Ecology and Natural History of The Endangered Upland Burrowing Treefrog, Smilisca dentata
Abstract
The Upland Burrowing Treefrog, Smilisca dentata, classified as endangered by the International Union for the Conservation of Nature, are primarily threatened by land-use change. This study presents the results of the first systematic, quantitative analysis of one of its populations, conducted in 2005 within a 1.3 km2 area in southern Aguascalientes, Mexico. We analyzed the influence of environmental and habitat conditions on the population. We estimated a population size of 595.1 ± 197.8 individuals; over 50% were adults, and the sex ratio biased toward females. About 25% of females had ovarian oocytes, although no reproductive activity was observed. Surface activity occurred from July to September, gradually declining over that period, with most individuals found far from disturbed areas. No spatial segregation among age or sex categories was detected. Activity was restricted to moderate temperatures, no more than 26 °C, and high humidity. Males were active over a wider temperature range, whereas nongravid females were active over a broader humidity range; gravid females showed the narrowest humidity range preference. The presence of rain, high wind, and lack of cloud cover limited activity. The main diet items were in the orders Coleoptera and Hymenoptera. Habitat characterization revealed signs of disturbance, particularly changes to the dominant plant species and composition of superficial soil. This study provides new ecological context for future fieldwork, monitoring, and updated conservation risk assessments related to management of this endangered species. We recommend studies on the climate resilience of individuals to assess their vulnerability to global warming. In addition, we emphasize the need to conserve as much intact and connected habitat as possible in the region.
Resumen
La rana arborícola de madriguera de tierras altas, Smilisca dentata, clasificada como amenazada por la Unión Internacional para la Conservación de la Naturaleza, esta principalmente amenazada por el cambio de uso de suelo. Este estudio presenta los resultados del primer análisis sistemático y cuantitativo de una de sus poblaciones, realizado en 2005 en un área de 1.3 km² en el sur de Aguascalientes, México. Analizamos la influencia de las condiciones ambientales y del hábitat sobre la población. Estimamos un tamaño poblacional de 595.1 ± 197.8 individuos, con más del 50% adultos y una razón sexual sesgada hacia hembras; alrededor del 25% de las hembras presentaron ovocitos ováricos, aunque no se observó actividad reproductiva. La actividad superficial ocurrió de julio a septiembre, disminuyendo gradualmente, y la mayoría de los individuos se encontraron lejos de áreas perturbadas. No se detectó segregación espacial entre categorías de edad o sexo. La actividad se restringió a temperaturas moderadas, no mayores a 26 °C, y alta humedad. Los machos estuvieron activos en un rango más amplio de temperaturas, mientras que las hembras no grávidas usaron un rango más amplio de humedad; las hembras grávidas mostraron la preferencia más estrecha en cuanto a humedad. La presencia de lluvia, vientos fuertes y la falta de nubosidad limitaron la actividad. Los principales elementos en la dieta fueron de los órdenes Coleóptera e Himenóptera. La caracterización del hábitat reveló signos de disturbio, particularmente cambios en la dominancia de especies de plantas y en la composición superficial del suelo. Este estudio aporta un nuevo contexto ecológico para futuros trabajos de campo, monitoreo y actualizaciones en las evaluaciones de riesgo para la conservación. Recomendamos estudios sobre resiliencia climática de los individuos para evaluar su vulnerabilidad al calentamiento global. Además, enfatizamos en la necesidad de conservar la mayor cantidad posible de hábitat intacto y conectado en la región.
It is estimated that at least half of the world's vertebrate species are facing significant population declines and potential extinction (Ceballos et al., 2017; Luedtke et al., 2023). This trend has been documented for amphibians since the last century, with many studies beginning in the 1970s and increasing substantially in the early 1990s onward (Angulo, 2002; Lips et al., 2005; Butchart et al., 2010). Currently, amphibians represent the vertebrate group with the highest proportion of species classified as at risk. This figure is likely an underestimate as it excludes species for which there are insufficient data for a proper assessment (Hoffmann et al., 2010; Ceballos et al., 2017).
The Upland Burrowing Treefrog, Smilisca dentata (Smith, 1957), member of the Hylidae family, is classified as endangered (EN) by the Mexican government (NOM-059-SEMARNAT-2010) and the International Union for the Conservation of Nature (IUCN). In 2009, the IUCN Amphibian Specialist Group estimated the extent of occurrence (EOO) to be 983 km², based on at least five localities defined by the degree of threat, and identified habitat loss as the main threat leading to population decline (IUCN SSC Amphibian Specialist Group, 2020). Recently, Encarnación-Luévano et al. (2025) quantified the impact of this loss, and they estimated the potential distribution based on climatic requirements that match natural grasslands and crasicaule scrub habitats for the species was at only 472 km2 in a highly fragmented and discontinuous landscape.
Although some ecological data exist locally, these are not widely accessible. At the level of the S. dentata's range, information remains largely limited to historical records of occurrence. Since its initial description by Smith (1957), multiple studies have documented 21 localities from Jalisco, Aguascalientes, and Zacatecas (Smith, 1957; Chrapliwy et al., 1961; Vázquez-Díaz and Flores-Villela, 1991; Ávila-Villegas et al., 2009; Quintero-Díaz and Vázquez-Díaz, 2009; Ávila-Villegas and Flores de Anda, 2017; Villalobos-Juárez, 2023; Encarnación-Luévano et al., 2025). Trueb (1969) and Duellman (1970; 2001) contributed morphological descriptions based on the holotype (Smith, 1957) and additional specimens collected by Chrapliwy et al. (1961), while also offering brief notes on natural history. Subsequently, Vázquez-Díaz and Flores-Villela (1991) reported the mating call from a population in Jalisco. From this same site, Rodríguez-Torres and Vázquez-Díaz (1996) documented abundance data, recording 23 breeding individuals at a 10 m² temporary pond and 26 individuals at a 437 m² pond. Later, Quintero-Díaz and Vázquez-Díaz (2009) published a popular science book summarizing aspects of natural history based on fieldwork and student research, focusing mainly on a population within the rural community of Buenavista de Peñuelas, south of the city of Aguascalientes.
From these references, it is known that individuals exhibit morphological adaptations for fossorial habits, such as incomplete integumentary-cranial co-ossification, the presence of a metatarsal tubercle, and narrower digit tips compared to more arboreal congeners (Duellman, 1970; 2001). Males do not differ significantly in snout vent length (SVL) from females (Quintero-Díaz and Vázquez-Díaz, 2009), and the minimum reported adult SVL was 47.6 mm (n = 28; Duellman, 2001). Individuals breed in shallow temporary ponds of approximately no more than 10 cm in depth, typically located in short grass plains scattered with xeric shrubs (Smith, 1957; Chrapliwy et al., 1961). Finally, Quintero-Díaz and Vázquez-Díaz (2009) described additional aspects of the species’ biology and ecology including adult and larval morphology, fossorial behavior, diet components, habitat characteristics, biological interactions, and environmental preferences for activity. Individuals dig underground galleries about 30 cm deep where they remain sheltered during the day and throughout extended dry periods. During droughts, they enter a state of dormancy by forming a protective cocoon that withstands harsh conditions. Reproductive activity is triggered only by torrential rains; males call at night but may also call during the day if humidity levels are high. During the wet season, surface activity remains limited, occurring mostly for foraging. Over a four-year period, the reported minimum population size was 330, and the maximum was 595 (Quintero-Díaz and Vázquez-Díaz, 2009).
To build on these foundational observations, this study provides data derived from the first systematic, quantitative analysis of population structure and activity patterns. Conducted in 2005 within the rural community of Buenavista de Peñuelas, our research examined how habitat characteristics, temperature, humidity, qualitative variables, and seasonality influence occurrence and abundance of individuals. We also explored the relationship between demographic parameters (such as abundance, sex ratio, and age composition) and environmental factors to better understand spatial and temporal population dynamics. These novel data complement earlier descriptive publications and offer a more comprehensive understanding of the population ecology of the Upland Burrowing Treefrog with implications for conservation and management of this endangered species.
Materials and Methods
Study Area
We conducted the study in an area 16 km to the south of the capital of the state of Aguascalientes (21.7206° N, 102.3001° W; WGS84 datum) in the rural community of Buenavista de Peñuelas (Fig. 1). The area was characterized by the presence of natural grassland and xerophytic scrub vegetation. The region is semi-arid with a dry period from December to June (Fig. 2A) and a rainy season from July to November (Fig. 2B) (INEGI, 2008). In 2005, the nearest meteorological station recorded a total rainfall of 445.5 mm, a mean annual maximum temperature of 27.3 °C, and a mean annual minimum temperature of 8.8 °C (CONAGUA, 2022). The study area, covering 1.3 km2, was near temporary agricultural plots, rural houses, and a four-lane highway. A water reservoir had been constructed for livestock, but naturally occurring areas of temporary waterlogging were observed during the rainy season.
Citation: Journal of Herpetology 59, 3; 10.1670/23-036R3
Citation: Journal of Herpetology 59, 3; 10.1670/23-036R3
Field Work
The monitoring period extended from December 2004 to December 2005. Field surveys were conducted monthly, with each survey encompassing a walk by two persons along four established transects, each 400 m by 10 m. Transects T1 and T2 crossed natural grassland areas and were separated by a large floodable zone during the rainy season; T3 crossed several floodable zones behind the embankment of the pond El Jagüey and adjacent to temporary agricultural plots; finally, T4 ran alongside houses and was closest to human activities (Fig. 1C). Monitoring began at 1930 h in the winter and 2030 h in the summer. The following data were recorded for each encounter: time, geographic coordinates, activity, sex, SVL, weight, and quantitative and qualitative variables related to local climatic conditions. For individual identification, we applied a photo identification method as a capture-recapture technique (Sutherland, 1996). The dorsum of each frog was photographed using digital cameras (SONY DSC-P71 and NIKON COOLPIX 8800 VH) (Figs. 2C, D, E). Individual identity was determined by visually matching unique dorsal spot patterns across capture events. Photographs from each sampling occasion were compared with previously cataloged images. Comparisons focused on shape, size, number, and relative position of spots and other distinctive dorsal marks. This matching process was conducted manually by AEL to ensure consistency. Excreta were collected from individuals who defecated during handling, and a small number of individuals were also selected for stomach lavage. Variables related to local climatic conditions were relative humidity, ambient temperature, absence or presence of rain, degree of cloud cover (classified as 0/4, 1/4, 3/4, or 4/4), and wind intensity (classified as high, moderate, or zero). We used the minimum area method to define sampling units (Lot and Chiang, 1986) needed to characterize the tree-shrub and herbaceous stratum based on importance values in the community (Elzinga et al., 1998). To characterize soil, we took 450 cm3 samples from three random points in each transect.
Data Analysis
To estimate population size, we assumed a closed population due to habitat alteration of surrounding areas. We applied the Lincoln-Petersen method (N = M * n / R; Southwood and Henderson, 2000) to individuals located within transects where N represents the population size, M signifies the number of individuals captured and photographed initially, n is the number of captured individuals in later samples, and R refers to the number of recaptured individuals based on photo match (Southwood and Henderson, 2000). Remaining analyses included individuals observed both inside and outside tran-sects. All individuals were classified into age categories: adult and juvenile (J). Classification was based on SVL following species descriptions of Duellman (2001) and Quintero-Díaz and Vázquez-Díaz (2009). Among adults, we identified males (M) and females, further distinguishing nongravid females (F) from gravid females (FG) by the presence of vocal sacs in males and mature eggs in gravid females (Duellman and Trueb, 1994).
To assess whether abundance was influenced by temperature, relative humidity, or their interaction, we performed nonpara-metric local polynomial regression using the “loess” function from the “stats” package in R (RStudio Team, 2022). Analysis of abundance by age category was conducted both spatially (i.e., between transects, n = 69) and temporally (i.e., between months, n = 88). To test for significant differences in temperature and relative humidity preferences among categories of individual frogs, we compared kernel density estimates (violin plots) and their respective means using Student's t-test (n = 64), implemented with the “ggplot2” (Wickham, 2016) and “ggpubr” (Kassambara, 2020) packages. We also performed redundancy analysis to assess response of age and sex categories to categorical environmental parameters using the “vegan” package (Oksanen et al., 2022). Prior to these analyses, we tested for correlation, normality (“stats” package), homoscedasticity (“olsrr” package; Hebbali, 2020), and linearity (“lmtest” package; Zeileis and Hothorn, 2002). Additionally, we used bar graphs to visualize qualitative microclimate data (n = 88).
We identified invertebrate remains from excreta and stomach lavages following Borror et al. (1989). Frequency of occurrence (FO) of each food item was calculated as FO = (Fs / N) * 100, where Fs is the number of samples in which a food type occurs. Results are expressed as percentages. We identified plant species following de Rzedowski and Rzedowski (2001). The importance value (IVi) per stratum was calculated as IVi = RDi + Rfi + RCi, where RDi is relative density, Rfi is relative frequency, and the RCi is relative cover. IVi ranges from 0 to 3, indicating each species’ influence within the community, and results are presented as percentages. Soil analysis followed Robledo and Maldonado (1997) and the Mexican standard NOM-021-RECNAT-2000, evaluating texture, saturation percentage, pH (in water), electrical conductivity, organic matter, moisture retention, soil color, mulch layer, and bulk density.
Results
Abundance and Population Size
We recorded a total of 69 individuals within the four transects: 20 males, 27 females, and 22 juveniles. Among the 27 females, 9 were gravid. We made eight recaptures, determined by on photo identification; we recaptured two individuals twice each (Table 1); they had moved approximately 750 m from their initial capture points. The estimated population size was 595.1 ± 197.8 individuals, with adults comprising 87% and juveniles 13% (F: 30, FG: 9, M: 21, J: 9). We found a similar proportion when we included individuals observed outside the transects in counts (F: 38, FG: 13, M: 27, J: 10). Mean SVL of females (including gravid females) was 53.2 mm (range: 45.1–71.7 mm; n = 50), whereas males had a mean SVL of 50.1 mm (range: 47.8–61.8 mm; n = 20). The overall sex ratio (females:males) was 0.54, indicating a population biased toward females. Approximately 26% of females were gravid. Formation of oocytes appears to occur within a short period as we recorded their presence in two recaptured females only 16 days after the initial capture (Table 1). Additionally, we counted the number of developing oocytes in a gravid female killed on the adjacent highway. We counted a total of 2,988 oocytes: 1,094 in the left ovary and 1,894 in the right ovary. Each ovary was divided into 10 and 9 lobes, respectively, separated by thin membranes. The number of oocytes varied among lobes.
Spatial and Temporal Requirements
Most individuals (94%) were concentrated in transects T1–T3, and only juveniles were not found in T4. Female abundance was highest in T1, whereas males and juveniles were more abundant in T3 compared to other transects (Fig. 3A). Regarding temporal activity patterns of individuals, surface activity was recorded only during three months of the monitoring year, with overall abundance peaking in July, gravid females observed until August, and nonjuveniles found in September. Throughout all months, the number of females exceeded that of males and juveniles (Fig. 3B).
Citation: Journal of Herpetology 59, 3; 10.1670/23-036R3
Nonparametric regression analysis (span = 0.5, degree = 1) showed that both temperature and relative humidity significantly influenced monthly abundance. The full model including both predictors fit the data substantially better than models with only one predictor. Temperature had a significant effect when added to the model with relative humidity only (F1, 51.4 = 8.99, P < 0.0001), and relative humidity had a significant effect when added to a model with temperature only (F1, 51.4 = 16.62, P < 0.0001).
Regarding influence of temperature and relative humidity on age and sex categories, the global redundancy analysis (RDA) model was not statistically significant (F2, 61 = 1.29, P = 0.27). When examined individually, humidity showed a weak, non-significant trend (F1, 61 = 2.08, P = 0.11), whereas temperature had no explanatory effect (F1, 61 = 0.51, P = 0.69). Nevertheless, the variables acted in opposite directions along the first canonical axis (RDA1), with humidity strongly and positively associated (0.99) and temperature negatively associated (-0.79) with abundance although positively aligned with the second axis (RDA2, 0.61).
Pairwise comparisons among age and sex categories of relative humidity values at capture sites showed that gravid females occurred at significantly lower mean humidity (mean ± standard deviation = 66.8 ± 6.8%) than nongravid females (73.6±8.7%) and males (74.7 ± 7.9%) (t30.4 = 2.616, P < 0.05; t28.01 =-2.985, P < 0.001, respectively). In contrast, comparisons of temperature preferences revealed no significant differences among age categories (all P > 0.1). Gravid females occurred at the highest mean temperature (20.7 °C), whereas males had the lowest (18.9 °C). Additionally, the activity of gravid females occurred within a narrower humidity range (FG: 51–77%) compared to nongravid females (F: 20–82%) (Fig. 4B). Juveniles and males were recorded at the lowest temperatures, but males exhibited the widest temperature range (15–26 °C). The maximum temperature recorded for active individuals was 26 °C, observed in adults (Fig. 4A).
Citation: Journal of Herpetology 59, 3; 10.1670/23-036R3
In analysis of qualitative environmental variables and activity, no individuals were observed during periods of high wind, clear skies, or rain. Highest activity was recorded under moderate or low wind conditions, with at least 3/4 cloud cover, and absence of rain (Fig. 5).
Citation: Journal of Herpetology 59, 3; 10.1670/23-036R3
Diet Characterization
The diet of the observed S. dentata consisted of insects from the orders Coleoptera (49.9%), Hymenoptera (38.5%), Hemiptera (7.7%), and Lepidoptera (3.8%). Scarabaeidae (38.5%; in Coleoptera) and Formicidae (in Hymenoptera) were the most frequently consumed families. Within Scarabaeidae, the subfamily Melolonthinae was present in all samples, whereas Aphodiinae appeared less frequently; identified genera included Diplotaxis and Aphodius, respectively. For Coleoptera, we also recorded remains from the sub-families Carabinae (Carabidae), Cincindelinae (Caraboidea), and Chauliognathinae (Cantharidae), including one individual identified as Chaulignathus. The families Lygaeidae and Eriocraniidae were the only ones recorded for Hemiptera and Lepidoptera, respectively (Fig. 6A).
Citation: Journal of Herpetology 59, 3; 10.1670/23-036R3
Habitat Characterization
We documented 10 plant orders, 16 families, 41 genera, and 48 species. Lamiales had the highest family diversity (4), while Poales (14) and Asterales (10) had the highest species richness (see Table S1). Within the shrub-tree stratum, Mimosa monancistra had the highest importance value (40.3%), followed by Neltuma laevigata (17%), Vachellia farnesiana (16.8%), Vachellia schaffneri (9.4%), and unidentified species of Opuntia spp. (6.7%) (Fig. 6B). In the herbaceous stratum, Poales dominated (75.3%), with Bouteloua curtipendula (35.7%) and Microchloa kunthii (20.8%) most representative. Other orders contributed smaller proportions: Asterales (1.7%), Fabales and Lamiales (2.3%), Malpighiales (7.4%), Malvales (0.8%), and Solanales (5.1%) (Fig. 6C).
Soil bulk density ranged from 1.2 to 1.4 g/cm3. Texture was predominantly loamy in T1 and T2 and clayey in T3 and T4. Soil was moderately acidic across all sites, with low electrical conductivity (0.4–1.1 dS/m). T2 had the lowest saturation (22.7%) and T3 the highest (34%). Field capacity peaked in T4 (38.1%). Permanent wilting point values were lowest in T1 and T2, increasing by up to seven more units in T3 and T4. Organic matter content was high in all transects except T1. The mulch layer was thickest in T4 (120.6 g/m2) and thinnest in T3 (8.2 g/m2). The soil color was darker when wet (see Table S2).
Discussion
Long-term data on abundance and population size are essential to detect population trends (Blaustein et al., 2001; Semlitsch, 2003; Blackwell et al., 2004) and assess conservation risk (Cushman, 2006). Among annual estimates made between 2005 and 2008 (Quintero-Díaz and Vázquez-Díaz, 2009), the highest population size estimate was recorded in 2005 (595.1 ± 197.8 individuals, this study). Years later, G. E. Quintero-Díaz (Gaceta Universitaria, 2022) reported that 1,400 individuals were recorded in the study area in 2010, a result he attributed to dedicated conservation initiatives supported by federal government resources between 2008 and 2010. By 2020, however, only about 60 individuals were observed, a decline that may be associated with urbanization driven by the establishment of international automobile companies, which led to residential housing, commercial centers, and supporting industries catering to the needs of these new economic activities (Gaceta Universitaria, 2022).
Since 2005, the sex ratio has deviated from the expected 1:1 ratio commonly observed in amphibians (Martof, 1956; Duellman and Trueb, 1994; Nakamura, 2009). From 2005 to 2008, Quintero-Díaz and Vázquez-Díaz (2009) reported a sex ratio of 63% females and 37% males. Skewed recruitment toward either sex has been associated with environmental and genetic factors (Nakamura, 2009) and may affect sexual selection (Trivers, 1972), genetic diversity (Wright, 1938; Cole, 1954; Blackwell et al., 2006), and ultimately, population dynamics including reproduction and survival rates (Schmeller and Merilä, 2007).
The gradual decline in abundance during the activity season is likely associated with the gradual decrease in precipitation and temperature (see Fig. S1). Previous studies have shown that emergence and surface activity in temperate amphibians is influenced by temperature (e.g., Bufo bufo: Reading, 2003; Anaxyrus fowleri: Green et al., 2016) and seasonal precipitation (e.g., Anaxyrus retiformis, Gastrophryne olivacea, Smilisca fodiens; Sullivan et al., 1996; Sullivan and Fernandez, 1999). Amphibian activity outside burrows is also affected by soil moisture (Seebacher and Alford, 1999), ambient humidity (Penman et al., 2006), and wind intensity (Seebacher and Alford, 1999; Pough et al., 2004; Penman et al., 2006). Also, reproductive activity appears to be influenced by a combination of environmental and social factors. Male activity often responds to rainfall and the chorus of conspecifics, whereas female behavioral patterns appear to be exclusively associated with social factors, such as conspecific calls (Höbel, 2017; O’Brien et al., 2021). Our analysis highlights clear differences in microclimatic preferences among age and sex categories of S. dentata. Males were active across a broader temperature range and showed a positive correlation with relative humidity, and gravid females were active within narrower temperature and humidity ranges. Nongravid females exhibited the widest humidity tolerance, suggesting greater flexibility in their activity patterns. Juveniles and males tolerated lower temperature limits compared to females. Individual activity peaked under moderate or minimal wind, substantial cloud cover, and lack of rain. In general, we found that individuals restricted activity to narrower values of temperature and humidity. Notably, individuals of all age categories preferred moderate temperatures, and no individuals were observed at temperatures above 26 °C. Conversely, activity increased with higher relative humidity.
Environmental constraints on activity outside burrows should be carefully monitored given the projected increases in temperature and decreases in precipitation in the near future as reported by multiple climate research centers through the Intergovernmental Panel on Climate Change (IPCC, 2023). Moreover, Encarnación-Luévano et al. (2024) found that the area historically occupied by S. dentata is projected to experience, by 2070 and under an intermediate greenhouse gas emission scenario (RCP 4.5), an increase in temperature of at least 3 °C and a reduction in precipitation of approximately 60 mm relative to the baseline period. As a result, the optimal conditions for peak activity of S. dentata will progressively disappear throughout the next several decades within their historically known range. Even under the most optimistic scenario, suitable conditions for surface activity outside burrows are projected to persist only during a single month, August.
The reproductive pattern for S. dentata resembles that of desert anurans with explosive breeding, such as S. fodiens in the Sonoran Desert, as well as Anaxyrus cognatus, A. punctatus, Incilius alvarius, and Scaphiopus couchii (Sullivan et al., 1996; Sullivan and Fernandez, 1999; Esparza-Orozco et al., 2020). This pattern is characterized by reproduction triggered by intense but short rainfall events. However, despite evidence of developed oocytes in females, we observed no breeding in 2005, suggesting that successful reproduction depends on reaching specific rainfall thresholds. We observed even Hypopachus variolosus, also highly seasonal anurans that share temporary ponds with S. dentata, in amplexus during the same year. Higher total precipitation appears necessary to trigger mating and oviposition: in 2004, we recorded oviposition, tadpoles, and meta-morphs, coinciding with rainfall exceeding 200 mm in July and remaining above 100 mm through September.
The population structure observed aligns with the high natal fidelity typical of anurans, which often remain within 1 km of breeding sites (Sinsch, 1990). The largest movements we recorded, involving two individuals, likely correspond to nearby inundated areas where tadpoles and metamorphs had been previously observed (Quintero-Díaz, personal observation). We detected no spatial segregation by age or sex, suggesting a potential metapopulation dynamic where most juveniles stay near natal ponds until some disperse while adults generally remain within 100–200 m (Semlitsch, 2003). This type of population structure is particularly vulnerable to habitat loss resulting from expansion of agricultural and urban frontiers. Such processes can lead to smaller, more isolated populations and a subsequent loss of genetic diversity (Hamer and McDonnell, 2008; Arntzen et al., 2017).
Urbanization and its indirect effects impact amphibian diversity more strongly than grazing, deforestation, or agriculture alone (Cordier et al., 2021). In this study, we confirmed that areas modified for anthropogenic activity reduce or hinder the presence of individuals. Proximity to houses, paddocks, and public areas in the rural community Buenavista de Peñuelas reduced abundance, probably due to scattered waste, deforestation, campfires, and livestock presence. Furthermore, although we did not observe reproductive activity in 2005, it has been reported that no individuals or tadpoles make use of the community's livestock watering pond (El Jagüey) (Quintero-Díaz and Vázquez-Díaz, 2009), possibly due to unsuitable biological and physicochemical conditions for egg and tadpole development (Carr and Fahrig, 2001; Cushman, 2006; McKinney, 2006). On the contrary, temporary ponds are critical breeding sites for individuals; even their proximity to croplands along the largest floodable transect (T3) did not seem to threaten the presence of individuals during the study period. However, edge effects from agricultural areas are known to alter microclimates and could indirectly affect population dynamics (Harper et al., 2005); therefore, it would be important to study these effects over time.
Most amphibians feed on invertebrates (Santana et al., 2019). However, anurans often show a preference for insects (Dorcas and Gibbons, 2011). Diet analysis showed that S. dentata fed exclusively on insects, a result which aligns with observations in northern populations of the closely related S. fodiens (Winter et al., 2007). Notably, remains of Coleoptera and Hymenoptera were most frequently found in both species, an occurrence widely reported in the diets of frogs and toads (Quiroga et al., 2009; Crnobrnja-Isoilavic et al., 2012; Moser et al., 2017; Ofori et al., 2021). This pattern has been described as a convergent niche element among anurans (Ofori et al., 2021). Despite these similarities, some differences were noted: S. fodiens also consume Orthoptera and Odonata, whereas S. dentata include Hemiptera in their diet. Additional and more extensive analyses are needed to confirm these findings as adult diets may vary seasonally or spatially, particularly in species with distinct breeding and foraging areas (Menin et al., 2015).
The study area falls within the potential distribution of natural grasslands (INEGI, 2008), characterized by a shrub-tree stratum (de la Cerda-Lemus, 2008). Prior to 2005, natural grasslands to the east and south of the Valle de Aguascalientes region had already been undergoing extensive degradation due to agriculture and livestock expansion. In our 2005 survey, we recorded two grass species indicative of soil disturbance: Aristida adscensionis and Bouteloua chondrosioides. Furthermore, we found that B. curtipendula was more abundant than B. gracilis, although the opposite would be expected since B. gracilis is the dominant species in undisturbed natural grasslands; however, it is highly preferred by cattle for grazing (de la Cerda-Lemus, 2008). Additionally, the observed dominance of the smallest spiny shrub, Mimosa monancistra, over the mesquite (Neltuma laevigata) and acacias (also known as huisaches, Vachellia spp.) is considered a sign of disturbance. In an undisturbed natural grassland, the typical structure of Fabaceae-dominated vegetation is characterized by mesquites as the main woody component, followed by acacias, and finally the smaller spiny shrubs, such as mimosas (also known as catclaw) (Herrera-Arreola et al., 2007). During monitoring, we frequently observed clandestine burning of woody parts of trees and shrubs to facilitate movement of cattle and people. Agriculture and livestock likely influence distribution and abundance of fossorial herpetofauna (Wong et al., 2021), although the magnitude of their impact on population size remains unclear.
Past studies as well as this one have reported recordings of S. dentata in areas with phaeozem, planosol, and xerosol soils, which are generally flat, rich in organic matter, and minimally rocky (INEGI, 2008). These soils are firm and friable when dry but become malleable when wet (FIRA, 1987), making them suitable for burrowing. Local variations in organic matter and soil pH acidification may correlate with lower individual abundance. Acidification alters soil structure by disrupting essential element balances (e.g., iron, calcium, potassium; Porta Casanellas et al., 1999). Understanding how human activities affect these soil properties is critical for assessing their impact on the persistence of amphibian populations. Because burrowing anurans rely heavily on soil type and structure for shelter, their abundance and distribution are closely tied to these factors (Carisio et al., 2014; Wong et al., 2021).
Human activities are recognized as major drivers of biodiversity loss (Gardner et al., 2007; Hof et al., 2011), with documented evidence of disproportionate impacts on amphibians (Nori et al., 2015; Cordier et al., 2021), particularly those with specialized habitats and life cycles (Futuyma and Moreno, 1988; Badillo-Saldaña et al., 2016). Despite conservation actions undertaken by local communities, researchers, academics, and government agencies to protect S. dentata, habitat fragmentation and land-use change continue to threaten viability of their populations.
Despite the absence of field-based monitoring across most of this species’ historical range, available data from Buenavista de Peñuelas indicate that habitat disturbances can lead to substantial reductions in population size and abundance. The primary challenge confronting this species is the imminent threat to its habitat, which is among the most endangered due to expanding urban and agricultural frontiers of the region. Although additional field studies are needed to clearly establish this relationship throughout the species’ distribution, current estimates of habitat loss cannot be underestimated. Prior to major land-use changes of the 19th century, the potential distribution of S. dentata, based on their climatic requirements and association with natural grassland and cracicaule scrub habitats, was about 6,269 km2. By 2018, however, 472 km2 remained in a primary state of conservation, and even this area was highly fragmented and discontinuous due to ongoing urban and agricultural expansion (Encarnación-Luévano et al., 2025).
Our results contribute to understanding Upland Burrowing Treefrogs’ ecological requirements and highlighting their vulnerability. Future work should update the 2005 baseline data we present here with more recent surveys and consistent field monitoring. We recommend studies on climate resilience of individuals to assess their vulnerability to climate change. In addition, we emphasize the need to conserve as much intact and connected habitat as possible in the region.
Aguascalientes is located in north central Mexico (A), and the study area is situated at the southern edge of the municipality of Aguascalientes (red circle) (B) within communal lands of the rural community of Buenavista de Peñuelas (C). Four transects for population monitoring (yellow lines) were established within the study area (C). Satellite image: 27 July 2008, © 2022 Maxar Technologies (Google Earth).
Seasonal landscape within study area: dry season (December–June; A) and wet season (July–November; B). Activity was recorded only from July to September in 2005. To identify individuals in subsequent recaptures, we photographed their unique dorsal spot patterns (C–E).
Abundance by age category across transects (A) and by month during the monitoring period (B). Total individuals recorded in transects, n = 69; total active individuals observed during monitoring, n = 88. Age categories: nongravid females (F: pink), gravid females (FG: green), males (M: blue), and juveniles (J: purple).
Temperature (A) and relative humidity (B) measured for different age categories: nongravid females (F: pink), gravid females (FG: green), juveniles (J: purple), and males (M: blue). Violin plots illustrate distribution of measurements, with thick black lines indicating mean values. Significant differences in humidity were observed between F and FG (P < 0.05) and FG and M (P < 0.01); sample size: n = 64.
Number of individuals recorded (blue bars) under different combinations of weather conditions: rain (absent/present), wind intensity (high, moderate, zero), and cloud cover (0/4, 1/4, 3/4, and 4/4).
Diet composition and vegetation. Diet items are shown by family (%), grouped by order: Coleoptera (Co), Hemiptera (He), Hymenoptera (Hi), and Lepidoptera (Le) (A). Tree-shrub stratum by species (%) (B). Herbaceous stratum by family (%), grouped by order: Asterales (As), Fabales (Fa), Lamiales (La), Malvales (Ma), Malpighiales (Malp), Poales (Po), and Solanales (So) (C).
Contributor Notes
