Constrained flexibility of parental cooperation limits evolutionary responses to harsh conditions

Parental care is predicted to evolve to mitigate harsh environments, thus adaptive plasticity of care may be an important response to climate change. In biparental species, fitness costs may be reduced with plasticity of behavior among partners. We investigated this prediction with the burying beetle, Nicrophorus orbicollis, by exposing them to contrasting benign and harsh thermal environments. We found strong fitness costs under the harsh environment, but rather than select for more care, visualized selection was stabilizing. Examining different components of care revealed positive directional selection gradients for direct care and strong stabilizing selection gradients for indirect care, resulting in constrained evolutionary responses. Further, because males and females did not coordinate their investments, the potential for adaptive plasticity was not enhanced under biparental care. Females cared at capacity with or without male partners, while males with partners reduced direct care but maintained indirect care levels. Decision rules were not altered in different environments, suggesting no shift from sexual conflict to cooperation. We suggest that the potential for parenting to ameliorate the effects of our climate crisis may depend on the sex-specific evolutionary drivers of parental care, and that this may be best reflected in components of care.


Introduction 24
Parental care is expected to evolve to mitigate hostile and unpredictable environments 25 (Wilson 1975). However, the extent that ecological conditions further modify parenting once it 26 evolves may depend on plasticity of parental care in response to environmental stress; that is, the 27 phenotypic variation exposed to selection. One potential source of plasticity of care where such 28 variation might be exposed is biparental cooperation. Theoretically, the default of biparental 29 systems is sexual conflict over which parent cares (Lessells 2012), which can lead to overall care 30 deficits (McNamara et al. 2003, Lessells andMcNamara 2012) and, ultimately, to one parent 31 being as effective or more effective at caring for offspring than two parents (Clutton-Brock 1991, 32 Smiseth et al. 2005, Trumbo 2006). However, the joint rearing of offspring may also allow 33 parents to breed under harsh conditions that would otherwise constrain single parent breeding 34 (Wilson 1975, Emlen 1982. This is because: 1) with more than one caregiver there is more 35 scope for increasing total care allocation (i.e., additive care; Ratnieks 1996 coincide with expansion into increasingly harsh environments (Wesolowski 1994(Wesolowski , 2004. Burying beetles (Genus: Nicrophorus) provide an ideal complement to avian systems for 55 investigating the mechanisms of cooperation and conflict over offspring care (Smiseth 2019), 56 particularly in the context of environmental stress and plasticity. First, burying beetle parental 57 care reflects their ecology. The beetles breed on an ephemeral and widely desirable resource, a 58 dead vertebrate, resulting in both rapid development and parental care involving direct 59 provisioning of predigested food and defense of the developing young Müller 1997, 60 Scott 1998a). Burying beetles are also subsocial; they do not form social associations outside of 61 additive and/or load-lightening effects), then members of the more flexible sex should also be 78 less inclined to withhold care in response to a generalized environmental stressor, which may 79 compromise the states of both parents. 80 Here, we use Nicrophorus orbicollis -a primarily biparental species and among the few 81 members of the temperate species complex to have successfully expanded into the warmer 82 climate of the US southeast (Trumbo 1990) -to examine the role that plasticity of parental 83 investment plays in mitigating harsh ambient conditions. High temperatures, as occur at low 84 latitudes, are generally implicated in more costly and less profitable reproduction in burying 85 beetles (Meierhofer et al. 1999

Study System 101
Nicrophorus orbicollis is a large-bodied, ecological generalist that breeds on small (~20 102 g) to medium (~100 g) vertebrate carcasses in North American woodlands. The species has a 103 large latitudinal distribution (from southern Canada to northern Texas), with breeding seasons at 104 the southern margins characterized by higher temperatures (3-8°C on average) and a greater 105 frequency of reproductive failure (Trumbo 1990). As with most members of the genus, parental 106 care is elaborate and extends into the post-hatching stage Müller 1997, Scott 1998a). 107 During pre-hatching stages, parents work together to clean and prepare the carcass by removing 108 hair and applying anal secretions to prevent microbial growth. During the post-hatching stage, 109 parents continue to maintain the brood ball and also directly provision to begging young via and both sexes perform the full repertoire of parenting behaviors (Scott and Traniello 1990, 116 Trumbo 1991, Scott 1998a. However, as is the case with any reproductive systems studied in 117 detail, individual investment is highly flexible and subject to environmental and social pressures 118 (Trumbo 1991, Scott 1998b, Creighton et al. 2015. 119 120

Field collection and husbandry 121
Nicrophorus orbicollis were captured from Whitehall Forest, Athens GA, in the summer 122 of 2020. Beetles were baited into hanging traps with salmon and collected twice weekly to breed 123 an outbred laboratory colony. Simultaneously, Thermochronâ iButton temperature loggers 124 (©Maxim Integrated Products, Inc., San Jose, CA, U.S.A) were deployed ~10-12 cm 125 underground at trap locations throughout our collection site to estimate the range of temperatures 126 beetles likely experience in their subterranean brood chambers. Nicrophorus orbicollis begin 127 emerging from hibernation in early spring and reach peak densities around midsummer (between 128 late June and early August; Ulyshen and Hanula 2004). In 2020, mean daily temperatures during 129 these two potential breeding windows -late spring/early summer (31 May-03 July) and mid/late 130 summer (15 July-22 August) -ranged between 21.71±1.38°C and 23.82±0.77°C, respectively 131 ( Fig S1). Diurnal temperature fluctuations were between 0.75°C and 7.71°C. To capture this 132 variation in the laboratory, we programmed two incubators to ramp between set points of diurnal 133 temperatures over the course of 10:14 hour reverse light: dark cycles, simulating early and late 134 summer breeding conditions, respectively. The first treatment, hereafter the 'benign' thermal 135 environment, was set to ramp between 19°C (night) and 20°C (day), while the second, hereafter 136 the 'harsh' thermal environment, was set to ramp between 23°C (night) and 24°C (day). Focal 137 individuals for the experiment were selected from F01 and F02 colony lines, which were bred on 138 countertops at room temperature (20±0.5°C). Larvae were divided evenly between the treatment 139 incubators on the third day of pupal development to facilitate acclimation (adults eclosed into the 140 environment in which they would ultimately breed) while controlling for possible early 141 developmental effects of temperature. All virgins selected for the experiment were at least 14 142 days of age. 143 144

Breeding trials 145
Breeding trials were carried out between October 2020 and January 2021. We used a 146 mixed factorial design as outlined in Figure 1, in which social condition (uniparental or 147 biparental) was measured as a within-subject factor and thermal environment (Benign or Harsh) 148 was measured as a between-subject factor. The goal was to achieve a balanced experimental 149 design with respect to the number of individuals undergoing repeated trials (N = 20 males and 150 females per thermal environment), which would allow us to explicitly quantify differences in 151 individual plasticity between the two thermal environments. To facilitate this, we randomized the 152 order in which focal individuals were exposed to either social condition. To create the biparental 153 condition, individuals were paired to a focal individual of the opposite sex within the same 154 thermal environment. To create the uniparental condition, individuals were paired to a random 155 unrelated beetle of the opposite sex (also within the same thermal environment) who would be 156 removed between egg laying and hatching. Individuals who successfully completed their first 157 trial would continue on to a second trial in the opposite social condition, while individuals that 158 failed their first trial within seven days of pairing were restarted. Beetles were allowed one 159 failure on their first attempt. individuals were assigned randomly to a starting social condition (uniparental or biparental) and 166 restarted in the opposite social condition upon successful completion of a first trial (social 167 condition = within-subject factor). In uniparental trials, non-focal parents were removed after egg 168 laying. All eggs were collected prior to hatching and each widowed parent or biparental pair was 169 allocated a standardized number of larvae (N=10) of random genetic origin. Behavioral and 170 performance measures were collected starting 24 hours into care. 171 172 containing a freshly thawed mouse carcass between 40 and 45 g (RodentPro, Evansville, IN, 175 USA). Boxes were returned to the incubator where they were kept on a darkened shelf beneath 176 blackout curtains to simulate an underground breeding environment. From pairing, breeding 177 boxes were checked twice daily for eggs. Pairs with no eggs after seven days were restarted on a 178 new mouse. Two days after eggs were first recorded, the brood ball and focal beetle(s) were 179 transferred to a new breeding box (non-focal parents were removed) such that eggs could be 180 collected and counted. This step was performed to facilitate brood standardization, which 181 ensured that comparisons of performance would be attributed to parental care rather than 182 differences in fertility or genetic quality. Eggs were placed in petri dishes with damp filter paper 183 and monitored every 8 hours until larvae appeared. At this stage, synchronously hatching broods 184 were randomly mixed, and each pair of fertile parents was given exactly 10 larvae. Broods that 185 failed to hatch within five days of laying were recorded as unfertilized, and the pair was 186 restarted. 187

188
Behavior and performance measures 189 Behavioral observations were carried out 24 hours after introducing larvae, as previous 190 work indicates that offspring provisioning peaks around this time (Smiseth et al. 2003). Breeding 191 boxes were placed in a dark, temperature-controlled observation room (20°C) and allowed to 192 acclimate for 30 minutes, ensuring that observed differences in parenting could not be attributed 193 solely to temperature-dependent activity. Observations took place under red light over a 30-194 minute period. Behaviors were recorded every minute via instantaneous scan sampling. These 195 included any instances of direct provisioning (i.e., mouth-to-mouth contact suggesting cavity or self-feeding to facilitate subsequent regurgitations), offspring association (i.e., in 198 physical contact with larvae but not provisioning), carrion maintenance (i.e., positioned under 199 brood ball or walking over brood ball exuding antimicrobial secretions), self-grooming, and off 200 brood ball. Individual behaviors were then grouped into two broad categories -direct care (direct 201 provisioning, oral pre-treatment of feeding substrates, and offspring association) and indirect 202 care (carrion maintenance) -to arrive at 2x budget scores (ranging between 0 and 30) for each 203 parent in each trial. "Self-grooming" and "off brood ball" were regarded as non-caring behaviors 204 and were interpreted only in the inverse. For biparental pairs, total direct and indirect care were 205 calculated by summing the time budget scores of the two parents. 206 After completing observations, brood boxes were returned to incubators and subsequently 207 checked three times per day for parental desertion. Desertion was inferred when beetles were 208 observed buried in the dirt away from the brood ball for three consecutive observations 209 recorded over the course of our experiment, our first analysis was of parental longevity. We used 222 a Cox proportional hazard regression model implemented in the R package 'survival' (Therneau 223 and Lumley 2015) to test the association between thermal environment and mortality, adjusting 224 for sex. Our second analysis was of reproductive parameters. We used two-tailed t-tests to 225 contrast the means of fecundity, fertility, and development time across all trials. To compare 226 realized performance (number and mean mass of dispersing larvae), we used simple linear 227 regression with thermal environment as a main effect and breeding history (binary specifying at 228 least one previous breeding success between parents) as a covariate, to account for variation in 229 parental experience within our study design. 230 After identifying costs associated with thermal stress, we split the dataset by thermal 231 environment and examined evidence for variation in behaviors under selection. We explored 232 how care influenced fitness through environment-specific performance gradients. We first 233 plotted the relationship between standardized offspring mass and care allocation by social 234 condition. Overall selection was visualized using total care (sum of all direct and indirect care).    For plasticity of biparental care to help mitigate increased environmental costs of parenting, 344 males and females should adjust investment strategies to complement, rather than overlap, with 345 care components of their partner. We found sex-specific tradeoffs in parenting behaviors, 346 consistent with covariance underpinning potential responses. Repeated measures ANOVA tests 347 identified social condition as an important factor explaining within-subject variation of parenting 348 behavior, but these effects were sex-specific (Table 3). Females were not plastic and appeared to 349 care at capacity regardless of thermal environment or social condition. Conversely, male 350 parenting effort varied significantly between social conditions, most dramatically in terms of 351 duration of care (Table 4). Further, males in the harsh environment significantly reduced their 352 provisioning effort when paired with a female but maintained the same levels of indirect care 353 (Table 4)  In this study, we investigated the potential of plasticity of biparental care to ameliorate a 373 harsh environment in a burying beetle, Nicrophorus orbicollis. Our prediction was that offspring 374 receiving more care through additive or load lightening benefits of multiple caregivers would 375 fare better under harsh environmental conditions. We tested this by exposing families with 376 different parental compositions to thermal stress and identifying behavioral correlates of 377 performance using standardized selection gradients. The patterns that emerged were opposite to but the type of care was important, and components were not independent of each other. 380 Plasticity of biparental care did not help overcome constraints because decision rules for 381 investment were sex-specific and were unaltered by generalized stress on the family. The 382 consequence was that in biparental pairs under environmental stress, females overinvested, and 383 males contributed only to care types that caused a decrease in fitness in excess. These results 384 challenge our understanding of the adaptive role of biparental care in hostile environments. 385 The thermal stress we imposed had strong deleterious fitness effects compared to a more 386 benign temperature. Not only did adults acclimated to the warmer environment suffer reduced 387 lifespans and lower reproductive potential per bout, but offspring were also less likely to survive 388 to dispersal and attained lower body mass than counterparts in the benign environment. confronted with more extreme environments (Wilson 1975, Wesolowski 1994, 2004. Burying 397 beetles are known for the level of plasticity they show in parental care, especially in response to 398 social condition, with uniparental female, uniparental male, and biparental care possible for 399 many species including N. orbicollis (Trumbo 1991, Scott 1998a, Smiseth and Moore 2004 to this expectation, our high-stress environment did not induce strong and consistent directional 403 selection relative to the benign environment. Instead, overall care was associated with significant 404 nonlinear effects -an indication of strong stabilizing selection (Schluter 1988). This translated to 405 fewer and smaller offspring among caregivers with both the lowest and the highest cumulative 406 behavioral investments (Fig 4B). We detected no improvements in performance among families 407 with two caregivers as opposed to one (Fig 4B). In fact, because two caregivers are effectively 408 capable of twice the sum total effort, biparental pairs accounted for much of the performance 409 reduction in the upper tails of the care distribution. These results are consistent with independent 410 investigations carried out in Oregon (Feldman 2020) and Canada (Ong 2019), which report 411 significantly reduced performance and limited compensation among biparental pairs exposed to 412 experimental warming treatments. Our study provides a mechanism for these effects: reduced 413 offspring performance at higher temperatures does not result from biparental care per se, but 414 from temperature-dependent thresholds in optimal care allocation, which are most likely to be 415 exceeded when two parents are active at the nest. 416 Our original prediction -that parents should compensate for environmental stress by 417 increasing the amount of time they spend caring -was based on an assumption of unconstrained 418 potential to evolve more care to offspring.  Parker et al. 2015), at least in terms of a lack of a response by 456 males to offspring need. In burying beetles, the prevailing theory for why males of some species 457 join females is that mating opportunities are limited elsewhere and pairs fare better than single 458 parents in defense of the valuable but temporally and spatially unpredictable resources (Scott 459 1990, 1998a, Trumbo 1991). If true, then the impetus for the evolution of care may be 460 different for males and females. While females should focus efforts on meeting the needs of 461 offspring, males should remain impervious to offspring needs except in the extreme case that the 462 female dies or abandons the nest. Thus, males adopt an 'insurance policy' strategy for 463 participation of care (Parker et al. 2015). This is supported by evidence in a variety of burying 464 beetle species for compensatory responses to partner removal (Smiseth et al. 2005 The prediction that cooperative parental strategies enhance resilience in harsh or hostile 472 environments is not novel (Wilson 1975, Emlen 1982), but climate change has afforded new 473 urgency to understanding its practical significance (Lucey et al. 2015, Manfredini et al. 2019 Henriques and Osmond 2020). Our study has shown that burying beetles at southern range 475 margins will face steep reproductive challenges associated with rising temperatures alone, and 476 that these will not be alleviated through biparental cooperation. Despite the predominance of 477 biparental social structures in this species strategies for coordinated care are unrefined. The 478 implication of our work is that the potential for parenting to ameliorate the effects of climate 479 change is likely to depend on the evolutionary drivers of parental care, which may be sex 480 specific and be best reflected in components of care.