Ski'TEMBi;r 5. 1895] 



NA TURE 



44; 



gear has been altered so that the driving rod performs its revolu- 

 tion in a second, and the motion is then communicated to the 

 driving screw through a small worm wheel. The driving rod 

 is vertical and in two parts, the lower ])orlion ending in a faced 

 ratchet wheel, 3 inches in diameter, and with 200 teeth. The 

 upper part of the rod ends in an arm at right angles to itself, and 

 this arm carries a ratchet of suitable shape held down by an 

 adjustable spring. An electro-magnet connecteil with the con- 

 trolling pendulum, is arranged so as to only permit the ratchet to 

 pass it once a second (see Fig. 42). If the clock be driving too 



Pit;. 42.— Electrical control for lo-inch equatorial. 



quickly, the ratcliet i-; held until tlij stop is raised by the 

 ]iendulum. When held in this way the ratchet is lifted out of 

 the teeth, and the driving clock itself is not affected. 



In order that this form of control may be effective, it is essen- 

 tial that the clock should be going too quickly, as it is only 

 ca]iablc of retarding the driving-rod. 



The contrulling |)LMKhiluni is, of course, regulated to the rale 

 required for the particular star which is being photographed. 



In Mr. Russell's form of control the two parts of the driving 

 rod are connected by friction plates. It was found, however, 

 on testing this arrangement, that when the upper porticm was 

 held by the electro-magnet the rate of the governors was seriously 

 retarded ; hence I introduced a ratchet wheel, and its working 

 leaves nothing to be desired. 



Eiilargei/ients'iof the Negatives. 



Many of the negatives taken have been enlarged about nine 

 times on glass, and further copies have been taken on bromide 

 paper, bringing the enlargement up to about twenty-five times 

 the size of the original. 



Owing to various causes the photographic spectra obtained Ijy 

 the metho<l of trails show irregularities resembling the lines along 

 the spectrum observed when the slit of a spectroscope is partly 

 clogged with dust. It has been noticed that the jjeriod of the 

 irregularities is equal to the time of revolution of the main 

 driving screw of the telescope, and hence they may be accounted 

 for by supposing the driving gear to l)e mechanically imperfect. 

 In that case some of the parallel lines which, liy their juxta- 

 position form the broadened spectrum, are superposed, while 

 others are drawn apart, thus giving rise to dark and bright lines 

 parallel to the length of the spectrum. These lines are more 

 apparent in the ca.se of bright stars than fainter ones. If the 

 telescope were driven with perfect regularity and the almosphore 

 were quite steady, we should obtain a spectrum of uniform in- 

 tensity along its width. This condition has very nearly been 

 obtained in some cases. 



The irregularities above described are eliminated in the en- 

 larged negatives by giving them a vcrj- slight up-and-down 



NO. 1349, VOL. 52] 



motion during exposure in a direction parallel to the lines of 

 the spectrum. This was originally done by hand, but a negative 

 holder has been constructed in which the necessary motion is 

 given to the negative by a small driving clock. 



-V diagram of the arrangement is given below. The only 

 drawback to this method is that defects of the film are apt to 

 produce, by a succession of their images on the enlarging plate, 

 lines (generally very faint) which have a semblance of the true 

 spectrum lines. 



To distinguish the real lines from the artificial ones, a direct 

 enlargement of the spectrum is made on the same 

 plate alongside the other, the to-and-fro motion 

 being disjiensed with. By a comparison of the 

 two enlarged strips, one can see at a glance 

 which are the true lines of the spectrum, and 

 which are those produced by small irregidarities 

 on the film. It may be stated that Dr. Scheiner 

 has also used a somewhat similar method to 

 the one described, the only difference being that 

 he caused the plate on which the enlargement 

 was to be taken to have the oscillating motion, 

 instead of the original negative. The method 

 employed by me. though no account of it had 

 been published, had been in use for some time 

 before Dr. Scheiners method was announced.' 



My object was not so much to obtain photo- 

 graphs of the spectra of a large number of stars, 

 as to study in detail the spectra of compara- 

 tively few ; hence many of the stars have been 

 photographed several times with special exposures 

 and foci for different regions of the spectrum. 



.'\s in the case of stellar spectra observed by 

 eye, the photographic spectra vary very con- 

 siderably in passing from star to star. 



In the classification of stars adopted from a 

 consideration of the visual observations, only the 

 broader differences in the spectra have been 

 taken into account. I'rof. Pickering, however, 

 has suggested a provisional classification in con- 

 nection with the Henry Draper Memorial photo- 

 graphs of stellar spectra, but this chiefly relates 

 to photographs taken with small dispersitm. 

 Now that i^ has t)ect>me jiossible to obtain large dispersion 

 photographs of the sjiectra, much more detail is revealed, 

 and hence I determinetl to deal with the ]iresence, or absence, or 

 changes o f intensity, of individual lines to a greater extent than 

 Prof. Pickering has done in his observations so far published. 



I ■-. ) ;. — Negative tiolder used ill cnlarKiiii;" 



In the first instance, I arranged the various stars of which the 

 spectra have been ])hotographed in tables, without reference to 

 any of the existing cla.sssfications, and not taking into account 

 the finer details. 



The basis upon which this first grouping was founded is the 

 extent of the continuous absorption at the blue end of the 

 spectrum. Such ^a distinction was not possible in the case of 



' Xatckk vnl. xiii. p. -01 8qo. 



