Breeding Behaviors of the Endangered Prairie Butterfly Oarisma poweshiek (Lepidoptera: Hesperiidae) in Relation to Environmental Factors in an Ex Situ Conservation Setting

Abstract

The Poweshiek skipperling Oarisma poweshiek (Parker, 1870) (Lepidoptera: Hesperiidae) is an endangered prairie obligate butterfly native to the north central United States and southern Canada. Conservation efforts for this species rely on ex situ approaches for population augmentation and reintroductions. As such, improving our understanding of the behaviors of Poweshiek skipperlings and maximizing their reproductive output in an ex situ setting are critical for the success of associated conservation initiatives. In this study, we examined the frequency of breeding behaviors exhibited by Poweshiek skipperlings in relation to various environmental factors: sunlight intensity (measured in lux), ambient temperature, and age. Sunlight intensity was a significant predictor of breeding behavior frequency, but we did not detect an effect of ambient temperature on breeding behavior. We also documented a generally negative relationship between age and breeding behavior frequency for both sexes. The results of our study underscore the importance of natural sunlight in encouraging breeding behaviors in an ex situ conservation environment. Ex situ observations also can help identify environmental conditions that promote high levels of Poweshiek skipperling activity, which could be used to optimize the timing of in situ population surveys.

1. Introduction

Conservation-centered managed rearing and breeding (ex situ) programs are increasingly recognized as key tools to support the recovery of Lepidopterans in the USA [1,2,3]. For instance, population reinforcement and reintroduction programs to help stabilize depleted populations and reestablish extirpated sites are derived from ex situ populations and have been identified as a need to achieve recovery of the globally Critically Endangered [4] Poweshiek skipperling (Oarisma poweshiek) (Parker, 1870) (Lepidoptera: Hesperiidae) [5,6,7]. This prairie obligate butterfly, native to the north central United States and southern Canada, has experienced rapid declines over the past several decades, and is extirpated from more than 99% of its known historic locations [8,9]. It is listed as Endangered in the United States [10] and Canada [11].

Given that ex situ conservation tools are central to ongoing restoration efforts benefitting Poweshiek skipperlings, it is important to understand their reproductive behaviors in a captive setting to maximize reproductive potential. We can also leverage these ex situ actions to expand our species-specific knowledge when such details are lacking and otherwise difficult to study in situ for endangered species [12]. For instance, sunlight plays an important role in thermoregulation and vision for butterflies. Butterflies absorb heat from solar radiation to increase their thoracic temperature above the threshold for autonomous flight [13,14,15,16]. Flight frequency, in turn, can affect both mating frequency and host finding behavior, therefore increasing reproductive success [15,17,18]. Autonomous flight and vision also can influence reproductive success indirectly. A previous study conducted with European skippers Thymelicus lineola (Ochsenheimer, 1808) (Lepidoptera: Hesperiidae) observed that males always flew and never walked to perched females, likely relying on visual cues to locate mates [19]. Understanding the relationship between sunlight intensity and breeding behaviors is critical for ex situ conservation programs targeting imperiled Lepidopterans.

In this study, we examined the frequency of breeding behaviors of Poweshiek skipperlings in relation to sunlight intensity, measured in lux, and ambient temperature. Due to the importance of sunlight for butterfly behavior documented elsewhere, we hypothesized a positive relationship between sunlight intensity and breeding behavior frequency. Since butterflies absorb heat from solar radiation, not ambient air, we further hypothesized there would be weaker and non-linear associations between temperature and breeding behavior frequency, with breeding behaviors peaking at an optimal temperature and dropping off on either side of this optimal temperature. We also explored how the age of males and females affected breeding behavior frequency.

2. Materials and Methods

We leveraged the Minnesota Zoo’s large ex situ population of Poweshiek skipperlings that serves as conservation insurance and as a source for wild population augmentations and reintroductions, e.g., [7,20]. All butterflies utilized in this experiment were reared and maintained following established husbandry protocols developed by the Minnesota Zoo’s Pollinator Conservation Initiative. Activities were authorized under the U.S. Fish and Wildlife Service Native Endangered and Threatened Species Recovery Permit ES64079B and the State of Minnesota Special Permit 32847.

Between 2 July and 10 July 2023, we set up a total of 10 breeding cages outside, with one male and one female Poweshiek skipperling in each. Our sample size was constrained by the authorizing permits, breeding recommendations and a limited number of individuals that were not designated for concurrent conservation translocations into the wild. Once a pair mated, the female was placed into a separate cage for oviposition and retired from the experiment, and the male was paired with a new female, if possible. If a male died before it mated, the remaining female was paired with a new male. The fine mesh pop-up breeding cages measured 30.48 cm on each side and contained a 10.16 cm pot of prairie dropseed grass and a plastic cup holding flower cuttings as a nectar source, which were replaced daily (Figure S1). We used a HOBO model MX2202 data logger (Onset, Bourne, MA, USA) that was set up inside of a breeding cage to collect sunlight intensity data every minute. We deployed and planned to integrate ambient temperature data from a temperature probe set up inside a breeding cage, but due to equipment malfunction, we instead utilized ambient temperature data from a nearby Weather Underground station (The Weather Company, LLC, Brookhaven, GA, USA), station KMNAPPLE108, for our analyses.

Every day, we observed up to five cages at a time from approximately 9:00 to 17:00 and documented every time we observed breeding behaviors. Additionally, we collected general behavioral data every five minutes to document each individual’s behavior at that moment to develop a behavioral time budget. One of the breeding behaviors we looked for was male abdominal J’ing, which is when a male curves its abdomen towards a female and extends its claspers [19]. Additional breeding behaviors included male wing flicks, female wing vibrations when the male was near (suggesting receptivity), and actual mating. The non-breeding behaviors we observed included flying, perching, nectaring, and one observation of oviposition. Behaviors were either categorized as 1 (breeding behaviors as defined above) or 0 (other, non-breeding behaviors).

We analyzed the data using R Version 4.4.1 [21]. We counted the number of breeding behaviors in 5 min intervals, and assumed, conditional on random intercepts associated with each individual, that the number of breeding behaviors followed a Negative Binomial distribution:

$$Y_{i,t}|{\tau }_i\sim \mathrm{Negative} \mathrm{Binomial}\left({\mu }_{i,t},\theta \right)$$

$$\mathrm{l}\mathrm{o}\mathrm{g}\left({\mu }_{i,t}\right)={\beta }_0+{\beta }_1{\mathrm{Sunlight}}_{i,t}+{\beta }_1{\mathrm{Temperature}}_{i,t}+{\beta }_3{\mathrm{Temperature}}_{i,t}^2+{\tau }_i$$

$${\tau }_i\sim N\left(0,{\sigma }_{\tau }^2\right)$$

where $Y_{i,t}$ gives the number of breeding behaviors for individual i during interval t, ${\mu }_{i,t}$ and $\theta$ are mean and overdispersion parameters of the Negative Binomial distribution, respectively, ${\beta }_0, {\beta }_1, { \beta }_2$ and ${ \beta }_3$ are regression parameters relating the mean (i.e., expected count) to the mean sunlight intensity and temperature associated with each observation interval, and ${\tau }_i$ are Normally distributed random intercepts associated with each interval. We considered a non-linear effect of temperature, as we expected breeding behavior frequency to peak at an optimal temperature value and to decrease on either side of that optimal temperature. We scaled the sunlight intensity predictor prior to analysis to improve numerical stability [22]. We fit models separately for males and females, as some of the breeding behaviors were sex-specific (i.e., abdominal J’ing and wing flicks being unique to males, wing vibrations being unique to females). We used the glmmTMB package for model fitting [23,24], and we performed model diagnostic checking using the performance package [25] and the DHARMa package [26]. Additionally, we tested for temporal autocorrelation in the residuals using the DHARMa package [26]. We created effect plots using the ggeffects package [26] to visualize relationships between the explanatory variables and the expected number of breeding behaviors. We also conducted null hypothesis tests using the ‘Anova’ function in the car package [27].

We examined associations between age and breeding behavior frequency by plotting the mean number of breeding behaviors and their 95% confidence intervals exhibited by each age group for males and females using ggplot2 [28]. We note that we did not use any day 0 males in this experiment due to previous ex situ observations of males not demonstrating any breeding behaviors on the same day as they eclosed (ER, CN pers. obs.).

3. Results

We collected data from 10 breeding cages using a total of 14 males and 11 females, resulting in 5154 behavioral observations. 66 of these were classified as breeding behaviors (

Table 1. Total number of behavioral observations documented over the course of the study.

Behavior	Number of Observations
Perched	4975
Nectaring	102
Flying	10
Ovipositing	1
Male abdominal J’ing *	8
Male wing flicks *	45
Female wing vibrations *	7
Mating *	6

* Classified as a breeding behavior

).

Residual diagnostic plots largely suggest model assumptions were reasonable and appropriate, and there was little evidence of autocorrelation in the residuals [29]. We observed a significant positive association between sunlight intensity and breeding behavior frequency for both males and females (Males: ${\chi }_1^2$ (Wald Chi-square test) = 16.77, p-value < 0.0001; Females: ${\chi }_1^2$ = 3.95, p-value = 0.047) (Figure 1). Neither males nor females reacted strongly to ambient temperature (Males: ${\chi }_1^2$ = 5.08, p-value = 0.079; Females: ${\chi }_1^2$ = 1.08, p-value = 0.58) (Figure 1).

We documented that breeding behavior frequency declined with increasing age for both sexes (Figure 2), particularly with males who exhibited more frequent breeding behaviors at young ages than females.

4. Discussion

We observed a significant positive association between sunlight intensity and breeding behavior frequency for both males and females (Figure 1). This suggests that, at least in an ex situ setting, brighter conditions are ideal for maximizing Poweshiek skipperling breeding behavior frequency. We note, however, that the data loggers used in our study measured sunlight intensity in lux, which does not measure the full range of butterfly vision. We elected to use these tools as they are easily accessible by other husbandry practitioners and can be used to conduct similar studies with other species.

Liao et al. demonstrated that an increase in sunlight intensity ultimately led to an increase in mating success, and they suggest that it may be possible to regulate butterfly reproduction by controlling light intensity [18]. From an ex situ conservation standpoint, it would be advantageous to replicate light conditions conductive to maximizing butterfly breeding behaviors. However, as noted above, natural light conditions relating to intensity, spectral composition, and the UV environment all contribute to various breeding, oviposition, and mate selection behaviors [30,31,32,33,34] and may not be replicated perfectly by artificial light. Additionally, we have previously observed that Poweshiek skipperlings have failed to mate under artificial UV lighting conditions, perhaps due to a lack of certain components present in natural sunlight [35]. With our current technology, it may be difficult to replicate the effects that natural sunlight has on butterfly behavior, but the results of this study point to the importance of providing ample natural sunlight to encourage breeding behaviors in an ex situ conservation environment to maximize reproductive success. Future studies could explore how artificial visible light conditions affect Poweshiek skipperling breeding behaviors differently than artificial UV lighting conditions.

We did not detect an association between breeding behaviors and temperature for males or females (Figure 1). However, differences between thermal conditions inside the breeding cages and thermal conditions recorded by the Weather Underground station from which we pulled temperature data may have limited our ability to detect biologically meaningful relationships. Despite the weak statistical evidence to support an association between ambient temperature and breeding behavior frequency, we observed that the skippers were perched and largely inactive during the hottest parts of the day and were most active when the temperature was less extreme. The 2023 summer in Minnesota brought especially high temperatures, and in a few instances, we needed to halt observations and bring the breeding cages into cooler, more shaded areas for the safety of the skippers. Additionally, it should be noted that our study period did not include low daytime temperatures, which may have limited our ability to discover a drop off in breeding behavior frequency at low temperature values. Our results emphasize the importance of managing heat conditions appropriately, as hotter conditions not only could yield a decrease in breeding behavior activity but could also result in unnecessary stress for the animals. In an ex situ setting, it is possible to control the temperature that skippers experience, to an extent, to avoid heat-related stress or mortality, and to encourage more activity to promote breeding behaviors and maximize reproductive potential.

Only 66 of the 5154 behavioral observations were classified as breeding behaviors (Table 1). While this reflects the natural rarity of these behaviors, such a low event frequency can reduce sensitivity to detect weaker effects, such as those related to temperature. We recommend future studies to further analyze the potential effects of ambient temperature on breeding behavior frequency of Poweshiek skipperlings in an ex situ setting. Additionally, we acknowledge that courtship activity (i.e., male abdominal J’ing) is not the same as realized reproduction (mating), but due to our limited sample size, we did not have sufficient data to assess only realized reproduction in relation to sunlight intensity and ambient temperature in a modeling framework. However, assessing all courtship activity and realized reproduction allows us to gain a better understanding of the levels of Poweshiek skipperling activity in an ex situ setting and conditions ideal for maximizing reproductive potential.

We further note that other environmental factors, such as relative humidity, may also impact breeding behavior frequency of Poweshiek skipperlings in an ex situ setting. Future work could explore environmental variables other than sunlight intensity or ambient temperature in relation to Poweshiek skipperling breeding behavior frequency.

Both males and females tended to exhibit fewer breeding behaviors with increasing age (Figure 2). The trends we observed in this study are preliminary, and future work should look at how age may affect the likelihood to breed in Poweshiek skipperlings, but our initial results suggest that pairings should target young adults (such as Day 1 males and Day 0 females) to maximize the reproductive potential of these endangered butterflies. Maximizing the reproductive potential of the globally Critically Endangered Poweshiek skipperling in ex situ conservation settings is vital to its recovery so that we can produce as many individuals as possible for sustainable insurance populations and releases back into the wild. Studying their behaviors in an ex situ environment can also shed light on how these butterflies may behave in the wild, which can inform in situ work such as ideal conditions for conducting surveys to assess the status of wild populations.