CRYSTALLOGRAPHICAL RELATIONSHIP 253 



T1C1 6 2H 2 ; (3) Rb 3 TlBr 6 2H 2 ; (4) K 3 InCl 6 2H 2 ; and 

 (5) K 3 InBr 6 2H 2 0. It is rather remarkable that while 

 Rb 3 TlBr 6 2H 2 O has been isolated, neither K 3 TlBr $ 2H 2 and 

 (NH 4 ) 3 TlBr 6 2H 2 on the one hand, nor Cs 3 TlBr 6 2H 2 on 

 the other, can be obtained from their solutions. Although 

 the order of stability in a series of salts of similar formula 

 dependent as it is in the main on the varying solubilities of the 

 different salts which may be precipitated seems as a rule to 

 have a definite relationship to the order of atomic weights, 

 there are evidently markedly exceptional cases. 



II : The series R 2 MX 5 H 2 



Crystallographical data have already been published on 

 two salts belonging to this series, i.e. on Rb 2 TlCl5 H 2 and 

 Cs 2 TlCl 5 H 2 (see Pratt, loc. cit.). An isomorphous salt where 

 X =Fe has been examined by Johnson( 15) namely (NH 4 ) 2 FeCl 5 

 H 2 O. In the investigation of this series it was found that the 

 crystals should be placed in a different position from that 

 adopted by Pratt and Johnson, and it may be well to explain 

 here why an interchange of the crystallographical axes has 

 been made. The crystal system of the isomorphous series is 

 rhombic bipyramidal, but there is a close resemblance to 

 tetragonal habit. In the case of Cs 2 InCl 6 H 2 O, for instance, 

 the relative values of the crystallographical axes are "9841 : 1. 

 Unless there were some valid objection, the third axis, which 

 differs markedly from these two, would be taken as the c axis, 

 and the two specified axes as the a and b axes respectively. 

 Now, on examining the crystals of Rb 2 InCl 5 H 2 O, it was found 

 that practically every individual was twinned on the unit 

 prism face between the two almost equal axes as twinning 

 plane, so that the twinned individual was turned almost 

 through a right angle round the axis parallel to the prism face, 

 thus accentuating the pseudo-tetragonal symmetry. It was 



