METHODS AND APPARATUS 15 



G, are placed between the corrugated plates and the perforations in 

 the plates are so arranged that the water passes through the filter 

 papers in its course through this chamber. The water then flows up- 

 ward through the passage indicated at I and is discharged from the 

 bowl at the point marked J. 



About 5.5 liters of water remain in the bowl at the end of a run 

 and this is siphoned off and used to wash oif the material that is de- 

 posited on the conical discs and on the sides of the bowl. A little 

 chloroform is added and the whole is then evaporated in porcelain 

 dishes at a temperature of about 60° C. About three quarters of the 

 evaporation is done on a sand bath after which the dishes are trans- 

 ferred to a Hearson electric evaporating oven for the completion of 

 the process. The dry material is then carefully removed from the 

 porcelain dishes, ground, and placed in a weighing dish. After stand- 

 ing in a desiccator for forty-eight hours it is weighed and transferred 

 to a bottle to await the chemical analyses. 



The water used for the centrifuge was obtained from the regular 

 station in the deep portion of Lake Mendota and it consisted of a cer- 

 tain quantity from each meter down to a depth of 20 meters. It was 

 strained through the large plankton net and was caught in containers 

 (milk cans) as it flowed from the large can in which the net was sus- 

 pended. It w^as then conveyed to the laboratory dock in the launch 

 where it was pumped into the large tank shown in figure 3. This tank 

 is made of galvanized iron and has a capacity of about 1,200 liters. It 

 is mounted on a framework which rests on a platform scale so that the 

 quantity of water used for each run is readily ascertained by weighing. 

 The scale also enables one to determine the rate at which the water is 

 centrifuged, the usual rate being about 10 liters per minute. The 

 framework elevates the tank to such a height above the centrifuge that 

 the water flows from the former to the latter through a hose, the rate 

 of flow being regulated by a valve. 



By far the greater portion of the work done on the centrifuge plank- 

 ton, or nannoplankton, as well as on the net plankton, was done on 

 Lake Mendota ; but for purposes of comparison some observations were 

 also made on Lakes Monona and Waubesa. The quantity of water 

 centrifuged at each run varied from about 1,100 liters to a little over 

 1,500 liters on Lake Mendota, but from the other two lakes it varied 

 from about 700 liters to a little over 1,500 liters. Figure 5 shows the 

 launch at the laboratory dock with a cargo of about 500 liters of water. 



During the open season, that is, from April to December, both in 

 1915 and in 1916, two centrifuge runs per week were made on Lake 

 Mendota when the weather was favorable. In most instances the ma- 

 terial obtained in the two runs was combined into one sample for the 



