OP ARTS AND SCIENCES. 73 



that the capillary ends should be drawn out anew for each analysis This 

 bulbous tube containing the liquid to be analysed was used in the following 

 manner. A cork fitting the combustion tube was perforated to fit one end of 

 the bulbous tube, and the latter tightly inserted in the perforation so that the 

 whole of the long capillary end projected beyond the cork. This end of the 

 bulbous tube was then introduced into the combustion tube, and the cork firmly 

 pressed into its place. The cork being a suitably flexible one, the end of the 

 capillary tube, at the proper moment, was readily broken off against the inner 

 surface of the combustion tube, simply by a lateral pressure against the outer 

 half of the cork. 



The distillation of the material from the bulb was better controlled by 

 means of a heated copper bar, as described in my paper " On a Process of 

 Organic Elementary Analysis," etc , already referred to. When no longer a 

 trace of liquid remained in the bulbous tube, connection with the oxygen gas- 

 ometer was readily made bj' means of a caoutchouc tube, and communication 

 established by breaking off the capillary end within the caoutchouc connecting 

 tube by means of pliers against the sides of the tube. Before the latter was 

 done, however, the oxygen was turned on, in order to produce pressure from 

 behind, and thereby prevent a possible loss from passage of vapor backward. 



When making several analyses in succession, the posterior end of the com- 

 bustion tube was generally found too hot at the close of an analysis to admit of 

 being so tightly clenched with the hand as is necessary in attaching the bulbous 

 tube ; and attempts to make a good connection in this manner under such cir- 

 cumstances were generally unsuccessful. To overcome this difficulty, I pro- 

 cured an iron clamp lined with cork, the lower half of which was attached with 

 screws to the top of the combustion furnace in such a position that the com- 

 bustion tube, at a convenient point near the end, would rest on the cork in the 

 lower half of the clamp. The upper half of the clamp being then laid on and 

 tightly screwed down, served to hold the tube so firmly that the whole force of 

 the hand could be applied to the cork, and thus insure a perfectly tight joint 

 without producing the slightest disturbance of the rest of the apparatus. 



* C = 0.831 ; H = 0.0693. [Mem.] " Repeat this determination by my 

 process." 



