482 
BULLETIN OF THE BUREAU OF FISHERIES 
With the temperature of the upper strata in the coastal waters of the Gulf only 
about 1 to 0.5° at its annual minimum when the vernal flowerings of diatoms com- 
mence, but rising to upwards of 18° off Massachusetts Bay and even to 20° locally in 
the center of the gulf in August, the viscosity decreases, say, by 40 per cent (from 
about 100 to about 60) during the spring and early summer. Consequently, other 
things being equal, a diatom would sink four-fifths faster in midsummer than during 
the first days of spring. Other things are not equal, however, because the specific 
gravity of the water as well as its viscosity decreases with the rising temperature and 
with diminishing salinity of spring. Thus, the surface stratum is not only a thinner 
fluid but a lighter one absolutely in summer than in winter, which makes for a still 
greater disparity between the tendency of diatoms to sink in the cold and in the warm 
seasons. 
It would, perhaps, be safe to say that differences in specific gravity of the water 
and in its viscosity would necessitate twice as active vertical circulation to hold any 
given object in suspension in summer as in early spring. As we have seen, however, 
(p. 481), the reverse actually obtains, the active vertical mixing characteristic of spring 
giving place to a condition of comparative vertical stagnation in midsummer, con- 
sequent on the increasing vertical stability of the water, which must increasingly hinder 
the flotation of diatoms in the gulf, just as happens in the fresh-water lakes described 
by Whipple (1905). Thus the seasonal cycle of viscosity and of vertical circulation 
combined tends to put a period to the seasonal multiplication of the species of diatoms 
which are characteristic of spring by increasing their tendency to sink. 
In the preceding pages I have tried to show that on theoretic grounds the 
gulf, taken as a whole, offers its most favorable environment for planktonic diatoms 
in spring, because of the following combination of circumstances: The supply of 
two of the nutrients on which it is probable that diatoms chiefly depend — nitrogen 
and silica — is then greatest. (European analyses suggest that this also applies to 
phosphoric acid.) The circulation of the water then tends to bring up a supply of 
nitrogen compounds and of dissolved phosphates most actively from below, the 
high viscosity of the water then most favors the flotation of diatoms, and the increas- 
ing strength of the sunlight from late winter on increasingly favors the processes 
of photosynthesis. It is probable that for abundant flowerings of diatoms all 
these requirements must be satisfied. Conversely, fluctuations in the amount of 
•any one of the essential foodstuffs may govern the amounts of diatoms actually 
present at any given time or place, and may even terminate the flowerings if it fall 
below the requisite minimum. 
The parallelism that has actually been shown to exist between the fluctuations 
in the concentration of silica in the sea water of Massachusetts Bay and of the diatoms 
there (p. 476, fig. 134) makes this our most suggestive illustration. Without the 
accumulation of this substance (which takes place during the winter when there are 
few diatoms to make use of it) the tremendously productive flowerings which we have 
encountered in spring probably could not take place, any more than they could 
unless there were enough nitrogen in available form to nourish them. But after the 
flowerings have abounded for a few weeks in this particular location they so reduce 
the supply of silica (as the analyses show) by converting it into an unavailable 
