BAXTER. — ATOMIC WEIGHT OF IODINE. 83 



burg's,* and my work. In this second paper two main criticisms of my 

 investigation are made. One of these concerns the validity of the con- 

 clusion that ordinary iodine does not contain an undiscovered halogen 

 element. This criticism is founded upon a mistaken understanding, how- 

 ever. My experiments were directed to prove the existence or non-exist- 

 ence of an element of Itigher atomic weight than iodine, which, if its 

 properties were those to be expected from the properties of the other 

 members of the halogen family, would have been set free from solutions 

 of its compounds by iodine, and hence would have accumulated in the 

 first fraction of the fractional separation (see page 422 of the first 

 paper). The existence of an unknown halogen of lower atomic weight 

 than iodine was not considered. 



Although Kothner assumes the possibility that such an element exists, 

 it is obviously improbable that any halogen of lower atomic weight could 

 have remained in the purified samples of iodine which were employed in 

 my experiments ; since those specimens which received even the least 

 purification were thrice distilled from an iodide, the iodide having been 

 made in each distillation from nearly one half the iodine from the pre- 

 vious distillation. 



In the second place, it is maintained that volatilization of silver halides 

 takes place when silver iodide is heated in a current of chlorine. This 

 is undoubtedly true, unless precautions are taken to prevent the volatili- 

 zation. Silver iodide is much more volatile at its fusing temperature 

 than silver chloride; and if the original silver iodide is fused before the 

 current of chlorine or bromide is begun, a loss may take place by volatili- 

 zation or possibly by spatteriny. In my own work, however, fusion of 

 the salt was always avoided until the greater part of the change had 

 taken place. Proof that no volatilization of silver halides actually 

 occurred in my experiments has already been given (page 75). It may 

 be added that the close agreement of the results of Series I and II is 

 additional evidence in the same direction ; for auy loss of silver bromide, 

 which is the most volatile of the three halides, would have tended to raise 

 the results of Series II ahove those of Series I, whereas such slight dif- 

 ference as exists is in the opposite direction. 



Kothner' s average result in the Series A«I : AgCl, when recalculated 

 upon the basis of chlorine 35.473, becomes 12G.974. The slight differ- 

 ence between this value and the final result of the research described in 

 this paper may be explained in several possible ways. A perusal of 



* Ber. d. d. chem. Gesell., 35, 2275. 



