Experimental design
To examine how shifts in relative emergence time varied with simultaneous changes in temperature and community contexts, we experimentally delayed the emergence of parasitoids relative to theDrosophila larvae in current and expected warming temperatures, and in high and low levels of resource competition in a fully factorial design with 4 emergence times x 2 temperatures x 2 levels of competition. Additionally, the two host species were reared in isolation or with the other host species to examine effects of an alternative resource. Thus, we had a total of three Drosophila species combinations and two wasp species for a total of 96 unique treatments. Each treatment consisted of at least five replicates and initiated over the course of five days, which were represented as blocks in our statistical analyses.
We reared Drosophila larvae in current ambient (22.9°C ± 0.47SD with 69% ± 4.34SD relative humidity) and predicted warming (27.4°C ± 1.15SD with 60% ± 10.1SD relative humidity) temperatures (climate change models predict a 1- 6°C increase in temperatures by 2070 in Australia (Hughes 2003)) in 12-hour light/12-hour dark photoperiod. Levels of intraspecific competition among Drosophila larvae were manipulated by providing 2ml or 20ml of fly medium. A total of 100 Drosophila eggs were added to all vials resulting in densities of 50 individuals/ml for the high competition, and 5 individuals/ml for the low competition treatments (Nouhaudet al. 2018). We decided to manipulate the volume of food instead of the abundance of larvae to avoid any frequency dependent parasitism effects. Thus, we established four temperature/competition treatments that varied host development rates: (1) slowest development (24°C with 2ml food); (2) slow development (24°C with 20ml food); (3) fast development (28°C with 2ml food); and (4) fastest development (28°C with 20ml of food). This process was repeated for both Drosophilaspecies in isolation and when grown together for alternative host species treatments, where the 100 total eggs were composed of equal proportions of the two species (e.g., 50 eggs D. sulfurigasterand 50 eggs D. birchii ).
We manipulated four levels of phenological relationships between host and parasitoids in 2-day intervals. The parasitoid either “emerged” at the same time (e.g., 0 day) as its host or 2-, 4-, or 6-days later. At each parasitoid emergence time, Drosophila larvae were exposed to three mated female and three male parasitoids of a single species of parasitoid. After 48-hours the parasitoids were removed from the vials and then maintained under the same temperature and light conditions described above. This allowed a more precise measurement of the vulnerability window and largely prevented the complete mortality of allDrosophila . To measure development times, emerges were recorded daily and stored in 95% ethanol until all hosts or parasitoids emerged (Fig. 1).
All experiments were set up in 2.8cm (diameter) x 9cm (height) glass vials, with a 4ml base layer of 1.5% agar gel in each vial to reduce excess desiccation of the fly medium and provide potential refuges from parasitoid attack. Control treatments consisted of ten unexposed vials for each host species combination (n = 3; D. birchii, D. sulfurigaster, D. sulfurigaster -D. birchii ). These unexposed vials were used to calculate the average development time (egg-to-adult) and host survival in the absence of parasitoids across treatments and species.