VARIANCE IN ZOOPLANKTON SAMPLES - EVALUATION OF A PREDICTIVE MODEL

We derived mean-variance relationships for zooplankton samples taken in the Hudson River and a series of Quebec lakes. Hudson River and Quebec lakes data were used to test a published model by comparing observed with predicted variance. In addition, we examined patterns of mean-variance relations (1) at higher levels of taxonomic organization, (2) between riverine and open water systems, and (3) between taxa common to the Hudson River and Quebec lakes data sets. There was an excellent correspondence between observed variance for the Hudson River and Quebec lakes data sets and the variance predicted using the model. Mean-variance relationships for higher taxa (cladocerans, copepods, rotifers) were similar to the general models. Zooplankton variance in the Hudson River was comparable with patterns observed in lentic systems contrary to our initial hypothesis that riverine zooplankton, because of rapid turbulent mixing, would have lower variance. Mean-variance relationships for taxa common to the Hudson River and Quebec lakes were comparable in three of six cases, but for the remaining three cases, regression intercepts were marginally different (P = 0.03-0.05) between systems. We conclude that zooplankton mean-variance relationships are highly consistent and predictable for a variety of aquatic systems and levels of taxonomic organization.