Barker.]

[Nov. 15,

spectroscope, with two inch telescope, and finally a full set of chemicals for Anthony's lightning collodion process, which in my experience is fully three times quicker than any other process."

"The arrangement of the phototelespectroscope requires farther description, for success in the work it was intended to do, viz., photographing the diffraction spectrum of the corona, was difficult and in the opinion of many of my friends impossible. In order to have every chance of success it is necessary to procure a lens of large aperture and the shortest attain able focal length, and to have a grating of the largest size adjusted in such a way as to utilize the beam of light to the best advantage. Moreover, the apparatus must be mounted equatorially and driven by clockwork so that the exposure may last the whole time of totality and the photographic work must be done by the most sensitive wet process. After some experi ments during the summer of 1877 and the spring of 1878, the following form was adopted.

"The lens being of six inches aperture and twenty-one inches focal length, gave an image of the sun less than one-quarter of an inch in diameter and of extreme brilliancy. Before the beam of light from the lens reached a focus it was intercepted by the Rutherford grating set at an angle of sixty degrees. This threw the beam on one side and produced there three images-a central one of the sun and on either side of it a spectrum; these were received on three separate sensitive plates. One of these spectra was dispersed twice as much as the other, that is, gave a photograph twice as long. This last photograph was actually about two inches long in the actinic region. If, now, the light of the corona was from incandescent gas giving bright lines which lay in the actinic region of the spectrum, I should have procured ring-shaped images, one ring for each bright line. On the other hand, if the light of the corona arose from incandescent solid or liquid bodies, or was reflected light from the sun I was certain to obtain a long band in my photograph answering to the actinic region of the spectrum. If the light was partly from gas and partly from reflected sunlight a result partly of rings and partly a band would have appeared.

"Immediately after the totality was over and on developing the photographs, I found that the spectrum photographs were continuous bands without the least trace of a ring. I was not surprised at this result, because during the totality I had the opportunity of studying the corona through a telescope arranged substantially in the same way as the phototelespectroscope and saw no sign of a ring.

"The plain photograph of the corona taken with my large equatorial on this occasion shows that the corona is not arranged centrally with regard to the sun. The great mass of the matter lies in the plane of the ecliptic but not equally distributed. To the eye it extended about a degree and a half from the sun toward the west, while it was scarcely a degree in length toward the east. The mass of meteors, if such be the construction of the corona, is therefore probably arranged in elliptical form round the sun. "The general conclusion that follows from these results is that on

1878.]

[Barker,

this occasion we have ascertained the true nature of the corona, viz: it shines by light reflected from the sun by a cloud of meteors surrounding that luminary, and that on former occasions it has been infiltrated with materials thrown up from the chromosphere, notably with the 1474 matter and hydrogen. As the chromosphere is now quiescent this infiltration has taken place to a scarcely perceptible degree recently. This explanation of the nature of the corona reconciles itself so well with many facts that have been difficult to explain, such as the low pressure at the surface of the sun, that it gains thereby additional strength."

TASIMETRIC OBSERVATIONS.

As this eclipse is the first in which any attempt has been made to measure the heat of the solar corona, Dr. Edison's report to Dr. Draper on this subject is here quoted in full. He says:

"The instrument which I used at Rawlins, Wyoming, during the solar eclipse of July 29th, 1878, for the purpose of measuring the heat of the sun's corona, was devised by me a short time only before that event, and the time was insufficient to give it as thorough a test as was desirable to ascertain its full capabilities and characteristics.

"This instrument I have named the tasimeter, from the Greek words, Taots, extension, and perpov, measure, because primarily the effect is to measure extension of any kind. The form of instrument which I used is shown in the annexed wood cut (Fig. 2.)

"With this instrument was used a Thomson's reflecting galvanometer on a tripod, having a resistance of three-fourths of an ohm. The galvanometer was placed in the bridge wire of a Wheatstone balance, two of the branches of which had constant resistances of ten ohms cach, while of the

Barker.]

[Nov. 15,

other two one had a constant of three ohms, and the other contained the tasimeter which was adjusted by means of the screw to three ohms. When thus balanced, if the strip of vulcanized rubber A (seen in Fig. 3), placed between the fixed point B and the carbon button C, was exposed to heat from any source, it expanded, placing pressure upon the carbon button, decreasing in this way its resistance and destroying the balance; thus allowing a current to pass through the bridge wire containing the galvanometer, the amount of this current of course being proportional to the expansion of the rubber and to the strength of the battery.

"The form of instrument here described was finished only two days before leaving for the west; hence, I was unable to test it. However, I set it up upon my arrival at Rawlins, but found that it was a very difficult matter to balance so delicate an instrument as a reflecting galvanometer with one cell of battery, through such small resistances. In fact, I did not succeed in balancing it at all in the usual way. Nor could it be balanced in any way until I devised a method which I may designate fractional balancing,' when it became very easy to accomplish the result and also to increase the effect by using two cells in place of a single one. This device

consisted of a rheostat formed of two rows of pins. The rows were about one-half an inch apart. A wire was connected from a pin on one row to a pin on the other row and so on, so that the current had to pass through the whole length of the wire, which was No. 24 gauge and four feet long. This was used as a shunt around the galvanometer. A copper wire connecting all the pins of one row served to reduce the resistance to zero. When the galvanometer was thus shunted, a very feeble current passed through it. If the spot of light was not at zero it was brought there by either increasing or decreasing the pressure upon the vulcanite of the tasimeter by the adjusting nut. When thus brought to zero the copper wire of the shunt rheostat was taken off of one pin, thus increasing the resist

ance of the shunt perhaps to one-fiftieth of an ohm. The spot of light was generally deflected nearly off of the scale. The light was again brought to zero by varying the resistance of the tasimeter, and another one-half inch of wire included in the shunt, another deflection and another balance was obtained by the tasimeter. Thus by gradually increasing the delicacy of the galvanometer by increasing the resistance of the shunt and balancing at every increase, the whole of the current was allowed to pass through the galvanometer and the shunt taken off. When this point was reached the damping magnet or director was in close proximity to the case of the galvanometer. To increase its delicacy to the fullest extent it became necessary to raise the director to the top of the rod. This was done by raising it cautiously a quarter of an inch at a time, bringing the spot of light to zero each time by the tasimeter.

"In order to form some idea of the delicacy of the apparatus when thus adjusted, a preliminary experiment was made on the evening of the 27th, with the star Arcturus. The tasimeter being attached to the telescope, the image of the star was brought on the vulcanized rubber. The spot of light from the galvanometer moved to the side of heat. After some minor adjustments, five uniform and successive deflections were obtained with the instrument, as the light of the star was allowed to fall on the vulcanite to produce the deflection, or was screened off to allow of a return to zero. "It was in this condition when the eclipse occurred. The tasimeter was placed in a double tin case, with water at the temperature of the air between the walls. This case was secured to a Dollond telescope of four inches aperture. No eye piece was used. At the moment of totality the spot of light was slowly passing towards cold. When I withdrew a tin screen and allowed the edge of the luminous corona to fall upon the rubber, the spot of light stopped, went gradually off of the scale towards heat, its velocity accelerating as it approached the end. The time required for the light to leave the scale was from four to five seconds.

"I interposed the screen and endeavored to bring the light back to zero, but I was unsuccessful. Had I known that the heat was so great I should have used a platinum strip in place of the vulcanite, and decreased the delicacy of the galvanometer by the approach of the damping magnet.

"I would then doubtless have succeeded in getting two or more readings, and afterwards by comparison with bodies of known temperature would have obtained a near approach to the temperature of the sun's corona."

TELESPECTROSCOPIC OBSERVATIONS.

My own results, obtained with an analyzing spectroscope attached to the telescope, seem to be almost unique in this eclipse. This fact must be my apology, if any be needed, for introducing here at such length, the facts of the case as contained in my report.

The instruments and apparatus used in the observations were loaned for the purpose from the physical cabinet of the University of Pennsylvania. They consisted (1) of an equatorially mounted achromatic telescope of four inches aperture made by Jones of London; (2) a direct vision astronom

ical spectroscope by Merz of Munich; (3) a second direct vision spectroscope by Hoffmann of Paris; and (4) a pocket spectroscope by Geo. Wale & Co. Beside this spectroscopic outfit, a second four-inch achromatic telescope by Dollond was taken for use with the tasimeter by Dr. Edison, and a Savart, a Senarmont, and an Arago polariscope, for determining the polarization of the corona. The Merz spectroscope above mentioned is described in the "Philosophical Magazine," IV., xli., Feb., 1871. It is provided with two compound direct vision prisms, of which one or both can be used at pleasure, each consisting of five single prisms, two of flint glass with a refracting angle of 84°, and three of crown; one of these having a refracting angle of 84°, the others of 87. The dispersive power of each of these compound prisms is about equal to that of two equilateral prisms of flint glass. The instrument has a collimating and an observing telescope, each furnished with an object glass two-thirds of an inch in aperture and four inches in focal length. The prism-tube is attached to the collimator by two centres, giving it a lateral motion about a line passing through these centres, which constitutes an axis parallel to the slit. The observing telescope is similarly attached to the tube carrying the prism. These motions serve to alter the incidence of the rays upon the surface of the prism, and also to bring any special part of the spectrum into the middle of the field. The observing telescope is provided with a positive eyepiece of an equivalent focal length of one inch, and also with a needle micrometer, having an eye-piece of one-half inch focus. The graduations upon this micrometer are strongly cut, enabling the positions and the distances of the lines measured with it to be easily read even in a faint light. The spectroscope was firmly attached to the draw tube of the equatorial telescope by means of an open frame made by Zentmayer, so that the posi tion of the image with reference to the slit could be readily observed.

The time from the date of our arrival at Rawlins until the eclipse, was occupied in setting up the instruments, in getting them into adjustment, and in practice with them. It was found that with only one of the compound prisms of the Merz spectroscope, the slit being placed radially, it was easy to observe the lines C and F reversed in the chromosphere, and also the bright line D. On the morning of the day of the eclipse, the solar edge was examined for protuberances, in order to locate them in advance of totality. But a single one was noticed, this being on the southwestern edge of the sun. As the time of first contact approached, the spectroscope was removed and a paper screen was attached to the drawtube, an image of the sun being formed on this screen by means of the eye-piece; thus enabling the time of this contact to be approximately determined and the subsequent progress of the eclipse to be conveniently observed. No spots were seen under these circumstances, though this could hardly have been expected since the solar image was so small, scarcely three inches in diameter, unless the spots were of large size. As the time of second contact drew near, the spectroscope was replaced upon the equatorial. Since you deemed it of importance to pay special attention to the oxygen lines in the vicinity of G, the micrometer of this instrument