We predicted a response to time constraints (TC) imposed by late egg hatching in life history and behavioural traits, and this response to be stronger under warming compared to the current temperature, and to be stronger in the absence of predator cues and presence of invasive alien predator (IAP) cues than in the presence of native predator cues. Our results partially supported these predictions. While we found the predicted partial compensation for a late-season oviposition date in egg and larval development times, and the associated higher larval boldness and activity rate, compensation in development time was only pronounced under warming and in the absence of predator cues. Shorter development times, however, resulted in a lower mass at emergence, an expected fitness cost for the accelerated life history under TC, yet no cost in terms of increased mortality rate was detected. As expected, predator cues delayed development time and, surprisingly, the negative carry-over effect of predator cues experienced in the egg stage was more pronounced in time-constrained damselflies, further delaying life history transition dates. Unexpectedly, cues from the invasive alien predator (IAP) caused a stronger effect than native predator cues and further delayed development, providing no support for the naïve prey hypothesis.
Seasonal compensation under warming and predator cues in the egg stage
Our results demonstrate partial compensation for late-season oviposition dates in egg development time, particularly under warming conditions and in the absence of predator cues. Faster egg development under time-stress situations has been reported in a few other temperate ectotherms 56–58. For example, regardless of temperature conditions, eggs laid during the late season generally had shorter incubation periods compared to those laid early in the season in the lizard Anolis sagrei 58. We report that such compensation also occurs in species with variable voltinism, where time stress can be alleviated by the addition of an extra growth season for juvenile development. This response appears to be adaptive, as earlier egg hatching might provide more time for larval development and growth before wintering, likely increasing winter survival and may reduce larval mortality caused by cannibalism, specifically between individuals from the same hatching date cohort 8,59,60.
The observed slight temperature differences during the first two (warming treatment) and three weeks (current temperature treatment) of egg rearing experienced by early and late eggs are unlikely to have triggered the difference in egg development rate (Table S1). However, the temperature effect was supported by the results based on thermal units, degree-days (DDs), as a larger difference in hatching dates between early and late groups was observed when egg development time was expressed in DDs (compare Fig. 3a vs. Fig. 3b). The somewhat shorter photoperiod encountered by late eggs could be a contributing factor. Seasonally changing day length indicates the remaining time window available before the end of the growth season. Previous studies have highlighted the important role of seasonal changes in photoperiod in shaping hatching phenology 57,61–63. Based on these observations, we suggest that the joint effect of temperature and especially photoperiod triggered earlier hatching in time-constrained eggs. In addition, adaptive maternal effects might also have played a role in determining egg traits, as this initial developmental stage is more likely to be influenced by maternal contributions compared to later ontogenetic stages—larval and adult—as shown in previous studies 64,65.
Predator cues prolonged egg development time, especially in late eggs, hence predator cues annihilated and even reversed the response to TC. Changes in egg traits driven by predator cues were previously reported across different taxa, e.g., amphibians 66,67, salamanders 68, fish 68, molluscs 69 and invertebrates 70, including I. elegans 12,28,31,44. Yet, the finding that the predator cues can further delay hatching when combined with time stress is, to our knowledge, novel. A possible explanation lies at the physiological level 71. A slower development in response to predation pressure could result from physiological stress in the eggs as these are potential prey. This stress might prompt a reallocation of energy towards costly defence mechanisms to maintain homeostasis, such as increased production of heat shock proteins, rather than a rapid development rate 72. Especially delayed egg development under predation pressure in late individuals may be attributed to increased predation risk late in the growth season. Ectothermic predators, such as fish and crayfish, tend to exhibit increased energy requirements in terms of prey consumption as the season advances in order to accumulate and save energy for wintering. Nonetheless, delayed hatching under predation pressure in time-stressed individuals might be adaptive when the fitness cost of predator-induced delayed hatching is lower than the cost of individuals being consumed by predators at hatching.
Surprisingly, damselfly eggs exhibited a more prolonged development under the cues of the IAP signal crayfish than under native perch cues, a pattern particularly pronounced in late individuals reared under current temperatures. We propose that the stronger response to IAP cues is influenced by historical experience with cues from the native Danube crayfish, Astacus leptodactylus, which has been observed in the study pond until 2015 (Bonk M., pers. comm.). This is plausible given that the chemical signals released by the two closely related crayfish species 73 are similar in content 74. Indeed, our previous studies on geographically nearby populations of I. elegans showed that eggs respond similarly to cues produced by two different species of crayfish 44. Finally, the recognition of cues from predators not present in the study pond may also be attributed to the strong between-population gene flow in I. elegans 75. The observation that crayfish cues induced the strongest egg response might also suggest that this developmental stage is more susceptible to predation by crayfish than by fish. Omnivorous crayfish commonly consume substrates where immobile damselfly eggs could be laid, whereas perch primarily prey on mobile targets. Consequently, the magnitude of the prey response might depend on whether the predator signals come from egg or larval predators 76.
Warming reduced the impact of predator cues on egg development time, likely because of accelerated development, leading to a shorter exposure time to the predator cues under elevated temperature. Additionally, the faster biodegradation of predator cues at higher temperatures, as noted in previous case studies 77, may have contributed to this reduction in the predator effect.
Immediate and carry-over effects from egg and larval stage up to adult emergence
Under TC, animals are expected to accelerate development and growth rate in the larval stage, which may come at the cost of an increased mortality due to less allocation of energy to somatic maintenance and repair 6,10,78. The current results confirmed the accelerated development in time-stressed individuals (discussed below), and an associated increase in mortality, but only at the current temperature. This implies that mild warming seems to benefit late hatchers in terms of survival. A possible explanation might be that the experimental warming temperature did not deviate considerably from the real temperature recorded at the sampled pond and could lie within a range of optimal temperatures for survival. This was especially true during the first growth season, i.e., the pre-winter period (Fig. S2, Table S1), when the larval food acquisition rate is likely the highest 79. Due to its characteristics, Płaszowski pond represents a relatively warm freshwater habitat, and I. elegans from this site show shorter development times across temperatures when compared to individuals from a nearby colder site, i.e., cogradient variation in the trait 12. Note that the low survival rate for the early control group reared under the current temperature should be interpreted with caution, as several individuals in this treatment group were accidentally lost during the course of the experiment, leading to a relatively low sample size and hence a large error within this group.
Development time until emergence was reduced for the late group, and more so under warming. Along with theory 80 and similar experimental results 5,81,82, this outcome is likely adaptive because more time-stressed individuals face a shorter time available before wintering and post-winter development up to emergence. Alternatively, late hatchers could take another growth season for development and postpone emergence 41,42, but this did not happen in the current experiment. However, when development time was expressed in DDs, it appeared that late hatchers required fewer DDs than early ones, especially under warming, whereas early individuals under warming needed the most DDs. This further supports the scenario of compensation for TC cued by photoperiod in combination with immediate temperature. We also note that only a four-week-long difference in egg-laying date sufficed in generating the compensation mechanism in a species that experiences relatively long growth and flying season, which lasts approximately 28 and 18 weeks, respectively 83 (Fig. S1, Table S1). In a study by Tüzün et al. 34 on I. elegans from Belgium, larvae from field-collected mothers during the second half of July responded to TC induced by a six-week-long delayed photoperiod in a qualitatively similar way as late hatchers in the current study, indicating the importance of photoperiod.
The shorter development times in late individuals did not fully compensate for TC in terms of emergence date. On average, later hatchers delayed emergence by 10 days compared to early individuals (while the late clutches were collected ca. 30 days later), with late individuals weighing 12% less than the early ones. The absence of full compensation in development time and a lower mass were reported earlier in other ectotherms (reviewed in Dmitriew 2011), including other damselfly species 84,85. Such results might be linked to environmental factors such as suboptimal quality and/or quantity of food provided 21,86,87, genetic constraints in terms of genetic correlations between life history traits 88 and maternal effects (discussed above). Interestingly, similar differences between early and late season damselflies were found in the size of the field-collected mothers that laid the early and late clutches (current results) and in several other, field-sampled odonate species 89,90. Current and previous results show a trade-off between age and mass at emergence 80 which might eventually lead to an adult fitness cost in term of lower mating success 23,91,92, but see 93.
Negative effects of exposure to predator cues in the egg stage were apparent on larval survival until day 14 and less so for survival until emergence, indicating a gradually decreasing but still present carry-over effect from the egg stage. While very few studies demonstrated that predation risk can kill prey in the larval stage 26,94, this has never been shown for carry-over effects of predation risk imposed in the egg stage. The carry-over effect of predator cues in the egg stage until emergence remained only in late individuals reared under warming and treated with signal crayfish cues. This confirms that IA crayfish not only delayed hatching but also led to increased mortality under warming through a carry-over effect. This adds to the negative effect of signal crayfish on native macroinvertebrate abundance and reduced species number in natural freshwater ecosystems 47. Weaker effects of predator cues through time in damselflies exposed all the time to predator cues probably reflected acclimation to biotic stressors. Such acclimation in predator-treated groups was reported in other organisms, e.g., in fish 95 and insects 96, and confirms previous finding in I. elegans 12,31.
Divergent predator cue effects during the egg and larval stages revealed that the variation in development time under predation pressure primarily stemmed from the egg, and not larval reactions to predator cues. The fact that the larval stage was less sensitive in this trait to predator stress could be explained by an acclimation effect (discussed above). Nonetheless, when the total development time was considered (combined egg and larval phases), the carry-over effects of predator cues from the egg up to emergence remained significant. Interestingly, previous studies on I. elegans showed that predator cues prolonged larval development time and/or led to lower mass under warming conditions, but this occurred when individuals were reared in groups and were therefore exposed to other stressors, cannibalistic pressure and alarm cue from conspecifics 8,32, which was absent in the present experiment.
Warming is expected to reduce mass at emergence, the so-called temperature-size rule 97. Instead, warming did not affect mass at emergence in late damselflies, and early individuals became even heavier at the warmer temperature. This might be because their growth trajectories changed under different temperatures. In this scenario, the difference could indicate the up-regulation of gene expression governing development in the early damselflies experiencing mild warming. This pattern of up-regulated genes categorized into different developmental processes such as anatomical structure development in individuals raised under elevated temperatures has been observed in larvae of I. elegans from several populations in southern Poland 32. The fact that time-stressed individuals did not differ in mass across temperatures indicates a fixed response to TC at the phenotypic and gene expression levels. Such a fixed response might constrain the evolution of this key fitness trait in populations facing rapid environmental change 98,99.
Life history responses to TC could at least be partially explained by the adjustment of larval behaviour. Higher boldness and activity rates in time-stressed individuals likely increased prey acquisition and, hence, development rates, as earlier reported in other ectotherms, including odonates 33,34,100,101. However, elevated ecologically risky behaviours, such as increased activity to locate food resources 9,102, could lead to negative consequences such as physiological costs in terms of a lower investment in immune function 103 and elevated mortality 10. In addition, an increased digestive physiology may have contributed to the accelerated life history under TC 104.
Warming negatively affected boldness in both early and late larvae. In general, the rise in temperature increases metabolic rate exponentially in ectotherms, which often influences behaviour 105,106. Here, we obtained a result that remains in contradiction with other studies on I elegans 102,107. However, Debecker and Stoks 102 and 107 conducted their studies under constant thermo-photoperiods, which may yield different results from rearing conducted in a thermo-photoperiod that follows the natural progression of temperature and photoperiod, as demonstrated in the damselfly Lestes sponsa33. Another difference is the more varied diet of the larvae, which from the pre-final instar prior to emergence received additional supplementation with chironomid larvae. This could have led to better conditions of the here studied individuals, and consequently a lower food requirement even under conditions of elevated temperature and increased metabolism. Results similar to ours, with decreased boldness with rising temperatures, have been shown in the invasive crayfish Procambarus clarkii 108.
We expected to see reduced activity in response to predation risk, as earlier indicated in other studies 109,110, including the study species 111. However, we did not find such a response. We suggest that this happened because of acclimation in groups that experienced predation risk during both the egg and larval stages (discussed above). Individuals exposed to predator cues during the egg stage only may have retained this acclimation until the short trial period in the larval stage.