PLANKTON OP THE GULF OP MAINE 
475 
suppose to have been present originally in the sea water employed in his tests, and 
second, because distilled water exercises much less solvent action than do the land 
waters with their load of dissolved organic compounds, humus acids, and Co 2 , which 
actually do the work of erosion on their way to the rivers and so to the sea; but, 
however slowly rock silicates are degraded in the sea, they are so degraded in the 
end. Indeed, all minerals, given time, finally succumb to the combined action of 
water, oxygen, and carbonic acid. Where a constant and rapid interchange of water 
between the bottom and the upper layers is kept up by vertical circulation (water, 
too, of low alkalinity — -that is, of comparatively high carbonic acid tension — as is 
the case on Georges Bank (p. 481) and in the Bay of Fundy) degradation of silicates 
will be more rapid than in the deeps, where, as Murray (1912, p. 187) points out, 
“the soluble by-products are removed and the supply of oxygen and carbonic acid 
maintained by diffusion only.” 
Furthermore, we must bear in mind that in the case of the degree of concentra- 
tion of silica we are dealing with solutions so attenuated that although the destructive 
action of sea water on felspathic rock fragments is almost inconceivably slow, it 
may be sufficient under hydrographic conditions as favorable as Georges Bank 
offers to yield the very small extra amount of silica needed to favor the active growth 
and multiplication of diatoms when added to what is in all sea water. Finally, the 
frustules of dead diatoms are themselves a potential store of this element and in one 
of its less insoluble forms. 
It is still to be proved that there is not always a sufficient supply of silica at all 
times and in all parts of the sea for the growth and multiplication of diatoms. But 
stress has often been laid on the apparent parallelism between the seasonal fluctua- 
tions in the concentration of dissolved silica which Raben (1905) reported for the 
waters of the North Sea and of the Baltic (Murray and Irvine’s earlier analyses 
are open to criticism) and the ebb and flow of the diatoms. Indeed, the corre- 
spondence between the two sets of phenomena, as it appears on Johnstone’s dia- 
gram (1908, fig. 30), is striking enough. Subsequent analyses made by Raben 
himself during the years 1904 to 1912 (Raben, 1905a to 1914), both for the central 
and eastern North Sea and for the western part of the Baltic, show that the seasonal 
fluctuations in the amount of silica are less regular than his earlier work suggested. 
But he again found maxima in February and November over the periods of years 
covered by the tests, the silica (Si0 2 ) content varying in the Baltic from 0.53 to 1.76 
milligrams per liter in February, to 0.40 to 0.93 in May, 0.20 to 1.49 in August, and 
0.93 to 1.36 in November, averaging as follows: 
Average silica (S 1 O 2 ) content in the western Baltic, 1902 to 1912 
Month 
Silica, 
milli- 
grams 
per liter 
Num- 
ber of 
analyses 
Month 
Silica, 
milli- 
grams 
per liter 
Num- 
ber of 
analyses 
February 
0.97 
19 
June 
0. 80 
2 
March... 
.83 
5 
August 
.86 
14 
April 
.65 
4 
November 
1. 17 
23 
May 
.69 
17 
