Multiple Effects of Temperature, Photoperiod and Food Quality on The

Ecological Entomology (2013), 38, 201–208 DOI: 10.1111/een.12005

Multiple effects of temperature, photoperiod and food quality on the performance of a pine sawﬂy IDA KOLLBERG,1 HELENA BYLUND,1 AXEL SCHMIDT,2 JONATHAN GERSHENZON2 andCHRISTER BJORKMAN¨ 1 1Department of Ecology, Swedish University of Agricultural Sciences, Uppsala, Sweden and 2Department of Biochemistry, Max-Planck-Institute for Chemical Ecology, Jena, Germany

Abstract. 1. Many entomologists strive to understand what consequences climate change will have on insect performance. Such understanding is important, not least, when trying to predict the future impact of pest insects. 2. In this study, it was reported how the multiple effects of temperature, photoperiod, and food quality affected the survival, development, and the final weight of the European pine sawfly (Neodiprion sertifer Geoffr.), an herbivorous outbreak species in boreal pine (Pinus spp.) forests. ◦ 3. Sawfly larvae were reared in two different temperatures (15 and 20 C) and under two different light regimes (20 and 18 h light). The larvae were fed pine needles either low or high in diterpene content. ◦ 4. A 5 C higher temperature did not affect the survival of the larvae, but reduced the development time by 37–41%. The final weight was reduced by 22% in the warmer temperature, but only in combination with a short day length. A high content of diterpenes in the needles reduced the susceptibility to the virus by 31%, but did not otherwise affect the performance of the larvae. 5. This study shows that the larval development could be shortened in a warmer climate and thereby decreasing the risk of predation. This per se may increase the risk for insect outbreaks, but the interactive effects of warmer temperatures with other abiotic and biotic factors such as day length and food quality (indicated in this study), and potential better performance of natural enemies and pathogens, illustrate the possibility for complex outcomes in a climate change perspective.

Key words. Development time, Diprionidae, diterpenes, insect herbivore, Neodiprion sertifer, Pinus sylvestris, plant–insect interactions, survival, temperature-size rule, virus.

Introduction 1994; Bale et al., 2002). As a consequence of recent warming we see species expanding their distributional ranges towards As an effect of a changing climate, the global mean tem- the poles (Parmesan, 1996) and species emerging earlier in perature is steadily increasing (IPCC, 2007). The largest the spring (Parmesan, 2007). This means that insects will be temperature increase is estimated to be towards the poles. In exposed to a different photoperiod in addition to enhanced Fennoscandia, the predicted increase in temperature is between ◦ temperatures. Abiotic changes will affect insect performance 2and6 C within the next 100 years (Lind & Kjellstrom,¨ not only directly, but also indirectly as an enhanced temper- 2008). For ectothermic species such as insects, the surrounding ature will affect food quality and natural enemies (Berggren temperature is highly inﬂuential because they have limited pos- et al., 2009). It is commonly anticipated that a warmer climate sibilities to thermoregulate. Temperature directly affects insect survival, development time, and the ﬁnal weight (Atkinson, may lead to more frequent or more severe outbreaks of insects but to fully understand how insect survival and performance Correspondence: Ida Kollberg, Department of Ecology, Swedish will be affected by global warming it is important that we University of Agricultural Sciences, Box 7044, S-750 07 Uppsala, understand how these multiple factors, both abiotic and biotic, Sweden. E-mail: [email protected] interact (Bale et al., 2002; Bjorkman¨ et al., 2011).

For insects, the direct effects of increased temperatures are Herbivore performance, in particular survival, is strongly often beneficial. Increased temperatures are usually related to influenced by natural enemies, including diseases. Because increased fitness, higher survival rates, and shorter develop- natural enemies also are likely to respond to changing weather ment times (Bale et al., 2002). However, in many previous conditions, increasing temperatures may indirectly affect an studies it has been shown that increased temperatures result herbivore’s survival and performance. When an outbreak of in a reduced final weight (the temperature-size rule; Atkinson, N. sertifer ends, and the density of the population abruptly 1994), possibly as an effect of temperature having different declines, virus infection is commonly a major larval mortality effects on development and growth (Van der Have & de Jong, factor (Olofsson, 1987). Hence, in a climate change perspective 1996; Karl & Fischer, 2008). If development is relatively faster the temperature effect on the interaction of virus, day length, than growth in higher temperatures then individuals will stop and food quality could be highly important. growing earlier and get a lower final size. In this study, we evaluate the multiple effects of three Day length also influences the performance of plant-feeding important abiotic and biotic factors on the performance of insects (Beck, 1980). While many environmental factors are N. sertifer. We report how a difference in temperature, highly variable, day length is a reliable cue regulating life photoperiod, and plant secondary compounds (diterpenes) cycles in many species. Photoperiod is mainly involved in the affected the survival, development time, and final weight (as process when insects are about to enter diapause but changes in a proxy for fecundity) of the larvae. The overall aim was to day length could potentially also be an important indicator for assess which consequences a changing climate will have on the larvae to determine when to stop feeding in preparation plant–insect interactions to better understand the future impact for pupation. Often the response acts in combination with of pest insects. We predicted that the cost in terms of decreased temperature which means that photoperiod can modify thermal performance owing to the sequestration of the diterpenes would reaction norms by accelerating or slowing down insect growth be less in the higher temperature treatment. Moreover we and development (Gotthard et al., 2000; Blanckenhorn & wanted to evaluate the role of day length to assess the effect Demont, 2004; Kutcherov et al., 2011). of increased temperature within N. sertifer’s distributional Food quality is, in addition to temperature and photoperiod, range in Sweden from the southern part to the northern another important factor determining the performance of border. We hypothesised that, in accordance with the general herbivorous insects (Schoonhoven et al., 2005). Still we understanding, the larvae would develop faster but have a know little about how abiotic factors affect physiological and lower final weight in the higher temperature. However, the behavioural interactions between insects and plants. Expected response to temperature would be influenced by photoperiod. changes in climate will, through enhanced temperatures, The season is shorter for individuals living at the northern CO2, and drought, have diverse impacts on plant quality border of its range and therefore the larvae would have a faster such as nitrogen content and secondary chemicals (White, development and consequently a lower weight in the longer day 1984; Sallas et al., 2003) which will indirectly affect insect length because they would act as they were restricted by time. performance. Effects on plant secondary compounds appear The study was not designed to study the effect of the virus. to be species-specific (Bidart-Bouzat & Imeh-Nathaniel, 2008) However, as it is difficult to avoid virus infections in laboratory and depend on resource availability (Herms & Mattson, 1992). rearing, we got the opportunity to study another important As temperatures increase, plant resistance might become less interaction, namely how the virus was affected by temperature important if the cost of sequestration for insects goes down and food quality, and hence affected sawfly survival. owing to an increased rate of metabolic processes. The European pine sawfly, Neodiprion sertifer Geoffr. Material and methods (Hymenoptera: Diprionidae), is a typical outbreak species and considered as a major pest in pine [Pinus spp. (Pinaceae)] Study species forests and the target species in this study. An outbreak may cover several hundred thousands of hectares (Lyytikainen-¨ The European pine sawfly (N. sertifer) is a gregarious out- Saarenmaa & Tomppo, 2002) but it rarely kills the trees. break species that as larvae feed on pine (Pinus spp.) needles. Still, the sawfly causes large economical losses for the forest Its native range is in Eurasia but it is also common in North industry (Lyytikainen-Saarenmaa¨ & Tomppo, 2002) as it America, where it has been introduced. In Sweden, larvae of substantially reduces the growth of the pines (Austara˚ et al., N. sertifer feed on P. sylvestris and P. contorta.Thelarvae 1987). The larvae feed gregariously on pine needles and hatch in early summer (May to June) and feed between 4 and diterpene content affects N. sertifer larval performance and 7 weeks depending on climatic conditions. Most commonly it possibly the population dynamics of this species (Larsson has a univoltine life cycle, but can be semi-voltine at colder et al., 2000). In N. sertifer, larvae feeding on needles with localities. Males have four larval instars and females five. high diterpene content show increased mortality and reduced growth compared with larvae feeding on low diterpene needles Experimental design (Larsson et al., 1986). However, as the larvae sequester the diterpenes, store them, and use it as an antipredator defence by Abiotic settings regurgitating resin droplets when attacked by natural enemies, feeding on high diterpene needles may increase survival if Neodiprion sertifer larvae were reared in two constant ◦ predators are present (Bjorkman¨ & Larsson, 1991). temperatures, 15 and 20 C, and in two constant light

© 2013 The Royal Entomological Society, Ecological Entomology, 38, 201–208 Multiple effects on sawfly performance 203 regimes, day length LD 18:6 h and LD 20:4 h. The lower gas chromatography–mass spectrometry (GC-MS) analysis or ◦ ◦ temperature treatment (15 C) was chosen to represent the stored in −20 C until analysis. mean temperature during the period the larvae are feeding Analysis of diterpene constituents was performed on a naturally at present conditions. The enhanced temperature of 5 Hewlett-Packard 6890 GC-MSD system fitted with an HP- ◦ C is within the span of the predicted increase in temperature 5 column (0.25 mm × 0.25 μm × 30 m; Hewlett-Packard, over the next 100 years. The different day length treatments Agilent Technologies, Santa Clara, California). Injections were were chosen to reflect the light situation in the southern and made with 1 μl of the ethereal extract. The GC-MS was set at ◦ the northern part of Sweden in the beginning of the summer, an injector temperature of 270 C. The instrument programme ◦ when the larvae are active. At this time of the year the days started with an initial temperature of 150 C for 3 min and ◦ ◦ are longer in the northern parts. Owing to limited facilities, increased then with a rate of 3.5 Cmin−1 to 280 Cand ◦ the treatments were combined as follow: 15 C and LD 20:4 was held for 4 min. Helium was used as the carrier gas at ◦ ◦ h, 20 C and LD 20:4 h and 20 C and LD 18:6 h. By a constant flow of 1 ml min−1. Identification of diterpenes ◦ comparing larval performances in 15 and 20 C at a day length was based on comparison of retention times and mass spectra of 20:4 h we were able to study the effect of temperature in with authentic standards or with mass spectra in the Wiley light conditions resembling the northern distribution range. By 275.L or National Institute of Standards and Technology comparing larval performance at a day length of LD 18:6 h and 98.1 MS libraries. Diterpene contents in the samples were ◦ LD20:4hin20 C, we could study the effect of day length quantified relative to the internal standard and their fresh in a future temperature scenario. Logistic limitations forced us weights. Five diterpenes were identified in the pine needles to omit one of four possible treatments and the combination (names and mean relative concentrations for the high and low ◦ 15 C and LD 18:6 h was chosen not to be included. The clones, respectively): pinifolic acid (high: 41.2%, low: 38.4%), reason was that the study focuses on future scenarios and this neoabietic acid (high: 29.4%, low: 23.3%), levopimaric acid combination correspond to the ambient mean temperature and (high: 11.8%, low: 11.5%), abietic acid (high: 3.2%, low: day length in the area where egg batches were collected. 2.3%), and dehydroabietic acid (high: 2.2%, low: 1.5%). There was also one diterpene that we were not able to identify, and which differed in the relative concentration more than the other Chemical analysis and characteristics of the needles compounds, for the clone high in diterpenes: 12.1% and for the clone low in diterpenes: 23.0%. The total diterpene content in Prior to rearing, pine needles were analysed for terpene the high diterpene clone was 1.8 times higher than in the low content with special emphasis on diterpenes as these are known diterpene clone (ratio based on mean ± SD: 1.6–1.9). to influence the performance of sawfly larvae. The nitrogen and The needles for the nitrogen and carbon analyses were dried ◦ carbon content in the needles were also analysed. In September in 50 C for 48 h and then ground. After that they were dried ◦ 2009, two different genotypes of Scots pine (P. sylvestris) again in 70 C for 24 h and kept in a desiccator until analysis. ◦ ◦ were selected in a seed orchard (Hade, 60 19 N, 17 30 E). Total C and total N in the needles were determined with a The previous year’s needles were collected from branches Carlo-Erba NA 1500 Elemental analyzer. In the low diterpene at two adjacent whorls facing south-west. Two trees from clone total C and N (mean ± SD) were 493.3 ± 5.9 and 13.6 ± each of the genotype were sampled, in total four samples per 1.2 mg g–1 respectively and in the high diterpene clone 500.3 genotype. Even although concentrations of diterpenes can vary ± 3.3 and 14.3 ± 0.6 mg g–1. within genotypes and between years owing to environmental conditions, the relative difference between genotypes seems to be stable (Gref & Tenow, 1987). Further, we consider Biotic settings variations, both quantitative and qualitative, within genotypes to be small as the trees grow under similar environmental Sawfly eggs were collected in May 2010 from an outbreak ◦ ◦ conditions in a homogenous seed orchard. Immediately after area in mid-Sweden (Ramnas,¨ 59 46N, 16 16E). The mean removing the needles from the shoot they were frozen on dry temperature in June for the period 1961–1990 in that area was ◦ ◦ ice and stored in the laboratory at −80 C. Half of the material 14.6 C and the day length during summer is approximately was used for diterpene analysis and half for the N- and C- 18 h. The eggs were stored at room temperature until hatching. analyses. To reduce confounding effects owing to genetic dissimilarities, For the diterpene analysis, the needles were ground in each egg batch was split into three groups and distributed liquid nitrogen and extracted in 1 ml of tert-butyl methyl to the treatments. Each replicate consisted of 10–15 newly ether (TBME) including the internal standard tetrahydroabietic hatched larvae siblings that were transferred to a pine shoot acid (0.045 mg ml−1) and shaken constantly for 14 h. (P. sylvestris) into a 14-cm-diameter Petri dish. Feeding larvae The ethereal extracts were washed with 0.3 ml of 0.1 m with detached shoots was justified as we know that there are (NH4)2CO3 (pH 8.0) and transferred to a fresh vial. This no induced responses in Scots pine to sawfly feeding (Niemela¨ sample was filtered through a Pasteur pipette column filled et al., 1991) and that the needle quality of detached shoots with MgSO4 andwashedwith1mlofTBME.Asampleof remains stable for a week (Larsson et al., 1986). Twenty- 0.5 ml of the combined ethereal extracts was methylated by four replicates per treatment were provided with needles with adding 50 μl of 0.2 m N-trimethylsulfoniumhydroxid at room high content of diterpenes and 24 replicates per treatment temperature for 2 h. Derivatised samples were subjected to with needles low in diterpene content. Pine shoots were

© 2013 The Royal Entomological Society, Ecological Entomology, 38, 201–208 204 Ida Kollberg et al. collected from the same two genotypes of Scots pine that cocoons in different treatments, including those that were virus were chemically analysed the year before the experiment. killed in the larval stage, was analysed using a generalised Seasonality affects the concentration of diterpenes in seedlings linear model with quasi binomial error distribution. In the (Nerg et al., 1994) and in young needles (Gref, 1982), but other analysis, the larvae that were obviously killed by the has little effect on mature needles older than 1 year (Gref, virus were removed from the data and the survival over time 1982). Even although the exact concentration of the needles from newly hatched larvae to the fourth instar was examined. used for larval food may differ from the needles that were In the later analysis replicate was treated as a random effect as analysed, the relative difference between the genotype high the measurements were repeated in time on the same subject and low in diterpene content is likely to be constant. What and therefore not independent. The replicates were only ran- can affect diterpene concentrations, however, is if needles are domised for its slope (over time) as they all had approximately growing in the sun or in the shade (Gref & Tenow, 1987). the same starting value. A binomial error distribution was used To reduce the risk of accidentally collecting shoots with non- as survival was measured as proportions (the glmer function representative needle quality, we collected shoots from two in the lme4 package, Bates et al. 2011). In this model larval south-west facing branches on two trees per genotype. stage was also included in the model using the proportion of The larvae were counted every third to fourth day, and were the larval group that had passed a certain larval stage as a at the same time provided with fresh pine shoots, until all variable. Model selection was done by backward elimination larvae had spun cocoons. If signs appeared that the larvae were of non-significant factors from the full model including all infected by the virus or fungi, the Petri dish was excluded interactions. Results from the different analyses are presented from the climate chamber; however, larvae infected by the given the final model with only significant effects. virus during the later larval stages are still able to spin a cocoon and the disease is hard to discover. Typical signs of virus disease are that the larvae become less active and get a Results reddish colour. To confirm that the virus was a major cause of death a random sample of the dead larvae was dissected. Survival Virus-infected larvae have virus polyeders in their gut cells which can be detected under a microscope (Olofsson, 1988). Larval survival (from newly hatched larvae to fourth instar) It is difficult to completely rule out that some larvae died as a was not affected by temperature or by day length. However, in combination with food quality, temperature affected the result of other causes or combination of causes, but our long- ◦ term experience of work with this insect makes us confident survival (Table 1). Larvae raised in 20 C on the low diterpene that virus infection was the major cause of death. The cocoons clone had a lower survival than in the other treatments but the were left for 2 weeks in their experimental treatments to harden difference was only obvious in the later larval stages indicated and then the cocoons were weighed and the sex ratio for each by the interaction between pine clone and larval stage (Table 1, larval group (replicate) was calculated. It is first when the Fig. 1). Needle quality seemed to influence the susceptibility cocoons have been formed that the sex of the sawflies can of larvae to virus reflected in a lower proportion, 44 ± 5% visually be determined by comparing the size of the cocoons. (mean ± SE), surviving to spinning a cocoon after feeding on Percentage survival, development time (number of days until the low diterpene clone compared with 62 ± 5% (mean ± SE) half of the larval group had spun cocoons), the relative growth on the high diterpene clone (t =−2.6, d.f. = 95, P = 0.01). rate (ln cocoon weight/development days), and mean cocoon weight were calculated for each replicate. Development time

Data analysis The development time was affected by both temperature and ◦ day length but not by food quality. In 15 C the average number The statistical analyses were performed using R (version of days from newly hatched larva to pupa was 12 days longer ◦ 2.14.2, R Development Core Team, 2011). The survival, than in 20 C (mean ± SE; 32 ± 0.3 days and 20 ± 0.3 development time (number of days from newly hatched larva days respectively in day length L20:D4; t =−29.2, d.f. = 72, to cocoon), relative growth rate, and final cocoon weight P < 0.001), and 1 day longer in 20 h day light compared with ◦ were analysed separately by fitting linear models. As the 18 h (20 ± 0.3 and 19 ± 0.2 days respectively in 20 C; t = 3.4, experimental design was not fully factorial, the effects of d.f. = 72, P = 0.001). temperature and day length were tested separately for the life history traits. In addition, food quality was used as a main explanatory variable in all the models. In the analysis of the Final weight development time, the sex ratio was used as a covariable to take into account that females have one additional larval stage The final cocoon weight was also affected by temperature and hence a longer development. When analysing the cocoon and day length (Fig. 2) but not by food quality. The female ◦ ◦ weights, separate analyses were made for females and males weight was higher in 15 Cthanin20 C(t =−2.5, because of the bimodal distribution of the weight variable. d.f. = 70, P = 0.02). The female weight was also higher in For the survival, two separate analyses were performed. In the longer day length in comparison with the shorter (t = 2.6, the first analysis the proportion of larvae that survived to spin d.f. = 65, P = 0.01). Male weight followed the same pattern

Table 1. Estimates, SE, z-, and P-values from a generalised linear mixed model over the survival of Neodiprion sertifer from newly 0.08 * hatched larvae to the fourth instar. * Fixed effects Estimate SE z-value p-value

(Intercept) 2.2 0.2 11.4 <0.001 0.06 ◦ Temp 20 C −0.5 0.2 −2.1 0.03 Clone low 0.4 0.3 1.4 0.18 Larval stage 0.6 0.1 5.5 <0.001 Time −0.5 0.1 −8.0 <0.001 0.04 ◦ Temp 20 C : clone low 0.8 0.3 2.4 0.01 Clone low : larval stage −0.4 0.1 −2.9 0.004

◦

Larvae feeding in 15 C and on needles with high content of diterpenes Female pupal weight (g) 0.02 are set as a reference. Number of replicates was 96.

100 0.00 15°C LD 20:4h 20°C LD 20:4h 20°C LD 18:6h 90 Treatment

Fig. 2. Female pupal weight (mean ± SE) of Neodiprion sertifer ◦ 80 raised under different temperature (15 and 20 C) and day length (20 and 18 h light) conditions in climate chambers. Both temperature and day length had an effect on female weight. Brackets indicate pairwise comparisons performed and stars indicate signiﬁcant differences 70 Survival (%) Survival (P < 0.05).

20°C High diterpene 60 15°C High diterpene the reduced virus susceptibility of sawfly larvae feeding on 20°C Low diterpene pines rich in diterpenes compared with larvae feeding on 15°C Low diterpene low diterpene pines. The larvae reduced their development 50 times as expected in the higher temperature but we did not 1.0 1.5 2.0 2.5 3.0 3.5 find any interactive effect of food quality and temperature, as Larval stage other previous studies have shown (Stamp & Bowers, 1994; Stamp et al., 1994; Yang et al., 1996). However, an increased Fig. 1. Survival (in %) of Neodiprion sertifer larvae as a function of ◦ variation in weight (SDweight in 15 C = 8.2 mg vs. SDweight developmental stages 1–4. The experiment started with 10–15 larvae ◦ = per group. The larvae were reared in two different temperatures (15 and in 20 C 13.0 mg, Fig. 3) among individuals raised in the ◦ 20 C) and under two different light regimes (20 and 18 h light). They warmer temperature indicates large plasticity. were fed with pine needles either low or high in diterpene content. This study shows that for the temperatures chosen, develop- ◦ Mortality was highest for larvae raised in 20 C on pine needles with ment time was the performance trait that was most affected by a low diterpene content. the temperature treatment. The shorter development time at a higher temperature is in accordance with what has been shown =− = for a range of insects (Atkinson, 1994). A possible reason for as the females regarding the temperatures (t 2.6, d.f. 49, ◦ P = 0.01), but day length did not have any significant effect the longer duration of the larval period in 15 C is because of a (t = 1.6, d.f. = 50, P = 0.125). The weight gained per day slower molting process between instars (Matsuki et al., 1994). ◦ (the relative growth rate) was faster in 20 C compared Contradictory to others, we did not find that poor food qual- ◦ with 15 C (females: t = 25.0, d.f. = 70, P < 0.001, males: ity (in our study in terms of high diterpene content) and low t = 18.8, d.f. = 49, P < 0.001), but was not affected by day temperature interacted by increasing the molting time and con- length (females: t =−1.1, d.f. = 65, P = 0.29, males: t =−0.8, sequently also the development time (Stamp & Bowers, 1994; d.f. = 50, P = 0.41). Stamp et al., 1994; Yang et al., 1996). In a climate change perspective, with warmer temperatures, the larval development would be less restrained by the duration of moltings. Reduc- Discussion ing the time spent in vulnerable stages such as molting, could have a positive effect on the survival by shortening the expo- The main goal of this study was to understand how the sure time to natural enemies when the larva is more or less combined effects of temperature, food quality, and photoperiod immobilised and not capable of defensive behaviour such as affect the performance of the sawfly larvae in order to better regurgitation. predict possible consequences of climate change for this It is clear from this study that diterpene content affected outbreak species. The most novel result in our study was virus susceptibility of the larvae. High diterpene content

0.10 prolonged development for feeding on needles with a higher content of diterpenes could be negligible. Female weight is commonly positively associated with fecundity (Stearns, 1992; Roff, 2002). For a species that cannot 0.08 have more than one generation per year, such as N. sertifer, there should be a selection for larvae to feed to become as large as possible (up to some physiological limit), thereby maximising their egg loads. The larvae in this study had a 0.06 ◦ higher relative growth rate in 20 C but as a result of a considerable shorter development time they pupated at a lower weight, following the temperature-size rule. Even although Female weight (g) pupal 0.04 maximising weight should be the best option in theory, it is not always the best option in nature. What would be the beneﬁts

20°C LD 18:6h of having many eggs, if you are dead? That is probably why ◦ 20°C LD 20:4h we see such a large variation in weight in 20 C (Fig. 3). It 15°C LD 20:4h 0.02 seems as if the larvae experienced a multiple choice situation in 15 20 25 30 35 40 which individuals responded differently. Some larvae seemed Development time (days) to choose to pupate before their potential maximum weight Fig. 3. Female pupal weights of Neodiprion sertifer in relation to was reached, perhaps to avoid larval natural enemies. Such larval development time. The larvae were raised under different a variation in response within a larval group of siblings to a ◦ temperatures (15 and 20 C) and different day length conditions novel situation could be adaptive as some individuals would (20 and 18 h light). Symbols represent, for each larval group, the do well in situations with low predation pressure and other interpolated number of days when half of the larvae had spun cocoons individuals do better in high predation conditions. Photoperiod (x-axis) and the corresponding mean pupal weight of the females (y- also seemed to influence the final weight. In this study we axis). used two light regimes to mimic the day length differences in northern and southern Sweden in early summer. Larvae in ◦ seemed to increase larval resistance against virus. To our 20 C had the same growth rate regardless of day length, but knowledge there are no previous reports of such an effect the larvae in the shorter day length pupated on average earlier on N. sertifer although it is known that the chemical and at a lower final weight compared with the larvae in the composition of the food can counteract the dissolution of longer day length. Most probably, this is explained by the day the virus in the midgut in other systems. Tannins are, for length to which the larvae (or eggs) were originally adapted example, known to inhibit virus infection in gypsy moth to. The eggs used in the study were collected in an area where larvae (Keating et al., 1990). There is also evidence that good the day length is about LD 18:6 h during the larval stage. It food quality makes lepidopteran larvae more resistant against is possible that the larvae slowed down development when the virus infection (Sosa-Gomez´ et al., 1991; Forschler et al., light duration was extended compared with what they were 1992). It would be extremely interesting to further study this adapted to. This is important when evaluating how climate plant–pathogen interaction to understand its relevance under change can affect the spread of forest pest species. If higher natural field conditions. In accordance with other previous temperatures reduce the final female weight and hence the studies (Tvermyr, 1969; Frid & Myers, 2002), temperature fecundity, a longer photoperiod can modify the response by not in itself did not affect the susceptibility to the virus but the reducing the weight as much. This means that the northward infected larvae died earlier in the warmer temperature (on spread of phytophagous invertebrate species can be enhanced ◦ average starting from developmental stage 3.3–3.5 in 20 C if females get larger and more fecund when they get exposed ◦ whereas in 15 C the larvae were in developmental stage 3.8). to a longer photoperiod than they are adapted to. Mortality owing to a virus infection is one major contribut- When looking at the response of insect herbivores to ing factor to end outbreaks of N. sertifer andinthatrespect enhanced temperatures, not only the plant interaction will have it is important to understand how global warming can affect an impact on the performance, but also the interaction with virus spread in sawfly populations. If increasing temperatures natural enemies. The response that we have seen from the enhance virus spread by infecting the larvae earlier, a virus lab experiment could be altered if we also include predators has the potential to have a stronger impact on the population and parasitoids as a disturbance for the larvae. It might for dynamics of N. sertifer in the future. No other performance example be possible that larvae will stop feeding earlier traits of the larvae were affected by the diterpene content of if they are exposed to a high predation pressure (Lima & the needles. Larsson et al. (1986) showed a strong response Dill, 1990). It would be a better strategy to mature at a in larval performance when they used pine genotypes with a smaller size and produce fewer eggs, than to be killed and larger difference in diterpene content than used here. However, miss the opportunity to produce any offspring at all. The as Bjorkman¨ and Gref (1993) indicates, variation in natural results from this study illustrate complex interactions between pine stands is not that large and the difference used in this temperature, food quality, and photoperiod on sawfly larval study is probably more representative for natural conditions. performance, but to be able to make accurate predictions If that is the case, the cost in terms of increased mortality and about how temperature affects a certain species we need to

© 2013 The Royal Entomological Society, Ecological Entomology, 38, 201–208 Multiple effects on sawfly performance 207 study the whole system, including tritrophic interactions, here parasites. PhD thesis, The Swedish University of Agricultural exemplified by the sawfly larval performance in relation to Sciences, Uppsala, Sweden. virus and food quality. Gref, R. & Tenow, O. (1987) Resin acid variation in sun and shade needles of Scots pine (Pinus sylvestris L.). Canadian Journal of Forest Research, 17, 346–349. Acknowledgement Herms, D.A. & Mattson, W.J. (1992) The dilemma of plants: to grow or defend. The Quarterly Review of Biology, 67, 283–335. We are grateful to Karin Eklund and Lisa Fors for their IPCC (2007) Contribution of Working Group I to the Fourth Assessment assistance in the lab. We also thank Marion Staeger at the Max Report of the Intergovernmental Panel on Climate Change. Climate Change 2007: The Physical Science Basis. Cambridge University Planck Institute for Chemical Ecology for technical support Press, Cambridge, U.K., New York, New York. during the terpene analysis. Richard Hopkins kindly provided Karl, I. & Fischer, K. (2008) Why get big in the cold? Towards a us with valuable feedback on the manuscript. The study was solution to a life-history puzzle. Oecologia, 155, 215–225. funded by FORMAS, the Mistra-program ‘Future Forests’ and Keating, S.T., Hunter, M.D. & Schultz, J.C. (1990) Leaf phenolic the EU-project BACCARA. inhibition of gypsy moth nuclear polyhedrosis virus – role of polyhedral inclusion body aggregation. Journal of Chemical Ecology, 16, 1445–1457. References Kutcherov, D.A., Lopatina, E.B. & Kipyatkov, V.E. (2011) Pho- Atkinson, D. (1994) Temperature and organism size – a biological law toperiod modifies thermal reaction norms for growth and develop- for ectotherms? Advances in Ecological Research, 25, 1–58. ment in the red poplar leaf beetle Chrysomela populi (Coleoptera: Austara,˚ Ø., Orlund, A., Svendsrud, A. & Veidahl, A. (1987) Growth Chrysomelidae). Journal of Insect Physiology, 57, 892–898. loss and economic consequences following two years defoliation of Larsson, S., Bjorkman,¨ C. & Gref, R. (1986) Responses of Neodiprion Pinus sylvestris by the sawfly Neodiprion sertifer in West-Norway. sertifer (Hym., Diprionidae) larvae to variation in needle resin acid Scandinavian Journal of Forest Research, 2, 111–119. concentration in Scots pine. Oecologia, 70, 77–84. Bale, J.S., Masters, G.J., Hodkinson, I.D., Awmack, C., Bezemer, T.M., Larsson, S., Ekbom, B. & Bjorkman,¨ C. (2000) Influence of plant Brown, V.K., et al. (2002) Herbivory in global climate change quality on pine sawfly population dynamics. Oikos, 89, 440–450. research: direct effects of rising temperatures on insect herbivores. Lima, S.L. & Dill, L.M. (1990) Behavioral decisions made under the Global Change Biology, 8, 1–16. risk of predation: a review and prospectus. Canadian Journal of Bates, D., Maechler, M. & Bolker, B. (2011) lme4: linear mixed- Zoology, 68, 619–640. effects models using S4 classes. R package version 0.999375-42 . Lind, P. & Kjellstrom,¨ E. (2008) Temperature and Precipitation [WWW document] URL http://CRAN.R-project.org/package=lme4 Changes in Sweden, a Wide Range of Model-Based Projections for [accessed on 3 December 2012]. the 21st Century. Reports Meteorology and Climatology No 113. Beck, S.D. (1980) Insect Photoperiodism. Academic Press, New York, SMHI, Norrkoping,¨ Sweden. New York. Lyytikainen-Saarenmaa,¨ P. & Tomppo, E. (2002) Impact of sawfly Berggren, A.,˚ Bjorkman,¨ C., Bylund, H. & Ayres, M.P. (2009) The defoliation on growth of Scots pine Pinus sylvestris (Pinaceae) and distribution and abundance of animal populations in a climate of associated economic losses. Bulletin of Entomological Research, 92, uncertainty. Oikos, 118, 1121–1126. 137–140. Bidart-Bouzat, G.M. & Imeh-Nathaniel, A. (2008) Global change Matsuki, M., Ayres, M.P. & MacLean, S.F. (1994) Temperature effects on plant chemical defenses against insect herbivores. Journal effects on growth and molt of Nematus calais (Hymenoptera: of Integrative Plant Biology, 50, 1339–1354. Tenthredinidae). Environmental Entomology, 23, 719–725. Bjorkman,¨ C. & Gref, R. (1993) Survival pf pine sawflies in cocoon Nerg, A., Kainulainen, P., Vuorinen, M., Hanso, M., Holopainen, stage in relation to resin acid content of larval food. Journal of J.K. & Kurkela, T. (1994) Seasonal and geographical variation of Chemical Ecology, 19, 2881–2890. terpenes, resin acids and total phenolics in nursery grown seedlings Bjorkman,¨ C. & Larsson, S. (1991) Pine sawfly defence and variation of Scots pine (Pinus sylvestris L.). New Phytologist, 128, 703–713. in host plant resin acids: a trade-off with growth. Ecological Niemela,¨ P., Tuomi, J. & Lojander, T. (1991) Defoliation of the Entomology, 16, 283–289. Scots pine and performance of diprionid sawflies. Journal of Animal Bjorkman,¨ C., Bylund, H., Klapwijk, M.J., Kollberg, I. & Schroeder, Ecology, 60, 683–692. M. (2011) Insect pests in future forests: more severe problems? Olofsson, E. (1987) Mortality factors in a population of Neodiprion Forests, 2, 474–485. sertfier (Hymenoptera: Diprionidae). Oikos, 48, 297–303. Blanckenhorn, W.U. & Demont, M. (2004) Bergmann and converse Bergmann latitudinal clines in arthropods: two ends of a continuum? Olofsson, E. (1988) Dispersal of the nuclear polyhedrosis-virus of Integrative and Computational Biology, 44, 413–424. Neodiprion sertifer from soil to pine foliage with dust. Entomologia Forschler, B.T., Young, S.Y. & Felton, G.W. (1992) Diet and sus- Exprimentalis et Applicata, 46, 181–186. ceptibility of Helicoverpa-zea (Noctuidae, Lepidoptera) to a nuclear Parmesan, C. (1996) Climate and species’ range. Nature, 382, polyhedrosis-virus. Environmental Entomology, 21, 1220–1223. 765–766. Frid, L. & Myers, J.H. (2002) Thermal ecology of western tent Parmesan, C. (2007) Influences of species, latitudes and methodologies caterpillars Malacosoma californicum pluviale and infection by on estimates of phonological responses to global warming. Global nucleopolyhedrovirus. Ecological Entomology, 27, 665–673. Change Biology, 13, 1860–1872. Gotthard, K., Nylin, S. & Wiklund, C. (2000) Individual state Development Core Team, R. (2011) R: A Language and Environment controls temperature dependence in a butterfly (Lasiommata maera). for Statistical Computing. R Foundation for Statistical Computing, Proceedings of the Royal Society of London, Series B, 267, 589–593. Vienna, Austria. Gref, R. (1982) Diterpenoids and some water-soluble compounds of Roff, D.A. (2002) Life History Evolution. Sinauer Associates, Inc., Pinus sylvestris and their possible relationship to resistance to fungal Sunderland, Massachusetts.

Sallas, L., Luomala, E.-M., Utriainen, J., Kainulainen, P. & Tvermyr, S. (1969) Effect of nuclear polyhedrosis virus in Neodiprion Holopainen, J.K. (2003) Contrasting effects of elevated carbon diox- sertifer (Geoffr.) [Hymenoptera: Diprionidae] at different tempera- ide concentration and temperature on Rubisco activity, chlorophyll tures. Entomophaga, 14, 245–250. ﬂuorescence, needle ultrastructure and secondary metabolites in Van der Have, T.M. & de Jong, G. (1996) Adult size in ectotherms: conifer seedlings. Tree Physiology, 23, 97–108. temperature effects on growth and differentiation. Journal of Schoonhoven, L.M., Van Loon, J.J.A. & Dicke, M. (2005) Insect-plant Theoretical Biology, 183, 329–340. Biology. Oxford University Press, Oxford, U.K. White, T.C.R. (1984) The abundance of invertebrate herbivores in Sosa-Gomez,´ D.R., Alves, S.B. & Marchini, L.C. (1991) Variation in relation to the availability of nitrogen in stressed food plants. the susceptibility of Bombyx mori L. to nuclear polyhedrosis virus Oecologia, 63, 90–105. when reared on different mulberry genotypes. Journal of Applied Yang, Y.L., Stamp, N.E. & Osier, T.L. (1996) Effects of temperature, Entomology, 111, 318–320. multiple allelochemicals and larval age on the performance of a Stamp, N.E. & Bowers, M.D. (1994) Effect of temperature and leaf age specialist caterpillar. Entomologia Experimentalis et Applicata, 79, on growth versus molting time of a generalist caterpillar fed plantain 335–344. (Plantago lanceolata). Ecological Entomology, 19, 199–206. Stamp, N.E., Temple, M.P., Traugott, M.S. & Wilkens, R.T. (1994) Accepted 20 December 2012 Temperature allelochemical interactive effects on performance of Manduca sexta caterpillars. Entomologia Experimentalis et Applicata, 73, 199–210. Stearns, S.C. (1992) The Evolution of Life Histories. Oxford University Press, Oxford, U.K.