440 Mr. J. Joly .
On the [ Nov. 20 , the electrodes of all dust , and seems to dispel all dust from the gas through which the discharge occurs .
It is well known that an electric discharge in a vessel of air has the effect of clearing out of the air all the particles that serve as nuclei for the condensation of water ; and we made several experiments with a view to determine whether a similar effect was produced on the dust in our tubes .
The gas from the sparking tube was carried through a glass globe , and so on to the jet where it was burned ; a wire connected with one pole of a Toss or Wimshurst electric machine projected into the interior of the globe , and a patch of tinfoil on the outside of the globe was connected with the other pole of the electric machine .
So long as the Yoss machine was not worked , the gas carried the dust from the sparking tube through the globe , and it was seen in the spectrum of the flame , or simply in the colour of the flame when lithium was one of the electrodes ; but , on working the machine so as to produce a silent discharge inside the globe , the flame , in one or two seconds , suddenly ceased to show the spectrum of the dust , and in the case of the lithium lost its red colour .
When the machine was no longer worked , the spectrum or colour speedily reappeared , to vanish again suddenly when the machine was started afresh .
When a narrow tube , with a piece of tinfoil outside and a wire inside , was substituted for the globe , the like results ensued .
It appears , then , not only that dust , however fine , suspended in a gas will not act like gaseous matter in becoming luminous with its characteristic spectrum in an electric discharge , but that it is driven with extraordinary rapidity out of the course of the discharge .
If , then , the spectrum of the aurora be due , not to the ordinary constituents of our atmosphere , but to adventitious matter from planetary space , we conclude that such matter must be in , or must be brought into , the gaseous state , or at least have its properties entirely altered from those it possesses at ordinary temperatures , before it becomes luminous in the electric discharge .
III .
" On the Specific Heats of Grases at Constant Volume .
Part I. Air , Carbon Dioxide , and Hydrogen .
" By J. Joly , M.A. , B.E. , Assistant to the Professor of Civil Engineering , Trinity College , Dublin .
Communicated by Professor Fitzgerald , M.A. , F.R.S. , F.T.C.D. Received September 2 , 1890 .
( Abstract .
) In this first notice the specific heats , at constant volumes , of air\gt ; carbon dioxide , and hydrogen are treated over pressures ranging from 7 to 25 atmospheres .
The range of temperature is not sensibl ) i 1890 .
] Specific Heats of Gases at Constant Volume .
j yaried .
It is found that the specific heats of these gases are not I constant , but are variable with the density .
In the case of air the j departure from constancy is small and positive ; that is , the specific heat increases with increase of the density .
The experiments afford directly the mean value O'1721 for the specific heat of air at the absolute density of 0'0205 , corresponding to the pressure of 19*51 atmospheres .
A formula based on the variation of the specific heat with density observed in the experiments ascribes the value 0*1715 for the specific heat at the pressure of one atmosphere .
The formula assumes the specific heat to be a linear function of the density , which must as yet be regarded only as an approximation , the exact nature of the relation being concealed by variations among the experiments .
These results appear to be in harmony with the experiments of Wiedemann on the specific heat at constant pressure , and of Rowland on the mechanical equivalent of heat , from which the value 0*1712 is deduced for C\#187 ; at 760 mm. The experiments on carbon dioxide reveal a more rapid variation of the specific heat with density , the variation in this case being again positive in sign .
The formula Cv = Px 0*2064+0*16577 appears with considerable reliability to express the relation between specific heat and density .
The relation between specific heat and density in the case of hydrogen is of a negative character ; that is , the specific heat diminishes with increase of density .
The experiments are chiefly directed to elucidate this point , for , owing to the difficulty of preparing pure hydrogen , it was found that variations in the quantitative results of experiments on different samples of the gas were unavoidable .
Accordingly the experiments were directed to a comparison of the specific heats of like samples of the gas at different densities .
The variation with density is small , but ( with one exception ) all experiments on the purer hydrogen ascribe a negative character to it .
The nature of these variations of specific heat with change of density is , in the case of the three gases , in accord with their behaviour as regards Boyle 's law , within the limits of pressure .
The experiments were effected in the steam calorimeter , a differential f Method being used in which an empty or idle vessel is thermally comil Pared with the vessel holding the gas at high pressure .
The vessels possessing approximately the same calorific capacity , the result , theoretically , is as if the gas was dealt with isolated from any con-1 hdng vessel .
Although practically this is not attained , many sources j error are eliminated by the procedure adopted .

