Shaped by the Sun: the effect of exposure to sunlight on the evolution of spider bodies

Abstract

Body temperature can strongly influence fitness. Some Sun-exposed ectotherms thermoregulate by adjusting body posture according to the Sun's position. In these species, body elongation should reduce the risk of heat stress by allowing the exposure of a smaller body area to sunlight. Therefore, selection should favour more elongated bodies in Sun-exposed than in Sun-protected species. Diurnal orb-web spider species that sit on their webs are more likely to be Sun-exposed, on average, than nocturnal or diurnal shelter-building species. We measured the body elongation of orb-web spiders (Araneae, Araneidae) across 1024 species and classified them as Sun-protected or exposed based on the literature. We found that Sun-exposed species evolved more elongate bodies than Sun-protected ones. Further, we built a model combining traditional heat transfer models with models of thermoregulatory postures in orb-web spiders and meteorological data. The model indicates that body elongation in large orb-web spiders decreases the risk of high body temperatures. Overall, our results suggest that Sun exposure influenced the evolution of body shapes of orb-web spiders.

1. Introduction

Internal body temperature can have a major impact on an individual's fitness [1–3]. Primarily, body temperature is directly associated with the metabolic rate by affecting the rate of biochemical reactions [2,3]. As such, it impacts the rate of development [1,4–6] and critical behavioural activities [3,7–9]. In ectotherms, low body temperatures result in slow growth, foraging and egg production, ultimately reducing fitness [3,10,11]. High body temperatures may reach threshold values at which most of these activities are compromised, possibly leading to death [3,7,11–13]. Therefore, maintenance of body temperature within optimal boundaries was likely a major force in the evolution of metazoans.

Sun-exposed species are particularly prone to heat stress [14–16]. Not surprisingly, several species behaviourally reduce exposure to the Sun by adopting nocturnal activity, burrowing and resting in shaded areas [9,15,17,18]. However, these behaviours may entail opportunity costs, such as reduced foraging [9,17]. In addition, body size also affects thermoregulation efficiency. Smaller bodies have a larger surface-to-volume ratio, which causes a higher rate of heat transfer via convection than larger bodies and, consequently, lower equilibrium temperatures [14,16]. However, larger body sizes generally have individual selective advantages, such as increased survival and fecundity [19–25]. In arthropods, in particular, larger bodies are closely linked to fecundity in females [20–24,26].

These costs and benefits of body size may lead to alternative body shapes. For Sun-exposed species, the body shape may impact the heat transfer rate via convection and, more importantly, heat transfer via solar radiation [14,16]. Several animals, such as grasshoppers, dragonflies and spiders, adjust their body orientation according to the Sun's position [17,27–31]. A cylindrical body shape in these animals allow behavioural thermoregulatory postures that expose different amounts of body surfaces to sunlight [29,32] (figure 1a,b).

Figure 1.

(a) Araneus horizonte (left), a Sun-protected species with a round abdomen (roundness = 0.99), and Trichonephila clavipes (right), a Sun-exposed species with body elongation (roundness = 0.37). (b) As the Sun moves in the sky, orb-weavers exhibit thermoregulatory postures by pointing the abdominal apex towards the Sun. (c) We modelled spider bodies as cylinders with varying sizes, shapes and angles of exposure to sunlight. At 90°, cylinders are minimally exposed to direct sunlight, but an elongated cylinder (roundness = 0.25) has a smaller exposed area than a non-elongated cylinder (roundness = 1.00) of the same volume. The model assumes an air temperature of 30°C, wind speed of 0.1 m s⁻¹ and 1000 W m⁻² of solar irradiance. (d) Model estimates of the equilibrium temperature of a spider pointing its abdomen to the Sun at an angle of minimum exposure for a given body shape (roundness) and size (volume; x-axis in log-scale) of Sun-protected and Sun-exposed species. Photos by P.C.M. (A. horizonte) and N. G. Ximenes (T. clavipes).

Orb-web spiders are a diverse and widespread group of sit-and-wait predators that vary in exposure to the Sun (figure 1a). Nocturnal or diurnal shelter-building species avoid direct Sun exposure by resting on the web centre only at night or hiding in leaf retreats and shaded web margins [33–35]. By contrast, diurnal species that sit at the centre of their webs during the day are more likely to be Sun-exposed. Additionally, several diurnal species exhibit thermoregulatory postures by pointing their abdominal apex towards the heat source [28–31] (figure 1b). Experimental data confirm that these animals reach lower equilibrium temperatures than exposing a larger body area to sunlight [31].

Hence, we expect to find a pattern of more elongated bodies in larger Sun-exposed species as they are, on average, more likely to suffer from thermal stress when compared with Sun-protected ones. An elongate body should facilitate thermoregulation and reduce the chance of heat stress which would not be possible with a spherical shape [16,36]. We evaluated whether the evolution of body elongation in Neotropical orb-web spiders of the Araneidae family is explained by the likelihood of Sun exposure. Further, we built a heat transfer model to estimate the effect of body size and elongation on the body temperature of Sun-exposed and Sun-protected spiders.

2. Material and methods

(a) . Orb-web spider data

We estimated body elongation and area of spiders by measuring dorsal images extracted from taxonomical articles on Neotropical species (electronic supplementary material, table S1). These images were processed using ImageJ software [37] (v. 1.8.0). We drew over the abdomen with the polygon feature and recorded two outputs: total area and roundness [38]. Roundness may vary between 0 and 1, with the latter being a perfectly round shape (equal minor and major axes; electronic supplementary material, equation S6). Other typical spider size metrics (e.g. cephalothorax width) do not consider the shape variation and are, therefore, not suited to our analyses.

We classified spider genera as either Sun-exposed or Sun-protected primarily based on the scientific literature (electronic supplementary material, table S2). We supplemented this information with specimen data from arachnological collections (DZUB at Universidade de Brasília, Brasília, Brazil and Universidade Federal de Minas Gerais, Belo Horizonte, Brazil), and our personal knowledge of these species. We classified diurnal spiders that sit on the web hub during the day as Sun-exposed and nocturnal and diurnal spiders that hide from the Sun (by shelter construction or waiting on vegetation) as Sun-protected (electronic supplementary material, table S2). Other variables such as habitat structure or building webs in the shade may cause variation in Sun exposure. However, our assumption is that diurnal species that sit on their webs are more likely to be Sun-exposed, on average, than a nocturnal or diurnal shelter-building species. Our genus-level classification may have missed some intra-genus variations, but detailed data on the species-level are scarce, and the available information suggests little intra-genus variation. We did not include in the analysis species for which there was no dorsal illustration or information on activity or behaviour.

We used a Bayesian multilevel approach to analyse the effects of Sun exposure and body size on body elongation. Roundness (square-root arcsine transformed) entered as the response variable. Sun exposure (Sun-protected or Sun-exposed), abdominal area (log-transformed then centred and scaled by 1 s.d.) and their interaction entered as predictor variables (there was no evidence of covariance between these variables; electronic supplementary material, figure S4). For the taxonomic model, we included the spider taxonomic genus as a group-level factor. For the phylogenetic model, the matrix of phylogenetic distances entered as a group-level factor. For that, we used a published time-calibrated phylogenetic tree of araneid spiders [39]. We pruned non-Neotropical genera and used the phylogeny as a genus-level backbone tree. We then added species to the tree into their genera as zero-length branches. The phylogenetic tree presented polyphyletic genera (Araneus, Eriophora, Larinia and Parawixia). For Eriophora, Larinia and Parawixia genera, we added species to the clades already containing Neotropical species in the original tree. For Araneus, we placed species into the clade containing Neartic Araneus species, which are morphologically similar to Neotropical species. The taxonomic model included 1024 species in 49 genera. For the phylogenetic model, we excluded data from the species whose genera were not present in the phylogenetic tree (70 species in 20 genera). We used a Student's t distribution in the taxonomic statistical model to reduce the influence of extreme values and a Gaussian distribution in the phylogenetic model.

(b) . Heat transfer model and meteorological data

We built a heat transfer model to estimate the equilibrium temperature of animals of varying elongations (roundness), volumes and angles of exposure to sunlight (figure 1c). We coupled classical models on heat exchange of ectotherms [14,16,40–43] with a model on thermoregulatory postures of orb-web spiders [32,44], plus a body elongation parameter. A detailed description of the model is provided in the electronic supplementary material. We validated the model estimates using published experimental data (electronic supplementary material, figures S1–S3).

Orb-web spiders sit on their suspended webs. Therefore, heat exchange via conduction is likely to be small. Similarly, in small invertebrates, heat generation due to the metabolic rate and heat loss via evaporation are orders of magnitude lower than heat exchange via other sources [14–16]. Therefore, the equilibrium temperature of an orb-web spider can be obtained by solving the equation:

$$W_c+W_{\mathrm{lw}}+W_{\mathrm{sun}}=0.$$

The equilibrium temperature depends on the solar radiation reaching the animal's body (W_sun), the convective heat transfer between the air and body (W_c) and heat transfer via longwave radiation (W_lw; thermal radiation) (figure 1c). Depending on the angle, the body may expose a larger or smaller surface area to direct sunlight (figure 1c). We modelled the animal body as a cylinder and parametrized our model based on Neotropical orb-web spider body sizes and shapes (see below and electronic supplementary material).

Further, we estimated the number of days per year Sun-exposed orb-web spiders would reach greater than or equal to 35°C and greater than or equal to 40°C. The former is usually above the preferred temperature of spiders, and the latter circa the critical maximum temperature of many spiders [13,28]. Only two out of 16 spiders species showed preferred temperatures greater than or equal to 35°C, and eight out of nine species exhibited heat stupor or leg jerking at body temperatures ranging from 39°C to 43°C (see [13] for the review). We applied the heat transfer model to 10 years of meteorological data (2011–2020) collected hourly (wind speed, solar radiation and air temperature) in automatic stations (−34° to 0° latitude and −36° to −73° longitude). We excluded years with less than 200 sampled days per year and stations that did not have data for at least 8 years. To avoid spatial autocorrelation, we excluded stations less than 4° distant from each other. We checked the spatial autocorrelation by visualizing variograms and comparing AIC values of linear models with and without a correlation structure (corExp and corGaus from the nlme package [45] v. 3.1-152 in R). After exclusions, we gathered data from 34 stations.

We assumed spiders were Sun-exposed and pointing their smaller surface area towards the Sun. We estimated the hourly body temperature for the largest (volume = 2500 mm³), median (volume = 24 mm³) and the smallest spiders (volume = 0.5 mm³), with an elongate (roundness = 0.25) and non-elongate body (roundness = 1.0). These volumes are based on the actual abdominal areas we measured (see above; we converted the area to volume assuming the abdomen was a sphere). Then, we calculated the maximum spider body temperature per day per station per year with the station hourly wind speed, solar radiation and air temperature data.

We analysed data in two Bayesian multilevel models [46]. The proportion of days per year in which spider temperatures reached greater than or equal to 40°C or greater than or equal to 35°C entered as the response variable. Our response variables were proportions and had zeros. Therefore, we used a zero-inflated beta distribution. Spider size, shape and interaction entered as the predictor variables to model both the proportion and the zero-inflation. We included station ID as a group-level factor. We multiplied the back-transformed model estimates by 365 to get the estimated number of days per year.

All Bayesian models were run in R [47] (v. 4.0.2) using the package brms [48] (v. 2.13.3 and 2.13.5), which implements Bayesian models in Stan [49]. We ran four independent chains for 25 000 iterations, 5000 as a warm-up for the meteorological model and 5000 iterations, 2500 as a warm-up, for the spider data models. We set weakly informative priors scaled to our data (electronic supplementary material, table S3). We evaluated model fits by checking chain convergence, the presence of transitions with diverging errors, visual posterior predictive checks and leave-one-out cross-validation.

3. Results

Data on Neotropical orb-web spider species suggest an effect of Sun exposure on body elongation (figure 2). The analyses confirmed that Sun-exposed spiders evolved more elongated bodies than Sun-protected ones (table 1). The estimated difference in roundness between Sun-protected and Sun-exposed for an average-sized spider was 0.18 (0.09–0.27; mode and 95% credible interval) for the taxonomic analysis and 0.16 (0.06–0.25) for the phylogenetic analysis.

Figure 2.

Sun-exposed orb-web spider species have more elongate abdomens (lower roundness) than Sun-protected species. Each box represents a Neotropical genus of the Araneidae family (Sun-protected, N = 551 species; Sun-exposed, N = 473 species). Traced lines denote the global median value of Sun-protected and Sun-exposed species.

Table 1.

Coefficients of the two Bayesian statistical models used to test the effect of Sun exposure and body size (area) on body shape (roundness; roundness was square-root arcsine transformed) of neotropical araneid spiders.

predictors	estimate	95% credible interval
taxonomic model
group-level
s.d. (intercept)	0.15	0.11, 0.19
population-level
intercept	0.92	0.83, 1.00
Sun exposure (Sun-protected)	0.20	0.10, 0.30
areaa	0.00	−0.02, 0.01
Sun exposure (Sun-protected): areaa	0.03	0.00, 0.05
phylogenetic model
group-level
s.d. (intercept)	0.21	0.15, 0.29
population-level
intercept	0.92	0.67, 1.17
Sun exposure (Sun-protected)	0.18	0.08, 0.28
areaa	0.00	−0.02, 0.01
Sun exposure (Sun-protected): areaa	0.03	0.00, 0.05

^aArea (mm²) was log-transformed and then centred and scaled to 1 s.d.

We found an interacting effect of body size and Sun exposure on the evolution of body elongation (table 1). There was a positive relationship between body size and roundness for Sun-protected species but not for Sun-exposed species (table 1). Consequently, we found that the larger the spider, the larger the difference in body elongation between Sun-exposed and Sun-protected spiders. For a minimally sized species, the estimated difference was 0.11 (0.01–0.22) points of roundness for the taxonomic and 0.10 (−0.01 to 0.21) for the phylogenetic model. By contrast, for a maximally sized species, the estimated difference was 0.25 (0.14–0.35) points of roundness for the taxonomic and 0.22 (0.10 to 0.34) for the phylogenetic model.

The heat transfer model indicated a large effect of body elongation and angle of exposure on the equilibrium temperature of Sun-exposed species (figure 1d; electronic supplementary material, figure S5). At an air temperature of 30°C, a large species with an elongate body that directs its smaller abdominal area towards the Sun has an equilibrium temperature of 33°C. By contrast, for a spider with the same volume but with a non-elongate body, the equilibrium temperature is 40°C (figure 1d). Conversely, for Sun-protected spiders of the same size and shape, the effect of body elongation is small (31 and 33°C; figure 1d). For a tiny Sun-exposed spider, the equilibrium temperature is 31°C for an elongate and non-elongate body (figure 1d).

The statistical analysis indicated that a large non-elongate spider pointing its smaller area towards the Sun reaches body temperature greater than or equal to 35°C for 85 d yr⁻¹ (56–126 d yr⁻¹; mode and 95% credible interval; table 2; electronic supplementary material, table S4), and greater than or equal to 40°C for 8 d yr⁻¹ (4–12 d yr⁻¹). By contrast, a large elongated spider is estimated to reach greater than or equal to 35°C for 36 d yr⁻¹ (22–57 d yr⁻¹) and greater than or equal to 40°C for 0 d yr⁻¹ (0–1 d yr⁻¹). These values are dependent on the local climate (table 2; electronic supplementary material, figure S6). For some regions, the estimated greater than or equal to 40°C d yr⁻¹ are 39 (32–46) versus 16 (13–21) for large non-elongate versus elongate spiders. The magnitude of the overall effects and the difference between body shapes is reduced for the median-sized and especially for the smallest species (table 2; electronic supplementary material, figure S6).

Table 2.

Statistical estimated number of days body temperature of Sun-exposed spiders would reach greater than or equal to 35°C and greater than or equal to 40°C, for large (abdominal volume = 2500 mm³), median (abdominal volume = 24 mm³) and small (abdominal volume = 0.5 mm³) spiders with non-elongate (roundness = 1.0) and elongate (roundness = 0.25) abdomens.

spider size	spider shape	estimated number of days per year (95% credible interval)	estimated range (d yr−1)
≥35°C
large	non-elongate	85 (56, 126)	1–324
large	elongate	36 (22, 57)	0–271
median	non-elongate	46 (27, 71)	0–288
median	elongate	29 (17, 47)	0–253
small	non-elongate	33 (20, 54)	0–264
small	elongate	26 (15, 42)	0–244
≥40°C
large	non-elongate	8 (4, 12)	0–39
large	elongate	0 (0, 1)	0–16
median	non-elongate	1 (0, 3)	0–22
median	elongate	0 (0, 1)	0–15
small	non-elongate	0 (0, 1)	0–16
small	elongate	0 (0, 0)	0–13

4. Discussion

The finding that Sun-exposed orb-web spiders evolved more elongated bodies than Sun-protected ones (figure 2 and table 1) is consistent with the heat transfer model data. The model's prediction for temperature differences between large elongated (33°C) or non-elongated (40°C) Sun-exposed species is coherent with other heat transfer models of invertebrates [41] and empirical measurements of spiders [31] and insects [14]. At 40°C, many spiders exhibit heat stupor and spasmodic leg jerking [13].

In comparison to Sun exposure, we found a small effect of body size on elongation (table 1). This is possibly attributable to other factors influencing body size evolution, such as the viability costs due to predation or parasitism [50]. In spiders particularly, body size may be influenced by season duration and developmental time length [23,24,51]. Locomotion efficiency via bridging and climbing are also affected by body size [52,53]. Furthermore, body size can be confounded with individual condition [54] as it varies due to food intake during development [55]. Additionally, the positive relationship between body size and roundness for Sun-protected species may reflect the higher costs of body elongation due to the larger cuticular surface invested for a given volume (surface-to-volume ratio). In this scenario, body elongation would trade-off with body size.

Our analyses indicate that the local climate is determinant for the risk of heat stress (table 2). There are certainly other variables that might influence Sun exposure, e.g. type of habitat and microhabitat preferences. Individuals in forest gaps, for instance, will be subjected to similar irradiance levels to those in open areas, whereas individuals in highly shaded areas will receive approximately one-third of the irradiance [56]. These variations in Sun exposure may have added unexplained variance in our analyses. Future studies on a smaller set of species to which detailed information on microhabitat preference exists (or could be measured) may give a more nuanced perspective of the findings in this study. Still, variables influencing Sun exposure should not bias results because, on average, diurnal spiders are more likely to suffer from heat stress than nocturnal or shelter-building orb-web spiders. Therefore, body elongation may expand the climate niche of orb-web spiders.

Body elongation in orb-web spiders may have other functions. For instance, alternative body shapes may provide camouflage among twigs, branches and debris [35,57,58]. Moreover, the thermoregulation efficiency of body elongation may interact with other variables such as colour [14–16,31,59] and pilosity [60].

In conclusion, diurnal orb-web spiders have, on average, more elongated bodies than nocturnal and shelter-building species. The thermoregulatory benefit of body elongation may explain this pattern. Therefore, the degree of Sun exposure may have been a significant factor in the evolution of orb-web spider body shapes.

Data accessibility

Data, code and data description are available from the Dryad Digital Repository: https://doi.org/10.5061/dryad.70rxwdbz8 [46] and in the electronic supplementary material [61].

Authors' contributions

F.M.G. and L.F.-S. conceived the idea. L.F.-S. and P.N.R. collected the data. P.C.M. provided scientific collection specimen data. L.F.-S., P.N.R. and P.C.M. performed the Sun exposure classification. F.M.G. created the model and performed the statistical analyses. L.F.-S., P.N.R., P.C.M. and F.M.G. wrote and edited the manuscript. P.C.M. and F.M.G. supervised the study. All authors approve the final version and agree to be held accountable for the work performed therein.

Competing interests

We declare we have no competing interests.

Funding

This study was supported by Fundação de Apoio à Pesquisa do Distrito Federal (FAP/DF, Brazil; grant no.: 00193-00002164/2018-5), and Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq, Brazil; grant no.: 428141/2016-1).