543
-78

U45

UC-NRLF

INSTRUCTIONS

OBSERVING

TOTAL SOLAR ECLIPSE

1878.
/

ttENT

ISSUED BY THE/U. S. NAVAL OBSERVATORY

WASHINGTON:

GOVERNMENT PRINTING OFFICE.
1878.

INSTRUCTIONS

OBSERVING

THE

TOTAL SOLAR ECLIPSE

ULY 2(), 10J

818.

ISSUED BY THE/U. S. NAVAL OBSERVATORY.

* ' -_,-

WASHINGTON:

GOVERNMENT PRINTING OFFICE.
1878.

'72

UNS

OF CONTENTS.

Page
I. Observations with the Naked Eye _

II. To Photographers

III. The Photoheliograph .!..'! !o

IV. The Equatorial Camera

V. Telescopic Observations

VI. Spectroscopic Observations -_ 2O

VII. Polariscopic Observations

VIII. Photometric Observations

IX. Therino-Electric Observations

76<)

These instructions were, by my direction, written by Prof. William Harkness,
United States Navy.

Suctions I, II, V, and Vlll describe such observations as can be made with the
apparatus usually possessed by amateurs ; while the other sections relate mostly to
observations which can only be carried out by persons who are able to command
expensive apparatus, and who are' skilled in astronomy and physics.

JOHN RODGERS,
Bear-Admiral, U. S. N., Superintendent.

INSTRUCTIONS FOR OBSERVING THE TOTAL SOLAR ECLIPSE OF

.JULY 29, 1878.

The following suggestions are intended to direct the attention of persons who
may witness the total solar eclipse of July 29, and who may desire to co-operate
with the United States Naval Observatory, to some of the phenomena which, in the
present state of science, it is most desirable should be carefully observed on that
occasion.

In writing these suggestions, much assistance has been derived from the instruc-
tions issued to the government eclipse expeditions sent out by various nations during
the last ten years ; and free use has also been made of the reports of these expeditions.

SECTION I -OBSERVATIONS WITH THE NAKED EYE.

i. At all places on the central line of the eclipse, the following phenomena will
be seen during totality, if the sky is clear, viz: Just as the sun's last ray fades
out, some glowing points of light will start into view, hanging upon the edge of the
black moon and glistening like rubies ; while surrounding the whole will be seen a
halo whose mild radiance reminds the beholder of the glories with which the painters
of old adorned the heads of saints. The glowing points are the red prominences, and
the halo is the corona ; both of which are now known to be solar appendages. The
corona consists of a comparatively bright inner part lying close to the sun, surrounded
by a much fainter mass of luminous matter of vast extent, and generally of most
irregular form. Accurate drawings of it will be exceedingly valuable, and for-
tunately, inaccuracy, such as affects the scientific value of the drawings, can be easily
avoided. With this view, persons intending to make sketches should provide them-
selves with a sheet of paper about 9 inches wide by 1 2 inches long, having upon it a
black disk, i^ inches in diameter, to represent the moon, from the center of which
straight lines are drawn at angles of thirty degrees, the whole being an exact copy of
Plate I. Let a weight be suspended by a string in such a position that the observer
can see it hanging over the sun's center ; and let the diagram upon which the draw-
ing is to be made be so placed upon any convenient stand that the line marked " top "
" bottom" shall be in the plane passing through the observer's eye, the string, and the
sun's center, the end marked "top" corresponding to the top of the string. Also, let
a lighted lantern be ready to illuminate the paper, in case the light of the corona is
insufficient. Finally*, two or three minutes before the beginning of totality, let the
observer close his eyes and cover them with a handkerchief, so as to make them as
sensitive as possible. A friend must be ready to warn him when totality begins, and
then, removing the handkerchief, he must rapidly sketch in the outline of the faint
outer parts of the corona, taking special care to preserve their correct proportions

relatively to the moon, and above all being scrupulously particular to place the
prominently marked branches and rifts in their proper positions. In doing this,
-real assistance \\ill lie afforded by the line- drawn upon the diagram from the center
of the black disk. Having completed the outer portions of the corona, the outline of
the bright inner part should be sketched : and in doing this it will probably be best
to view the eclipse through a spectacle glass of a light green tint. Finally, if any
time remains, the whole drawing should be carefully revised; with the aid of an opera-
Lj-lass. if possible.* It is not worth while to insert the red prominences. The duration
of totality being only three minutes in the most favorable localities, it is useless to
attempt making more than a single drawing: in fact, such an attempt would show
that the work had been so hurriedly done as to deprive it of all value.

Plate II is a drawing of the eclipse of April 16, 1874, given by .Mr. K. .). Stone,
director of the Royal Observatory at the ( 'ape of ( iood Ilope.f It shows the style in
which the drawings should be made : <t, , ", being the outline of the faint outer portion,
and b, l>, //, the outline of the bright inner part, of the corona. Probably many parts of
the outer corona fade away so gradually that it will be impossible to say precisely
where they terminate. This uncertainty should be indicated by a serrated outline,
as shown at a'. The original sketches must not be altered after the totality is over:
but if the observer desires to make a finished drawing from memory, it may In-
done upon another piece of paper.

2. Portions of the corona may probably be seen for a few seconds, both before
and after the totality. Note the exact positions and appearances of these portions,
observing particularly whether the parts seen before totality are the same as those
seen after totality, and how long they are visible before and after totality.

3. Note the character of the extreme outer boundary of the corona. hoes it
fade gradually away, or has it a definite edge f

4. The moon's shadow should be bordered by diffraction bands; and some ob-
s( rvers claim to have seen them as dark shadows -living over the earth's surface imme-
diately In-fore and immediately after totality. Probably they would be most \isible
from an eminence commanding an extensive landscape, and it would be a matter of
interest to look for them.

5. As the truth of Leverrier's discovery of an apparently unexplained motion of
the perihelion of Mercury is now established beyond all doubt, it is important to renew
the search for an intra-Merciirial planet or planets. With this view, the parts of the
ecliptic near the sun should be carefully examined during the totality of the coining
eclipse, and if an\ thing hitherto unknown is discovered, its bearing and distance from
the sun should be accurately measured. The bearing can be defined by holding a
watch between the eye and the corona, and giving the hour which points toward tin-
object, it being understood that the XII is toward the zenith. The distance should be
li'iveu in inches and parts of an inch, as measured by a rule held exactly eighteen inche s
from the eye To insure the rule being <o held a string should be tied to it, in which

' Tin- out, -i i-oron:i i- MI t'.iiut that (liDrrnnl i>i'i>,.ii< may MM- it i|iiiti- ilitl'rri-ntl.v. It i* I'M llir )>iir|iosi- of iliDiiuJHli-

illj- till-. |IITSOM.-|]||\ tliat till- IIM- III' III,' ll|H-):l nla~^ I- HTIMIMIICIIlll'll.

Mi-moils iil'lhr l.'\:il Aitroiiiiniiriil Soi-irty. Vol. I'.', p. 1:1. Mr. Slime uilvnratril t hr mi-tlioil nl' i-Ui-l. Inn.; whirl)

.-; lii'i u ,1. -i-i ilii-il.

there is a knot at a point just eighteen inches from the rule. By holding this knot
between his teeth, and keeping the string stretched, the observer will be sure that the
rule is at the proper distance from his eye, and then he has only to be careful to hold
it at right angles to the string, lint one eye should be used in making the measures,
and no attention need be paid to any object more than five and a half inches from the
sun. Mercury will be quite conspicuous at a distance of seven and a half inches, and
Mars will be seen at about five and a half inches.

6. Very good observations of the third and fourth contacts, the difference between
which is the duration of totality, can be made by the naked eye; the only instrument
requisite being a watch having a seconds hand. The importance of such observations
and the method of making them are full}' explained in Section V.

7. Note what part of the sky is darkest during totality, the form of this dark
part, and how the light varies from it to the brighter parts.

SECTION II. TO PHOTOGRAPHERS.

Photographs of the corona of very considerable scientific value can probably be
made with the apparatus in ordinary use by photographers. With this view, the camera
should be fitted with the largest and longest focused portrait lens at the operator's dis-
posal, and then, the diaphragm being removed, it should be most carefully focused
upon some distant object. The image will be well defined only over a small space in
the center of the ground glass, but that is of no consequence. Two or three minutes
before totality, some half-sized plates, prepared with the most sensitive chemicals ob-
tainable, should be in readiness for immediate nse. The number of these plates should
be equal to the number of half minutes in the predicted duration of totality at the
place occupied by the operator. For example, if the predicted duration is 2 m 30", then
five plates should be prepared. The camera should also be in readiness, directed to
the sun, and held there in any convenient manner. As the sun will be high in the
heavens, the ordinary camera stand may riot be available, but with a little ingenuity
some rough contrivance can be made which will answer the purpose.

As many of the plates as possible should be put in holders and laid beside the
camera, so that they can be used without delay. Then, the focus of the camera re-
maining as it was when adjusted upon the distant object, as soon as totality commences,
and the operator has assured himself that the image of the corona is in the center of
the ground glass, he should insert a plate and expose it. The ground glass should
next be replaced, when, owing' to the diurnal motion of the earth, it will be seen that
the image of the corona has moved away from the center of the field. It should be
piit back, after which another plate should be inserted and exposed; and this routine
should be continued till the totality is over, when the plates can be developed. With a
quick-acting portrait lens, the first plate should be exposed three seconds, the next six
seconds, the next three seconds, and so on, alternately. With an ordinary portrait
lens the exposures should be alternately five and ten seconds; but it must be under-
stood that diaphragms are not to be used in any case. The development must be car-
ried far enough to bring out all the details ; but extreme care must be taken not to push
it solar as to clog up the delicate outlying portions of the corona. On account of the

motion of the earth while the plate is liein^ exposed, the intake of the sun or rather
of the moon, which covers the snn will he a little blurred, and somewhat oval in
shape; and to prevent disappointment, it must lie borne in mind that the portrait lenso
in ordinary use have such short foci that they can yive but very small pictures of the
moon. The actual si/e in inches will be very nearly as follows, viz: With a 1-4
lens, 0.04; with a 1-2 lens, 0.07; with a 4-4 lens, 0.09 : and with an extra 4-4
lens, 0.14. The diameter of the extreme outlying portions of the corona may be four
or five times these figures.

It is hoped that photographers making pictures in this manner will send the
original negatives to the Naval Observatory; but, if they prefer to retain them, it is
earnestly requested that they will, at least, send positive copies on <rlass. 1'aper
prints are useless for scientific purposes. Each plate must be accompanied by a state-
ment of the kind of lens used in its production; the focal distance and aperture of
that lens; the chemical process employed in the dark room ; whether the bath was
neutral or acid; how lou<r the plate was exposed; and the means of identifying its
top. The latter point is of special importance, and the upper side of every plate
must be carefully marked when it is put into the holder, because the character of the
corona is such that without this precaution it will be impossible to say certainly which
is the top of the negative.

SKCTIOX III. Til K I'lIOTolIFLIooKAIMI.

Ft. is known that very accurate positions of Mercury and Venus relati\ely to the
sun's renter can be obtained from photographs of the latter body taken during transits
of these planets; and it inav be inferred that equally accurate positions of the moon
ini^lit be obtained from photographs of the partial phases of a solar eclipse. I n for-
tunately such is probably not the case. The clearness of the atmosphere, the sensi-
tiveness of the chemicals employed, the length of exposure of the plate, and the hour
of the day, all affect the photographic diameter of the sun to such an extent that it is
practically impossible. to make a series of pictures in which it shall be even approxi-
mately constant. In the case of transits of .Mercury or Venus this is of no consequence,
because it mav be assumed that the change of diameter takes place uniformly all
around the circumference; and having a complete outline of the sun, it is always pos-
sible to find its center with accuracy. In a similar way it is possible to find the center
of the planet, whose diameter is also variable, and thus the positions of the two bodies
relatively to each other become known. When the diameter of a celestial body is
unknown, at least a semi-circumference of its photographic ima^e is necearv from
which to find the exact position of its center. In the case of a solar eclipse the diam-
eter- of the snn and moon are nearly equal, and therefore any picture which shows a
larire part of the circumference of 01 ..... f these bodies necessarily shows but a small
part of tin- circumference of the other. For that reason it seems impossible to find
accurately the center of more than one of the two bodies : and consequently it seems
impossible to find accurately the positions of the two bodies relatively to each other.
On the other hand, contact observations afford a very accurate means of finding the
error- of the lunar tallies : and as Mich observations can be made with ordinary instrn-

ments, at a comparatively trifling cost, it does not seem wise to incur the trouble and
expense of erecting more than a single experimental horizontal photoheliograph
specially to observe an eclipse. At [daces where such instruments are already in
position it may be well enough to use them, but it is extremely doubtful if the pic-
tures will be of much value.

SECTION IV. THE EQUATORIAL CAMERA.

Good photographs of the corona will be exceedingly valuable, but they require
a special form of apparatus for their production. That best adapted to the purpose is
a camera, fitted for using either large or small plates, and furnished with a quick-
acting objective, the whole being equatorially mounted and driven by clock-work, so
that long exposures can be readily made. To resist the strain consequent upon the
manipulation of the plates, the mounting and driving clock must be of the most sub-
stantial kind; and to depict the fainter parts of the corona the objective must be as
rapid as possible. In the latter matter our choice will be practically restricted to the
lenses named in the following table, which embraces all the most rapid varieties in the
market. The names of these lenses, and the corresponding intensity ratios, are varied
somewhat by the different makers. Those given in the table have been taken from
Dalhneyer's catalogue.

-_ .

o
a

.- rt

3 C
~ ^

Description o! Photographic Objective.

.2 a

"5 3
Cfi

Q " J

v- O

3 ^

t/l

O ^.

o Cr

-3

Indus.

/itt'/ies.

s. s.

o 3 to I 6

Quick Acting Portrait

. 126

0.7 to 3.6

Ordinary Portrait

.224

1.3 to 6.4

Portrait and Group (D)

,',

.308

2.9 to 14.4

Rapid Rectilinear

33 i

o . T i ?

^ . i to 25 . 6

If F is the equivalent focal distance of a photographic objective ; d its working
aperture; C the exposure constant, whose value depends upon the intensity of the
light and the sensitiveness of the chemicals employed; ami the time of exposure

required to produce a good negative; then, the intensity ratio is , and

The data from which to determine an approximate value of (' for the corona are
very limited. In August, 1869, M'r. J. A. Whipple obtained a negative of the corona
at Shelbyville, Ky., by placing the sensitive plate at the principal focus of an
ordinary astronomical telescope. In this instrument the value of i'-^tl was 16.4, and
as the exposure was 40 seconds it results from the formula above that (7 330.149. In

December. iS70, Mr. A. Mrothers. ;it Syracuse, Sicily, obtained a very tliin negative
of the corona, with a Dalhmver rapid rectilinear lens, in S seconds. For this objective
/'-=- d= 8, and consequently, (' = 0.125. During the same eclipse, Mr. (). II. Willard,
at ,lere/., Spain, obtained a negative of the corona in 90 seconds. He used an astro-
nomical tele-cope, precisely as Mr. Whipple did in August, 1869. If we assume /'-=-</
to lie 15, as is usual in such instruments, then ^'1= 0.400. The negative did not seem
overexposed. In December, 1871, .Mr. ,1. IJoesinger, at Ootacamund, India, olitained
a negative of the corona in 3 seconds \vith a Dallmeyer No. 6 D portrait and group
lens. In this case /'-=- <l =. 6, and, therefore, ('=. 0.084. In August, 1869, Dr. Kdward
Curtis, at Des .Moines, Iowa, used values of (' which were respectively O.Oi6 and
O.Oii ; but on that occasion the actinic force of the sun was diminished from ten to
tueiity times by a dense haze, and therefore these values were not equivalent to more
than O.OOI6 and O.OO 1 1 in an ordinarily clear atmosphere. With the lirst exposure
the prominences and the lowest layer of the chromosphere were well shown. With
the second exposure the prominences were still depicted, but the chromosphere only
appeared around part of the sun's circumference. The corona, properly so called,
was entirely absent from both pictures.

From all these facts it is probably safe to conclude that, with a clear sky and a
moderately high sun, exposures in which the value of C"is about 0.002 will give only
the prominences and the outline of the moon. When C becomes o.oS the corona will
begin to appear, and will increase in extent as the exposure increases, at least up to
the point where C becomes 040. Beyond that 1 am not aware that we have any ex-
perience. Accordingly, the shortest exposure specified in the table above corresponds
to G' 0.08, and the longest to (' 0.40.

Jn choosing a lens from those mentioned in the table, Xos. i and 2 may be at
once dismissed from consideration, as giving too small images of the sun. Xos. _j and
5 are practically of the same focal length, and as both of them are entirely free from
distortion and flare, No. 4 should be preferred on account of its larger aperture.
Thus our choice is narrowed down to Nos. 3 and 4, No. 3 having the greater value of
/'-=-'/, but neither so large nor so Hat a field as No. 4. If the lens were required
solely for photographing the corona, it might be dillicult to choose between these two,
a- either of them will cover a plate far larger than can ever be required for that p;,r-
pose ; but it will presently appear that there are questions in eclipse photography for
the solution of which plates of the largest si/.e must be used, and on that account it is
dearly best to choose No. 4.

Having adopted a lens of 33 inches focus, we have next to consider what si/.e of
plates should be used with it. The diameter of the sun's image will be 0.31 of an
inch, and, taking .Mr. Urothers's picture of the eclipse of December, iS;o, as a guide,
the diameter of the corona may be live or six times as great, or. in other words, about
1.7 inches, to receive which a plate measuring 4 '. ; by 5' . inches will lie ample. Hut
an attempt should be made to use the h-ns for another purpo>e. e\ eu more important
than photographing the corona. The magnitude of its intensitv ratio enables it to
depict taint objects rapidly, and the extent of its angle of \ie\v is such as to embrace
a field of more than forty degrees. It. therefore, >eeius peculiarly fitted to assist in

the search for intra-Mercurial bodies, provided the illumination of the sky does not
prove too great. Mercury's maximum elongation from the sun scarcely reaches
twenty-nine degrees, and, if we postulate Bode's so-called law, the greatest elongation
of the hypothetical intra-Mercurial planet or planets should not exceed fifteen degrees.
Double this distance is the space along the ecliptic which must be searched; and to
photograph it a plate 17.7 inches long will be required. Our lens will cover a plate
measuring 20 by 22 inches, but it is very desirable to keep the apparatus as light as
possible, and on that account it seems best to have a camera twenty inches square,
and to use plates measuring i 7 by 20 inches, which will suffice to cover a space of
thirty-three and a half degrees along the ecliptic.

Assuming the adoption of an equatorial camera twenty inches square, provided
with a lens whose intensity ratio is one-sixth and whose focal distance is about thirty-
three inches, it yet remains to consider how this apparatus should be managed during
a totality lasting only three minutes. As the illumination of different parts of the
corona varies greatly, there can evidently be no certainty of getting all the details of
the phenomenon unless a series of plates are taken, in which the exposures vary from
the shortest possible up to the point where it is certain that an increase of time does
not improve the picture. On. this account it will be desirable to take as many as six
plates, the exposures being, respectively,

3 s , 5 s . io s , 2O S , 40", and 6o s .*

The first four of these plates will receive such short exposures that it is unlikely they
will show anything but the corona, and therefore their size should be 4^ by y/ 2 inches.
With the last two plates the case is different. Their size should be 17 by 20 inches,
because their longer exposures will probably suffice to bring out upon them any bright
points which may exist within their field. A lens such as is here under consideration
should depict an eighth-magnitude star. in about one minute, but of course the inten-
sity of the sky-illumination during totality will determine the limit of brightness at
which faint luminous points will cease to impress themselves upon the negatives, and
what this limit may be it is impossible to predict, The necessity for at least two
large plates is evident when it is remembered that the image of a small bright point
could not be distinguished from an accidental blemish in the film, and it wquld only
l)e by finding it upon both plates that its true character could be unmistakably recog-
nized. It is exceedingly desirable to determine accurately the maximum exposure
that the corona will bear with advantage, and it is hoped that on at least one of the
large plates it will [trove to be over-exposed.

For the management of the equatorial camera two persons will be required, desig-
nated respectively as No. i and No. 2, and the following routine will be found conven-
ient. The operations in the dark room must be so managed that the six sensitive
plates shall be placed beside the camera, in readiness for instant use, a minute or two

' It is questionable whether or not all these exposures can be made in tbiee, minutes, but a few preliminary trials
will settle that point. Experiments should also be made upon the stars in the evening twilight to ascertain if an ex-
posure so long as sixty seconds is really required. In any event, it will probably be best to reverse the order given
above, and to work oft' first those plates requiring most time, thus reducing to a minimum the damage which would
accrue from the sun bursting out while an exposure was in progress.

tin- predicted time of Mvoiid contact. Six sand-glasses must also lie pro-
vided, each adjusted to run for the number of seconds that the plate to which it
belongs is to lie exposed. Then, the cap being on the lens, the first plate must he
placed in the camera, and the slide must lie withdrawn from the holder: after which
operator No. i will take position with his hand upon the cap covering the lens, and
No. 2 will seize the proper sand-glass and hold it in such a way that the sand shall
all he in the bottom of the glass. At the instant totality begins, No. I will say,
" Ueady?" No. 2 will answer, "Ready." No. I will begin the exposure of the
plate bv careful! v removing the cap from the lens, exclaiming as he does so, "Turn;"
when No. 2 will instantly turn the glass. Moth operators will stand motionless
while the sand is running, No. 2 watching carefully for the escape of the last
grain, when he will exclaim, "Out," and No. i will instantly terminate the expos-
ure by capping the lens. Without a moment's delay No. i will push the coyer into
the plate-holder, remove the exposed plate, place the next holder, containing a fresh
plate, in the camera, and draw the slide; being assisted in all this by No. 2, if neces-
sary. Then No. i will again take position- with his hand upon the cap of the lens;
No. 2 will seize the second sand-glass; and this plate will be exposed in the same
manner as the preceding. This routine will be continued till all the plates have been
exposed, and after totality is over they n'ill be developed in the usual way.

With regard to the plate-holders, it may be remarked that the last two plates
must necessarily he in holders as large as the camera, but the first four plates can he
more conveniently handled if they are in half-sized holders. To render such an
arrangement possible a reducing frame must be provided, which, upon being applied
to the camera, will reduce its opening to that proper for receiving the small holders.

As to the photographic process to be employed, if only small pictures are desired,
dry plates, prepared with washed emulsion, are recommended; but if any large
plates are to be taken the wet collodion process will be preferable. In either case it
is specially important to use well bromized collodion, because the bulk of the coronal
light has probably a wave length of 5315.9 (the 14/4 line), and to such light silver
bromide is far more sensitive than either silver chloride or silver iodide. For wet
plates there is reason to think that fineness of deposit and richness in details can be
best secured by a free use of sugar in the iron developer.

To facilitate the study of the red prominences, it has been proposed to photograph
them on a scale of ten seconds of arc to a millimeter. The optical apparatus for the
production of such pictures must have an equivalent focal distance of 2062.7 centi-
meters, or 8 i 2. i inches, and if we take (' equal to O.OO2, which is probably very near

the truth,* the values of t for lenses of various apertures can be found from the
formula already given, and will be as follows :

l _ i

O o

'+ *r-t

2 -

3 H

c c *

3 o

( J

in v "
3

. o

'-* * .

V (? /

0. f

<u ^

o *

Indies.

J".

135-3

18306

36.6

20.1

IOI.5

10302

20. 6

H-3

Si. 2

f>593

13-2

7-2

67.7

45S3

9.2

5-0

54-1

2927

5-8

3-2

40.6

1648

3-3

1.8

31-2

973

1.9

I.O

The last column of the table gives the moon's motion during an exposure of the
length specified in the preceding column. A prominence one minute high is a large
one, and yet it could scarcely be photographed with a six-inch objective, because twenty
seconds of its height would be covered by the advancing moon before the exposure
was over. In short, it does not seem possible to photograph prominences during
eclipses on the scale here contemplated with an aperture much less than ten inches,
and twelve woiild be far better. Under these circumstances the practicability of the
scheme is questionable.

SECTION V. TELESCOPIC OBSERVATIONS.

Ti'lexcojH's. A proper telescope for observing the eclipse is very desirable, but
any telescope at all is better than none ; and owners of small spy-glasses, or opera-
glasses, should bear in mind that their instruments are capable of doing good service.
( )pera-glasses can be held in the hand with sufficient steadiness, but spy-glasses, how-
ever small, will require a stand to produce the best effect. This need not be an
elaborate affair, but may be made of wood in the simplest manner, provided only that
it is capable of holding the telescope steadily pointed in any required direction. Some
artificial support is absolutely necessary, and, in default of anything better, the tele-
scope might even be lashed to a round post.

Sh(i</<' (ildSM'.i. Fit anticipation of the late transit of Mercury, several letters Avere
received at the Naval Observatory inquiring about shade glasses, and on that account
it seems desirable to notice the subject here. The best form of shade is undoubtedly
.a long wedge of colored glass, achromatized by a similar wedge of white glass firmly
cemented to it. The most usual, and by tar the most objectionable, color for a shade
is red. Either neutral tint or green is excellent, but it is difficult to obtain the latter
sufficiently dark. Deep blue has been advocated by some observers, and dec}) }'ellow
was much used in Germany, but is now apparently losing favor. Shades consisting
of a single piece of thick glass are very apt to split, and thus endanger the observer's

*Tho exposures given by Mr. Do la Rue to his pictures of tho eclipse of July iS, 1860, indicate a value of C some-
where between 0.0007 all( l 0.0012 ; while Dr. Curtis's exposures on the eclipse of August 7, 1869, give (' =0,0016.

eye, it' tin- rays of ;i bright sun arc concentrated upon them Ity an objective of lame
aperture. To avoid this, such objectives must have their aperture reduced by a cap,
t\\n inclio hciujr tho proper si/.c for a telescope of thirty inches focus, and three inches
for one of live or six feet focus. My making the shade of three thicknesses of "-lass,
the piece next the eye heiny thickest and darkest in color, the other pieces bciii^-
successively thinner and lighter in color, and all hcinir fitted loosely into their cell so
that they can expand freely, it will be possible to use with safety an aperture of'five
inches upon a telescope of six feet focus. When no proper shade is at hand, an
excellent substitute can be made by smoking a slip of thin Hat ii'lass, about an inch
wide and four inches Ion** 1 , in such a way that its tint may increase uniformly, from
very li-i'ht at one end to a d'-pth at the other end stillicient for viewing the brightest
sun without inconvenience. The smoked surface should be protected b\- a covering
jjlass, of the same si/.e as the original slip, and separated from it by bits of thin card
inserted at the ends, the whole heiii^- bound together by a narrow strip of paper pasted
around the ed^es. For observing the second contact of the eclipse, a neutral tint, or
li-reeii. spectacle <, r lass will be found convenient,

Of course it will be understood that the necessity for caps to diminish the aper-
ture is confined to common telescopes. Instruments intended for observing the sun
are always provided with arrangements for ^ettin- rid of the excessive li^ht and heat
without diminishing the aperture, and often without usin^ shades

lli'dthif/ <>f TelescojH'. When observing a brio-lit sun the eye end of the telescope
soon becomes quite hot, and thus air currents an* set up in the tube which interfere
seriously with the definition. To remedy this the tube must occasionally be cooled,
either by shading the objective or by turning the telescope away from fly sun As
it is important to have the best possible definition while observing the contacts and
totality of the eclipse, special care must be taken that the temperature of the lube at
these times is the same as that of the surrounding air.

Tlic ( oittm-lt;. The first touch of the moon upon the sun's limb is technically
called the first contact, the complete extinction of the last rav of sunlight, is simulta-
neous with the second contact, the reappearance of the minutest possible portion of
the sun marks the third contact, and tho disappearance of all trace of the moon from
the sun's disk constitutes the fourth or last contact. The first and fourth contacts are
sometimes called the external, and the second and third the internal, contacts. The
observation of these contacts consists in noting the exact instants of their occurrence,
and for that purpose a time-piece with a seconds hand and a telescope are usually
employed, hut if the latter instrument is not obtainable, the internal contacts can be
well observed with the naked eye.

There is generally some uncertainty about the first contact, because the only evi-
dence of its occurrence is a slight indentation in the sun's limb, which, unless the ob-
server knows precisely where to look for it, naturally attains some si/e before it is seen,
ami thus the contact is noted too late. If the telescope has a position-micrometer, its
wire may be set at the computed point of first contact, and then there \\ill be no dilli-
culty ; but in the absence of such an arrangement, the only way is to use a power of not
more than sixty diameters, and to watch carefully a portion of the sun's circumference

large enough to include with certainty the point in question. The shade-glass enl^-
ployed should be sufficiently dark to destroy all glare, and to permit of the sun's im-
age being viewed without effort.

The internal contacts are phenomena of such a definite character that the instants
of their occurrence "can be noted within a small fraction of a second. For a minute
or two before the predicted time of second contact the sun's thin, and now fast wan-
ing, crescent should be carefully watched through a neutral tint or green shade glass
as light as the eye can conveniently bear. Presently the crescent will become a mere
thread of light, which will rapidly shorten, and suddenly disappear. This is the second
contact, and there can be no mistake about the exact instant of its occurrence because
the ruddy glow of the chromosphere, which remains at the point last covered by the
moon, is totally different from the dazzling white light of the photosphere which has
just disappeared. Through a red shade glass the contrast between these two kinds of
light might be partly lost, but through a neutral tint or green glass it will be suffi-
ciently apparent. Finally, there is yet another indication of the second contact, and
that is the seemingly miraculous appearance of the complete outline of the moon,
round and black, reposing upon the wondrous radiance of the corona.

The approach of the third contact will be heralded by the rapid brightening of
the chromosphere at the point of the moon's limb where the sun is about to reappear,
and two or three seconds later a sudden burst of light will announce the contact itself
and with it the termination of totality. This contact may be observed without any
shade glass.

The observation of the fourth contact is a very simple matter. The segment
cut out of the sun by the retreating limb of the moon is carefully watched as it be-
comes less and less, and the instant of its final disappearance is noted as the fourth
contact. The shade glass used must be so dark as to destroy all disagreeable glare.

To observe the contacts, a station must be chosen where there is an uninterrupted
view of the sun, and the party must consist of only two persons, one of whom must
have a time-piece and writing materials and the other must have a telescope. With
regard to the latter instrument, it may be remarked that if it is a common one a spy-
glass, for instance there will probably be no choice as to its magnifying power; but
if it belongs to the better class, and has a battery of eye-pieces, the power chosen
should be equal to fifteen or twenty times the diameter of the objective in inches. To
secure the best results, the party must be entirely alone; and above all, it must not
Ite within hearing of other observing parties. A minute or so before the expected
time of contact the person with the time-piece will fix his eyes upon its seconds hand
and begin silenth* to count the seconds, while the observer at the telescope will care-
fully watch the sun. Upon the occurrence of contact the observer will cry "Mark,"
and the person at the time-piece will immediately write down, first, the second; then,
the minute ; and, finally, the hour. Before beginning to observe, care must be taken
to see that when the seconds hand points to 60 the minute hand is exactly over a
minute mark; and in observing, if the seconds are near sixty, special care will be re-
quired to avoid recording a wrong minute. For example, when the time is 4"' 58" the
minute hand will be almost exactly over 5"' ; and special care must be taken not to
3

record 5 58* instr;id of 4 58'. Practice in calling and noting time previous to the
eclipse will greatly diminish the risk of such mistakes. A little before totality a lan-
tern should be lighted, because without it there may be difliculty in seeing the hands
of the time-piece at the approach of third contact.

Observations of the internal contacts can be utilized without a knowledge of the
error of the time-piece ; but with the external contacts the case is dill'erent. They
will be of no value unless the error of the time-piece is accurately known, either on
local or on Washington time. To meet tliis difliculty, it is hoped that arrangements
may be made by which the Western Union Telegraph Company will furnish, free of
cost, a signal from the Naval Observatory, indicating the instant, of Washington mean
noon, to all proper persons who may apply for it at any of their ollices on July 27,
29, and 30. The intending observer can best utili/e these signals by making sure
that the minute and seconds hands of his watch arec with each other, as explained
above, and then carrying the watch to the telegraph otlice on each of the days men-
tioned and noting the hour, minute, and second indicated by it when the signal is re-
ceived. No matter how the watch may run, it must not. be meddled with in any wav,
either by moving its hands or touehing its regulator, between July 26 and 31. With-
out this precaution, neither its rate, nor the degree of dependence to be placed upon
it, can be determined.

The final record of a set of observations of contacts should contain the following
information :

1. The name of the station; including town, county, and State.

2. The date.

3. A description of the location of the station with reference to the nearest promi-
nent landmarks; as, for example, its bearing and distance from the nearest court-house,
railroad station, or church ; or its position with respect to the lines of the I'nited States
land surveys; or the bearing and distance of a prominent mountain peak; or the
bearings of three or more such mountain peaks; or, in the case of a citv, the street
and number, together with the names of the nearest cross streets on each side.

. 4. A description of the time-piece, and a statement of the aperture, focal length,
and magnifying power of the telescope emploved.

5. The times indicated by the watch at the reception of the Washington noon
signal on July 27, 29, and 30.

6. The time of each of the contacts; just as read from the face of the watch,
without the application of any correction.

7. An estimate of the uncertainty to which these times are liable.
S. Any remarks which may be thought necessarv.

9. The signatures of the two observers, thus: A. 1!., observer with telescope; ('
1 > .. time observer.

It is particularly requested that the original pencil memoranda of the times of
contact lie s,. n f to the Naval Observatory, inclosed in the above-described report.

If several parties observe the eclipse in the .same neighborhood, it is quite likely
that their recorded times may differ a little. Such differences occur even in the work
of the most practiced astronomers; and thev must be allowed f<> remain. It is never

permissible to alter the record. A mere suspicion that sucli a thing has been done
will insure the entire rejection of the observation.

Limits of the Shadow Path. There is yet so much uncertainty in the best solar
and lunar tables that in the eclipse of July next the British Ephemeris places the path
of the moon's shadow, while crossing Colorado, about three and a half miles south-
west of the position assigned by the American Ephemeris. It is therefore important
to determine accurately the true position of the path, and this may be readily accom-
plished by observing the duration of totality at points situated from one to ten miles
within the shadow. As the duration of totality is the interval between the second and
third contacts, it is determined by observing these contacts in the manner already
described; except that for this purpose a knowledge of the error of the watch is unnec-
essary, and the observations may be made with the naked eye if a telescope is not
available. At places very near the edge of the shadow the totality may be so brief
that the time observer would risk losing the third contact if he stopped to record the
second. At such places it will be necessary to have a third person to make the record,
and then both contacts can be readily observed, even if they follow each other at an
interval as short as a single second.

. Sketches of the Corona. In making a general sketch of the corona the magnifying
power employed should not exceed from twenty to forty diameters, and some me-
chanical means should be adopted to insure accuracy of outline. Owing to the faint-
ness of the corona, it is doubtful if the camera lucida can be successfully used; and
the most promising plan seems to be a system of squares, ruled on glass, in the focus
of the eye piece of the telescope, and a corresponding system drawn upon the paper
on which the sketch is .to be made. It is known that spider lines in a micrometer
sometimes cease to be visible under the feeble illumination of the corona, and to avoid
the possibility of this happening with the squares, the "lines defining them should be
rather coarsely cut. More than once attempts have been made to use a telescope after
the manner of a camera obscura, by drawing out the eye piece a little, and thus
causing it to throw an enlarged image of the corona upon a piece of finely ground
glass, or semi-transparent paper, placed a few inches beyond ; the whole being so
arranged that the image could be seen through, and traced upon, the glass or paper.
Owing to the faintness of the image, such attempts have not hitherto been very suc-
cessful, but perhaps they might do better if for the plate of ground glass a thin sheet
of gelatine, or mica, were substituted, and the drawing were traced upon it with a
needle point. The proper mode of executing such sketches has been already described
in Section I.

Miit/i/c Examination of the Corona. With as high a magnifying power as the
telescope will readily bear, examine the structure of the corona; noting particularly
whether it is homogeneous, nucleated like a star cluster, or filamentous. If any nuclei
are seen, describe their shape, size, color, and mode of distribution. If filaments are
discerned, state their length, diameter, color, and mode of distribution ; whether they
are straight, curved, or contorted ; in what direction they lie relatively to the sun ; and,
if curved, specify how the curvature faces. Be sure to observe how the structure
varies in passing from the sun to the outermost visible limit of the corona.

Since the spectroscope furnishes an efficient means of studying tin- red promi-
nences at any time, it will be very umlesiraltle to waste a single one of the precious
nioineiits of totality in examining them.

Intr<i-MiT<:nrinl llotlic*. To facilitate the work of such astronomers as may desire
to .-earch for intra-Mercurial planets with considerable telescopic power, ;i chart will
lie found at the end of these instructions showing every star so large as the seventh
magnitude in that portion of the heavens which will be occupied by the sun on the
sgth of July next. The black circle in K. A. X 1 ' 36'", Dec. + 18 39' indicates the
position of the sun. Mercury, Kegulus, and Mars will be pretty close together, and
probably quite conspicuous during totality, but they are so far to the eastward that
only the last-named conies within the limits of the chart. Venus may also be seen,
but she will be low in the western sky. While looking-for planets, the possibility of
discovering a small comet, or a meteor stream, should be borne in mind.

I'dftia! J'/KISCS. The corona forms a luminous background upon which the moon's
limb is sometimes seen projecting beyond the sun; and a little before totality it is
even possible that the complete outline of the moon may become visible. Look for
these phenomena, and note the time of their occurrence. It is difficult to assign any
reason for the existence of rays, or brushes, of light at the cusps of the sun, lnit.it is
saiil thev have been seen. If any such appearances present themselves, they should he
earefullv scrutinized to ascertain if they change either their position or intensity ; and
the interior of the telescope should be examined to make sure that they do not ori^i
uate in reflections, either from the tube, or from the lenses

SKCTION VI. SPECTROSCOPIC OHSKUVATloNS.

This section has been largely compiled from the writings of Young and l.ockyer,
some of the paragraph* being almost in the exact language of these authors.*

Among the numerous- arrangements of spectroscopic apparatus devised by scientific
men, experience has shown that those mentioned below are best adapted to the obser-
vation of eclipses, viz :

Aini///:iin/ IiiNtritiiH'iit, A. A spectroscope having a long slit and a long collimator,
mounted at right angles to the optical axis of a reflecting telescope of large aperture
and short focus; the slit of the spectroscope lying in the focus of the mirror as usual
and the telescope being provided with a large finder. In this combination the ima^e
of the corona is very bright, and small enough in proportion to the length of the slit
to allow of the simultaneous visibility of the spectrum on both sides of the dark moon ;
while the long collimator permits the use of a wide slit, and thus insures the admission
of the maximum of light.

Iiitri/fiiliiii/ Inxtnitiiritf, 11. A spectroscope, having a long collimator, mounted so
that the light falling upon its slit is received from the eve-piece of a telescope of
large aperture and short focus, previously adjusted for distinct vision on a distant

" See tin- following: Keport on Observations of the Total Solar Kelipse of December 22, 1870. Jiy Pi of. (.'. A. YOIIII-J
U. 8. Coast Silrvpy Re|ort for 1X70, p. 141. Inst ructions for Observers at tin- Kn<;lisli (iovei mnent l-'.clipse Kxpedilion
1-71. Nature. 1X71, vol. 4, p. 517. Knglixli (ioveriiincnt Kelipue r.\pedilion, 1X75 Instniriions to Observers. Nairn.-,
875, vol. II, p. 351. Report on tin' Total Solar Keli|ise of April 6, 1X75. l!y .1. \. l.oekyer and Arllnir Schuster. I'liil.
Tiauit. Tbu Coming Total Solar Kelipse. liy .1. X. l.oekyei. Nature. |S;X, vol. 17. pp. 481 and 501.

object. The area of the heavens from which this telescope collects its light should be
three or four degrees in diameter, which will be the case if it magnifies about fifteen
times. To insure accurate pointing it must be provided with a finder.

Analyzing Instrument, 0. An automatic spectroscope of large dispersive power ;
attached to an equatorial telescope in the usual manner.

Aini/i/.i/i/(i Instrument, D. A spectroscope of moderate or small dispersive power;
attached to ci telescope as above described.

[utcyratiny Instrument, E. A spectroscope, either mounted upon a stand or held
in the hand, receiving the light of the corona directly upon its slit without the inter-
vention of any optical apparatus whatever.

Iitti'f/i-dfh/i/ litxtnimeiit, F. A meteor spectroscope, having neither slit nor colli-
mator, but receiving the light directly upon its prism.

The telescopes to which the above-described spectroscopes are attached should
nil be equatorially mounted ; and they must either be driven by clock-work, or 'an
assistant must be employed to point th'em. When a spectroscope slit is covered from
view the only available mode of bringing an object upon it is by means of the finder;
but if it is exposed, as it should always be, the image of the object upon the slit-plate
will be the best guide. For observing the corona the width of the slit may be such
jis to show but two of the three h lines when the spectroscope is directed to a faint
cloud. The instruments best adapted for making the observations suggested below
have in every case been indicated by the use of the reference letters employed in the
descriptions of apparatus just given.

/irttxltcs oflif/ht af ciixjtH. Some observers report having seen brushes of red light
at. the solar cusps during the partial phases of eclipses. It is difficult to understand
how such phenomena can occur, but if they present themselves their spectra should
be carefully examined. Instrument, C or D.

Yoiuiifs llei'ervhif/ Layer. -Close to the sun's limb look for the layer in which the
Fraunhofer lines originate. Just before totality it should give a nearly continuous
spectrum, and just at the instant of totality it should show reversed all the dark lines
of the spectrum, except those of terrestrial origin. The dispersive power of the spec-
troscope employed cannot be too high. The image of the sun should not be less than
one inch in diameter, and since the thickness of the layer in question does not exceed
2", it is evident that, extreme care must be exercised in adjusting the focus, in making
the slit very narrow, and in placing it precisely in the plane of, and rigorously tan-
gent to, the solar image. Instrument, C or D.

An attempt should be made to measure" the thickness of the reversing layer by
noting how long the lines continue bright. But for this purpose a chronograph capa-,
ble of indicating the one hundredth part of a second will be necessary.

The Prominence b^n'cfnnn. Just before and just after totality the atmospheric
glare is greatly reduced, and advantage should be taken of this circumstance to look
carefully for new lines. The search should be directed particularly to that part of the
spectrum lying above F, as it deserves, and will probably repay, careful study, and is
very difficult to deal with in the absence of an eclipse. Instrument, C.

During totality, attend to the following points:

In passing outward from the sun, are the li linos replaced by a diffuse band?.
Does the /' line extend higher than the other hydrogen lines .' Instruments, J, (', l>.

It !> known that the prominences consist mainly of hydrogen; but do they ever
contain large quantities of calcium .' In other words, are the // lines sometimes con-
spicuous in them .' Instruments, .1, (', I).

Probably the // line is produced only by very hot hydrogen. As giving a clew to
the temperature of the prominences, look for it in their spectrum, and note how high
above the photosphere it exists. Instruments, A, C, />.

Look for new lines in the ultra- violet part of the spectrum. Instrument, Cor />,
provided with a fluorescent eye piece.

The Cm -(iiin. Direct attention specially to the following points :

l>oes. the line 1474 stop short of the chromosphere, or does it enter that forma-
tion f If the latter, does it broaden out as it nears the sun :' Instruments, .1, <\ I).

Are there any parts of the corona which give a spectrum containing bright lines
between //, and 14/4 ' In trying to answer this question use rather a wide slit, and,
if any lines are found, locate the part of the corona giving them. Instrument, C or />.

IJesides the well-known bright line 1474, the corona gives a faint spectrum, which
most observers have thought to be continuous. Is it really continuous, or is it crossed
by the Fraunhofer lines ? Look particularly for />, 7-,', 1>, and /'', and if any of them
are seen note in what part of the corona. Instrument, (.' or I).

Examine the dark rifts in the corona very carefully to make sure whether or not
thev give any spectrum. Instrument, C or 1).

Does the corona give any ultra-violet lines ? Instrument, C or I), provided with
a fluorescent eye piece.

Examine the brightest parts of the outer corona in order to determine how far
from the chromosphere any spectrum is visible, and note also what that spectrum is.
Instrument, C or />.

Examine what part of the coronal spectrum is polari/.ed. This may be done by
noting the effect produced upon the spectrum when a Nicol's prism placed before the
slit is rotated ; but the results thus obtained will be complicated by the polarization due
to the passage of the light through the dispersing prisms, and to eliminate this the slit
must be placed at different angles to the vertical, and upon various parts of the corona.
The following seems a more satisfactory mode of proceeding: If it is borne in mind
that the polarization of the corona is partly radial to the sun and partly in the plane
passing through the zenith, the sun, and the observer's eye, it will be evident that the
parts of the corona cut by a vertical circle passing through the sun's center are polar-
i/.ed only in the plane of that circle. Hence, if the >lit of the spectroscope is placed
in that plane, and the image of the sun is allowed to fall upon it in such a wav that
the sun's center coincides with the axis of the slit; and if the Nicol's prism is placed
with its principal section at right angles to, and immediately before', or immediatelv
behind, the slit, then it is evident that all verticaliv polari/.ed light will be excluded,
and the spectrum visible will be due solely to the unpolari/.ed light of the corona.

Examine the thermal properties of the coronal spectrum. Eor this purpose a
delicate thermoelectric pile ami galvanometer will be required: and as glass is coin

paratively opaque to heat waves, a reflecting 1 telescope should be employed, the lenses
of the spectroscope should be of rock salt, and the dispersion should be obtained, either
by means of a rock-salt prism, or by means of a reflecting diffraction grating.

As totality approaches, observe with an integrating spectroscope what rays fade
away and then brighten ; and observe particularly whether all the Fraunhofer lines
are reversed at the instant of totality During totality note the intensity of (7 relatively
to F, and the intensity of 1474 relatively to F, C, 7> ;! , and b. Bear in mind that the
relative intensities may change during totality, and pay attention to that point! Watch
for the reversal of the Fraunhofer lines at the close of totality. Instrument, 7? or E.

Examine the appearance of the eclipsed sun with a meteor spectroscope, having
neither collimator nor slit. So far as the corona is monochromatic it will be distinctly
seen notwithstanding the prism, while those portions of it which shine only by re-
flected sunlight will be indistinct, their light being dispersed. The same object may
be attained to some extent by merely looking at the corona through an ordinary
prism, or through a direct-vision combination. Instrument, F.

Photof/rajtliy. The duration of totality is so brief that it is exceedingly desirable
to photograph the spectra then seen, in order that they may afterward be examined at
leisure. Attempts to accomplish this have been made with two very different kinds of
apparatus, but thus far with only moderate success. One variety of the apparatus,
known as the prismatic camera,, belongs to the class of slitless spectroscopes, and con-
sists of a photographic camera before whose objective a prism is placed, the whole
being equatorially mounted and driven by clock-work. The adjustment of the prism
to minimum deviation is effected by the aid of a collimator temporarily placed before
it ; but on some accounts it would be advantageous to dispense with the prism entirely,
and to use in its stead a diffraction grating placed behind the objective. The photo-
graphs produced by this instrument exhibit a dispersed series of spectral images of
the corona and prominences, each image being due to light of a definite degree of
refrangibility. The other variety of apparatus consists of a spectroscope of the ordi-
nary form, except that it is provided with a small camera instead of a reading tele-
scope, the whole being either attached to an equatorially-mounted reflector of short
focus, or else so arranged upon a table that an image of the sun may be thrown upon
its slit by means of a suitable heliostat and lens. With the prismatic camera, pictures
were obtained during the eclipse of April 6, 1875 ; but thus far it is believed that no
results have been got from the other form of apparatus. The reason for this is not
apparent, because the brightness of the image depends mainly upon the intensity ratio
of the lens by which it is depicted upon the sensitive plate, and there is no reason why
this ratio may not be the same in the two forms of apparatus. Theoretically, the slit-
less form should work slightly the quickest, because it has fewest optical surfaces ; but,
as an offset to this, there is great difficulty in determining the wave-lengths of the
.somewhat complicated series of spectral images which it produces.

A careful search in the library of the Naval Observatory failed to reveal any
data upon which to found an estimate of the value of the exposure constant for the
1474 line. Under these circumstances, if spectroscopes with slits are used it will
probably be best to expose a single plate during the whole three minutes of totality;
and as prisms give brighter spectra than diffraction gratings, the former will be pref-
erable for any moderate degree of dispersion ; unless, indeed, the object is to photo-

irraph tin- ultra violet r;ivs. If prismatic cameras are employed, the exposure constant
should probably IK- from 0.234 to 0.469 at least, these were the values used l>y the
Kuglish c\])edition to Siaiu in their work on the eclipse of April 6, 1875.

SECTION VII. I'OLAUISCOPIC OBSERVATIONS.

( Hiservations of the polarization of the corona wore first undertaken at a time
when it was still uncertain whether the corona was a true solar phenomenon, or was
in some way produced l>y the earth's atmosphere ; and their primary object was to
decide between these two hypotheses. Since then the spectroscope and photography
have placed the solar origin of the corona beyond doubt; but the polariscopic obser-
vations have proved so delicate that the results hitherto obtained from them cannot be
regarded as final. Besides, the details of the polarization, if sufficiently well defined,
may reveal something of the condition of the matter emitting the coronal light; and
if photography can be substituted for eye-observations, information may even be ob-
tained concerning regions further from the sun's surface than any of which we have
at present cognizance.

Dfsrrijition of fiifttntiiicittfi. The polariscope has assumed many forms, but those
best adapted to the investigation now under consideration may be briefly described,
as follows :

Iiixti-nn'iit A. A Savart polariscope, consisting of a compound quart/ plate and
a Nicol's prism or tourmaline: preferably the latter. This instrument indicates the
presence of polarized light by exhibiting a series of bands, either parallel or perpen-
dicular to the plane of polarization ; and it is so delicate that these bands become
visible whenever the polarized light amounts to so much as one per cent, of the total
illumination. To discriminate between the plane of polarization and a plane perpen-
dicular to it, we must know whether the central band is black or white, and as that is
a difficult question when the bands are faint, it will be desirable to make a small addi-
tion to the apparatus. The Savart polariscope being attached to the eye-piece of a
telescope in the usual way, a light-colored tourmaline should be inserted in the eve-
piece in such a position as to cut, off a small segment of the field of view: the axis of
the tourmaline being parallel to the edge or chord of the segment, and the bands being
perpendicular to that chord. Of course the tourmaline must be so placed as to be dis-
tinctly visible through the eye-piece, and it must also maintain its position relatively
to the Savart when the eye-piece is rotated. With this arrangement, the plane of po-
larization will be determined by noting whether the bands continue across the tour-
maline or break off at its edge

liixti-iniii'iit I'>.\ pair of l.abinet's quartz compensating wedges, placed at the
common focus of the objective and eye-piece of a telescope, the eve-piece being pro-
vided with an analyzer consisting of a Nicol's prism or tourmaline. This instrument
indicates the presence of polarized light by exhibiting a series of bands, which make
an angle of forty-live degrees with the plane of polarization. It nearly equals the

Savart in delicacy, and excels it in the facility it affords for determining the plai f

polarization, because the position of the middle band can be marked upon the quartz,
and then a glance suffices to show whether it is black or \\hite.

/ l,',iiiii-iit ('. A biquartz, arranged in the same manner as the I'.abinet's wedge.-

just described. This instrument indicates the presence of polarized light by the colora-
tion of the quartz ; which, however, does not become perceptible until the polarized
light amounts to from 5 to 15 per cent, of the total illumination. If the biquartz
is turned till its two halves are colored alike, both being of the purple hue known as
the sensitive tint, or tint of passage, then the line of junction of the two pieces will be
in the plane of polarization, and as a very slight rotation suffices to destroy the uni-
formity of the tint, this position can be found with great exactness. Thus it appears
that a biquartz is inferior to the instruments A or B in sensitiveness to feebly polar-
ized light, but superior to them in the accuracy with which it will determine the plane
of polarization.

Instrument D. An Arago polariscope, which consists either of a double-image
prism and a plate of quartz cut parallel to the axis, or of a double-image prism and
a plate of selenite. It will be most convenient to mount the prism and plate of quartz,
or selenite, in a brass cell, which can either be used alone, or can be inserted in one
end of a tube about a foot long which contains at its opposite extremity a diaphragm
of such size that the two images of it, seen on looking through the polariscope, may
appear just in contact with each other. This instrument indicates the presence of
polarized light by the coloration of the field of view, and has about the same degree
of sensitiveness as the biquartz, but it is not well adapted to the exact determination
of the plane of polarization.

Instrument E. An achromatic double-image prism.

Instrument F. A Nicol's prism, or a tourmaline. These instruments, as well as E,
are specially valuable when the light is faint, but they will scarcely indicate the pres-
ence of polarization unless its amount is at least 10 per cent, of the total illumination.

A Polarimctcr will be necessary to determine the percentage of polarized light
present. It should consist of four plates of glass, turning upon an axis lying in a
plane parallel to their surfaces and at right angles to their length, and carrying an index
moving over a graduated circle arranged to show the angle through which the plates
are turned. The object whose light is to be tested is viewed through the polarimeter
with a Savart's polariscope, so set that the bands seen in it are parallel to the axis of
rotation of the plates ; then the whole instrument is turned until the bands are per-
pendicular to the plane of polarization of the light to be tested; and, finally, the plates
are rotated until the bands disappear. When this happens, the amount of polarization
present is just equal to that produced by the plates, the amount of which is known
from the reading of the index.

With the exception of D and E, which are best used in the hand, all the above-
described polariscopes should be attached to telescopes provided with positive eye-
pieces magnifying about twenty diameters, and giving pencils of light sufficiently large
to fill the pupil of the eve completely. Before using the telescopes their lenses should
be examined by the aid of polarized light, to make sure that they are free from strain,
and after putting the instruments together again special care must be taken to see that
all reflections are stopped out in the tubes. Any neglect of these precautions may
lead to serious errors.

Points to fee investigated. On the whole,' the observations hitherto made seem to
lead to the conclusion that the corona is strongly polarized in planes passing through
4

the sun's center ; or, in other words, radially ; anil that upon this radial polarization
there is superposed a much feebler vertical polarization, which is nearly uniform over
tin- whole corona. The radial polarization is probably due to reflections taking place
within a few hundred thousand miles of the sun's surface, and Professor Pickering- lias
shown that if it is produced in accordance with Fresnel's theory it should be strongest
in the outer corona, and should diminish to nothing close to the sun's limb. The ver-
tical polarization is probably due to the action of our own atmosphere.

Theoretically there should be no difference between the indications of an Arago
polariscope with a quartz plate and one with a selenite plate; and upon ordinary
objects the results given by these two forms of the instrument ai m e certainly identical;
but Professor Pickering thinks that his observations upon the eclipses of August, 1869,
and December, 1870, show that there is some peculiarity about the corona which
causes its light to behave differently toward a quartz plate from what it does toward
a selenite one.

From these considerations, it will be gathered that the questions demanding special
attention are the following :

1. Is the light of the corona polarized radially or vertically, or in both these ways?

2. If radial polarization exists, does it extend with unabated intensity down to
the sun's limb, or does it continually diminish, and finally vanish, in passing from the
outer corona inward to the chromosphere ?

3. What is the percentage of polarized light present ?

4. In examining the corona with an Arago polariscope, do the results given by
an instrument with a quartz plate differ from those given by an instrument with a
selenite plate?

MctJiods of Observation. If a beam of light polarized in a vertical plane is exam-
ined through any of the polariscopes described above, the appearances presenting
themselves as the instrument is rotated will be as indicated in the following table :

1<2

c .J.

o c '

- 3

CU *O

03 rt

J UO 3

U .

!_, 3 O

Is.

1.2 8

o K
e c in

ts)

c t:

O C3

M .2

o o H

.fl <u

o -

B "^

.2 no

w 5-

*> o

u HH "M a

g "^ W5

'Sea

.s c "a

uiBbcg. S S2P

- c "3 ,:

III

^ CJ 0>

o > >

lali

H -S J

^ o u i;
DO V r ^

Maximum.

Disappear.

Violet.

Red.
Green.

Maximum.
Minimum.

Maximum.

Disappear.

Maximum.

Green, Red.

Colorless.

Equal.

Maximum.

Disappear.

Yellow.

Red, Green.

Max., Min.

Minimum.

Disappear.

Maximum.

Green, Red.

Colorless.

Equal.

1 80

Maximum.

Disappear.

Violet.

Green.
Red.

Minimum.

Maximum.

225

Disappear.

Maximum.

Red, Green.

Colorless.

Equal.

270

Maximum.

Disappear.

Yellow.

Green, Red.

Min., Max.

Minimum.

3'5

Disappear.

Maximum.

Red, Green.

Colorless.

Equal.

Of course the changes indicated in the table take place, not suddenly, but
gradually and continuously, as the polariscope is rotated. With the Savart and Babinet
instruments, if the bands are black centered at the first and third maxima they will
be white centered at the second and fourth maxima, or vice versa. In describing the
action of the biquartz and Arago instruments the colors red and green are mentioned;
but it must be clearly understood that the two complementary tints actually given by
any instrument are determined solely by the thickness of the quartz or selenite plate
employed.

At first sight the table seems to indicate that no two polariscopes are governed
by the same law ; but such a conclusion would be a mistake. They all exhibit two
perfectly distinct test phenomena, or reactions ; as for example, bands with a white
center and bands with a black center, or a given tint and its complementary, or a
maximum and a minimum of light ; and if in any given position of the principal sec-
tion of the polariscope one of these reactions is at a maximum, then the same reac-
tion will also be at a maximum when the position of the polariscope is 1 80 different,
while for positions 90 or 270 different the other reaction will be at a maximum.

Now, supposing the eclipsed sun to be viewed through a telescope armed with a
polariscope, let us apply the principles just enunciated to the solution of the first of
the questions proposed above. The corona will be brought to the center of the field
of view, and the polariscope will be turned until it exhibits a maximum of one of the
reactions. Then, if the polarization is vertical, this reaction will be exhibited over the
whole corona ; but if the polarization is radical, the corona will appear divided into
four quadrants, the first and third of which will exhibit one of the reactions while the
second and fourth will exhibit the other. Next, the polariscope will be slowly rotated,
while the corona is kept steadily in the center of the field of view. Then, if the
polarization is vertical, the first reaction will gradually fade out and be replaced by
the second, which in its turn will be again replaced by the first, and so on, as indicated
in the table above ; but if the polarization is radial, the two reactions exhibited in the
alternate quadrants of the corona will not undergo any change, except that they will
rotate with the polariscope.

The observations necessary for the solution of the second question proposed
above may be made with almost any polariscope by noting whether or not the reactions
just described continue undiminished quite down to the moon. It is probable, how-
ever, that the results obtained from a biquartz will be most satisfactory. It should be
set so that the normal to the moon's limb makes an angle with the line of junction of
its two halves sufficiently large to cause their colors to contrast vividly. Then the
observation will consist in noting whether this contrast continues with unabated force
all the way to the moon's limb, or whether it gradually diminishes, and finally
fades out.

The amount of polarized light present in the corona can only be determined by
measurements with the polarimeter, the method of using which has been already
described.

In dealing with the fourth question proposed above, it is desirable that the two
Arago polariscopes employed should give as nearly as possible the same colors. The

method of proce durewill be as follows : Firstly, the corona will be examined through
one of the polariscopefl, furnished with its tube and diaphragm, and the reactions, both
upon the corona and upon the sky, will be noted; secondly, the tube and diaphragm
will be removed from tin- brass cell containing the prism and plate, and the corona
will be again observed. In this case, the images of two portions of the sky, distant
from each other about t\vo and a half degrees, will be superposed, thus eliminating
their polarization, and therefore it will only be necessary to note the reaction upon
the corona itself; thirdly, all the observations will be repeated with the other polari-
scope.

In order to familiarize themselves with the appearances likely to be presented
during the eclipse, intending observers will do well to practice beforehand upon an
artificial corona radially polarized, such as has been described by Professor Picker-
ing in the United States Coast Survey Report for 1870, page 167. The polariscopes
which seem likely to give the best results are Babinet's compensating wedges and the
biquartz.*

Photography. Among the various forms of polariscopes the Nicol's and don hi e-
image prism seem almost the only ones available under the peculiar conditions which
surround attempts to photograph the corona. If a Nicol is employed, its aperture will
probably be comparatively small, and on that account it should be placed behind the
photographic objective, and near its focal point; but if a double-image prism is used,
its aperture should be sufficient to permit its being placed immediately before the
objective. On the whole, the simplest form of photographic polariscope, and that most
likely to yield satisfactory results, seems to be an ordinary camera, provided with a
photographic objective having an intensity ratio of about one-eighth and a focal dis-
tance of about twelve inches, immediately before which is mounted a double-image
prism with a clear aperture equal to that of the objective. With this apparatus wet
collodion plates of ordinary sensitiveness should be exposed about thirty seconds, and
after each exposure the prism should be rotated thirty degrees, the motion being in
the same direction as that of the hands of a watch.

SECTION VIIL PHOTOMETRIC OBSERVATIONS.

As there is now reason for suspecting great variability both in the size and in the
brightness of tlie corona, exact photometric observations during total solar eclipses
become of much importance ; and in making them, attention may be directed either
to the general illumination of the atmosphere, or to the quantity of light emitted by
the corona.

For investigating the first of these points the apparatus devised by Prof. J. II.
Eastman, and used by him in observing the eclipse of August 7, 1869, will be found

For farther details consult the following: Report on Observations of the Total Solar Eclipse of December 22, 1870.
By Prof. E. C. Pickering. U. 8. Coast Survey Report for'iS/o, p. 165. List of Observations on tin- I'ol:iri/.ation of the
Corona. By Prof. E.G. Pickering. Journal of the Franklin Institute, 1871, vol. 61, p. 58. Instructions for Observers at
tin- English GoviTimiriit Eclipse Expedition, 1871. Nature, 1871, vol. 5, p. 18. Suggestions to Observers of t lie Solar
Eclipse of DcctMnber next. By A. C. Ranyard. Nature, 1871, vol. 4, p. 327. English Government Eclipse Expedition,
1875. Instructions to Observers. Nature, 1875, vol. n, p. 352.

very satisfactory.* On that occasion the light during totality proved about equal to
that on a clear moonless evening, at the time Avhen the sun has sunk so far below the
horizon that third magnitude stars are just becoming easily visible.

For measuring directly the amount of light emitted by the corona, some form of
Bunsen's photometer will be required, and probably that adopted by Professor Pick-
ering will be found as convenient as any.f It consists of a box 9 inches wide, 1 8
inches high, and 6 feet long, within which a lighted candle can be moved backward
and forward by means of a long projecting rod. One end of the box is covered by a
piece of thin white unruled writing paper, in the center of which is a greased spot
about half an inch in diameter. The observation consists in pointing the box toward
the corona, so that the rays of the latter may fall squarely upon one side of the paper
while those of the candle fall upon the other; moving the candle by means of the rod
until the greased spot disappears ; and then marking upon the rod by a stroke of a
lead-pencil the distance between the candle and paper. A number of disappearances
should be observed, and after totality is over the distances between the marks and the
candle must be measured and recorded. During the eclipse of December 22, 1870,
Professor Pickering found that, on the average, the candle had to be 18.5 inches from
the paper to make the spot disappear, but the observations were much embarrassed by
flying clouds.

Owing to the difference of color between daylight and candle-light, no adjust-
ment of the candle will suffice to make the greased spot disappear completely ; it can
only be reduced to a minimum of visibility.! Care must be taken to have the interior
of the photometer-box painted dead black with a mixture of lamp-black, shell-lac, and
alcohol. The candle must be of wax or sperm, of the kind known as six to the
pound, and experiments must be made to determine exactly how many grains it burns
per hour ; or, better, the candle may be sent to the Naval Observatory and these expe-
riments will be made there.

Persons who are inexperienced in photometry should have some practice previous
to the eclipse. The necessary apparatus will be, two lighted candles placed three or four
feet apart, and a sheet of thin white writing-paper, having in its center a spot half an
inch in diameter, rendered transparent by the application of a little grease. The practice
will consist in holding the paper in, and at right angles to, the line joining the two
candles, and then moving it backward or forward until the greased spot disappears,
which it will do when the illumination on the two sides of the paper is exactly equal.
When this experiment has become quite familiar, the photometer which is to be em-
ployed during the eclipse should be taken into a room dimly lighted by daylight, and
the intensity of the illumination should be repeatedly measured by making the spot
disappear as completely as possible, and then noting the interval between the candle
and the paper. Excellent practice may also be had by measuring the intensity of
twilight in the early evening.

* Reports on Observations of the Total Eclipse of the Suu, August 7, 1869, Appendix II to the Washington Observa-
tions for 1867, p. 100.

t See United States Coast Survey Report for 1870, p. 172.

t The greased spot can be made to disappear completely by using paper which is blue on one side and white_on the
other ; but this cannot be recommended, because it introduces serious theorectical difficulties.

SECTION IX. THERMO-ELECTRIC OBSERVATIONS.

The principal points to be investigated are the total amount of heat emitted by
the corona and chromosphere, and the relative temperatures of the prominences and of
different parts of the corona. The galvanometer employed should be very sensitive,
and it will be advisable to have some small resistances which can be inserted in the
circuit to reduce the deflections of the needle in case the currents given by the pile
prove unmanageably strong.

For investigating the first of the points mentioned above, an ordinary thermo-
pile, provided with the usual conical reflector, will be the best instrument ; and it
should be used by pointing it to the corona without the intervention of any condensing
lens or mirror whatever. After measuring the heat in that way, its so-called quality
may be tested by noting the effects produced by the interposition of screens of glass
or other transparent material of various known thicknesses.

For investigating the relative temperature of the prominences, and of different
parts of the corona, a thermopile placed in the focus of a telescope, which should be
a reflector if possible, will be required; and as we are ignorant of the degree of heat
to be expected, the pile should not be too delicate. A very suitable one can be made
by taking a piece of iron wire about six inches long, soldering to each of its extremi-
ties pieces of german-silver wire, and then bending the whole into the form of an M,
the solderings being at the points forming the top of the M. The wire should not be
more than one or two hundredths of an inch in diameter, and the whole should be
so mounted in a piece of cork that, the junctions may project only a short distance
above its surface. The currents given by this pile will be due solely to the differences
of temperature of its two junctions, and, consequently, the galvanometer will be
insensible to any flow of heat which affects both junctions alike. The mode of using
the instrument will be to place one of its junctions upon the moon, or upon a promi-
nence, and the other upon the point whose relative temperature is to be measured.
One of the most important points to be ascertained is how much more heat comes
from the prominences than from the neighboring portions of the corona.

As soon as possible after the totality is ended, the constants of the apparatus
should be most carefully determined, so as to furnish the means of converting the gal-
vanometer readings into ordinary thermometric degrees.

Plate 1.

fioftont