1. It is well known that animals can increase their stress resistance by prior exposure to sub-lethal conditions, but this acclimation process is often accompanied by deleterious fitness effects. These may reflect a cost of acclimation, or more general costs arising from damage due to sub-lethal exposures.
2. Acclimation for increased adult resistance to a 40 °C heat shock was tested in the egg parasitoid, Trichogramma nr. brassicae, following exposure of immature stages to 33 °C for different periods. Prepupal exposure did not increase heat resistance, but daily exposures at the pupal stage for 2 h day−1 and 3 h day−1 over 4 days increased survival of the shock.
3. Pre-pupal exposure for 3 h day−1 over 4 days led to decreased parasitism reflecting general damage costs. Pupal exposure for 3 h day−1 and 4 h day−1 also decreased parasitism rates, but there were no deleterious effects when pupae were acclimated for 2 h day−1.
4. These findings indicate that acclimation can occur without costs in Trichogramma nr. brassicae and suggest a general phenotypic approach for separating acclimation costs from other fitness costs caused by sub-lethal exposures. The potential for using acclimation to increase field parasitism success of this important group of parasitoids is discussed.
Costs that occur following exposure to sub-lethal conditions may be associated specifically with the acclimation process (‘acclimation costs’) or result from general damage unrelated to this process. For instance, insects exposed to high temperatures may suffer water loss when their epicuticular layer is damaged, an effect unrelated to heat acclimation costs such as those associated with heat shock protein production. One way of identifying acclimation costs is to relate them directly to the physiological mechanism underlying the acclimation response. For instance, in Drosophila, Krebs & Feder (1997) linked variation in one of the heat shock proteins (hsp70) to variation in larval survival; lines which showed increased expression of hsp70 had increased larval thermotolerance, but also suffered a reduction in larval viability. Unfortunately, this approach is rarely feasible because mechanisms and natural variation underlying acclimation responses are usually unknown.
Another approach for testing acclimation costs is to examine the acclimation response in detail and determine if there are conditions that lead to acclimation without any costs. If such conditions exist, then any deleterious effects due to sub-lethal exposures are likely to have been caused by general damage rather than the acclimation process. This approach is particularly useful for testing costs associated with thermal acclimation where fitness benefits may only require a short exposure to sub-lethal conditions.
In this study the second approach is used to examine costs for heat resistance acclimation in an egg parasitoid, Trichogramma nr. brassicae. This species has been reared commercially for control of pest moths, in particular Helicoverpa species, and is responsible for high rates of field parasitism of Helicoverpa in tomatoes and other crops ( McLaren & Rye 1983). There has been little work on acclimation in Trichogramma, even though these parasitoids form an important component of integrated pest management (IPM) strategies in many countries and despite the fact that high temperatures decrease Trichogramma parasitism rates ( Chihrane et al. 1991 ). In Trichogramma carverae, Scott et al. (1997) found that exposure of pupae to 33 °C for a total of 13 h or more led to an increase in resistance to heat at the adult stage. They also found that adult resistance to heat could be increased by hardening the adults at sub-lethal temperatures. However, pupal treatments decreased adult parasitism at 25 °C, suggesting that any benefit of acclimation could be offset by costs.
The focus of this study is on acclimation across life-cycle stages rather than short-term hardening of adults. Three specific questions are considered. Firstly, can adult resistance to a high temperature shock be increased by prepupal acclimation and/or by pupal acclimation in T. nr. brassicae? Secondly, are there other benefits of acclimation? Thirdly, are treatments that increase resistance associated with costs in terms of decreased parasitism or decreased longevity, and can these costs be overcome by modifying acclimation treatments?
Materials and methods
STRAIN AND ACCLIMATION
Experiments were undertaken with a mass-bred strain of T. nr. brassicae which had been initiated by combining 17 isofemale lines started by offspring from single females collected from a number of locations in south-eastern Australia: Rochester, Yarra Ridge, Tatura, Shepparton and Swan Hill. This strain took 9–10 days to complete development at 25 °C, comprising an egg-larval stage lasting 4–5 days and a pupal stage lasting 4–5 days. Wasps were reared on eggs of the Grain Moth, Sitotroga cerealella, at 25 °C.
The acclimation treatments involved exposures to a sub-lethal temperature (33 °C) for a few hours per day over several days. This procedure was favoured over a single continuous acclimation period because (1) wasps in southern Australia where this species occurs typically encounter maximum temperatures for only a few hours per day during the afternoon, and (2) a long exposure to a high temperature induces deformities in offspring as well as sterility. Acclimation was undertaken at the early stage of development or after pupal formation at the end of development.
For prepupal acclimation, wasps were left to parasitize S.cerealella eggs for 24 h. Parasitized host eggs were then exposed to 33 °C for a total of 0, 2 and 3 h day−1 over the first 4 days of development (i.e. 0, 8 and 12 h in total). A longer period of 4 h per day was also initially considered, but this treatment led to a high percentage (> 75%) of wasps emerging with deformed wings and was therefore abandoned.
For pupal acclimation, parasitized host eggs were exposed to 33 °C during the last 4 days of development (7–10 days after parents had parasitized host eggs) for 0, 2, 3 or 4 h per day (i.e. 0, 8, 12 and 16 h in total). To acclimate pupae in host eggs, the eggs were placed in plastic 38-ml vials sealed with Parafilm and placed in a water bath. In control treatments, host eggs were also transferred to vials but were not placed in the water bath.
SURVIVAL OF HEAT SHOCK
Females (1–20 h postemergence) were stressed for 4·75 h at 40 °C because pilot tests showed that this stress resulted in intermediate mortality levels. Wasps were stressed in sealed 12-ml glass vials (6–10 per container) as for acclimation. Survival was scored after 24 h. A few (< 5%) of the wasps in vials were alive but immobile at this time; these were scored as dead because they never recovered. For the experiment on prepupal acclimation, all treatments were scored at the same time. For the experiment on pupal acclimation, each of the treatments was tested at a different time against a separate control and data therefore could not be combined across controls. There were seven to ten replicates per treatment.
Wasps emerging from host eggs were tested for two fitness components, parasitism rate and longevity. Parasitism rate was considered under three temperature conditions: 25 °C, 30/25 °C and 37/25 °C. In the 30/25 °C and 37/25 °C conditions, wasps were placed at the higher temperatures for 5 h per day and at 25 °C the rest of the time. Mated females (1–20 h postemergence) were placed individually in 12-ml glass vials with a drop of honey and at least 100 S. cereallela eggs on a card. For the 25 °C and 30/25 °C conditions, cards were replaced after 2 days. For the 37/25 °C condition, a new card was introduced each day when vials were placed at 37 °C because eggs on cards tended to dry out more quickly in this treatment compared with the others. Because more eggs were provided than the wasps could have parasitized over 2 days, replacing cards daily or every 2 days is unlikely to have influenced total parasitism rate.
Parasitism was assessed 5 days after cards had been exposed to the wasps. Parasitism was quantified by counting the number of black unhatched eggs (indicative of parasitoids developing inside hosts). For the experiment on prepupal acclimation, there were 12 replicates per acclimation treatment and 24 replicates per unacclimated control treatment for each of the three temperature conditions. For the experiment on pupal acclimation, there were 25 replicates per acclimation treatment and 25 per control treatment for each of the three temperature conditions.
Longevity of mated females was scored in similar temperature conditions. However, longer exposure times to high temperatures were used in an attempt to increase the stress to which wasps were exposed. A longer exposure time could be used than for parasitism rate; for parasitism there had been a concern that deformities arising from high-temperature exposure may have led to an underestimate of this trait. For the 30/25 °C condition, wasps were placed for 10 h per day at 30 °C in the case of both the prepupal and pupal treatments. For the 37/25 °C condition, wasps were placed for 8 h (prepupal acclimation) or 5 h (pupal acclimation) per day at 37 °C. To score longevity, wasps were placed individually in 12-ml glass vials with honey, and examined daily for mortality. There were 30 wasps per treatment for prepupal acclimation, and 25–35 wasps per treatment for pupal acclimation.
Survival following prepupal acclimation could not be evaluated with parametric statistics because survival was low for all treatments ( Fig. 1). A non-parametric Kruskal–Wallis test showed that survival did not differ significantly (P = 0·22) between treatments. Hence there was no evidence that prepupal treatments increased heat resistance.
In contrast, pupal acclimation influenced resistance ( Fig. 2). For the 2 h day−1 exposure, differences between treatments were significant by a t-test on arcsin transformed proportions (t = 8·59, df = 7, P < 0·001 when corrected using the Dunn-Sidák method for the three comparisons of pupal acclimation). Treatments also differed significantly (t = 6·62, df = 7, P < 0·001 after correction) for the 3 h day−1 comparison, but the 4 h day−1 exposure did not influence adult survival (t = 1·17, df = 7). Pupal acclimation therefore increased heat resistance but only for exposure periods of 2 h day−1 and 3 h day−1.
In the experiment on prepupal acclimation, exposure of the parasitized eggs to a high temperature for 2 h day−1 did not decrease parasitism rates, in contrast to the 3 h day−1 treatment ( Fig. 3). A two-way anova (with temperature condition and acclimation treatment as the factors) indicated a significant effect of acclimation (F2,135 = 31·28, P < 0·001) and a marginally significant effect of temperature condition (F2,135 = 3·69, P = 0·03) on parasitism rate, while the interaction between these factors was not significant (F4,135 = 1·63, P = 0·17). Posthoc tests (Tukey B) showed that while the control and 2 h day−1 treatments did not differ significantly, the 3 h day−1 treatment resulted in a lower parasitism rate regardless of testing conditions. Posthoc tests also indicated a significant difference between the 37/25 °C condition and the other two conditions, suggesting that exposure to 37 °C for 5 h day−1 had a negative effect on parasitism rate.
In the experiment on pupal acclimation, parasitism rates tended to be higher in the 30/25 °C condition than the other two conditions. Heat exposure longer than 2 h day−1 decreased parasitism ( Fig. 4). The anova indicated a significant effect of acclimation (F3,288 = 9·20, P < 0·001) and temperature condition (F2,288 = 6·28, P = 0·002) on parasitism rate, while the interaction between these terms was not significant (F6,288 = 1·04, P = 0·40). Posthoc tests (Tukey B) indicated that parasitism rates did not differ significantly for the 3 and 4 h day−1 treatments, while the controls and 2 h day−1 treatments had a higher parasitism rate and did not differ significantly from each other. Hence short pupal acclimation periods do not appear to adversely affect parasitism rates under any of the tested temperature conditions.
For the prepupal experiment, longevity was reduced by daily exposures to 30 °C and particularly 37 °C ( Fig. 5). Unfortunately all the data could not be analysed in a single anova because the low longevity scores from wasps placed at 37/25 °C fell into a narrow range. The anova on data from the 25 °C and 30/25 °C conditions indicated a significant effect of temperature condition (F1,171 = 11·83, P < 0·001) but no effect of acclimation (F2,171 = 0·27, P = 0·76) or interaction between these factors (F2,171 = 1·03, P = 0·36). For the 37/25 °C data, there was also no difference between the treatments in a non-parametric Kruskal–Wallis test (χ2 = 1·81, df = 2, P = 0·40).
For the longevity experiment following pupal acclimation, a single anova could be carried out because longevity scores were higher in the 37/25 °C condition than in the previous experiment (the 37 °C exposure period was shorter) and fell into a wider range. To remove an association between means and variances, data were log-transformed prior to analysis. The anova indicated a significant effect of temperature condition (F2,385 = 134·5, P < 0·001) and acclimation (F3,385 = 8·05, P < 0·001), while the interaction between condition and acclimation was also significant (F6,385 = 2·97, P = 0·007). The posthoc tests indicated significant differences between the 2 h day−1 treatment and the other three treatments including the controls, as well as differences among the three temperature conditions ( Fig. 6). Thus there was a beneficial effect of acclimation for 2 h day−1 on longevity. This effect was noticeable under the 30/25 °C condition but was small for the other two temperature conditions.
The results indicate that there are beneficial effects of acclimation across life-cycle stages in T. nr. brassicae. Survival under heat stress was increased when pupae (but not eggs/larvae) were exposed to 33 °C. There were no beneficial effects of acclimation on parasitism rate. However, longevity was increased by pupal acclimation for 2 h day−1, particularly when wasps were exposed to periods of 30 °C. In another Trichogramma species, T. carverae, exposure to 33 °C at the pupal stage increased adult survival ( Scott et al. 1997 ), although in this case a cumulative exposure longer than 8 h was required to produce a beneficial effect. Conditions leading to beneficial acclimation effects therefore differ between Trichogramma species.
Four of the five treatments provided evidence for costs whereas a beneficial effect of acclimation was detected in only two of them. These results generally support the notion that exposures to specific conditions do not necessarily increase fitness under those conditions ( Leroi et al. 1994 ; Huey & Berrigan 1997) and contradict the Beneficial Acclimation Hypothesis. Nevertheless in one treatment (pupal acclimation for 2 h day−1) there was a beneficial effect without fitness costs as measured by parasitism rate and longevity. Acclimation without costs may also occur in other Trichogramma; while Scott et al. (1997) found that pupal acclimation in T. carverae always resulted in a reduced parasitism rate at 25 °C, recent data (M. Robinson and L. Thomson, unpublished data) suggest that costs may disappear when acclimation is confined to a late pupal stage.
Although the results suggest that costs are common when organisms encounter sub-lethal conditions, they have implications for the notion that acclimation is invariably associated with fitness costs ( Hoffmann 1995). Instead the results suggest that it may be possible to identify conditions leading to acclimation without costs, and that costs may be associated with general damage rather than the effects of the acclimation response per se. Whether this will apply to other insects is unclear. The findings of Krebs & Feder (1997) associating Hsp expression with decreased larval survival suggest that increased thermotolerance in Drosophila larvae often involves a cost because Hsp 70 expression plays a major role in larval acclimation.
A problem with the present study concerns the number of treatments that were considered in establishing conditions leading to benefits (and costs), which could lead to significant results being obtained by chance. In the experiments, there were 42 treatment–trait–environment combinations and 12 treatments for heat resistance. To counter this problem, some adjustment needs to be made to significance levels when testing effects. In the present case, overall effects of acclimation were tested in four anovas, two Kruskal–Wallis tests and three t-tests. To obtain a significant effect at the 5% level, probabilities would have to be less than 0·005 after correction for the number of comparisons (9) by the Dunn-Sidák method. All acclimation effects remain significant at this level and even if an adjustment is made for the total number of treatments (54).
Another problem with this study and other acclimation studies is that costs and benefits have been tested for only a limited number of traits and under laboratory conditions. Only parasitism rate and longevity were examined. These are important aspects of fitness, but other fitness components could also be affected by acclimation responses such as host location and mating success. In Trichogramma field fitness will depend on the ability of adults to mate and find moth eggs as well as on parasitism rate and longevity. It should be possible to test some of these components with field releases using acclimated wasps. There are established designs for assessing field host location and parasitism (see Kazmer & Luck 1995; Bennett & Hoffmann 1998).
A further limitation of this study is that it was not possible to separate costs due to environmental effects on parasitoids from their effects on the hosts. Although the acclimation treatments used did not influence the hatch rates of unparasitized host eggs, it is possible that there were sub-lethal host changes that diminished their suitability for the parasitoids. It is difficult to separate these effects unless developing parasitoids can somehow be acclimated independently of their hosts. One solution is to acclimate them on a completely artificial medium, but this option is not yet available for Trichogramma.
Although further work is required, the possibility of acclimation without costs has implications for the success of commercially produced Trichogramma in controlling moth pests. In south-eastern Australia, T. nr. brassicae is released for the control of pests on tomatoes, sweet corn and other crops. The Trichogramma are typically released as parasitized host eggs just prior to adult emergence, under air temperatures that can exceed 40 °C during summer months. The results of this study suggest that parasitism success under such conditions could be improved by exposing pupae to 33 °C just prior to release. This simple process could be undertaken by the commercial producer prior to shipment, or by growers after they have obtained the parasitized eggs.
This work was supported by grants from the Australian Tomato Processing Industry Research Council, the Horticultural Research and Development Corporation and the Australian Research Council.
Received 17 March 1999; revised 2 June 1999;accepted 1 July 1999