( 879 ) 
From this it was evident that heat was being conducted along 
the metal valve during the formation of the jet, and that as long as 
a valve made of conducting material was employed the proper adia- 
batie conditions of expansion could not be fulfilled. 
§ 8. It remained to repeat the experiments using a valve of non- 
conducting material. Preliminary attempts to construct a wooden 
valve failed, for the wood used was too porous to withstand a 
pressure of more than a few atmospheres without leakage. Glass was 
next tried, and gave considerable trouble, for though very carefully 
annealed, the glass valves frequently broke before a pressure of 20 
atmospheres was reached. Glass pins ground so as to close the valves 
proved quite useless, for the points got constantly broken off and 
plugged the valve. Eventually a glass valve closed by a wooden pin 
was used; it was tested up to 70 atmospheres, and although not 
quite pressure-tight, it closed sufficiently well for the purpose. The 
metal valve, A, was removed, and the glass one soldered in its place. 
The glass-copper junction was made according to the method of 
Carrerer *), by first platinising the glass in the blowpipe, then 
coppering it electrolytically, and finally, soldering it in the usual 
way. The junction proved most effective, and successfully resisted 
a pressure of 70 atm. 
$ 9. It is evident from the following results which were obtained 
with this form of apparatus, that the warming effects previously 
observed at low pressures have been eliminated by the use of a 
non-conducting valve. 
TABLE IV. 
Series XII. Glass valve. Air at 0°C. 30/3/08. 
(p,—I) of Sipe 
6 atm. 0:92 1E 
14. Sy i be pee 
16 Be yet KOM 
21 us, 3:04 i), 
31 A SAT tie 
41 3 Tal 
That these results, although heat conduction has been successfully 
eliminated, by no means agree with the Jourr-KeLvin experiments 
1) Comm. Phys. Lab. Leiden. No. 27. (Zitt.versl. Mei en Juni 1896). 
