Royal Society. 367 



different from those they bear to the other kinds of spectra. For 

 the successful prosecution of this curious investigation, the first step 

 must consist in the minute examination of the chemical actions of 

 aU the parts of a pure spectrum, not formed bymaterial prisms, 

 and he points out, for that purpose, one formed mFraunhofer s me- 

 thod, by the interference of the rays of light themselves in passing 

 through gratings, and fixed by the heliostat. _ 



He notices a curious phenomenon respecting the action ot light 

 on nitrated paper ; namely, its great increase of intensity, under a 

 certain kind of glass strongly pressed in contact _mth it ; an efiect 

 which cannot be explained either by the reflection of light, or the 

 presence of moisture ; but which may possibly be dependent on the 



evolution of heat. c- t i, tr 



Twenty-three specimens of photographs, made by bir John tler- 

 schel, accompany this paper : one, a sketch of his telescope at 

 Slouo-h, fixed from its image in a lens ; and the rest copies of engra- 

 ving^and drawings, some reverse, or first transfers ; and others se- 

 cond transfers or re-reversed pictures. 



March 21 .—The following papers were read :— , , , 



I. " Description of a Compensating Barometer, adapted to ^Me- 

 teorological purposes, and requiring no corrections either for Zero, 

 or for Temperature." By Samuel B. Hewlett, Esq., Chief Military 

 Draftsman, Ordnance. Communicated by Sir John F. W. Herschel, 

 Bart., K.H., V.P.R.S., &'c. . 



In the instrument here described, there is provided, in addition to 

 the ordinary barometric tube (inverted, in the usual way, m a cistern 

 of mercury,) a second tube of the same dimensions, placed by the 

 side of the former, and Ukewise filled with mercury, but only to the 

 heio-ht of twenty-eight inches above the level of the mercury of the 

 ci^t'em This tube is closed at its lower end, and fixed to a float 

 supiiorted by the mercury in the cistern : and it bears, at its upper 

 end an ivory scale, three inches in length. The elevation ot the 

 mercury in the barometric tube is estimated by the difference be- 

 tween its level and that of the mercury in the closed tube ; and is 

 measured on the ivory scale by the aid of a horizontal index, em- 

 bracing both the tubes, and sUding vertically along them. As the 

 float which bears the closed tube, to which the scale is attached, 

 re«ts freely on the mercury in the cistern, and consequently always 

 adiusts itself to the level of that fluid, no correction for the zero 

 point is needed ; and as every change of temperature must similarly 

 affect the columns of mercury in both the tubes, after the scale has 

 been adjusted so as to read correctly at any given temperature, such 

 as S-^" which may be effected by comparison with a standard baro- 

 meter 'every other reading will correspond to the same temperature, 

 and will require no correction. The author considers the eiTor 

 arising from the difference of expansion corresponding to the dit- 

 ferent lengths of the two columns of mercury, and which will rarely 

 amount to one four-hundredth of an inch, as too small to deserve 

 attention in practice, beinfj, in fact, fur within the hmits of error m 

 ordinary obser^'ations. 



