THE LITE OF DEVILS LAKE 39 



the chain, the plug was jerked from the cork and the water allowed 

 to enter in a steady stream as the air escaped thru the rubber hose. 

 "When oxygen samples were taken at the same time as the plancton, 

 the sample of the latter was taken from the larger bottle of the 

 Winkler apparatus used in collecting the former. 



Prior to September 4, 1913, surface and shore collections were 

 taken by immersing a bottle below the surface. Thereafter they were 

 made by scooping it up and transferring to a jar. Comparative 

 tests of the methods of sampling by means of the scoop, the bottle, 

 the Winkler apparatus and the pump have been made for the 

 surface, and in the case of the two latter methods, for the bottom 

 as well. In these tests from 48 to 100 cc. of the samples collected by 

 each method was carefully measured and the animals contained in 

 it counted by eye in a white porcelain dish. The method is not a 

 very accurate one, but is sufficiently so to demonstrate any con- 

 siderable or constant variation in the different methods. With the 

 scoop water was taken both from the surface and from seven to 

 ten centimetres below. With the Winkler apparatus and pump 

 the intake was necessarily held a few centimetres below the surface, 

 and as nearly as possible at the same level in all collections. The 

 samples were taken as nearly as possible in the same spot and with 

 brief intervals between each collection. The results of these tests 

 show no constant difference between the various methods, and such 

 differences as occur are smaller in general than those found by 

 Moberg (r.c.) in the horizontal distribution of the plancton. In 

 the shore samples, however, where the number of Crustacea was 

 large, there is a distinct advantage for the scoop method. 



A comparison of the results obtained by the plancton trap, 

 described by Juday (1916), the pump, Sedgwick-Rafter, centrifuge 

 and filter paper methods indicates no constant difference for the 

 Crustacea and rotifers in favor of either the Sedgwick-Rafter, pump 

 or trap methods, such differences as exist being inconstant and 

 doubtles due to variable distribution already referred to.* 



With the bacteria and algae the results vary. For the fila- 

 mentous forms (Nodularia and Lyngbya) the Sedgwick-Rafter 

 method gives distinctly larger results. For Coelosphaerium, 

 Oocystis, Merismopedium and Micrococcus there is a distinct 

 advantage in favor of the centrifuge ; while in the case of Dictyos- 

 phaerium, Chroococcus and Chaetoeeros, there is but little difference 

 one way or the other. 



In these experiments a small hand centrifuge, holding 15 cc. 

 of water in each tube and run for one minute at approximately 

 3,000 revolutions was employed and it was very difficult to throw 

 down all of the Nodularia. With a larger machine, operated at 



♦See pp. 36-8. 



