390 Bee GA EING 
sea-water was made having as nearly as possible the composition of 
natural sea-water. It was, of course, also possible to make up artificial 
sea-waters of different compositions, one or other of the constituent 
salts being increased or diminished in amount, or omitted altogether. 
What we may call the normal artificial sea-water, that is, the one with 
a composition as nearly as possible that of natural sea-water, had the 
following constitution :— 
Sodium chloride. : : . 28-13 grams per litre. 
o 
oO 
Potassium chloride ; 5 “Aut * 
Calcium chloride. ; er ee. 1-2 Ap ¥ 
Magnesium chloride 2D Wass : 
Magnesium sulphate —. ; eos) * ; 
Sodium bicarbonate : : é 2 oil 
ss 
To this there was added in some of the experiments a trace of Potassium 
iodide and Potassium bromide, but the results did not seem to be affected 
by this addition. 
There is another substance which requires careful consideration when 
we are dealing with diatom cultures. The skeleton of a diatom is composed 
of Silica, so that to get a healthy growth that substance must be supplied. 
An Austrian botanist, Richter, has shown that when cultures are made 
in glass vessels, enough silica dissolves from the glass to supply the dia- 
toms with all they require. In my experiments [ found that the addition 
of silica in other forms to the culture solutions appeared to make no 
difference to the growth. We may therefore in what follows disregard 
the silica, remembering that all the needful supply of it could be obtained 
from the glass of the flasks in which the experiments were made. — - 
Having made the artificial sea-water in the way described the essential 
constituents of Miquel’s solutions—Potassium nitrate, Sodium phosphate 
and iron——-were added. After boiling and cooling, the flasks containing 
the solution were inoculated by adding a small quantity of a culture of 
the diatom Thalassiosira, which was already growing in natural sea- 
water. The flasks were then placed in a good light and the cultures 
given an opportunity to develop. 
In the early experiments the results were very uncertain and difficult 
to understand. In most cases there was an entire failure of growth, but 
every now and then quite a good growth was obtained. It was noticed 
also that a good growth more frequently resulted when a flask which 
had failed was inoculated a second or a third time. It remained for a 
long time a puzzle why the cultures should generally fail but occasion- 
ally succeed, until it occurred to me that the quantity of natural sea- 
