622 ANDERSON AND LENAT 



MATERIALS AND METHODS 



Collection Techniques 



Zooplankton were collected through September of year 5 with a 

 12.6-cm diameter 20-mesh Birge closing net. Monthly duplicate 

 samples were collected at the 10 stations shown in Fig. 1 by taking 

 vertical hauls through the euphotic zone (defined as the depth to 

 which 1% of the surface illumination penetrated as measured by a 

 selenium cell underwater photometer). The depth of the euphotic- 

 zone collections encompassed and usually surpassed the depth of the 

 epilimnion (7 to 8 m) at all stations except 1 and 7. The entire 

 water column was sampled at these shallower stations. A larger 

 (24-cm diameter) 20-mesh net was used in October and November, 

 and, beginning in December and continuing throughout the rest of 

 the study, a 30-cm diameter 20-mesh net was used. The change in net 

 diameter was made to increase the chances of collecting zooplankton 

 that might have avoided the smaller nets. 



The zooplankton collected in each year were preserved in the 

 field with a 4% Formalin solution. A sample from one representative 

 station had a 1% Neo-Synephrine solution added before preservation 

 to narcotize the zooplankton. This sample was used to compile a 

 preliminary species list for the month and to aid in identifying 

 organisms that would contract when preserved with Formalin. 



Statistical Techniques 



We normalized the euphotic-zone zooplankton data first using a 

 log(n + 1) transformation of the averaged duplicate counts from each 

 monthly collection. This procedure aids in stabilizing variance. Two 

 separate sets of statistical analyses were performed. The first set of 

 tests examined significant differences between years for 19 taxa. The 

 second set of tests examined spatial and seasonal homogeneity for 

 these same taxa. All tests were conducted with data from eight 

 stations that were consistently sampled over the period of years 4 

 to 6. 



Differences between years were first tested for each taxon by 

 treating each station separately with a two-way analysis of variance 

 (ANOVA) using year and month as variables. This was followed by a 

 three-way ANOVA, using stations, year, and month as variables, to 

 assess lake-wide differences between years for each taxon tested. The 

 year x month interaction term was used as the error term in these 

 tests. The replicate data were tested for most of the species and did 

 not add significantly to the source of variation. Therefore the 



