REPORT ON THE COMPOSITION OF OCEAN- WATER. 
229 
unless these are executed with the highest attainable precision. All the components must 
be determined in the style adopted for the lime (in the supplementary work) and for the 
bromine. I could not possibly have determined all the saline components in my 77 
waters by similarly refined methods for sheer want of material, and besides, the large 
number of analyses required would have rendered the work almost impracticable. 
What ought to be done is to collect waters at different times throughout the year 
at two stations, one might be selected somewhere in the middle of the Pacific, and a 
second at some place in the middle of the Atlantic Ocean. In each case two large 
samples should be taken, one from a little below the surface (to preclude abnormal 
dilution with rain-water), another at some 50 fathoms above the bottom to avoid ad- 
mixture of solid bottom matter, which in the bottle would gradually dissolve. 
Supposing we had, from each of the stations, six surface and six bottom samples, or 
twenty-four samples in all, we should begin by determining the chlorine in each sample 
a haute precision, and then do the same with the lime. The six samples from each 
place should then be mixed together (in equal volumes), so as to produce four samples, 
each representative of one of the four places. In each of them the chlorine, lime, 
sulphuric acid, magnesia, potash, and alkalinity should now be determined by at least 
triplicate analyses executed with the highest precision. Ships which happen to pass the 
localities might be instructed to collect samples as indicated, and bring them home. 
This would enable us, before trying to find out the difference between Atlantic water 
on the one hand and Pacific on the other, to inform ourselves as to the extent to which 
Pacific or Atlantic water at a given place is liable to vary. But before even this can 
be done successfully we must have sufficiently exact methods for the execution of the 
analyses. Hence, first and foremost, a chemist should be appointed to work out (by 
synthetical experiments in the first instance, and repeated analyses of some one sea- 
water in the second) a series of methods by means of which the sulphuric acid, magnesia, 
and potash could be determined with at least that degree of precision which I attained 
in regard to the lime. 
Another useful investigation would be the exact determination of the minor 
components (iodine, silica, fluorine, iron, aluminium, manganese) in a large mass of some 
one kind of sea-water. If a chemist succeeded in devising easy and yet sufficiently exact 
routine methods for determining one or other of these components, its comparative 
determination in different sea-waters might be undertaken. 
2. Alkalinity. — Tornpe’s method is sufficiently exact, and if applied to a very large 
number of judiciously-selected samples would be sure to give valuable results. 
3. Carbonic Acid . — In regard to the methods for determining the carbonic acid, 
there is room for much improvement. For oceanographic purposes, carbonic acid deter- 
