236 Proceedings of Royal Society of Edinburgh. [sess. 
formity when compared with those of the maximum group. It will 
he seen that the analyses of the filtered waters show silicic acid is 
present either in traces or only in quantity equal to one part in from 
220,000 to 460,000 parts of sea- water, or even in still more minute 
quantities. 
The determinations made by us at the Scottish Marine Station 
with carefully-filtered waters from different parts of the ocean led to 
similar results. The amount of soluble silica was so minute that it 
was difficult to believe it to he the exclusive source from which 
Diatoms and Radiolarians procured the silica for their frustules and 
skeletons, the results showing only one part of soluble silica present 
in from 200,000 to 500,000 parts of sea-water. 
In all attempts to determine the silicic acid, we filtered the sea- 
water through several folds of ashless filter-paper, or we added to 
it, in the cold, a solution of pure albumen, thereafter raising the 
liquid to a temperature of 212° E., so that the coagulated albumen 
which collected as a scum on the top of the boiling fluid carried 
with it any mechanically suspended matter present in the water. 
Even the ash from the apparently ashless filter may give rise to a 
profound error in the result, and in these determinations, in order 
to secure absolute accuracy, platinum vessels should be used.* 
The accurate determination of silicic acid in sea-water is compli- 
cated by another difficulty. This arises from the presence of 
fluorides, which in the ordinary methods for determining silicic acid 
would tend to form volatile fluoride of silicon, thus vitiating the 
results of the analyses. The results referred to above, showing the 
quantity of silicic acid in sea-water to amount to not more than one 
part in 200,000 to 500,000, were obtained by evaporating a weighed 
quantity of carefully-filtered sea-water to dryness with hydrochloric 
acid. The dried salts were drenched with hydrochloric acid and 
again heated to dryness, the insoluble residue left being taken as 
silicic acid. It is evident that in the presence of fluorides, if in suffi- 
cient amount, the whole of the silicic acid would he driven off, and 
pass away during the evaporation and subsequent drying and ignition, f 
* The balance used by us was not very delicate ; the results can only be 
relied on to the third place of decimals. 
+ Exp. (A). To determine this point we added silicic acid in a soluble form to a 
litre of artificial sea- water (which water was practically free from silicic acid), 
