Transgenerational effects increase the vulnerability of a host-parasitoid system to rising temperatures

Abstract

Transgenerational effects, non-evolutionary processes by which environmental conditions in one generation influence the performance in subsequent generations, are hypothesised to have substantial consequences for population dynamics under stochastic environments. However, any direct apparent detriment or advantage these processes generate for a focal species may be counteracted by concurrent effects upon interacting species. Using an experimental Drosophila-parasitoid model system, we determined how the previous generation's thermal environment impacts the thermal performance of both hosts and parasitoids. We found substantial responses in both trophic levels, with potential evidence for both condition-transfer effects and adaptive transgenerational plasticity. We used these results to parameterise discrete-time simulation models to explore how transgenerational effects of thermal conditions and temporal autocorrelation in temperature are expected to impact the time to extinction for this host-parasitoid system under climate change. The models predicted that transgenerational effects would significantly hasten the time to extinction, largely through a reduction in estimated average performance. Under the assumptions of one of the population dynamics models trialled, we identified an additional hastening of extinction from the combined effect of both host and parasitoid transgenerational effects. Our research demonstrates how community-level consequences of transgenerational effects may impact a population's sensitivity to climate change under a fluctuating environment and highlights the need to quantify and contextualise thermal transgenerational effects in their ecological setting.

Keywords: Drosophila; climate change; host‐parasitoid; modelling; population dynamics; transgenerational.

FIGURE 1

Summary of experimental design for the fly temperature treatment experiment. G0 = Generation 0. G1 = Generation 1. Large numbers of G0 bottles were initiated with approximately 20 flies and randomly assigned to incubators set at one of three G0 temperatures (19, 23 and 27°C). After 24 h to lay eggs, the adult flies were removed. Once G0 flies emerged and matured, two males and two females were counted into each acclimatisation vial. Acclimatisation vials were randomly assigned to incubators set at one of five temperatures (19, 21, 23, 25 and 27°C) in a fully factorial design (n = 40 per combination). After 24 h of acclimatisation, flies were transferred into standardised test vials that remained in their assigned G1 test temperature to lay eggs for 24 h. G1 emergences were regularly removed and counted.

FIGURE 2

Summary of experimental design for wasp temperature treatment experiment. G0 = Generation 0. G1 = Generation 1. To initiate G0, two male and two female wasps were added to standardised vials of hosts. The vials were stored in incubators at either 19, 23 and 27°C. Two male and two female wasps that recently emerged were transferred to G1 host vials randomly assigned to G1 temperatures (19, 21, 23, 25 and 27°C), in a full factorial design. Each treatment combination had between 16 and 20 replicates. As G1 wasps and flies emerged, they were removed, stored and then counted.

FIGURE 3

Fly reproduction best fit thermal performance curves and raw data. Each panel shows the effect of temperature during that generation (G1) on the number of emerged offspring, for each temperature in the preceding generation (G0). See (Figure 1) for how generations were delineated. Black points represent raw data from each vial and are jittered horizontally. Means and standard errors for each treatment combination are given in Figure S3. Blue lines represent central expectations based on the best fit thermal performance curve model. Parameter estimates for the thermal performance curves and 95% confidence intervals are in Supporting Information (Table S1).

FIGURE 4

Wasp emergence rate best fit thermal performance curves and raw proportion data. Each panel shows the effect of temperature on the proportion of wasps (wasp count/total count, where total count includes both emerged flies and wasps) for each previous generation temperature (G0). See Figure 2 for delineation of generations. Black points represent raw data from each vial and are jittered horizontally. Means across the vials and approximations of standard errors for each treatment combination are given in Figure S4. Blue lines represent predictions based on best fit model. Fitted model parameter estimates and 95% confidence intervals are listed in Supporting Information (Table S2).

FIGURE 5

Identification of emergent dynamic impact of combined transgenerational effects. Density histograms of the observed distribution of extinction points over 100 simulations of the host–parasitoid system under different scenarios and models. The central temperature on which noise is imposed rises by 1°C per 400 generations after an initial 500 generation burn‐in period to allow the system to settle. Differences arise with joint inclusion of autocorrelation and both transgenerational effects. There is a near perfect overlap between scenarios with transgenerational effect modelled for ‘wasp‐only’ and ‘none’ (blue and green, combining to dark turquoise), as well as ‘both’ and ‘fly‐only’ (yellow and red, combining to orange), with the exception of the top left panel (simpler functional response and autocorrelation). Geometric means are given in Table S4. Note the logarithmic x‐scale.