NANNOPLANKTON OF LAKE MENDOTA 83 



to indicate that some of the suspended material gradually settles out 

 of the water under these conditions because the ice keeps the water from 

 being disturbed by the wind. The results obtained on March 9 show a 

 further decrease in the quantity of silt thus indicating that the process 

 of settling continued during the intervening period of time. In some 

 winters, however, there is a very substantial contribution to the 

 amount of silt during this season. 



Thaws accompanied by rain sometimes occur in late January or in 

 February and much drainage water, well laden with silt, may reach the 

 lake at such times. This drainage water contains very little dissolved 

 material so that its specific gravity is less than that of the lake water. 

 For this reason the drainage water does not mix readily with the lake 

 water, but tends to spread out and form a layer of turbid water just 

 under the ice. The thickness of this stratum ranges from half a meter 

 to a meter, or perhaps a little more at times. In some years the spread- 

 ing continues until the entire lake becomes covered with this stratum of 

 turbid water, which results in a wide distribution of the suspended silt. 



The second part of table 23 gives the results of some analyses of the 

 ash of the centrifuge material. It will be noted that the addition of the 

 bowl water to the samples in 1916 and 1917 increased the percentages 

 of CaO and MgO very materially, but that this made practically no dif- 

 ference in the percentage of silica nor in that of iron and alumina. 

 Analyses of the solids of 14 samples of bowl water in 1915 showed that 

 CaO constituted from 16.2 per cent to 19.5 per cent of the total solids, 

 while MgO ranged from 10.7 per cent to 17.4 per cent. The percentage 

 of silica in the solids of the bowl water, as well as that of iron and 

 alumina, was relatively low; the former varied from 1.8 per cent to 3.4 

 per cent with one sample unusually high, namely, about 9.0 per cent, 

 and the latter ranged from 1.1 per cent to 1.7 per cent. These results 

 indicate that the diatoms and silt are pretty thoroughly removed from 

 the bowl water in the centrifuging process. Since the ash of the cen- 

 trifuge catches had a complex origin, any further discussion of its 

 chemical composition would not be profitable from a biological stand- 

 point. 



The Number of Organisms 



After being pumped into the tank, the water used for a centrifuge 

 run was thoroughly stirred and a representative sample of about two 

 liters was taken out for a numerical study of the organisms ; about 75 

 cubic centimeters of thfs sample were used for the enumerations and the 

 remainder was then returned to the tank. A few of the forms were 

 usually present in sufficient numbers to be counted directly without any 

 concentration; for the enumeration of the others it was necessary to 

 concentrate the material with a centrifuge. An electric centrifuge car- 



