SOLAR ECLIPSE, MAY 0, 18S;5. 103 



telescope before our arriviil at the island, bat Mr. Upton took witli liini a small telescope aud its 

 inoiuitiug, which there formed a part of the complete apparatus. After the oljservations on the 

 day of the eclipse it was found that the mounting was too unsteady to admit of the recpiisit** 

 accuracy in pointing, so that, although Mr. Prkston found polarization in all parts of the corona, 

 the angles of polarization were thoroughly discordant. Of the greatest theoretical importance, 

 however, was tlie observation that with greater brightness of the field the Savart's bands were 

 more intense. As the instrument is specially devised for the examination of isolated ]>ortions of 

 the corona, the inference is that the intensity of polarization does not decrease in approaching the 

 moou's limb. 



2. Prismatic telescope. — This was composed of a large flint-glass prism of 30° before the ob- 

 jective of a telescope lent by Mr. Rockwell. The essential constants were as follows: Clear ap- 

 perture, 2i inches; apparent length of spectrum (A to H), 32|o ; magnifying power used, 20. 



The instrument was used by Mr. Upton, aud the results are given in his report. 



3. Integrating spectroscope. — The instrument belongs to the writer; its constants are given in 

 the report of Ensign Brown, who observed with it. Tiiere is, however, one novelty in its con- 

 struction which is worthy of notice, because it seems to me better adapted to the end in view than 

 any arrangement before used. 



The angular aperture of a small telescope lens, such as would be used for a spectroscope col- 

 limator, is rarely less than five or six degrees ; consequently, if such a collimator is directed towards 

 a source of light as small as the suu, only -^ or i\ of the whole aperture is utilized. The ordinary 

 method used for correcting this fault since the eclipse of 1870, when Professor Young employed it 

 for the lirst time, has been to place a small telescope before the slit directed towards the sun. The 

 effect of this addition is obviously to increase the apparent magnitude of the object ; it is also 

 evident that the magnifying power of the auxiliary telescope should be, in tlie case of a total eclipse, 

 about three or four diameters. The only objections to the method are that it requires at least two 

 lenses, aud that the system admits of no range of adjustment. I used a single concave lens in- 

 troduced between the slit and objective of the collimator. It is clear that such a lens would 

 reduce the apparent magnitude of the objective as seen from the slit, an<l therefore the angular 



* 



V 



apertTure of the system. At the same time, the condition that the distance from the objective to 

 the virtual image of the slit formed by the concave lens shall equal the focal length of the collima- 

 tor is readily satisfied, and this is the only condition for distinct N'ision. If we set (Fig. 1) 



F = focal length of collimator objective, 



A — angular aperture of same, 



(t = angular aperture desired, 

 — / = focal length of concave lens, 



M = distance from concave lens to slit, 



V ^distance from concave lens to image of slit, and 



L = total length of collimator, 

 the solution is as follows: 



