Physiological Entomology
Timing of diapause te rmination in varying winter climates
Journal: Physiological Entomology
ManuscriptFor ID Draft Review Only
Wiley - Manuscript type: Special Issue Article
Date Submitted by the Author: n/a
Complete List of Authors: Lehmann, Philipp; Stockholms Universitet Van Der Bijl, Wouter; Stockholms Universitet Nylin, Soren; Stockholm University, Wheat, Christopher; Stockholms Universitet Gotthard, Karl; Stockholms Universitet
Keywords: biological clock, butterfly, climate change, phenology, Pieridae, stress
In temperate insects, winters are typically endured by entering diapause, a deep resting stage. Correct timing of diapause termination is vital for synchronization of emergence with conspecifics and for mobilizing resources when conditions for growth and reproduction become favourable. Although critical to survival, the intrinsic and extrinsic drivers of diapause termination timing are poorly understood. Here we investigated diapause development under a range of durations (10 to 24 weeks) spent at different temperatures (-2 to 10°C) in the pupal diapausing butterfly, Pieris napi. We determine 1) the maximum temperature for diapause development, 2) whether diapause termination is distinct or gradual and 3) if pupae in diapause count cold days or cold sums. Results indicate large and idiosyncratic effects of nonlethal thermal stress on diapause Abstract: development in P. napi. While all temperatures tested lead to diapause termination, a thermal optimum between 2 and 4°C was observed. Lower temperatures lead to decreased hatching propensity while higher temperatures slowed development and increased emergence desynchronization. These data suggest that rather than a simple cold- summing process with a distinct diapause termination point, there are trade-offs between time and temperature at the low and high end of the thermal range, resulting in a nonlinear thermal landscape showing a ridge of increasing hatching propensity. The current study suggests that the effects of temperature on diapause development should be included in projections on post-winter phenology models of insects, including pest species.
Page 1 of 20 Physiological Entomology
For Review Only Physiological Entomology Page 2 of 20
1 Timing of diapause termination in varying winter climates
2 3 Special Issue in Physiological Entomology: Diapause and seasonal morphs 4 5 Philipp Lehmann, Wouter van der Bijl, Sören Nylin, Christopher W. Wheat, Karl Gotthard 6 7 Department of Zoology, SE 106 91, University of Stockholm, Sweden 8 9 Corresponding author:For Philipp Lehmann,Review philipp.lehm [email protected]
1
Page 3 of 20 Physiological Entomology
10 Abstract 11 12 In temperate insects, winters are typically endured by entering diapause, a deep resting stage. 13 Correct timing of diapause termination is vital for synchronization of emergence with 14 conspecifics and for mobilizing resources when conditions for growth and reproduction 15 become favourable. Although critical to survival, the intrinsic and extrinsic drivers of 16 diapause termination timing are poorly understood. Here we investigated diapause 17 development under a range of durations (10 to 24 weeks) spent at different temperatures ( 2 18 to 10°C) in the pupal diapausing butterfly, Pieris napi . We determine 1) the maximum 19 temperature for diapauseFor development, Review 2) whether diapause Only termination is distinct or gradual 20 and 3) if pupae in diapause count cold days or cold sums. Results indicate large and 21 idiosyncratic effects of nonlethal thermal stress on diapause development in P. napi . While 22 all temperatures tested lead to diapause termination, a thermal optimum between 2 and 4°C 23 was observed. Lower temperatures lead to decreased hatching propensity while higher 24 temperatures slowed development and increased emergence desynchronization. These data 25 suggest that rather than a simple cold summing process with a distinct diapause termination 26 point, there are trade offs between time and temperature at the low and high end of the 27 thermal range, resulting in a nonlinear thermal landscape showing a ridge of increasing 28 hatching propensity. The current study suggests that the effects of temperature on diapause 29 development should be included in projections on post winter phenology models of insects, 30 including pest species. 31 32 Keywords: biological clock, butterfly, climate change, phenology, Pieridae, stress
2
Physiological Entomology Page 4 of 20
33 Introduction 34 35 Seasonally varying environments are characterized by temporal heterogeneity in the 36 distribution of resources and stress. Therefore organisms need to synchronize their life cycles 37 with periods of local resource abundance (Tauber et al. , 1986). During periods of low 38 resource availability and abiotic stress, many insects enter diapause, which is a deep, 39 preprogrammed, neurohormonally regulated resting stage characterized by developmental 40 and reproductive arrest, metabolic suppression and increased abiotic stress tolerance (Koštál, 41 2006). 42 For Review Only 43 In many insects diapause entails two timing related decisions, the entry into, and the 44 termination of, diapause. Entry into diapause occurs during summer when conditions 45 generally still are permissive for growth and development, and thus the activation of the 46 diapause pathway precedes deterioration in the environment (Tauber et al. , 1986; Koštál, 47 2006). There are many studies showing adaptive variation and clines in timing mechanisms 48 that regulate diapause entry (e.g. Sadakiyo & Ishihara, 2011; Urbanski et al. , 2012; Aalberg 49 & Gotthard, 2015; Lehmann et al. , 2015a), ensuring that a maximal proportion of the growth 50 season can be utilized without risking mortality of sensitive growth stages in late summer and 51 autumn. 52 53 Termination of diapause is a dynamic developmental process, best viewed as subdivided into 54 endogenous and exogenous diapause. Endogenous diapause is defined as the part of diapause 55 when insects are nonresponsive to external conditions in the maintenance of the diapause 56 phenotype (Koštál, 2006). At some point in diapause progression during winter, the 57 endogenous state shifts to exogenous diapause, during which insects can respond 58 developmentally to external conditions. This means that the diapause phenotype, while in 59 itself often relatively unchanged until conditions improve, shifts from being internally 60 maintained to being externally maintained (Hodek, 2002). Once the insects have transitioned 61 to exogenous diapause, increasing temperatures or photoperiods lead to post diapause 62 development and the emergence of a new generation of reproductively competent adults. 63 Although the termination of endogenous diapause is a critical component of the seasonal 64 timing of life cycles, the influence of overwintering conditions on the transition from 65 endogenous to exogenous diapause remains poorly understood. 66
3
Page 5 of 20 Physiological Entomology
67 The green veined white Pieris napi Linnaeus (Lepidoptera:Pieridae), is a common butterfly 68 with facultative diapause that shows a strong diapause phenotype (Forsberg & Wiklund, 69 1988; Lehmann et al. , 2016). A previous study has shown latitudinal variation in the duration 70 of endogenous diapause duration, which becomes longer with increasing latitude 71 (Posledovich et al. , 2015a). This might be a consequence of longer winter durations, acting to 72 keep populations synchronized with local season lengths. For the diapausing population, 73 synchronization of post diapause development with the return of permissive conditions can 74 be important from both energetic (Hahn & Denlinger, 2007) and thermal tolerance 75 perspectives (Denlinger & Lee, 2010), with the synchronous emergence of short lived adults 76 critical for successfulFor mating (HodekReview & Hodková, 1988), Only as well as the synchronization of 77 life cycles with host availability in host specialist species (Posledovich et al. , 2015a, b; 78 Stålhandske et al. , 2015). 79 80 In many insects, the termination of diapause is critically dependent on having experienced 81 enough chilling (Hodek, 2002), in a process similar to cold vernalization in plants (Brunner et 82 al. , 2014). Similarly, it is known that the period spent at low temperature affects post winter 83 hatching propensity in several species of insects, including P. napi (Forsberg & Wiklund, 84 1988; Posledovich et al. , 2015a) and closely related butterflies (Xiao et al. , 2013; 85 Stålhandske et al. , 2015). It is also known that cold shocks can speed up diapause 86 development in P. napi (Posledovich et al. , 2015a). These data suggest that a shorter time 87 spent at lower temperature might equal a longer time spent at higher temperature, and 88 indicate the presence of some form of cold degree day counting mechanism (Hodek & 89 Hodkova, 1988; Hodek, 2002). It is also possible that diapause development operates 90 optimally under a certain temperature and that low temperatures lead to thermal stress and 91 high temperature to energetic stress. 92 93 In the current study we investigate these aforementioned aspects of diapause development by 94 rearing P. napi pupae across a range of diapause temperatures for varying amounts of time. 95 Thus we study pupal diapause development across a landscape of cold degree days. This 96 landscape is designed around a temperature and duration that leads to 50% termination in the 97 studied population (Forsberg & Wiklund, 1988; Posledovich et al. , 2015a). We ask the three 98 following general questions: (1) What is the maximum temperature that leads to termination? 99 (2) Is diapause termination best viewed as a distinct point or rather as a period? (3) Do pupae
4
Physiological Entomology Page 6 of 20
100 seem to accumulate cold degree days, i.e. does a shorter time at a lower temperature equal a 101 longer time spent at a higher temperature? 102 103 Materials and Methods 104 105 Study animals and rearing conditions 106 Experiments were performed between April 2015 and July 2016 at the Department of 107 Zoology at Stockholm University, Sweden. Eggs of P. napi were collected from wild plants 108 from a two sites (~20 km apart) in Skåne, southern Sweden (Kullaberg; 56°18’N, 12°27’E 109 and Vejbystrand; 56°18’N,For 12°46’E) Review and brought to the Onlylaboratory during the 110 summer/autumn in 2013 and 2014. Butterflies of both sexes (P1) were kept in cages 111 (0.8×0.8×0.5 m) and provided with a host plant, Horseradish Armoracia rusticana Gaertner, 112 Meyer & Scherbius (Brassicales:Brassicaceae) for oviposition at 25°C and 60% relative 113 humidity (RH) under long day conditions (next to large windows and under 400 W metal 114 halide lamps) and fed sugar water. Their offspring (F1) were mass reared to adulthood on A. 115 rusticana at a long day photoperiod (22L:2D, 20°C, 80% RH) and re mated under similar 116 conditions as their parents. 117 118 For the experiments, the F2 generation was mass reared to pupation in groups of 250 on A. 119 rusticana leaves under conditions inducing diapause (12L:12D, 17°C, 80% RH) in climate 120 controlled rooms. A total of 1500 pupae were reared. Pupae were left at 17°C for 1 month, 121 then split among dark climate cabinets (KB8400L, Termaks, Bergen, Norway) set to 2, 0, 2, 122 4, 6, 8 and 10°C. For all treatments between 2 and 6 degrees we used 10, 12, 13, 14, 15, 16 123 and 22 weeks. For temperature treatments 8 and 10°C we only used a duration of 24 weeks. 124 Of these, the animals having that spent 10 weeks at 6°C had the lowest cold accumulation 125 and the ones that spent 22 weeks at 2°C had the highest cold accumulation. Sample sizes per 126 temperature and time group was between 36 and 40 individuals with roughly equal amounts 127 of females and males. 128 129 Diapause occurred in dark cabinets with stable temperature (set temperature ± 0.5°C). 130 Cabinets were randomly re programmed and switched every other week to avoid cabinet 131 effects. After the designated holding time at the respective temperatures the pupae were 132 moved to a 20°C cabinet at 18L:6D, 80% RH until they hatched or died (mortality is easily 133 identified since dead pupae become black or dry up). Pupae were checked daily until day 20
5
Page 7 of 20 Physiological Entomology
134 and then every other day. Hatching was marked as either normal or abnormal. Normal 135 hatching entailed complete exit from the exuvia, successful wing formation, the capacity to 136 fly and drink water within 48 hours. Only normal hatching individuals were included since 137 other studies indicate that latent damage faced during diapause (e.g. cold shocks) can lead to 138 difficulties in completing metamorphosis (Lehmann P, unpublished). The traits investigated 139 were: mortality during diapause, mortality during development, hatching propensity, 140 malformation propensity and development time (days from removal from cold to adult 141 hatching). Of these, mortality during diapause and development were very low (0.1% and 0% 142 respectively), and therefore omitted from the analyses. 143 For Review Only 144 Statistical analyses 145 For the statistical analysis of the data, factorial generalized linear models (GLM) were 146 employed. Model improvement was tracked through the Akaike Information Criterion (AIC) 147 and nonsignificant terms were removed starting from highest order interactions (Sokal & 148 Rohlf, 2003). While the full treatment range is shown in figures 2 and 3 (i.e. 2 to 10°C and 149 10 to 24 weeks), statistical analyses are performed only on the 2 to 6°C and 10 to 22 week 150 groups to uphold orthogonality. The statistical tests shown below were performed in SPSS, 151 version 24.0 (IBM SPSS Inc., Chicago, Illinois). Malformation propensity was analyzed with 152 a binary logistic GLM using malformation (yes/no) as dependent variable and holding time 153 (10 to 22 weeks) as well as diapause temperature ( 2 to 6°C) as continuous explanatory 154 variables. Holding time was not significant and removed during model selection. 155 Development time was analyzed with a GLM using the logarithm of development time (days) 156 as dependent variable and holding time (10 to 22 weeks) as well as diapause temperature ( 2 157 to 6°C) as continuous and sex as categorical explanatory variables. Hatching propensity was 158 investigated with a binary logistic GLM using hatching (yes/no) as dependent variable and 159 holding time (10 to 22 weeks) as well as diapause temperature ( 2 to 6°C) as continuous and 160 sex as categorical explanatory variable. Only adults that hatched without apparent flaws 161 within 30 days upon removal from the holding temperature were included (see results). All 162 interactions were removed during model selection. 163 164 Results 165 166 The maximum temperature leading to diapause termination
6
Physiological Entomology Page 8 of 20
167 All diapause temperatures, including both the lowest ( 2°C) and highest (10°C), led to 168 successful diapause termination in some individuals. The only group from which no adults 169 hatched was the 0°C group after 10 weeks. The highest hatching proportion was 93%, seen in 170 the group kept at 6°C for 24 weeks. 171 172 The effect of holding time and temperature on malformation propensity 173 Malformation propensity increased with temperature in a time independent fashion (Table 174 1a). While no malformed individuals hatched from the 2 and 0 day groups, the 2, 4 and 6 175 degree groups had 2 ± 1.1% (mean ± SE), 3 ± 1.2% and 6 ± 1.8% malformed individuals 176 respectively. For Review Only 177 178 The effect of holding time and temperature on development time 179 There were strong effects of time, temperature and sex on development time (Table 1b). 180 Generally, development time decreased with both time spent in diapause and temperature, 181 excepting the lowest temperature, where development time and its variance was very high, 182 especially at the short diapause durations (Fig. 1). It is important to note that model power is 183 low at the lowest temperature due to low hatchability in this group. Overall males had shorter 184 development times and seemed to be influenced by diapause duration to a smaller, as 185 compared to females (Table 1b). The majority of pupae, 87% (552/636), hatched within the 186 first 30 days following removal from diapause holding temperatures. 187 188 The effect of holding time and temperature on hatching propensity 189 There was a strong effect of time, temperature and sex on hatching propensity (Table 1c and 190 Fig. 2). The longer pupae stayed in diapause, at any given temperature, the higher the 191 hatching propensity became. Also temperature had a strong effect on hatching propensity, 192 hatching propensity was lowest at low temperatures and increased with diapause holding 193 temperature (Fig. 2). Furthermore, the sexes differed in hatching propensity, with more males 194 hatched at any given time and temperature than females (Table 1c). 195 196 To further investigate the temperature effects, the data was split by holding time and curve 197 fitting metrics employed to investigate the relationship between temperatures and hatching 198 propensity. Cubic functions consistently best explained the within component of the holding 199 time data (Table 2). These data suggest that temperature affects hatching propensity across all 200 holding times in a non linear fashion both at the low and the high end of the thermal range.
7
Page 9 of 20 Physiological Entomology
201 Since cubic functions seem to fit the data best, there were features of the temperature effect 202 that the binary logistic GLM could not detect. There appears to be a detrimental effect of high 203 temperatures beyond 4 degrees since hatching propensity decreased between 4 and 6 degrees 204 at all holding times. A simple two way binomial test (0.5 7 x 2) indeed suggests that the 205 probability of this consistent effect occurring randomly is 0.0156 (i.e. < 0.05). 206 207 Discussion 208 209 The present study sets out to investigate diapause development under a range of durations 210 spent at different temperaturesFor inReview the pupal diapausing butterfly,Only Pieris napi . We show that, 211 in a population from Southern Sweden, diapause development proceeds optimally at 2 to 4°C. 212 All temperatures lead to some degree of diapause termination, including the lowest ( 2°C) 213 and highest (10°C) tested temperature. There are strong effects of time and temperature on 214 hatching propensity. The effect of time is rather straightforward and linear with an increase in 215 hatching propensity with time spent in diapause, regardless of temperature. Overall, the effect 216 of time is stronger than the effect of temperature, as also noted in other species, such as the 217 Chinese citrus fly Bactrocera minax Enderlein (Diptera:Tephritidae) (Dong et al. , 2013). 218 Furthermore, males overall have higher hatchability than females, as well as shorter 219 development time. This result is in agreement with previous data on protandry in P. napi 220 (Forsberg & Wiklund, 1988). The effect of temperature is more complex, as hatching 221 propensity decreases from the optimum both towards the lower and higher end of the thermal 222 range. Nevertheless, these data suggest that rather than a simple cold summing process, there 223 are trade offs between time and temperature at the low and high end of the thermal range, 224 that produces a thermal landscape showing a ridge (Fig. 3) where the highest hatching 225 propensity is seen at relatively high temperatures at start, after 14 to 16 weeks, and 226 subsequently shifts increasingly to favour lower temperatures after 22 and 24 weeks. These 227 data suggest that low and high temperatures incur stress that varies with time. 228 229 While low temperatures (< 2°C) are associated with lowered hatching propensity, there are 230 no apparent cost in terms of development time. In fact, the shortest development times were 231 found in lower, rather than higher temperatures at most holding times (Fig. 1). The lowered 232 hatching propensity is unlikely to be due to direct cold tolerance related stress for two 233 reasons. First, since mortality during diapause is very low (0.1%) no clear association 234 between mortality, temperature and time is suggested (data not shown). Second, unpublished
8
Physiological Entomology Page 10 of 20
235 data shows that diapausing P. napi pupae survive periods of up to 30 days at 10°C and still 236 hatch normally (Lehmann P., unpublished). 237 238 In other studies low constant temperatures have been shown to disrupt diapause development, 239 although the optimal thermal range depends on conditions that are normally faced in local 240 habitats (Pullin & Bale, 1989; Nomura & Ishikawa, 2000). The findings of the current study 241 are in agreement, and it seems more likely that too low and constant temperatures could 242 disrupt time keeping mechanisms associated with diapause development, perhaps by 243 preventing cycling. This disruption seemed to be non linearly compensated by spending 244 longer time at low Fortemperature, Review shown by the high hatching Only propensity in the 22 week pupae 245 at 2°C. Thus since overall hatching propensity increased with time it is possible that some 246 offset of high or low temperature could occur given enough time, but the current data 247 suggests this is more likely to occur at the lower than upper end of the thermal range. 248 249 The higher diapause temperatures (> 2°C) incur a different impact on diapause development 250 than the lower temperatures, described above. While diapause at higher temperatures leads to 251 a small decrease in hatching propensity, there is a much clearer detrimental effect in the form 252 of development time extension. Both the 8 and 10°C groups had a longer mean development 253 time (36 and 39 days respectively) than the chosen 30 day cut off limit. Not only does high 254 temperature lead to increasing development time, it also leads to strong desynchronization of 255 emergence (increased variation in hatching), that was stronger in females than males (Fig. 1). 256 Indeed, standard deviation of mean development time was lowest in the group that had the 257 shortest development time in each time group (Fig. 1) indicating that faster development 258 equals more synchronized emergence. This indicates that there could exist a thermal limit to 259 diapause development that lies between 6 and 8°C within the range of diapause durations 260 tested here. Importantly, since pupae still developed and hatched at the high temperatures, 261 this thermal limit reflects an extrinsic ecological rather than intrinsic physiological limitation 262 (Hodek, 2002). 263 264 The extension of development time might be due to metabolic depletion of primary energy 265 stores during diapause at high temperature (Irwin & Lee, 2003; Lehmann et al. , 2016) that 266 either depletes energy stores, or forces pupae to switch to other energy sources. This might 267 entail the mobilization of sub optimal substrates both for anabolic and katabolic processes. In 268 many diapausing insects, including P. napi (Lehmann et al. , 2016), only small energy
9
Page 11 of 20 Physiological Entomology
269 depletion occurs during winter, and stored energy is primarily saved for post diapause 270 processes such as growth, migration and reproduction (Hahn & Denlinger, 2007). 271 272 This study strongly suggests that diapause termination is best viewed as a period of 273 increasing competence for, or probability of, termination, rather than a distinct population 274 level point (Fig. 3). The thermal landscape covers a range of winter conditions, and while 275 constant temperatures do not reflect the extent of daily, weekly and monthly variability in 276 temperature seen in the field (Tauber et al. , 1986), the current setup shows reaction norms 277 against varying temperature sums. Since hatching occurs gradually with time towards the 278 optimum temperaturesFor (2 to 4°C), Review the notion of a distinct Only termination point receives weak, if 279 any, support in the current study. 280 281 A general question that arises is whether and to what extent thermal variability simply moves 282 the termination period forward or backward, or if the actual phenotype at termination is 283 changed? Put another way, how robust is the diapause phenotype against thermal variability 284 (Hodek, 2002; Koštál, 2006; Williams et al. , 2012)? Further studies need to target fluctuating 285 temperatures and cold as well as heat shocks experienced during diapause. These further 286 experiments hinge critically on obtaining better estimates on microclimate variability in the 287 field. Current models are generally based on air temperature measurements that likely 288 correlate poorly with sheltered microhabitats (Lenoir et al. , 2016). Also, while the current 289 study suggests that diapause can be terminated under a range of temperatures and conditions, 290 it says nothing about adult viability and fecundity that might be compromised especially at 291 higher diapause temperatures (Andrewartha, 1952). 292 293 In conclusion, the current study shows that there are large and idiosyncratic effects of 294 nonlethal thermal stress on diapause development in P. napi . Low temperatures lead to 295 decreased hatching propensity while high temperatures lead to slowed development and 296 emergence desynchronization. Studies investigating diapause ecophysiology often focus on 297 the effect of temperature on direct thermal or energetic stress (Irwin & Lee, 2003, Williams et 298 al. , 2012, Koštál et al. , 2014, Lehmann et al. , 2015b, Marshall & Sinclair, 2015). The effect 299 of temperature on diapause development per se is less studied (Xiao et al. , 2013), partially 300 due to lack of good model systems where diapause termination is well characterized. The 301 current study suggests that the effects of temperature on diapause development can be subtle 302 but of large impact and should be included in projections on post winter phenology models
10
Physiological Entomology Page 12 of 20
303 (Tauber et al. , 1986; Stålhandske et al. , 2015) of for instance pest species. Further, this study 304 also shows that predicting the consequences of climactic variability and climate change is 305 difficult since small differences in the thermal landscape can have large consequences on 306 diapause and thus post diapause development and ultimately spring phenology of temperate 307 insects. 308 309 Acknowledgements 310 311 We thank Shin Goto and Daniel Hahn for inviting us to contribute to this special issue. Björn 312 Rogell is thanked forFor statistical Reviewconsulting. The work was Only financed by the Strategic Research 313 Program Ekoklim at Stockholm University, the Knut and Alice Wallenberg Foundation 314 (KAW 2012.0058) and the Swedish Research Council grants (to SN) VR 2012 3715 and (to 315 CWW) VR 2012 4001. The authors declare no conflicts of interests.
11
Page 13 of 20 Physiological Entomology
316 References 317 318 Aalberg Haugen, I. & Gotthard, K. (2015) Diapause induction and relaxed selection on 319 alternative developmental pathways in a butterfly. Journal of Animal Ecology , 84 , 464 472. 320 Andrewartha, H.G. (1952) Diapause in relation to the ecology insects. Biological Reviews , 321 27 , 50 107. 322 Brunner, A.M., Evans, L.M., Hsu, C. Y. & Sheng, X. (2014) Vernalization and the chilling 323 requirement to exit bud dormancy: shared or separate regulation? Frontiers in Plant Science , 324 5, 732. 325 Denlinger, D.L. & ForLee, R.E.Jr. Review(2010) Low temperatu reOnly biology of insects. Cambridge 326 University Press, UK. 327 Dong, Y., Wang, Z., Clarke, A.R., Pereira, R., Desneux, N. & Niu, C. (2013) Pupal diapause 328 development and termination is driven by low temperature chilling in Bactrocera minax . 329 Journal of Pest Science , 86 , 429 436. 330 Forsberg, J. & Wiklund, C. (1988) Protandry in the Green Veined White Butterfly, Pieris 331 napi L. (Lepidoptera; Pieridae). Functional Ecology , 2, 81 88. 332 Hahn, D.A. & Denlinger, D.L. (2007) Meeting the energetic demands of insect diapause: 333 nutrient storage and utilization. Journal of Insect Physiology , 53 , 760 773. 334 Hodek, I. (2002) Controversial aspects of diapause development. European Journal of 335 Entomology , 99 , 163 173. 336 Hodek, I. & Hodková, M. (1988) Multiple roles of temperature during insect diapause: a 337 review. Entomologica Experimentalis et Applicata , 4 9, 153 165. 338 Irwin, J.T. & Lee, R.E.Jr. (2003) Cold winter microenvironments conserve energy and 339 improve overwintering survival and potential fecundity of the goldenrod gall fly, Eurosta 340 solidaginis . Oikos , 100 , 71 78. 341 Koštál, V. (2006) Eco physiological phases of insect diapause . Journal of Insect Physiology , 342 52 , 113 127. 343 Koštál, V., Miklas, B., Doležal P, Rozsypal J. & Zahradníčková, H. (2014) Physiology of 344 cold tolerance in the bark beetle, Pityogenes chalcographus and its overwintering in spruce 345 stands. Journal of Insect Physiology , 63 , 62 70. 346 Lehmann, P., Lyytinen, A., Piiroinen, S. & Lindström, L. (2015a) Latitudinal differences in 347 diapause related photoperiodic responses of European Colorado potato beetles ( Leptinotarsa 348 decemlineata ). Evolutionary Ecology , 29 , 269 282.
12
Physiological Entomology Page 14 of 20
349 Lehmann, P., Kaunisto, S., Koštál, V., Margus, A., Zahradníčková, H. & Lindström, L. 350 (2015b) Comparative ecophysiology of cold tolerance related traits: assessing range 351 expansion potential for an invasive insect at high latitude. Physiological and Biochemical 352 Zoology , 3, 254 265. 353 Lehmann, P., Pruisscher, P., Posledovich, D., Carlsson, M., Käkelä, R., Tang, P., Nylin, S., 354 Wheat, C.W., Wiklund, C. & Gotthard, K. (2016) Energy and lipid metabolism during direct 355 and diapause development in a pierid butterfly. Journal of Experimental Biology , 219 , 3049 356 3060. 357 Lenoir, J., Hattab, T. & Pierre, G. (2016) Climatic microrefugia under anthropogenic climate 358 change: implicationsFor for species Review redistribution. Ecography Only, In Press, doi: 10.1111/ecog.02788 359 Marshall, K. & Sinclair, B. (2015) The relative importance of number, duration and intensity 360 of cold stress events in determining survival and energetics of an overwintering insect. 361 Functional Ecology , 29 , 357 366. 362 Nomura, M. & Ishikawa, Y. (2000) Biphasic effect of low temperature on completion of 363 winter diapause in the onion maggot, Delia antiqua . Journal of Insect Physiology , 46 , 373 364 377. 365 Posledovich, D., Toftegaard, T., Wiklund, W., Ehrlén, J. & Gotthard, K. (2015a) Latitudinal 366 variation in diapause duration and post winter development in two pierid butterflies in 367 relation to phenological specialization. Oecologia , 177 , 181 190. 368 Posledovich, D., Toftegaard, T., Wiklund, C., Ehrlén, J. & Gotthard, K. (2015b) The 369 developmental race between maturing host plants and their butterfly herbivore – the influence 370 of phenological matching and temperature. Journal of Animal Ecology , 84 , 1690 1699. 371 Pullin, A.S. & Bale, J.S. (1989) Effects of low temperature on diapausing Aglais urticae and 372 Inachis io overwintering physiology. Journal of Insect Physiology , 35 , 277 281. 373 R Core Team (2016) R: A language and environment for statistical computing. R Foundation 374 for Statistical Computing, Vienna, Austria. URL https://www.R project.org/. 375 Sadakiyo, S. & Ishihara, M. (2011) Rapid seasonal adaptation of an alien bruchid after 376 introduction: geographic variation in life cycle synchronization and critical photoperiod for 377 diapause induction. Entomologia Experimentalis et Applicata , 140 , 69 76. 378 Sokal, R.R. & Rohlf, F.J. (2003) Biometry: The Principles and Practice of Statistics in 379 Biological Research.WH Freeman and Company, New York, New York. 380 Stålhandske, S., Lehmann, P., Pruisscher, P. & Leimar, O. (2015) Effect of winter cold 381 duration on spring phenology of the orange tip butterfly, Anthocharis cardamines . Ecology & 382 Evolution , 5, 5509 5520.
13
Page 15 of 20 Physiological Entomology
383 Tauber, M.J, Tauber, C.A. & Masaki, S. (1986) Seasonal adaptations of insects. Oxford 384 University Press, New York. 385 Urbanski, J., Mogi, M., O’Donnell, D., DeCotiis, M., Toma, T., Armbruster, P. (2012) Rapid 386 adaptive evolution of photoperiodic response during invasion and range expansion across a 387 climatic gradient. The American Naturalist , 179 , 490 500. 388 Wickham, H. (2009) ggplot2: Elegant Graphics for Data Analysis. Springer Verlag New 389 York. 390 Williams, C.M., Marshall, K.E., MacMillan, H.A., Dzurisin, J.D.K., Hellmann, J.J. & 391 Sinclair, B.J. (2012) Thermal variability increases the impact of autumnal warming and 392 drives metabolic depressionFor in anReview overwintering butterfly. Only PLoS ONE , 7, e34470. 393 Wood, S.N. (2011) Fast stable restricted maximum likelihood and marginal likelihood 394 estimation of semiparametric generalized linear models. Journal of the Royal Statistical 395 Society (B) , 73 , 3 36. 396 Xiao, H., Wu, S., Chen, C. & Xue, F. (2013) Optimal low temperature and chilling period for 397 both summer and winter diapause development in Pieris melete : based on a similar 398 mechanism. PLoS ONE , 8, e56404.
14
Physiological Entomology Page 16 of 20
399 Figure legends 400 401 Fig. 1. The developmental time in days of Pieris napi pupae until adult hatching as a factor of 402 diapause holding temperature (°C) and time in diapause (weeks). For each combination, the 403 median and interquartile range are shown by a circle and line. Additionally, the kernel density 404 estimates are shown as split violins for each sex. That is, the deviation (height) of the filled 405 curves approximates the number of observations for that developmental time. A more narrow 406 peak indicates that individuals are more synchronous in their development. Generally, males 407 emerged earlier and were more synchronous in their emergence than females. The area of the 408 violins is scaled proportionallyFor Reviewto the total number of observations Only within that group. For 409 some groups ( 2°C, 10 to 15 weeks) very few individuals hatched, for these only the median 410 and quartiles are plotted. 411 412 Fig. 2. The amount of Pieris napi adults hatching after diapausing for different amounts of 413 time (weeks) at different diapause holding temperatures (°C). The current figure only 414 includes individuals that hatched without apparent flaw in under 30 days. Note that the 415 statistical model described in the text was run only on the 2 to 6°C and 10 to 22 week 416 groups. The 24 week group is included only for visual comparison. N = 36 40 per point. 417 418 Fig. 3. Summary figure of the diapause thermal landscape in Pieris napi from southern 419 Sweden modeled on the hatching percentages (excluding malformed individuals and using a 420 developmental cutoff of 30 days). The smooth surface was estimated by fitting a generalized 421 additive model (GAM) with penalized thin plate regression splines to the available data, and 422 generating predictions from this model for the complete time by temperature grid. The 423 analysis was performed in R (v3.3.2, R Core Team 2016), using the mgcv package for GAM 424 estimation (v1.8 15, Wood, 2011) and the ggplot2 (v2.2.0, Wickham, 2009) package for 425 visualization. Note the areas with lacking data.
15
Page 17 of 20 Physiological Entomology
426 Table 1 Final GLM models of the effect of time and temperature on the (a) malformation 427 propensity (b) development time and (c) hatching propensity of Pieris napi pupae in 428 diapause. In (c) the effect of time and temperature on the hatching propensity of pupae that 429 developed normally and hatched in under 30 days can be seen. 430 Effect Wald Chi Square df Sig. (a) Malformation propensity Intercept 8.447 1 0.004 Temperature 7.977 1 0.005 b) Development timeFor Review Only Intercept 997.228 1 < 0.001 Time 66.192 1 < 0.001 Temperature 17.317 1 < 0.001 Sex 2.152 1 0.142 Temperature * Time 12.919 1 < 0.001 Sex * Temperature 13.719 1 < 0.001 (c) Hatching propensity Intercept 240.689 1 < 0.001 Time 196.374 1 < 0.001 Temperature 160.215 1 < 0.001 Sex 69.941 1 < 0.001 431
16
Physiological Entomology Page 18 of 20
432 Table 2 Coefficients for cubic regressions on hatching propensity of Pieris napi as a function 433 of temperature on the data split by week. Cubic regressions consistently showed the best fit 434 (full data not shown). B1 is the linear, B 2 the quadratic and B 3 the cubic polynomial 435 coefficient. Malformed adults and adults hatching after 30 days have been removed. 436 Week R2 Intercept B1 B2 B3 10 0.99 0.22 2.59 0.88 0.17 12 0.95 8.85 7.64 0.85 0.26 13 0.95 27.51 8.86 1.39 0.08 14 For 0.99 Review 27.14 12.56Only 0.33 0.27 15 1.00 39.99 16.66 1.34 0.16 16 0.98 53.32 15.03 2.43 0.03 22 0.91 84.53 9.29 3.43 0.28 24 0.99 100.50 6.63 0.84 0.10 437
17
Page 19 of 20 Physiological Entomology
For Review Only
Physiological Entomology Page 20 of 20
For Review Only
Page 21 of 20 Physiological Entomology
For Review Only