56 
results arrived at by the Professor differed so much from 
those obtained by the author that a careful examination 
of the Professors paper became necessary before any 
progress could be made in the reduction of the author’s 
observations, as, if the results of the Professor were correct, 
the observations must involve such errors that their reduc- 
tion would be useless. The examination was therefore 
undertaken, and was extended to the results of the Ameri- 
can observations, which are shown in a curve given among 
the plates to Maury’s Physical Geography of the Sea (9th 
edition). The American observations were found to accord 
with those of the author and to differ from the results obtained 
by the Professor to an extent in some cases of one per cent 
of the total salinity. The apparent cause of the discrepancy 
seemed to be the length of time during which some of the 
Professor’s specimens were kept before being subjected to 
analysis, the paper in the Philosophical Transactions showing 
that in one case at least this interval amounted to some 
twenty years. A further examination of the paper showed, 
as the author conceived, great errors in arrangement, and 
especially in the determination of the areas, all the water 
taken in which, was assumed to be from the same region, 
and all the results of which specimens were included in a 
common average. There was a strange deficiency in the 
observations for certain regions, but the boundary lines of 
the areas were so arranged that the Gulf Stream was 
included in three regions, the L’ Agulhas current in two, 
and the Equatorial current in two at least ; so that if the 
results had not been drawn from too few specimens, the 
effect of the high specific gravity of these currents would 
have been entirely masked by the observations of other 
water not the subject of current, but included in the 
averages as belonging to the same regions. As it was, the 
same result, viz. the disappearance of the effect on the 
average of each region of the high specific gravity of the 
