

G72 



itEroRT--18S4. 



i 



When tlie hydrofron is loss timn double tlio oxvfrcii, tlie excess of nxv(,'eu cannot 

 r(>act with any ol' the three otlier pises present — carbonic oxide, carbonic acid and 

 steam, but has to wait until an equal volume of steam is reduced to hydrofen by 

 the carbonic ox'de. 'i'he excess of inert oxyg(^n has the same ellect'as the ir.i I't 

 nitropen in favouring the formation of carbonic acid. 



The variations in tlie coelluient of atlinity found by IForstmann with dill'ereiit 

 quantities of oxyj^en, ar(> due jiartly to this cau.M>, but chieliy to the varyinir 

 amounts of steam condensed by the cold eudiometer durinpf the reaction in (htlerem 

 experiments. 



(7) As tlie peiievnl result of these experiments it has be(>n .shown, that when a 

 mixture of carbonic oxide a'ld hydrop-en is exploded with insullicient oxvpen lor 

 complete combustion, at a temperature at wliicli no condensation of steam can take 

 place durinp: the reaction, and at a pressure prreater than the critical pressure, an 

 equilibrium between two opposite chanires is establislieil, which is independent of 

 the quantity of oxy^'en tahen, so lont,' as tliis quantity is less than half tiu- 

 liydropen. Within tlie limits marked out abovr,tl.e Law of Mass is conqiletelv 

 verifieil for the jraseous system conii-.osod of carb";iic oxide, carbonic acid, livdrj- 

 gen and steam at a hiirh ienipi'ralnrt'. 



'k SprcJro.'U'ripic Stiidir.-^ of I'J.rphsioils. 

 T,tj Professors Livking, F.U.S., and J3i;\var, F.Pi.S.^ 



The ox])losions observed were cliieOy tho->e of ]iydroi,i-en with oxyf^cn, ami ni 

 carbonic oxide with oxy pen, and wer(? made in an iron lub(> fitted with quartz umls. 

 The spectra were botii olisorved witli the eye and iilioto^raphed. Lininfrs oi' tliiii 

 -sheet metal of various l<inds were introduced into the tube, and in some caso 

 metallic salts in fine powder were ]mt in. W'lien the tube was clean, many iron 

 lines were seen in tlie flash, nine lines in the p-reen were identified, and fovtv-iiine 

 more in the blue, violet and ultra-violet. Only one line m<n'e refranp'ible than wns 

 observed, and that was 'J'. Twenty-five line< of ni'.diel and twent\-two of cobalt, 

 all ^yinji; between (t and P, were photoirraidied. No other metal gave any tliin;.' 

 like so many lines as these three, but magnesium gave the h group, copper gave one 

 green and two ultra-violet lines, manganese the violet tri|ilet: cliromiiuii, three 

 triplets in tlie green, indigo and uUra-violet respectively ; silver, two ultra-violer 

 lines; sodium gave D and the pair near Q ; potassium, the violet lines and tlie jmir 

 near O. On tlie otlier hand, zinc, cadmiuu., mercury, aluminium, tin, bismuth, 

 iiutimony and arsenic developed no lines in the flash ; marsh ga*, sulphuretted, 

 arsenetted and ant imoniuretted hydrogen exploded with oxygen gave no peculiar 

 lines, merely a more continuous spectrum. It a]ipear3 to be proved that iron, nickel 

 and cobalt are volatile in some degree at 3,000°, which, according to Bunsen and 

 Berthelot, appoaiv to bo about the temperature of the exploding gas ; and ibis may 

 help to exjilain tlie appearance of iron lines Inthe highest parts of solar prominence.-'. 

 It might bo possihle to establish a .spectroscopic scale of temperatures if the lines 

 successively developed with increasing temperatures were noted. Thus the iron line, 

 T .seems to bo bnrely developed at JJ.OOO", the aluminium lines at H only come out 

 at a somewhat higher temperature, the lithium blue line may be just seen in tlie 

 inner green cone of a Bunsen burner, while the green line comes out in the 

 explosion Hash. 



FRTDAY, AUGUST 29. 



The following Papers were read: — 



1. On the Constitution of tlie Elements. By Professor Dewar, F.E.S, 



' For details, see Pial. Mag. for September 1884, 



