322 Proceedings of Royal Society of Edinhergli. [sess. 
On the other hand, about one-third of the samples give values 
for D above 1*464, and ranging up to 1*483, while only some twelve 
or fifteen samples give values for D falling below 1 *449. To put 
the matter in another light : In the case of more than two-thirds 
of the 315 “Challenger” samples with which Dittmar made deter- 
minations of Xi there is to all appearance an entire absence of 
grounds for suspecting either that Buchanan’s specific gravity deter- 
minations were inaccurate, or that appreciable evaporation or 
chemical change had taken place during the years which elapsed 
between their collection and their examination by Dittmar. 
With regard to the remaining samples, we have no choice between 
assuming either an inaccuracy in the specific gravity determinations 
or some chemical change occurring between the time of their collec- 
tion and their examination by Dittmar, or, finally, that they were 
originally of appreciably different chemical composition. 
This latter assumption appears to me to have by far the greatest 
probability on experimental grounds alone, but this probability is 
increased when we consider the case in the light of a critical 
examination of the results of Dittmar’s full analyses of 77 of the 
“ Challenger ” samples of sea-water and in connection with geo- 
graphical and with vertical distribution. The full discussion of this 
side of the question I must reserve for a future communication, when 
I have completed all the necessary calculations, and have had time 
to consider each case separately. I have, however, made rough 
diagrammatic representations of the distribution of the “Challenger ” 
waters giving normal and abnormal values for D and f ~ X* 
charts used for this purpose, showing the course of H.M.S. “ Chal- 
lenger,” were kindly given to me by Dr John Murray. These 
I lay before the meeting. 
