Table S2 Regression Values Used to Calculate Zooplankton Biomass. Mean Length Values Were

Table S1 Information for all lakes involved in the experiment. Data was collected from May 2012 to August 2012 (S.E. Arnott, unpublished data). Lake Opinicon was the source of all mesocosm water, while zooplankton were collected from the other four lakes.

Table S2 Regression values used to calculate zooplankton biomass. Mean length values were derived from a N. Yan et al. unpublished database in which over 300 Ontario lakes were sampled, and all counted zooplankton were measured. References for regression coefficients are provided under the ‘Reference’ heading, and full references are listed following the table. A flat value of 0.2µg dry weight per individual was applied to the rotifers as we did not have mean length values for each species.

Bottrell, H.H., Duncan, A., Gliwicz, Z.M., Grygierek, E., Herzig, A., Hillbricht-Ilkowska, A., Kurosawa, H., Larsson, P. and Weglenska, T. 1976. A review of some problems in zooplankton production studies. Norw. J. Zool. 24: 419–456.

Culver, D.A., Boucherle, M.M., Bean, D.J., and Fletcher, J.W. 1985. Biomass of freshwater crustacean zooplankton from length-weight regressions. Can. J. Fish. Aquat. Sci. 42(8): 1380–1390. doi: 10.1139/f85-173.

McCauley, E. 1984. The estimation of the abundance and biomass of zooplankton in samples. In A manual on methods for the assessment of secondary production in fresh waters. Edited by J.A. Downing and F.H. Rigler (2nd edition). IBP Handbook 17, Blackwell Scientific Publications. pp. 228–265.

Rosen, R.A. 1981. Length-dry weight relationships of some freshwater zooplankton. J. Freshwat. Ecol. 1(2): 225–229. doi: 10.1080/02705060.1981.9664034

Table S3 Minimum adequate model (MAM) structure and transformations for all experimental metrics and species from the final sampling date. ‘PA’ and ‘Gradient’ entries under the ‘MAM’ heading represent the categorical and continuous predictors for, respectively, EPA and EGrad that were retained in the MAMs. R2 values represent the variance explained by the overall model. For EGrad, ‘Gradient’ indicates a retained linear term for Hemimysis abundance, ‘Gradient2’ a retained quadratic term, and ‘*’ a retained interaction.

Table S4 ANOVA tables for the minimum adequate models used for data from EPA.

Table S5 Regression tables for the minimum adequate models used for data from EGrad.

Fig. S1 Daytime average temperature (solid line) and overall average (dashed line) air temperature for the QUBS area over the period in which the experiment was conducted. Note that the field site was shaded, and cover was provided to all tanks, which would lead to cooler temperatures than those shown and reduce daily fluctuations.

Fig. S2 Mid-day temperature variability of all mesocosms from EPA (n=12) and EGrad (n=20) on the initial and final sampling dates for each experiment.

Fig. S3 Changes in abundance (initial – final individuals L-1) of rotifers from EGrad across the Hemimysis treatment gradient (0.011, 0.022, 0.033, 0.044, 0.055, 0.067, 0.078, 0.089, 0.01, 0.011 individuals L-1). Abundance values below zero indicate species that increased in density between the initial (day 0) and final (day 5) sampling dates, while values above zero indicate a decrease in density. Best fit lines are only shown for species whose density was related to Hemimysis abundance (P<0.1, plotted based on the minimum adequate models for these species). Note the changing y-axes scales for each panel.