Beneath the surface: co-habitation of recruits of the land crab Johngarthia lagostoma and its relevance for conservation on oceanic islands

Abstract

Gecarcinid crabs have their life cycles in antagonistic scenarios, with their larvae developing in the sea and the juvenile/adult phases occurring on land. Adults migrate from land to sea to release their larvae, which return to land upon reaching the megalopa stage. Recruitment and early instar traits in gecarcinids crabs remain unknown, leading to some species lacking age-specific information. Despite massive recruitment observed in some insular gecarcinid species (e.g. Gecarcoidea natalis), recruits are generally expected to be few and exhibit cryptic behaviour, potentially occupying the burrows of conspecifics. We evaluated whether recruits of Johngarthia lagostoma on Trindade Island, Brazil, co-inhabit larger conspecific burrows, analysing this occurrence and examining their growth patterns, density, and body size across different lunar phases. Johngarthia lagostoma recruits inhabit conspecific burrows, either abandoned or occupied by adult crabs, but always with leaves stored in the inner chamber. Recruits in co-inhabiting behaviour reach a maximum carapace width of 7.3 mm, and after that, they are likely detected by the adults and possibly cannibalized or leave burrows naturally. During the full moon, the higher density and smaller size of recruits were recorded, indicating a recruitment lunar phase. It is crucial to ascertain the prevalence of co-inhabitation behaviours in other land crab species to expand the knowledge about recruitment patterns in these key community species.

Introduction

Currently, around 4900 crustacean species have been recognized for their terrestrial conquest, among which crabs from the Brachyura infraorder are included. These crabs are part of the Order Decapoda, one of six crustacean lineages known to have successfully colonized terrestrial ecosystems (Marin and Tiunov, 2023). According to Wolfe et al. (2022), crabs transitioned from marine to non-marine habitats between 5 and 15 times, while making the reverse journey 3–4 times. This diversification primarily occurred in the Triassic period, with family-level divergences happening in the late Cretaceous and early Paleogene periods. The most derived crabs, commonly referred to as ‘true crabs’ (Eubrachyura), diverged from their ancestral brachyurans during the Cretaceous period (Tsang et al., 2014; Luque et al., 2021; Watson-Zink, 2021), and present some of the highest levels of terrestrial adaptations among the six grades proposed by Watson-Zink (2021), which were possible due to morphological, reproductive (e.g. aerial respiration, moulting, and development) and physiological changes (e.g. osmoregulation, nitrogen excretion, desiccation resistance, and thermoregulation).

Terrestrial and semi-terrestrial crabs comprise approximately 300 species. A high adaptation to a terrestrial lifestyle can be observed in the family Gecarcinidae (Marin and Tiunov, 2023), where the species colonize land through marine environments (including intertidal mudflats, sandflats, mangrove forests, etc.) (Watson-Zink, 2021). These crabs have a life history characterized by juvenile and adults occurring in terrestrial habitats, while their larvae undergo a planktotrophic development that spans approximately 15–30 days in the marine environment (five to six zoeal stages and one megalopa) (Colavite et al., 2021). The success of these species depends equally on the capacity of migration between the land-sea gradient in both phases and on their adaptations to survive in both environments (Bliss and Mantel, 1968; Burggren and McMahon, 1988; Hartnoll, 2010; Watson-Zink, 2021; Marin and Tiunov, 2023). Some gecarcinid species are endemic to oceanic islands, which poses an additional challenge for their migratory behaviour since the residence sites are even further from the sea (Doi et al., 2019; João et al., 2021), and sometimes reach more than 1000 metres of altitude (e.g. Gecarcinus ruricola in Caribbean Islands - Hartnoll et al., 2006). Indeed, studying insular gecarcinid species can be challenging due to their isolated habitats; however, there is existing well-documented research focused on understanding the biology and behaviours of crabs within this family (Bliss et al., 1978; Hicks, 1985; Foale, 1999; Adamczewska and Morris, 2001; Green, 2004; Hartnoll et al., 2006, 2007, 2009, 2017; Liu and Jeng, 2007; López-Victoria and Werding, 2008; Turner et al., 2011; Perger, 2014; Sanvicente-Añorve et al., 2016; Tavares and Mendonça, 2022; João et al., 2023a). However, there is a knowledge gap concerning how these species transition from the larval phase to land. The process of recruitment in these species has received limited attention, with only a few documented events and scarce information available (Hartnoll and Clark, 2006; Hartnoll et al., 2014).

There are two well-documented recruitment processes for insular gecarcinids in the literature, the Christmas Island red crab Gecarcoidea natalis (Hicks, 1985; Hicks et al., 1990) and the Caribbean black crab Gecarcinus ruricola (Hartnoll and Clark, 2006). In both cases, a notable similarity is the mass return of megalopae, which creates a visually striking phenomenon where the tideline is painted with a red colour. This mass return event greatly facilitates the understanding of the overall recruitment process in both species. Furthermore, it appears that the behaviours of egg release and subsequent return of recruits in gecarcinid crabs are connected to the phases of the full or new moon (Hicks, 1985; Liu and Jeng, 2005, 2007; Hartnoll and Clark, 2006). Among all other gecarcinid species where there is some recorded information about recruitment in the literature, the only common characteristic is the presence of megalopae on land (Johngarthia lagostoma (H. Milne Edwards 1837) and J. weileri – Hartnoll et al., 2014; J. planata – Erhardt and Niassaut, 1970; Gecarcoidea lalandii, erroneously named as G. natalis in the study – Webb, 1922; and Tuerkayana celeste – Hicks et al., 1990). Although these works provide only basic records, they do suggest that gecarcinid megalopae could live on land, even venturing more than 100 metres away from the shoreline. Unfortunately, the recruitment events can be infrequent and sporadic, as observed for Gecarcinus ruricola presenting an interval between each recruitment every 5 or 6 years (Hartnoll and Clark, 2006), which makes documenting and studying these events difficult.

Population studies on insular gecarcinids have highlighted the absence of records on first crab instars and juvenile individuals. As a result, these studies have indicated a concerning pattern of population aging (Hartnoll et al., 2009; Turner et al., 2011); however, this could partially be attributed to methodological limitations. The habitat preferences and distribution of juvenile crabs remain unclear, which can lead to sampling biases that primarily capture adult individuals (Turner et al., 2011). In addition, there is evidence that juveniles of certain crab species occupy specific habitats such as crevices and areas under rocks (as Tuerkayana hirtipes – Hicks et al., 1990) or could be associated with adult burrows (as described for T. hirtipes – Hicks et al., 1990; and for Cardisoma carnifex – Vannini et al., 2003). Cleary, associating with adult burrows can be considered an adaptive strategy because burrows remain for at least five years in some cases (Green, 2004), providing a humid and thermally stable habitat (Greenaway, 1989; Berti et al., 2008), with chambers where leaves are stored by the owner crab (O'Dowd and Lake, 1989; Vannini et al., 2003).

Johngarthia lagostoma, commonly known as the yellow land crab, is an endemic species of insular land crab found in four islands around the world, in the South Atlantic Ocean (Rocas Atoll, Fernando de Noronha, Ascension Island, and Trindade Island, according to Melo, 1996). Few studies were conducted about their biology, mostly on Ascension and Trindade islands (Ascension: Hartnoll et al., 2009, 2010; Trindade: João et al., 2021, 2022, 2023a, 2023b; Tavares and Mendonça, 2022; Entringer and Srbek-Araujo, 2023), where J. lagostoma population structure was considered skewed for adult individuals (Hartnoll et al., 2009; João et al., 2023a). In addition, the lack of clear information about the recruitment of J. lagostoma is of particular concern for Brazilian islands (Rocas Atoll, Fernando de Noronha, and Trindade Island) where this species is categorized as Endangered – EN (Santana and Coelho, 2018; MMA, 2022), following the IUCN criteria (IUCN, 2012). A pressing concern that requires immediate investigation is the recruitment of J. lagostoma to each island, as this information is crucial for assessing the species' demographics and informing conservation strategies (Pinheiro et al., 2016). The only note about the species recruitment in Trindade Island was made in 1987, where initial crab stages were observed in galleries constructed by adults (Tavares and Mendonça, 2022). So, in this study, we evaluate the association between the first instars of J. lagostoma and adult crab burrows in Trindade Island (Brazil) to co-inhabiting behaviour, analysing recruits' relative growth, population density, and size in function of the lunar phases (full, waning, new, and waxing moons).

Materials and methods

Study area and recruits sampling

All the samples were carried out on Trindade Island (20° 29″ S- 29° 20.7″ W), a volcanic island located in the South Atlantic, approximately 1200 km off the Brazilian coast (Figure 1A, B). A portion of Trindade Island has been designated as a large marine protected area since 2018 (ICMBio, 2018). The island is permanently inhabited by the Brazilian Navy, with a human population of around 40 people, including military and researchers. The sampling took place on the eastern face of Trindade Island, specifically at Andradas Beach (Figure 1C), which is known to be an important reproductive site for J. lagostoma (João et al., 2023a). To evaluate the co-habitation of recruits on galleries of adult conspecific, we conduct samples in January 2020, during the reproductive season of J. lagostoma, which typically occurs between December and May for Rocas Atoll and Ascension Island (Teixeira, 1996; Hartnoll et al., 2010), and between October to April for Trindade Island (Tavares and Mendonça, 2022; João et al., 2023a).

Figure 1. Geographic location of Trindade Island (Brazil). Where: (A) Southeast Brazilian coast showing the position of the Trindade Island; (B) general view of the Trindade Island indicating the study area location (gray circle); (C) general view of Trindade Island; (D) general view of the study area (Andradas Beach); (E) frontal view of a Johngarthia lagostoma adult in the Andradas Beach; and (F) dorsal view of a J. lagostoma recruit (scale: 1 cm).

At Andradas Beach (Figure 1D), J. lagostoma is the exclusive species known to construct galleries in the supralittoral zone, primarily associated with sand hill vegetation. These burrows are called ‘transit burrows’ by Hartnoll et al. (2010) due to be constructed during migration or larval release but subsequently abandoned. For each lunar phase (full, waning, new, and waxing), we systematically examined a minimum of 25 random and visibly active burrows, characterized by the absence of debris accumulation and the presence of other biogenic signals (e.g. tracks and faeces) close to the opening. During the day period, when crabs typically remain within their burrows, each gallery was carefully and manually excavated until reaching its end or until to attain an adult of J. lagostoma (Figure 1E). In this process, all excavated sediment was collected and placed in a plastic tray and then sieved. All adult crabs collected during the excavation were identified based on their respective gallery numbers, sexed, and reserved in plastic boxes to be released back onto the beach at the end of these procedures. The sex was verified by inspection of abdominal dimorphism (males, subtriangular; and females, semi-rounded) and the number of pleopod pairs (males, two uniramous pairs; and females, four biramous pairs). Following the sieving of the sediment, the recruits (Figure 1F) of each burrow were carefully placed in labelled individual plastic tubes with their corresponding gallery number. These tubes were transported to the laboratory, where the recruits were crioanesthetized and subsequently preserved in 70% alcohol. Finally, after excavation, the sediment surrounding the burrows collapsed, mitigating the risk of sampling the same burrow other times. Moreover, during each lunar phase, samples were taken from different locations within the sand dune vegetation to prevent potential bias due to previous sampling.

Recruits: recognition, measurements, and biometric relationships

In the laboratory, all recruits were evaluated to be identified as megalopae or juveniles of J. lagostoma species, using diagnostic characters informed for Gecarcinids by Cuesta et al. (2007), respectively. Following, each exemplar was measured using an image analysis system (KS-300^®-Zeiss^®) integrated to a stereomicroscope (Axiolab^®-Zeiss^®, 0.01 mm). The carapace width (CW, mm) was determined as the maximum cephalothorax distance between the lateral margins and used as the reference for body size in crabs. Frequency histograms were constructed using 1 mm CW size classes, from which the Fisher asymmetry coefficient (SK, skewness) was calculated according to Sokal and Rohlf (2012), as recommended by Pinheiro et al. (2022), and categorizing the size distribution as symmetric (−0.5 ≤ SK ≤ 0.5), positive asymmetric (SK > 0.5) of negative asymmetric (SK<−0.5).

Biometry of juveniles was registered in some morphological structures (also in millimetres), represented by: carapace length (CL, distance between the frontal to posterior margin of carapace); major cheliped propodus length (PL, distance between end of the fixed finger and the tooth at the propodus-carpus joint); and abdominal width (AW, greatest width in the fifth somite). Additionally, the weight of each recruit (WT, in grams) was recorded using a digital analytical scale (Ohaus, 0.0001 g). Individuals with carapace damage or missing appendages were excluded from the biometric analysis. Since sexual dimorphism recognized was not present in the recruits, they were not sexed.

To test the fit between all biometric measures, morphological relationships were examined by using carapace width (CW) as the independent variable and the other measures (CL, PL, AW, and WT) as the dependent variables. Regression analysis was employed to determine the fit, employing the power function (Y = aX^b) in each biometric relationship. To accomplish this, all variables were log-transformed to calculate the linear models for allometric growth rates. The slope value (constant ‘b’) was used to determine relative growth to each relationship (CL × CW, PL × CW, and AW × CW), where growth patterns could be categorized as isometric (b = 1), positive allometric (b > 1), or negative allometric (b < 1). The same was conducted with the WT × CW relationship, in this case with b-values characterizing these weight's growth patterns in relation to 3 (Pinheiro and Fiscarelli, 2009). Student's t-tests were conducted to assess departures from isometry in all relationships (α = 0.05), using b-values of 1 or 3 in each case.

Relationship between juvenile recruits' traits with lunar phases and adult body size

To assess the influence of lunar phases on J. lagostoma recruitment, the density of juvenile recruits (DE, individuals per burrow) and corporal measures (CW and WT) values were used as dependent variables, while the lunar phases (full, waning, new, and waxing) were treated as factors. The normality assumption and homogeneity of variances for all dependent variables were assessed using Shapiro–Wilk (W) and Levene (L) tests, respectively. Since the data did not meet the assumptions of parametric distribution, a Kruskal–Wallis test (Sokal and Rohlf, 2012) was conducted. In cases where statistically significant differences were observed, post-hoc Dunn tests were performed to compare the median values across different lunar phases. All these statistical analyses were performed using R version 4.2.1 (R Core Team, 2022) with the ‘dunnTest’ function available in the ‘FSA’ package (Ogle et al., 2023).

Evaluation of a possible association between juvenile recruits' density/burrow (DE) and mean juvenile recruits' body size (CW_J), both in relation to adults' body size (CW_A), was submitted to regression analysis by linear regression analysis, represented by DE vs CW_J and CW_J vs CW_A, respectively. This association was also evaluated by Pearson's linear correlation coefficient (r), considering ‘n-2’ degrees of freedom.

Results

Recruits' occurrence and biometric relationships

A total of 128 burrows were examined, of which 54.7% (n = 70) were empty without any recruits or owners. In 23.4% of the burrows (n = 30), both J. lagostoma owner and recruits were present, while in 21.9% (n = 28) only the recruits were found inhabiting the galleries. Among the burrows with both owners and recruits, 53.3% had ovigerous females (n = 16), 43.3% had males (n = 13), and only 3.3% had non-ovigerous females (n = 1). A total of 113 J. lagostoma recruits were sampled, with a density (DE) ranging from 1 to 10 ind./burrow (mean ± standard deviation: 1.7 ± 0.9 ind./gallery; and variation coefficient: 52.9%), and their carapace width (CW) ranged from 2.9 to 7.5 mm (4.9 ± 0.7 mm; and 14.3%, respectively) (Table 1). The size frequency histogram indicated a symmetric distribution for all sampled recruits (SK = 0.3, Figure 2).

Table 1. Regression analysis using all biometric relationships of Johngarthia lagostoma recruits from Trindade Island, Brazil

Where: AW, abdominal width; CL, carapace length; CW, carapace width; PL, cheliped propodus length; R ², coefficient of determination; t, calculated t-value evaluating departure from isometry (b = 1), expressed by *P ≤ 0.05 or ^nsP > 0.05; and WT, weight.

Figure 2. Abundance of Johngarthia lagostoma recruits in a size-frequency histogram showing a symmetric distribution (SK = 0.28). Where: CW, carapace width; and numbers above the bars, N of each size-class.

Inside the burrows, the presence of both owners and recruits was not random. Even in the absence of an owner, the burrows appeared to have been recently abandoned, as indicated by the absence of debris and the well-structured entrance and tunnels. The owner was consistently found at the deepest part of the gallery, which varied in distance from approximately 20 to 100 cm from the ground surface. Regardless of the presence or absence of the owner, the recruits were never found in this deepest part of the gallery but were commonly associated with small crevices along the sides of the tunnels. These crevices seemed to be used by the owners for storing food, represented by the accumulation of leaves and bioturbated sediment.

All biometric relationships showed statistical significance (P < 0.05), displaying a positive correlation (ρ ≥ 0.85; P < 0.001), and demonstrating good fits (R ² ≥ 0.70) in the regression analysis (Table 1). The slopes calculated for CL × CW (b = 0.93) and AW × CW (b = 0.93) indicated isometry, meaning there was proportional growth between the dependent variables and CW. On the other hand, the PL × CW relationship exhibited a lower slope (b = 0.77) and confirmed negative allometry, indicating that there was greater growth in CW compared to PL. Lastly, the WT × CW relationship confirmed isometry (b = 3.11), showing that the weight (WT) and CW grew proportionally in the recruited individuals.

Relationship between juvenile recruits' traits with lunar phases and adult body size

Firstly, all dependent variables (DE, CW, and WT) were significantly influenced by the lunar phases (DE: KW = 28.3, P = 0.003; CW: KW = 21.4, P = 0.0001; WT: KW = 22.3, P = 0.0001 – Figure 3). For density of juveniles the highest values were recorded during the full moon phase (DE: 1 to 10 ind./gallery = 2.0 ± 2.5 ind./gallery), which decreased significantly during subsequent moon phases (Figure 3A). In the case of CW and WT values (Figure 3B, C), an opposite pattern was observed, with the lowest values registered during the full moon phase (CW: 3.2 to 5.7 mm, 4.7 ± 0.5 mm; WT: 0.003 to 0.06 g, 0.03 ± 0.01 g), followed by an increase, with the highest values observed during the waxing moon phase (CW: 5.4 to 7.5 mm = 6.4 ± 0.9 mm; WT: 0.06 to 0.1 g = 0.1 ± 0.04 g).

Figure 3. Variation of the density (ind./burrow: A), size (CW, carapace width: B) and weight (WT, wet weight: C) of the Johngarthia lagostoma recruits co-inhabiting burrows with conspecific crabs along the lunar phases. Where: line inside the box, median values; rhombus dot, mean values; box, interquartile range (IQRs); whiskers, lowest and highest values within 1.5×IQRs; circle dot, original data on which a random noise was added to avoid overlap. Distinct letters indicate significant differences in the dependent variables between the lunar phases (P ≤ 0.05).

Overall, recruits increased an average size of 1.4 times between full and waxing moons (=39.4% per month), corresponding to an average increase in weight of 3.23 times (323% per month) for the same period. Furthermore, a higher frequency of occurrence of recruits in the galleries (65.9%) was recorded during the full moon, being reduced by 50.5% after one month (lunar cycle), in relation to the lowest percentage recorded on the waxing moon (15.4%). Finally, only one megalopa was registered during the studied period, found inside the galleries inspected in 7.1% (n = 1) in the full moon.

The linear regression analysis for DE vs CW_A relationship was not significant (DE = 0.039 CW_A – 0.394; R ² = 0.028; n = 25) with a positive but not significant association between them (r = 0.168; P > 0.001). The same was verified for CW_J vs CW_A relationship, which was not significant for the regression (CW_J = 0.0033 CW_A – 4.73; R ² = 0.002; n = 25) and association among these variables (r = 0.045; P > 0.001).

Discussion

The knowledge of gecarcinid crabs recruitment is generally limited (Vannini et al., 2003; Hartnoll and Clark, 2006; Hartnoll et al., 2014), with few reports available for some species, such as J. lagostoma (von Fimpel, 1975; Hartnoll et al., 2014). Specifically, research on J. lagostoma recruitment has mainly focused on Ascension Island, where megalopae and first instar crabs (4.2 mm CW) were found over 100 metres from the sea (Hartnoll et al., 2014).

In Brazilian islands (Fernando de Noronha, Rocas Atoll, and Trindade), information is limited and sometimes unclear. A report mentions a 7 mm dark brown crab leaving the sea on Trindade Island (von Fimpel, 1975), raising doubts since other gecarcinid crab recruitment records have found megalopae on land (see Lafaix, 1969; Klaassen, 1975; Hicks, 1985; Hartnoll and Clark, 2006; Hartnoll et al., 2014), including the anomuran Birgus latro (Drew et al., 2010). Thus, we support the pattern suggested by previous studies, where gecarcinid megalopae enter land and then metamorphose into the first instar stage (Hartnoll et al., 2014).

Chemical cues influence the metamorphosis from megalopa to the first juvenile instar in both aquatic and semi-terrestrial brachyurans (see Christy, 1989; Wolcott and de Vries, 1994; Andrews et al., 2001; Diele and Simith, 2007; Simith et al., 2010; Christy, 2011). These studies suggest that metamorphosis is faster and more targeted when megalopae encounter conspecific cues in sediment. However, its relevance to gecarcinid crabs, especially as their megalopae transition to land, remains unclear. In terrestrial environments, detecting and interpreting chemical cues presents unique challenges (Krång et al., 2012; Waldrop et al., 2016). Although it has been hypothesized that the odour within a species' burrow is more concentrated (Schmidt and Diele, 2009), this aspect was not studied for J. lagostoma.

The lack of records on the early instars of land crab species can be attributed to three main factors, as suggested by Vannini et al. (2003): (1) irregular recruitment patterns, which complicate the timing of first instar detection; (2) previous sampling efforts focused primarily on adults, potentially overlooking juvenile habitats; and (3) juvenile recruits, although present, often remain concealed within refuges. Our findings suggest that the latter two factors are particularly relevant to the recruitment of J. lagostoma. The co-habitation observed in our study indicates that J. lagostoma recruits occupy the same habitat as adults but with varying levels of concealment. However, no recruits were found in higher elevation environments (e.g. non-flooded supralittoral or mountainous areas), with or without vegetation.

Although our field research predominantly identified conspecific burrows as habitats for recruits, this does not imply they are the exclusive refuges. Due to the scope of our investigation on Trindade Island, other potential habitats such as crevices and spaces beneath rocks in the supralittoral zone were not explored, leaving open the possibility of additional juvenile habitats. Nevertheless, the information from our study, along with previous records for J. lagostoma (Tavares and Mendonça, 2022) and similar findings for C. carnifex and Gecarcinus lateralis (see Vannini et al., 2003 and Klaassen, 1975, respectively), offers valuable insights into where recruits might undergo moulting stages before constructing their own burrows.

The transition to terrestrial environments poses significant challenges for crabs, with desiccation being a major concern, especially on oceanic islands near the equator and tropics. Trindade Island is characterized by intense east winds and frequent heavy rains, often escalating into tropical storms that flood its valleys (Pires et al., 2013). These rain events, known as ‘Pirajás,’ occur almost daily during the summer, triggered by the humid oceanic air rising over the island's highest peak, ‘Desejado’ (~600 m), resulting in storm clouds that precipitate around the island (SECIRM, 2017). This study was conducted in January, one of the island's driest months, with an average rainfall of 65 mm (SECIRM, 2017). Although Trindade is the only Brazilian oceanic island with perennial watercourses and springs (Marques et al., 2019), beaches like Calhetas and Andradas have type IV drainage (direct to the ocean or with limited flow). This highlights the importance of ‘Pirajás’ in maintaining sediment and undergrowth moisture during the key reproductive period of this species, a factor that must be considered.

Gecarcinidae crabs dig burrows in sediment to cool off and may inhabit areas influenced by tides or vegetation (Watson-Zink, 2021). Bliss (1963) observed that terrestrial crabs often occupy elevated areas, creating shallow burrows that avoid groundwater and help maintain low internal moisture. Gecarcinid species within the genera Johngarthia, Gecarcinus, and Gecarcoidea possess a ventral tuft of hydrophilic setae that aids in water retention (Bliss, 1963, 2014; Guinot et al., 2018). In J. lagostoma, these setae are located between the 5^th pereopod and the margins of the 1st–2nd pleonal somites (Oliveira, 2014). More terrestrial species, such as Cardisoma guanhumi and J. lagostoma, also utilize arthrodial membranes for water absorption and have adaptations to minimize water loss, including a strong seal of the branchial chambers (Wolcott, 1984).

After these rainy events, the water can be retained in the interstices of the sediment, particularly at greater depths, as well as in the larger biomass of undergrowth vegetation (e.g. the grass Cyperus atlanticus). Even during the drier summer months, this vegetation can provide a humid, shaded environment as its dried tussocks droop over the sediment, forming so-called ‘vegetation crowns’. These areas frequently harbour adult J. lagostoma, but notably not their juvenile counterparts. This suggests that factors such as moisture levels and potential chemical signals from adults may not be as influential in recruit behaviour as initially hypothesized, given that recruits are not found in association with adults in these ‘vegetation crowns’. Further investigation is necessary to fully understand these dynamics.

Adults of J. lagostoma are more active at night, likely due to lower temperatures and higher humidity (Hartnoll et al., 2006). Daytime activity varies, with individuals seen at dawn and dusk, but rarely during intense sunlight and low humidity, particularly in the absence of ‘Pirajás’. This susceptibility to desiccation likely affects megalopae and juveniles, suggesting a reliance on summer rains. Additionally, their activity may be synchronized with moon phases that generate higher tidal flooding amplitudes, a pattern common among semi-terrestrial crabs. The association of juvenile recruits with adult burrows could be due to chemical attraction to plant material stored within the burrows, a behaviour also noted in other gecarcinid crabs [e.g. C. carnifex and C. guanhumi as reported by Micheli et al. (1991) and Novais et al. (2021), respectively], as well as in some semi-terrestrial crabs [e.g. Ucides occidentalis and Ucides cordatus as noted by Twilley et al. (1997) and Schories et al. (2003), respectively]. The plant material, rich in nitrogen and carbon, undergoes decomposition, which increases nitrogen content through microbial activity (Nordhaus et al., 2017; Tongununui et al., 2021; Gao and Lee, 2022). The decomposition process is further enhanced by J. lagostoma during ingestion, breaking down complex molecules into more absorbable forms (Johnston et al., 2005). Microbial volatile organic compounds (M-VOCs – e.g. alcohols, phenols, etc.) are produced during decomposition (Gray et al., 2010; Tongununui et al., 2021), with their composition depending on the vegetable matrix, microorganisms involved, and fermentation conditions (Rajendran et al., 2023). The odours emitted during this process can provide information about food quality and potential benefits or dangers to consumers (Price et al., 2011; Davis et al., 2013). Gecarcinid crabs are particularly sensitive to these odours, as evidenced by their frequent capture using traps baited with aromatic or strong-smelling fruits (e.g. lemon, pineapple, banana, jackfruit) (Krång et al., 2012). Despite this common attraction in adults, no studies have yet confirmed olfactory attraction in juvenile Gecarcinidae species, highlighting an area for future research.

A question that remains unanswered pertains to the mechanisms through which recruits enter the burrows of larger crabs. Possible factors could include the attractiveness of a moist or shaded environment, the release of pheromonal signals by adults, or the availability of pre-processed food within the burrows, among other factors not addressed by us. While these aspects remain speculative, answers to them could significantly contribute to our understanding of the species' ecology.

The symmetric size distribution of recruits suggests the range at which co-inhabiting behaviour begins and ends, likely around 7 mm CW (Figure 2), when recruits either leave the burrow or are detected and consumed by adults. A similar pattern is observed in U. cordatus, where recruits co-inhabit conspecific burrows until reaching a size that risks detection and cannibalism (Vannini et al., 2003; Schmidt and Diele, 2009). Cannibalism is common among gecarcinid species (Erhardt and Niassaut, 1970; Bliss et al., 1978; Hicks, 1985; Wolcott, 1988). This risk increases when food is scarce or during stressful periods, such as when females await larval release (Wolcott and Wolcott, 1984, 1987; Hartnoll et al., 2010). Consequently, crabs larger than 7 mm CW likely leave the burrow to find alternative refuges.

Biometric analysis of recruits reveals trends typical of terrestrial crabs but distinct from adult gecarcinids. For the CL × CW relationship, recruits have a more square-shaped carapace, reflecting a transitional morphology. While megalopae generally have a longer carapace (Cuesta and Anger, 2005; Hartnoll and Clark, 2006; Cuesta et al., 2007; Hartnoll et al., 2014), adults show a wider carapace due to negative allometry (Hartnoll et al., 2006). This shift likely aids in adapting to terrestrial habitats, where a broader carapace improves gill chamber space for respiration (Bliss and Mantel, 1968; Vannini et al., 2003). In contrast to typical positive allometry seen in juvenile gecarcinids (Hartnoll et al., 2006; Molina-Ortega and Vázquez-López, 2018; Doi et al., 2019; João et al., 2022), our study found negative allometry for chelipeds and isometry for the abdomen in recruits. This suggests that growth during this phase prioritizes overall body size rather than specific structures linked to sexual maturity (Hartnoll, 1982). The WT × CW relationship in recruits also displayed isometry, unlike the negative allometry seen in adults (Hartnoll et al., 2006; Molina-Ortega and Vázquez-López, 2018), reinforcing the focus on balanced body growth during recruitment.

Our findings regarding the density of recruits and their body measurements (CW and WT) reveal a distinct pattern associated with lunar phases, at least during our sampling period (January 2020). We observed an inverse relationship, whereby higher recruit densities were observed during the full moon, which subsequently decreased in the following phases. Furthermore, we found that the smallest recruits were sampled during the full moon, while their mean size and weight increased in subsequent moon phases. These observations suggest that there was a peak in recruitment activity during the full moon, followed by growth of the recruits within the owner burrows during the subsequent phases. Worth noting, that only one megalopa was captured in this study during the full moon, confirming the starting point of the recruitment process of this species. The reproductive process of many gecarcinids is known to be influenced by lunar phases, particularly during periods of larger tidal amplitudes, such as the full and new moon (Klaassen, 1975; Liu and Jeng, 2005, 2007). However, the specific timing of recruitment for J. lagostoma remains largely unknown, with only two instances associated with the new moon reported for individuals from Ascension Island (Hartnoll et al., 2010). In the case of other gecarcinid species, such as G. lalandii and Epigrapsus notatus, a significant release of larvae by females has been observed during the new moon (Liu and Jeng, 2007) and full moon (Liu and Jeng, 2005), respectively.

Recruitment in gecarcinid species is generally sparse, raising concerns for their conservation (Hicks, 1985; Hartnoll and Clark, 2006). Demographic studies of the Gecarcinidae family show low juvenile proportions, ranging from 8.5% for T. hirtipes (Turner et al., 2011) to 26% for Gecarcinus ruricola (Hartnoll et al., 2006) and 36.4% for E. notatus (Doi et al., 2019). Recruitment of individuals under 10 mm CW is rare, typically seen only in species with massive recruitment events (Hicks, 1985; Hartnoll and Clark, 2006). Juvenile J. lagostoma populations vary across islands: 0.7% on Ascension Island, 4.0% at Rocas Atoll (Teixeira, 1996; Hartnoll et al., 2009), and 16.4% on Trindade Island (João et al., 2023a). This percentage for Trindade could increase if individuals from this study are included, showing a relatively better recruitment scenario. Replicating this study's methodology on other islands could confirm co-inhabiting behaviour and reveal changes in population structure, contributing significantly to the management and conservation of J. lagostoma in Brazilian islands.

Our findings offer valuable insights into the recruitment of J. lagostoma, focusing on three key aspects: the behaviour of recruits in conspecific burrows, growth patterns in morphometric traits, and the influence of lunar phases on recruitment. While some findings align with existing literature, others reveal new aspects, such as co-inhabiting behaviour. These insights enhance our understanding of the species’ biology and have important implications for conservation efforts.