Figure 1.
Feeding, lipid content, and survival of Gammarus pulex under constant exposure to imidacloprid.
Imidacloprid concentrations in medium (I), cumulative food consumption (II), mobile fraction of individuals, and lipid content (% of total wet weight) of Gammarus pulex (III) in the constant treatments (C) and controls of 14-day and 21-day experiments. Pie charts show the percentage of dead and immobile individuals amongst those removed from the beakers (non-mobile individuals = immobile+dead).
Figure 2.
Feeding and survival of Gammarus pulex under pulsed exposure to imidacloprid.
Imidacloprid concentrations in medium (I), cumulative food consumption (II), and mobile fraction of Gammarus pulex (III) in the pulsed treatments (A, B) and controls of 14-day and 21-day experiments. Pie charts show the percentage of dead and immobile individuals amongst those removed from beakers (non-mobile individuals = immobile+dead).
Figure 3.
Toxicokinetic model validation.
Internal concentrations were measured from immobile individuals in 14-day experiment (open squares) and from mobile individuals in additional beakers which were not used for observing mortality (crosses). These values are plotted with predictions of internal concentration (black line) by a previously published and calibrated toxicokinetic model [1].
Table 1.
Feeding activity of Gammarus pulex under constant (treatment C) or pulsed (treatments A and B) exposure to imidacloprid.
Table 2.
Mean percentage error (%) of individual tolerance (IT) and stochastic death (SD) model when pulsed (PT), constant (CT), or all data was used for calibration of the survival model for Gammarus pulex exposed to imidacloprid.
Figure 4.
Parameter estimates and fraction of mobile animals simulated with the individual tolerance distribution model.
Parameter estimates of individual tolerance models calibrated with all data, pulsed data only, and constant treatments only (I). Calibration to pulsed treatments (gray lines and gray symbols) and using these parameters to simulate the fraction of mobile animals in the constant scenario (black lines) are shown in the green box (II). Vice versa, calibration to constant treatments (black lines and black symbols) and using these parameters to simulate the fraction of mobile animals in the pulsed scenario (grey lines) are shown in the blue box (III). Symbols represent the data: black triangles are the mobile fraction in constant treatments (C), gray squares are data from pulsed treatments A and gray circles are from pulsed treatments B. Closed symbols are data from 14-day experiment and open symbols from 21-day experiment.
Figure 5.
Calibration of the starvation model.
The table shows the calibrated parameter values and their standard deviation.
Figure 6.
Simulation of survival of Gammarus pulex in constant imidacloprid exposure according to the chemical stress model (I), starvation model (II), and multiple stressor model (III).
In the starvation model (II), lack of food (LF) for the 14-day experiment was set to 1.0 and for the 21-day experiment 0.5 due to differences in feeding activity (no feeding in 14-day experiment, ca. 50% reduced feeding in the 21-day experiment). The chemical stress model was GUTS calibrated with pulsed toxicity data sets.