d20 Dr. E. J. Mills on Elective Attraction, 



stantly unlike in their povrer of producing a residue in ^vhicll 

 chlorine and that oxide are relatively retained. The numbers to 

 ^Yhich a corresponds are ratios ; and the formal process vrhereby 

 they are obtained may therefore be termed '^ the method of 

 latios/^ It must be distinctly understood that « is a ratio and 

 not an absolute quantity ; it may be compared to interest, while 

 the nitrate corresponds to a stock from which interest has been 

 derived. The dynamical z/se that can be made of a nitrate is re- 

 presented by the svmbol «. The data from which a has been 

 deduced (namely, certain weights of argentic chloride and mag- 

 nesic pyrophosphate) are, if singly considered, new with each 

 experiment ; they depend on time, rate of heating, the state of 

 division of the nitrate, and on the incidence of other and even 

 minuter conditions. But assuming the results to have been 

 brought about under a law of chemical action, the value of a 

 must be independent of those circumstances, which could only 

 pari passu affect the primitive numerator and denominator; it 

 is essentially related only to the actual occurrence of the reac- 

 tion. This is very well shown in the case of baric nitrate_, from 

 which neither chloride nor phosphate was obtained, and hence 

 a = 0, in accordance with which result it is found that no other 

 substance was formed ; in other words, no reaction occurred. 

 In the following Table !a stands for the symbolic value (atomic 



weight) of a nitrate, and Q= — • 



a. 2. Q. 



fThallous nitrate . . 8-76 265-30 SO^O"! 



■I Ar-entic ,, . . 5-48 169-94 31-01 l 



[Pliimbic „ . . 5-17 165-56 32-02] 



Rubidic „ . . 2-38 147-40 61-93 



Csesic „ . . 2-21 195-01 88-24 



JPotassic „ . . 1-99 101-14 50-821 



i Scdic ,, . . 1-70 85-05 50-03/ 



Lithic „ . . 1-61 69-00 42-86 



With regard to the above Table, I may remark that it com- 

 plies with the logical requirement of a common series by refer- 

 ring all the nitrates to the same weight of oxynitryl (^0^); for 

 the salts employed have only been considered as nitrates^ and 

 not in any other relation. Hence the value for plumbic nitrate 



is ^ — - =165-56. Again, by reference to the analysis of 



the thallic residue, it will be observed that in that particular case 

 a dichloride has been produced instead of a mcncchloride as in 

 the other cases. Hence thallous nitrate has simulated thallic 

 dinitrate. But the formula of the nitrate T1(IS0^;'^ corresponds 



