Discussion
We found strong support for Circadian Rhythm and Thermal Acclimation Hypotheses, indicating that colonies of Ectatomma ruidumdynamically manage constitutive and induced mechanisms of thermal tolerance. When we altered the thermal regime experienced by these ant colonies, the thermal persistence of workers responded accordingly, consistent with the Thermal Acclimation Hypothesis. Simultaneously, colonies demonstrated overt diel shifts in thermal persistence in the absence of external temporal cues, supporting the Circadian Rhythm Hypothesis. While we found some differences in thermal performance associated with behavioral roles, the Division of Labor Hypothesis could not account for differences in thermal persistence based on time of day, even though that was the prior finding that led us to conduct the present study (Esch et al., 2017; Nelson et al., 2018). The authors of the earlier studies wondered how colonies of E. ruidum managed the thermal tolerance of foragers associated with diel changes in temperature. Our answer to that question is that a constitutive division of labor is not as important of a mechanism as circadian rhythms and proximate acclimation to recent thermal exposure.
We were surprised by the small magnitude of the findings. The statistical significance of the results is robust, so we have little doubt that these are real biological phenomena. Nevertheless, the differences among sample populations were not so great. This is consistent with the hypothesis that hotter and less seasonal environments result in narrow thermal limit ranges (Kaspari et al., 2015). It does not appear that the changes are driven by a number of outlying individuals with extreme differences, because in Figures 1 and 2, it appears that the entire distributions are shifted. We think that the biological significance of our findings is not that these tropical ants are responding to their thermal environment with massive physiological changes, but rather that these ants have the capacity to demonstrate flexibility in a relatively complex manner. We did not press hard on thermal levers to exert stresses on these ants. The marking experiment simply measured existing differences among ants occurring in the wild, and the colony fragments in the circadian rhythm experiment were only held in the laboratory for up to four days. The only treatment with a more-than-slight thermal stress (cooling a tropical rain forest species to a constantly air-conditioned environment) produced an effect of substantial magnitude. We expect that subjecting these ants to greater thermal extremes will result in more marked effects.
The discovery of a circadian rhythm appears to be a relatively novel result. Thermal tolerance is known to follow circadian rhythms in some prokaryotes and plants (Rensing & Monnerjahn, 1996), but to our knowledge has not yet been documented in any insects. Because some animals regulate heat shock proteins following the seasons (Arad, Mizrahi, Goldenberg, & Heller, 2010; Bujan, Roeder, Yanoviak, & Kaspari, 2020), it makes sense that tropical rain forest insects are also capable of recapitulating this activity over a 24 hour period, =he time scale during which they experience the greatest range of thermal challenges (Janzen, 1967).
In all treatments, thermal persistence was greater at dawn (the daily thermal minima) than the early afternoon (the daily thermal maxima). We were primed to expect the opposite result, because Esch et al. (2017) and Nelson et al. (2018) reported that ants foraging in the heat of the afternoon demonstrated higher thermal tolerances. In the circadian rhythm experiment, our controls were foragers freshly sampled from the field, which constitutes a near-replication of those earlier studies with a greater sample size, and we found a non-significant result in the opposite direction. After having had an opportunity to digest the results from the treatments in the Circadian Rhythm experiment, we have landed on a working hypothesis. The well documented mechanism of thermal tolerance for these ants is the production of heat shock proteins, which are both energetically expensive and are consumed as they are used to prevent tissue damage from heat (Feder & Hofmann, 1999; Moseley, 1997). The production of heat shock proteins must happen prior to the moment of heat exposure, so their production must be made in anticipation of future need. Ants living in a tropical rain forest, including E. ruidum , live with absolute certainty that evenings will bring cool temperatures and the following day will bring heat. Based on our findings, we expect that these ants upregulated heat shock proteins at the start of the day, which effectively anticipates the thermal challenges of a new day. While they know that it will get hot, the extent of the heat and the timing of it throughout the day are not as predictable (Sanford Jr, Paaby, Luvall, & Phillips, 1994), so acclimation matters too. This working hypothesis is consistent with our finding that ant colonies subjected to a constant cool temperature have extremely low thermal persistence in the afternoon. We interpret this to mean that level of heat shock protein production, while governed in part by a circadian rhythm, can be downregulated once colonies experience consistently low temperatures. To gain a greater understanding of how this circadian rhythm works, further work using more extreme thermal manipulations and more frequent time steps will flesh out the nature of this daily rhythm, as well as work to measure heat shock protein and heat shock protein gene expression. Other ant species have already been shown to upregulate heat shock protein gene expression in response to heat exposure (Nguyen, Gotelli, & Cahan, 2016), so this is a tenable future line of investigation.
While our finding of higher thermal tolerance in the early morning contradicts earlier published results (Esch et al., 2017; Nelson et al., 2018), we believe this makes sense in light of our other findings. While we conducted our project in the same field site as Nelson et al. (and even used the same heating/cooling device), both earlier papers were conducted over a single 24 hour period during the dry season (which has cooler temperatures), while the present study was conducted over a longer duration in the wet season (which is characteristically hotter). Moreover, the low-temperature sampling by Esch et al. (2017) happened several hours before the thermal maxima, and the sampling by Nelson et al. (2018) was between 04:00 and 04:30, whereas our sampling occurred 04:30-06:30. Our slightly later sampling was marginally closer to the thermal minima, but not to an extent that we think the temperature difference would matter. Nonetheless, if our working hypothesis that heat shock protein production is upregulated at the start of the day is correct, then this time difference could possibly explain why our early morning ants demonstrated greater thermal tolerance.
We found that foragers have greater thermal persistence than non-foragers. However, is not the essential prediction of our Division of Labor hypothesis, which was that there would be differences thermal persistence between foragers that ventured into the heat of the afternoon and those that foraged in the cool of the early morning. The difference between foragers and non-foragers is more parsimoniously explained by Thermal Acclimation. In the marking experiment, we noticed that the mean difference in thermal persistence between foragers and non-foragers was three minutes, which is same magnitude of difference in the thermal minima assays for the circadian rhythm experiment, between freshly caught control foragers and laboratory colony ants. This suggests that both the thermal treatment on colonies as well as the artifacts of the laboratory manipulation did not alter thermal persistence in the thermal minima time step, and that the effect of the treatment was contingent on the circadian rhythm.
While our experiment was not designed with the purpose of assessing intercolonial differences in thermal persistence, these comparisons were possible with our sampling design. We separately compared the thermal tolerances of colonies from the two experiments, and found that workers from 3 of 21 colonies demonstrated significantly greater thermal persistence. By random chance, 1 out of 20 colonies is expected to generate a p-value below 0.05, and we don’t think that adding just 2 more colonies constitutes compelling evidence, as it looks more like a few lucky rolls on a 20-sided die. We selected our colonies intentionally so that they were in as similar environments to one another as possible (with respect to light exposure in particular), because this was a lurking variable that we sought to control. That we detected any differences between colonies even though we sought to prevent them from occurring suggests there is greater ambient variation, especially considering the abundance and extremely broad geographic range of this species. Perhaps if we intentionally sampled for thermal heterogeneity in the field and subjected colonies to thermal treatments, this would not only affect colony movement behavior as found in an earlier study (McGlynn et al., 2010), but also shifts in the thermal persistence of colony members.
With respect to climate change, one central concern is the extent of behavioral flexibility, ecological plasticity, and evolutionary lability available to animals facing new thermal challenges. Heat genuinely prevents ants from foraging, as ants will avoid resource-rich areas if they cannot withstand the thermal stresses (Spicer et al., 2017). Thermal limits are a major factor in structuring species distributions across broad environmental gradients (Diamond et al., 2012). Ants also adapt their thermal tolerance to local conditions (Bujan & Kaspari, 2017; Diamond, Chick, Perez, Strickler, & Martin, 2017; Villalta et al., 2020); we have known this ever since Mary Talbot invented equipment and protocols to measure critical thermal maxima (Talbot, 1943). In temperate environments, it appears that the ability to withstand cold may be more predictive of distributions than the ability to withstand heat (Bishop, Robertson, Van Rensburg, & Parr, 2017; Bujan, Roeder, de Beurs, Weiser, & Kaspari, 2020). In the tropical climates where the most species occur, however, we have many open questions about how biodiversity will respond to rising temperatures (Jenkins et al., 2011). We argue that an integrative understanding of organisms, including the functional ecology at the colony level, will be critical for developing more informative models as the world continues to heat up.
Animals living in social groups are capable of leveraging the modularity of their colonies to efficiently organize behavioral responses to thermal challenges. While this is well known from other social insects, particularly honey bees (for example, Cook & Breed, 2013; Cook et al., 2016; Kaspar et al., 2018; Stabentheiner, Kovac, & Brodschneider, 2010), we think that a greater research investment into intracolonial mechanisms of thermal tolerance in ants is necessary to understand the distribution of thermal tolerance at higher levels of organization. With the projected levels of warming that we will experience over at least the next few generations (IPCC, 2014), and the criticality of ants for ecosystem processes (Del Toro, Ribbons, & Pelini, 2012), we need to understand colony-level processes of thermal management appears foundation to conservation planning for these animals.