view both sides at once. Two plane mirrors reflect rays 
 from the two sides into a double glass prism. This consists 
 of two separate right-angled prisms, the largest face of one 
 being partly beveled away and the two being cemented to- 
 gether by Canada balsam, which has the same optical density 
 as the glass, and therefore reflects no light. The central 
 rays from the left pass through the double prism to the tele-. 
 
 126 
 
PHOTOMETRY. 127 
 
 scope while the marginal rays are totally reflected by the 
 beveled edge. The marginal rays from the right are totally 
 reflected and reach the telescope, but the central rays pass 
 through. Thus the eye sees a circular portion of the left 
 side of the opaque disk and a surrounding rim of the right 
 side. 
 
 To eliminate error from lack of symmetry, the lamps 
 compared should be interchanged in the course of the readings 
 or the screen should be rotated 180. 
 
 The lights to be compared are 
 mounted at opposite ends of a 
 graduated bar 3 meters long, which, 
 with a parallel bar and suitable sup- 
 ports, constitutes the photometer 
 bench. The screen is mounted on a 
 carriage movable along the bench. 
 
 Many light standards have been 
 employed. A candle of certain care- FlG 
 
 fully specified dimensions was long 
 employed, and the illuminating power of such a candle is still 
 regarded as the unit and called "one candle-power," but, 
 for practical purposes in testing, some other standard is 
 usually employed. The best such standard is a lamp, with 
 a wick of specified form and dimensions, burning amyl 
 acetate with a flame of specified height. (See references.) 
 Its relation to the "candle-power" is i c. p. = 1.14 amyl 
 acetate units. For most purposes an incandescent lamp 
 that has been standardized is the most useful standard, 
 especially in the study of incandescent lamps; but it must 
 not be used any great length of time without being re-stand- 
 ardized, since its illuminating power changes with prolonged 
 use. 
 
 The chief difficulty in comparing two different forms of 
 light is due to the fact that a difference of quality of the 
 two lights renders perfectly equal and similar illumination 
 of the two sides of the screen impossible. This difficulty 
 is still more marked in the study of arc-lights (for mechan- 
 
128 LIGHT. 
 
 ical arrangements see Stine, p. 236), for which it is best to 
 use as an intermediate unit a very powerful incandescent 
 lamp. The latter may be standardized by comparison with 
 an ordinary incandescent lamp, which again is compared 
 with an amyl acetate standard. Before connecting a lamp 
 to a circuit, ascertain that the voltage is not excessive for that 
 particular lamp. 
 
 (A) Carefully standardize an incandescent lamp, for 
 use as a working standard, by comparison with either an 
 amyl acetate lamp or a standardized incandescent lamp. 
 If the latter is used, the lamps should be in parallel, that the 
 voltage may be the same, and a variable resistance should 
 also be in the circuit, by varying which the voltage across 
 the lamps is maintained at the value prescribed for the 
 standard. See that the filament of the standard is in the 
 marked azimuth and note the position of the filament for 
 which the other lamp is standardized. 
 
 In each case several settings of the screen should be 
 rapidly made, and then the screen reversed and several more 
 made. The calculations may be facilitated by Table XXL 
 
 (B) The law of inverse squares should be tested by com- 
 paring two somewhat different incandescent lamps (i) when 
 3 m. apart, (2) when 2.5m. apart, (3) when 2 m. apart 
 on the photometer bench. The ratio of their illuminating 
 powers, as deduced in the three cases, should be a constant. 
 
 (C) The horizontal distribution of candle-power about 
 an incandescent lamp should be studied. This incandes- 
 cent lamp should be connected in parallel with the working 
 standard and the voltage maintained at the value for which 
 the latter was standardized. The lamp should be mounted 
 on the revolving lamp-holder of the photometer, care being 
 taken to have the center of the filament at the same height 
 as the center of the screen. The lamp is first turned to the 
 standard position, i. e., the position in which the plane of 
 the shanks of the filament is at right angles to the photom- 
 eter bench, and a marked face of the lamp is toward the 
 screen. The candle-power of the lamp is to be found in 
 
PHOTOMETRY. 1 29 
 
 this position and at positions 30 apart as the lamp is ro- 
 tated through 360. Two careful readings should be made 
 at each angle and the c. p. deduced from the mean. The 
 mean of all these values of the c. p. gives the mean horizon- 
 tal candle-power, A curve should be plotted, giving the 
 distribution of c. p. in polar co-ordinates. The mean hori- 
 zontal candle-power is more easily determined by continu- 
 ously rotating the lamp about a vertical axis by means of a 
 small motor. 
 
 (D) Efficiency of an Incandescent Lamp. Keeping the 
 potential of the working standard at the proper point (or 
 calibrating and using a lamp which may be connected to the 
 lighting circuit if the potential of that is constant), apply 
 various potentials to the lamp used in (C) at intervals be- 
 tween about 25% below the normal voltage to 25% above. 
 For each potential, determine the candle-power and current. 
 Calculate the watts consumed and the watts per candle- 
 power. 
 
 In the report, plot in three curves, with volts as abscissae, 
 (a) current, (b) candle-power, (c) watts per candle-power. 
 The scales of the three curves should be shown on the 
 vertical axis. 
 
 (E) Mean Spherical Candle-power. With the lamp in the standard 
 position of (C), find the c. p. at intervals of 30 in a vertical circle by 
 rotating the lamp about a horizontal axis. After this, start again 
 from the standard position and first turn the lamp through 45 in 
 azimuth (or around a vertical axis), and then, as before, find the c. p. 
 at intervals of 30 in the vertical circle of 45 azimuth, and so for the 
 vertical circles of 90 and 135 azimuth. As before, plot the curves 
 of distribution in polar co-ordinates. 
 
 To find the mean spherical candle-power omit any repetitions 
 and take the mean of the readings in the following positions : 
 
 1 . At tip i 
 
 2. At 60, 120, 240, 300 on the vertical circles of o and 90 
 
 azimuth 8 
 
 3. At 30, 150, 210, 330 on the vertical circles of o, 45, 90, 
 
 135 azimuth 16 
 
 4. 12 equidistant positions on horizontal circle 12 
 
 5. At base (o) i 
 
 Total, 38 
 
 These directions are chosen because they are nearly uniformly 
 distributed in space. 
 
 9 
 
130 LIGHT. 
 
 (F) If time permit, study the differences of quality of 
 light given by different sources; e. g., compare an oil lamp and 
 an incandescent lamp using interposed colored glasses: 
 (i) a pair of red glasses, (2) of yellow glasses, (3) of blue 
 glasses. Calculate the relative illuminating powers in each 
 case. 
 
 The possible error may be deduced as usual from the 
 mean deviation of the readings in a set. 
 
 Questions. 
 
 1. Explain the deviation of the current-voltage curve from a 
 straight line. 
 
 2. What is the advantage in increasing the voltage applied to an 
 incandescent lamp? Disadvantage? 
 
 XXXIII. SPECTROMETER MEASUREMENTS. 
 
 Text-book of Physics (Duff), pp. 387, 436, 437, 440; Watson's Phys- 
 ics, pp. 468, 469, 493; Antes' General Physics, pp. 459, 460, 
 505, 506; Crew's General Physics, p. 510; Edser's Light, pp. 86-91. 
 
 A spectrometer consists of a framework supporting a 
 telescope and a collimator, both movable about a vertical 
 axis, and a platform movable about the same axis. The 
 platform is for supporting a prism or grating. The colli- 
 mator is a tube containing an adjustable slit at one end and 
 a lens at the other end. The purpose of the collimator is 
 to render light coming from the slit parallel after it leaves 
 the lens. (Only when the light that falls on a prism is par- 
 allel light, that is, light with plane wave front, does it seem 
 when emerging from the prism to come from a clearly de- 
 fined source. When it is not parallel, there is spherical 
 aberration.) Hence the slit of the collimator should be in 
 the principal focus of the lens. The telescope is for the 
 purpose of viewing the light that comes from the collimator, 
 either directly or after the light has been refracted or 
 reflected. Hence, since the light that comes from the colli- 
 mator is supposed to be parallel, that is, as if it came from 
 a very distant source, it follows that if the telescope is to 
 
SPECTROMETER MEASUREMENTS. 131 
 
 receive the light and form a distinct image of the slit, the 
 telescope must be focused as for a very distant object (theo- 
 retically an infinitely distant one) . 
 
 The first adjustment is to focus the telescope. First 
 focus the eye-piece of the telescope on the cross-hairs and 
 then focus the whole telescope on a distant object out of 
 doors. The telescope will now be in focus for parallel rays. 
 Turn the telescope to view the image of the slit formed by 
 the collimator and adjust the slit until its image is seen 
 most distinctly. 
 
 That the instrument should be in complete adjustment, 
 it is necessary that the telescope, collimator, and platform 
 should rotate about the same axis, and that the optical axes 
 of the telescope and the collimator should be perpendicular 
 to this axis of rotation. For fine work spectrometers are 
 made with all these parts separately adjustable, but simpler 
 instruments have the telescope and collimator put into per- 
 manent adjustment by the instrument maker. In any case 
 the telescope and collimator should not be adjusted for level 
 without the advice of an instructor. 
 
 Adjustment of Prism. The refracting edge of the prism 
 must be made parallel to the axis of the instrument. Place 
 the prism on the platform with one of the faces perpendicular 
 to the line joining two of the leveling screws. Turn the 
 collimator slit horizontal and place the telescope so as to 
 receive the image of the slit reflected from this face of the 
 prism. Adjust these two leveling screws until the image of 
 the stationary edge of the slit coincides with the horizontal 
 cross-hair. Then observe the image reflected from the 
 other face and adjust the third leveling screw until the edge 
 of this image is on the horizontal cross-hair. A little con- 
 sideration will show that, when these two adjustments have 
 been made, both faces, and therefore the refracting edge, 
 are parallel to the axis. Restore the collimator slit to the 
 vertical position. 
 
 Measurement of the Angle of a Prism. Method (A). The 
 prism should be so placed that the faces are about equally 
 
132 LIGHT. 
 
 inclined to the collimator. To secure good illumination, the 
 edge of the prism should be near the axis of the instrument. 
 The telescope is turned to view the image of the slit in the 
 two faces alternately, and the scale and vernier read when 
 the slit and cross-hair coincide, the slit being narrowed until 
 barely visible. If the scale is provided with two verniers, 
 to eliminate error from eccentricity, always read them both. 
 Half of the angle between the two positions of the telescope 
 gives the angle, A, of the prism, as may be readily seen by 
 drawing a diagram. The readings on each side should be 
 repeated three times. 
 
 Method (B). The following method, which is some- 
 times easier than the preceding, may be used if the platform 
 that carries the prism can be rotated and the rotation read 
 by a scale. Turn one face of the prism so as to reflect the 
 image of the slit into the telescope. Adjust the telescope 
 until the vertical cross-hair coincides with the slit and then 
 read the platform scale. Now rotate the platform until the 
 other face of the prism reflects the slit and again read the 
 platform scale. The difference of the readings is 1 80 A, as 
 may readily be seen by drawing a figure. The observation 
 should be repeated at least three times. 
 
 We are now in a position to make a final measurement 
 for finding the index of refraction of the glass of the prism 
 for any particular light of the spectrum, for instance, 
 sodium light (see p. 124). The only additional measure- 
 ment necessary is the deviation produced by the prism when 
 it is in such a position that it gives a minimum deviation to 
 the light refracted through it. 
 
 Minimum Deviation. The position of minimum devia- 
 tion is such that the image of the slit seen in the telescope 
 moves in the same direction (that of increasing deviation) 
 no matter which way the platform carrying the prism is 
 turned. There are, of course, two positions in which the 
 deviation can be obtained, one with the refracting edge turned 
 toward the right of the observer, and the other with it 
 toward the left. The deviation in each case is the angle 
 
SPECTROMETER MEASUREMENTS. 133 
 
 between the corresponding position of the telescope and its 
 position when looking directly into the collimator, the prism 
 being removed. But it is not necessary to remove the 
 prism, for it is easily seen that the minimum deviation must 
 also be equal to half of the angle between the two positions 
 of the telescope when observing the minimum deviation. 
 
 FIG. 35. 
 
 These two positions should be observed three times success- 
 ively, and the mean value for the minimum deviation, D, 
 taken. From A and D the index of refraction may be cal- 
 culated by the formula 
 
 sin 
 
 If time permit, determine the index of refraction for as 
 many other wave-lengths (colors) as possible (see p. 124). 
 
 The possible error of the determination of the refractive 
 index can be calculated by means of formulae deduced by 
 the calculus, as explained on pp. 7, 8. A simple, but 
 less accurate method is to recalculate n with A and D, 
 increased by their mean deviations and to consider the 
 difference between this value and the original value as the 
 possible error. 
 
 It is probable that in this experiment there are other 
 sources of error that exceed mere error in reading the 
 scale; e. g., (i) The faces of the prism may not be true 
 planes, (2) the divided circle may not be uniform, (3) the 
 center of the circular scale may not coincide with the cen- 
 ter of the instrument, (4) the various adjustments may not 
 be perfect, (5) there may be difficulty in fixing the position 
 
134 LIGHT. 
 
 of minimum deviation. These errors might be eliminated by 
 repeating all the adjustments and observations many times 
 and using different parts of the divided scale. There is no 
 other way of allowing for them. 
 
 Questions. 
 
 1. Give both physical and mathematical definitions of the refract- 
 ive index. 
 
 2. Why is monochromatic light used? 
 
 3. Why is the minimum deviation chosen? 
 
 XXXIV. MEASUREMENT OF RADIUS OF CURVATURE. 
 
 Glasebrook and Shaw, Practical Physics, pp. 339-343; Edser, Light, 
 pp. 116-121; Koklrausch, pp. 174-176. 
 
 The radius of curvature of a surface may be determined 
 from the size or position of the image which the spherical 
 surface, regarded as a mirror, forms of a definite object. 
 Method (A) below is especially applicable to the measure- 
 ment of the radius of curvature of convex surfaces, and 
 method (B) to concave surfaces. 
 
 (A) Two bright objects (see Fig. 36) are placed on a line 
 at right angles to the axis of the spherical surface, the 
 intersection of the line and the axis being at a considerable 
 distance A, from the surface, and each object being at a 
 distance L/ 2 from the axis. If the apparent distance be- 
 tween the images of the two objects be /, the radius of curva- 
 ture of the surface is 
 
 ~ 
 
 the + sign being used for a concave surface and the sign 
 for a convex. 
 
 Proof. 
 
 (For convex mirror.) 
 
 Let <i = true distance between the images, x = distance of images 
 from mirror. 
 By geometry 
 
 L A+r d_A+x . L (A+r)(A+x) 
 d~r-x' l~ A ' '' I "' (r-x)A ' 
 
MEASUREMENT OF RADIUS OF CURVATURE. 135 
 
 By the equation for spherical mirrors (see p 125). 
 \ 112 
 
 x r A r 
 
 r x_A+r 
 
 roc ~ Ar 
 
 L A + x A 
 
 .'. -y = -- =1 + . 
 / X X 
 
 From ( i ) = i H 
 
 OC T 
 
 L iA 
 
 
 The radius of curvature should be found for both surfaces 
 of a double convex lens. The lens, preferably the one used 
 in Exp. XXXV, if that has already been performed, 
 is fitted in a clamp in a darkened recess. At some distance 
 are two vertical slits illuminated from behind by incan- 
 descent lamps (or the lamp filaments themselves may be 
 used) and between them a *> 
 telescope. The telescope and 
 the lens are adjusted until, on 
 looking through the telescope 
 toward the lens, the illumi- 
 nated slits are seen reflected 
 from the near surface of the 
 lens. Distinguish these im- 
 
 ages from the images produced by the rear surface of the lens 
 by the change of focus necessary to make one pair of images 
 most distinct, and then to make the other pair most distinct, 
 or, by observing the two images of a light held just outside 
 one of the slits. Remember that the telescope inverts. 
 A paper scale is pinned over the lens so that the upper edge 
 is just below the center of the lens. The telescope is focused 
 upon the scale, and rotated until the vertical cross-hair 
 bisects one of the images from the near surf ace 'of the lens, 
 and the scale read where crossed by the cross-hair. (Esti- 
 
136 LIGHT. 
 
 mate tenths of millimeters as always.) A similar reading 
 is made for the other image. The difference between the 
 two readings gives the apparent distance, /, between the 
 images. At least six independent determinations of this 
 distance should be made. Measure the distance, L, between 
 the slits, the distance, A, from the lens surface to the line 
 joining the slits and substitute in the formula. 
 
 From the two radii of curvature and the focal length, if 
 known, calculate the refractive index, n, of the glass of the 
 lens by means of the formula 
 
 (B) As a concave mirror we may use one of the surfaces 
 of a concave lens, mounted in a lens-holder. To reduce 
 reflection from the other surface, the latter may be covered 
 by moist filter paper. The radius is determined from u, 
 the distance of the object, v, the distance of the image and 
 the formula 
 
 112 
 
 -+-=-, (seep. 125). 
 
 u v r 
 
 In locating the image, use is made of the fact that if the 
 eye is a considerable distance off, a real image can be seen 
 in space as well as a virtual image, and a wire, needle, or 
 pointer is moved about until there is no parallax between it 
 and the image; i. e., until, when the eye is moved about, 
 there is no relative motion of the two. 
 
 A vertical wire illuminated by a lamp, behind which is 
 a sheet of white paper, is a convenient object, and a second 
 mounted wire is moved about until it coincides with the 
 image of the first (see (B) Exp. XXXV). The image 
 should be found for at least the following three typical 
 positions of the object. For each position make several 
 settings and from the means determine u and v, and from 
 them determine r. 
 
 (i) Let the object be at a considerable distance from the 
 mirror. 
 
FOCAL LENGTH OF LENS. 137 
 
 (2) Let the object be at the center of curvature of the 
 mirror. In this position the image and the object coincide. 
 
 (3) Let the object be within the principal focus. For this 
 position the wire locating the image must be on the other 
 side of the lens. This wire is moved about until the pro- 
 longation above the lens of the image of the 
 
 first wire coincides with what is seen of the 
 second wire above the lens. 
 
 In the report, sketch the relative positions 
 of mirror, image, and object, and state whether ' 
 
 the image was magnified or diminished, erect 
 or inverted. 
 
 (C) If time permit, check your results with a 
 spherometer (see p. 16). The spherometer 
 should be read alternately on a plane surface 
 and on the lens. Let a = difference in the 
 two readings (see Fig. 37) and r = radius of the circle of the 
 legs. Then the radius of curvature of the lens is 
 
 as may be easily shown. 
 
 Questions. 
 
 1. For what lenses would the first method of determining the radius 
 of curvature be preferable, and when would the spherometer be 
 preferable ? 
 
 2. What objection is there to determining the radius of curvature 
 of the farther face of a convex lens, considering it a concave surface? 
 
 3. What advantages has the method used in (B) for locating real 
 images over the use of a screen ? 
 
 4. How could you directly determine with a screen the center of 
 curvature of a concave mirror? 
 
 XXXV. FOCAL LENGTH OF A LENS. 
 
 Text-book of Physics (Duff), pp. 392-398; Watson's Physics, pp. 471- 
 479; Ames' General Physics, pp. 470483; Crew's General Physics, 
 pp. 466-469; Edser, Light, pp. 110-116; Glazebrook and Shaw, 
 Practical Physics, pp. 343-352. 
 
 The focal length of a lens is the distance from the optical 
 center of the lens to the focus for rays of light from an 
 
138 LIGHT. 
 
 infinite distance; i. e., for plane waves. If / is the focal 
 length, u the distance of the object from the lens, and v 
 that of the image, then, with the convention respecting signs 
 given on page 125, for all lenses 
 
 iii 
 
 v u f 
 
 (A) Real Image. (li Exp. XXXIV has preceded, use 
 the same lenses.) An "object," the lens, and a screen 
 for receiving the image of the object, are mounted so that 
 they can be moved along a graduated scale. A convenient 
 form for the object is a wire cross or gauze, mounted in a 
 black wooden support, and illuminated from behind by an 
 incandescent lamp. The lens is clamped in a wooden frame 
 movable along the scale. This should grasp the lens on the 
 sides, leaving the top and bottom clear. The distance from 
 the center of the lens to some point on the support must be 
 determined once for all and applied as a correction to the 
 readings. With object and screen in fixed positions that are 
 recorded, the lens is adjusted until the image 
 on the screen is as distinct as possible and 
 its position is then recorded. This should 
 be done several times, and the mean taken 
 for the position of the lens. Keeping object 
 and screen fixed and moving the lens about, 
 , another image will be found, for which similar 
 
 observations should be made. Calculate / from the averages 
 of all the values of u and v. The object and screen should 
 then be shifted and the observations repeated. From the 
 two sets of observations a mean value of / is deduced. 
 
 Study of Spherical Aberration. Determine / for the central 
 part of the lens by covering, with a pasteboard screen, 
 all but a central disk of about one-third the diameter of 
 the lens. Similarly determine / for the edge of the lens, 
 using a diaphragm covering all but the edge. 
 
 Study of Chromatic Aberration. Using the entire lens, 
 
FOCAL LENGTH OF LENS. 139 
 
 determine / for red light by placing red glass before the lens 
 or object, and similarly for blue or green light. 
 
 In the report, tabulate, for comparison, the different 
 mean values of /. 
 
 (B) Virtual Image. In the preceding a real image was 
 observed, but the focal length may also be found from 
 observations of a virtual image. The following directions 
 apply to a divergent lens. A vertical dark line on a white 
 background serves as object. The image (between the lens 
 and the object) is located with a short vertical wire, which 
 is moved back and forth until a position is found where 
 the image of the dark line seen through the lens (a in Fig. 38) 
 appears at the same distance as the portion of the wire seen 
 just below or above the lens (b in Fig. 38). This is secured 
 when there is no relative motion of the image and this wire 
 as the eye is moved horizontally, i. e., the wire appears as 
 the prolongation of the. image of the dark line or remains 
 equidistant from such a prolongation, v will be the distance 
 from the center of the lens to this wire which locates the 
 image. Using a longer wire as the object and the dark line 
 to locate the image, this method may be applied to the 
 virtual image of a convergent lens. 
 
 Estimate the possible error of a typical measurement 
 of /. . Since practically all the error is in the location of the 
 lens, the distance between the object and the screen may 
 be considered free from error. If this distance is desig- 
 nated by w , the formula becomes 
 
 , 
 
 u wu j 
 
 from which a formula may easily be derived for the possible 
 error in / in terms of the possible error in u. The latter may 
 be taken as the mean deviation from the mean in the location 
 of the lens (see p. 4). 
 
 If Exp. XXXIV has preceded, determine the refract- 
 ive index of the glass from the focal length and the radii 
 of curvature. 
 
140 LIGHT. 
 
 Questions. 
 
 1. What is the minimum distance between object and screen to 
 secure a real image? The maximum distance between object and 
 lens to secure a virtual image? 
 
 2. What advantage is there in covering with a diaphragm all but 
 the central portion of a lens? What disadvantage? 
 
 3. What is the cause of chromatic aberration? 
 
 4. What sort of a lens would show large spherical aberration? 
 Large chromatic aberration? 
 
 XXXVI. LENS COMBINATIONS. 
 
 Edser, Light, Chaps. VI, VII, X; Watson's Practical Physics, pp. 358- 
 367; Drude's Optics, pp. 44-46, 66-72; Hastings' Light, Appendix 
 A. Antes' General Physics, pp. 488, 493, 494. 
 
 (A) Determination of Principal Foci. Calculation from 
 Focal Lengths and Separation. Let two lenses of focal lengths, 
 /!, and/ 2 , be separated a distance d. An object at a distance 
 u from the first lens forms an image at a distance v deter- 
 mined by the equation 
 
 v ^ u uf l ' 
 
 This image acts as an object for the second lens at a 
 distance d v. Hence the distance of the final image from 
 the optical center of the second lens is given by the equation 
 
 i i i i u ~fi 
 
 v' /a d ~ v / 2 du-dfi-ufi 
 
 If u is infinite, v r is the distance, V, of the principal focus 
 from the optical center of the second lens, 
 
 Experimental Location. Determine experimentally the 
 position of this principal focus by finding the position in 
 which an object must be placed for clear vision when it is 
 viewed through the combination with a telescope focused 
 for a very distant object. Compare with the calculated 
 position. Determine similarly the other principal focus. 
 
LENS COMBINATIONS. 141 
 
 (B) Determination of Focal Length. To determine the 
 focal length, the position of the "principal points" must 
 be known, as well as the principal foci. With thin lenses, 
 the principal points practically coincide at the so-called 
 "optical center." In thick lenses or lens combinations they 
 may be considerably separated. The focal length is the 
 distance from either principal focus to the nearer principal 
 point. The equations for locating the image, 
 
 iii 
 -+-=7, 
 it v f 
 
 and for finding the linear magnification, 
 
 7' 
 
 M = - 
 
 u 
 
 are applicable in all cases, if u and v are measured from the 
 principal points. 
 
 Since it is difficult to locate the principal points, a method 
 is often employed for determining the focal length which 
 eliminates their position. Suppose that the linear mag- 
 nification is M lt when the distance of the object from the 
 principal plane is u lt and M 2 when the object is moved 
 until its distance is u. 
 
 If the focal length is small, the magnification should 
 be determined with a micrometer microscope. A carefully 
 graduated scale is a convenient object and the size of the 
 image of one or more divisions is measured for two positions 
 of the object a known distance apart. (Principle of Abbe's 
 Focometer.) 
 
 Thus determine the focal length of the combination 
 used in (A). Determine also the focal length of both the 
 objective and the eye-piece of a telescope, using the same 
 
142 LIGHT. 
 
 one as employed in Exp. XXXVII, if that has pre- 
 ceded, and calculate the magnifying power for great 
 distances. 
 
 In the report, draw careful figures of the lens combina- 
 tions, representing the principal foci and principal points 
 as calculated from the final mean results. 
 
 (C) If the principal points of a convergent system are 
 close together, i. e., if the lens or lenses may be said to have 
 an optical center, we may use the following approximate 
 method: If w is the distance between object and image, 
 and x that between the two positions of the lens for real 
 images, u=(wx)/2, v=(wx)/2. Substituting these 
 values in the formula we get 
 
 Zf 2 X 2 
 
 4W 
 
 If time permit, try this method for the combination of 
 lenses. 
 
 Describe a lens combination which (i) magnifies without 
 distortion; (2) magnifies without chromatic aberration; 
 (3) inverts without magnifying. (See references.) 
 
 XXXVII. MAGNIFYING POWER OF A TELESCOPE. 
 
 Glazebrook and Shaw, pp. 358363; Watson's Practical Physics, pp. 
 
 367, 368. 
 
 The magnifying power of a telescope is the ratio of the 
 angle subtended at the eye by the image as seen through 
 the telescope to the angle subtended by the object viewed 
 directly. (If Exps. XXXVI and XXXVIII have already 
 been performed, use the telescope employed in those 
 experiments.) 
 
 (A) Direct Method. A minute mirror is attached to the 
 telescope by wax so as to make an angle of about 45 with 
 the axis and partly cover the aperture of the eye-piece. 
 The telescope is focused upon a scale. A second scale is 
 mounted parallel to the first and near the eye-piece, in such a 
 
MAGNIFYING POWER OF A TELESCOPE. 143 
 
 position that the observer's eye sees, side by side, the image 
 of the scale viewed though the telescope and the image of 
 the other scale reflected in the small mirror. From the 
 ratio of the images of one or more scale divisions, and their 
 distances, the angles are calculated, and from their ratio the 
 magnifying power is deduced. 
 
 Find the magnifying power for at least six distances, 
 making several observations for each. Also determine 
 the angular field of view of the telescope by determining for 
 each distance, r, the total distance on the scale, n, visible 
 in the telescope. The angular field of view, in degrees, will be 
 
 The magnifying power defined above is very approximately 
 equal to (i) the ratio of the magnitude of the image to the 
 magnitude of the object when the two are in the same 
 plane, and, for great distances, is equal to (2) the ratio of 
 the focal length of the objective to the focal length of the 
 eye-piece. 
 
 .* 
 
 ~~---K)* 
 ---'' f> 
 
 FIG. 39. 
 
 The eye-piece is of such short focus that the angle subtended by 
 its image is practically the same as if the image were at infinity. For 
 convenience we will consider the virtual image P' Q' (see Fig. 39) 
 produced by the eye-piece to be at the same distance as the object 
 
 Since the telescope usually views objects at distances great com- 
 pared with its own length, the angle subtended by the object viewed 
 directly is practically P a Q = p a q, and that subtended by the image 
 is P' b Q f = p b q. The ratio of these two angles, which may be taken 
 as the ratio of the tangents, since the angles are small, = a c -r- c b = the 
 ratio of the focal length of the objective to the focal length of the 
 eye-piece; and also, since the length of the telescope is short compared 
 with the distance of the object, this ratio = P' Q' + P Q, or the ratio 
 of the magnitude of the image to the magnitude of the object. 
 
144 LIGHT. 
 
 (B) The first approximate statement of the magnifying 
 power furnishes another method for determining the mag- 
 nifying power for different distances of the object. The 
 telescope is directed toward a horizontal scale. The scale 
 is viewed through the telescope with one eye and is also 
 observed with the other eye by looking along the outside 
 of the telescope. The eye-piece is moved in or out until the 
 image appears at the same distance as the scale as viewed 
 outside the telescope with the other eye, i. e., until there is 
 no parallax between the scale and its image (no relative 
 motion of the two as the eye is moved about). It may 
 require some practice to secure this. Determine the number 
 of divisions on the scale which, as viewed directly, are 
 covered by the image of one or two large divisions as viewed 
 through the telescope. If it is difficult to read the division 
 on the scale viewed directly, two black strips may be moved 
 along the scale until they include the image of one or more 
 divisions as seen through the telescope, and the distance 
 between these strips read off. Repeat the measurements 
 of (A). 
 
 (C) The second method of defining the magnifying power 
 of a telescope is useful in determining the magnifying power 
 
 'for very distant objects. Focus the telescope on some very 
 distant object. Without changing the focus, remove the 
 object glass and substitute for it a diaphragm with a rec- 
 tangular opening. The ratio of the focal length of the 
 objective to the focal length of the eye-piece is the ratio of a 
 linear dimension of the aperture of the diaphragm, L, to the 
 corresponding dimension, /, of the image of this aperture 
 produced by the eye-piece. 
 
 Since the telescope was focused for parallel rays, the distance, u, 
 of the object, L, from the eye-piece is numerically very nearly the 
 sum of the focal lengths, P+f (Fig. 40). .'.the distance of the 
 image, /, formed by the eye-piece, is determined by 
 
 i = _-i i -f+(F+f) F 
 
 v (F+fVf- /(F+/) /(F+/)' 
 Hence, 
 
RESOLVING POWER OF OPTICAL INSTRUMENTS. 145 
 
 To measure /, a micrometer microscope (see p. 15) may 
 be used, the microscope being in line with the axis of the 
 telescope and focused upon the real image in space. L may 
 
 FIG. 40. 
 
 be measured with vernier calipers, or the same micrometer 
 microscope may be placed opposite the other end of the 
 telescope and L measured in the same way as /. 
 
 Questions. 
 
 1. Explain why the magnifying power should vary as you have 
 found it to do with the distance of the object. 
 
 2. Which is preferable to gain magnifying power by increasing 
 the focal length of the objective, or by decreasing the focal length of 
 the eye-piece? Why? 
 
 XXXVIII. RESOLVING POWER OF OPTICAL 
 INSTRUMENTS. 
 
 Text-book of Physics (Duff), pp. 421, 422; Watson's Practical Physics, 
 pp. 335338; Antes' General Physics, pp. 483487; Mann, 
 Advanced Optics, pp. 11-18; Drude's Optics, pp. 235-236; Hast- 
 ings' Light, pp. 7072. 
 
 The magnification obtained with an optical instrument 
 depends upon the focal lengths of its lenses, as has been 
 seen in the case of the telescope. The ability to distinguish 
 details of the image, i. e., the "resolving power," depends 
 on the diameter of the aperture through which light enters 
 the instrument. 
 
 If d is the distance between two details of an object at 
 a distance D from an aperture whose width parallel to these 
 details is a, they may be distinguished if 
 
 d I 
 
 D>a 
 
 10 
 
146 LIGHT. 
 
 where / is the wave-length of the light employed. If the 
 aperture is circular, the equation is 
 
 d I 
 
 where a is the diameter of the aperture. 
 
 (A) Resolving Power of Telescope. Metal gauze answers 
 as a very satisfactory object for studying the resolving 
 power, as it gives a great amount of uniform detail. Since 
 this detail consists of rectangular lines, and the aperture of 
 the object glass is circular, the determination of the maxi- 
 mum distance at which the lines are discernible will be more 
 definite, if the aperture is made rectangular by placing a slit 
 in front of the object glass. 
 
 Determine carefully by several settings, the maximum 
 distance, D, at which the lines of the gauze, parallel to the 
 slit, are perceived. The gauze should be illuminated from 
 behind by monochromatic light (p. 124). Measure carefully 
 the distance, d, between the centers of the adjacent wires 
 of the gauze, and the width of the slit, a. Compare d/D 
 with //a; / may be obtained from Table XVIII. Repeat 
 with other slits and other gauzes. 
 
 (B) Resolving Power of Eye. With Porter's apparatus 
 the resolving power of the eye may be determined for 
 various apertures. 
 
 Four different gauzes, 37 .6, 27, 20, and 14, meshes to the 
 cm., respectively, may be viewed through four different 
 apertures of diameters i.oo mm., 0.65 mm., 0.53 mm., 
 and 0.35 mm., respectively. The resolving power is de- 
 temined by finding the distance, D, from a slit of diameter 
 a to a gauze, of which the distance between the centers of 
 two adjacent wires is d, when the wires are separately dis- 
 cernible. From the mean position of several settings of a 
 particular gauze for a particular aperture, d/D should be 
 calculated and compared with 1.2 I /a. If ordinary light is 
 used, / may be taken as 0.00006 cm. Use each aperture 
 and gauze in succession. 
 
WAVE-LENGTH OF LIGHT BY DIFFRACTION GRATING. 147 
 
 Questions. 
 
 (1) Upon what does the illumination of the image of an optical 
 rnstrument depend? 
 
 (2) When the diameter of the pupil of the eye is 4 mm., how far 
 away may two points be distinguished which are o. 2 mm. apart? 
 
 XXXIX. WAVE-LENGTH OF LIGHT BY DIFFRACTION 
 
 GRATING. 
 
 Text-book of Physics (Duff), pp. 423, 437; Watson's Physics, pp. 
 529-532; Ames' General Physics, pp. 530-537; Crew's General 
 Physics, pp. 488491; Edser, Light, pp. 448-458; Wood, Physical 
 Optics, pp. 168-180. 
 
 A diffraction grating consists of a great many lines ruled 
 parallel and equidistant on a plane (or concave) surface. 
 If the surface be that of glass, the grating is a transmission 
 grating; if of metal, a reflection grating. If a transmission 
 grating be placed perpendicular to homogeneous parallel 
 light from a collimator (see Exp. XXXIII) and with 
 the lines parallel to the slit, a series of spectra will be formed 
 on either side of the beam of light transmitted without 
 deviation. If n be the number or order of a particular 
 spectrum counting from the center, 6 the deviation or angle 
 that the rays forming the spectrum make with the original 
 direction of the light, a the grating space or average distance 
 between the centers of adjacent lines, and / the wave-length 
 of the light 
 
 I a sin 6. 
 n 
 
 The deviation 6 may be observed by placing the grating 
 on a spectrometer (see Exp. XXXIII where the ad- 
 justments of the spectrometer are described). The 
 position of the telescope when in line with the collimator is 
 read. The grating is adjusted parallel to the axis of rotation 
 of telescope and collimator as one face of a prism is adjusted. 
 For convenience in adjusting, the plane of the grating should 
 be perpendicular to the line of two of the leveling-screws. 
 This enables us to adjust the lines of the grating parallel 
 
148 
 
 LIGHT. 
 
 to the slit by means of one leveling-screw without altering 
 the plane of the grating. The lines are parallel to the slit 
 when the spectrum of some homogeneous light, e. g., from a 
 sodium flame (p. 124), is as distinct as possible. When the 
 plane of the grating is perpendicular to the incident light, the 
 deviations (on opposite sides) of the two spectra of the same 
 order should be equal. This adjustment is also secured 
 when that part of the beam which is 
 reflected back to the collimator appears 
 co-axial with its object glass. 
 
 Determine first the wave-length of 
 sodium light. For a final measure- 
 ment of the deviation of any spectrum 
 the mean of at least three measure- 
 ments on each side should be taken. 
 The deviations of all the spectra clearly 
 visible should be obtained. 
 
 If the grating space be not too small 
 it may be obtained by measurements 
 on a dividing engine (p. 17), or with a 
 micrometer microscope (p. 15). In determining the grating 
 space with the dividing engine, secure the best possible 
 illumination of the lines. Set the cross-hair of the micro- 
 scope on a line and read the position of the divided head 
 (circular scale). Watching the lines through the micro- 
 scope, turn the screw, always in the same direction, until, 
 for example, the tenth line is under the cross-hair, and read 
 the circular scale. Then turn the screw until the tenth line 
 from this is under the cross-hair, read the scale, and so on. 
 Take ten such groups in different parts of the grating. 
 Find the average grating space from the mean. When the 
 grating space is very small, the wave-length of some well- 
 known spectrum (e. g., sodium) is assumed in order that the 
 grating space may be derived by reversing the process of 
 finding the wave-length. 
 
 If time permit, determine the wave-length of as many 
 other lights (colors) as possible (see p. 124). 
 
 FIG. 41. 
 
INTERFEROMETER. 149 
 
 Questions. 
 
 1. What do you observe as regards the width of spectra of 
 different orders? What would this indicate as regards the disper- 
 sion if mixed light or light from an incandescent solid were used ? 
 
 2. What is a normal spectrum and wherein does a prismatic 
 spectrum differ from a diffraction spectrum? (See references.) 
 
 XL. INTERFEROMETER. 
 
 Text-book of Physics (Duff), p. 438; Watson's Physics, p. 540; Mann, 
 Advanced Optics, Chap. V; Wood, Optics, Chap. VIII; Michelson, 
 Light Waves and Their Uses. 
 
 The interferometer is an instrument for determining the 
 number of wave-lengths of a monochromatic light con- 
 tained in a given distance. For a description of the inter- 
 ferometer and the adjustments, see the references. 
 
 The interferometer will be used to determine the wave- 
 length of sodium light, assuming a knowledge of the true 
 
 pitch of the screw. This will | i 
 
 illustrate the more practical 
 and common, but also more 
 difficult, utilization of the in- 
 terferometer in determining a 
 length, assuming a knowledge 
 of the wave-length of the light 
 employed. 
 
 The light 5 (see figure) 
 had best be monochromatic, FlG 42 
 
 e. g., a sodium flame. Initially, 
 
 place the mirror D at approximately the same distance 
 from the rear face of A as the distance from this surface 
 to C. Adjust C until its image coincides with either of 
 the images from D. (There will be two images owing to 
 reflection from the two faces of A.) A slight adjustment 
 will now give the fringes (alternate light and dark bands, 
 preferably arcs of circles) . The observer must look at A in 
 a direction parallel to AD. 
 
 Move the mirror D by means of the worm, and count 
 
150 LIGHT. 
 
 the number of fringes which pass over the field of view. 
 A needle in front of A may help as an index. 
 
 From the number of turns and fractional turns of the 
 screw and the value of the pitch of the screw, find the 
 distance D has moved and from this and the number of 
 fringes which have passed, calculate the wave-length. 
 Notice that the length of the path of the light changes by 
 twice the displacement of the mirror D. 
 
 XLI. ROTATION OF PLANE OF POLARIZATION. 
 
 Text-book of Physics (Duff), pp. 476-479; Watson's Physics, pp. 580- 
 582; Ames' General Physics, pp. 563-565; Watson's Practical 
 Physics, pp. 370377; Edser, Light, pp. 503509; Wood, Optics, 
 Chap. XIV; Ewell, Physical Chemistry, pp. 217-223. 
 
 Plane polarized light is obtained by passing light through 
 a Nicol prism. If the light be then allowed to fall on a 
 second Nicol prism that can be rotated, there will be two 
 positions of this second prism in a complete rotation in 
 which no light will pass through. If an optically active 
 substance, such as a solution of cane sugar, be then intro- 
 duced between the two Nicols, it will rotate the plane of 
 polarization of the light which falls on the second prism, 
 and then, to quench the light, the second prism must be 
 rotated through an equal angle. Thus the rotation pro- 
 duced by the sugar is measured. 
 
 Monochromatic light must be used and a sodium flame 
 is most convenient (see p. 124). The light rays must be 
 made parallel before they fall on the polarizing prism, other- 
 wise rays in different directions would pass through different 
 thicknesses of the sugar and would consequently be rotated 
 by different amounts. Parallel light may be obtained by 
 putting the source at the principal focus of a convex lens 
 through which the light has to pass before falling on the 
 polarizing Nicol. The light must also be parallel to the axis 
 of the Nicol. 
 
ROTATION OF PLANE OF POLARIZATION. 151 
 
 The empty tube intended to contain the sugar solution is 
 first placed in position between the prisms and the position 
 of the analyzing Nicol noted, on the circular scale, when the 
 light is quenched. This setting will be facilitated by using 
 a screen to cover all but a small central part of the prism. 
 It may be found that the Nicol can be rotated through an 
 appreciable angle without the light reappearing. The best 
 that can be done is to take the middle of this space as the 
 position of extinction. The observation should be repeated 
 a number of times and the mean taken. The analyzing 
 Nicol should then be rotated through 180 and the zero 
 reading in that position also noted. Sugar solutions of 
 different strengths (which should be carefully made up and 
 recorded) are then introduced in succession into the tube and 
 the rotations they produce observed. The zero readings 
 should be frequently repeated. The length of the tube 
 should also be obtained, so that the rotation per decimeter 
 may be deduced. With the results obtained, a curve 
 should be plotted, rotations per decimeter being ordinates 
 and concentrations abscissae. 
 
 FIG 
 
 43- 
 
 The apparatus here described is simple but very imperfect 
 in its action. The sensitiveness is greatly increased by 
 introducing between the polarizer and the specimen a so- 
 called biquartz, two parallel, abutting, plates of quartz, 
 one with left rotatory power and the other with right. A 
 source of white light must be employed, for example, a 
 frosted incandescent bulb, and the analyzer is set for equality 
 of color in the two halves. There are two common colors, 
 but the darker is preferable. 
 
 Fig. 44 explains the color changes. R and L are the two halves 
 of the bicjuartz, viewed from the analyzer. The two halves are of 
 such a thickness (3.75 mm.) that the plane of polarization of yellow 
 
152 
 
 LIGHT. 
 
 light is rotated through 90. Owing to the rotatory dispersion 
 the other colors will be rotated different amounts as shown by the 
 letters R (red) and B (blue). If the analyzer is set to transmit light 
 vibrations parallel to those which left the polarizer, the yellow light 
 will be omitted and each half of the biquartz will appear of a purplish 
 color ("tint of passage"). If the analyzer 
 is displaced slightly clockwise, more of the 
 red component on the right will be trans- 
 mitted and less of the blue, and therefore 
 this half will appear red and the other 
 half will appear blue. 
 
 If a dextrorotatory specimen is placed 
 between the biquartz and the analyzer, 
 the directions of vibration of the different 
 colors will be rotated to the positions 
 indicated by the dotted lines and the 
 analyzer must be rotated to a new position 
 (/') , perpendicular to the emerging yellow 
 vibration, in order to have the two halves 
 the same color. With the help of the 
 biquartz the analyzer can be set within 
 about a tenth of a degree. 
 
 The effective thickness of the 
 biquartz will not be correct (90 
 degrees rotation of sodium light), 
 FIG. 44. unless it is perpendicular to the axis 
 
 of the Nicol prisms. This may be 
 
 secured by using sodium light and analyzer set for extinc- 
 tion, and then placing the biquartz in such a position that 
 there is still extinction when the analyzer is rotated 90. 
 
 Repeat all the measurements with white light and the 
 biquartz and plot the results on the same sheet with the 
 preceding. 
 
 QUESTIONS. 
 
 1. How can the rotation be partially explained ? (See references.) 
 
 2. What is the chemical characteristic of substances that are 
 optically active in solution ? 
 
 3. Wherein does the rotation produced by a solution differ from 
 that produced by a magnetic field? 
 
 4. What would be the effect of using white light in the first part 
 of the experiment? 
 
ELECTRICITY AND MAGNETISM. 
 
 29. Resistance -boxes. 
 
 A resistance-box consists of a number of resistance coils 
 joined so that each one bridges the gap between two of a 
 series of brass blocks placed in line on the cover of the box 
 within which the coils are suspended. For each gap a plug 
 or connector is also provided, and when the plug is inserted 
 into the gap the resistance at the gap is " cut out " or practi- 
 cally reduced to zero. The coils are wound so as to be free 
 from self-induction. The successive resistances are arranged 
 in the same order, and are of the same relative magnitudes 
 as the successive weights in a box of weights. By removing 
 the proper plugs any combination of resistances can be 
 obtained from the smallest to the sum of all. Before begin- 
 ning work, it is advisable to clean the plugs with fine emery- 
 cloth so that they may make good contacts, and thereafter 
 care should be taken not to soil them with the fingers. 
 
 One important precaution in regard to the use of the 
 resistance-box should be observed. If any of the plugs 
 are in loosely, there will be some resistance at the contact. 
 Hence, the plug should be screwed in firmly, but not vio- 
 lently. When any* one plug has been withdrawn, the others 
 should be tested before proceeding, for the removal of one 
 may loosen the contact of the others. This precaution is 
 especially important in making a final determination. 
 
 30. Forms of Wheatstone's Bridge. 
 
 The practical measurement of a resistance consists in 
 comparing it with a known or standard resistance. Wheat- 
 stone's Bridge is an arrangement of conductors for facilitat- 
 ing this comparison, and consists essentially of six branches 
 
 153 
 
ELECTRICITY AND MAGNETISM. 
 
 which may be represented by the sides and diagonals of a 
 parallelogram (see Fig. 45). The unknown resistance, 
 R, and the known resistance, 5, form two adjacent sides. 
 The other two sides are formed by two conductors of resist- 
 ance P and Q, which, however, do not need to be known 
 separately, provided their ratio be known. One of the- 
 
 diagonals contains a battery 
 and the other a galvanom- 
 eter. If the ratio of P to 
 Q is adjusted until no current 
 flows through the galvanom- 
 eter, R:S::P:Q. (See refer- 
 ences under Exp. XLIV.) 
 
 Two forms of the Wheat- 
 stone " Bridge arrangement 
 are in common use. One is 
 called the Wire (or meter) 
 Bridge; the other, which uses 
 
 /r 
 
 FIG. 45. 
 is called a Bridge Box. 
 
 a box of adjusted resistances, 
 In the wire bridge the "ratio 
 arms" (whose resistances are P and Q) are the two parts of 
 a uniform wire i meter long, and the ratio of P to Q is that 
 of the lengths of the corresponding parts of the wire. The 
 known resistance, 5, may be that of a standard coil or one 
 of the known resistances of a resistance-box. 
 
 The Bridge Box, or "Post-office Bridge," consists of a 
 resistance-box with three series of resistances in line, forming 
 three arms of the Wheatstone Bridge, the unknown resist- 
 ance forming the fourth arm. The "ratio arms" consist of 
 resistances of i, 10, 100, 1000 (all of which are not always 
 necessary), so that the calculation of the ratio is very 
 simple. Keys for closing the battery and galvanometer 
 branches are also usually mounted on the box. 
 
GALVANOMETERS. 155 
 
 31. Galvanometers. 
 
 Text-book of Physics (Duff), pp. 559-563; Hadley's Electricity and 
 Magnetism, pp. 273-284; Watson's Practical Physics, 170-174; 
 Ames and Bliss, Appendix iii. 
 
 There are two chief types of reflecting galvanometers. 
 In both the principle at basis is that if a magnet be placed 
 in the plane of a coil of insulated wire, on passing a current 
 through the coil both magnet and coil become subject to 
 forces that tend to set them at right angles to each other. 
 In the Thomson type the coil is fixed and the magnet 
 suspended within the coil is free to turn, while in the 
 d'Arsonval type the magnet, of a horseshoe form, is fixed, 
 and the coil, suspended between the poles of the magnet, 
 is free to turn. 
 
 The sensitiveness of the Thomson galvanometer is greatly 
 increased in two ways: first, two magnetic needles, forming 
 an astatic pair, are attached to the same axis of rotation, 
 second, an external control magnet is used to weaken the 
 restraint of the earth's magnetic force or even to overcome" 
 the earth's field and produce a suitable field of its own. 
 The chief difficulty in greatly increasing the sensitiveness 
 by means of the control magnet is that slight variations of 
 the whole magnetic field, due to outside currents or move- 
 ments of magnetic materials in or near the laboratory, 
 disturb the needle. 
 
 On the d'Arsonval galvanometer variations of the external 
 magnetic field have practically no effect, since its own 
 magnetic field is very strong. On the other hand, the 
 torsion of the fine suspending wire through which the 
 current has to pass changes somewhat with the temperature, 
 so that the zero reading of the galvanometer is subject to 
 some change. The sensitiveness can be increased by in- 
 creasing the strength of the magnet, but there is a limit to 
 this, since small traces of iron are always present in the 
 wire and insulation of the coil, and this, acted on by the 
 magnetic field, exercises a magnetic control that is pro- 
 portional to the square of the strength of the field. When 
 
156 ELECTRICITY AND MAGNETISM. 
 
 an extremely sensitive galvanometer for very accurate 
 work is required, the Thomson type must be used. 
 
 A ballistic galvanometer is a reflection galvanometer of 
 either type, so made that its period of swing is very long, 
 so that it starts into motion only very slowly. If this con- 
 dition be fulfilled, and if it be subject to only very slight 
 damping of its motion, the galvanometer may be used for 
 comparing quantities of electricity suddenly discharged 
 through the coils of the galvanometer, for, practically 
 speaking, all the electricity will have passed before the 
 swinging system has appreciably moved from this position 
 of rest. In these circumstances it can be shown that the 
 quantity of electricity is proportional to the sine of half the 
 angle of the first swing, or (since the angle is very small) 
 practically to the deflection as read on the scale. 
 
 Two methods of reading the deflection of a galvanometer 
 are in common use. In one, called the English or objective 
 method, a beam of light reflected from the mirror of the 
 galvanometer falls on a scale, forming a spot of light which 
 moves as the needle or coil is deflected. In the other, 
 called the German or subjective method, the image of a scale 
 formed by the mirror of the galvanometer is read by a 
 telescope with a cross-hair. 
 
 Devices for Bringing a Galvanometer to Rest. For bringing 
 to rest the needle of a ballistic Thomson galvanometer a 
 coil is mounted on the outside of the galvanometer in front 
 of the lower needle. The terminals of the coil are brought 
 to a reversing switch by which the current from a cell can 
 be sent through the coil in either direction. By suitably 
 choosing the direction and duration of the current, the 
 needles and mirror may be brought to rest. (A current 
 in this coil affects the needle in the same manner as would a 
 current in one of the regular galvanometer coils, but it is 
 much more convenient to use a separate coil like this, which 
 is readily accessible and which does not interfere with the 
 other connections.) 
 
 The suspended system of either type of galvanometer 
 
GALVANOMETER SHUNTS. 157 
 
 may also be brought to rest by short-circuiting the galvan- 
 ometer by a simple key directly connected to the terminals. 
 For, by Lenz's Law, the currents induced are such as to 
 bring the moving coil or needle to rest. If the resistance 
 of the coils is high, this method is slow, and the following 
 more rapid method may be used. A coil in which a small 
 bar magnet can be moved is placed in series with the short- 
 circuiting key. By suitably moving the magnet in and out, 
 currents are induced which will quickly bring the suspended 
 system to rest. 
 
 32. Correction for Damping of a Ballistic Galvanometer- 
 
 Kohlrausch's Physical Measurements, 51; Stewart and Gee's Practical 
 Physics, II, pp. 364369. 
 
 In considering the throw proportional to the charge 
 passing through the coils of a ballistic galvanometer, we 
 assume that the galvanometer is free from damping; i. e., 
 that the suspended system, needles, mirror, etc., experiences 
 no resistance to turning. Since this is never realized, a 
 correction must be applied to the throw. 
 
 The correction is not of importance where we compare 
 throws, since the correction cancels out, but in much 
 work with ballistic galvanometers this correction is very 
 important. 
 
 Set the needle vibrating and record n + i successive turn- 
 ing-points. From these we obtain by successive subtrac- 
 tion n successive full vibrations of the needle from one 
 side to the other. Call the first full vibration a l and the 
 last a n . Then the correction by which each throw should be 
 multiplied is (i -M/ 2 ) where 
 
 ^^loga 1 -logq M 
 
 n i 
 
 \ 
 
 33. Galvanometer Shunts. 
 
 If a galvanometer of resistance G is shunted by a shunt 
 of resistance 5 and if C is the whole current and C l the 
 current through the galvanometer 
 
ELECTRICITY AND MAGNETISM. 
 
 C G+S' 
 
 Galvanometers are frequently supplied with shunt-boxes 
 in which the ratio of 5: G are 1/9, 1/99, 1/999, so "that the 
 values of 5 :(G +S) are i/io, i/ioo, 
 i/iooo. Such a shunt-box cannot 
 easily be used with any galvanom- 
 eter except that for which it was 
 designed. 
 
 Universal shunt-boxes are now 
 made which can be used with any 
 galvanometer. Such a box consists 
 of a series of high, resistances con- 
 nected as indicated in the figure. AB 
 is a coil, of resistance S, connected to the galvanometer, of 
 resistance G. Let the current through the galvanometer 
 be C lt and let the whole current be C. Then as above 
 
 c,= cs 
 
 VVvNA/VV\AAAA/V j? 
 
 FIG. 46. 
 
 Now let the battery circuit be connected to A and P, where 
 the resistance of AP is S/n. Denoting the current through 
 the galvanometer by C/, and the whole current by C and 
 making the proper changes in the above equation, 
 
 Cy_ Sin _i 5 
 
 ~C ~S/n + (S-S/n)+G ~n S+G' 
 i CS 
 
 nS+G 
 
 Hence when a current is connected to A and P, the galvan- 
 ometer deflection is i/n as great as when the same current 
 is connected to A and B, or the sensitiveness is i/n as great. 
 By subdividing AB, the values of 3, 10, 100, etc., are given 
 to n. 
 
 Shunting a Ballistic Galvanometer. The formulas stated 
 above were deduced from Ohm's Law for steady direct 
 currents. It can, however, be shown that shunts like the 
 
STANDARD CELLS. 159 
 
 above may be used in the same way with ballistic galvan- 
 ometers through which charges of electricity are passed. 
 To prove this, all we need to do is to show that charges, 
 like steady direct currents, divide in a parallel arc into parts 
 inversely as the ohmic resistances. Consider any one of 
 several branches in a parallel arc. Let the part of the charge 
 that passes through it be q, and let the magnitude of the 
 instantaneous current through it, at time t after the begin- 
 ning of the discharge, be *\. The induced e. m. f. at that 
 moment is L^dijdt where L t is the self -inductance of the 
 branch. Suppose the discharge is caused by connecting an 
 e. m. f. to the parallel arc for a short time and then dis- 
 connecting it, and let the whole time of rise and fall of the 
 brief current be T. Then 
 
 E C T L C 
 
 -! dt-\ di v 
 
 R ijo R ijo 
 
 ET 
 
 Hence the charges through the various branches are in- 
 versely as their ohmic resistances. If the above proof be 
 carefully examined, it will be seen that it simply means 
 that the total quantity due to the induced e. m. f . is zero, since 
 the induced current in the first half of the process is op- 
 posite to that in the second half. 
 
 34. Standard Cells. 
 
 Text-book of Physics (Duff), pp. 584-585; Watson's Physics, pp. 806- 
 807; Watson's Practical Physics, 202-203; Bureau of 
 Standards Bulletin, Nos. 67, 70, 71; Henderson's Electricity and 
 Magnetism, pp. 176-182. Ewell, Physical Chemistry, pp. 334-336. 
 
 The standard Daniell cell consists of an amalgamated 
 zinc rod dipping into a porous cup containing a solution of 
 
l6o ELECTRICITY AND MAGNETISM. 
 
 sulphate of zinc, which, in turn, stands in a glass vessel 
 containing a copper sulphate solution and a copper plate. 
 To amalgamate the zinc rod, thoroughly clean it with sand- 
 paper, dip it in dilute sulphuric acid, and rub over it a few 
 drops of mercury with a cloth. The porous cup should be 
 thoroughly cleaned inside and out. The copper plate should 
 be cleaned bright with sand-paper. The porous cup is half- 
 filled from a stock bottle with a solution of zinc sulphate 
 (44. 7 g. of crystals of c. p. zinc sulphate dissolved in 100 c.c. 
 of distilled water). The zinc rod is introduced and the 
 porous cup is placed in the glass vessel, which is filled, 
 not quite up to the level of the zinc sulphate in the porous 
 cup, with copper sulphate solution (39.4 g. of c. p. copper 
 sulphate dissolved in 100 c.c. of distilled water). The 
 copper plate is also placed in the outer vessel. After being 
 set up, the cell should be short-circuited for 15 minutes 
 and then allowed to stand on an open circuit for 5 minutes. 
 The cell should not remain set up more than a few hours. 
 When it is no longer needed, pour the copper sulphate solu- 
 tion back into the stock bottle and the zinc sulphate solution 
 back into its bottle, unless the zinc has turned black, in which 
 case throw the zinc sulphate away. The e. m. f. of the 
 Daniell cell, prepared as above, is i . 105 international volts, 
 correct to o . 2 per cent. 
 
 The Clark cell, which differs from the above in the fact 
 that the copper is replaced by mercury and the copper sul- 
 phate by mercurous sulphate, is a more constant standard 
 than the Daniell cell, but it needs to be treated with much 
 greater care, since the passage of a very small current through 
 it will alter the e. m. f. Hence it can be used only for null 
 methods and kept in circuit for the briefest time possible. 
 At temperature / its e. m. f. in volts is 
 
 i.433-.ooi2(/-i5). 
 
 In the cadmium cell the zinc and zinc sulphate of the 
 above are replaced by cadmium and cadmium sulphate. 
 Its e. m. f. is 
 
 1.019 . 00004(2 17). 
 
DOUBLE COMMUTATOR. 
 
 35. Device for Getting a Small E. M. F. 
 
 In many experiments it is desirable to use an e. m. f- 
 much smaller than that of a single cell. To get such an 
 e. m. f., a box of very high resistance may be placed in 
 series with a constant cell and any desired fraction of the 
 whole e. m. f. may be obtained by tapping off from various 
 
 FIG. 47- 
 
 points; e. g., at the ends of a resistance r out of the total re- 
 sistance R of the box (Fig. 47). The e. m. f. thus obtained 
 may be found from Ohm's Law, but it must be noticed 
 that the resistance between the terminals of r is the resist- 
 ance of a parallel arc. If, however, the resistance of the 
 branch circuit be proportionally very large and that of the 
 cell proportionally very small, both may be omitted in the 
 calculation. 
 
 36. Double Commutator. 
 
 It is sometimes desirable to be able to reverse two parts 
 of a network repeatedly and at the same rate. For this pur- 
 pose a double commutator is con- 
 venient. It consists of two two-part 
 commutators mounted on a common 
 shaft; e. g., on opposite ends of the 
 shaft of a small motor. If, for ex- 
 ample, the battery used with a Wheat- 
 stone's Bridge be connected through 
 
 one commutator while the galvanom 
 eter is connected through the other, 
 ii 
 
 FIG. 48. 
 
1 62 ELECTRICITY AND MAGNETISM. 
 
 an alternating current will act in the arms of the bridge, 
 while a direct current (or a succession of unidirectional 
 pulses) will pass through the galvanometer. 
 
 37. Relation Between Electrical Units. 
 
 (E.S. = Electrostatic; E.M. = Electromagnetic.) 
 Ampere =io- 1 C.G.S.-E.M. units of current. 
 Coulomb =io- 1 C.G.S.-E.M. units of quantity. 
 Volt =io 8 C.G.S.-E.M. units of electromotive force. 
 
 Ohm = io 9 C.G.S.-E.M. units of resistance. 
 
 Farad =io- 9 C.G.S.-E.M. units of capacity. 
 
 Microfarad = io- 15 C.G.S.-E.M. units of capacity. 
 Henry =io 9 C.G.S.-E.M. units of inductance. 
 Volt =xio- 2 C.G.S.-E.S. units of electromotive force. 
 
 Coulomb =3X10" C.G.S.-E.S. units of quantity. 
 Microfarad = 9 Xio 5 C.G.S.-E.S. units of capacity. 
 
XLII. HORIZONTAL COMPONENT OF EARTH'S 
 MAGNETIC FIELD. 
 
 Ames 1 General Physics, pp. 609613; Watson's Physics, pp. 602-607; 
 Text-book of Physics (Duff), pp. 498-500; Crew' s General Physics , 
 pp. 319-323; Hartley's Electricity and Magnetism, pp. 92-98; 
 Watson's Practical Physics, pp. 403-414; Kohlrausch's Physical 
 Measurements, pp. 240-247; Stewart and Gee's Practical Physics, 
 pp. 284-309- 
 
 In this experiment the horizontal component of the 
 earth's magnetic field, at a point in the laboratory, is de- 
 duced from the period of vibration of a bar-magnet and the 
 deflection of a magnetic needle produced by this same bar- 
 magnet when placed at known distances E and W (mag- 
 netically) of the needle. The dimensions and mass of the 
 magnet must also be obtained in order that its moment of 
 inertia may be calculated. 
 
 If the period of vibration of the magnet be T in the 
 place in which we wish to determine the horizontal com- 
 ponent H, its magnetic moment be M, and its moment of 
 inertia /, then 
 
 d) 
 
 \M# 
 
 For, when the magnet is deflected through a small angle 6, the 
 restoring couple is MH sind = MH6. Hence if the angular accelera- 
 tion at that moment is a 
 
 -MH6=Ia 
 and 
 
 Since M, H, and 7 are constant, the motion is simple harmonic and 
 T is given by (i). 
 
 If a magnetic needle at a distance d, E or W of this same 
 bar-magnet, in line with it and the point where H is to be 
 
164 ELECTRICITY AND MAGNETISM. 
 
 determined, be deflected through an angle <p and, when 
 at a distance d^ be deflected through an angle (^ 
 
 M d 5 tan (bd* tan d>\ 
 
 (->) = L- l 
 
 H 2(d 2 -d, 2 ) 
 
 Equation (2) is deduced from the expression for the force F pro- 
 duced by a magnet of magnetic moment M at a distance d in the 
 direction of the axis of the magnet. For, if m is the strength of either 
 pole of the magnet and 2 1 its magnetic length, the resultant force 
 due to the two poles is 
 
 By expanding the'denominator we may also write this : 
 F = *4 
 
 in which K is approximately a constant. (If the length of the needle 
 were also taken account of, this expression would remain unchanged , 
 except that the value of the constant K would be different) . If, un- 
 der the force F and the component H of the earth's magnetic field, 
 a magnetic needle makes an angle with the magnetic meridian, 
 F = H tan 0. Hence, 
 
 M / K\ d* tan d> 
 
 If, now, the distance be changed to d lt and the deflection becomes 
 </> t another equation similar to the above will be obtained and the 
 elimination of K will give equation (2) above. 
 
 From equations (i) and (2) both H and M may be obtained 
 when the other quantities have been measured. 
 
 (A) To determine the period of vibration, remove all 
 movable iron (knives, keys, etc., included) to several 
 meters from the vicinity of the entire experiment. Suspend 
 the deflecting magnet, by means of a stirrup attached to a 
 single strand of silk thread, in a box which has glass ends 
 and sides and is surmounted by a glass tube through which 
 the suspension passes. Level until the thread hangs in the 
 axis of the tube. The magnet should be adjusted until it is 
 horizontal as tested by comparison with a leveled rod 
 attached to the outside of the box. Attach pointers to the 
 opposite glass sides of the box (or adjust those provided) 
 so that they are in line with the magnet at rest. Set the 
 magnet vibrating through an angle not exceeding 10. 
 
HORIZONTAL COMPONENT OF EARTH'S MAGNETIC FIELD. 165 
 
 Check any pendulum vibrations by judiciously pressing on 
 the top of the glass tube. Then determine the period by the 
 method of passages as in Exp. X. (see p. 54). 
 
 The magnet should be vibrated as near as is convenient 
 to the place where the needle is deflected in the second 
 part, i. e., where we wish to determine H. 
 
 (B) The instrument used in the deflection part of the 
 experiment is called a magnetometer. It consists of a box 
 with glass sides in which is suspended a mirror attached to 
 either a small magnetic needle with a damping vane or a 
 small bell magnet vibrating in a copper sphere. The sphere 
 is placed at the center of a graduated bar upon which can 
 be placed the deflecting magnet. Level until the suspending 
 fiber is at the center of the bottom of the suspension-tube. 
 If the needle or the damping vane does not swing free, a 
 little additional leveling will be necessary. 
 
 A specially mounted large compass needle is used to 
 adjust the magnetometer bar perpendicular to the magnetic 
 meridian. By means of it a rod is placed in the direction 
 of the magnetic meridian, and then, by means of a square, 
 the magnetometer bar is made perpendicular to the rod. 
 Place a telescope and scale about one meter from the mag- 
 netometer. See that the scale is perpendicular to the 
 telescope. Adjust until the scale reflected from the mirror 
 is clearly seen in the telescope (for directions for this adjust- 
 ment see p. 25). 
 
 Place the magnet whose period of vibration has been de- 
 termined on a small wood slide near one end of the magnet- 
 ometer bar. Note the scale-reading on the magnetometer 
 bar corresponding to the end of the magnet nearer the 
 needle. When the needle comes to rest, record the scale- 
 reading against the vertical cross-hair of the telescope. 
 Remove the magnet several meters and read the zero. 
 Replace the magnet at the same distance from the needle, 
 but reversed, and again read the scale division correspond- 
 ing to the vertical cross-hair. Make two similar readings 
 with the magnet at an equal distance on the other side of 
 
1 66 ELECTRICITY AND MAGNETISM. 
 
 the needle. Read the zero before or after each reading and 
 always estimate tenths of millimeters. Make four similar 
 readings with the magnet at about two-thirds the distance 
 on each side of the needle. 
 
 If the zero is somewhat unsteady, the following method 
 will be found better. Omit zero readings and obtain the 
 four deflection readings as rapidly as possible. Repeat this 
 twice so that twelve readings in all are obtained. Take 
 half the difference of each two successive readings as one 
 value of the deflection. The final result will be the mean 
 of all values so found. The extent to which they agree 
 will indicate the reliability of the mean. 
 
 Measure the distance from the center of the scale beneath 
 the telescope to the center of the suspension-tube of the 
 magnetometer (i. e., the distance to the mirror). From 
 this distance and the mean scale-reading for that distance, 
 tan 2$ is obtained (for it must be remembered that a re- 
 flected ray of light is turned through twice the angle that 
 the reflecting mirror is turned through). Since </) is a small 
 angle tan 2(f> = 2 tan (/> very nearly. The distances from the 
 needle to the near end of the magnet plus half the length 
 of the magnet give d and d^. At the close of the experiment, 
 measure the length of the magnet with vernier calipers, 
 and the diameter with micrometer calipers, and also weigh 
 it. If / be the length, r the radius, and m the mass, the 
 moment of inertia is : 
 
 In reporting, state the possible errors of the measure- 
 ments of r, /, d, d lt tan <, tan fa. 
 
 Questions. 
 
 i . How could the true length of the deflecting magnet be obtained ? 
 
 2. H having been obtained at one point in the room or building, 
 what would be the easiest way of finding its value at any other point ? 
 
 3. What are the other "elements" of the earth's magnetism? 
 
 4. If you have done Exp. XLIII, calculate the total force and 
 the vertical component. 
 
MAGNETIC INCLINATION OR DIP. 167 
 
 XLIII. MAGNETIC INCLINATION OR DIP. 
 
 Ames' General Physics, pp. 618619; Watson's Physics, pp. 605-607; 
 Text-book of Physics (Duff), p. 506; Crew's General Physics, p. 
 312; Hartley's Electricity and Magnetism, pp. 99102; Watson's 
 Practical Physics, pp. 4 1 5-4 1 7 ; Stewart and Gee's Practical Physics, 
 II, pp. 275-284. 
 
 (A) The dip, or inclination of the earth's magnetic lines 
 of force to the horizontal, is found by means of a dipping 
 needle or magnetic needle suspended on a horizontal axis 
 which passes as nearly as possible through the center of 
 gravity of the needle, with a vertical graduated circle for 
 reading the angle of inclination. Such an apparatus is 
 called a dip-circle, and includes a level and leveling screws 
 for making the circle vertical, knife-edges for bearing the 
 axis of the needle, a horizontal graduated circle for fixing 
 the azimuth of the vertical circle, and an arrestment, with 
 Y-shaped supports, for raising and lowering the needle and 
 placing it so that its axis of rotation passes as nearly as 
 possible through the center of the vertical circle. 
 
 The zero-line of the vertical circle must first be made 
 vertical. This adjustment is made by means of the level- 
 ing screws and level just as a cathetometer is leveled (see 
 p. 19). The circle must then be turned into the plane 
 of the magnetic meridian. To attain this, advantage is 
 taken of the fact that if the plane in which the needle is 
 free to rotate be at right angles to the magnetic meridian, 
 the needle must stand vertically; for in that position the 
 horizontal component of the earth's magnetic force is par- 
 allel to the axis of rotation of the needle, and hence has no 
 moment about that axis. The circle is, therefore, turned 
 approximately east and west and then adjusted until the 
 needle is vertical. This adjustment should be repeated 
 several times, and each position should be carefully read 
 with the assistance of a vernier if one is provided. A rota- 
 tion of the circle through 90 from the mean position, 
 as indicated by the horizontal circle, should then bring the 
 plane of the circle to coincidence with the plane of the 
 
1 68 ELECTRICITY AND MAGNETISM. 
 
 magnetic meridian. By raising and lowering the arrest- 
 ment, the needle is then placed on the knife-edges in the 
 proper position for indicating the dip. 
 
 A single reading of the needle in this position would give 
 a very imperfect value of the dip. Errors arise from 
 various causes: (i) the axis may not roll freely on the 
 knife-edges, owing to dust or friction. To remove any 
 dust the axis and knife-edges should be brushed with a 
 camel's-hair brush. The setting by means of the arrest- 
 ment and the readings should be made at least twice, and 
 both sets of readings recorded. (2) The axis of rotation 
 of the needle may not be exactly at the center of the divided 
 circle. This error may be eliminated by reading the posi- 
 tion of both ends of the needle, one reading being from this 
 cause as much too great as the other is too small. (3) The 
 line of zeros on the vertical scale may not be truly vertical, 
 and this would cause errors in the same direction in the 
 readings of the ends of the needle. These errors may be 
 eliminated by turning the vertical circle through 180 about 
 a vertical axis and repeating the readings, for in these read- 
 ings the quadrants on the other side of the zero line are 
 used. (4) The axis of rotation may not pass exactly through 
 the center of gravity of the needle. So far as the fault 
 lies in the fact that the axis of rotation is to one side of the 
 axis of figure of the needle, the error may be eliminated 
 by reversing the needle in its bearings and repeating the 
 readings; for in one position gravity will make the readings 
 as much too great as in the other case it makes them too 
 small. But gravity will also cause an error if the axis 
 of rotation be in the axis of figure, but not at the center of 
 the latter. The error will not be eliminated by reversing 
 the needle on its bearings, but it will be if the magnetism 
 of the needle is reversed and all of the preceding readings 
 repeated; for then the other end of the needle will be lower 
 and the error will be in the opposite direction. The re- 
 versal of the magnetism should be done under the direction 
 of the instructor, the method of double touch being used. 
 
MEASUREMENT OF RESISTANCE. 169 
 
 In recording these various positions and readings, the 
 side of the circle on which the scale is engraved may be 
 called the face of the instrument, and similarly one side of 
 the needle may be fixed upon as its face. Thus two readings 
 of each end of the needle are to be made in each of the fol- 
 lowing positions: 
 
 (1) Face of instrument E, face of needle E; 
 
 (2) Face of instrument W, face of needle W; 
 
 (3) Face of instrument W, face of needle E; 
 
 (4) Face of instrument E, face of needle W. 
 
 The magnetism of the needle having been reversed, 
 readings are to be again taken in the above positions. The 
 final result is taken as the mean of these 32 readings. 
 
 (B) Another method of determining the dip is by means 
 of an earth inductor in series with a ballistic galvanometer 
 (p. 156). The earth inductor is first placed with the plane 
 of its coils vertical and perpendicular to the magnetic 
 meridian. It is then rotated through 180 and the throw d^ 
 noted. Several readings should be made. The plane of 
 the coils is then placed horizontally and the throw d 2 on 
 rotation through 180 noted. The ratio of d, to d l is the 
 tangent of the dip. 
 
 Questions. 
 
 1. What other sources of error may there be in measurement by 
 the dip-circle? 
 
 2. Would you be justified in making a calculation of "probable 
 error" from the various readings with the dip-circle? 
 
 3. If you have performed Exp. XLII, calculate the total force 
 and the vertical component. 
 
 XLIV. MEASUREMENT OF RESISTANCE BY WHEAT- 
 STONE'S BRIDGE. 
 
 Ames' General Physics, pp. 725-727; Watson's Physics, pp. 685-687; 
 Text-book of Physics (Duff), p. 572; Hadley's Electricity and 
 Magnetism, pp. 306-310; Watson's Practical Physics, pp. 432- 
 437; Kohlrausch's Physical Measurements, p. 303. 
 
 The practical measurement of a resistance consists in 
 comparing it with a known or standard resistance. For 
 
170 ELECTRICITY AND MAGNETISM. 
 
 resistances of medium magnitude, Wheatstone's Bridge is 
 usually used (p. 153). 
 
 In joining the known and unknown resistances to the 
 bridge, connectors should be used whose resistance is negli- 
 gible; that is, less than the unavoidable error that may 
 occur in determining the unknown resistance. In con- 
 necting the battery and galvanometer, no such precaution is 
 necessary, for their resistances do not enter into the calcu- 
 lation. The galvanometer may be connected to either pair 
 of opposite corners; but, where the greatest sensitiveness 
 is required, if the galvanometer has a higher resistance 
 than the battery, it should be in the branch that connects 
 the junction of the highest two of the four resistances P, 
 Q, R, S to the junction of the lowest two; while, if the 
 battery has the greatest resistance, it should occupy that 
 position. Two spring keys should be included in the con- 
 nections, one in the battery arm and the other in the gal- 
 vanometer arm. When testing for a balance, the battery 
 key should be pressed first, then the galvanometer key. 
 If taken in the reverse order, there might be a small deflec- 
 tion due to the self-induction of the various parts. These 
 keys should be pressed for a moment only. Except for a 
 final determination, it is not necessary to wait until the gal- 
 vanometer has quite come to rest, for a lack of balance will 
 be indicated by a sudden disturbance of the swing when 
 the galvanometer key is pressed. The pressure of the gal- 
 vanometer key should be brief, sufficient merely to indicate 
 the direction of the initial movement. 
 
 In practice, it is best to use a box-resistance as nearly 
 as possible equal to the unknown resistance. This comes 
 to the same thing as saying that the box-resistance should 
 be varied until a balance is attained when the parts of the 
 meter wire are nearly equal. The reason for this preference 
 is that the sensitiveness is then a maximum, or a slight lack 
 of balance is most easily detected by the deflection of the 
 galvanometer. The exact ratio of P to Q for a balance 
 should be very carefully ascertained. At least six settings 
 
GALVANOMETER RESISTANCE BY SHUNT METHOD. 171 
 
 should be made ; and to secure independence of the settings, 
 the eye should be kept on the galvanometer-scale and the 
 reading of the bridge not examined until the setting has 
 been decided on. The mean of these six is then taken. R 
 and 5 should then be interchanged and six more settings 
 made. This interchange will serve to eliminate the effect 
 of lack of symmetry of the two sides of the wire bridge 
 and its connections. 
 
 The structure of the galvanometer to be used, its coils, 
 magnets, and connections, should be carefully examined and 
 care taken that it is thoroughly understood (p. 155). 
 
 Three unknown resistances should be measured sep- 
 arately and then all in parallel. From the separate resist- 
 ances the resistance of the conductors in parallel should be 
 calculated and compared with the measurement of the 
 same. The resistance of a wire should then be measured 
 and its length and mean diameter obtained. From these 
 data, the specific resistance of the material of the wire 
 should be deduced. The temperature at which the resist- 
 ance is measured should also be noted, and from the tem- 
 perature coefficient of the material (Table XXII) the specific 
 resistance at o C. calculated. 
 
 The possible errors of the measurements, and hence the 
 extent to which the calculations should be carried, may be 
 deduced from the mean deviation in each set of readings. 
 
 Questions. 
 
 1 . Does the battery need to be a constant one ? 
 
 2. What objections are there to allowing the battery circuit to 
 remain closed? 
 
 3. Why is it difficult by this method to measure very large or very 
 small resistances ? 
 
 XLV. GALVANOMETER RESISTANCE BY SHUNT 
 METHOD. 
 
 Kohlrausch's Physical Measurements, p. 325. 
 If a galvanometer of resistance G connected in series 
 with a battery of resistance B and e. m. f. E and a box 
 resistance R gives a deflection d and if C be the current 
 
172 
 
 ELECTRICITY AND MAGNETISM. 
 
 C = 
 
 R+B+G 
 
 = Kd 
 
 where K is a constant for the galvanometer. If now the 
 galvanometer be shunted by a resistance S and the deflection 
 be then d f and the current through the galvanometer C', 
 
 E S 
 
 /-*/ 
 
 R+B + 
 
 GS X G+S 
 
 = Kd'. 
 
 G + S 
 
 Hence 
 
 (R+B) (G+S)+GS_d 
 S (R+B+G) = ~d r> 
 
 and from this G is readily deduced provided B is known. 
 Usually a battery of such low resistance can be used that B 
 is negligible compared with R and may be omitted; otherwise 
 A ^ B must be obtained as in Exp. LI II. 
 
 The galvanometer should be connected 
 through a commutator and several read- 
 ings on both sides should be made. 
 
 If the e. m. f. of the cell supplied is too 
 great, a suitable fraction of it should be 
 employed (p. 161). 
 
 As a check, the determination of G 
 should be repeated, a different value for 
 5 being used. If the galvanometer is 
 very sensitive, its resistance must be found 
 from two readings with shunts. A suita- 
 ble formula is readily worked out. 
 
 If R should be very great compared with the other 
 resistances, the formula may be simplified. This will 
 usually be the case if the galvanometer is very sensitive or of 
 low resistance. The quantities added to R in the first two 
 equations may then be neglected and we get 
 
 G+S d 
 
GALVONOMETER RESISTANCE BY THOMSON'S METHOD. 173 
 
 XLVI. GALVANOMETER RESISTANCE BY THOMSON'S 
 
 METHOD. 
 
 Hadley's Electricity and Magnetism, p. 321; Kohlrausch's Physica 
 Measurements, p. 328; Stewart and Gee's Practical Physics, II, 
 p. 140-142. 
 
 The resistance of the coils of a galvanometer may be found 
 by means of Wheatstone's Bridge as the resistance of any 
 ordinary conductor is found. This would require the use 
 of a second galvanometer. The second galvanometer, for 
 detecting when the bridge is balanced, is frequently un- 
 necessary. The condition for 
 a balance is that, when the 
 branch in which the galvanom- 
 eter is usually placed is closed 
 by a key, no current shall flow 
 through it. If a current did 
 flow through it, a change 
 would take place in the cur- 
 rents in the other arms. Now 
 the presence of a galvanom- 
 eter in one of these other arms 
 enables us to test whether any 
 change in the distribution of 
 the currents takes place on 
 
 the key's being pressed. Hence, in Thomson's method for 
 galvanometer resistance the galvanometer is placed in the 
 "unknown" arm and a spring key, K, is placed in the 
 branch in which, in the ordinary arrangement of Wheat- 
 stone's Bridge, a galvanometer is found. A diagram to 
 illustrate the connections is given in figure 50. 
 
 From the above it will be seen that in this method a bal- 
 ance is obtained when the deflection of the galvanometer 
 does not change on the key, K, being pressed. Two practical 
 difficulties are met with. The first is that the deflection 
 of the galvanometer before the key is pressed may be so 
 large that it cannot be read. When the galvanometer is of 
 
 FIG. 50. 
 
174 ELECTRICITY AND MAGNETISM. 
 
 the Thomson type (p. 155), this difficulty may be overcome 
 by turning the control magnet until the deflection can be 
 read (the zero position of the galvanometer could, of course, 
 not then be read on the scale, but that is not necessary). 
 In the d'Arsonval type of galvanometer there is no such 
 way of overcoming this difficulty, and so this method is not 
 so easily applied to such a galvanometer. The second 
 difficulty is that if the battery be a variable one, the gal- 
 vanometer will not give a steady deflection. Hence, a 
 constant battery of the Daniell or Gravity type should be 
 used (p. 159). It may also be necessary to decrease the 
 current through the bridge and galvanometer by putting 
 considerable resistance in series with the battery, or a fraction 
 of the e. m. f. of the cell may be used (p. 161). 
 
 In the experiment it is better to use a bridge-box instead 
 of a wire bridge, for the condition for sensitiveness, that 
 the arms should be as nearly equal as possible, still holds, 
 and the resistance of a wire bridge is usually very small 
 compared with that of the galvanometer. Beginners some- 
 times find difficulty in deciding on the proper connections. 
 The best way is to consider what the connections would be 
 in the ordinary use of Wheatstone's Bridge, and then con- 
 sider the modifications introduced in the present method. 
 If possible, ratio arms of 1000 to 1000, 100 to 1000, and 10 
 to 1000 should be used in succession to obtain successive 
 approximations. The last should give the resistance to two 
 places of decimals (if one ohm is the least box-resistance), 
 but the decreasing sensitiveness may prevent the latter 
 ratios from giving more accurate results than the first. 
 
 If the galvanometer has more than one coil, the resist- 
 ance of each should be measured separately and then the 
 resistance of all in series. This will afford a check on the 
 work. 
 
 Question. 
 
 i. Describe carefully the swinging system, coils, control magnet, 
 and connections of the galvanometer used. 
 
MEASUREMENT OF HIGH RESISTANCES. 175 
 
 XLVII. MEASUREMENT OF HIGH RESISTANCES (i). 
 
 Watson's Practical Physics, pp. 460-461; Henderson's Electricity 
 and Magnetism, pp. 66-72. 
 
 The method of Wheatstone's Bridge is not suitable for 
 measuring very high resistances. One method is to con- 
 nect the unknown resistance X, a battery of negligible re- 
 sistance and e. m. f. E, and a sensitive galvanometer of 
 resistance G in series. If the current be C, 
 
 E 
 
 C = - , giving a deflection d. 
 X -\-G 
 
 Now replace X by a known resistance, R, and shunt 
 the galvanometer by such a resistance, 5, that the deflection 
 is readable. By considering the total current and the part 
 C' of the total current that passes through the galvanometer, 
 we readily find that 
 
 W S" 
 
 C' --- , giving a deflection d' , 
 GS G+S 
 
 + 
 
 Hence, 
 
 R(G+S)+GS _d 
 S(X+G) = J" 
 
 and from this X is readily deduced. G may be found as in 
 Exp. XLV or XLVI; but if (G+S)/S is known and G is 
 small compared with X and R, the resistance of the galvan- 
 ometer need not be determined. Many galvanometers are 
 provided with shunt boxes, for which S/(G+S) is o.i, 
 o.oi, or o.ooi. 
 
 The galvanometer should be connected through a com- 
 mutator, and several readings on both sides should be made 
 to obtain a reliable mean. 
 
 As a check, repeat the measurements with a different 
 value for R and a different value for S. 
 
7 6 
 
 ELECTRICITY AND MAGNETISM. 
 
 XLVIII. MEASUREMENT OF HIGH RESISTANCES (2). 
 
 Text-book of Physics (Duff), pp. 537-538; Watson's Physics, pp. 
 656-657; Ames' General Physics, pp. 658-659; Henderson's 
 Electricity and Magnetism, pp. 71-75; Hadley's Electricity and 
 Magnetism, pp. 202-210; Watson's Practical Physics, pp. 569- 
 57 1 - 
 
 A very high resistance, such as the insulation resistance of 
 a cable or the resistance of cloth, paper, wood, etc., may 
 be measured by finding the rate at which the electricity in a 
 charged condenser leaks through the conductor. An 
 electrometer is used to find the change of potential of the 
 condenser and from this the rate of loss of its charge is 
 deduced. The Dolezalek form of Kelvin's quadrant electrom- 
 
 FIG. 51. 
 
 eter is suitable. Its needle is kept charged to a high poten- 
 tial by being connected to one pole of a battery of small 
 cells, the other pole being grounded. 
 
 To find the insulation resistance of a cable the whole of 
 the cable except the ends is immersed in a tank of salt water 
 which is connected to the earth. One of the ends is carefully 
 paraffined to prevent surface leakage and the core of the 
 other end is connected to the insulated pair of quadrants. 
 If the cable is sheathed with metal, immersion is not neces- 
 sary. Other materials, such as those mentioned, are 
 pressed between sheets of tinfoil, one sheet being connected 
 to the earthed quadrants and the other to the insulated 
 quadrants. 
 
MEASUREMENT OF HIGH RESISTANCES. 177 
 
 L e t V\ = potential given the condenser on closing the 
 key K. The charge Q in the condenser and cable = C V^ 
 where C is their joint capacity. Upon opening K the charge 
 flows through a resistance R for a time, t, ^ t being the insul- 
 ation resistance of the cable, the condenser, and the elec- 
 trometer and keys in parallel. 
 
 Since the current at the time / equals V/R^ by Ohm's law, 
 and also equals the rate of decrease of Q or of CV 
 
 = -c dv - 
 
 R, dt 
 
 dV _dt 
 
 'T".s; 
 
 Integrating between the limits t = o when V=V 1 and / = t 
 when V=V 2 we get 
 
 R t ^-434* 
 
 C lg e ~ C 1~ * 
 
 where d l and d 2 are the initial and final deflections of the 
 electrometer from the zero position. 
 
 The zero should be determined both before V l is found 
 and after V 2 is found. As it is very apt to vary slightly, 
 more reliable results can be attained by continuing to read 
 V at intervals (e. g., every half -minute) until it has fallen 
 to about one-half of its original value. From a curve drawn 
 to represent V and /, two reliable points may be chosen 
 to give values for V l and V 2 to be used in the calculation. 
 
 A subdivided condenser is desirable in order that a 
 capacity giving a sufficiently rapid fall of potential may be 
 chosen. 
 
 The total insulation resistance, R 2) of the other parts in 
 parallel with the cable are found by disconnecting the 
 cable and making a second set of observations as above. 
 
 12 
 
178 ELECTRICITY AND MAGNETISM. 
 
 The insulation resistance, R, of the cable is then deducible 
 for 
 
 The capacity, C lt of the cable can be compared with 
 that of the condenser, C 2 , by the method of "divided 
 charge." First charge the condenser and observe its 
 potential by the electrometer and let the deflection be d r 
 Then connect in the cable and let d 2 be the new deflection. 
 Since the total charge Q remains unchanged, 
 
 and, since the deflections are proportional to the potentials, 
 
 C=C l 2 
 
 Questions. 
 
 1. Why should one pole of the battery that charges the needle be 
 grounded ? 
 
 2. Why must keys of specially high insulation be used in this 
 method ? 
 
 3. Calculate the capacity of the cable in electrostatic units from 
 rough measurements of its dimensions and reduce to microfarads 
 (see p. 162). 
 
 XLIX. MEASUREMENT OF LOW RESISTANCES (i). 
 
 Very low resistances cannot be measured by the Wheatstone 
 Bridge method, because the unknown resistances of the 
 connections are not small compared with the resistance 
 to be measured. The simplest method for low resistances 
 is a "fall of potential " method. A current is passed through 
 the resistance, the current is measured by an ammeter and 
 the difference of potential is measured by a voltmeter; then 
 the resistance is known from Ohm's Law. For very low 
 resistances the fall of potential will be very small and an 
 instrument much more sensitive than any commercial volt- 
 meter must be used. Instead of a voltmeter a sensitive gal- 
 vanometer of high resistance, or a low resistance galvan- 
 ometer in series with a high resistance, is used and the 
 
MEASUREMENT OF LOW RESISTANCES. 179 
 
 value of a scale division of the galvanometer regarded as a 
 voltmeter is found by a separate experiment. 
 
 Let the resistance to be measured be x, and let the differ- 
 ence of potential at its ends when current C passes through 
 it be e. Then 
 
 e 
 
 x= c 
 
 C, which should be large, may be measured by an ammeter. 
 To find e we must know the constant, K, of the galvanometer 
 considered as a voltmeter; that is, the number of volts per 
 unit deflection. If the deflection is D 
 
 e=K.D 
 
 To find K apply to the galvanometer a small fraction of 
 the e. m. f., E, of a Daniell'scell (p. 159). , E 
 
 For this purpose connect the cell in 
 series with a very high resistance box 
 and a box of moderate resistances and 
 join the galvanometer to the ends of 
 one of the small resistances, r, choosing 
 r so that the deflection, d, will not be 
 very different from D. Then if the re- 
 sistance of the galvanometer be great 
 compared with r (see p. 161) and 
 if the total resistance in series with FIG. 52. 
 
 the battery be R, the e. m. f. acting on the galvanometer 
 is Er/R. 
 Hence 
 
 As a check on the work redetermine K using a different 
 value of r. From the above equations x is found. 
 
 In the first part of the experiment place a commutator 
 in the main circuit so that C may be reversed and the effect 
 of thermo-electric forces at the contacts eliminated, and 
 connect the galvanometer through a second commutator so 
 that lack of symmetry in its deflection may be eliminated. 
 
i8o 
 
 ELECTRICITY AND MAGNETISM. 
 
 Thus D will be the mean of four readings. Exactly simi- 
 lar precautions should be observed in the second part. 
 Close the currents only for the shortest possible times 
 necessary to make the readings, otherwise heating may 
 occur and resistances (especially the unknown x) may change. 
 The determination should be repeated several times with 
 different values of C. If the work has been reliable, D should 
 be proportional to C. Note also the temperature of the 
 specimen and calculate its resistivity from its resistance and 
 dimensions. 
 
 Questions. 
 
 1. If r had not been negligible compared with the galvanometer 
 resistance how would this have appeared in the course of the work? 
 
 2. Find the equation that must replace the above if the resist- 
 ance of the battery is not negligible compared with R and if r is not 
 negligible compared with the galvanometer resistance. 
 
 L. MEASUREMENT OF LOW RESISTANCES (2). 
 
 Henderson's Electricity and Magnetism, pp. 5758; Stewart and Gee's 
 Practical Physics, II, pp. 177-181. 
 
 When a standard low resistance (o.oi or o.ooi ohm) is 
 available, a conductor of low resistance x may be connected 
 
 in series with it and a battery, 
 and a very sensitive voltmeter, 
 or a high-resistance galvanom- 
 eter, serving as a voltmeter, 
 may be used to compare the 
 falls of potential in x and the 
 standard. The resistances will 
 be proportional to the falls of 
 potential. 
 
 Connection with the battery 
 should be made through a com- 
 mutator to reverse thermal 
 effects at the connections, and 
 the galvanometer should be 
 connected through a second 
 commutator to eliminate asym- 
 
MEASUREMENT OF LOW RESISTANCES. 
 
 181 
 
 metry of the galvanometer readings. Thus each final read- 
 ing will be the mean of four separate readings. 
 
 The currents should be closed for the shortest times 
 sufficient for the readings, to avoid heating. Note the 
 temperature of the specimen. 
 
 Questions. 
 
 1. What are the comparative advantages and disadvantages of this 
 and the preceding method ? 
 
 2. Why is a high-resistance galvanometer to be preferred? 
 
 3. Will poor contact have as much effect as in a measurement of 
 low resistance by Wheatstone's Bridge? Why? 
 
 LI. MEASUREMENT OF LOW RESISTANCES BY THE 
 THOMSON DOUBLE BRIDGE. 
 
 Watson's Practical Physics, pp. 465-469; Stewart and Gee, II, 
 pp. 182-187. 
 
 In Thomson's Double Bridge the errors of the contacts 
 in the use of Wheatstone's Bridge are avoided. Its princi- 
 ple is, in fact, that of the fall of potential method (Exp. 
 L) the direct comparison of the falls 
 of potential being replaced by a null 
 method. This method is applicable to 
 extremely low resistances as well as to 
 medium resistances. 
 
 In the diagram x is the resistance to 
 be measured and r a standard known 
 resistance; a may be made 10 or 100; " 
 and b may be made 100, 1,000, 10,000. 
 Similarly, a' may be made 10 or 100 
 
 and b f 100, 1,000, 10,000. Now it can 
 
 FIG. 54. 
 
 be shown that if there is no current in the galvanometer, and 
 if the ratios a/ b and a'/b' are equal, 
 
 x a a' 
 
 ~r = b = l)'' 
 
 hence a value of r (which is variable) and values of a, b, 
 a' b', are sought which give no current in the galvanometer, 
 and from these x is calculated. 
 
l82 
 
 ELECTRICITY AND MAGNETISM. 
 
 The form of Double Bridge made by Hartmann and Braun 
 is very satisfactory. The correspondence of parts to parts 
 of the diagram is readily traced. The ratio of a to b can be 
 varied from 100 to 100 to 10 to 10,000; moreover, a and b 
 may be interchanged and so the ratio reversed. Similar re- 
 marks apply to a' and b'. Thus values of x/r varying 
 from 10/10000 to 10000/10 may be measured. The 
 variable r may be varied from o . 044 down to o, but can 
 hardly be read with an accuracy of i% below o.ooi. 
 
 Hence values of x between 
 o.ooi/ 1000 or o.oooooi and 
 0.044X1000 or 44, may be 
 measured by the bridge. 
 
 Care must be taken not to 
 injure the standardized bar by 
 scraping the contact maker along 
 it. The contact maker must be 
 raised for each movement. Do 
 not allow the sharp jaws of the 
 clamps to come down on the bar 
 too suddenly, for they might cut 
 into the bar somewhat. 
 
 Test as many as possible of the following materials: 
 (i) Brass. (2) Iron. (3) Copper. (4) Zinc. (5) Lead. 
 (6) Carbon. (7) Rail Bond, 
 and calculate the Specific Resistance of each. 
 
 Proof of Formula. 
 
 Regard Thomson's Double Bridge as a modified Wheatstone's 
 Bridge, the modification consisting in the paralleling of parts of the 
 arms AB and BC as indicated. Let G and D be points at the same 
 potential as indicated by the galvanometer. E B F is the heavy 
 conductor joining the unknown and the standard. Let B be a 
 point in it at the same potential as G. We may suppose B and G 
 permanently connected. Let the resistance of p and a' in parallel 
 be m, and that of q and b' be n. Then, by the Wheatstone Bridge 
 formula : 
 
 FIG. 55. 
 
 a _x + m 
 b~ r + n' 
 
COMPARISON OF RESISTANCES. 183 
 
 Now if we show that 
 
 it will follow that 
 
 a = x_ 
 b 7' 
 
 To show this, note that 
 
 a _ a' _ p 
 
 a' + p b'+q' 
 
 Now 
 
 = a'p ^b^q 
 
 m p a 
 
 Hence 
 
 x a 
 
 Questions. 
 
 1. Considering this as a modified fall of potential method, why 
 should a, b, a', b', be of very large and E B F of very small resistance? 
 
 2. Does the battery current need to be steady? Why? 
 
 3. Could an alternating current be used in any circumstances? 
 
 LII. COMPARISON OF RESISTANCES BY THE CAREY- 
 FOSTER METHOD. 
 
 Watson's Practical Physics, pp. 442-446; Schuster and Lees' Practical 
 Physics, pp. 307-309; Henderson's Electricity and Magnetism, 
 pp. 5357; Stewart and Gee's Practical Physics, II, pp. 158170. 
 
 To find very accurately the difference between two very 
 nearly equal resistances R and 5, connect them and two 
 other nearly equal resistances, P and Q, as indicated in the 
 diagram, where ab is a very uniform wire, which we shall 
 suppose to have a resistance of more than i ohm. Let the 
 resistance of unit length of the wire ab be p. Let the dis- 
 tance ad be x lt when a balance has been obtained in the 
 usual way. Then exchange R and 5, and again obtain a 
 balance. Denote the new value of ad by x 2 . Since P and Q 
 
1 84 
 
 ELECTRICITY AND MAGNETISM. 
 
 have not been changed and the total resistance R, S, and ab 
 was not changed, it is clear that R -i-x 1 p = S +x^p, or 
 R S= (x 2 xjp. To find the value of p, replace R by a 
 standard i-ohm coil, and 5 by a heavy connector of neg- 
 
 FlG. 
 
 ligible resistance, and proceed as above; then p(x 2 x l ) = i. 
 The exchange of R and 5 is made by means of a special 
 key designed so that the resistance of the connections will 
 remain the same (see Fig. 57). 
 
 FIG. 57 . 
 
 To compare a box of unknown errors with a standardized 
 box, the difference between each resistance of the former 
 and a corresponding resistance of the latter is found by the 
 above method. 
 
BATTERY RESISTANCE BY MANGE'S METHOD. 
 
 185 
 
 To calibrate two boxes, put one in place of R and replace 
 5 by a standard i-ohm coil and so find exactly the value 
 of each i-ohm unit in the box. Then replace the standard 
 by the other box, in position S, and compare the i-ohm 
 units of the second box with those of the first box. Then 
 compare a 2-ohm unit in one box with two i-ohms in 
 the other, and so on. Special care must be taken to avoid 
 confusion in making the calculations, and for this purpose 
 the box resistances may be denoted by I lf I 2 , II lt II 2 , etc., 
 for one box, and I\, I' 2 , II' 2 , etc., for the other. 
 
 Questions. 
 
 1. State the formula for Wheatstone's Bridge before and after R 
 and 5" are interchanged and therefrom deduce the above formula. 
 
 2. Do P and Q need to be exactly equal, and why? 
 
 LIII. BATTERY RESISTANCE BY MANGE'S METHOD. 
 
 Hartley's Electricity and Magnetism, p. 322 ; Watson's Practical Physics, 
 p. 475; Schuster and Lee's Practical Physics, pp. 303-306. 
 
 The resistance of a battery may be determined by plac- 
 ing it in the "unknown arm" .R of a Wheatstone's Bridge 
 (p- I 53)- I n "this case there 
 will be a current through 
 the galvanometer when the 
 bridge battery is not con- 
 nected. But if P, Q and 5 
 be adjusted until there is no 
 change in the deflection when 
 the key of the bridge battery 
 is pressed, the points to 
 which the galvanometer is 
 connected will be at the same 
 potential so far as the effect 
 of the bridge battery is con- 
 cerned. Since, when the 
 
 FIG. 58. 
 
 adjustments are right, the bridge battery sends no current 
 through the galvanometer, this battery may be removed and 
 the key alone will serve to test the adjustment of P, Q, and S. 
 
1 86 ELECTRICITY AND MAGNETISM. 
 
 If the deflection of the galvanometer is too great to be 
 readable, the control magnet (in the case of a Kelvin gal- 
 vanometer) may be used to bring the needle back, or the 
 galvanometer may be shunted or a resistance put in series 
 with it. Most cells vary slightly in resistance and e. m. f. 
 when on closed circuit; hence, the keys should not be 
 pressed longer than is necessary. 
 
 In Lodge's modification of Mance's Method a condenser 
 is placed in series with the galvanometer. There will then 
 be no continuous current through the galvanometer; but, if 
 the adjustments of P, Q, and 5 are not right, on pressing 
 the key by which the adjustment is tested the galvanometer 
 will be momentarily deflected. 
 
 Questions. 
 
 1. Why is there a slow movement of the galvanometer needle 
 when the keys are kept pressed ? 
 
 2. Should the condenser be of large or small capacity? Would a 
 Ley den jar do? 
 
 LIV. TEMPERATURE COEFFICIENT OF RESISTANCE. 
 
 Text-book of Physics (Duff), p. 564-566; Ames' General Physics, pp. 
 731-732; Watson's Physics, pp. 681-682; Hadley's Electricity 
 and Magnetism, p. 294; Henderson's Electricity and Magnetism, 
 pp. 95-101. 
 
 The resistance of most solids increases as the tempera- 
 ture rises; carbon is one of the exceptions, for its resistance 
 decreases. For moderate ranges of temperature the resist- 
 ance is approximately a linear function of the temperature or, 
 if ^ be the resistance at o and R that at t, 
 
 R=R Q (i+at) 
 
 The constant a is called the temperature coefficient of the ma- 
 terial. It may be defined as the change per ohm, referred 
 to the resistance at o, per degree change of temperature. 
 
 The change of resistance can be most conveniently studied 
 by the box form of Wheatstone's Bridge (p. 154). 
 
 (A) For finding the temperature coefficient of a wire such 
 as copper, a length sufficient to give several ohms resistance 
 
TEMPERATURE COEFFICIENT OF RESISTANCE. 187 
 
 should be used. The determination of the temperature 
 coefficient does not require that the dimensions of the 
 specimen should be known, but the specific resistance of 
 the specimen might as well be determined at the same time. 
 Hence the length and mean diameter of the wire should be 
 carefully measured. The wire should then be soldered to 
 heavier lead wires and immersed in a bath of oil, and its 
 resistance determined at intervals of about 10 as the tem- 
 perature is raised. The thermometer should be placed inside 
 the coil so as to be as nearly as possible at the temperature 
 of the latter. It will be an improvement if the coil and 
 thermometer are in a test-tube that is immersed in the 
 bath, the opening of the tube being closed with cotton-wool. 
 
 To keep the temperature constant, while measuring the 
 resistance, would be difficult. The following method will 
 be found to give much better results: Having measured 
 the resistance at the temperature of the room, adjust the 
 known resistance of the bridge so that there would be a 
 balance if the resistance of the wire were increased 4 or 5 
 per cent. The galvanometer will be deflected. Now heat 
 the wire very slowly and the galvanometer reading will 
 begin to drift toward zero. When it just reaches zero, read 
 the thermometer and continue the process step by step. 
 
 The various resistances and temperatures should then 
 be plotted in a curve that should be approximately a straight 
 line. If exactly a straight line is obtained, the temperature 
 coefficient should be calculated from two reliable and widely 
 separated points on the curve. Let R and R' be the resist- 
 ances at t and t', respectively. Substituting these values in 
 the above equation we shall get two equations from which 
 R can be eliminated. 
 
 If the plotted readings give a distinct curve, the resistance 
 must be expressed as a quadratic function of the temperature. 
 
 R=RQ(I + at + bt 2 ) 
 
 From three points on the curve three equations may be 
 written down and from these a and b may be calculated. 
 
1 88 ELECTRICITY AND MAGNETISM. 
 
 (B) For finding the temperature coefficient of carbon 
 an incandescent lamp may be used. As it would be diffi- 
 cult to determine accurately the temperature of the filament 
 in the exhausted bulb by the preceding method, water may 
 be used for the bath and two careful determinations of the 
 resistance made, the first being while the water is at about 
 the temperature of the room, and the other when the water 
 is boiling. In each case the final determination of the 
 resistance should not be made until the temperature of the 
 filament has become constant, as is indicated by its resist- 
 ance becoming quite constant. The leads, where they are 
 immersed in the water, should be carefully insulated with 
 tape. 
 
 LV. SPECIFIC RESISTANCE OF AN ELECTROLYTE. 
 
 Watson's Practical Physics, pp. 475-486; Henderson's Electricity and 
 Magnetism, pp. 8084; EwelVs Physical Chemistry, pp. 54-57; 
 Kohlrausch's Physical Measurements, pp. 316-321. 
 
 The object of this experiment is to determine the specific 
 resistance of an electrolyte for instance, solutions of copper 
 sulphate of different concentrations. The box form of 
 Wheatstone's Bridge is most suitable for the purpose (p. 1 54). 
 
 A steady current from a battery and a galvanometer to 
 determine when there is a balance, as ordinarily used with 
 Wheatstone's Bridge, cannot satisfactorily be used in meas- 
 uring the resistance of an electrolyte, for a steady current 
 produces in a short time polarization at the electrodes. 
 This polarization leads to too high an estimate of the resist- 
 ance of the electrolyte, for when no current flows through 
 the galvanometer, the three other arms of the bridge are 
 balancing the potential difference necessary to overcome 
 the true resistance of the electrolyte plus the potential 
 difference required for overcoming the polarization potential 
 difference at the electrodes. This difficulty is obviated by 
 using the rapidly alternating current from the secondary of an 
 induction coil instead of a steady current from a battery. 
 
SPECIFIC RESISTANCE OF AN ELECTROLYTE. 
 
 189 
 
 The time that the current continues in one direction is so 
 short that no appreciable accumulation can form at the 
 electrodes to produce an opposing difference of potential. 
 An ordinary galvanometer would not be affected by an 
 alternating current, but a telephone which is a very delicate 
 detector of an alternating current may be substituted. 
 
 In the simplest form of apparatus a vertical glass tube of 
 known cross section, which may 
 be found by calipers (p. 14), 
 holds the electrolyte. The elec- 
 trodes are connected to wires 
 that pass through the stoppers; 
 the upper electrode can be 
 raised or lowered as desired. 
 The resistances corresponding 
 to two different distances of 
 separation of the electrodes 
 should be determined. From 
 the difference we get the re- 
 sistance of a column whose 
 length is the difference in the 
 two lengths and thus eliminate uncertainty as to remaining 
 polarization at the electrodes and the exact ends of each 
 column. 
 
 Two tubes should be available for the work. In one 
 measurement should be made of the resistance of samples 
 of several solutions of different concentrations, the samples 
 being obtained from stock bottles. The other tube is for the 
 purpose of determining the temperature coefficient of a 
 solution. It is placed in a steam heater such as is used in 
 Exp. XIX (p. 88). As the heating to a steady temperature 
 will require considerable time, the tube should be prepared 
 at the beginning of the work. The resistance of the electro- 
 lyte in it having been determined at room temperature, 
 the heating may be allowed to proceed while the measure- 
 ments with the other tube are made. 
 
 A tube of small cross section may also be used. If its 
 
1 90 ELECTRICITY AND MAGNETISM. 
 
 ends pass through stoppers into much larger tubes that 
 contain the electrolyte and electrodes, the resistance 
 measured will be virtually that of the electrolyte in the 
 small tube. The diameter of the tube may be found as in 
 Exp. XI (p. 58). Substituting a different length of the 
 same tubing and taking differences we may as before 
 eliminate residual polarization effects. 
 
 In measuring resistances it will probably be desirable to 
 use equal resistances, e. g., 100 ohms, in the ratio arms of the 
 bridge. It may be impossible to obtain a balance for 
 which there is no sound, for even though there were a 
 balance for steady current, there would not in general be a 
 balance for varying currents such as are used in this experi- 
 ment, owing to the inductive electromotive forces of capacity 
 and self-induction in the resistance coils. When there is 
 uncertainty as to whether a small resistance should be added 
 or cut out, the ear is often assisted by adding and cutting 
 out a larger resistance about which there is no doubt. On 
 comparing the change of tone on a variation of this latter 
 resistance with the variation of tone with the uncertain 
 resistance, one can often decide whether the small resistance 
 should be added or not. With a little practice one should 
 determine resistances within i per cent. 
 
 Calculate the specific resistance of each solution at each 
 temperature and tabulate the results. Find also the 
 temperature coefficient of the solution which was heated 
 and calculate its specific resistance at o. 
 
 Questions. 
 
 1 . Why should we expect the resistance to decrease with increased 
 temperature ? 
 
 2. What is supposed to be the nature of electric conduction in an 
 electrolyte? 
 
 3. Are the specific resistances inversely proportional to the con- 
 centrations ? Why ? 
 
COMPARISON OF E. M. F/S BY HIGH-RESISTANCE METHOD. 191 
 
 LVI. COMPARISON OF E. M. F.'S BY HIGH-RESISTANCE 
 
 METHOD. 
 
 Watson's Practical Physics, pp. 430-431; Stewart and Gee's Practical 
 Physics, II, pp. 101-102. 
 
 The readiest method of comparing the electromotive forces 
 of cells is by means of a galvanometer of sufficiently high 
 resistance. If the deflections are (by the use of added 
 resistances) kept small the deflections of the galvanometer 
 will be closely proportional to the currents that pass through 
 it or i = k.d where k is a constant. Two methods may be 
 employed for comparing two cells. In the first, called the 
 "equal resistance" method, the total resistance R is kept 
 constant (the resistance of the cells being supposed negligi- 
 ble). Hence, by Ohm's Law, the e. m. f.'s are proportional 
 to the currents, that is, to the deflections, or 
 
 In the other or "equal deflection " method, such resistances 
 are used in the circuit that the cells cause equal deflections of 
 the galvanometer. Hence by Ohm's Law, since the currents 
 are equal, the electromotive forces must be proportional to 
 the resistances, or 
 
 EjJZi 
 E 2 R, 
 
 Both methods should be employed to find the e. m. f.'s 
 of several cells by comparing them with that of a standard 
 Daniell cell (p. 159). Directions for the adjustment of the 
 telescope and scale are given on p. 25. 
 
 (A) Equal Resistance Method. Make R such that the 
 standard Daniell cell gives a deflection of about 10 cm. on 
 a scale about i m. from the mirror. Make a reading of the 
 zero; i. e., when no current passes through the galvanometer. 
 Send the current through the galvanometer and read the 
 division now on the cross-hair. In this way make at least 
 six readings on one side and, reversing, make six on the 
 
I Q2 ELECTRICITY AND MAGNETISM. 
 
 other. Read the zero often, as it is liable to change. In 
 reading, use, if necessary, the method of vibration (see p. 
 23). If the vibrations are irregular on account of trolley 
 currents or other disturbances, estimate the position of equi- 
 librium from the vibrations without actually making read- 
 ings. With some galvanometers the damping is so great 
 that the system comes to rest instead of vibrating about the 
 position of equilibrium. In this case the true reading can 
 be made at once. Always, if possible, estimate tenths of 
 the smallest divisions. When you have thus found the 
 mean deflection for the standard, find similarly the deflec- 
 tion for as many different types of cells as time allows. 
 With the other cells, three readings on a side will be suffi- 
 cient. The internal resistance of the different batteries 
 varies, but the differences are negligible compared with the 
 total resistance of the circuit. Express in volts your final 
 values of the e. m. f.'s of the cells tested. 
 
 (B) Equal Deflection Method. With a resistance which 
 gives a deflection of about 10 cm., make at least six careful 
 readings of the deflection on each side given by the standard 
 Daniell cell. Replace the standard by one of the cells 
 to be tested and vary the resistance of the circuit until the 
 deflection is the same as you found it on this side for the 
 standard. Similarly find the resistance which will make 
 the deflection on the other side the same as that given by 
 the standard on that side. We can neglect, in comparison 
 with the resistances in the boxes, the resistance of the bat- 
 tery and connecting wires, but not the resistance of the gal- 
 vanometer. Take the mean of the two resistances deter- 
 mined above, plus the resistance of the galvanometer, as the 
 resistance required to give the same deflection as the stand- 
 ard cell gave through the box-resistance used with it, plus 
 the galvanometer resistance. The resistance of the galvan- 
 ometer, G, must be determined as in Exp. XLV (last 
 paragraph) . 
 
 In determining the possible error of your results, estimate 
 the possible error of resistances from the least change in 
 
COMPARISON OF E. M. F.'S BY CONDENSER METHOD. 193 
 
 resistance which will have an appreciable effect, and the 
 possible error of deflections from the mean deviation from 
 the mean in your readings. 
 
 Questions. 
 
 1. What are the advantages and disadvantages of the type of 
 galvanometer used in this experiment compared with other types 
 used in the laboratory? 
 
 2 . Which of the two methods do you consider the better ? Why ? 
 
 3. How could this method be used for finding the internal resist- 
 ance of a cell? 
 
 4. Are the deflections of a galvanometer strictly proportional to 
 the currents? Why? 
 
 LVII. COMPARISON OF E. M. F.'S AND MEASUREMENT 
 
 OF BATTERY RESISTANCE BY CONDENSER 
 
 METHOD. 
 
 Text-book of Physics (Duff), pp. 530-534, 562; Watson's Physics, pp. 
 634; Watson's Practical Physics, p. 526; Henderson's Electricity 
 and Magnetism, pp. 185-187. 
 
 When a condenser of capacity C is connected to a battery 
 of e. m. f. E it receives a charge Q = CE. If it be then con- 
 nected to a ballistic galvanometer, the 
 throw, d, will be proportional to Q, or Q = 
 K . d, where K is a constant. We shall apply 
 this to (A) compare e. m. f. 's and (B) measure 
 the resistance of cells. We shall describe 
 these separately, but in practice they may 
 be combined. 
 
 (A) Suppose the condenser is first charged 
 by a battery of e. m. f., E v and the deflec- 
 
 tion when connected to the ballistic galva- 
 
 , 
 nometer is a u and suppose that when this 
 
 same condenser has been charged by a battery of e. m. f., 
 E 2 , the deflection is d 2 ; then 
 
 Kd,; Q 2 =CE 2 =Kd 2 
 E d d 
 
 FIG. 60. 
 
194 
 
 ELECTRICITY AND MAGNETISM. 
 
 Use a key with an upper and a lower contact. The 
 condenser should be connected to the battery when the key 
 is down and to the galvanometer when the key is up. Be 
 very careful never to connect the battery directly to the 
 galvanometer. When a discharge is sent through a ballistic 
 galvanometer, the needle swings over to one side and then 
 swings back. Observe the reading of the scale on the verti- 
 cal cross-hair of the telescope when the needle stops and 
 turns back. Always in such work estimate tenths of the 
 smallest division. Before each throw bring the needle as 
 nearly as possible to rest. The zero is likely to change; 
 
 | AUS 
 
 
 P 
 
 
 \\ | 
 
 
 F 
 
 
 | EIN 
 
 FIG. 61. 
 
 therefore, before each throw, record the zero, and, after each 
 throw, record both the turning-point and the difference 
 between this turning-point and the zero; i. e., the amount 
 of the throw. 
 
 Always charge the condenser for approximately the 
 same length of time, for instance, five seconds. With a 
 standard Daniell cell (p. 159), record six throws on one side. 
 Reverse the battery connections and record six throws on 
 the other side. Let the mean of these be d r Replace the 
 Daniell cell by one of another type and find as before the 
 mean throw d 2 . If E 2 is the e. m. f. of the latter cell 
 
 Measure in this way the e. m. f. of as many cells of different 
 type as possible. 
 
 (B) This method of finding the resistance of a cell depends 
 on the fact that when the poles of a cell of resistance B are 
 
COMPARISON OF E. M. F.'s BY CONDENSER METHOD. 195 
 
 joined by a conductor of resistance r, the difference of 
 potential of the poles depends on the ratio of r to B. Let 
 the current be i. Then the difference of potential of the 
 poles is ri by Ohm's Law applied to the part r of the circuit. 
 If in this condition the poles be joined to a condenser of 
 capacity C it will receive a charge Cri and if this when 
 discharged through a condenser causes a throw d f 
 
 Cri=K-d' 
 
 Now if E be the e. m. f. of the cell, E=(B +r)i by Ohm's 
 Law applied to the whole circuit. Hence, if the condenser be 
 charged by the cell when it is not short-circuited as above, 
 the charge C E also equals Ci(B +r) and if the deflection 
 when the condenser is discharged is d, 
 
 Dividing one equation by the other and solving for B, 
 
 d-d' 
 B = r ~^- .. 
 
 The connections are the same as when comparing e. m. f.'s 
 with the addition of a circuit containing a resistance 
 and a very low resistance key (e. g., a mercury key), con- 
 necting the poles of the cell. The battery should be short- 
 circuited just before the charging key is depressed, and the 
 short-circuiting key should be released immediately after 
 the other, otherwise the battery will run down. Choose 
 such a short-circuiting resistance that the galvanometer 
 throw is reduced to about half the value which it has without 
 the short-circuit. Do not use the plug-box resistances 
 for this work, on account of the danger of burning them out, 
 but use open wound resistances of large wire.. Find the 
 internal resistance of cells of several different types. 
 
 In estimating the possible error of your results, estimate 
 the possible error of your mean readings from the mean 
 deviation from the mean in the individual readings. 
 
 Additional exercises (to be performed if time permit.) 
 
 (C) Study the effect of length of time of charge by means 
 
196 ELECTRICITY AND MAGNETISM. 
 
 of the throws obtained with the condenser charged for 
 different lengths of time with the same battery. 
 
 (D) Study the leakage of the condenser by comparing the 
 throws when the condenser has been successively charged 
 with the same e. m. f., and has remained charged for different 
 intervals of time. 
 
 (E) Study the electric absorption of the condenser by 
 charging for several minutes, discharging and reading the 
 throw and immediately insulating; after one minute, again 
 discharge and insulate. Continue discharging for several 
 minutes, the condenser being insulated during the minute 
 intervals. 
 
 Questions. 
 
 1. What are the peculiarities and requirements of a good ballistic 
 galvanometer ? 
 
 2. What is the construction of a condenser and what do absorption 
 and leakage mean? 
 
 3. How could you find the resistance of the galvanometer used, em- 
 ploying a condenser and a known resistance ? 
 
 LVIII. CALIBRATION OF A VOLTMETER. 
 
 Ames 1 General Physics, pp. 674-675; Text-book of Physics (Duff), 
 PP- 57~57 1; Watson's Practical Physics, pp. 493-498; Hender- 
 son's Electricity and Magnetism, pp. 200-208; Hadley's Electricity 
 and Magnetism, p. 320. 
 
 A voltmeter may be calibrated by balancing a part of the 
 e. m. f. applied to the terminals of the voltmeter against 
 the e. m. f. of one or more standard cells. To do this a 
 very high resistance circuit, consisting of -a high resistance- 
 box in series with an ordinary resistance-box, is placed in 
 parallel with the voltmeter. The fall of potential in part 
 of the low resistance-box is measured by a side-circuit con- 
 sisting of the standard cell, a sensitive galvanometer and a 
 key, the standard cell being so turned that it tends to 
 send a current through the galvanometer in the opposite 
 direction to the fall of potential in the box. 
 
 Let e be the e. m. f. of the cell, E the potential difference 
 at the terminals of the voltmeter, r 1 the resistance across 
 
CALIBRATION OF A VOLTMETER. 
 
 197 
 
 which the galvanometer circuit is connected, and r^ the 
 remaining resistance in the high-resistance circuit. When 
 r l is adjusted so that there is no deflection when the key is 
 pressed 
 
 A special fuse-wire for very low currents should be placed 
 immediately adjacent to the battery to prevent the possi- 
 bility of injury to the resistance-boxes. The main circuit 
 should be closed through a spring-key only a sufficient 
 length of time to enable the voltmeter or the galvanometer 
 to be read. A high resistance should be placed in series 
 with the standard cell to prevent any considerable current 
 passing through it. In first 
 performing this experiment 
 it is well to use a simple 
 and inexpensive f6rm of 
 standard cell, and the 
 Daniell cell (p. 159) will be 
 suitable. For later and 
 
 more accurate work, either 
 the Clark or the Weston cell 
 should be used. By vary- 
 ing the number of cells in 
 the main circuit or using different resistances in the main 
 circuit, different voltages at the terminals of the voltmeter 
 may be obtained. 
 
 The above method will not apply if the voltmeter is to 
 be tested at voltages less than the e. m. f. of the standard 
 cell. In this case, an inversion of the connections may be 
 used. Instead of balancing a variable part of the voltage 
 against the e. m. f. of the cell a variable part of the e. m. f. 
 of the cell is balanced against the voltage. A little con- 
 sideration will indicate the necessary change of connection. 
 
 Instead of the above temporary arrangement of circuits, 
 a Potentiometer, which consists essentially of the several 
 
198 ELECTRICITY AND MAGNETISM. 
 
 circuits with the necessary resistances and keys in perma- 
 nent connection, may be used. With its aid the work may 
 be performed more rapidly and more accurately. Its parts 
 and connections should be carefully traced out with the 
 assistance of a large diagram (which may be attached to the 
 wall near the instrument) and additional explanation will be 
 supplied by the instructor. 
 
 A calibration curve, consisting of true volts plotted 
 against scale readings, should be drawn. 
 
 Questions. 
 
 1. Prove the above formula by applying Kirchoff's laws. 
 
 2. Draw a diagram showing the connections when a millivoltmeter 
 has to be calibrated. 
 
 3 . Draw a diagram to show how the above method could be adapted 
 to compare the e. m. f.'s of cells. 
 
 LIX. CALIBRATION OF AMMETER. 
 
 Hadley's Electricity and Magnetism, p. 325; Watson's Practical 
 Physics, pp. 516517; Henderson's Electricity and Magnetism, p. 
 205. 
 
 A method somewhat similar to that used for the volt- 
 meter may be employed. The current from a storage 
 battery that passes through the ammeter passes also through 
 a conductor of large current capacity and of measured 
 resistance and a switch. The potential difference at the ends 
 of this conductor is found by a shunt circuit, consisting of 
 a high resistance-box in series with a box containing low 
 resistances. In parallel with the latter is a circuit containing 
 a Daniell cell, a sensitive galvanometer and a key. A 
 special very fine fuse-wire should be used in series with the 
 two boxes, and its resistance should be known and taken 
 account of in calculating the current. The large conductor 
 should be immersed in oil and its temperature kept as nearly 
 as possible at the temperature at which its resistance is 
 determined. To prevent heating, the main current should 
 be closed for short intervals only. 
 
CALIBRATION OF AMMETER. 
 
 199 
 
 The galvanometer should be protected by a shunt during 
 the first adjustments. Notice first in which direction the 
 galvanometer moves when the key in its circuit is depressed. 
 The deflection should be reversed or reduced when in 
 addition the switch is closed. If this is found not to be so, 
 the connection of either the Daniell cell or the storage 
 batteries should be reversed. The resistance in the box 
 nearest the galvanometer and, if necessary, in the other 
 box also, should be varied 
 until there is no deflection if 
 the galvanometer key is de- 
 pressed when the switch is 
 closed. 
 
 When the adjustment has 
 been obtained as closely as 
 possible, the fall of potential 
 between the points of the 
 high-resistance circuit to 
 which the standard cell circuit 
 is attached equals the e. m. f. 
 of the cell (p. 159). From this 
 and the resistances of the boxes the fall of potential between 
 the ends of the large conductor is found and then from the 
 resistance of the large conductor the current through it 
 and the ammeter is calculated. The total resistance in 
 the two boxes must be kept high. A preliminary calculation 
 will show about how large the resistance of large capacity 
 should be. A number of currents distributed over the 
 range of the ammeter should be used and from the results a 
 calibration curve should be drawn. 
 
 In the above we have assumed that the voltage applied 
 to the conductor exceeds that of the cell. If the reverse is 
 the case, the arrangement must be inverted; i. e., part of 
 the e. m. f. of the cell must be balanced against the voltage 
 applied to the conductor. This method must be applied 
 when the current is less than the quotient of the e. m. f. of 
 the cell and the resistance of the conductor. 
 
200 ELECTRICITY AND MAGNETISM. 
 
 Instead of the arrangements of circuits above, the Po- 
 tentiometer referred to in Exp. LVIII may be used to 
 measure the fall of potential in the conductor of large cur- 
 rent capacity. 
 
 Questions. 
 
 1. Why must the resistance in the shunt circuit be large? 
 
 2. Storage batteries giving an e. m. f. of 50 volts are available for 
 calibrating an ammeter whose range is from 5 to 25 amperes and 
 whose resistance is 0.2 ohms, (a) What is the least possible value 
 for the resistance of large capacity? (b) What is the greatest 
 value? 
 
 LX. COMPARISON OF CAPACITIES OF CONDENSERS. 
 
 Henderson 1 s 
 's Practical 
 
 Hadley's Electricity and Magnetism, pp. 331334; Henderson's 
 Electricity and Magnetism, pp. 235241; Watson' j 
 
 Physics, pp. 530-535. 
 
 Two or more condensers are to be compared by three 
 methods. 
 
 (A) First Method. Each condenser is charged in turn 
 by the same battery and then discharged through a ballistic 
 galvanometer. Let the capacities of the two condensers be 
 C\ and C 2 . The charges which they receive when connected 
 to a battery of e. m. f. E, are Q 1 =C 1 E, and Q 2 .=C 2 E. Let 
 the throws of the galvanometer when the condensers are 
 discharged through it be d l and d 2 , respectively. Then 
 
 (Exp. LVII). The connections are the same as in Exp. LVII, 
 with the addition of one or more keys to charge alter- 
 nately two or more condensers. If either deflection be too 
 small, additional cells should be added. Storage cells may 
 be used if connections are made through special very fine 
 fuse-wires to protect the resistances. 
 
 If either deflection be too large the galvanometer should be shunted 
 by a known resistance, S. Let G be the resistance of the galvan- 
 ometer determined as in Exp. XLV (last paragraph), d' the 
 throw obtained with the galvanometer shunted, d the throw which 
 
COMPARISON OF CAPACITIES OF CONDENSERS. 
 
 201 
 
 would have been obtained without the shunt, q r the quantity of 
 electricity passing through the galvanometer, q" the quantity passing 
 through the shunt. Then 
 
 9"_5 
 q' S 
 
 for charges of electricity, like steady direct currents, divide inversely 
 as the resistances (p. 158). Hence 
 
 Since 
 
 _* 
 
 q' d" 
 
 (B) Bridge Method. The two condensers to be com- 
 pared, C\ and C 2 , form two arms of a Wheatstone's Bridge, 
 two high non-inductive 
 resistances, R 1 and R 2 (see 
 figure), preferably several 
 thousand ohms, forming 
 the other two arms. These 
 two resistances are ad- 
 justed until on closing the 
 battery circuit at a the 
 galvanometer is not dis- 
 turbed. Then during both r~^ 
 charge and discharge the 
 farther poles of the con- 
 denser (A and B) must 
 remain at the same potential as well as the nearer poles 
 (joined at D). Hence the charges Q^ and Q 2 in the con- 
 densers must have the ratio, Q t : Q 2 ' : C\ : C 2 . But the quanti- 
 ties which have flowed into the condensers will be inversely 
 proportional to the resistances through which the charges 
 have flowed, that is, Q^\ Q 2 : : R 2 : R^ Hence 
 
 FIG. 65. 
 
 C 2 R,' 
 
 The battery key has an upper and lower contact; the 
 upper contact (b), against which the lever ordinarily rests, 
 
2O2 
 
 ELECTRICITY AND MAGNETISM. 
 
 short-circuits the battery terminals of the bridge, thus keep- 
 ing the condenser uncharged. The sensitiveness may be 
 increased by increasing the number of cells in the battery, 
 and also by using a double commutator (see p. 1 6 1 j . Instead 
 of a galvanometer a telephone may be used in this method, 
 the battery being replaced by a small induction coil. 
 
 (C) Thomson's Method of Mixtures. The connections 
 are as shown in the figure. K l is a Pohl's commutator, K 2 
 an ordinary single contact switch. When the swinging 
 
 arm of the commutator is in the position aa', the two con- 
 densers are charged, C l to the difference of potential at the 
 extremities of R lf C 2 to the difference of potential at the 
 extremities of R 2 . The swinging arm of the commutator is 
 now placed in the position b'b' and the two charges are 
 allowed to mix. If they are exactly equal, being of opposite 
 sign, the galvanometer will not be affected when K 2 is 
 depressed. R l and R 2 (which are large resistances, pref- 
 erably several thousand ohms), are adjusted until this is 
 secured. The charges being equal, C 1 V l = C 2 V 2 , and since 
 
ABSOLUTE MEASUREMENT OF CAPACITY. 203 
 
 Questions. 
 
 1. What is the composite capacity of three microfarad condensers 
 in parallel? In series? 
 
 2. Which of these three methods do you consider best? 
 
 3. State briefly in words (without formulae) why charges divide 
 like steady currents; i. e., inversely as the ohmic resistances. 
 
 4. Why cannot series resistance be used in (A) to reduce the 
 sensitiveness of the galvanometer. 
 
 LXI. ABSOLUTE MEASUREMENT OF CAPACITY. 
 
 The magnitude of a capacity can also be found without 
 the use of a known capacity with which to compare it. This 
 can be done in different ways. The following is one of the 
 simplest. 
 
 Let Q be the charge received by the condenser of unknown 
 capacity C when connected to a cell of known e. m. f., E. 
 Then 
 
 To find Q discharge the condenser through a ballistic galvan- 
 ometer the constant of which has been found by the method 
 of Exp. LXIV. If the constant be K and the deflection 
 D t Q = KD. 
 
 LXII. COEFFICIENTS OF SELF-INDUCTION AND OF 
 MUTUAL INDUCTION. 
 
 Text-book of Physics (Duff), pp. 609-611; Ames' General Physics, 
 PP- 743-745; Watson's Practical Physics, pp. 543-548; Parr's 
 Practical Electrical Testing, p. 207; Hadley's Electricity and 
 Magnetism, pp. 417-422. 
 
 The coefficient of self-induction of a circuit is the number 
 of magnetic lines of force which link with the current when 
 the circuit is traversed by unit current. Owing to the 
 difficulty of calculating this important quantity from the 
 dimensions of the circuit, experimental methods of deter- 
 mination have much value. 
 
204 
 
 ELECTRICITY AND MAGNETISM. 
 
 (A) Probably the best method (using direct currents) 
 is Anderson's Modification of Maxwell's Method. The 
 connections are shown in the figure. The coil of self-in- 
 duction L and resistance Q is made one arm of a Wheat- 
 stone's Bridge (preferably Post-office Box form). Ob- 
 tain a balance for steady currents by proper 
 variation of 5, so that when K l is closed 
 and then K 2 , the galvanometer is not dis- 
 turbed. For delicacy of adjustment it is 
 well to either have a resistance which can 
 be varied continuously form a part of 5, 
 or make the ratio arms P and R such that 
 5" is large. Vary r, the resistance in the 
 battery circuit, and if necessary, vary the 
 capacity of the condenser C until there is 
 a balance for transient currents; i. e., 
 until the galvanometer is not disturbed 
 
 FIG. 67. 
 
 when K 2 is depressed and K l depressed afterward. Then 
 if C is the capacity of the condenser. 
 
 For at time t\et x = current in branch AB, y = current in AD = cur- 
 rent in D E, = current in BE .'. current in r = y + z. q = charge in 
 condenser, e = potential difference of its poles = Rz + r (y + z) . 
 
 Since there is a balance for transient currents we may equate the 
 e. m. f. in AD to that in AB. Hence 
 
 The current in the branch containing the condenser is (x z) ; but 
 it can also be expressed as 
 
 dq r de 
 dJ mC di 
 
 Hence 
 
 
 Now since there is a balance for steady currents RQ = PS and 
 since Rz = Sy, it readily follows that 
 
 = Qy + C[r 
 
COEFFICIENTS OF INDUCTION. 
 
 If the resistances are expressed in ohms and the capac- 
 ity in farads, the results will be in henries. 
 
 Measure the self-induction of a coil whose length is 
 great compared with its diameter and compare the result 
 with that calculated. To calculate the coefficient of self- 
 induction it is necessary to know the number of lines of 
 force passing through the coil. This number multiplied by 
 the number of turns will give the number which link with 
 the current; i. e., the self-induction. If A be the area of 
 the cross section of a solenoid of practically infinite length, 
 with n Q turns per cm. of 
 length, the number of lines 
 is 4xn A for unit current. 
 The number of turns in a 
 length d is n Q d; hence the 
 coefficient of self-induction of 
 this length is 47iAn Q 2 d in C. 
 G. S. units. Reduce to 
 henries by dividing by io 9 
 (p. 162). ' 
 
 To secure greater sensitive- 
 ness in making the balance FlG 
 for transient currents, replace 
 
 the battery by the secondary of a small induction coil and 
 the galvanometer by a telephone, or use a double commu- 
 tator (see p. 161). 
 
 (B) Comparison of Two Coefficients of Self-induction. 
 The two coils of self-induction L lf L 2 , and resistances R v 
 R 2 , are placed in two arms of a Wheatstone's Bridge, a 
 variable resistance, r, being included in one arm. By vary- 
 ing r, and, if possible, by varying one of the self-inductances, 
 if not, by varying r, P, and Q, find a balance for both steady 
 and transient currents. 
 
 Then for steady currents 
 
206 ELECTRICITY AND MAGNETISM. 
 
 and for transient currents 
 
 P 
 
 where <o = 271 X frequency. Hence 
 
 The above adjustment is obtained by first securing a 
 balance for steady currents. A balance for transient cur- 
 rents is then sought by varying L l or L 2 . If this cannot 
 be secured, r and P must be both increased or decreased 
 and a balance for steady currents again obtained and then 
 
 one for transient currents found by 
 varying L t or L 2 . If neither L l 
 nor L 2 can be varied, a balance can 
 only be obtained by a series of trials 
 as above, the ratio of P and (7^ +r) 
 being kept constant, so that the 
 steady current balance may not be 
 disturbed. To increase the sensi- 
 tiveness with transient currents, a 
 double commutator (p. 161) should 
 be used or the battery may be re- 
 placed by the secondary of an induction coil and the galva- 
 nometer by a telephone. 
 
 (C) The coefficient of mutual induction of two coils is the 
 number of lines of force which link with the turns of the other 
 when the first is traversed by unit current. Pirani's 
 method is perhaps the most satisfactory for the experimental 
 determination of coefficients of mutual induction. The 
 connections are shown in figure 69. If M be the required 
 coefficient, C the capacity of the condenser, and r^ and r 2 
 the values of the variable resistances for which the galvan- 
 ometer is not disturbed, 
 
 FIG. 69. 
 
STRENGTH OF A MAGNETIC FIELD. 207 
 
 For let the steady current in the battery circuit = i. The potential 
 difference at the terminals of r lt = ir^ The charge of the condenser is 
 Cir r If t time required to establish or destroy the battery current 
 the average current in the condenser branch during this time = 
 Cir l /t and the potential difference at the terminals of r. 2 = Cir l r. J /t. 
 Opposing this e. m. f. in the galvanometer circuit is that due to M, 
 the average value of which = Mi/t. If the galvanometer is not dis- 
 turbed on making or breaking the battery circuit, Cir^r^t^Mi/t 
 .'. M = Cr l r. 2 . 
 
 The secondary of M is acted on by two e. m. f.'s one due 
 to its connection with the main circuit in which there is an 
 e. m. f., the other due to mutual induction between the 
 primary and secondary. If these two do not oppose one 
 another a balance cannot be found. If such is found 
 to be the case the connections of the secondary to the 
 galvanometer circuit must be reversed. To increase the 
 sensitiveness of the method a double commutator may be 
 used or, better still, the battery and galvanometer may be 
 replaced by a small induction coil and telephone. 
 
 When an approximate adjustment has been found C and 
 r l should be altered until the sensitiveness is a maximum, 
 and then r l and r 2 treated in the same way. 
 
 Find the coefficient of mutual induction of two coils, one 
 wound upon the other, and one of which is long compared 
 with its diameter, and compare the result with that calculated 
 from the definition. 
 
 QUESTIONS. 
 
 1. Coils with iron cores do not have definite induction coefficients. 
 Explain. 
 
 2. How are resistance coils in boxes wound so as to be free from 
 self-induction ? 
 
 LXIII. STRENGTH OF A MAGNETIC FIELD BY A 
 BISMUTH SPIRAL. 
 
 Ames' General Physics, p. 758; Hadley' s Electricity and Magnetism, 
 
 p. 296. 
 
 The electrical resistance of a bismuth wire is changed 
 when it is placed transverse to a magnetic field and the 
 magnitude of the change depends on the strength of the field. 
 
208 ELECTRICITY AND MAGNETISM. 
 
 When a curve representing the resistance of a flat spiral 
 of bismuth as a function of the strength of the magnetic 
 field has been obtained the spiral may, in connection with a 
 Wheatstone's Bridge, be used to measure the strength of any 
 magnetic field within the range of the calibration. For 
 instance, it may be used to study the magnetic field of an 
 electromagnet. The following three points may be examined : 
 
 (A) Find how the magnetic field between the poles varies 
 when the strength of the current actuating the electro- 
 magnetic is varied by means of a rheostat. 
 
 (B) Find how the strength of the field midway between 
 the pole-pieces changes when the distance apart of the pole- 
 pieces is varied, the current being kept constant. 
 
 (C) Find how the strength of the field in an equatorial 
 plane varies with the distance from the axis of the pole-pieces. 
 
 In each case represent the results by means of a curve. 
 
 LXIV. CONSTANT AND RESISTANCE OF A BALLISTIC 
 GALVANOMETER. 
 
 Text-book of Physics (Duff), pp. 556, 562; Ames' General Physics, p. 
 674, 712; Watson's Practical Physics, pp. 518-524; Pierce, Am. 
 Acad. Arts and Sc. Vol. 42, pp. 159160. 
 
 When quantities of electricity are discharged through a 
 ballistic galvanometer (p. 156) the throws are proportional 
 to the quantities or, 
 
 Q=K.d, 
 
 where K is the constant of the ballistic galvanometer. To 
 determine the value of K a known quantity must be dis- 
 charged through the galvanometer and the throw noted. 
 
 This known quantity might be obtained from a condenser 
 of known capacity, charged to a known potential, or by 
 turning an earth inductor (Exp. XLIII) in a field of 
 known strength. Both of these methods, however, re- 
 quire that other constants (capacity and e. m. f. in the first, 
 strength of field in the second) be determined. 
 
BALLISTIC GALVANOMETER. 2OQ 
 
 A simpler method is to use a so-called calibra ting-coil ; 
 i. e., an induction coil of known winding without a magnetic 
 core. The primary is a long, straight helix, so long that 
 there is no appreciable leakage near the center. Over the 
 center there is wound a secondary. If the primary be of n 
 turns per cm. and the secondary be of n' total turns, then 
 the magnetizing force produced by a current of i amperes 
 in the primary is 
 
 10 
 
 and the quantity induced in the secondary by making or 
 breaking i is 
 
 han' 
 
 where a is the area of cross section of the cylinder on which 
 the primary is wound, r is the total resistance of the second- 
 ary circuit and the factor io 8 is required when q is in 
 coulombs and r in ohms (p. 162). 
 
 The number of lines of force that pass through the secondary is 
 a h. Hence when h is increasing the induced e. m. f. is in absolute 
 units 
 
 ,d(ah) 
 
 The quantity induced is I i dt and i equals e/r. Hence 
 
 Jn' d(ah) n'ak 
 7~dT ' 
 
 From the above expression for q and the throw d, K can 
 be calculated. If the throw is small it may be doubled by 
 reversing i and the half of the double throw taken for d. 
 Several currents should be tried. The value of K thus 
 found is in coulombs per scale division. If the distance of 
 the scale from the galvanometer be changed in any pro- 
 portion K will be increased in the same proportion. Hence 
 the distance should be noted unless it is not liable to change. 
 
 It is often necessary to change the sensitiveness of a ballistic 
 galvanometer by shunting it or putting resistance in series 
 14 
 
210 ELECTRICITY AND MAGNETISM. 
 
 with it. To allow for this we must know the resistance of 
 the galvanometer. 
 
 The resistance of a ballistic galvanometer when used 
 ballistically on a closed circuit is different from its resistance 
 when used as an ordinary galvanometer. This is due to the 
 fact that the galvanometer coil moves in a magnetic field, 
 and thus an induced e. m. f. is produced, which opposes the 
 applied e. m. f. and has the same effect as an added resist- 
 ance. To find the effective resistance of a ballistic gal- 
 vanometer we may use the same apparatus and connections 
 as in finding the constant of the galvanometer. If a 
 current be reversed in the primary of the calibrating coil, 
 the quantity of electricity that will flow through the second- 
 ary will vary inversely as the total secondary resistance. 
 Hence, by observing the throw with a certain primary 
 current, and then increasing the secondary resistance by the 
 insertion of a box-resistance and repeating the reversal of 
 the primary, we can, by proportion, find the resistance of 
 the galvanometer when used ballistically. 
 
 The resistance of the galvanometer should also be found 
 by the method of Exp. XLVI or that of Exp. XLV and 
 compared with the above. 
 
 LXV. MAGNETIC PERMEABILITY. 
 
 Text-book of Physics (Duff), pp. 500503; Ames' General Physics, pp. 
 609, 615; Watson's Physics, pp. 712-714; Hartley's Electricity 
 and Magnetism, pp. 384-392; Watson's Practical Physics, pp. 
 563-568; Henderson's Electricity and Magnetism, pp. 282-284; 
 Swing's Magnetism in Iron, Chapter III. 
 
 A current in a long solenoid of wire will produce near 
 the center of the solenoid a magnetic force H, which may 
 be specified by the number of lines of force per unit of area 
 at right angles to the lines. If a long iron rod be now 
 thrust into the solenoid, the number of lines of force (now 
 called lines of induction) will be much greater, say B per 
 unit of area. The permeability of the iron is defined as 
 u=B+H. 
 
MAGNETIC PERMEABILITY. 211 
 
 If this experiment were performed with comparatively 
 short iron rods, it would be found that B would be less the 
 shorter the rod. One consistent way of explaining this is 
 to consider the free poles developed at the ends of the rod 
 when magnetized. A little consideration will show that 
 they of themselves would produce a magnetic force in the 
 space occupied by the iron, this magnetic force being opposed 
 to the original magnetizing force, and so we may say that 
 the effective magnetic force, H, is the original magnetic 
 force diminished by the demagnetizing force of the poles. 
 It is this effective magnetic force that we should divide into 
 the induction to get the permeability. The calculation of 
 the demagnetizing force is usually difficult and uncertain, 
 and so it is better to take some method of eliminating it. 
 
 One such way is that implied in the statement at the out- 
 set, to use a long rod, for that will diminish the magnitude 
 of the demagnetizing force at the center. But the necessary 
 length makes it inconvenient to test specimens in this way. 
 Another method is to join the ends of the rod by a heavy 
 yoke of iron, for opposite poles developed in the yoke neu- 
 tralize the effect of the poles in the rod. (This is one way of 
 stating the case. Another way is to say that the yoke carries 
 around the lines of force. A third way is to say that the 
 yoke diminishes the magnetic resistance of the circuit.) 
 The difficulty with a yolk method is in getting a satisfactory 
 contact between yoke and rod. A very small gap will re- 
 sult in the neutralization being not quite complete (or in 
 leakage of lines of force or in magnetic resistance) . 
 
 A more satisfactory method is to take an endless speci- 
 men; i. e., a ring. Then there are no free poles and no de- 
 magnetizing force. On the ring a magnetizing coil of N 
 turns per cm. is wound. When a current of / amperes passes 
 through it, the magnetizing force produced is 
 
 IO 
 
 For finding the value of. B a, secondary coil is wound on the 
 
212 
 
 ELECTRICITY AND MAGNETISM. 
 
 ring and put in series with a ballistic galvanometer. Sup- 
 pose the iron initially free from magnetism. The setting up 
 of the field B produces a discharge, Q, of electricity through 
 the secondary. If A be the area of cross section of the ring 
 and N' the total number of turns, 
 
 N'AB 
 
 R being the total (ohmic) resistance of the secondary circuit. 
 (The factor io 8 is not necessary if Q and R are in absolute 
 units. It must be used when Q is in coulombs and R in 
 ohms (p. 162). If the throw of the galvanometer is D 
 
 Q=K-D 
 
 when K is the ballistic constant (Exp. LXIV). If K is 
 not known, a calibrating coil for determining it should be 
 included in the arrangement of the apparatus. From the 
 above formulae and the data, H, B and fi can be calculated. 
 
 VVVVVVVS.'Vv'V'S 
 
 
 FIG. 70. 
 
 In making the connection for the practice of the method, 
 it is much better to have a clear understanding of the plan 
 and purpose of each part and to proceed systematically 
 than to copy the connection from a diagram. In the first 
 
MAGNETIC PERMEABILITY. 213 
 
 place, the secondaries of both coil and ring should be kept 
 permanently in series with the galvanometer. Then a switch 
 is to be so arranged that the current can be passed through 
 either the primary of the calibrating coil or that of the ring. 
 A suitable rheostat and ammeter are needed in the primary 
 circuit. If, as the primary current is increased, the deflec- 
 tions of the galvanometer become too great to be read, a 
 resistance must be put in series or in parallel with the gal- 
 vanometer. The former is preferable. In choosing this 
 added series resistance, it is well to so choose it that the 
 whole new secondary resistance is made a simple multiple 
 of the former resistance. If this is done the throw will be 
 reduced in the proportion in which the resistance is increased, 
 and all throws may be reduced to what they would have been 
 with the original resistance by multiplying the actual throw 
 by the proportion in which the secondary resistance was 
 increased. For methods of bringing the galvanometer to 
 rest, see p. 156. 
 
 Before readings are begun the ring should be demagnetized 
 as thoroughly as possible. This can be done by passing an 
 alternating current through the primary and reducing it 
 from a large value to zero by means of a rheostat, or, by 
 rapidly commutating and at the same time reducing a 
 direct current. Also at each new value of the magnetizing 
 current, before readings are taken, the commutator should be 
 reversed several times, so that the iron may come to a steady 
 cyclical state. Instead of attempting to get the throw on 
 making the primary current, the double throw on reversing 
 the current is taken with both calibrating coil and ring 
 and divided by 2. 
 
 At least three throws that agree well should be read for 
 each strength of the primary current. The magnetizing 
 current should be increased at first by small steps to bring 
 out the characteristic features of the curve of magnetization, 
 afterward by larger steps. The work need not be continued 
 after the readings begin to differ in a much smaller propor- 
 tion than the successive magnetizing currents, for this shows 
 
214 ELECTRICITY AND MAGNETISM. 
 
 approaching saturation. The throw at break of current 
 should also be carefully noted as a means of estimating the 
 permanent magnetism; for from the throw at break the 
 diminution of B, and, therefore, the residual value of B, 
 can be calculated as above. 
 
 In the report the various values of /, H, Q, B, and /* 
 should be tabulated and a curve drawn with B as ordinates 
 and H as abscissae (B-H curve or curve of magnetization). 
 On the same sheet a B-fi curve should also be drawn and a 
 third curve showing the permanent magnetism as deduced 
 from the throws at break of the current. 
 
 Questions. 
 
 1. Why is only the ohmic and not the self -inductive resistance of 
 the secondary considered ? 
 
 2. What is the effect of the windings being closer together on the 
 inside of the ring than on the outside? 
 
 3. What is meant by intensity of magnetization? Susceptibility? 
 Calculate a few values from your results. 
 
 LXVI. MAGNETIC HYSTERESIS. 
 
 Text-book of Physics (Duff), pp. 503-504; Watson's Physics, pp. 716 
 722; Hartley's Magnetism and Electricity, pp. 393-395. Watson's 
 Practical Physics, p. 561; Henderson's Electricity and Mag- 
 netism, p. 294; Ewing's Magnetism in Iron, Chapter V. 
 
 Let a magnetizing force applied to a specimen of iron as in 
 the preceding experiment be increased step by step and let 
 the resulting increases of magnetization be observed. At 
 some stage let the process be stopped and then the mag- 
 netizing force decreased by the same steps. It will be found 
 that the steps of decrease of magnetization are less than 
 those by which it at first increased, or the magnetization 
 lags behind the magnetizing force. This is called hysteresis. 
 For a complete view of the process a cycle must be com- 
 pleted, i. e., the magnetizing force must be decreased step 
 by step to zero, then increased to a negative value equal 
 (numerically) to the positive value at which the decreases 
 were begun, then decreased again to zero, and finally 
 
MAGNETIC HYSTERESIS. 
 
 215 
 
 increased again to the highest positive value. Thus a 
 hysteresis loop will be obtained. 
 
 With a ring specimen, over which primary and secondary 
 coils are wound, there are two methods of procedure. 
 
 (A) Step by Step Method. This method follows closely 
 the general description given above. The successive steps 
 are indicated in figure 71. The increases or decreases of 
 7 must be made without break of the current. The steps 
 must not be too large or the points on the curve will be 
 too far apart, and they must not be too small or the work 
 
 -I -I, 
 
 FIG. 71. 
 
 will become tedious. To satisfy these conditions, place in 
 the primary circuit a special rheostat consisting of suitable 
 resistances in parallel, each of which can be short-circuited 
 by a knife-edge switch. Such a rheostat may be made up 
 with resistances permanently connected in position, but a 
 better plan is to use removable resistances. In the latter 
 case a considerable collection of units should be supplied, 
 and from these, by a preliminary trial, units that will 
 produce suitable changes of I (e. g. , from 4 to o . 5 amp. by steps 
 of o. 5 amp.) should be chosen and placed in position in the 
 rheostat. 
 
 It is not necessary to start from zero magnetization. 
 
2l6 ELECTRICITY AND MAGNETISM. 
 
 Beginning with the highest current to be used, reverse 
 several times to produce a cyclical state and then find the 
 throw on reversal. From this the maximum value of B can 
 be calculated as in Exp. LXV. Then dimmish the current 
 by steps and note the throw in each step. After the step 
 that reduces the current to zero, the current must be 
 reversed and the resistances increased step by step. The 
 rest of the process needs not be described. From each throw 
 the corresponding change of induction, AJ5, is calculated 
 as in Exp. LXV. When the cycle has been completed the 
 algebraic sum of the throws should be zero. It should not 
 be necessary to change the sensitiveness of the galvanometer; 
 it will give the smaller throws with less accuracy, but they 
 are less important. This "step by step" method of meas- 
 uring hysteresis is the most instructive and is not difficult 
 after some initial practice. It has, however, the disad- 
 vantage that an error in one reading of the galvanometer 
 vitiates the whole. 
 
 (B) The Ewing-Classen Method. The last-mentioned dis- 
 advantage is avoided in the method by starting each 
 step from the maximum value of B. As before, we first find 
 by reversals the value of B corresponding to the maximum 
 value of /. We then diminish I (without breaking the cur- 
 rent) and from the throw we calculate the diminution of 
 B. This gives us a second point on the curve. We then 
 return to the maximum current and, after several reversals, 
 to re-establish the cyclical state, we again decrease /, but by 
 a larger amount than before. From the throw we again 
 calculate the diminution of B and thus get another point 
 on the curve. Proceeding in this way, we reach the stage 
 at which I is decreased from its maximum to zero. This 
 gives us the point at which the curve crosses the axis on 
 which B is plotted. 
 
 A simple method of producing the above changes of 7 is to 
 connect the rheostat described under (A) in parallel with one 
 of the cross-bars of the Pohl's commutator used for reversing 
 / (Fig. 72). If this cross-bar be suddenly removed, the 
 
MAGNETIC HYSTERESIS. 
 
 217 
 
 resistance in the rheostat will be thrown into the circuit 
 without breaking the current. 
 
 By the above process, we have obtained that part of the 
 descending branch of the hysteresis loop, which lies to the 
 right of the B axis. To obtain the remainder of the branch, 
 we again proceed by steps from the positive maximum value 
 of B, but, since each change of / will carry it from its positive 
 maximum to a smaller negative value, we must simultane- 
 ously diminish and reverse the current. To be able to do this, 
 remove the cross-bar of the commutator which is in parallel 
 with the rheostat and turn the commutator so that the 
 
 r 
 
 FIG. 73. 
 
 current flows to the ring, but does not pass through the 
 rheostat (Fig 73). If the commutator be now reversed, the 
 current will be reversed and will be diminished by passing 
 through the rheostat. Thus we get another point on the curve 
 and, by a series of such steps with decreasing resistances in 
 the rheostat, the descending branch of the loop is completed. 
 To trace the other branch we might proceed as above, begin- 
 ning each step from the negative maximum of B. This, 
 however, is unnecessary, since we would evidently be merely 
 repeating the previous readings. The loop is symmetrical 
 about the origin, and the co-ordinates of the ascending branch 
 
2l8 ELECTRICITY AND MAGNETISM. 
 
 are equal to those of the descending branch but with signs 
 reversed. 
 
 It can be shown that the energy expended in such a cyclical 
 change of magnetization is 
 
 C 
 
 xJ 
 
 fidB 
 
 4xJ 
 
 ergs per c.c. of the iron. The integral also represents the 
 area of the loop, due allowance being made for the scale on 
 which it is plotted. Hence if the area be found by means 
 of a planimeter (the use of which will be explained by an 
 instructor), the energy loss per c.c. per cycle can be cal- 
 culated. 
 
 The total number of lines of induction through each turn of the 
 magnetizing coil is AB. Since the total number of turns is IN , 
 when B is being increased there is induced in the magnetizing coil an 
 e. m. f. 
 
 _, d(lNAB} jlB r n c ., 
 
 E = -- - JT - = VN-jT C. G. S. units. 
 at at 
 
 V being the volume of the core ( = IA). The work done by the 
 battery in time dt in overcoming this opposing e. m. f. is 
 
 dW=IEdt=INVdB ergs 
 Now the area of the hysteresis loop is the integral of HdB and 
 
 ,W - I HdB 
 
 Questions. 
 
 1. What rise of temperature would 1000 cycles produce in the iron 
 if no heat were lost ? 
 
 2. How much less would the energy loss be if the maximum 
 magnetization were half as great as in your cycle? 
 
THE MECHANICAL EQUIVALENT OF HEAT. 2IQ 
 
 LXVII. (A) THE MECHANICAL EQUIVALENT OF HEAT. 
 
 (B) THE HORIZONTAL INTENSITY OF THE 
 
 EARTH'S MAGNETISM. 
 
 Text-book of Physics (Duff), pp. 557, 587-588; Ames' General Physics, 
 pp. 664-665, 688; Watson's Physics, pp. 692-693, 674-677, 775- 
 776; Crew's General Physics, 289, 319; Hadley's Electricity 
 and Magnetism, pp. 336340, 458; Watson's Practical Physics, 
 pp. 508-512. 
 
 If Q calories of heat be produced in a conductor by the 
 passage of a current i for time t, and if no other work, 
 chemical or mechanical, be performed, then 
 
 JQ=?Rt t 
 
 ] being the mechanical equivalent. If i be expressed in am- 
 peres, R in ohms and Q in calories, i 2 Rt will be in joules 
 (one joule being io 7 ergs), and J will be obtained as the 
 number of joules in a calorie. 
 
 Q can be measured by immersing the conductor of resistance 
 R in a known mass of water contained in a vessel of known 
 water equivalent. The mass of water may be obtained 
 with sufficient accuracy by measuring it from a burette. 
 To eliminate the effects of radiation, conduction and con- 
 vection, the water should be at the beginning of the passage 
 of the current as much below the temperature of the room 
 as it finally rises above it, for the current is kept steady and 
 the temperature of the water therefore rises steadily, so 
 that it is as long above the room temperature as below. 
 
 The resistance R may be measured against a standard 
 ohm coil by Wheatstone's Bridge, and, since it will be 
 found necessary to use a wire of comparatively small re- 
 sistance, R should be measured with great care. Leads of 
 large size and small length should be employed for con- 
 necting the wire to the bridge. While being measured it 
 should be immersed in the calorimeter in water at the 
 temperature of the room, so that the mean resistance 
 throughout the experiment is obtained. To reduce to 
 absolute units the resistance in ohms is multiplied by io 9 . 
 
220 
 
 ELECTRICITY AND MAGNETISM. 
 
 The current, i, may be obtained from its chemical effect 
 in another part of the circuit. Careful measurements have 
 shown that unit current (C. G. S.) flowing through a solution 
 of copper sulphate of a certain strength between copper 
 electrodes deposits 0.00326 gms. of copper per second on 
 the cathode. 
 
 The form of copper voltameter employed consists of a 
 glass vessel containing a solution of copper sulphate into 
 which dip three plates. The two outer are of heavy copper 
 
 Calon'meter 
 
 Rheostat 
 
 Voltameter 
 FIG. 74. 
 
 and are both joined, directly or indirectly, to the positive 
 pole of the battery, forming the anode. The intermediate 
 plate is thin and light, and is connected to the negative 
 pole of the battery, forming the cathode. A satisfactory solu- 
 tion consists of 15 grams of copper sulphate dissolved in 
 100 grams of water, to which are added 5 grams of sulphuric 
 acid and 5 grams of alcohol. (The alcohol is easily oxidized, 
 thus preventing the oxidization of the deposit on the 
 cathode and the formation of polarizing compounds at the 
 anode.) 
 
THE MECHANICAL EQUIVALENT OF HEAT. 221 
 
 Clean the two anode plates with sand-paper and fasten 
 them in the two outside binding posts of the top of the 
 voltameter. Clean with sand-paper a cathode plate, wash 
 with tap water and then with alcohol. When dry, weigh 
 on one of the chemical balances, weighing to milligrams 
 with the rider. Wrap in paper and set aside. Be very 
 careful not to touch with the fingers any part of the plate 
 which will be in the solution, after it has been cleaned. 
 Clean with sand-paper a trial cathode and mount it on the 
 middle binding post. Before putting the voltameter in the 
 circuit, dip the two wires which are to be connected to the 
 voltameter in the solution of the voltameter. Decrease the 
 variable resistance until you have a moderate current, but 
 do not entirely cut it out. Notice on which wire copper is de- 
 posited as a brown powder. Connect this wire to the cathode 
 of the voltmeter and the other wire to the anode plates. 
 
 It is important that the current be kept constant. It is 
 true that even if the current vary, the deposit will give the 
 true mean value of the current. But what is needed is the 
 mean value of i 2 , and this is not necessarily the same as the 
 square of the mean value of i. If a storage battery in good 
 condition be used as the source of current, the current will 
 not vary much; nevertheless, a tangent galvanometer or an 
 ammeter should be included in the circuit to test the con- 
 stancy of the current. There is also another reason for 
 including a current meter of some form. The difference 
 of potential at the terminals of the heating coil, or iR, 
 must not be as great as the e. m. f. (1.6 V) that will elec- 
 trolyze water, otherwise some part of the energy of the 
 current will be spent in chemical work. Knowing R, one can 
 choose a safe value for i. If the constant of the galvan- 
 ometer be not known, it can be calculated roughly from 
 the dimensions of the coils and the approximate value (say 
 o.i 8) for the horizontal component of the earth's field (see 
 Exp. XLII), and so the deflection corresponding to a safe 
 value of i deduced. An ammeter, if available, affords a 
 still simpler means. 
 
222 ELECTRICITY AND MAGNETISM. 
 
 If a tangent galvanometer be used a fairly reliable value 
 for the horizontal component' of the earth's field may be 
 deduced from the results of the experiment. For this pur- 
 pose the dimensions of the galvanometer should be care- 
 fully measured and the current through it frequently re- 
 versed and carefully read. The Helmholtz form of tangent 
 galvanometer may be used. This consists of two coils sepa- 
 rated a distance equal to their common radius, with the 
 needle on their common axis midway between them. This 
 arrangement of two coils produces a very uniform field 
 over quite an area where the needle is located, allowing the 
 use of a longer needle. The formula for the galvanometer 
 (the proof for which will be found in text-books on physics) is 
 
 r / # 2 \f 
 i =H 1 i H I tan . 
 
 (If a simpler type of tangent galvanometer, with but 
 one coil, is used, x is the distance from the plane of this coil 
 to the suspension of the needle.) 
 
 Equating this expression, which involves the dimensions 
 of the galvanometer and the deflection, to the current as de- 
 termined by the voltameter, the horizontal component is 
 deduced. 
 
 When the adjustments have been completed, open the 
 switch, remove the trial cathode and put in place the other 
 cathode, which has been kept wrapped in paper. Take 
 care that there is no metallic connection between the 
 cathode and the anode plates. Remove all iron from the 
 neighborhood of the tangent galvanometer and from your 
 pockets. All wires must be close together to avoid stray 
 induction and the galvanometer had best be at some distance 
 from the other apparatus. 
 
 After reading the temperature of the calorimeter every 
 minute for five minutes note the exact second on an ordinary 
 watch, and close the switch. As soon as possible, read both 
 ends of the needle. Reverse the current, making the reversal 
 quickly, and again read both ends of the needle. Always 
 
THERMOELECTRIC CURRENTS. 223 
 
 estimate tenths. Reading both ends of the needle eliminates 
 error due to the axis about which the pointer turns, not 
 coinciding with the center of the graduated circle, and 
 reversing eliminates uncertainty about the reading for the 
 zero position. Keep the current constant with the variable 
 resistance. At intervals of three minutes (approximately) 
 read both ends of the needle and, reversing, again read 
 both ends. Read the temperature every minute to tenths 
 of the smallest division. Allow the current to flow until 
 the temperature has risen to the extent desired. Note the 
 exact second of breaking the circuit. Continue to observe 
 the temperature at minute intervals for five minutes. Re- 
 move the cathode, being very careful not to touch the 
 copper deposit. Wash it gently with tap water and then 
 with alcohol, allowing the liquid to simply flow over the 
 surface. When it is dry, weigh as before. Measure very 
 carefully the diameter of the coils in a number of directions, 
 and from the mean determine r. Count n, the total number 
 of turns in both coils, and measure 2 x, the distance between 
 the centers of the two coils. Weigh the inner calorimeter 
 vessel and note of what metal it consists. Plot the 
 temperature readings and correct for radiation (p. 63). 
 
 Questions. 
 
 1 . Calculate what the exact voltage at the terminals of the heating 
 coil was. 
 
 2. What sources of error remain uneliminated ? 
 
 3. Calculate the mean activity of the current in the coil during the 
 experiment. 
 
 4. What are the peculiarities and advantages of the tangent 
 galvanometer ? 
 
 5. What chemical actions take place in the voltameter? To what 
 is the deposition of any metal proportional? 
 
 6. Why is it advantageous to have the deflection about 45? 
 
 LXVIII. THERMOELECTRIC CURRENTS. 
 
 Text-book of Physics (Duff), pp. 593-598; Ames' General Physics, pp. 
 679-683; Watson's Physics, pp. 696-705; Hartley's Electricity 
 and Magnetism, pp. 359-381. 
 
 To the ends of wires of iron, nickel, silver etc., cop- 
 per wires are soldered and brought to binding posts on a 
 
224 ELECTRICITY AND MAGNETISM. 
 
 board. Below the ends of the board are vessels contain- 
 ing sand or oil in which two test-tubes are supported. The 
 junctions are placed in these test-tubes as indicated in 
 figure 75. The binding posts are connected, by copper 
 wires, to a key of as many parts as there are wires to be 
 tested, so that each circuit may be completed through a 
 sensitive galvanometer. Thermometers are placed in the 
 test-tubes to note the temperatures as one vessel is being 
 heated by a burner. 
 
 It is especially important that the temperature should 
 be ascertained accurately. Hence heat should be applied 
 
 Cu. 
 
 Cu. 
 
 FIG. 75. 
 
 cautiously, especially at first, and, when observations are 
 to be made, the source of heat should be removed, and time 
 should be allowed for the temperatures to become fairly 
 constant. 
 
 The galvanometer reading should be noted with the great- 
 est care and the zero should be frequently tested. After 
 each reading of the galvanometer, the temperature should be 
 noted. 
 
 If a high resistance galvanometer of sufficient sensitive- 
 ness is available, the other resistances may be neglected 
 and the various e. m. f.'s will then be proportional to the 
 deflections. Or a sensitive low resistance galvanometer, with 
 a constant high resistance permanently in series, may be used 
 with similar simplicity. The constant of the galvanometer, 
 considered as a voltmeter, may be found by applying to it 
 a fraction of the e. m. f. of a standard cell (pp. 161, 179). 
 
 With this arrangement (which will be readily under- 
 stood from the figure) the thermo-electric force of each 
 circuit, consisting of copper and another* wire, may be de- 
 
ELEMENTARY STUDY OF ALTERNATING CURRENTS. 225 
 
 termined. Curves representing the results should be plot- 
 ted with the differences of temperature of the junctions as 
 abscissae and the e. m. fs as ordinates. 
 
 If a low resistance galvanometer of low sensitiveness is 
 used, it will be necessary to consider it as an ammeter. 
 In this case, the resistances of the various circuits and of 
 the galvanometer must be found and the constant of the 
 galvanometer, in amperes per unit deflection, must be ob- 
 tained by connecting it in series with a standard cell and a 
 sufficient known resistance. Thus, the currents and the 
 resistances being known, the thermo-electric forces can be 
 calculated. 
 
 Questions. 
 
 1. State what would be observed if the temperature of the hot 
 junction were increased steadily beyond the highest temperature 
 used in this experiment. 
 
 2. Is the effect observed here due solely to differences of potential 
 produced at the contacts ? 
 
 LXIX. ELEMENTARY STUDY OF RESISTANCE, SELF- 
 INDUCTION AND CAPACITY. 
 
 Text-book of Physics (Duff}, pp. 623624; Watson's Physics, p. 758; 
 Hadley's Electricity and Magnetism, pp. 441450. 
 
 In the following exercises, which are intended for students 
 who have not made a study of the theory of alternating 
 currents, some of the properties of such currents are studied 
 and compared with those of direct currents. 
 
 Ohm's Law for steady currents states that 
 
 E 
 
 =a constant, R, 
 
 i 
 
 where R is called the resistance of the conductor. 
 
 (A) Apply various e. m. f.'s to anon-inductive conductor. 
 Measure the current by an ammeter, and the voltage by a 
 voltmeter (of any type) and calculate R for each value of 
 the e. m. f. The latter may be varied by means of a series 
 rheostat. 
 
226 ELECTRICITY AND MAGNETISM. 
 
 (B) Apply the same method to (i) a large coil, (2) the 
 large coil and the non-inductive resistance, (a) in series, and, 
 (b) in parallel. Compare the results of (a) and (b) with the 
 calculated values. 
 
 (C) Repeat (A) and (B), using alternating currents and 
 an electrostatic voltmeter. Corresponding to Ohm's Law we 
 have 
 
 E / 
 
 - =constant = V R 
 i 
 
 where the constant is called the impedance and L is the 
 coefficient of self-induction and n the frequency. Find the 
 value of the constant for different e. m. f.'s and currents, 
 and, from the mean and the values of R and n, calculate L. 
 
 Contrast the results in series and 
 parallel combinations with the 
 N values calculated by treating im- 
 
 - 1 pedance in the same way as re- 
 
 Jtheo I i . ., . 
 
 A I sistance in direct currents. 
 
 Tabulate all results so that they 
 may be readily compared. 
 
 (D) When an alternating cur- 
 F rent is applied to a condenser, it 
 
 is charged, discharged, charged 
 
 oppositely and discharged during each alternation. Evi- 
 dently the total quantity that traverses the leads in each 
 unit of time is proportional to the frequency and to the 
 product of the capacity and the voltage (since q = CV) and 
 it can be shown that the current is given by 
 
 i = 2nnCV. 
 
 Measure i for various values of V and calculate C. Do 
 this for several condensers (i) separately, (2) in parallel, 
 (3) in series. Compare the results of (2) and (3) with the 
 calculated values. 
 
 Questions. 
 
 i. What is meant by the effective value of an alternating current 
 and what ratio does it bear to the maximum value? 
 
INDUCTION AND CAPACITY, ALTERNATING CURRENTS. 227 
 
 2. How, by means of a diagram, would you find the impedance 
 when given the ohmic resistance R and the inductance L ? 
 
 3. Supposing the alternating e. m. f. resolved graphically into 
 two parts, one to overcome the ohmic resistance and the other to 
 overcome the inductance, what relation between the phases of these 
 two parts does question (2) suggest? 
 
 LXX. SELF-INDUCTION, MUTUAL INDUCTION AND 
 CAPACITY, ALTERNATING CURRENTS. 
 
 See references to LXIX. J . J. Thomson's Electricity and Magnetism, 
 233, 244-245; Jackson's Alt. Cur., pp. 90-91, 151-200; Parr's 
 Electrical Eng. Testing, pp. 222224, 228231, 234235. For 
 Electrostatic Voltmeter, see Parr, 367-371. 
 
 This exercise, which is somewhat more advanced than the 
 preceding, is intended, for students who have made some 
 study of the theory of alternating currents. 
 
 Let E be the alternating e. m. f. in a circuit of resistance 
 R, capacity C, and self-induction L. If i is the current, 
 . E 
 
 *>- i /Co) 2 
 
 We can test the above formula by calculation, after 
 measuring i, E, R, C, and L. An inductance with a magnetic 
 core has a variable value of L, the magnitude of which 
 depends on the strength of the current. Hence, for this 
 experiment, an inductance consisting of a very large coil 
 containing no iron is used. 
 
 (A) Measurement of C. If, in the general formula, L 
 be zero, C can be deduced from the values of i, E, and R, 
 assuming that <o, which equals 2 it times the frequency n, 
 is known. E, the e. m. f. across the terminals of the con- 
 denser, is measured by an electrostatic voltmeter, i by an 
 alternating current ammeter. Initially a high resistance 
 of large current capacity must be included. This may later 
 be cut out. A fuse of lower capacity than the range of the 
 ammeter must be permanently in circuit. 
 
 (B) Measurement of L. The value of L is found by 
 observing the values of i and in a circuit containing the 
 self-inductance coil and then applying the general formula. 
 
228 ELECTRICITY AND MAGNETISM. 
 
 Sufficient additional resistance must be placed in the cir- 
 cuit, but the value of E required is that across the terminals 
 of the inductance coil. R, which in this case is the resistance 
 of the coil, is best found by Wheatstone's Bridge. 
 
 (C) Test of General Formula. Connect the condenser 
 and self-induction in series. Measure the current and the 
 total e. m. f. ; also the e. m. f. across each part. Connect the 
 condenser and inductance coil in parallel. Measure the 
 common e. m. f., the total current and the current in each 
 branch. 
 
 Calculate i in the series arrangement from the above 
 formula and compare with the experimental value. If you 
 are familiar with the method of complex quantities and 
 graphical methods, apply these also to calculate the 
 currents in both series and parallel arrangements. 
 
 (D) Measurement of Mutual Inductance. Measure the 
 mutual inductance, M, of the two coils of a transformer 
 (with iron core) by observing the e. m. f., E, across one coil 
 when a measured current, i, is applied to the other. 
 
 E=Mi<o. 
 Vary i several times and find how M varies. 
 
 Questions. 
 
 1 . Why is an electrostatic voltmeter necessary ? 
 
 2. Does the self-induction depend upon the frequency? Why 
 does the latter enter into the equation ? 
 
 LXXI. DIELECTRIC CONSTANTS OF LIQUIDS. 
 
 Text-book of Physics (Duff), pp. 533, 535; Ames' General Physics, pp. 
 641, 66 1 ; Watson's Physics, p. 637; Hartley's Electricity and 
 Magnetism, Chapter X. 
 
 The dielectric constant of a liquid, or the ratio of the 
 capacity of a condenser with that liquid as dielectric to its 
 capacity when its dielectric is air, can be determined by 
 a comparison of capacities by the Bridge Method of Exp. 
 LX. For this purpose it is convenient to use a condenser 
 consisting of two parallel plates, the distance between which 
 
DIELECTRIC CONSTANTS OF LIQUIDS. 
 
 229 
 
 is adjustable, as shown in figure 7 7 . The distance between 
 the plates can be measured by means of a scale, B, attached 
 to the movable plate A, and a vernier attached to the 
 framework. The plates hang in a vessel for holding the 
 dielectric. Two methods can be used. In one the distance 
 between the plates is not varied; in the other it is varied. 
 
 A A' 
 
 B' 
 
 p' 
 
 FIG. 77. 
 
 The first method consists in comparing the capacity of the 
 above condenser with that of a Ley den jar (i) when the 
 dielectric is the liquid to be tested; (2) when it is air. From 
 these comparisons the ratio of the capacities of the condenser 
 in the two conditions is deduced and this equals the dielectric 
 constant of the liquid. Instead of a battery and galvan- 
 ometer, an induction coil and a sensitive telephone are used. 
 
 The second method assumes the (approximate) formula 
 
 Ae 
 
 C 
 
 for the capacity of such a plate condenser (in electrostatic 
 units, see p. 162), where A is the area of each plate, e is the 
 dielectric constant of the surrounding medium, and d is 
 
230 ELECTRICITY AND MAGNETISM. 
 
 their distance apart. If C be the same with two dielectrics, 
 but with different values of d 
 
 e 2 d 2 
 
 Having obtained a balance for air as dielectric, leave 
 the resistances in the bridge unchanged and again obtain 
 a balance, after filling the jar with the liquid to be tested, 
 by adjusting the distance between the plates. This second 
 method is less accurate, since the formula assumed is only 
 approximate and the distances cannot be determined as 
 accurately as the resistances. An accurate formula will be 
 found in Kohlrausch's Physical Measurements, p. 379. 
 
 The above methods should be applied to two highly 
 insulating liquids, such as kerosene and benzol. 
 
 Questions. 
 
 1. Calculate the capacity of the two plates when separated by 
 (a) air; (b) liquid, the distance apart being the same as in the first 
 part of the experiment. 
 
 2. Calculate the charge for each case if (a) 100 electrostatic units 
 of potential are applied; (b) 100 volts (p. 162). 
 
 LXIII. ELECTRIC WAVES ON WIRES. 
 
 Dielectric Constants of Liquids. 
 
 Ames' General Physics, pp. 752754; Watson's Physics, pp. 856-858, 
 870871; Text-book of Physics (Duff}, pp. 635-639; Hadley's 
 Magnetism and Electricity, pp. 541-548, 585-586; J. J. Thom- 
 son's Electricity and Magnetism, 243; Kohlrausch's Physical 
 Measurements (on capacity of a plate condenser), p. 379. Drude, 
 An. der Phys., Vol. 8, p. 336. 
 
 In this experiment electric waves on a wire, AD, are 
 excited by electric oscillations in a neighboring circuit or 
 "exciter," E, which contains an inductance, L, and a 
 capacity, C. The period of such oscillations is T=2x\/LC. 
 The inductance is that of two thick semicircular wires. 
 The ends e and e' of these wires carry small spheres and 
 the wires are so bent that the spheres are beneath the 
 
ELECTRIC WAVES ON WIRES. 
 
 231 
 
 surface of kerosene in a small cup and form a spark- 
 The length of this spark-gap can 
 be adjusted by means of a mi- 
 crometer screw attached to one 
 of the ebonite posts, H H', on 
 which the semicircular wires are 
 supported. The condenser is of 
 the variable form shown in figure 
 77 and is connected between the 
 other two ends of the semicircular 
 wires. The impulses that start 
 the oscillations in the exciter are 
 produced by a Tesla coil, the 
 secondary of which is connected 
 across the spark-gap ee' t while the 
 primary of the Tesla coil is con- 
 nected through another spark-gap, 
 Z, to the secondary of an induc- 
 tion coil /. 
 
 The wire, AD, on which the 
 waves are formed is bent to a 
 U-form and lies in a horizontal 
 plane above the plane of the 
 exciter. The oscillations in the 
 exciter act by induction on the 
 part, A, of the wire and produce 
 waves that move along the wire 
 toward D; between any two cor- 
 responding points, such as d and 
 d', there is an oscillating differ- 
 ence of potential and the trans- 
 mission of this oscillation consti- 
 tutes the wave-motion. At the 
 free end, D, these waves are 
 reflected and interfere with the 
 direct waves. If the wire is of 
 proper length, this interference 
 
 gap. 
 
 B 
 
 B 
 
 D 
 
 FIG. 78. 
 
232 ELECTRICITY AND MAGNETISM. 
 
 produces stationary waves; that is, the wire "resonates" 
 to the exciter. If the wire be also sufficiently long, there 
 will be one or more nodes on the wire, that is, places of no 
 potential difference, with intervening antinodes, or places of 
 maximum oscillating potential difference. If a small wire 
 "bridge," B, be placed across the wire at a node it will not 
 interfere with the stationary waves ; but it will destroy them 
 if it is placed at any other point. (It will be instructive to 
 compare the above with the formation of stationary sound 
 waves in a resonance tube such as that of Exp. XXX, the 
 tuning-fork being the exciter.) When the bridge has been 
 placed at a node the part of the wire between it and the free 
 end, D, could be changed in length or removed without 
 appreciably diminishing the oscillations between A and B, 
 (just as the lower part of a violin string that is touched by 
 the finger does not interfere with the vibrations of the upper 
 part). The part A of the wire is (approximately) a node, 
 although it is the part where the oscillations are excited. 
 (Compare with this the fact that when a tuning-fork, 
 connected to a long thread as in Melde's experiment, throws 
 the latter into stationary vibrations, the point of connection 
 is a node). 
 
 Various means have been used for detecting such station- 
 ary vibrations. The simplest is a "vacuum" tube, V, 
 which contains helium at a very low pressure. An alter- 
 nating potential difference between the ends of such a tube 
 will cause oscillating discharges accompanied by a glow. 
 If placed across the wire when there are no stationary 
 oscillations the tube will not glow; but, if stationary oscil- 
 lations exist, it will glow brightly at an antinode, less 
 brightly between an antinode and a node, not at all at a node. 
 By the aid of the tube the bridge can be adjusted to each 
 node. If the wire be long enough to permit of more than 
 one node, twice the distance between two adjacent nodes 
 will equal the wave length. 
 
 Since the wire is in resonance with the exciter, the period 
 of oscillation of the wire equals that of the exciter. The 
 
ELECTRIC WAVES ON WIRES. 233 
 
 latter, and therefore the former, can be changed by changing 
 C. Hence 
 
 The electric waves, while directed by the wire, are really 
 waves of oscillation of electric force in the medium between 
 the two branches of the wire. Such waves travel with a 
 velocity that is independent of the wave length, and, if 
 A! be the wave-length when the period is T lt ^ that when 
 the period is T 2 , v = ). l /T l =A 2 /T 2 . Hence 
 
 By determining the wave-length with air as the dielectric 
 in C and then with a liquid as dielectric, we can evidently 
 find the dielectric constant of the liquid. 
 
 In the practice of the method most trouble is likely to be 
 due to the spark-gap, ee'. It must be adjusted until the 
 spark occurs under the kerosene. To avoid danger to the 
 tube by accidental dropping, it may be attached loosely to 
 the wire by a loop of thread. Each node should be deter- 
 mined several times. The distance between the plates of C 
 should be varied four or five times with air as dielectric 
 and a curve drawn with d as abscissa and X as ordinate. 
 From this curve and the value of X for each liquid, the value 
 of the dielectric constant for that liquid is readily deduced. 
 
 For a complete proof of the formula works on electricity and 
 magnetism must be consulted (e. g., J. J. Thomson's Electricity and 
 Magnetism, 243). The following considerations suggest the 
 formula. When an alternating e. m. f., E, is applied to a circuit 
 containing capacity, self-inductance, and ohmic resistance in series, 
 
 The potentials across the self-inductance (including the resistance) 
 and the capacity, respectively, are 
 
 L = /-\/r 2 + a> 2 L 2 and E c = ~^ 
 
234 ELECTRICITY AND MAGNETISM. 
 
 Evidently the potentials across the parts of the system may be greater 
 than the total e. m. f. which is the resultant obtained by geometrical 
 (i. e.,' vector) addition of the parts. This constitutes resonance. 
 It is complete when 
 
 Substituting for w its value 2x/T, we get T = 2n'\/LC. This, then, 
 is the period of the applied e. m. f. when resonance results. It is 
 also, therefore, the period of the free natural vibrations of the system. 
 The approximate formula for the capacity of a plate condenser is 
 given in Exp. LXII, a more exact one in Kohlrausch, p. 379. 
 
 Questions. 
 
 1. Assuming the velocity of the waves to be that of light, calculate 
 the frequency of the oscillations for one value of ^. 
 
 2. From this and the approximate formula for C calculate L. 
 
 3. If a sufficient amount of liquid were available, how could its 
 dielectric constant be found by immersing the wire AD in it? 
 
 4. What would be observed if AD were contained in a vacuum 
 tube? 
 
TABLES 
 
236 
 
 TABLES. 
 
 TABLE I. 
 Logarithms of Numbers from i to 1000. 
 
 No. 
 
 o 
 
 i 
 
 2 
 
 3 4 
 
 567 
 
 8 
 
 9 
 
 10 
 
 oooo 
 
 0043 
 
 0086 
 
 0128 
 
 0170 
 
 0212 
 
 0253 
 
 0294 
 
 0334 
 
 0374 
 
 ii 
 
 0414 
 
 4 C 3 
 
 0492 
 
 Q53 1 
 
 0569 
 
 0607 
 
 0645 
 
 0682 
 
 0719 
 
 0755 
 
 12 
 
 0792 
 
 0828 
 
 0864 
 
 0899 
 
 0934 
 
 0969 
 
 1004 
 
 1038 
 
 1072 
 
 1 106 
 
 13 
 
 "39 
 
 "73 
 
 I2O6 
 
 1239 
 
 1271 
 
 J 33 
 
 J 335 
 
 1367 
 
 *399 
 
 143 
 
 14 
 
 1461 
 
 1492 
 
 1523 
 
 *553 
 
 1584 
 
 1614 
 
 1644 
 
 1673 
 
 !73 
 
 1732 
 
 IS 
 
 1761 
 
 1790 
 
 1818 
 
 1847 
 
 .1875 
 
 1903 
 
 I93 1 
 
 1959 
 
 1987 
 
 2014 
 
 16 
 
 2041 
 
 2068 
 
 2095 
 
 2122 
 
 2148 
 
 2175 
 
 2201 
 
 2227 
 
 2253 
 
 2279 
 
 17 
 
 2304 
 
 233 
 
 2355 
 
 2380 
 
 2405 
 
 2430 
 
 2455 
 
 2480 
 
 2504 
 
 2529 
 
 18 
 
 2553 
 
 2577 
 
 2601 
 
 262 5 
 
 2648 
 
 2672 
 
 2695 
 
 2718 
 
 2742 
 
 2765 
 
 19 
 
 2788 
 
 2810 
 
 2833 
 
 2856 
 
 2878 
 
 2900 
 
 2923 
 
 2945 
 
 2967 
 
 2989 
 
 20 
 
 3010 
 
 3032 
 
 354 
 
 3075 
 
 3096 
 
 3118 
 
 3139 
 
 3160 
 
 3181 
 
 3201 
 
 21 
 
 3222 
 
 3243 
 
 3263 
 
 3284 
 
 3304 
 
 3324 
 
 3345 
 
 3365 
 
 3385 
 
 3404 
 
 22 
 
 3424 
 
 3444 
 
 3464 
 
 3483 
 
 3502 
 
 3522 
 
 3541 
 
 356o 
 
 3579 
 
 3598 
 
 23 
 
 3617 
 
 3636 
 
 3 6 55 
 
 3674 
 
 3692 
 
 37" 
 
 3729 
 
 3747 
 
 3766 
 
 3784 
 
 24 
 
 3802 
 
 3820 
 
 3838 
 
 3856 
 
 3874 
 
 3892 
 
 3909 
 
 3927 
 
 3945 
 
 3962 
 
 25 
 
 3979 
 
 3997 
 
 4014 
 
 4031 
 
 4048 
 
 4065 
 
 4082 
 
 4099 
 
 4116 
 
 4133 
 
 26 
 
 4150 
 
 4166 
 
 4183 
 
 4200 
 
 4216 
 
 4232 
 
 4249 
 
 4265 
 
 4281 
 
 4298 
 
 27 
 
 43*4 
 
 4330 
 
 4346 
 
 4362 
 
 4378 
 
 4393 
 
 4409 
 
 4425 
 
 4440 
 
 44 5 6 
 
 28 
 
 4472 
 
 4487 
 
 4502 
 
 4518 
 
 4533 
 
 4548 
 
 4564 
 
 4579 
 
 4594 
 
 4609 
 
 29 
 
 4624 
 
 4639 
 
 4654 
 
 4669 
 
 4683 
 
 4698 
 
 47 J 3 
 
 4728 
 
 4742 
 
 4757 
 
 30 
 
 477 1 
 
 4786 
 
 4800 
 
 4814 
 
 4829 
 
 4843 
 
 4857 
 
 4871 
 
 4886 
 
 4900 
 
 31 
 
 4914 
 
 4928 
 
 4942 
 
 4955 
 
 4969 
 
 4983 
 
 4997 
 
 5011 
 
 5024 
 
 5038 
 
 32 
 
 505 1 
 
 5065 
 
 5079 
 
 5092 
 
 5io5 
 
 5"9 
 
 5132 
 
 5M5 
 
 5159 
 
 5172 
 
 33 
 
 5i85 
 
 5198 
 
 5211 
 
 5224 
 
 5237 
 
 5250 
 
 5263 
 
 5276 
 
 5289 
 
 5302 
 
 34 
 
 5315 
 
 5328 
 
 5340 
 
 5353 
 
 5366 
 
 5378 
 
 539i 
 
 5403 
 
 54i6 
 
 5428 
 
 35 
 
 544i 
 
 5453 
 
 5465 
 
 5478 
 
 549 
 
 5502 
 
 55*5 
 
 5527 
 
 5539 
 
 555 1 
 
 36 
 
 55 6 3 
 
 5575 
 
 5587 
 
 5599 
 
 5611 
 
 5623 
 
 5635 
 
 5 6 47 
 
 5658 
 
 5670 
 
 37 
 
 5682 
 
 5694 
 
 575 
 
 5717 
 
 5729 
 
 574 
 
 5752 
 
 5763 
 
 5775 
 
 5786 
 
 38 
 
 5798 
 
 5809 
 
 5821 
 
 5832 
 
 5843 
 
 5855 
 
 5866 
 
 5877 
 
 5888 
 
 5899 
 
 39 
 
 59 11 
 
 5922 
 
 5933 
 
 5944 
 
 5955 
 
 5966 
 
 5977 
 
 5988 
 
 5999 
 
 6010 
 
 40 
 
 6021 
 
 6031 
 
 6042 
 
 6053 
 
 6064 
 
 6075 
 
 6085 
 
 6096 
 
 6107 
 
 6117 
 
 4i 
 
 6128 
 
 6138 
 
 6149 
 
 6160 
 
 6170 
 
 6180 
 
 6191 
 
 6201 
 
 6212 
 
 6222 
 
 42 
 
 6232 
 
 6243 
 
 6253 
 
 6263 
 
 6274 
 
 6284 
 
 6294 
 
 6304 
 
 6314 
 
 6325 
 
 43 
 
 6335 
 
 6345 
 
 6355 
 
 6365 
 
 6375 
 
 6385 
 
 6395 
 
 6405 
 
 6415 
 
 6425 
 
 44 
 
 6435 
 
 6444 
 
 6454 
 
 6464 
 
 6474 
 
 6484 
 
 6493 
 
 6503 
 
 6513 
 
 6522 
 
 45 
 
 6532 
 
 6542 
 
 6551 
 
 6561 
 
 657i 
 
 6580 
 
 6590 
 
 6599 
 
 6609 
 
 6618 
 
 46 
 
 6628 
 
 6637 
 
 6646 
 
 6656 
 
 6665 
 
 6675 
 
 6684 
 
 6693 
 
 6702 
 
 6712 
 
 47 
 
 6721 
 
 6730 
 
 6 739 
 
 6749 
 
 6758 
 
 6767 
 
 6776 
 
 6785 
 
 6794 
 
 6803 
 
 48 
 
 6812 
 
 6821 
 
 6830 
 
 6839 
 
 6848 
 
 6857 
 
 6866 
 
 6875 
 
 6884 
 
 6893 
 
 49 
 
 6902 
 
 691 1 
 
 6920 
 
 6928 
 
 6937 
 
 6946 
 
 6955 
 
 6964 
 
 6972 
 
 6981 
 
 50 
 
 6990 
 
 6998 
 
 7007 
 
 7016 
 
 7024 
 
 733 
 
 7042 
 
 7050 
 
 759 
 
 7067 
 
 51 
 
 7076 
 
 7084 
 
 793 
 
 7101 
 
 7110 
 
 7118 7126 
 
 7135 
 
 7 J 43 
 
 7152 
 
 52 
 
 7160 
 
 7168 
 
 7177 
 
 7185 
 
 7 J 93 
 
 7202 
 
 7210 
 
 7218 
 
 7226 
 
 7 2 35 
 
 53 
 
 7 2 43 
 
 7251 
 
 7259 
 
 7267 
 
 7275 
 
 7284 
 
 7292 
 
 7300 
 
 7308 
 
 73 J 6 
 
 54 
 
 7324 
 
 7332 
 
 7340 
 
 7348 
 
 7356 
 
 7364 ', 7372 7380 
 
 7388 
 
 7396 
 
 No. 
 
 I 5 
 
TABLES. 
 
 237 
 
 TABLE I. Continued. 
 Logarithms of Numbers from i to 1000. 
 
 No. 
 
 
 
 1 1 * \ 3 
 
 4 
 
 5 | 6 
 
 7 
 
 8 
 
 9 
 
 55 
 
 7404 
 
 7412 
 
 7419 
 
 7427 
 
 7435 
 
 7443 
 
 745i 
 
 7459 
 
 7466 7474 
 
 56 
 
 7482 
 
 7490 
 
 7497 
 
 7505 
 
 75*3 
 
 7520 
 
 7528 
 
 7536 
 
 7543 755 1 
 
 57 
 
 7559 
 
 7566 
 
 7574 
 
 7582 
 
 7589 
 
 7597 
 
 7604 
 
 7612 
 
 7619 7627 
 
 58 
 
 7^34 
 
 7642 
 
 7649 
 
 7657 
 
 7664 
 
 7672 
 
 7679 
 
 7686 
 
 7694 7701 
 
 59 
 
 7709 
 
 7716 
 
 7723 
 
 7731 
 
 7738 
 
 7745 
 
 7752 
 
 7760 
 
 7767 7774 
 
 60 
 
 7782 
 
 7789 
 
 7796 
 
 7803 
 
 7810 
 
 7818 
 
 7825 
 
 7832 
 
 7839 ! 7846 
 
 61 
 
 7853 
 
 7860 
 
 7868 
 
 7875 
 
 7882 
 
 7889 7896 
 
 7903 
 
 7910 ! 7917 
 
 62 
 
 7924 
 
 793 1 
 
 7938 
 
 7945 
 
 7952 
 
 7959 7966 
 
 7973 
 
 7980 7987 
 
 63 
 
 7993 
 
 8000 
 
 8007 
 
 8014 
 
 8021 
 
 8028 
 
 8035 
 
 8041 
 
 8048 8055 
 
 64 
 
 8062 
 
 8069 
 
 8075 
 
 8082 
 
 8089 
 
 8096 
 
 8102 
 
 8109 
 
 8116 8122 
 
 65 8129 
 
 8136 
 
 8142 
 
 8149 
 
 8156 
 
 8162 
 
 8169 
 
 8176 
 
 8182 8189 
 
 66 8195 8202 
 
 8209 
 
 8215 
 
 8222 
 
 8228 8235 8241 
 
 8248 8254 
 
 67 8261 
 
 8267 
 
 8274 
 
 8280 
 
 8287 
 
 8293 8299 
 
 8306 
 
 8312 8319 
 
 68 8325 
 
 8331 
 
 8338 
 
 8344 
 
 8351 
 
 8357 
 
 8363 
 
 8370 
 
 8376 8382 
 
 69 
 
 8388 
 
 8395 
 
 8401 
 
 8407 
 
 8414 
 
 8420 
 
 8426 
 
 8432 
 
 8439 | 8445 
 
 70 
 
 8451 
 
 8457 
 
 8463 
 
 8470 
 
 8476 
 
 8482 
 
 8488 
 
 8494 
 
 8500 ! 8506 
 
 7i 8513 
 
 8519 
 
 8525 
 
 8531 
 
 8537 
 
 8543 
 
 8549 
 
 8555 
 
 8561 ; 8567 
 
 72 8573 
 
 8579 
 
 8585 
 
 859i 
 
 8597 
 
 8603 
 
 8609 
 
 8615 
 
 8621 8627 
 
 73 1 8633 
 
 8639 
 
 8645 
 
 8651 
 
 8657 
 
 8663 
 
 8669 
 
 8675 
 
 8681 ; 8686 
 
 74 8692 
 
 8698 
 
 8704 
 
 8710 
 
 8716 
 
 8722 
 
 8727 
 
 8733 
 
 8739 8745 
 
 75 
 
 875i 
 
 8756 
 
 8762 
 
 8768 
 
 8774 
 
 8779 
 
 8785 
 
 8791 
 
 8797 
 
 8802 
 
 76 j 8808 
 
 8814 
 
 8820 
 
 8825 
 
 8831 
 
 8837 
 
 8842 
 
 8848 
 
 8854 
 
 8859 
 
 77 
 
 8865 
 
 8871 
 
 8876 
 
 8882 
 
 8887 
 
 8893 
 
 8899 
 
 8904 
 
 8910 
 
 8915 
 
 78 
 
 8921 
 
 8927 
 
 8932 
 
 8938 
 
 8943 
 
 8949 
 
 8954 
 
 8960 
 
 8965 
 
 8971 
 
 79 
 
 8976 
 
 8982 
 
 8987 
 
 8993 
 
 8998 
 
 9004 
 
 9009 
 
 9oi5 
 
 9020 
 
 9025 
 
 80 
 
 9031 
 
 9036 
 
 9042 
 
 9047 
 
 9053 
 
 9058 
 
 9063 9069 
 
 9074 
 
 9079 
 
 81 
 
 9085 
 
 9090 
 
 9096 
 
 9101 9106 
 
 9112 
 
 9117 i 9122 
 
 9128 
 
 9*33 
 
 82 
 
 9138 
 
 9M3 
 
 9149 
 
 9154 9159 
 
 9165 
 
 9170 9175 
 
 9180 
 
 9186 
 
 f 3 
 
 9191 
 
 9196 
 
 9201 
 
 9206 9212 
 
 9217 9222 i 9227 
 
 9232 
 
 9238 
 
 84 
 
 9243 
 
 9248 
 
 9253 
 
 9258 
 
 9263 
 
 9269 
 
 9274 
 
 9279 
 
 9284 
 
 9289 
 
 85 
 
 9294 
 
 9299 
 
 934 
 
 9309 
 
 9315 
 
 9320 
 
 9325 
 
 9330 
 
 9335 
 
 9340 
 
 86 
 
 9345 
 
 935 
 
 9355 
 
 9360 
 
 93 6 5 
 
 9370 
 
 9375 I 938o 
 
 9385 
 
 9390 
 
 87 
 
 9395 
 
 9400 
 
 9405 
 
 9410 
 
 9415 
 
 9420 
 
 9425 i 943 9435 
 
 9440 
 
 88 
 
 9445 
 
 9450 
 
 9455 
 
 9460 
 
 9465 
 
 9469 i 9474 
 
 9479 9484 
 
 9489 
 
 89 
 
 9494 
 
 9499 
 
 9504 
 
 9509 
 
 95 J 3 
 
 9518 ! 9523 
 
 9528 9533 
 
 9538 
 
 90 
 
 9542 
 
 9547 
 
 9552 
 
 9557 
 
 9562 
 
 9566 
 
 957i 
 
 9576 9581 
 
 9586 
 
 9i 
 
 959 
 
 9595 
 
 9600 
 
 9605 i 9609 
 
 9614 
 
 9619 
 
 9624 9628 
 
 9633 
 
 92 
 
 9638 
 
 9643 ! 9647 
 
 9652 9657 
 
 9661 9666 
 
 9671 1 9675 
 
 9680 
 
 93 
 
 9685 
 
 9689 i 9694 
 
 9699 9703 
 
 9708 ! 9713 
 
 9717 
 
 9722 9727 
 
 94 
 
 973i 
 
 9736 
 
 974i 
 
 9745 
 
 975 
 
 9754 
 
 9759 
 
 9763 
 
 9768 
 
 9773 
 
 95 
 
 9777 
 
 9782 
 
 9786 
 
 9791 
 
 9795 
 
 9800 
 
 9805 
 
 9809 
 
 9814 
 
 9818 
 
 96 
 
 9823 
 
 9827 
 
 9832 
 
 9836 
 
 9841 
 
 9845 
 
 9850 
 
 9854 
 
 9859 
 
 9863 
 
 97 
 
 9868 
 
 9872 
 
 9877 
 
 9881 
 
 9886 
 
 9890 
 
 9894 
 
 9899 
 
 9903 
 
 9908 
 
 98 
 
 9912 
 
 9917 
 
 9921 
 
 9926 
 
 9930 
 
 9934 
 
 9939 
 
 9943 
 
 9948 
 
 9952 
 
 99 
 
 9956 
 
 9961 
 
 9965 
 
 9969 
 
 9974 
 
 997 s 
 
 9983 
 
 9987 
 
 9991 
 
 9996 
 
 No. 
 
 o 
 
 z 
 
 2 
 
 3 
 
 4 I 5 
 
 6 
 
 7 
 
 8 
 
 9 
 
238 
 
 TABLES. 
 
 TABLE II. 
 Natural Sines and Cosines. 
 
 Sine 
 
 Cosine 
 
 4, 
 
 o.oooo 
 
 
 90 
 
 I.OOOO 
 
 
 I 
 
 0.0175 
 
 175 
 
 89 
 
 0.9998 
 
 02 
 
 2 
 
 0.0349 
 
 1 74 
 
 88 
 
 0.9994 
 
 04 
 
 3 
 
 0.0523 
 
 174 
 
 87 
 
 0.9986 
 
 08 
 
 4 
 
 0.0698 
 
 175 
 
 86 
 
 0.9976 
 
 10 
 
 5 
 
 0.0872 
 
 174 
 
 85 
 
 0.9962 
 
 14 
 
 6 
 
 0.1045 
 
 
 84 
 
 0.9945 
 
 
 7 
 
 0.1219 
 
 174 
 
 83 
 
 0.9925 
 
 20 
 
 8 
 
 0.1392 
 
 173 
 
 82 
 
 0.9903 
 
 22 
 
 9 
 
 0.1564 
 
 172 
 
 81 
 
 0.9877 
 
 26 
 
 10 
 
 0.1736 
 
 172 
 
 80 
 
 0.9848 
 
 29 
 
 ii 
 
 0.1908 
 
 172 
 
 79 
 
 0.9816 
 
 32 
 
 12 
 
 0.2079 
 
 171 
 
 78 
 
 0.9781 
 
 35 
 
 13 
 
 0.2250 
 
 171 
 
 77 
 
 0.9744 
 
 37 
 
 14 
 
 0.2419 
 
 169 
 
 76 
 
 0.9703 
 
 
 15 
 
 0.2588 
 
 169 
 
 75 
 
 0.9659 
 
 44 
 
 16 
 
 0.2756 
 
 1 68 
 
 74 
 
 0.9613 
 
 46 
 
 J 7 
 
 0.2924 
 
 168 
 
 73 
 
 0.9563 
 
 50 
 
 18 
 
 0.3090 
 
 166 
 
 7 2 
 
 0.9511 
 
 52 
 
 19 
 
 0.3256 
 
 166 
 
 7 1 
 
 0.9455 
 
 56 
 
 20 
 
 0.3420 
 
 164 
 
 70 
 
 0-9397 
 
 58 
 
 21 
 
 0.3584 
 
 164 
 
 69 
 
 0.9336 
 
 61 
 
 22 
 
 0.3746 
 
 162 
 
 68 
 
 0.9272 
 
 64 
 
 2 3 
 
 0.3907 
 
 161 
 
 67 
 
 0.9205 
 
 67 
 
 24 
 
 0.4067 
 
 1 60 
 
 66 
 
 -9 I 35 
 
 70 
 
 25 
 
 0.4226 
 
 159 
 
 65 
 
 0.9063 
 
 72 
 
 26 
 
 0.4384 
 
 158 
 
 64 
 
 0.8988 
 
 75 
 
 27 
 
 0.4540 
 
 156 
 
 63 
 
 0.8910 
 
 78 
 
 28 
 
 0.4695 
 
 155 
 
 62 
 
 0.8829 
 
 81 
 
 29 
 
 0.4848 
 
 153 
 
 61 
 
 0.8746 
 
 83 
 
 30 
 
 0.5000 
 
 I 5 2 
 
 60 
 
 0.8660 
 
 86 
 
 3 1 
 
 0.5150 
 
 I 5 
 
 59 
 
 0.8572 
 
 88 
 
 32 
 
 0.5299 
 
 149 
 
 58 
 
 0.8480 
 
 92 
 
 33 
 
 0.5446 
 
 147 
 
 57 
 
 0.8387 
 
 93 
 
 34 
 
 0.5592 
 
 146 
 
 56 
 
 0.8290 
 
 97 
 
 35 
 
 0-5736 
 
 144 
 
 55 
 
 0.8192 
 
 98 
 
 36 
 
 0.5878 
 
 142 
 
 54 
 
 0.8090 
 
 102 
 
 37 
 
 0.6018 
 
 140 
 
 53 
 
 0.7986 
 
 104 
 
 38 
 
 0.6157 
 
 139 
 
 52 
 
 0.7880 
 
 106 
 
 39 
 
 0.6293 
 
 136 
 
 
 0.7771 
 
 109 
 
 40 
 
 0.6428 
 
 X 35 
 
 5 
 
 0.7660 
 
 in 
 
 41 
 
 9-6561 
 
 J 33 
 
 49 
 
 0-7547 
 
 XI 3 
 
 42 
 
 0.6691 
 
 130 
 
 48 
 
 Q-743 1 
 
 116 
 
 43 
 
 0.6820 
 
 129 
 
 47 
 
 0.7314 
 
 117 
 
 44 
 
 0.6947 
 
 127 
 
 46 
 
 0-7*93 
 
 121 
 
 45 
 
 0.7071 
 
 124 
 
 45 * 
 
 0.7071 
 
 122 
 
 Ctfsine 
 
 Sine 
 
TABLES. 
 
 239 
 
 TABLE III. 
 For Reduction of Time of Oscillation to an Infinitely Small Arc. 
 
 4 4 64 4 
 
 If t = observed time and 
 T = true reduced time 
 T-t-kt. 
 
 a 
 
 fc 
 
 a 
 
 k 
 
 a 
 
 & 
 
 
 
 o.ooooo 
 
 7 
 
 o .00023 
 
 14 
 
 O.OOOQ3 
 
 i 
 
 ooo 
 
 8 
 
 030 
 
 J 5 
 
 107 
 
 2 
 
 002 
 
 9 
 
 39 
 
 16 
 
 122 
 
 3 
 
 004 
 
 10 
 
 048 
 
 J 7 
 
 138 
 
 4 
 
 008 
 
 1 1 
 
 058 
 
 18 
 
 154 
 
 5 
 
 012 
 
 12 
 
 069 
 
 *9 
 
 172 
 
 6 
 
 Oi; 
 
 !3 
 
 080 
 
 20 
 
 I9O 
 
 7 
 
 023 
 
 14 
 
 093 
 
 
 
 TABLE IV. 
 Reduction of Barometer Readings to o. 
 
 (The corrections below are in mm. and are to be subtracted. The 
 uncorrected height is in cm.) 
 
 Temp. 
 
 Brass Scale 
 
 Glass Scale 
 
 72 
 
 73 
 
 74 
 
 75 
 
 76 
 
 77 
 
 78 
 
 74 
 
 75 
 
 76 
 
 77 
 
 78 
 
 15 
 
 r -75 
 
 1.77 1.81 1.83 
 
 1.86 
 
 1.88 
 
 1.91 
 
 1.92 
 
 1.94 
 
 1.97 
 
 2.0O 
 
 2. 02 
 
 1 6 
 
 1.87 
 
 1.89 1.93 
 
 1.96 
 
 1.98 
 
 2.01 
 
 2.03 
 
 2.05 
 2.17 
 
 2.07 
 
 2. 2O 
 
 2.10 
 
 2.1 3 
 
 2.16 
 2 29 
 
 i7 
 
 1.98 
 
 2.01 
 
 2.05 
 
 2.08 
 
 2.IO 
 
 2.13 
 
 2.16 
 
 2.23 
 
 2.26 
 
 18 
 
 2.IO 
 
 2-13 
 
 2.17 
 
 2. 2O 
 
 2.23 
 
 2.26 2.29 
 
 2.30 
 2-43 
 
 2-33 
 
 2.36 
 
 2-39 
 
 2-43 
 2.56 
 
 "9 
 
 2.22 
 
 2.25 
 
 2.29 
 
 2.32 
 
 2-35 
 
 2.38 2.41 
 
 2.46 
 
 2-49 
 
 2.53 
 
 20 
 
 2 -33 
 
 2.37 2.41 
 
 2-44 
 
 2.47 
 
 2-5 1 
 
 2-54 
 
 2.56 
 
 -59 
 
 2.62 
 
 2.66 
 
 2.69 
 
 2 I 
 
 2-45 
 
 2.48 2.53 
 
 2.56 
 2.69 
 
 2.6O 
 
 2.63 
 2.76 
 
 2.67 
 
 2.68 
 2.81 
 
 2.72 
 
 2. 7 6 
 2.89 
 3.02 
 
 3.15 
 3.28 
 
 2.79 
 
 2.8 3 
 2.96 
 
 22 
 
 2-57 
 
 2.60 2.65 
 
 2.72 
 
 2.79 
 
 2.8 5 
 
 2.92 
 
 2 3 
 
 2.68 
 
 2.72 
 
 2-77 
 
 2.81 
 
 2.8 4 
 
 2 88 
 
 2.92 
 
 2.94 
 
 2.98 
 
 3- 11 
 3.23 
 
 3.06 
 3- J 9 
 
 3- J o 
 3-23 
 3-36 
 
 24 
 
 2.80 
 
 2.84 2.89 
 
 2.93 
 
 3.05 
 
 2.97 
 
 3-9 
 
 3.01 
 3-!3 
 
 3-05 
 3-J7 
 
 3.06 
 
 3- J 9 
 
 i 
 
 25 
 
 2.92 
 
 2.96 
 
 3.01 
 
 3-32 
 
240 
 
 TABLES. 
 
 TABLE V. 
 
 Density and Volume of One Gram of Water at Different 
 Temperatures. 
 
 Temp. 
 
 Density 
 
 Vol. of i. gr. 
 
 Temp. 
 
 Density 
 
 Vol. of i . gr. 
 
 
 
 0.999878 
 
 I.OOOI22 
 
 J 
 
 21 
 
 0.998065 
 
 .001939 
 
 i 
 
 Q-999933 
 
 .000067 
 
 22 
 
 0.997849 
 
 .0021 56 
 
 2 
 
 0.999972 
 
 .000028 
 
 2 3 
 
 0.997623 
 
 .002383 
 
 3 
 
 0.999993 
 
 .000007 
 
 24 
 
 0.997386 
 
 .002621 
 
 4 
 
 I.OOOOOO 
 
 .000000 
 
 25 
 
 0.997140 
 
 .002868 
 
 5 
 
 0.999992 
 
 .000008 
 
 30 
 
 0-99577 
 
 .00425 
 
 6 
 
 0.999969 
 
 .000031 
 
 35 
 
 0.99417 
 
 .00586 
 
 7 
 
 Q-999933 
 
 .000067 
 
 40 
 
 0.99236 
 
 .00770 
 
 8 
 
 0.999882 
 
 .000118 
 
 45 
 
 0-99035 
 
 .00974 
 
 9 
 
 0.999819 
 
 .000181 
 
 50 
 
 0.98817 
 
 .01197 
 
 10 
 
 0-999739 
 
 .000261 
 
 55 
 
 0.98584 
 
 .01436 
 
 1 1 
 
 o 999650 
 
 .000350 
 
 60 
 
 0-98334 
 
 .01694 
 
 12 
 
 0.999544 
 
 .000456 
 
 65 
 
 0.98071 
 
 .01967 
 
 *3 
 
 0.999430 
 
 .000570 
 
 7 
 
 0.97789 
 
 .02261 
 
 U 
 
 0.999297 
 
 .000703 
 
 75 
 
 0-97493 
 
 .02570 
 
 i5 
 
 0.999154 
 
 .000847 
 
 80 
 
 0.97190 
 
 .02891 
 
 16 
 
 0.999004 
 
 .000997 
 
 85 
 
 0.96876 
 
 -03225 
 
 17 
 
 0.998839 
 
 .001162 
 
 90 
 
 0.96549 
 
 03574 
 
 18 
 
 0.998663 
 
 .001339 
 
 95 
 
 0.96208 
 
 .03941 
 
 19 
 
 0.998475 
 
 .001527 
 
 IOO 
 
 0.95856 
 
 04323 
 
 20 
 
 0.998272 
 
 .001731 
 
 
 
 
 TABLE VI. 
 
 Density of Gases (o, 76 cm.). 1 
 
 Hydrogen 00008987 
 
 Oxygen 0014290 
 
 Nitrogen .0012507 
 
 Air 0012928 
 
 Chlorine 003167 
 
 Carbon monoxide 0012504 
 
 Carbon dioxide 0019768 
 
 Ethane 001341 
 
 Ethylene 001252 
 
 Steam (at 100) 00060315 
 
 Largely from Guye, J. Ch, Phys., 1907, p. 203. 
 
TABLES. 
 
 2AI 
 
 TABLE VII. 
 Density (o), Specific Heat (o), and Coefficient of Linear Expansion 
 
 Element 
 
 Density 
 
 Specific 
 Heat 
 
 Coef. of Lin. Exp. 
 Multiplied by io 6 
 
 \luminum 
 
 2 60 
 
 2 2 
 
 231 
 
 Antimony 
 
 6.62 
 
 O4Q 
 
 ^ j. x 
 
 Bismuth 
 
 9.8 
 
 .OT, I 
 
 
 Cadmium 
 
 8.61 
 
 O ? ? 
 
 -2Q.7 
 
 Carbon, diamond 
 Carbon, graphite 
 Carbon gas carbon 
 
 3-52 
 2.25 
 
 i .00 
 
 .10 
 15 
 
 3 r.?8 
 7.8 
 
 5 A 
 
 Cobalt 
 
 8.8 
 
 . I 06 
 
 124 
 
 Copper 
 Copper sulphate (crys ) 
 
 8.92 
 S-S8 
 
 .094 
 
 16.8 
 
 Gold 
 
 IQ.3 
 
 .032 
 
 M-4 
 
 Iron 
 
 7.8 
 
 . I I 
 
 I 2.1 
 
 Lead 
 
 n.^6 
 
 O2 O 
 
 202 
 
 Magnesium . 
 
 1.74. 
 
 
 
 Mercury 
 
 I "? . ^06 
 
 QT.T.T. 
 
 181 (cub. exp ) 
 
 Nickel 
 
 
 .IO8 
 
 12.8 
 
 Phosphorus yellow . 
 
 1.83 
 
 .20 
 
 
 Phosphorus, red 
 
 2.IQ 
 
 I 7 
 
 
 Phosphorus, metallic. 
 
 2.^4 
 
 
 
 Platinum 
 
 214. 
 
 O 3 3 
 
 90 
 
 Potassium chloride 
 
 I 08 
 
 
 
 Silver 
 
 I O ^ 3 
 
 .0 ^6 
 
 I Q. 2 
 
 Sodium chloride 
 
 2.1 C 
 
 
 
 Sodium sulphate. . . . 
 
 2.6 t? 
 
 
 
 Tin 
 
 7.3 
 
 .0 ^6 
 
 22.3 
 
 Zinc 
 
 7-2 
 
 OQ4 
 
 20. 2 
 
 Zinc sulphate (anhy ) 
 
 3 A Q 
 
 
 
 
 4V 
 
 
 
 16 
 
242 
 
 TABLES. 
 
 TABLE VIII. 
 
 Density, Specific Heat, and Coefficient of Expansion of 
 Miscellaneous Substances (o). 
 
 Substance 
 
 Density 
 
 Specific 
 heat 
 
 Coef. of Lin. Exp. 
 (Xio 6 ) 
 
 Castor oil 
 
 .060 
 
 
 
 Glass, green 
 Glass crown 
 
 2.6 
 
 2 . 7 
 
 .19 
 . 19 
 
 8.9 
 8.8 
 
 Glass crystal 
 
 2 
 
 . 18 
 
 7 . 7 
 
 Glass flint 
 
 3. I ^ 3 O 
 
 . 10 
 
 7 3 
 
 Hard rubber 
 Marble 
 Paraffin 
 
 I-I5 
 
 2-75 
 .89 
 
 
 7-7 
 11.7 
 
 Quartz, crystal II .... 
 Quartz, crystal - 1 - 
 Quartz fused 
 
 2-653 
 2-653 
 
 2 2O 
 
 .19 
 
 7.2 
 13.2 
 
 54 
 
 Alcohol (ethyl) 
 Benzol 
 
 .81 
 .800 
 
 54 
 .^8 
 
 i .048 1 
 I.I76 1 
 
 Carbon bisulphide. . . . 
 Chloroform 
 
 1.293 
 I *\3 
 
 .24 
 
 2 3 
 
 I.I4 1 
 i n 1 
 
 Ether (ethyl) 
 Glycerine 
 
 74 
 i 26 
 
 S3 
 
 . s8 
 
 i-Si 1 
 
 
 
 
 
 TABLE IX. 
 Average Value of Elastic Moduli. 
 
 Shear Modulus. 
 
 Coefficient of cubical expansion Xio 3 . 
 
 Young's Modulus. 
 
 
 
 
 Brass 
 
 V7Xio u 
 
 10. 4X10" 
 
 Iron 
 
 7.7X10" 
 
 I9-6XI0 11 
 
 Steel 
 
 8.2X10" 
 
 22 Xio 11 
 
 
 
 
TABLES. 243 
 
 TABLE X. 
 
 Surface Tension T (15), Temperature Coefficient of Surface 
 Tension c', and Angle of Contact a. 
 
 
 T 
 
 c' 
 
 
 
 Ethyl ether 
 
 10 
 
 . i j 
 
 1 6 
 
 
 
 
 
 Ethyl alcohol 
 
 2 ^ 
 
 087 
 
 
 
 
 
 
 
 Benzol 
 
 31 
 
 13 
 
 
 
 Water 
 
 7 6 
 
 '5 
 
 small 
 
 Mercury 
 
 527 
 
 --38 
 
 i35 
 
 TABLE XL 
 
 Coefficient of Viscosity (20). * 
 
 Water oioo 
 
 Mercury o 1 59 
 
 Acetic acid 0122 
 
 Methyl alcohol 00591 
 
 Ethyl alcohol .0119 
 
 Ethyl ether 00234 
 
 Benzol 00649 
 
 1 "\Vinkelmann, 1908, I, 2, p. 1397. 
 
244 
 
 TABLES. 
 
 TABLE XII. 
 Specific Heats of Gases.' 
 
 Temp. 
 
 s v 
 
 Argon... 20 .1205 
 
 Helium 20 1.25 
 
 Mercury 27 5~3 56 .0246 
 
 Hydrogen o-2oo 3.406 
 
 Nitrogen -3o-20o .244 
 
 Oxygen o-2oo .217 
 
 Air o-2oo .2375 
 
 Chlorine i9-343 - 1 J 5 
 
 Iodine 2oo--377 -0336 
 
 Bromine 85-228 .0555 
 
 Water i 3 o -25o .480 
 
 Hydrogen sulphide... io-2oo .245 
 
 Carbon dioxide i 1 00 .217 
 
 Ammonia j 2o-2io .512 
 
 Chloroform ! 28-n8 .144 
 
 Ethyl alcohol i io-22o .453 
 
 Ether 7o-225 .480 
 
 Benzol n6-2i8 .375 
 
 1.66 
 
 1.64 
 
 1.66 
 
 1.396 
 
 1.405 
 
 1.40 
 
 1-405 
 
 1.32 
 
 1.29 
 
 1.29 
 
 1.287 
 
 1.28 
 
 1.28 
 
 1.14 
 1.07 
 1.187 
 
 1 Jiiptner, Phys. Chem. I, pp. 71-73- 
 

 TABLES. 245 
 
 
 TABLE XIII. 
 
 Pressure of Saturated Water Vapor (Regnault). 
 
 
 (mm.) 
 
 Temp. 
 
 Pressure Temp. 
 
 Pressure 
 
 
 | 
 
 
 
 Ice Water 29 
 
 29.782 
 
 
 
 30 
 
 3I-548 
 
 
 
 10 
 
 1.999 2.078 31 
 
 33-405 
 
 
 3 2 
 
 35 359 
 
 8 
 
 2-379 2.456 33 
 
 37-4io 
 
 
 
 34 
 
 39 565 
 
 
 
 6 
 
 2.821 2.890 35 
 
 41.827 
 
 
 
 40 
 
 54.906 
 
 
 
 4 
 
 3-334 3-387 45 
 
 7J-39 1 
 
 
 
 50 
 
 91.982 
 
 
 
 2 
 
 3-925 3-955 55 
 
 117.479 
 
 
 
 60 
 
 148.791 
 
 
 
 
 65 
 
 186.945 
 
 O 
 
 
 4.600 
 
 70 
 
 233-093 
 
 + I 
 
 
 4.940 
 
 75 
 
 288.517 
 
 2 
 
 
 5-302 
 
 80 
 
 354-643 
 
 3 
 
 
 5.687 
 
 85 
 
 433-41 
 
 4 
 
 
 6.097 
 
 90 
 
 525-45 
 
 5 
 
 
 6-534 
 
 9 1 
 
 545-78 
 
 6 
 
 
 6.998 
 
 92 
 
 566.76 
 
 7 
 
 
 7.492 
 
 93 
 
 588.41 
 
 8 
 
 
 8.017 
 
 94 
 
 610.74 
 
 9 
 
 
 8-574 
 
 95 
 
 633-78 
 
 10 
 
 
 9.165 
 
 96 
 
 657-54 
 
 ii 
 
 
 9.792 
 
 97 
 
 682.03 
 
 12 
 
 
 10.457 
 
 98 
 
 707.26 
 
 13 
 
 
 11.062 
 
 98-5 
 
 720.15 
 
 14 
 
 
 11.906 
 
 99-o 
 
 733-91 
 
 15 
 
 
 12.699 
 
 99-5 
 
 746.50 
 
 16 
 
 
 !3- 6 35 
 
 IOO.O 
 
 760.00 
 
 17 
 
 
 14.421 
 
 100.5 
 
 773-7 1 
 
 18 
 
 
 15-357 
 
 IOI.O 
 
 787-63 
 
 19 
 
 
 16.346 
 
 IO2.O 
 
 816.17 
 
 20 
 
 
 i7-39i 
 
 IO4.O 
 
 875.69 
 
 21 
 
 
 18.495 
 
 105 
 
 906.41 
 
 22 
 
 
 19.659 
 
 no 
 
 1075.4 
 
 23 
 
 
 20.888 
 
 I2O 
 
 i49!-3 
 
 24 
 
 
 22.184 
 
 I 3 
 
 2030.3 
 
 25 
 
 
 23-550 
 
 X 5 
 
 358i-2 
 
 26 
 
 
 24.998 
 
 1 7S 
 
 6717 
 
 27 
 
 
 26.505 
 
 200 
 
 1 1690 
 
 28 
 
 
 28.101 225 19097 
 
246 
 
 TABLES. 
 
 TABLE XIV. 
 Boiling Point of Water, t, at Barometric Pressure, p,(ui m.) . 
 
 p> 
 
 t. 
 
 P> 
 
 t. 
 
 P- 
 
 t. 
 
 740 
 
 99.26 
 
 750 
 
 99-63 
 
 760 
 
 100.00 
 
 41 
 
 .29 
 
 5i 
 
 .67 
 
 61 
 
 .04 
 
 42 
 
 33 
 
 52 
 
 .70 
 
 62 
 
 .07 
 
 43 
 
 37 
 
 53 
 
 74 
 
 63 
 
 .1 1 
 
 44 
 
 .41 
 
 54 
 
 .78 
 
 64 
 
 15 
 
 45 
 
 .44 
 
 55 
 
 .82 
 
 65 
 
 .18 
 
 46 
 
 .48 
 
 56 
 
 85 
 
 66 
 
 .22 
 
 47 
 
 52 
 
 57 
 
 .89 
 
 67 
 
 .26 
 
 48 
 
 56 
 
 58 
 
 93 
 
 68 
 
 .29 
 
 49 
 
 59 
 
 59 
 
 .96 
 
 69 
 
 33 
 
 75 
 
 99-63 
 
 760 
 
 100.00 
 
 770 
 
 100.36 
 
TABLES. 
 
 247 
 
 TABLE XV. 
 Wet and Dry Bulb Hygrometer. 
 
 (Actual vapor pressures (mm.) for different temperatures of dry 
 thermometer and various differences of temperature between the two 
 thermometers. 
 
 The first vertical column gives the temperature of the dry-bulb 
 thermometer. The first horizontal line gives the difference between 
 the two thermometers. Since the difference is zero if the air is satu- 
 rated, the second vertical column gives the saturated vapor pressure 
 for the corresponding temperatures in the first column.) 
 
 tc. 
 
 o 
 
 I 
 
 2 
 
 3 
 
 4 
 
 5 
 
 6 
 
 7 
 
 8 
 
 9 
 
 10 
 
 I I 
 
 o 
 
 4.6 
 
 3-7 
 
 2-9 
 
 2.1 
 
 J -3 
 
 
 
 
 
 
 
 
 i 
 
 4-9 
 
 4.0 
 
 3-2 
 
 2.4 
 
 1.6 
 
 0.8 
 
 
 
 
 
 
 
 2 
 
 5-3 
 
 4-4 
 
 3-4 
 
 2.7 
 
 1.9 
 
 I.O 
 
 
 
 
 
 
 
 3 
 
 5-7 
 
 4-7 
 
 3-7 
 
 2.8 
 
 2.2 
 
 1 '3 
 
 
 
 
 
 
 
 4 
 
 6.1 
 
 5- 1 
 
 4.1 
 
 3-2 
 
 2-4 
 
 1.6 
 
 0.8 
 
 
 
 
 
 
 5 
 
 6-5 
 
 5-5 
 
 4-5 
 
 3-5 
 
 2.6 
 
 1.8 
 
 I.O 
 
 
 
 
 
 
 6 
 
 7.0 
 
 5-9 
 
 4-9 
 
 3-9 
 
 2. 9 
 
 2.0 
 
 i.i 
 
 
 
 
 
 
 7 
 
 7-5 
 
 6.4 
 
 5-3 
 
 4-3 
 
 3-3 
 
 2-3 
 
 1.4 
 
 0.4 
 
 
 
 
 
 8 
 
 8.0 
 
 6.9 
 
 5-8 
 
 4-7 
 
 3-7 
 
 2.7 
 
 *-7 
 
 0.8 
 
 
 
 
 
 9 
 
 8.6 
 
 7-4 
 
 6-3 
 
 S- 2 
 
 4.1 
 
 3- 1 
 
 2.1 
 
 i.i 
 
 O.2 
 
 
 
 
 10 
 
 9.2 
 
 8.0 
 
 6.8 
 
 5-7 
 
 4.6 
 
 3-5 
 
 2-5 
 
 i-5 
 
 o-5 
 
 
 
 
 ii 
 
 9.8 
 
 8.6 
 
 7-4 
 
 6.2 
 
 5- 1 
 
 4.0 
 
 2.9 
 
 1.9 
 
 0.9 
 
 
 
 
 12 
 
 10.5 
 
 9.2 
 
 8.0 
 
 6.8 
 
 5-6 
 
 4-5 
 
 3-4 
 
 2-3 
 
 i-3 
 
 
 
 
 *3 
 
 II. 2 
 
 9.8 
 
 8.6 
 
 7-3 
 
 6.2 
 
 5-o 
 
 3-9 
 
 2.8 
 
 i-7 
 
 
 
 
 14 
 
 II.9 
 
 10.6 
 
 9.2 
 
 8.0 
 
 6.7 
 
 56 
 
 4-4 
 
 3-3 
 
 2.2 
 
 i.i 
 
 
 
 15 
 
 12.7 
 
 n-3 
 
 9-9 
 
 8.6 
 
 7-4 
 
 6.1 
 
 S-o 
 
 3-8 
 
 2.7 
 
 1.6 
 
 o-5 
 
 
 16 
 
 J 3-5 
 
 12. 1 
 
 10.7 
 
 9-3 
 
 8.0 
 
 6.8 
 
 5-5 
 
 4-3 
 
 3-2 
 
 2.1 
 
 I.O 
 
 
 I 7 
 
 14.4 
 
 13.0 
 
 n-5 
 
 IO.I 
 
 8-7 
 
 7-4 
 
 6.2 
 
 4-9 
 
 3-7 
 
 2.6 
 
 !-5 
 
 0.4 
 
 18 
 
 i5-4 
 
 I 3 .8 
 
 12.3 
 
 10.9 
 
 9-5 
 
 8.1 
 
 6.8 
 
 5-5 
 
 4-3 
 
 3- 1 
 
 2.O 
 
 0.9 
 
 19 
 
 16.4 
 
 14.7 
 
 13-2 
 
 11.7 
 
 10.3 
 
 8.9 
 
 7-5 
 
 6.2 
 
 4-9 
 
 3-7 
 
 2-5 
 
 1.4 
 
 20 
 
 17.4 
 
 15-7 
 
 14-1 
 
 12.6 
 
 ii. i 
 
 9-7 
 
 8-3 
 
 6.9 
 
 5-6 
 
 4-3 
 
 3-i 
 
 1.9 
 
 21 
 
 18.5 
 
 16.8 
 
 i5-i 
 
 13-5 
 
 12. 
 
 10.5 
 
 9.0 
 
 7-6 6.3 
 
 5- 
 
 3-7 
 
 2-5 
 
 22 
 
 19.7 
 
 17.9 
 
 16.2 
 
 14.5 
 
 I2. 9 
 
 11.4 
 
 9-9 
 
 8. 4 | 7.0 
 
 5-7 
 
 4-4 
 
 3-i 
 
 2 3 
 
 20.9 
 
 19.0 
 
 J 7-3 
 
 15-6 
 
 13-9 
 
 12.3 
 
 10.8 
 
 9-2 
 
 7-8 
 
 6.4 
 
 5- 1 
 
 3-8 
 
 24 
 
 22.2 
 
 20.3 
 
 18.4 
 
 16.6 
 
 14.9 
 
 13-3 
 
 11.7 
 
 IO.I 
 
 8-7 
 
 7-2 
 
 5-8 
 
 4-5 
 
 25 
 
 23.6 
 
 21.6 
 
 19.7 
 
 17.8 
 
 16.0 
 
 14-3 
 
 12.7. 
 
 II. I 
 
 9-5 
 
 8.0 
 
 6.6 
 
 5-2 
 
 26 
 
 25.0 
 
 22.9 
 
 21. 
 
 19.0 
 
 17.2 
 
 i5-4 
 
 J 3-7 
 
 12. 1 
 
 10.5 
 
 8.9 
 
 7-4 
 
 6.0 
 
 27 
 
 26.5 
 
 24.9 
 
 22.3 
 
 20.3 
 
 18.4 
 
 16.6 
 
 14.8 
 
 I3- 1 
 
 11.4 
 
 9.8 
 
 8-3 
 
 6.8 
 
 28 
 
 28.1 
 
 25-9 
 
 23-7 
 
 21.7 
 
 19.7 
 
 17.6 
 
 16.0 
 
 14.2 
 
 12.5 
 
 10.8 
 
 9-2 : 7-7 
 
 29 
 
 29.8 
 
 27-5 
 
 25-3 
 
 23.1 
 
 21. 1 
 
 19.1 
 
 17.2 
 
 15-3 
 
 13.6 11.9 
 
 IO.2 
 
 8.6 
 
 3 
 
 3 1.6 
 
 29.2 
 
 26.9 24.6 
 
 22.5 20-5 
 
 18.5 
 
 16.6 
 
 14.7 
 
 13.0 
 
 I 1.2 
 
 9.6 
 
248 
 
 TABLES. 
 
 TABLE XVI. 
 Vapor Pressure of Mercury (mm.). 
 
 Temp. 
 
 Pres. 
 
 Temp. 
 
 Pres. 
 
 o 
 
 O.O2 
 
 170 
 
 8.091 
 
 + 20 
 
 O.O4 
 
 1 80 
 
 I I.OOO 
 
 40 
 
 O.o8 
 
 190 
 
 14.84 
 
 60 
 
 0.16 
 
 200 
 
 19.90 
 
 80 
 
 o-35 
 
 210 
 
 26.35 
 
 IOO 
 
 0.746 
 
 22O 
 
 34 7 
 
 no 
 
 1.073 
 
 230 
 
 45-35 
 
 I2O 
 
 *-534 
 
 240 
 
 58.82 
 
 I 3 
 
 2-175 
 
 250 
 
 75-75 
 
 140 
 
 3-059 
 
 260 
 
 96.73 
 
 ISO 
 
 4.266 
 
 270 
 
 123.01 
 
 I 60 
 
 5.900 
 
 280 
 
 I55-I7 
 
 TABLE XVII. 
 
 Melting Point of Metals. (Holborn and Day and 
 Waidner and Burgess. 1 ) 
 
 Tin . . 232 
 
 Cadmium 321 
 
 Lead 
 
 Zinc 
 
 Antimony 
 
 Aluminum 
 
 Silver 
 
 Gold 
 
 Copper .... 
 
 327 
 419 
 
 63 1 
 
 657 
 961 
 1063 
 1084 
 
 Platinum 1 770 
 
 Rev., 1909, p. 467; Compt. Rend., 1909, cxlviii, p. 
 
 1177. 
 
TABLES. 
 
 249 
 
 TABLE XVIII. 
 Wave Lengths in Angstrom Units (io- 8 cm.). 
 
 Line 
 
 Element 
 
 Wave Length 
 
 Color 
 
 C H a . . .. 
 
 Hydrogen. 
 
 6 ^63 o ?4 
 
 Red 
 
 Dr 
 
 Sodium 
 
 c8o6 i ? ^ 
 
 Yellow 
 
 D, 
 
 Sodium 
 
 5890 182 
 
 Yellow 
 
 F H^ 
 
 Hydrogen 
 
 4861. <C27 
 
 Blue 
 
 G' H v 
 
 Hydrogen 
 
 4. -i 4.0. 6^4. 
 
 Violet 
 
 H T " . 
 
 Calcium 
 
 ^068 62 s 
 
 Violet 
 
 
 Helium 
 
 7O6 ^.2 
 
 Red 
 
 
 Helium 
 
 6678.1 
 
 Red 
 
 
 Helium 
 
 c8? ?.6 
 
 Yellow 
 
 
 Helium 
 
 t;oi <.7 
 
 Green 
 
 
 Helium 
 
 402 i.o 
 
 Blue 
 
 
 Helium 
 
 471 ^.2 
 
 Blue 
 
 
 Helium 
 
 4-47 I. ^ 
 
 Violet 
 
 
 Mercury 
 
 623 2 O 
 
 Red 
 
 
 Mercury 
 
 C7QO.7 
 
 Yellow 
 
 
 Mercury 
 
 ^760-6 
 
 Yellow 
 
 
 Mercury 
 
 ^460.7 
 
 Green 
 
 
 Mercury 
 
 40 ^0.7 
 
 Green-Blue 
 
 
 Mercury 
 
 4916.4 
 
 Blue 
 
 
 Mercury 
 
 4358.3 
 
 Blue 
 
 
 Mercury 
 
 4078.1 
 
 Violet 
 
 
 Mercury 
 
 4046.8 
 
 Violet 
 
 K~ 
 
 Potassium 
 
 7699.3 
 
 Red 
 
 Ko 
 
 Potassium 
 
 =:8^2.2 
 
 Yellow 
 
 ^p 
 
 fr ; 
 
 La 
 
 Li R 
 
 Potassium 
 Lithium 
 Lithium 
 
 4047.4 
 6708.2 
 6103.8 
 
 Violet 
 Red 
 Orange 
 
 
 Cadmium 
 
 6438.5 
 
 Red 
 
 
 Cadmium 
 
 5085.8 
 
 Green 
 
 
 Cadmium 
 
 4799.9 
 
 Blue 
 
2sO 
 
 TABLES. 
 
 TABLE XIX. 
 
 Refractive Indices. 
 
 [Yellow light, (D lines) 20.] 
 
 Glass, light crown (density = 2 . 50) 
 
 Glass, heavy crown (density = 3 .00) 
 
 Glass, light flint (density ==2 .87) 
 
 Glass, heavy flint (density =4.22) 
 
 Quartz, crystal, -*- , ord 
 
 Quartz, crystal, - 1 - , ext 
 
 Alcohol, ethyl 
 
 Benzol, 
 
 Carbon bisulphide 
 
 Chloroform 
 
 Ether, ethyl 
 
 Glycerine 
 
 Water 
 
 Air, o, 7 6 cm i 
 
 .5280 
 .5604 
 .5410 
 7102 
 5442 
 5533 
 .3614 
 .5014 
 .6277 
 .4490 
 356o 
 .4729 
 
 33 2 9 
 .000293 
 
 TABLE XX. 
 Specific Rotatory Power (20). Yellow Sodium Light (D).* 
 
 Active Substance 
 
 Concentration 
 (=c) (gr.iniooc.c.; 
 
 Cane-sugar R 
 
 / 3-28 
 
 66.639 .02o8<7 
 
 Invert sugar L 
 
 \ 10-86 
 
 I 14. 
 
 66.453 .000124(7 
 
 2O 07 O4l 
 
 Glucose (dextrose). R 
 (crystallized) 
 Fructose (levulose) L . . 
 Milk-sugar, R 
 
 Tartaric acid R ... 
 
 0-100% 
 0-40 
 
 f ".' 
 
 47-73 +.oisX% 
 100.3 -f.ioSc 
 
 52-53 
 15.06 .131(7 
 
 Quartz R or L 
 
 1 22-63 
 
 i3.436-.ii9C 
 21 70 (for i mm thickness) 
 
 
 
 
 Landolt and Bornstein. 
 
TABLES. 
 
 TABLE XXI. 
 Photometric Table. 
 
 251 
 
 
 -. . x 1 ui at, 
 
 ^<-xj-pcn \> s. n 
 
 VJbVJlllClilll/ U< 
 
 ( 3 oo-w) 2 ' 
 
 
 n 
 
 
 
 I 
 
 2 
 
 3 
 
 A 
 
 50 
 
 0.0400 
 
 O.42O 
 
 0.0440 
 
 0.0460 
 
 0.0482 
 
 60 
 
 .0625 
 
 -6 5 I 
 
 .0678 
 
 .0706 
 
 735 
 
 70 
 
 .0926 
 
 .0961 
 
 .0997 
 
 1034 
 
 . 1072 
 
 80 
 
 .1322 
 
 .1368 
 
 .1414 
 
 .1643 
 
 .1512 
 
 90 
 
 1837 
 
 .1896 
 
 !957 
 
 .2018 
 
 . 2082 
 
 IOO 
 
 25OO 
 
 .2576 
 
 2653 
 
 2734 
 
 .2815 
 
 I IO 
 
 335 2 
 
 3449 
 
 3549 
 
 3652 
 
 3756 
 
 I2O 
 
 4445 
 
 457 
 
 .4698 
 
 .4829 
 
 .4964 
 
 130 
 
 .5848 
 
 . 6009 
 
 6i73 
 
 6343 
 
 .6516 
 
 I4O 
 
 .7656 
 
 .7864 
 
 .8078 
 
 .8296 
 
 .8521 
 
 J 5 
 
 i .0000 
 
 i .027 
 
 1-055 
 
 1.083 
 
 1.113 
 
 1 60 
 
 1.306 
 
 1.342 
 
 1-379 
 
 i .416 
 
 !-454 
 
 170 
 
 i . 700 
 
 i-757 
 
 i. 806 
 
 1.856 
 
 1.907 
 
 1 80 
 
 2.250 
 
 2-313 
 
 2-379 
 
 2.446 
 
 2.516 
 
 190 
 
 2.983 
 
 3.070 
 
 3 .160 
 
 3-253 
 
 3-35 
 
 200 
 
 4.000 
 
 4.122 
 
 4-285 
 
 4-380 
 
 4.516 
 
 210 
 
 5-444 
 
 5.621 
 
 5-803 
 
 5-994 
 
 6.192 
 
 220 
 
 7-563 
 
 7.826 
 
 8.100 
 
 8.387 
 
 8.687 
 
 230 
 
 10.80 
 
 I I .21 
 
 11.64 
 
 12.09 
 
 I2 -57 
 
 240 
 
 16.00 
 
 16.68 
 
 17.41 
 
 18.17 
 
 18.98 
 
 250 
 
 25.00 
 
 
 
 
 
 n 
 
 5 
 
 6 
 
 7 
 
 8 
 
 9 
 
 5 
 
 0.0504 
 
 0.0527 
 
 0.0550 
 
 0.0574 
 
 0.0599 
 
 60 
 
 .0765 
 
 .0796 
 
 .0827 
 
 .0859 
 
 .0892 
 
 70 
 
 .mi 
 
 .1151 
 
 . 1 193 
 
 1235 
 
 .1278 
 
 80 
 
 i5 6 3 
 
 .1615 
 
 .1668 
 
 !7 2 3 
 
 .1779 
 
 90 
 
 .2148 
 
 .2215 
 
 .2283 
 
 2354 
 
 .2428 
 
 IOO 
 
 .2899 
 
 .2985 
 
 3074 
 
 .3164 
 
 3256 
 
 no 
 
 .3864 
 
 3974 
 
 .4088 
 
 .4204 
 
 4323 
 
 120 
 
 .5012 
 
 5244 
 
 5389 
 
 5538 
 
 .5691 
 
 I 3 
 
 .6694 
 
 .6877 
 
 .7064 
 
 7257 
 
 7454 
 
 140 
 
 .8752 
 
 .8988 
 
 .9231 
 
 .9481 
 
 9737 
 
 I 5 
 
 1-143 
 
 1.174 
 
 i .205 
 
 1-238 
 
 1.272 
 
 1 60 
 
 1.494 
 
 i-535 
 
 i-577 
 
 i .620 
 
 i .664 
 
 170 
 
 i .960 
 
 2 .OI4 
 
 2 .071 
 
 2.129 
 
 2.188 
 
 180 
 
 2.588 
 
 2.662 
 
 2-739 
 
 2.817 
 
 2.889 
 
 190 
 
 3-449 
 
 3-552 
 
 3-658 
 
 3.768 
 
 3.882 
 
 200 
 
 4.656 
 
 4.803 
 
 4-954 
 
 5.111 
 
 5-275 
 
 2IO 
 
 6.398 
 
 6.122 
 
 6-835 
 
 7.068 
 
 7.310 
 
 22O 
 
 9.000 
 
 9-327 
 
 9.670 
 
 10.03 
 
 10.40 
 
 2 3 
 
 13.07 
 
 13.60 
 
 14-15 
 
 14.74 
 
 15-35 
 
 24O 
 
 19.84 
 
 20.75 
 
 21.72 
 
 22.75 
 
 23.84 
 
 2 5 
 
 
 
 
 
 
2 5 2 
 
 TABLES. 
 
 TABLE XXII. 
 Specific Resistance at o C. and Temperature Coefficient. 
 
 
 Specific 
 Resistance 
 
 Temperature 
 Coefficient 
 
 Bismuth 
 Copper ( 
 Copper ( 
 German 
 Iron. . . 
 
 (hard) 
 
 132.6X10- 
 i .590X10- 
 1.622X10- 
 20.24X10- 
 10.43X10- 
 19.85X10- 
 94.07 X io- 
 8.957X10- 
 1.521X10- 
 1.652X10- 
 9.565X10- 
 
 0.0054 
 0.0043 
 
 annealed) 
 
 hard drawn) 
 
 silver (4Cu + 2 Ni + i Zn) 
 
 0.00027 
 0.007 
 0.0039 
 0.00089 
 0.0034 
 0.00377 
 
 Lead (pi 
 Mercury 
 Platinun 
 Silver (a 
 Silver (h 
 Tin 
 
 ressed) 
 i 
 
 nnealed) 
 
 ard drawn) 
 
 0.004 
 
 
 TABLE XXIII. 
 Specific Resistance and Temperature Coefficient of Solutions (18).* 
 
 
 Sp. Res. 
 
 ! Temp. 
 Coef. 
 
 
 Sp. Res. 
 
 Temp. 
 Coef. 
 
 wHCl 
 
 3-3 2 
 
 i o 0165 
 
 nNaCl 
 
 I -7 A f 
 
 O O2 2 6 
 
 o.inHCl 
 o.oinHCl 
 
 28.5 
 
 271. 
 
 
 o.mNaCl.... 
 o.oiwNaCl . . 
 
 108.1 
 974. 
 
 
 wHNO 3 . 
 o.iwHN0 3 ... 
 o.oiwHNO 3 . . 
 
 n iH 2 SO 4 
 
 3-23 
 28.6 
 
 272. 
 
 c o ^ 
 
 0.163 
 o . o 1 64 1 
 
 wKCl .. 
 o.inKCl 
 
 o.oiwKCl . . . 
 
 nAgNO 3 
 
 10.18 
 
 89.5 
 817. 
 
 14. . 7 c 
 
 0.0217 
 o . 02 1 6 
 
 o.iniH 2 SO 4 .. 
 o.oiw$H 2 SO 4 . 
 
 wC 2 H 4 O 2 
 
 44-4 
 325- 
 
 758 
 
 1 
 
 o. in AgNO 3 .. . 
 o.oiwAgNO 3 . 
 
 w^Pb(NO 3 ) 2 
 
 105-7 
 922 . 
 
 21. 8 
 
 
 o.mC 2 H 4 O 2 .. 
 o.oiwC 2 H 4 O 2 . 
 
 2170. 
 6990. 
 
 
 o.miPb(N0 3 ) 2 
 o.omiPb(NO 3 ) 2 
 
 129.4 
 967. 
 
 
 nNaOH . 
 o.iwNaOH .. 
 o.oiwNaOH. 
 
 6-25 
 
 54-7 
 500. 
 
 0.019 
 
 w^ZnSO 4 
 o.iZnSO 4 .. 
 o.omiZnSO 4 
 
 37-6 
 217. 
 1362. 
 
 0.02 5 
 
 wNH 4 OH .. 
 o.mNH 4 OH . 
 o.oiNH 4 OH 
 
 1125. 
 
 33- 
 10420. 
 
 
 niCuSO 4 . 
 o.iw JCuSO 4 .. 
 o.oiw ^CuSO 4 . 
 
 38.8 
 223. 
 1385- 
 
 0.0225 
 
 * A normal solution (designated by the subscript w), contains in one liter a 
 number of grams equal to the chemical equivalent (atomic or molecular weight 
 divided by the valency). A solution with the subscript o. in has one-tenth this 
 concentration, etc. For exmaple, o. iwHCl has 3 .65 grs. of HC1 (gas) in one 
 trile of solution, or that proportion. 
 
TABLES. 253 
 
 TABLE XXIV. 
 Dielectric Constants. 
 
 I 
 
 II 
 
 Hydrocyanic acid 
 Water . . 
 
 .... 96 
 80 
 
 Ether . . 
 
 4(T 
 
 Xylol 
 Benzol 
 
 2.26 
 
 2 2 
 
 Methyl alcohol 
 Ethyl alcohol 
 Ammonia (liquid) 
 Acetone 
 Sulphur dioxide 
 Pyridene 
 
 :::: % 
 
 .... 22 
 .... 17 
 14 
 .... 12 
 
 Toluol 
 
 2 2 
 
 Petroleum 
 
 2.O7 
 
 
 
 
INDEX, 
 
 Aberration, 138 
 Absorption, electric, 196 
 Acceleration of gravity, 36 
 Air, density of, 33 
 
 thermometer, 76 
 Alloys, melting-point, 109 
 Alternating current measurements, 
 
 225, 227 
 
 Ammeter, calibration of, 198 
 Anderson's method (self induction), 
 
 204 
 
 Angle of prism, 131 
 Angular field of view, 143 
 Apparent expansion of gas, 76 
 
 of liquid, 74 
 Arc of vibration, correction, 239 
 
 Balance, 21-25 
 
 correction for air buoyancy, 24 
 
 method of oscillations, 22-24 
 
 ratio of arms, 24 
 Ballistic galvanometer, 156, 208 
 Barometer, 21 
 
 table of corrections, 239 
 Battery, electromotive force, 191, 193 
 
 resistance, 185, 195 
 Beckman thermometer, 65 
 Biquartz, 151 
 Bismuth spiral, 207 
 Boiling-point of water (table), 246 
 Bridge, Wheatstone's, 153 
 Bunsen photometer, 126 
 
 Cadmium cell, 160 
 Calibrating coil, 209, 212 
 Calibration of ammeter, 198 
 
 of galvanometer, 179 
 
 of resistances, 183 
 
 of scale, 26 
 
 of thermometer, 67 
 
 of voltmeter, 196 
 
 Callender's equation (platinum ther- 
 mometer), 117 
 
 Calorimeter, for gases, 113 
 
 for liquids, 114 
 
 for solids, no, in 
 
 simple, 89 
 
 Candle-power, measurement of, 1 26 
 Capacity, absolute measurement, 203 
 
 divided charge method, 178 
 
 measurement (alternating cur- 
 rents), 225, 227 
 
 Capacities, comparison of, 200, 228, 
 230 
 
 different types, 194, 229 
 Carey Foster bridge, 183 
 Cathetometer, 19 
 Chemical hygrometer, 84 
 Chromatic aberration, 138 
 Clark cell, 160 
 Clement and Desermes' method 
 
 (specific heat of gases), 91 
 Coefficient of apparent expension, 74, 
 76 
 
 of expansion, 71, 241, 242 
 
 of friction, 41-45 
 
 of increase of pressure, 76 
 
 of mutual induction, 206 
 
 of self induction, 203 
 
 of viscosity, 56, 243 
 Coincidence method, 38 
 Commutator, double, 161 
 Comparator, 15 
 Condenser, see capacity. 
 Conductivity, thermal, 102 
 
 of electrolyte, 188 
 Copper voltameter, 220 
 Curves, plotting of, 1 1 
 
 Daniell cell, 159 
 Demagnetization of iron, 213 
 Density, of gases, 33, 240 
 
 of liquids, 29 
 
 of powders, 32 
 
 of solids, 28, 241, 242 
 
 of water, 240 
 
 255 
 
256 
 
 INDEX. 
 
 Dew-point, 83 
 
 Dielectric constant, 228, 230 
 
 (table), 253 
 
 Diffraction grating, 147 
 Dip circle, 167 
 Dividing engine, ry 
 Dolezalek electrometer, 176 
 Double bridge, 178 
 Double commutator, 161 
 Drude's apparatus (electric waves), 
 230 
 
 Earth inductor, 169 
 Elastic constants, 242 
 Electric absorption, 196 
 Electrical resonance, 232 
 
 units, 162 
 
 waves, 230 
 
 Electrolytes, resistance of, 188 
 Electrometer, quadrant, 176 
 Electromotive force, device for small, 
 161 
 
 measurement of, 191, 193 
 
 of various cells, 160 
 Equivalent, chemical, 31 
 Errors, 2-10 
 
 of weights, 27 
 Expansion, apparent, 74, 76 
 
 coefficient of, 71, 241, 242 
 Eutectic alloy, 109 
 
 Focal length of lenses, 137, 140 
 
 of mirrors, 134 
 
 Frequency of tuning fork, 44, 1 20 
 Friction, coefficient of kinetic, 42 
 
 coefficient of static, 41 
 
 "G," determination of, 36 
 Galvanometer, bringing to rest, 156 
 
 calibration of, 179, 208 
 
 damping, 157 
 
 different types, 155 
 
 resistance of, 171, 173 
 
 shunt, 157 
 
 study of ballistic, 158, 208 
 
 tangent, 222 
 
 Gas, coefficient of increase of pressure, 
 76 
 
 density of, 240 
 Grating, diffraction, 147 
 
 Heat, conductivity for, 102 
 Heat value of gas, 113 
 
 of liquid, 114 
 
 of solid, no 
 
 Hempel calorimeter, no 
 
 Hooke's law, 46 
 
 Horizontal component of earth's field, 
 
 163, 219 
 Hygrometry, 83 
 Hypsometer, 70 
 Hysteresis, 214 
 
 Incandescent lamp, study of, 128 
 
 Inclination, magnetic, 167 
 
 Index of refraction, measurement of, 
 
 132 
 
 table of, 250 
 
 Inertia, measurement of moment of, 5 2 
 Insulation resistance, 176 
 Interferometer, 149 
 Iron, permeability of, 210 
 
 Junker calorimeter, 113 
 
 Kundt's method (velocity of sound)' 
 122 
 
 Latent heat of fusion, 94 
 
 of vaporization, 96, 99 
 Lenses, combinations, 140 
 
 focal length, 137 
 
 rule of signs, 125 
 Light, filters, 124 
 
 monochromatic, 124 
 Logarithmic decrement, 157 
 
 tables, 236 
 Low resistance, measurement of, 178- 
 
 183 
 
 Lummer-Brodhun photometer, 126, 
 127 
 
 Magnetic field, measurement of, 207 
 of earth, dip, 167 
 of earth, horizontal component, 
 
 163, 219 
 hysteresis, 214 
 permeability, 210 
 Magnetometer, 165 
 Magnification, 141 
 Magnifying power of telescope, 142 
 Mance's method (battery resistance), 
 
 185 
 
 Mechanical equivalent of heat, elec- 
 trical method, 219 
 by friction, 105 
 Melting-point of alloy, 109 
 of metals (table), 248 
 
INDEX. 
 
 257 
 
 Mercury, vapor pressure of (table), 
 
 248 
 
 Michelson's interferometer, 149 
 Micrometer caliper, 14 
 
 microscope, 15 
 Minimum deviation, 132 
 Mirror and scale, adjustment of, 25 
 Mirrors, spherical, measurement of 
 focal length, 134 
 
 rule of signs, 125 
 Moduli, law of, 31 
 Mohr-Westphal balance, 29 
 Moment of inertia, 52 
 Monochromatic light, 124 
 Mutual induction, 206, 227 
 
 Optical lever, 48, 73 
 pyrometer, 116 
 
 Passages, method of, 54 
 
 Pendulum, correction for arc (table), 
 
 239 
 
 physical, 37 
 simple, 36 
 Permeability, 210 
 Photometric table, 251 
 Photometry, 126 
 Pirani's method (mutual induction), 
 
 200 
 
 Pitch of tuning fork, 44, 120 
 Planimeter, 218 
 Platinum thermometer, 116 
 Pohl commutator, 202, 216 
 Polarization, rotation of plane of, 150 
 Possible error, 4-9 
 Post-office box bridge, 154 
 Potentiometer, 197 
 Pressure, coefficient of increase of, 76 
 
 of mercury vapor, 248 
 
 of water vapor (measurement), 80 
 
 (table), 245 
 Primus burner, 114 
 Prism, angle of, 131 
 
 minimum deviation, 132 
 Probable error, 9 
 Pyknometer, 33 
 Pyrometry, 115 
 
 Quadrant electrometer, 176 
 
 Radiation correction, 63 
 
 pyrometer, 116 
 Radius of curvature of mirror, 134 
 
 Ratio of specific heats, measurement 
 of, 91, 122, 123 
 
 table, 244 
 Refractive index, measurement of, 132 
 
 of lenses, 136 
 
 table, 250 
 Regnault's apparatus, hygrometry, 83 
 
 vapor pressure, 81 
 Reports, 2 
 Resistance, boxes, 153 
 
 electrolytic, 188 
 
 high, 175-178 
 
 low, 178-183 
 
 measurement of, 169 
 
 of ballistic galvanometer, 208 
 
 of battery, 185, 195 
 
 of galvanometer, 171, 173 
 
 temperature coefficient of, 186 
 Resistances, comparison of, 183 
 Resolving power, of eye, 146 
 
 of telescope, 145 
 Rigidity of metals, 51 
 Rosenhain calorimeter, in 
 Rotation of plane of polarization, 
 measurement of, 150 
 
 table, 250 
 Rubber grease, 34 
 
 Saccharimetry, 150 
 Scale, calibration of, 26 
 construction of, 26 
 Self induction, alternating current 
 
 method, 225, 227 
 Anderson's method, 204 
 inductions, comparison of, 205 
 Shunts, galvanometer, 157, 171 
 Shear modulus, 51 
 Signs (mirrors and lenses), 125 
 Slide wire bridge, 154 
 Sound velocity of, 119, 122 
 Specific gravity bottle, 33 
 heat, of gases, 91 
 (table), 244 
 of metals, 85 
 (table), 241 
 of miscellaneous substances 
 
 (table), 241, 242 
 inductive capacity, see dielectric 
 
 constant, 
 resistance of electrolytes, 188, 252 
 
 of metals, 171, 252 
 rotatory power (table), 250 
 Spectrometer, 130 
 Spherical aberration, 138 
 Spherometer, 16 
 
258 
 
 INDEX. 
 
 Standard cells, 159. 
 Stroboscopic disk, 44 
 Surface tension, measurement of, 60 
 table, 243 
 
 Tangent galvanometer, 222 
 Telescope, adjustment of, 25 
 
 magnifying power of, 142 
 
 resolving power of, 145 
 Temperature coefficient of expansion, 
 
 7*> 74 
 
 of expansion (tables), 240-242 
 
 of resistance, 186 
 
 (table), 252 
 
 Thermal conductivity, 102 
 Thermocouple, 91, 116, 223 
 Thermometer, air, 76 
 
 Beckman, 65 
 
 calibration of, 67-7 1 
 
 fixed points, 69 
 
 platinum, 116 
 Thomson's double bridge, 181 
 
 method (galvanometer resistance), 
 
 method of mixtures, 202 
 Time of vibration, method of coin- 
 cidences, 38 
 method of passages, 54 
 reduction to infinitely small arc, 
 
 239 
 
 signals, 25 
 
 Torsion, modulus of, 51 
 Trigonometrical functions (table), 238 
 Tuning fork, pitch of, 44, 1 20 
 
 Units, electrical, 162 
 
 Vacuum, reduction of weighing to, 24 
 
 Valson's law of moduli, 31 
 
 Vapor pressure of mercury (table), 248 
 
 of water (measurement), 80 
 
 (table), 245 
 
 Velocity of sound, Kundt's method, 
 122 
 
 resonance method, 119 
 Vernier, 13 
 
 caliper, 14 
 
 Virtual image, 136, 139 
 Viscosity, measurement of coefficient 
 of, 56 
 
 table, 243 
 
 Voltameter, copper, 220 
 Voltmeter, calibration, 196 
 Volumenometer, 32 
 Water, boiling-point (table), 246 
 
 density (table), 240 
 
 equivalent, 89 
 
 vapor pressure, 80, 245 
 Wave length, of electric waves, 230 
 
 of light waves (measurement), 147 
 (table), 249 
 
 of sound waves, 119, 122 
 Weighing, by oscillations, 22 
 
 double, 24 
 
 reduction to vacuum, 24 
 Weight thermometer, 76 
 Weights, calibration of, 27 
 Weston (cadmium) cell, 160 
 Wet and dry bulb hygrometer, 84, 247 
 Wheatstone's bridge, 153, 169 
 
 Young's modulus, by bending, 47 
 by stretching, 46 
 table, 242 
 
THIS BOOK IS DUE ON THE LAST DATE 
 STAMPED BELOW 
 
 RENEWED BOOKS ARE SUBJECT TO IMMEDIATE 
 RECALL 
 
 LIBRARY, UNIVERSITY OF CALIFORNIA, DAVIS 
 
 Book Slip-Series 458 
 
167U18 
 
 Duff, A.W. 
 
 Physical measure- 
 ments . 
 
 Call Number: 
 
 QC37 
 
 D8 
 
 1910 
 
 QC37 
 
 167418