191 1.] IN THE TAIL OF HALLEVS COMET. 257 



demonstrated that the constituents of the tail varied markedly from 

 one period to another. Thus from April 29, 1910, to May 7 the 

 spectrum of the tail was almost wholly emissive. On May 11 it 

 had changed to one nearly continuous, while on ]\Iay 23 it had be- 

 come largely emissive again and grew more so as time went on. 



By comparing the photographic with the spectrographic series 

 of representations of the tail a striking fact came to light. To appre- 

 ciate this another point must be taken into account. In order to 

 compare properly a photograph and a spectrogram, both should be 

 made on the same brand of plate. Xo plate reproduces all parts of 

 the spectrum with eqiial intensity. One kind of plate will emphasize 

 certain rays and depreciate others ; the r.ext will reverse the estima- 

 tion. Great error will then be introduced unless the plates be 

 identical. 



Xow the photographs measured were taken with a Brashear 5-in. 

 doublet, an excellent lens, on a Lumiere 2 plate. The rays regis- 

 tered by this plate extend from 3500 in the violet to 5 160 in the 

 green where the sensitiveness ceases. Indeed the effect would have 

 stopped sooner had it not been for the h_\ drocarbon emission at this 

 point. The light, therefore, of the photograph would be exactly 

 differentiated into its constituents by a spectrogram taken on a 

 Lumiere 2 plate. The only difference between the two would be due 

 to the absorption of the objective prism, an absorption relatively 

 greater for the violet than for the blue or green. This would work 

 as much on one kind of light as on another of the same refrangibility 

 and as the two different kinds we are considering, the emission and 

 the continuous spectrum, are about equally spaced in the region of 

 the violet, the correction needed on this account is small. We may. 

 then, directly gauge the character of the photograph's light by that 

 of the objective prism spectrogram taken on the 2 plate. 



This w-e now proceed to do. On the exact date of the photo- 

 graph no 2 plate was used with the two objective prisms though we 

 have objective prism spectrograms on Cramer Iso. Instantaneous 

 on ]\Iay 2'^, 25, 26, 28 and 29. The nearest plate to the date in 

 question was on ]\Iay 29. Of this the best estimate gives for the 

 constituents of the light of the tail at a distance of 3° to 6° from 

 the head : 



