SEARCH

SEARCH BY CITATION

Keywords:

  • acclimation;
  • flight mill;
  • Grapholita molesta ;
  • insects;
  • parental temperature;
  • phenotypic plasticity;
  • range expansion

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References
  9. Supporting Information

Nongenetic parental effects may affect offspring phenotype, and in species with multiple generations per year, these effects may cause life-history traits to vary over the season. We investigated the effects of parental, offspring developmental and offspring adult temperatures on a suite of life-history traits in the globally invasive agricultural pest Grapholita molesta. A low parental temperature resulted in female offspring that developed faster at low developmental temperature compared with females whose parents were reared at high temperature. Furthermore, females whose parents were reared at low temperature were heavier and more fecund and had better flight abilities than females whose parents were reared at high temperature. In addition to these cross-generational effects, females developed at low temperature had similar flight abilities at low and high ambient temperatures, whereas females developed at high temperature had poorer flight abilities at low than at high ambient temperature. Our findings demonstrate a pronounced benefit of low parental temperature on offspring performance, as well as between- and within-generation effects of acclimation to low temperature. In cooler environments, the offspring generation is expected to develop more rapidly than the parental generation and to comprise more fecund and more dispersive females. By producing phenotypes that are adaptive to the conditions inducing them as well as heritable, cross-generational plasticity can influence the evolutionary trajectory of populations. The potential for short-term acclimation to low temperature may allow expanding insect populations to better cope with novel environments and may help to explain the spread and establishment of invasive species.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References
  9. Supporting Information

Phenotypic variation among individuals is the raw material upon which natural selection acts (Mousseau & Fox, 1998a). The investigation of the causes and consequences of phenotypic variation is therefore not only of functional interest, but also of evolutionary relevance. Phenotypic variation can result from either genetic variation or phenotypic plasticity (Pigliucci, 2001). Phenotypic plasticity refers to the ability of an organism to adjust its phenotype to prevailing environmental circumstances (Pigliucci et al., 2006). Many studies have investigated the effects of environmental variation on phenotypic expression within an individual's lifetime, especially in relation to alternative pathways and various (e.g. seasonal) polyphenisms (Nijhout, 2003; Simpson et al., 2011) or to the phenomenon of acclimation, that is the plastic phenotypic response to the environmental conditions experienced during development (Huey & Berrigan, 1996; Bennett & Lenski, 1997). However, it is becoming increasingly evident that the phenotype of an individual can be affected by environmental factors experienced by its parents (reviewed in Mousseau & Fox, 1998b), a phenomenon sometimes referred to as ‘soft heredity’ (Bonduriansky, 2012). Such nongenetic inheritance is expected to be favoured when the environment is spatially or temporally heterogeneous across generations, when the environment that the offspring are likely to encounter can be predicted from the parental environment and when the costs of acquiring information about the environment as well as the costs of the adequate phenotypic response are low for both parents and offspring (Uller, 2008). ‘Parental’ or ‘cross-generational’ effects are potentially adaptive: if the parental environment is correlated with that of their offspring, then parents may enhance their fitness by activating developmental programmes that tune their offspring's phenotype accordingly (Mousseau & Dingle, 1991; Rossiter, 1996; Fox et al., 1997; Donohue & Schmitt, 1999). Consequently, not only within-generation phenotypic plasticity, but also parental effects mediated by nongenetic mechanisms of inheritance need to be considered when investigating the dynamics of phenotypic evolution (Mousseau & Fox, 1998b; Bonduriansky et al., 2012).

Ambient temperature is a key environmental factor known to affect the phenotype of ectotherms and is considered an important selective agent (Bale et al., 2002; Clarke, 2003). In several insect species, developmental temperature affects adult life-history traits such as body size, reproductive investment and locomotory abilities (Atkinson, 1994, 1996; Huey et al., 1999; Fox & Czesak, 2000; Gibert et al., 2001; Bloem et al., 2006; Chidawanyika & Terblanche, 2011). Anecdotally, some studies have highlighted effects of temperature experienced by parents, especially by mothers, on offspring phenotype (Mousseau & Dingle, 1991; Zamudio et al., 1995). So far, several competing hypotheses that explain cross-generational phenotypic responses to temperature received experimental support. The ‘adaptive cross-generational’ hypothesis proposes that offspring experiencing the same environment as that of their parents will perform better than offspring exposed to a novel environment (Mousseau & Dingle, 1991; Gilchrist & Huey, 2001). The ‘colder parents are better’ hypothesis claims that parents that developed at low temperatures are larger and thus will produce larger offspring that, because large body size conveys fitness advantages, will perform well as adults in any thermal environment (Gilchrist & Huey, 2001). Reciprocally, the ‘hotter parents are better’ hypothesis predicts that parents that developed at high temperatures are smaller and thus will produce smaller offspring that, because small body size conveys fitness advantages, will perform well as adults in any thermal environment (Zamudio et al., 1995; Crill et al., 1996; Gilchrist & Huey, 2001).

The oriental fruit moth, Grapholita (= Cydia) molesta, is an economically important pest species that attacks fruit trees mainly of the Rosaceae family. Considered to be native to north-western China (Rothschild & Vickers, 1991), G. molesta has spread across a large geographic area over a relatively short timescale (Kirk et al., 2013) and is presently distributed throughout temperate regions of all continents. This multivoltine species infests its primary host, peach, early in the growing season and sometimes shifts to pear and apple after peach harvest (Rice et al., 1972; Rothschild & Vickers, 1991; Natale et al., 2004; Myers et al., 2006). Overwintering occurs as prepupae in sheltered sites, such as under bark scales or in leaf litter at the base of trees, and the overwintering generation pupates in spring (Rothschild & Vickers, 1991). The spatial and temporal environment experienced by individuals is therefore heterogeneous across generations. In addition, previous studies have shown that the oriental fruit moth is sensitive to environmental conditions, especially to temperature. Females prefer oviposition sites close to 30 °C, matching the temperature at which offspring performance is greatest in terms of larval survival and growth (Notter-Hausmann & Dorn, 2010). Moreover, we have recently demonstrated a beneficial acclimation effect on flight performance at low temperature (A. Ferrer, D. Mazzi & S. Dorn, unpublished). We showed that G. molesta females that developed at low temperature had similarly good flight abilities at low and at high ambient temperatures, whereas females that developed at high temperature had poorer flight abilities at low than at high ambient temperatures (A. Ferrer, D. Mazzi & S. Dorn, unpublished). Owing to the above reasons, G. molesta is an ideal organism to study the effects of parental temperature on the life-history and flight performance of offspring.

In the framework of the proposed hypotheses explaining the alternative patterns of response, we investigated the effects of the temperature experienced by the parents and by their offspring during development and adulthood on offspring life-history traits and, in particular, on offspring flight performance. Specifically, an adaptive cross-generational response predicts that offspring perform best when they experience an environment matching the parental one. In contrast, the ‘colder parents are better’ and the ‘hotter parents are better’ hypotheses are independent of a correlation between parental and offspring environment; rather, they predict that offspring of parents that experienced cold and warm, conditions, respectively, are at an advantage in any thermal environment. The incorporation of not necessarily hard-wired responses to prevailing environmental conditions expands the understanding of how phenotypes are determined in a multivoltine invasive species and may allow to identify the factors that contribute to its ongoing range expansion towards higher latitudes.

Materials and methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References
  9. Supporting Information

In this study, we implemented an experimental design to independently manipulate (i) the temperature at which the parental generation was reared, (ii) the temperature at which the offspring generation underwent development and (iii) the temperature at which the offspring generation was exposed during adulthood. Each of the three factors had two levels (low temperature: 20 °C and high temperature: 30 °C). The lower level is the temperature at which a beneficial acclimation effect within one generation has previously been documented (A. Ferrer, D. Mazzi & S. Dorn, unpublished); the higher level is the temperature favoured for oviposition, supporting highest larval survival and fastest growth (Notter-Hausmann & Dorn, 2010). Both temperatures are ecologically relevant with respect to the species distribution throughout temperate regions of the world.

Insect rearing

Stock culture

The moths used in this study were derived from larvae collected in 16 peach orchards in the region Emilia Romagna in northern Italy during May 2010. The colony was refreshed with the addition of 80 wild specimens collected in the same region in May 2011. Moths were reared in the laboratory for about 12 generations before refreshment, and for about seven generations thereafter, on an artificial diet (Najar-Rodriguez et al., 2013) under constant environmental conditions of 25 °C (±0.2 °C), 70% relative humidity (RH) and a photocycle of 16 h of light/8 h of dark (16 L : 8 D). Colony size was kept at a minimum of 100 adults per generation.

Experimental insects

Experimental insects were randomly sampled from the stock culture and reared for two generations, that is ‘parental’ and ‘offspring’ generations. Individuals of the parental generation were held during their entire lifetime at either 20 °C or 30 °C, respectively, termed ‘low’ and ‘high’ ‘parental temperatures’ (TP). Their offspring were split into two groups and reared from egg to adult emergence at either 20 °C or 30 °C, respectively, termed ‘low’ and ‘high’ offspring ‘developmental temperature’ (TD). Upon emergence, these offspring were again split into two groups and exposed to ‘low’ and ‘high’ offspring ‘adult temperature’ (TA) of 20 °C and 30 °C, respectively. Hence, we examined three factors (TP, TD and TA), with two levels each (low and high), resulting in eight temperature treatment groups.

To control for genetic relatedness among individuals allocated to different temperature treatments, we reared 30 replicated families separately. The rearing design for one exemplary replicate is illustrated in Fig. 1. A family was initiated with a random pair of adults of opposite sex from the stock culture. All the eggs laid by the female during the third and fourth days of adult life (i.e. the period of peak oviposition, Hughes & Dorn, 2002) were collected and randomly and evenly split into two groups allocated to either one of two parental temperatures (TP) to undergo development. Upon emergence, we sampled one female for each of the two groups and paired it with an unrelated male reared at the same temperature (TP). Again, the eggs laid by the female during the third and fourth days of adult life were randomly and evenly split into two groups allocated to either one of two developmental temperatures (TD) to undergo development. We randomly selected four female pupae. Upon their emergence, we paired each female with a male from the same family and from the same combination of TP × TD. We allocated two pairs to each of the two adult temperatures (TA).

image

Figure 1. Rearing design, exemplified for one of the 30 replicate families, with TP = parental temperature (temperature at which the parents were reared), TD = offspring developmental temperature (temperature at which the offspring were exposed throughout pre-imaginal development) and TA = offspring adult temperature (temperature at which the offspring were exposed after emergence). The experiment was designed to test the effects of the parental temperature and of the temperature experienced during pre-imaginal development and adulthood on the life-history traits of female offspring.

Download figure to PowerPoint

Apart from temperature, rearing conditions were identical across life stages and generations: 70% RH and 16 L : 8 D photocycle. Larvae were reared on artificial diet in plastic Petri dishes (diameter 14 cm, height 2 cm) in groups of 12. Strips of corrugated cardboard were provided as shelters for the larvae to pupate. Five days after pupation, pupae were removed from their cocoons and sexed (Beeke & De Jong, 1991). Female and male pupae were kept separately in plastic cylinders (diameter 4.5 cm, height 10 cm), sealed with cotton lids and placed in climate chambers (temperature accuracy of ±0.2 °C; Conviron, Controlled Environments System, Winnipeg, MB, Canada) set to the assigned temperature of either 20 °C or 30 °C. All matings occurred in 48-cm3 plastic box (6 × 4 × 2 cm) provided with a wet cotton wick, on whose walls the females deposited eggs.

Life-history traits of female offspring

For measurements of pupal mass and of developmental time (defined as the number of days from egg deposition to adult emergence), we used the four selected females per combination of TP × TD for each family. Just before emergence, pupae were weighed on a Sartorius microbalance (accuracy of ± 0.001 mg; Mettler MT5, Tagelswangen, Switzerland) and isolated in individual 48-cm3 plastic boxes. The boxes were checked daily, and the day of emergence was noted.

For measurements of lifetime egg production and of longevity, we used one of the two females allocated to each temperature treatment TP× TD× TA for each family. The boxes were checked daily, and the newly laid eggs were counted as long as the female was alive.

For measurements of flight parameters, we used the other one of the two females allocated to each temperature treatment TP × TD × TA for each family. Flight tests were performed at the peak flight period, which is 3–4 days post-emergence for mated females (Hughes & Dorn, 2002). The moths were flown at their assigned TA. A computer-linked tethered flight device was used to monitor flight parameters. Moths were anaesthetized on ice, the thoracic scales were removed, exposing the sclerite, and a tethering pin was fixed centrally on the thorax using a piece of sugar-free chewing gum. A static test flight was initiated by the application of a gentle air current. Only moths that were securely attached, and that displayed wing movement unimpeded by the tethering process, were monitored. Moths were thus allowed to fly voluntarily. The tethering pin was pushed perpendicularly through a flight arm. Another pin, which was also inserted vertically through the flight arm, served as central axis. The central axis pin was placed between two magnets, the magnetic field holding it in a vertical position. A reflector on the flight arm and a computer-linked infrared receiver mounted on the mill registered each revolution (illustrated in Wanner et al., 2006).

The flight mill units were placed in a climate chamber maintained at 70% RH. Flight was monitored over a 12-h period, encompassing five phases of photointensity corresponding to the natural light cycle: 1 h of full light (6000 lux), 2 h of simulated dusk conditions, 6 h of darkness, 2 h of simulated dawn conditions and 1 h of full light. The dusk simulation started at 20:00 h. The light intensity linearly decreased from 6000 to 60 lux between 20:00 h and 21:00 h (decrease of approximately 100 lux min−1) and from 60 to 6 lux between 21:00 h and 22:00 h (decrease of approximately 1 lux min−1). At 22:00 h, lights were switched off. The dawn simulation started at 4:00 h. The light intensity linearly increased from 6 to 60 lux between 4:00 h and 5:00 h and from 60 to 6000 lux between 5:00 h and 6:00 h. With this schedule, the middle of the simulated twilight conditions (21:00 h and 5:00 h) in the flight mill climate room corresponded to the time at which lights switched were off or on, respectively, in the rearing climatic chambers. All lighting was provided by high-frequency daylight spectrum fluorescent lamps (Sylvania, Laxline Plus, Cool white de luxe 840, Noida, India). A custom-developed software (M. Gernss, Applied Entomology, ETH Zurich, Switzerland) was used to compute flight parameters: the total duration of flight, the total distance flown, the longest single flight and the total number of single flights performed. The sequence of revolutions was interpreted in terms of single flights and breaks. A break was defined as 3 s or longer without a revolution, and a single flight was defined as the period between two breaks.

To control for potential confounding effects of differences in wing loading (i.e. a measure of the mass carried by a given unit of wing surface), females used in the flight tests were stored in alcohol at 4 °C until dissection under a binocular microscope. Forewings were detached and stuck with transparent tape on a sheet of millimetre paper. A picture of the most intact forewing was taken using a digital camera (DCM 510, ScopeTek, Hangzhou, Zhejiang Province, China) connected to a binocular microscope (Leica MS5, Leica Microsystems AG, Wetzlar, Germany). The forewing surface was estimated by photo analysis using Adobe Photoshop, version 12.0.4. (Adobe Systems Incorporated, San Jose, CA, USA). The forewing loading was defined as the ratio of the surface of one forewing surface to pupal mass (cf. Torriani et al., 2010).

Statistical analyses

We used generalized linear mixed models to test the effects of temperature regimes on the considered life-history traits. Pupal mass and developmental time were modelled assuming a normal distribution, with TP and TD as fixed factors, and family as a random factor. Lifetime egg production and longevity were modelled assuming a Poisson distribution, with TP, TD and TA as fixed factors, family as a random factor, and pupal mass as a covariate. However, for clarity, the raw data (not corrected for variation in pupal mass) are given in the Results section and plotted in Fig. 3. For the analysis of flight performance, we used data collected in the period of 2 h of simulated dusk conditions, which is known to match peak flight activity (Natale et al., 2003). Logarithmically transformed (Zar, 1999) total duration of flight, total distance flown, longest single flight and total number of single flights performed were modelled assuming a normal distribution, with TP, TD and TA as fixed factors, family as a random factor and forewing loading as a covariate. Again, for clarity, unadjusted data are given in Fig. 4.

All interactions among fixed factors were included in the models. Significant differences were investigated through least significant differences (LSD) post hoc tests. All statistical analyses were performed using SPSS version 19 (IBM SPSS Statistics, Chicago, IL, USA); generalized linear mixed models were performed using the ‘mixed models’ procedure.

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References
  9. Supporting Information

In the oriental fruit moth G. molesta, the temperature experienced by the parents had significant effects on the expression of three of four considered life-history traits of their daughters. Pupal mass was significantly affected by both TP and TD, but not by their interaction (Fig. 2; Table 1). Pupae were heavier when TP was low than when TP was high (mean ± SE, 15.193 ± 0.102 mg vs. 14.352 ± 0.133 mg) and were heavier at high TD than at low TD (15.219 ± 0.111 mg vs. 14.325 ± 0.124 mg).

Table 1. Results of generalized linear mixed models for the effects of parental temperature (TP) and offspring developmental temperature (TD) on pupal mass and developmental time in Grapholita molesta females.
 Pupal massDevelopmental time
df F P df F P
  1. Significant P-values (P <0.05) are given in bold type.

T P 1,2368.9 0.003 1,23682.1 <0.001
T D 1,2366.9 0.009 1,2366423.9 <0.001
TP × TD1,2360.00.8741,23657.5 <0.001
image

Figure 2. Pupal mass (a) and developmental time (b) (mean ± SE) for each combination of one parental temperature (high = 30 °C or low = 20 °C) and one developmental temperature (high = 30 °C or low = 20 °C). Standard error bars are smaller than the symbols in (b).

Download figure to PowerPoint

Developmental time was significantly affected by both TP and TD, as well as by their interaction (Fig. 2; Table 1). The effect of TP on developmental time varied depending on TD: at high TD, developmental time was similar regardless of TP (low TP: 22.2 ± 0.1 days; high TP: 22.4 ± 0.1 days); however, at low TD, developmental time was shorter when TP was low than when TP was high (34.1 ± 0.1 days vs. 36.8 ± 0.1 days). Overall, developmental time was shorter at high TD than at low TD (22.3 ± 0.1 days vs. 35.4 ± 0.1 days).

Lifetime egg production was significantly affected by both TP and TA (Fig. 3; Table 2). The main effect of TD was not significant, but the interactions between TP × TD and TD × TA were significant (Table 2). Females whose parents were reared at low temperature (‘low TP females’) laid more eggs (relative to their body mass) than females whose parents were reared at high temperature (‘high TP females’). The significant TP × TD interaction indicated that this effect was stronger for females that developed at low temperature (‘low TD females’) (low TP: 179 ± 4 eggs; high TP: 159 ± 5 eggs) than when females developed at high temperature (‘high TD females’) (low TP: 180 ± 6 eggs; high TP: 168 ± 8 eggs). Overall, females laid more eggs at low TA than at high TA. The significant TD × TA interaction indicated that the difference between the number of eggs laid at low TA and at high TA was larger for high TD females (low TA: 188 ± 7 eggs; high TA: 160 ± 6 eggs) than for low TD females (low TA: 176 ± 4 eggs; high TA: 162 ± 4 eggs).

Table 2. Results of generalized linear mixed models for the effects of parental temperature (TP), offspring developmental temperature (TD), adult offspring temperature (TA) and the covariate pupal mass on lifetime egg production and female longevity in Grapholita molesta.
 Lifetime egg productionLongevity
df F P df F P
  1. Significant P-values (P <0.05) are given in bold type.

T P 1,23128.0 <0.001 1,2310.00.833
T D 1,2311.20.2631,2315.1 0.024
T A 1,231200.6 <0.001 1,231209.4 <0.001
TP × TD1,2315.8 0.017 1,2310.50.487
TP × TA1,2310.10.7961,2310.30.596
TD × TA1,23122.5 <0.001 1,2310.10.747
TP × TD × TA1,2312.20.1421,2310.60.425
Pupal mass1,231459.5 <0.001 1,2314.7 0.031
image

Figure 3. Lifetime egg production (a) and longevity (b) (mean ± SE) for each combination of one parental temperature (high = 30 °C or low = 20 °C), one developmental temperature (high = 30 °C or low = 20 °C) and one adult temperature (high = 30 °C or low = 20 °C).

Download figure to PowerPoint

Longevity was significantly affected by TD and TA, but not by TP. None of the interactions were significant (Table 2). Low TD females had a shorter lifespan (relative to their body mass) than high TD females (12.9 ± 0.4 days vs. 14.2 ± 0.4 days), and lifespan was shorter at high TA than at low TA (10.1 ± 0.2 days vs. 17.0 ± 0.3 days) (Fig. 3).

Three of the four parameters describing offspring flight performance were significantly affected by both TP and TD: total flight duration, total flight distance and longest single flight (Fig. 4; Table 3). In addition to the main effects of TP and TD, the interaction between TD and TA was significant for all three above-mentioned flight parameters (Table 3). Low TP females flew longer and further than high TP females (2152 ± 206 s vs. 1906 ± 207 s, and 1496 ± 150 m vs. 1217 ± 140 m, respectively). Low TP females also performed longest single flights that lasted longer than those performed by high TP females (830 ± 116 s vs. 793 ± 133 s). Low TD females flew over twice as long and over twice as far as high TD females (2773 ± 225 s vs. 1286 ± 160 s, and 1872 ± 161 m vs. 843 ± 110 m, respectively). Low TD females also performed longest single flights that lasted over twice as long as those performed by high TD females (1123 ± 147 s vs. 500 ± 90 s). In addition, when taking wing loading into account, the flight performance of low TD females (in terms of total flight duration, total distance flown and longest single flight) was similar at low and high TA, whereas the flight performance of high TD females dropped at low adult temperature (Fig. 4). The total number of single flights performed was significantly affected by both TD and TA (Fig. 4; Table 3), but not by TP. Low TD females performed more single flights than high TD females (41.4 ± 5.2 flights vs. 23.4 ± 2.7 flights). Overall, females performed more single flights at high TA than at low TA (36.3 ± 4.3 flights vs. 28.4 ± 4.2 flights).

Table 3. Results of generalized linear mixed models for the effects of parental temperature (TP), offspring developmental temperature (TD), offspring adult temperature (TA), and the covariate wing loading on flight performance in Grapholita molesta females.
 Total flight durationTotal flight distanceLongest single flightTotal number of flights
df F P df F P df F P df F P
  1. Significant P-values (P <0.05) are given in bold type.

T P 1,2314.7 0.031 1,2313.9 0.048 1,2314.5 0.035 1,2311.10.299
T D 1,23132.7 <0.001 1,23132.7 <0.001 1,23128.6 <0.001 1,2317.4 0.007
T A 1,2311.30.2551,2312.80.0951,2310.30.5971,2319.6 0.002
TP × TD1,2310.70.4151,2310.00.8441,2310.00.9671,2310.70.387
TP × TA1,2310.20.6461,2310.10.7871,2310.00.9031,2310.40.536
TD × TA1,2318.5 0.004 1,2317.9 0.005 1,2315.7 0.018 1,2313.60.060
TP × TD × TA1,2310.10.7961,2310.20.6891,2310.20.8921,2311.60.211
Wing loading1,2310.80.3791,2311.40.2371,2310.70.4021,2310.80.378
image

Figure 4. Total duration of flight (a), total distance flown (b), longest single flight (c) and total number of flights (d) (mean ± SE) for each combination of one parental temperature (high = 30 °C or low = 20 °C), one developmental temperature (high = 30 °C or low = 20 °C) and one adult temperature (high = 30 °C or low = 20 °C).

Download figure to PowerPoint

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References
  9. Supporting Information

The parental rearing temperature, the temperature experienced during pre-imaginal development and the temperature experienced during adulthood all significantly impacted important life-history traits of G. molesta females. We demonstrated that a low parental temperature affected the body mass of female offspring, their developmental time at low temperature, as well as their egg production and flight performance. In addition to these cross-generational effects, we found a within-generation effect of acclimation to low temperature on flight performance.

The relationship between body size and temperature is not straightforward and still much debated (reviewed in Ashton, 2001), with contrasting patterns across and within taxa (Shelomi, 2012). In the present study, G. molesta parents reared at low temperature produced larger daughters than parents reared at high temperature, suggesting that body mass may increase across generations in cold environments, as known from other lepidopteran species (Nylin & Svärd, 1991). With respect to the within-generation response of body size to temperature, the most common pattern in ectotherms, known as the temperature–size rule (Angilletta & Dunham, 2003; Kingsolver & Huey, 2008), is that development at low temperature results in larger individuals than development at higher temperature (Atkinson, 1994, 1996; Fischer & Karl, 2010). Consistent with an earlier study (A. Ferrer, D. Mazzi & S. Dorn, unpublished), we found that within a single generation, the temperature–size rule does not apply to G. molesta. On the contrary, development at low temperature resulted in significantly lighter pupae than development at higher temperature. This phenomenon is known from a few other species within the Orthoptera and the Lepidoptera (Walters & Hassall, 2006; Diamond & Kingsolver, 2010), such as the geometrid moth Cabera exanthemata (Kivelä et al., 2012). However, the physiological mechanism underlying the reversal of the temperature–size rule is unknown, and therefore, the adaptive significance of temperature-dependent body size variation remains elusive (Davidowitz & Nijhout, 2004; Walters & Hassall, 2006; Diamond & Kingsolver, 2010). In G. molesta, the phenotypic responses of body mass to parental and developmental temperature were antagonistic and thus preclude generic predictions about the evolution of body mass as a by-product of temperature variation across space (e.g. among geographic populations) (Mousseau, 1997; Blanckenhorn & Demont, 2004) or time (e.g. among successive generations over a season).

Parental temperature also had a significant effect on the developmental time of G. molesta females: if parents were reared at low temperature, their daughters developed faster at low temperature than if parents were reared at high temperature. These faster-developing females also reached a higher pupal mass, demonstrating that a low parental temperature enhanced overall larval growth performance in a cold environment. Parental thermal environment has been shown to affect offspring growth performance in at least two other insect species, including a butterfly (Gilchrist & Huey, 2001; Steigenga & Fischer, 2007). However, in those former studies, the offspring of parents that experienced a cold environment had poorer growth performance than the offspring of parents that experienced a warm environment (Gilchrist & Huey, 2001; Steigenga & Fischer, 2007), indicating that the adverse effects of low ambient temperature on physiological performance carried over across generations. To the best of our knowledge, our study demonstrates for the first time a beneficial cross-generational acclimation to low temperature in a nonmodel, economically relevant insect species (see e.g. Rako & Hoffmann, 2006 for evidence from Drosophila melanogaster).

A low parental temperature also had a beneficial effect on fecundity and flight performance. Females whose parents were reared at low temperature had a significantly higher lifetime egg production than females whose parents were reared at high temperature, in spite of similar longevity. Contrary to the effect of developmental temperature, the effect of parental temperature on offspring egg production has only rarely been investigated in insects and has usually been found to be either negligible (Steigenga & Fischer, 2007) or restricted to slight paternal effects (Huey et al., 1995). In addition, flight mill assays showed that, irrespective of developmental temperature, females whose parents were reared at low temperature flew longer and further than females whose parents were reared at high temperature.

To our knowledge, this study is the first report of congruent effects of parental temperature on several fitness-relevant life-history traits in an insect species. In addition, taken together, these results on the cross-generational effect of temperature on life-history traits provide some support for the hypothesis that ‘colder parents are better’ (Gilchrist & Huey, 2001). Parents that developed in a cool environment produced larger and more fecund daughters with enhanced flight performance when compared with the daughters of parents developed in a warmer environment. However, the ‘colder parents are better’ hypothesis relies on the fitness benefits accrued from larger body size; in other words, the improved performance is assumed to be a direct consequence of the larger body size typically resulting from development at lower temperatures (Atkinson, 1994, 1995). This causality is violated in G. molesta, as, over one generation, females that developed at low temperature did indeed perform better in a cold environment than females developed at higher temperature, but were smaller.

Compared with conventional modes of inheritance, the potential for short-term acclimation mediated by the parental experience provides several assets (reviewed in Jablonka & Raz, 2009), which may contribute to the evolutionary success and ecological spread of the oriental fruit moth and of other taxa capable of cross-generational plasticity. For example, cross-generational plasticity generates adaptive, heritable variation on demand, as opposed to stochastic genetic variation; also, given environmental conditions can trigger the same adaptive phenotype in many individuals simultaneously, in contrast to the sporadic occurrence of novel phenotypes via mutations. Plasticity will also allow improved adaptive potential of populations depleted of genetic diversity by demographic declines, as typically occur during founding events. Indeed, short-term phenotypic plasticity is a key factor promoting invasiveness (reviewed in Chown & Terblanche, 2007; Davidson et al., 2011). Cross-generational plasticity may promote the establishment of viable populations in variable or stressful but predictable habitats, because the offspring are equipped to face the environment encountered by their parents, and hence allowed to skip the time lag of mounting the appropriate response. An increased environmental tolerance and an increased plasticity of environmental tolerance are known trumps for the survival, establishment and spread into a novel environment (Chown et al., 2007; Slabber et al., 2007). For example, the plasticity in acute thermal tolerance of the Mediterranean fruit fly, Ceratitis capitata, has been identified as a mechanism facilitating the survival upon introduction to novel thermal habitats (Nyamukondiwa et al., 2010). In general, the recognition that phenotypic adaptation may take a sort of shortcut around necessarily slow changes in allele frequency extends the scope of evolutionary thinking beyond the concept of natural selection and urges for an improved incorporation of development into the study of heredity in times of rapid environmental change.

Acknowledgments

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References
  9. Supporting Information

We thank Anna Drewek from the Department of Mathematics of ETH Zurich and Vincent Payet from ISARA Lyon for support with statistical analyses and Claudio Sedivy and two anonymous reviewers for valuable comments on the manuscript.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References
  9. Supporting Information

Supporting Information

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References
  9. Supporting Information
FilenameFormatSizeDescription
jeb12218-sup-0001-FigureS1.epsimage/eps1486KFigure S1 Flight performance traits for each combination of parental, developmental and adult temperature as given in Fig. 4, but plotted using the residuals from a regression on the covariate wing loading.
jeb12218-sup-0002-FigureS2.epsimage/eps1237KFigure S2 Flight performance traits for each combination of parental, developmental and adult temperature as given in Fig. 4, but plotted using log-transformed data.

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.