Breaking the cliché: sex reversal in size dimorphism and mobility in South American Allocosinae (Lycosidae) spiders

Abstract

Typically, females and males are expected to have characteristic sexual strategies and patterns of size dimorphism, but these generalizations are subject to exceptions. The occurrence of atypical cases has been related to species or populations from environments under strong physical, ecological and/or social constraints. Allocosa marindia and Allocosa senex are two coastal spiders (Lycosidae: Allocosinae) with reversal in sex roles and sexual size dimorphism. Males are larger than females, and females are the mobile sex that initiates courtship. It is unclear whether the occurrence of non-typical sexual traits in Allocosinae spiders is correlated with coastal habitats. Our aim was to study sexual size dimorphism and surface mobility in Allocosinae spiders from different habitats throughout South America. We revised specimens from scientific collections and performed 3-day samplings to collect individuals and determine nocturnal surface mobility. We analysed a total of 1071 Allocosinae adult individuals from 18 species and/or morphotypes. Our results revealed new species inhabiting coastal habitats with reversal in sexual size dimorphism and higher nocturnal surface activity in females; however, not all coastal species shared those characteristics. Future studies will focus on studying other ecological, physiological and/or phylogenetic factors that could be shaping the origin and maintenance of sex role reversal in Allocosinae.

activity patterns, sex role reversal, sex traits, sexual dimorphism, wolf spiders

INTRODUCTION

The energetic investment of each sex in reproduction is the outcome of costs associated with production and maintenance of gametes, courtship and mating effort, delivery of nuptial gifts and/or other resources related to reproduction, the occurrence of mate guarding and parental effort, among other social and environmental factors (Trivers, 1972; Parker, 1984; Gwynne, 1991; Bonduriansky, 2001; Fritzsche et al., 2021). Classically, owing to their higher investment in gametes and parental care, females are assumed to be the selective sex, whereas males compete among themselves and court potential sexual partners (Darwin, 1871; Parker, 1984; Andersson, 1994; Fritzsche et al., 2021). However, during recent decades several studies have shown that this general rule has numerous exceptions (Gwynne, 1991; Eens & Pinxten, 2000; Bonduriansky, 2001). Furthermore, the concepts of fixed female and male ‘sex roles’ are currently under debate and could be related to traditional views about gender stereotypes (Ah-King & Ahnesjö, 2013; Pollo & Kasumovic, 2022).

There are multiple examples within the animal kingdom of species in which males show similar or higher reproductive investment than females and in which sexual patterns differ markedly from those typically expected (Gwynne, 1991; Andersson, 1994; Bonduriansky, 2001; Aisenberg, 2014). In these cases, females search for potential sexual partners and initiate courtship, whereas males are selective when taking mating decisions (Gwynne, 1991; Fritzsche et al., 2021). Although these examples have been reported in taxa as diverse as birds, amphibians, fish, crustaceans, insects and spiders (Gwynne, 1991; Eens & Pinxten, 2000; Bonduriansky, 2001; Aisenberg, 2014; Fritzsche et al., 2021), there are few exhaustive studies on this topic. Several of these reports come from species or populations inhabiting environments that suffer variable and/or extreme climatic conditions and/or unpredictability in access to refuges and prey availability, among other factors (Gwynne, 1991; Karlsson et al., 1997; Queller, 1997; Lorch, 2002). These factors could shape the intensity of sexual selection on each sex, as has been shown in Mormon crickets by Gwynne (1993), Gwynne & Simmons (1990) and Hare et al. (2022) and reviewed in different animal groups by Gwynne (1991) and Eens & Pinxten (2000). Individuals adapted to these environmental conditions could require high reproductive investment to be viable, not only from the female but also from the male, in terms of mate search, courtship and mating effort, nest-site selection and construction and/or parental care, driving the reversal of behavioural patterns expected for that animal group.

Sexual dimorphism is the evolutionary response in each sex to the effects of natural and sexual selection under genetic, phylogenetic, physiological, behavioural and ecological restrictions (Shine, 1989; Blanckenhorn, 2005; Fairbairn et al., 2007; McLean et al., 2018). Sex role reversal can include the occurrence of non-expected sexual dimorphism patterns between females and males, including sexual differences in size, ornamentation, coloration or other morphological traits (Gwynne, 1991; Eens & Pinxten, 2000). Although body size differences between the sexes are well documented in vertebrates, information about invertebrates is scarce, particularly within arachnids (Abouheif & Fairbairn, 1997; McLean et al., 2018). In spiders, in general females are larger than males, in some cases reaching extreme sexual size dimorphism (Vollrath & Parker, 1992; Hormiga et al., 2000; Foellmer & Moya-Laraño, 2007; De Mas et al., 2009; McLean et al., 2018). The occurrence of sexual size dimorphism in spiders has been discussed in terms of fecundity and foraging advantages in females. In spiders, there is evidence that female body size is positively correlated with the number of eggs, hence females with larger body sizes and better body condition would have higher reproductive success (Prenter et al., 1999; Hormiga et al., 2000; Walker & Rypstra, 2002; Foellmer & Fairbairn, 2005). In males, a smaller size could provide advantages when looking for females, for avoiding predators and/or climbing female webs during the reproductive period (Vollrath & Parker, 1992; Coddington et al., 1997; Hormiga et al., 2000; Moya-Laraño et al., 2002; Foellmer & Fairbairn, 2005). Wolf spiders, in which both sexes are usually wanderers, typically show moderate sexual dimorphism, with males being ~10–20% smaller than females but with longer legs (Head, 1995; Walker & Rypstra, 2002; Framenau & Hebets, 2007; Aisenberg et al., 2010; Logunov, 2011). Studies testing the evolutionary pathways of sexual size dimorphism in spiders suggest that there have been diverse factors shaping its origin and maintenance according to the group, making it difficult to formulate global explanations (Hormiga et al., 2000; Kuntner & Elgar, 2014; McLean et al., 2018).

The wolf spider subfamily Allocosinae was proposed by Dondale (1986) and comprises 27 species (Dondale & Redner, 1983; Brescovit & Alvares, 2011; Brescovit & Taucare-Rios, 2013; Simó et al., 2017; Piacentini & Ramírez, 2019; Gonnet et al., 2021a; Laborda et al., 2022). Dondale & Redner (1983) revised the genus Allocosa from the Neartic and Central America and described several species, but our knowledge of South American representatives is scarce. In recent phylogenetic studies, the subfamily Allocosinae was recovered as a monophyletic group that comprises five genera, of which Allocosa is the most species rich, including a clade with South American coastal species (Simó et al., 2017; Piacentini & Ramírez, 2019; Gonnet et al., 2021a; Laborda et al., 2022; Laborda, 2023).

Allocosa marindia Simó et al., 2017 and Allocosa senex (Mello-Leitão, 1945) (Allocosinae, Lycosidae) are two nocturnal spiders that inhabit the sandy coasts of rivers, lakes and the seashore of the Atlantic Ocean in Southern South America (Capocasale, 1990; Simó et al., 2017). Both species show reversal of typical sex roles and sexual size dimorphism (Aisenberg et al., 2007; Aisenberg & Costa, 2008; Aisenberg, 2014). Males are larger than females (Aisenberg et al., 2007; Aisenberg & Costa, 2008), and they construct long, silk-lined burrows in the sand, where they stay for long periods waiting for female visits. Conversely, females construct small silk capsules, where they stay during the day (Aisenberg et al., 2007). They are the mobile sex, and during summer nights they search for male burrows and initiate courtship (Aisenberg et al., 2007; Aisenberg & Costa, 2008). Mating occurs inside male burrows, and both sexes are selective during sexual interactions: females on male burrow length, and males on female reproductive status and body condition (Aisenberg et al., 2007; Aisenberg & González, 2011). Only in A. senex, it was observed that rejected females could be attacked and cannibalized (Aisenberg et al., 2009; Bollatti et al., 2022). After mating, males leave their own burrows, whereas females stay inside and will oviposit there, emerging when it is time for spiderling dispersal (Postiglioni et al., 2008; Aisenberg, 2014).

Although the coastal spiders A. senex and A. marindia have been studied, little is known about size dimorphism and mobility patterns of each sex among other South American Allocosinae species. South American sandy beaches are subject to drastic changes of physical factors, such as temperature, gradients of salinity (i.e. estuarine and oceanic beaches), humidity, winds, tides and the possibility of floods, and other factors, such as reduction of their area, fragmentation, and variations in prey availability and potential refuges for animal species (Albín et al., 2017; McLachlan & Defeo, 2018; Lopes Costa et al., 2022). As mentioned above, environments with these characteristics could be associated with the origin and maintenance of sex role reversal.

Our aim was to study sexual size dimorphism and surface mobility in Allocosinae spiders from different habitats of South America. For that purpose, we compared sexual size dimorphism of South American species with information from scientific collections of other Allocosinae species from Central and North America. Based on the reports of sex role reversal dependent on environmental conditions of the habitat in a variety of species (Gwynne, 1991; Karlsson et al., 1997; Queller, 1997; Lorch, 2002), we predicted that only Allocosinae spiders inhabiting coastal habitats would show non-typical sexual traits, with larger body size in males and higher surface mobility in females. In addition, we evaluated the effects of habitat on variation in sexual dimorphism at the intraspecific level in A. senex, a species that inhabits two types of coastal habitats (fluvial coast and oceanic estuarine coast).

MATERIAL AND METHODS

Field samplings

We conducted nocturnal samplings in Uruguay (Melilla and Instituto de Investigaciones Biológicas Clemente Estable, Montevideo; Área Protegida Montes del Queguay, Paysandú), Argentina (Parque Provincial Ischigualasto, San Juan; Parque Nacional Lanín, Neuquén; Parque Nacional El Palmar, Entre Ríos), Chile (Parque Nacional Río Clarillo, Santiago Región Metropolitana) and Brazil (RPPN Pró-Mata-PUCRS, Rio Grande do Sul; Parque das Dunas Salvador, Bahia) between 2018, 2019 and 2020 (Table 1). We selected these locations based on preliminary observations by the authors (AA, AL, LBB, LNP) and on distribution and habitat data of American Allocosinae available from natural history collections. We included Allocosinae species and morphotypes from different habitats, such as fluvial and oceanic–estuarine coasts, volcanic sandy coasts, sandy valleys, sandy mountain coasts, grasslands and gardens (Table 1; Fig. 1). We categorized the habitats based on observations at each site of substrate characteristics, topography and revision of landscape maps, according to the classification of global ecosystems by Keith et al. (2022).

Table 1.

Field sampling localities of Allocosinae species from South America

Species	Locality	Coordinates		Habitat	Biome	Functional group
		Latitude	Longitude
Allocosa senex	Parque Nacional El Palmar, Entre Ríos, Argentina	31°52ʹ32.128″S	58°12ʹ31.76″W	Oceanic–estuarine–fluvial coast	Savannas and grasslands	Temperate woodlands
	Área Protegida Montes del Queguay, Paysandú, Uruguay	32°10ʹ43.4″S	57°14ʹ12.2″W	Oceanic–estuarine–fluvial coast	Savannas and grasslands	Temperate woodlands
Allocosa cf. senex	Parque das Dunas, Bahia, Brazil	12°55ʹ26.52″S	38°19ʹ46.98″W	Oceanic–estuarine–fluvial coast	Tropical–subtropical forests	Tropical/subtropical lowland rainforests
Allocosinae sp. 1 ‘Ischigualasto’	Parque Provincial Ischigualasto, San Juan, Argentina	30°10ʹ55.6″S	67°54ʹ03.7″W	Dry riverbed, sandy valley	Desert and semi-deserts	Semi-deserts steppe
Allocosinae sp. 2 ‘Ischigualasto’	Parque Provincial Ischigualasto, San Juan, Argentina	30°10ʹ55.6″S	67°54ʹ03.7″W	Dry riverbed, sandy valley	Desert and semi-deserts	Semi-deserts steppe
Allocosinae sp. 4 ‘Lanín’	Parque Nacional Lanín, Neuquén, Argentina	39°45ʹ10.4″S	71°30ʹ1.84″W	Volcanic sandy coast	Desert and semi-deserts	Cool deserts and semi-deserts
Allocosinae sp. 5 ‘Río Clarillo’	Parque Nacional Río Clarillo, Santiago Región Metropolitana, Chile	33°43ʹ47.3″S	70°28ʹ05.9″W	Sandy mountain coast	Temperate–boreal forests and woodlands	Warm temperate laurophyll forests
Allocosinae sp. 7 ‘Río Clarillo’	Parque Nacional Río Clarillo, Santiago Región Metropolitana, Chile	33°43ʹ47.3″S	70°28ʹ05.9″W	Sandy mountain coast	Temperate–boreal forests and woodlands	Warm temperate laurophyll forests
Paratrochosina amica	Parque Nacional El Palmar, Entre Ríos, Argentina	31°52ʹ3.47″S	58°12ʹ33.78″W	Grassland, garden	Savannas and grasslands	Temperate woodlands
	Instituto de Investigaciones Biológicas Clemente Estable (IIBCE), Montevideo, Uruguay	34°53ʹ15.6″S	56°08ʹ34.8″W	Grassland, garden	Savannas and grasslands	Temperate subhumid grasslands
	Melilla, Montevideo, Uruguay	34°43ʹ53″S	56°19ʹ23.4″W	Grassland, garden	Savannas and grasslands	Temperate subhumid grasslands
	Reserva Particular do Patrimonio Natural Pró-Mata-PUCRS, Rio Grande do Sul, Brazil	29°28ʹ51.21″S	50°10ʹ26.92″W	Grassland, garden	Tropical–subtropical forests	Tropical/subtropical lowland rainforests

Figure 1.

A, female of Allocosa senex (photograph: Marcelo Casacuberta). B, typical habitat of A. senex, Atlántica Beach, Rocha, Uruguay (photograph: Marcelo Casacuberta). C, male of Allocosinae sp. 4 ‘Lanin’ (photograph: Anita Aisenberg). D, Typical habitat of Allocosinae sp. 4 ‘Lanin’, Parque Nacional Lanín, Neuquén, Argentina (photograph: Anita Aisenberg). E, female of Paratrochosina amica (photograph: Marcelo Casacuberta). F, typical habitat of P. amica, Salto, Uruguay (photograph: Marcelo Casacuberta).

We sampled during late spring, summer and early autumn of the Southern Hemisphere (i.e. from November to May), the months of highest surface mobility for adults of Allocosa (Costa, 1995; Costa et al., 2006; Bidegaray-Batista et al., 2017). We started the fieldwork 1 h after sunset, according to the hours of activity reported for A. senex and A. marindia (Costa et al., 2006; Aisenberg, 2014). On three consecutive nights and for 2 h, three researchers using headlamps collected manually all adult Allocosinae spiders found walking or leaning out from burrow entrances (hereafter referred as ‘active individuals’). We sexed the adult individuals by observing the external genitalia with a portable dissecting microscope and recorded the proportion of active individuals of each sex. All the specimens were fixed in 95% ethanol. We deposited representatives of each Allocosinae species and morphotypes in the scientific collections of Museo Argentino de Ciencias Naturales ‘Bernardino Rivadavia’ (MACN; Argentina), Laboratorio de Biología Reproductiva y Evolución, Instituto de Diversidad y Ecología Animal (LABRE-Ar; Córdoba, Argentina), Facultad de Ciencias, Universidad de la República (FCE; Uruguay) and Museu de Ciências e Tecnologia da Pontifícia Universidade Católica do Rio Grande do Sul (MCTP; Brazil).

Identification of species and morphotypes

We examined the Allocosinae specimens collected during the fieldwork and those in the arachnological collections of California Academy of Sciences (USA), Museo Argentino de Ciencias Naturales (Argentina), Facultad de Ciencias (Uruguay) and Institute Butantan (Brazil) (Table 2). We used stereomicroscopy (FCE: Nikon D3000 attached to a microscope Nikon YS100 and Nikon SMZ 10 stereomicroscope; IIBCE: Olympus SZ61 stereomicroscope; MACN: Leica 165 and Leica A205 stereomicroscopes) and scanning electron microscopy (FCE: Jeol JSM-5900; MACN: FEI XL 30 TMP) for examination of the individuals. The categorization of habitats of specimens was based on observations obtained during our field samplings (Table 1) or on identification labels from collections when available. Old specimens from collections frequently did not include geolocation records. In cases in which data labels from collections did not include this information, we based the characterization on descriptions by Dondale & Redner (1983).

Table 2.

Allocosinae specimens from scientific collections, with details of localities provided in the identification labels

Species	Scientific collection	Localities	Habitat
Allocosa brasiliensis	Instituto Butantan, São Paulo, Brazil	BRAZIL: São Paulo	Oceanic–estuarine–fluvial coast
Allocosa funerea	California Academy of Sciences, CA, USA	USA: Arkansas, Colorado, Kansas, Massachusetts, Missouri, New Hampshire, Oklahoma, Texas	Grassland, garden
Allocosa furtiva	California Academy of Sciences, CA, USA	USA: Massachusetts	Oceanic–estuarine–fluvial coast
Allocosa marindia	Facultad de Ciencias, Montevideo, Uruguay	URUGUAY: Canelones	Oceanic–estuarine–fluvial coast
Allocosa noctuabunda	California Academy of Sciences, CA, USA	USA: Arizona, Arkansas, Colorado, Missouri	Fluvial coast
Allocosa panamena	California Academy of Sciences, CA, USA	COSTA RICA: Cartago, Heredia, Tarrialba	Rainforest
Allocosa parva	California Academy of Sciences, CA, USA	USA: California, Colorado,	Grassland, garden
Allocosa senex	Museo Argentino de Ciencias Naturales, Buenos Aires, Argentina; Instituto Butantan, São Paulo, Brazil; California Academy of Sciences, CA, USA	ARGENTINA: Neuquén; Río Negro; Entre Ríos URUGUAY: Colonia; Paysandú; Maldonado; Rocha BRAZIL: Rio Grande do Sul; Paraná; Bahía; São Paulo; Río de Janeiro; Espírito Santo; Santa Catarina	Oceanic–estuarine–fluvial coast
Allocosinae sp. 3 ‘Coquimbo’	Museo Argentino de Ciencias Naturales, Buenos Aires, Argentina	CHILE: Región de Coquimbo: Canela	Oceanic–estuarine–fluvial coast
Allocosinae sp. 4 ‘Lanín’	Museo Argentino de Ciencias Naturales, Buenos Aires, Argentina	ARGENTINA: Neuquén	Volcanic sandy coast
Allocosinae sp. 6 ‘Bahia’	Instituto Butantan, São Paulo, Brazil	BRAZIL: Alagoas; Bahía	Oceanic–estuarine–fluvial coast
Gnatholycosa spinipalpis	Museo Argentino de Ciencias Naturales, Buenos Aires, Argentina	ARGENTINA: San Juan	Sandy mountain coast
Paratrochosina amica	Museo Argentino de Ciencias Naturales, Buenos Aires, Argentina; Facultad de Ciencias, Montevideo, Uruguay	ARGENTINA: Buenos Aires; Ciudad Autónoma de Buenos Aires; Entre Ríos URUGUAY: Montevideo; Paysandú; BRAZIL: Río Grande do Sul	Grassland, garden
Allocosa cf. senex	Instituto Butantan, São Paulo, Brazil	BRAZIL:Bahía	Oceanic–estuarine–fluvial coast

The information about habitat corresponds to all the places where the species can be found. Collection codes are provided in the Supporting Information (Table S1).

Measurement of morphological characters

We measured the carapace width of all Allocosinae adults collected during fieldwork and available from scientific collections. Carapace width is a measure of body size in spiders (Eberhard et al., 1998; Moya-Laraño & Cabeza, 2003; Foellmer & Moya-Laraño, 2007). We did not include the measurements of Allocosinae sp. 5 (Río Clarillo, Chile) because we had only one male.

Sexual dimorphism index

We calculated sexual size dimorphism as the ratio between the average carapace width of the male and the average carapace width of the female for each species. We estimated the sexual dimorphism index (SDI) as the sexual size dimorphism minus one, for comparative analyses (Lovich & Gibbons, 1992; Foellmer & Moya-Laraño, 2007; Kuntner & Cheng, 2016). Positive SDIs indicate a larger carapace in males than in females (reversed sexual size dimorphism), whereas negative values show a larger carapace in females (‘typical’ sexual size dimorphism expected in spiders) (Lovich & Gibbons, 1992). We compared the SDIs between species and habitats qualitatively according to the following criteria: 0, no dimorphism; 0–0.5, slight male-biased dimorphism; 0.5–1, strong male-biased dimorphism; 0 to −0.5, slight female-biased dimorphism; and −0.5 to −1, strong female-biased dimorphism. We followed the comparative methodology for SDIs according to Prenter et al. (1999), Cheng & Kuntner (2014), Turk et al. (2018) and Walker & Holwell (2022).

Statistical analyses

We explored the variability in sexual size dimorphism among species and its relationship to the habitat by means of generalized linear models. We set the carapace width of each individual as the response variable, with the sex, species and interaction between them as fixed effects. In a separate model, we ran the same response variable, but considering the habitat of the species as a fixed effect. Additionally, for A. senex we compared the same variables between individuals from fluvial and oceanic–estuarine coasts.

We compared the surface mobility (probability of finding active individuals on the surface) of the sexes by coding as a binomial response variable the occurrence in the field of active females as one and of active males as zero. These models included the species and habitat as predictor variables in separate models. These analyses were carried out considering only the individuals collected during the fieldwork of this study (Table 1; Supporting Information, Table S1).

All statistical analyses were conducted in the R environment v.4.1.2 (R Core Team, 2021). In all models, normality and homogeneity of variances were assessed graphically using the ‘fitdist’ function and the Cullen and Frey graph of the fitdistrplus package (Delignette-Muller & Dutang, 2015) and using goodness-of-fit statistics of the same package. The variable carapace width was modelled with a gamma family distribution, using the ‘log’ link function for interspecific comparison, and modelled with a normal distribution, using the ‘identity’ link function for intraspecific comparison in A. senex. The variable probability of finding active individuals on the surface was modelled with a binomial family distribution, using the ‘logit’ link function.

The variables were modelled with the ‘glm’ functions of the stats package. Model residuals were checked for assumptions of normality and homoscedasticity. The statistical significance of the generalized linear models was calculated with the ‘Anova’ function of the car package (Fox et al., 2012). We considered the statistical significance with a level of α = 0.05. We used the function ‘cld’ of the multcomp package (Hothorn et al., 2016) for a posteriori testing. For the creation of graphics, we used the ggplot2 package (Wickham et al., 2016).

RESULTS

Carapace width variation between species and habitat

We analysed 1071 individuals from 18 species or morphotypes. We found a significant interaction between sex and species regarding carapace width (χ² = 164.00; P < 0.005; d.f. = 16), indicating sexual size dimorphism in some Allocosinae species (Fig. 2). Allocosa noctuabunda (Montgomery, 1904) (SDI = −0.19), Allocosa parva (Banks, 1894) (SDI = −0.17) and Allocosinae sp. 6 ‘Bahia’ (SDI = −0.09) showed sexual size dimorphism biased to females; whereas males were larger than females in A. senex (SDI = 0.24), A. marindia (SDI = 0.13), Allocosinae sp. 3 ‘Coquimbo’ (SDI = 0.2) and Allocosinae sp. 4 ‘Lanín’ (SDI = 0.27).

Figure 2.

Carapace width (mean ± SD, in millimetres) in females and males of Allocosinae species collected during our field samplings and from scientific collections, distinguished with different colours according to their habitat. Significant differences between the sexes in each species are highlighted with an asterisk (*).

We found an effect of the habitat on the carapace width in Allocosiane species (χ² = 1335.90; P < 0.005; d.f. = 6; Fig. 3). Smaller individuals were found in the rainforest, grasslands and gardens, whereas larger ones were found in volcanic sandy coast, oceanic–estuarine–fluvial coast and dry riverbed–sandy valley (Fig. 3; Supporting Information, Table S1). Individuals from volcanic sandy coasts and oceanic–estuarine–fluvial coasts exhibited positive and higher SDI, indicating a strong bias towards males in sizes (SDI = 0.27 and SDI = 0.18, respectively). Individuals from fluvial coast (SDI = −0.19) showed negative SDI, suggesting female-biased sexual size dimorphism. Within A. senex, the carapace width exhibited a stronger bias for males in individuals from fluvial coast compared with those from oceanic–estuarine coast (F = 17.233; P < 0.005; d.f. = 1; Fig. 4; Supporting Information, Table S2).

Figure 3.

Average carapace width (mean ± SD, in millimetres) of Allocosinae adults of each habitat. Letters denote significant differences between habitat categories.

Figure 4.

Comparisons between carapace width (mean ± SD, in millimetres) of Allocosa senex females and males found in fluvial (A) and oceanic–estuarine coasts (B), and between females (C) and males (D) from both habitats. Letters denote significant differences between the categories.

Nocturnal surface mobility

We found interspecific differences in surface mobility among Allocosinae species (χ² = 67.94; P < 0.005; d.f. = 11; Fig. 5). Allocosinae sp. 4 ‘Lanín’ (the only species living in volcanic sandy coast habitat) presented the lowest probability of finding females on the surface (P = 0.04), whereas Allocosa cf. senex showed a remarkably high value (P = 0.9). Allocosa senex also exhibited high variation according to the location. The probability of finding walking females of A. senex was higher in ‘Montes del Queguay’ (Uruguay) (P = 0.94), whereas in ‘El Palmar’ (Argentina) this proportion was lower (P = 0.62). In the global analysis, we found significant differences between the sexes in surface mobility according to the habitat (χ² = 50.75; P < 0.005; d.f. = 4). The volcanic sandy coast (including Allocosinae sp. 4 ‘Lanin’) showed the lowest probability of finding walking females (P = 0.04), whereas the oceanic–estuarine–fluvial coast (including A. senex and A. cf. senex) showed the highest probability (P = 0.78) (Fig. 5).

Figure 5.

Probability of finding active females walking or leaning out from burrow entrances during the field samplings of this study, distinguished with different colours according to their habitat. Letters denote significant differences between habitat categories.

DISCUSSION

Our results reveal differences in sexual size dimorphism and surface mobility in adults of Allocosinae spiders living in different habitats throughout South America. We found new reports of species (or morphotypes) inhabiting coastal habitats with reversal in the expected sexual size dimorphism and with higher nocturnal surface mobility in females, similar to those reported for A. senex and A. marindia (Aisenberg et al., 2007; Aisenberg & Costa, 2008). However, as discussed below, not all Allocosinae species from those habitats showed non-typical sexual patterns. Our findings suggest that coastal habitats could be related to the occurrence of non-expected sexual traits in Allocosinae spiders. Nevertheless, other unexplored factors (i.e. phylogenetic, ecological, physiological and behavioural constraints, among others) could also be involved, paving the ground for investigating their impact on sexual role reversal in future research. Furthermore, in this study we report cases of South American spiders (A. cf. senex, Allocosinae sp. 3 ‘Coquimbo’ and Allocosinae sp. 4 ‘Lanin’) showing non-typical sexual traits in size dimorphism and/or mobility patterns, suggesting them as good candidates to test sex role reversal hypotheses.

As previously stated, sexual size dimorphism in wolf spiders has been reported to be moderate or absent (Head, 1995; Walker & Rypstra, 2002; Framenau & Hebets, 2007; Aisenberg et al., 2010; Logunov, 2011); again, the findings of this study controvert the widespread patterns for this group. We found that Allocosinae sp. 3 ‘Coquimbo’ from sandy fluvial coasts and Allocosinae sp. 4 ‘Lanin’ from volcanic lake coasts presented reversal in sexual size dimorphism, with males being noticeably larger than females. Additionally, preliminary data from records of sexual behaviour suggested that Allocosinae sp. 4 ‘Lanin’ presented characteristics (mobile females that initiate courtship) that match those expected in spiders with non-typical sexual roles (A. Aisenberg, pers. obs.). In this study, we confirmed that A. senex showed the reversed pattern of sexual size dimorphism at all the sampled localities, as in former reports from localities in Uruguay and Argentina (Aisenberg et al., 2007; Aisenberg & Costa, 2008; Bollatti et al., 2017). In agreement with this result, these three Allocosinae coastal species and A. marindia showed the highest positive sexual SDI values, suggesting body size biased to males.

Two other species (A. cf. senex and Allocosinae sp. 7 ‘Clarillo’) that inhabit South American sandy coasts showed positive SDI values, but without statistically significant support, when we compared carapace width between sexes. Surprisingly, Allocosinae sp. 6 ‘Bahia’ sampled in habitats similar to those of A. cf. senex showed sexual size dimorphism biased to females and negative SDI values. This pattern was also found in the North American species A. noctuabunda from fluvial coasts and A. parva from grasslands and gardens. Although not all the wolf spiders inhabiting coastal habitats showed a reversal in sexual size dimorphism, it is noteworthy that we did not find any species with males larger than females in other habitats, such as grasslands, gardens or rainforests. For example, the recently redescribed species Paratrochosina amica (Mello-Leitão, 1941) (Gonnet et al., 2021a) from South American grasslands and gardens did not show significant differences in carapace width between sexes, and values of SDI were close to zero.

The results of the present study suggest that the habitat, although not exclusively, could be shaping non-typical sexual size dimorphism in South American Allocosinae spiders. In this sense, it is worth noting the case of Allocosa alticeps (Mello-Leitão, 1944), which inhabits the sandy coastal dunes of Southern Buenos Aires Province, Argentina (Simó et al., 2017). This species is an exception to the rule, because Guerra et al. (2022) reported that although A. alticeps does not exhibit reversal in sexual size dimorphism, it exhibits behavioural characteristics that match those of sex role reversal (mobile females that initiate courtship). This study alerts us that other factors could be also acting in shaping these striking sexual traits. Also, trait correlation should be investigated in a phylogenetic context, to control for the relatedness of species. Unfortunately, at present a phylogeny including all the Allocosinae species reported in this study is not available. Preliminary phylogenetic results of the subfamily suggest that A. alticeps, A. senex, A. cf. senex, A. marindia, Allocosinae sp. 3 ‘Coquimbo’ and Allocosinae sp. 4 ‘Lanin’, all of which are coastal species, form a monophyletic group (Laborda et al., 2020; Laborda, 2021, 2023). These coastal species showed reversal in sexual size dimorphism, SDI values biased towards males and/or behavioural traits expected for sex role reversal (Aisenberg et al., 2007; Aisenberg & Costa, 2008; Guerra et al., 2022). The sister clade of this group includes P. amica (Gonnet et al., 2021a) from South American grasslands and gardens, in which sexual size dimorphism was not detected, and according to preliminary data, this species would not show reversal in sex roles (Gonnet et al., 2021b). An explicit phylogenetic comparative analysis including all the species analysed in the present study and most of the North and South American species from different habitats will be necessary to reconstruct the evolution of sexual traits and to understand fully the origins and drivers of diversification in South American Allocosinae.

Our results suggest that body size in this subfamily can vary according to the habitat (Fig. 3). Larger Allocosinae adults were found on volcanic coasts, followed by those from oceanic–estuarine–fluvial coasts and dry riverbed valleys, whereas smaller adults occurred in rainforests, grasslands and gardens. These differences are in agreement with studies by Henschel (1997) in Namib desert spiders. Henschel (1997) found that psammophilous species exhibit adaptations to avoid desiccation through larger body size, which reduces the surface-to-volume ratio. In A. senex, we found that adults from fluvial coasts presented larger carapace width than those from oceanic–estuarine coasts and that sexual size dimorphism biased toward males was stronger in individuals from fluvial coasts (Fig. 4). Differences in body size among individuals from locations with different ecological dynamics have already been reported in A. senex (Postiglioni, 2015; Bollatti et al., 2017), A. marindia (Cavassa et al., 2022) and in other wolf spiders (Bonte et al., 2006) and arthropods, such as amphipods and decapods (Contreras et al., 2003; Defeo & Gómez, 2005). Factors such as coastal morphodynamics or salinity could be driving these differences as has been cited for other coastal arthropods (Witteveen & Joosse, 1987; Ehlinger & Tankersley, 2004; Pétillon et al., 2011; Foucreau et al., 2012). Nevertheless, other factors, such as flooding adaptations, prey abundances and sexual selection, could also be acting (Blanckenhorn, 2000; Bollatti et al., 2017; Albín et al., 2021; Mardiné et al., 2022).

In general, there was a higher probability of finding females than males in all the sampled coastal Allocosinae species. Conversely, Allocosinae sp. 4 ‘Lanin’ from volcanic coasts showed the lowest probability of finding females at the surface (22 males and one female). This was an unexpected result, because Allocosinae sp. 4 ‘Lanin’ also lives in a coastal habitat, the volcanic sandy shores of the Huechulafquen Lake in Parque Nacional Lanín (Argentina). As mentioned above, this morphotype presents reversal in sexual size dimorphism, the females initiate courtship and they mate inside the male burrow (A. Aisenberg, unpublished data). Perhaps, the sandy volcanic habitat of the Huechulafquen Lake, near Volcán Lanin, surrounded by glaciers and characterized by snowy periods from May to October and by extreme daily variations in temperatures in summer, could constrain female sexual activity to other periods of the year. These hypotheses require further testing in the future.

Our study is the first to analyse sexual size dimorphism from a comparative approach in Allocosinae species from North, Central and South America, in addition to nocturnal surface mobility of South American species. The results unveil new cases of South American spiders that exhibit reversal in typical body size (i.e. Allocosinae sp. 3 ‘Coquimbo’ and Allocosinae sp. 4 ‘Lanin’) and high mobility in females (i.e. A. cf. senex). Future research should aim to record sexual behaviour and other life-history traits of the Allocosinae spiders of this study. Additionally, inferring a thoroughly sampled, well-resolved phylogeny of Allocosinae including these species is an unavoidable step towards testing hypotheses regarding the number of times that sex role reversal has evolved in this subfamily and how it could be related to colonization and survival in coastal environments.

SUPPORTING INFORMATION

Additional supporting information may be found in the online version of this article on the publisher's website.

Table S1. Allocosinae specimens from scientific collections, with details on sex, identity code, habitat, locality and carapace width (in millimetres).

Table S2.Allocosa senex specimens from scientific collections, with details on sex, identity code, habitat (fluvial coast or oceanic–estuarine coast), locality and carapace width (in millimetres).

ACKNOWLEDGEMENTS

We are grateful to T. Casacuberta, D. Hagopián, N. Kacevas, C. Mattoni, P. Pintos and M. Trillo for their help during field samplings. We acknowledge Charles Griswold for providing specimens from the arachnological collection of California Academy of Sciences (USA). We also thank National Park Administration El Palmar (Argentina), National Park Lanin (Argentina), National Park Río Clarillo (Chile), San Juan Authorities of Ischigualasto Provincial Park (Argentina), Authorities of Parque das Dunas (Brazil), Reserva PUCRS Pró-Mata (Brazil) and Protected Area Montes del Queguay (Uruguay) through Sebastián Horta (DSNAP, MVOTMA, Uruguay) for the authorizations for samplings. This study was supported financially by the projects FCE_1_2017_1_136269 (Fondo Clemente Estable, ANII) and NATGEO WW204R-17 (National Geographic Society). A.A., L.B.B., M.S., R.P., V.G. and A.L. acknowledge financial support by Programa Desarrollo de Ciencias Básicas (PEDECIBA, Uruguay) and Sistema Nacional de Investigadores (SNI, ANII, Uruguay). A.D.B. acknowledges a grant from Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq, no. 303903/2019-8). F.B., M.I. and M.O.D. acknowledge financial support by Consejo Nacional de Investigaciones Científicas y Técnicas (CONICET). R.P. is grateful for POS_NAC_2017_1_140432 (ANII, Uruguay), V.G. for POS_NAC_M_2020_1_164074 and A.L. for POS_FCE_2018_1_1007751. We are grateful to anonymous reviewers and the Editor, Professor John A. Allen, for their comments and suggestions, which improved the final version of the manuscript. The authors do not have any conflict of interest to declare.

DATA AVAILABILITY

Data underlying this study are available online at the Biological Journal of the Linnean Society.