]\gt ; PR 1891 .
] fluence of Pressure on the Spectra of Flames .
217 The application of ( 46 ) gives . . . . . . . . . . . . . . . . . .
( 48 ) , and . . . . . . . . . . . . . .
( 49 ) .
We see then that the modulus of ( 43 ) will be a minimum , when . . . . . . . . . . . . . . . . . . .
( 50 ) , and in taking the square root the ambiguity must be so determined as make the real part of positive .
Equation ( 50 ) coincides with that by Mr. Heaviside for the case where all the quantities are real . .
' On { he Influence of Pressure on the Spcctra of Flames By .
D. , F.R.S. , Prol.essor of Chemistry , and J. DEWAR , M.A. , F.R.S. , Jacksonian Professo ] , of Cambridg .
Received , 1891 .
We have already described Phil. Trans remarkable spectrum of the ogen f burning at the ordinary atmospheric pressure .
Recently we have examined the } ) ectrum of the same at surcs : hydrogen in excess of oxygen to a ; ur of 40 , and oxygen in excess of hydrogen up to a pressure of atmospherei , , also .
thab of the mixed gases hurning in bonic acid gas .
The apparatus employed was an adaptation of cxne of the nbes used in our expel.iments on the absorption of essed gases ( ' Phil. Mag September , 1888 , and 'Roy . .
Proc vol. 46 , .
It consisted of a steeI cylinder , about 50 mm. in interllal diameter and mm. long , fitted at o1le end with a tz stopper , the annexed figure , and with a jet , , for ning the gas , adapted by a fitting urlion joint the } ) posite end .
) were two tubes , and , connected to the c.ylinder at the sides , of which one , the introduction of gas , the other , fitted a stopcock was used to draw off the water formed , or to reduco prsssure the gas in the cylinder if desired .
The flame was observed , nearly end on , th the quartz opper apparatus was kept cool by a of water on to a sponge cloth wrapped round the cylinder .
the course of the tube conveying gas to the jet was intel.posed a smau cylinder , , in which sodium was placed , and by heating thi the gas entering could be charged with sodium vapour .
Hydrogen in Oxygen .
Influence of Pressure on the Spectra of Flames .
from the rssidue of atmospheric air mixed with the condensed gases .
We took a photograph of , and on comparing this with a photograph of the , we found the two to be identica ] .
xcept f. the bands , and the ighb lines of sodium , the spectrum to be con tinuous , and to extend about 6200 to 4150 , with }ightcst p about 5150 .
It iucreased brilliance as the increased , as well as in extcnt , being visible at i3 atmosphercs from about to 4040 .
greater inCtIleSS of the bands at ] essures 1 due both to the greater brightness of contiuuous pectru and to the greater quantity of ormed .
A large quautity of water accumulated in the expe ] -tal tube , and when this was drawn off by the stopcock , it cffervesc.ed with escape of NO , crtlld was fonnd to be strongly acid .
A specimen ated found to contain vory nearly 3 per cent. of acid .
The tions were continued up to pressure of 40 atmospheres .
Thel e was no lication that the continuotls ) ecfrum ] any ccnllexion with the of hydrogen .
was no crease of brilliance in the hood of the , or lines of .
The chal a of the were , however , better seen in the absence , and will be described in the next section .
Oxygen IIyclrogen .
In this case the colour of the flame was very diflerent from that ' in oxygen .
Instead of bsing yellow , it appeared , to the unaided eye , to have a lave1lder hue .
the oscope it showed a perfectly spectrum , ighl e in the greer .
, about the region ) nhofer line , and very ally f away on either side .
On the red side it could be traced up to about , and ou the violet side to about 4285 , at ordina The sodium lines were absent .
With of pressure it increased much in ightltess , and at 8 atmospheres pressure it could be as low as and as high as 3990 .
The sion used was of a direct , vision spectroscope ( such as was by us , ' Roy .
Soc. ' vol. 41 , p. 449 ) , equivalent to three of white flint glass , but the and telescope very , so as to obtain of light .
less rsion , perhaPs , the pectrum might beeu traced further .
Photographs , Lowevel , showed thab it scarcely nded into the ult ra-vlolel The was no indication that this specbrum was due to expansion of the of eithcr the first , or second , spectrum of hydrogen It is truo that the maximum brightness ( which could not be determined any great accuracy ) was ) very far from , but uo of any second maximum in the neighboul.hood of bservations were continued with the eye up to 25 atmospheru pressure , but no trace of emission , or absorptio1l , corresponding to either of hydrogen could be detected , and it is doubtful it either spectrum can be produced in such a flame .
Since the of steam from its component gases is atbended with a diminutionW volume , increased pressure will increase the stability of the pound , and bhe flame will contain a larger proportion of steam , well as have a higher lture , bhan at ordinary pressures .
The water formed when the flame was jet of oxygen burnin9 hydrogen was found to be alkaline , and to contain ammonia .
Bnt the proportion of amnlonia was much less than the proportion of formed when the jet was hydrogen burning in oxygen ; 1 specimen titrated contained per cent. of ammonia .
liiffects of Pressure on the Sodium Spectrum .
1891 .
] Influence of Pressure on the Spectra of Flames .
edges to begin with , it is extremely difficult to judge whether there is any expansion .
At all events , may say that there is no expansion produced by pressure at all comparable with that produced in a at pressure by increasing the quantity of sodium in the tiame .
We noticed , however , that the presence of sodium , which produces a feebIe continuous spectrum in a flame at an ordinary pressure , seemed to increase the continuous spectrum of he flame under pressure , especially in the orange and green .
Oxy-hydrogen Jet in Carbonic Acid Gas .
For this experiment a two-branched ) ( the upper one in the ligure ) was used .
The jeb of mixed oxygen and hydrogen was first lighted and introduced into the experimental cylinder while the Iatter was full of air and the open .
The air was then replaced by enlering by the tubc .
The eifect of was at once to brighten the flame and its colour from yellow to blue .
Seeu in the spectl.oscope , the change consisted in an increase of continuous spectrum , especially towards the more refrangible eIid .
Wben the stopcock was closed so that the essure rose in the experimental cylinder , fiame increased in brightness , but there was no other change in the spectrum .
It remained continuous with no bright or dark lines , or bands , except the lines of sodium .
It resembled ordinary flame of O. jet would not burn i11 unlefis there was some excess of oxyge1 ] , and even with an excess of oxygen we could not get it to continue to burn in at a higher than 2 atmospheres .
Ethylene in Oxygen .
A jet of ethylene bnrning in oxygen gave , when bhe flame was small , the usual candle-flame , together with a band in the lndigo shading towards the vioIet ; but as the pressure was mcreased the continnous spectrum ightened and completely overpowered the bands , and at the same time the absorption spectrum of appeared .
We carried the pressure up to 33 atmospheres , and at that pressure ihe flame seemed to give nothing but a continuous spectrum , intersected by the absorption bands of .
In ollr tube , flame was viewed almost directly end on , and it is possible hat if we had seen the { lame sideways , we might have detected the hydroflame near the nozzle .
At the high pressure eparated .
We lried burning a mixture of eth lene and oxyge1l .
The mixedjet burnt well in air and , when the supply of oxygen was fufficient , ve the hydrocarbon flame spectrum .
In the experimental tnbe in oxygen , the jet burnt well at the atmospheric pressure , onogen and en .
1891 .
] Influence of Pressure on the Spectra of .
223 noticed to occur in the principal sections of the discontinuous spectrum of the oxy-hydrogen flame .
Whether the colltinuous spectrum is due to the mutual action of the molecules of the com- pressed gases may perhaps be best determ1ned by some photometric measures of the rate at which the brilliance lncreases with the pressure .
Frankland Researches , has made some such measures , but not sufficient solve the tion .
We have made attempt to measure , not tho total intensity of the but that of rays of definite refi.angibility .
Photo } of Oxy-Hydrogen Flame under Pressure .
The nsed for these measul.es was a -photometer of the pattern employed by Crova ( ' Annals de Chimie , ' , vol. 29 , p. 556 ) .
In this , the rays of one of the soul.ces of to be compared are passed two icol 's prisms , and then refiected one half of the slit of spect oscopc , when the light from the other passes directly into other half of the slit .
By turning one of the Nicol pl.isms , the light from the first source can be reduced at pleasure , and any section of the spectrum can be separately observed by cutting off the rest by means of a tter in the eye-piece .
We fou.nd it by no ]noalls easy to get good concordant observations .
A much large ] .
csscJ was used than for the eal.lier experiments , one which ined several litres , and we may presume a more uniform prcssure aintaill e withill it .
The of the best series of on the photometric intensity of jet of oxygen burning in hydl.ogen are given in the following tablo .
comparison light was a petrolenm lamp .
4 .
The first column gives the pressure of the , tho seco the of four to six rvations of the oular deviation of the Nicol 's prisms from the position of co1nplete , for each pressure .
The third colnnm gives the squares of the sines of the in the seoond column multiplied by 100,000 .
will be seen from the last column that the numbers in tho third column , which should be proportional to tho ) hotometric itensities A series of similar observations on hydrogen burning in somewhat diiferent resuIts , tabulated below:\mdash ; 1891 .
] Focometry of Lenses and Lens-Combinations .
radiation ; but the electric energy may well be directly conerted into the motion of radiation .
As a fact , we have never yet been able to obtain either the emission or the absorption spectrum of hydrogen without the aid of an electric current , so that , in reasoning on this spectrum , we are much more in a region of specuIation than when tresking of flames .
Whekher the .
lines , bright or dark , in the solar spectrun ] are dlrectly by the high temperature of the sun , may even be called in question .
may admit that the density of the hydrogen in the atmo , outsiae the photosphere , is but slight , it does not follof that the total pressure of all the gases forming that atmospkere is so very small as Messrs. ankland and Lockyer ( ' Roy . .
Pxoc vol. 17 , p. 288 ) have , from the width of the lines , concluded it.to be .
After all , it is not so easy to connect the temperature , even of a flame , with its radiation , for it is only when the condition of a gas is teady that we can assume that there is a definite relation between the motion of agitation , lvhich tenlperature depends , the vibratory motions , on which radiation depends .
In speculating on questions , chemical , as well ical , changes must not be lost sight of , although the latter be more directly concerned in radiation .
Experiments we have upon the in an atmosphere of })essed gas tend to the same concl t It does not appear that the metallic lines in the arc are sensibly aifected by a up to 15 atmospheres .
Tho details of these observatious , which are complicated by the variation of resistance change of pressure , we defer until the ments are finished .
III .
" " On the Focometry of Lenses and -Combinations , and on a new Focometer By SILVANUS P. THOMPSON , D.Sc .
, B.A. , Professor of Physics in the City and Guilds Technical College , Finsbury .
Communicated by PI ofessor G. CAREY FOSTER , , B.Sc. , F.R.S. uary 4 , 1891 .
( Abstract .
) Few of the accepted methods of focometry take into account the tance b ' the two principal points ( or Gauss points ) of a or afford the means of measuring this distauce , as ag Ah true focal length , and some of them are open to the objection that they Sitate troublesome double adjustments .
Of these methods the author gives a brief categorical review .
He bas dsvised a method in which there are no double adjustments , no uremenls of the size of optical images , no assumptions as to

