118 Yan Name and Brown — Tri- Iodide and 



ratio 7/7' shall be constant even when the ratio i^\)li^\^ 

 varies widely. This is at first sight an unexpected result, but 

 the correct explanation has probably been given by Bray and 

 MacKay* who show that it is in agreement with the principle 

 governing ionization in a mixture of two largely ionized salts 

 having an ion in common. The same reasoning and corre- 

 sponding mathematical relations must also hold in the case of 

 bromides. 



A number of direct determinations of the value of A^for 

 iodides have been made by Bray and MacKay,f who obtained 

 them by measurements of the conductivities of potassium 

 iodide solutions saturated with iodine, and independently fj'om 

 the conductivities of iodine solutions saturated with cuprous 

 iodide. These experiments were all made in rather dilute 

 solutions, (21) 4- (SIg) = O'l molar and below. Values for K 

 for potassium iodide solutions from O'l to 1*0 molar, satui'ated 

 with iodine, have also been calculated by Bray and MacKay 

 from experimental data given by Laurie.:}; Comparison of the 

 observed values of K and K^ for iodides shows that these two 

 constants are in close agreement in dilute solution, and probably 

 agree fairly well in solutions up to normal strength. For con- 

 centrations in the neighborhood of 0*1 normal iodide both 

 constants are nearly independent of the iodine concentration, but 

 in the stronger solutions the values of the constants decrease 

 very materially, as saturation with iodine is approached. 



The identity of K and K^ means that the ratio 7/7' in the 

 dilute solutions is not only constant but approximately unity, 

 or in other words, that the iodide and tri-iodide in the 

 equilibrium mixture are ionized to practically the same extent, 

 irrespective of the value of the ratio (2I)/(2I,). An important 

 consequence of this, from a practical standpoint, is that a 

 determination of K^ is in effect an approximate determination 

 of K. This, however, is only true in " normal " cases. 



Finally, reasoning by analogy, we are naturally led to the 

 conclusion that for bromides also, the values of ^ and K^ in 

 dilute solutions are probably identical, but until independent 

 determinations of K for bromides are available, this inference 

 can not be directly tested. 



Nature and Probable Cause of Abnormal Results. 



What we designate as "abnormality" with respect to the 

 tri-halide equilibrium actually consists, as a reference to the 

 above tables will show, in a lower capacity to unite with the 

 halogen, and one which varies greatly with the concentration 



*Jonr. Am. Chem. Soc, xxxii, 9l6, 1910. 



f Jour. Am. Cliem. Soc. xxxii, 914, also xxxii, 1207, 1910. 



iZeitschr. phys. Chem.; Ixvii, 627, 1909. 



