20 ROYAL SOCIETY OF CANADA 



greater than its true amount, both of which tend to reduce the discrep-- 

 ancy pointed out on page 19. On the other hand the absorption of 

 chlorine and possibly of potash by the glass will increase the weight of 

 the bulb, and therefore tend to increase the difference or discrepancy re- 

 ferred to. But, as it has been shown that the chlorine and potash set 

 free are almost certainly smaller in amount than the increase in the- 

 bulbs, there is still a deficiency in the following sum : — 



(Weight of oxygen for KCIO^ + loss in weight of bulb) to be 

 accounted for. It appears quite possible therefore that this deficiency 

 may be accounted for by the absorption of small amounts of oxygen. 



The accompanying Table No. 3, contains a number of the results • 

 obtained at different temperatures. The numbers given are the actual 

 weights found in connection with the 0-5 gm. of chlorate that was heated 

 in each case. Theoretically the difference between the numbers in 

 columns 9 and ^, given in column 5, should equal those in column 6, but 

 it will be noticed that the numbers in column 5 are always somewhat 

 greater than those in 6. At the same time it will be observed that the 

 increase in the weight of the glass varies both with the temperature and 

 the time of heating. These numbers in column 7 cannot be taken how- 

 ever as representing the total increase which the bulb originally received 

 owing to the nature of the treatment of the glass in cleaning. In this 

 process the bulbs were first heated in boiling water to heat the water 

 contained in them and thus ensure the dissolving of the whole of the 

 residue of KCIO^, etc. They were then filled with chromic acid solu- 

 tion and allowed to stand some time, after which they were washed suc-- 

 cessively with water, alcohol and ether. It is evident, therefore, that a 

 certain amount of loss is almost certain during this treatment, so that 

 the numbers in column 7 may frequently be too small. 



The increase seems to depend also upon the glass used. In experi- 

 ments conducted previous to the beginning of the series the results of 

 which are not given in the accompanying tables, tubes were used which 

 showed a greater increase than those of the latter series, although in both 

 cases, soda glass was used. 



Thus in the case of one bulb, the increments for five successive ex- 

 periments at temperatures from about 500° to 540° are the following,, 

 expressed in milligrams : — 



5-6; 2-9; 2-2; 56; 4-6; A total of 22-9 mg. 



The glass was not changed in appearance until after it had been 

 in use for some time, when portions of its inner surface showed a large 

 number of fine checks as though the inner surface had become changed 

 somewhat in its character. After these checks appeared the tube soon 

 cracked, when put into the furnace, and had to be thrown away. Ow- 



