Urban Facilitation of Reproductive Biology and Body Size in Invasive Boa constrictor on Aruba
Anthropogenic changes to ecosystems have dramatically altered environmental processes such that native landscapes may be more susceptible to invasive species due to urbanization. This may particularly be the case on islands, where anthropogenic habitat alterations may alter available niche space that allows invasive species colonization and expansion. On Aruba, a small desert island with limited natural surface water, we examined how reproductive development, body size, body condition, and prevalence of pentastome parasites vary in the invasive Boa Constrictor (or simply, “Boa,” Boa constrictor) between natural and urbanized habitats to test whether urbanization may alter the success of this species. Female B. constrictor from urban habitats had greater body mass, greater snout–vent lengths, greater ovary mass, and more ovarian follicles than those from natural habitats. Reproductive females were only found in urbanized habitats. Male body mass, body condition index, and testes mass also showed positive relationships with the number of buildings within their estimated activity range. Additionally, we compared the Boa anatomical dataset to recent prey consumption in the snakes we necropsied, which demonstrated a correlation between prey consumption and urbanization. Though limited by sample size of recently consumed prey items, large, endothermic, non-native prey (e.g., Gallus domesticus, Rattus rattus, and Columba livia domestica) were only found in B. constrictor living in urbanized habitats; very few Boas in the natural ecosystem were found to have recently consumed prey, and of those, only small, native prey species were consumed by Boas. This study provides evidence that urbanized habitats may facilitate growth and reproduction in this large-bodied invasive snake and suggests that urban prey availability may be an important resource subsidy.
ANTHROPOGENIC land-use change has rapidly degraded native ecosystems by altering both basic abiotic and biotic factors (Rizwan et al., 2008). Urbanization of natural habitats leads to more impervious surfaces and alters hydrology, vegetation cover, chemical composition of surrounding soil, nutrient cycling, and ultimately biodiversity (e.g., Poff et al., 1997; Pickett et al., 2001; Duh et al., 2008). Urbanized habitats possess significantly altered ecological communities (Piano et al., 2017) with the reduction or loss of sensitive native species (Pauw and Louw, 2012) and the bolstering of invasive species populations (Riley et al., 2005; Rodda and Tyrrell, 2008; Jesse et al., 2018; Clements et al., 2019). Urban areas drive survival of many invasive species because invasive species tend to be generalists, adapt quickly, and in some cases are pre-adapted to urban areas (Cadotte et al., 2017). Moreover, invasive species in urban areas encounter reduced competition in the presence of anthropogenically subsidized resources (Cadotte et al., 2017).
Some of the most dramatic examples of invasive species population facilitation by urbanization are in ecosystems where anthropogenic modifications cause an increase in a naturally limited abiotic resource, such as the addition of anthropogenic water to desert ecosystems. In the Sonoran Desert, native desert floral communities have been altered due to irrigation and human-dwelling sources of water from air conditioning condensation (Hope et al., 2006). In that resulting vegetation change, the desert was modified from naturally sparse and open habitat structure to more shaded habitats with broad-leaved trees and greater three-dimensional habitat heterogeneity (Hope et al., 2006). In South Africa, urban development caused the transition of native xeric grasslands into canopied forests, which, in turn caused an increase in species richness of invasive vertebrates, mostly small mammals (Potgieter et al., 2020). On Guam, the rapid expansion of secondary-growth forests following World War II was one likely mechanism that facilitated the rapid invasion of the island by the Brown Treesnake (Boiga irregularis; Rodda et al., 1999a). Paradoxically on Guam, though the early successional forests in the middle 20th century may have driven rapid expansion of the invasive snake, currently urban areas function as refugia for native species that are susceptible to predation by Brown Treesnakes (Rodda et al., 1999a). Thus, the temporal contingency of both habitat alteration and invasive species expansion is fundamental to the biology of invasive species.
Urbanization is also known to facilitate native snake populations through human-driven augmentation of prey resources. The Western Diamondback Rattlesnake (Crotalus atrox) responds positively to disturbance due to increased rodent populations surrounding cities (Sullivan et al., 2017), and agricultural areas create high-density rodent habitats that support unnaturally large populations of Habu (Trimeresurus flavoviridis) in the Ryukyu Islands of Japan (Mishima et al., 1999). Hauptfleisch et al. (2021) found that Namibian snakes were more often observed in affluent areas of a city which provided greater water access through irrigation.
Island ecosystems are among the most globally threatened and yet they contain disproportionately high levels of endemic biodiversity (Kier et al., 2009). Due to small surface area, islands face intense development pressure to convert natural ecosystems into modified urban zones. The Boa Constrictor (or simply, “Boa”), Boa constrictor, rapidly expanded across the small xeric island of Aruba, having been introduced in the mid-1990s from just a few individuals that were likely released pets (Bushar et al., 2015). Unlike the typical tropical forests where B. constrictor are primarily found in their native range, the dry thorn scrub ecosystem of Aruba (Reinert et al., 2008) is hypothesized to limit the species’ success (Reinert et al., 2021) and generally lacks large canopy trees and natural sources of freshwater. Relatively low prey and water availability are thought to limit B. constrictor in natural Aruban habitat (Reinert et al., 2021). However, B. constrictor successfully use the natural–urban interface and have been observed to have visibly higher levels of abdominal fat when found in close proximity to human dwellings (Quick et al., 2005). Similarly, invasive Burmese Pythons (Python molurus) in the Florida everglades have been found to prefer anthropogenic-edge habitats (Walters et al., 2016; Bartoszek et al., 2021).
A pre-existent, extensive eradication program run by the Aruban legal and wildlife authorities provided a unique opportunity for us to study a large multi-year sample of invasive B. constrictor from across the island. We were interested in testing whether urban habitats facilitate (i.e., allow population expansion and/or maintain a robust population size) this invasive snake across this desertified insular ecosystem; if urban areas facilitate Boas on Aruba, Boas collected from nearer to human structures would receive a measurable benefit. We examined 394 Boas to test the hypothesis that reproductive characteristics, among other morphometric parameters, of B. constrictor were affected by habitat urbanization. Specifically, if urban habitats facilitated reproduction in Boas, we predicted that we would see increased parameters of reproductive output (as described below) in Boas from more urbanized areas relative to those from natural areas. We also secondarily examined the role of recent prey consumption in necropsied snakes across a range of urbanization levels as a potential explanatory mechanism of facilitation in urbanized environments.
MATERIALS AND METHODS
Specimens.—
Boa constrictor was first observed on Aruba in 1999, and it was common and widespread by 2003 (Quick et al., 2005). For the past 24 years, there has been a partnership program between Aruban government and non-government entities (Aruba Veterinaire Dienst and Directie Landbouw, Veeteelt, Visseru, en Markthallen [LVV]) and the Fundacion Parke Nacional Arikok [FPNA]) designed to reduce the Boa population on the island. As part of that program, the staff of FPNA have regularly conducted mostly diurnal patrols of the 34 km2 national park and captured any Boas encountered. In addition, park staff have collected and removed Boas reported from anywhere across the island. The staff of FPNA immediately euthanized all Boas that they collected, which was overseen by the Aruba Veterinaire Dienst. The 394 Boas examined for this study were collected between December 2016 and March 2023. The collection date and geographic location (if known) of each snake was recorded, and the specimen was frozen for later evaluation.
Necropsies.—
Necropsies were performed in a standardized way for all snakes in this study after thawing, which were previously euthanized by the Aruban Boa management program. Snout–vent length (SVL) and total length (TL) were recorded by loosely extending the snake along a flat surface parallel to an outstretched tape measure (Blouin-Demers, 2003). Body mass was recorded using Pesola spring scales. Each snake was incised along the mid-ventral body wall by inserting scissors through the anterior border of the cloaca and cutting through the skin and ventral abdominal muscles anteriorly to the trachea. This dissection exposed the entirety of the coelomic cavity allowing for the determination of sex and an assessment of the gonads, abdominal fat bodies, stomach contents, and the presence of Pentastomida parasites in the lungs, which were morphologically consistent with Raillietiella orientalis.
Both the right and left gonads were removed and massed. We combined these two measures to generate the total gonad mass. If ovarian follicles were discernible, the total number of follicles contained in each ovary was recorded. Abdominal fat was scored on a 0–3 relative scale (Reinert et al., 2021); 0 indicated a snake with no visible abdominal fat, 1 indicated a snake with small and generally discontinuous abdominal fat, 2 indicated a snake with moderately sized continuous abdominal fat, and 3 indicated a snake with very large abdominal fat deposits that were continuous throughout the coelom and visibly distended the body cavity. To ensure continuity in the assessment of all anatomical parameters, the same investigator (JMG) was present during every necropsy.
Following reproductive and fat score assessment, the entirety of the lung (both respiratory and non-respiratory regions) was carefully dissected at the ventral midline starting at the trachea and cutting posteriorly to the end of the non-vascular air sac of the right lung. Once the lung was opened, we carefully examined the entire inner surface of the lung for the presence of parasitic Pentastomida. Although the dominant right lung was our main focus, we also examined the left lung for these parasites. The total number of pentastomes was recorded. From the data collected, we calculated a gonadosomatic index (GSI, = [total gonad mass/body mass] X 100) and a body condition index (BCI = standardized residuals from sex-specific regressions using the natural log of mass and natural log of SVL).
Lastly, prey items contained in the stomach were identified to the lowest taxonomic unit possible. We did not examine chyme or fecal material present in the alimentary canal below the stomach, representing prey items consumed within the prior three days, as estimated by gastric emptying rates for the species (Milken et al., 2020). Consequently, the prey items recorded for this study represented recently ingested prey.
Activity range estimation.—
To estimate activity range size for subsequent spatial analyses of urbanization level in which each snake lived, 31 Boas had been radio-tracked as part of a separate study during 2006 and 2007 in habitat surrounding FPNA (HKR, unpubl.). Transmitters were surgically implanted (Reinert and Cundall, 1982; Reinert, 1992), and each snake was located approximately 2–3 times per week for periods of up to 507 days. This resulted in a mean minimum convex polygon activity range estimate of 20 ha (SE = 4.8 ha, n = 31) for all tracked Boas. Resulting from the observed mean 20 ha activity ranges, a circular 20 ha activity range was then applied to each necropsied snake for GIS modeling. Because snakes were euthanized previously by Aruban authorities, we could not know the shape or size of each animal’s home range; thus, we modeled necropsied snakes as using a 250 m circular area (which equals 20 ha) that radiated around the point at which each snake was found. While this assumption of space usage may not equally apply to every specimen in the study, a standardized estimate of nearby habitat type was necessary to test the urbanization hypothesis.
GIS analyses of urbanization level.—
We performed GIS analyses in ArcGIS Online ver. 10.8.21 (ESRI, 2022) to quantify the area equal to a 250 m radius from each snake’s detected location. This area was constructed in ArcMap around the GPS coordinates taken at the point of each Boa capture. A GIS layer of buildings on Aruba was added to the map (from https://data.nextgis.com/en/region/AW/msbld/), and the number of buildings within each estimated Boa activity range was determined. If the area contained no buildings, the snake was categorized as occupying a “natural” habitat. If there were any buildings located within the area, the snake was classified as occupying an “urbanized” habitat, and the number of buildings in the prescribed area was recorded. We utilized two different statistical approaches (described below) that treated urbanization both as a discrete parameter (i.e., tested for differences in snakes between natural and urban habitats) as well as a continuous parameter using number of buildings within the 250 m radius surrounding each snake’s location. Buildings were counted as integers by which if a roof partly touched the 250 m radius, it was considered a single building within the snake’s range. Figure 1 shows a map of habitat type distribution on Aruba based on this classification.
Citation: Ichthyology & Herpetology 113, 4; 10.1643/h2024099
Statistical analyses.—
The proportions of females and males in natural and urbanized habitats were compared using Z-tests. The sex ratio of snakes was compared to equality using a Chi-squared goodness of fit test. We used two-way analysis of variance (ANOVA) with replication to examine the main effects of habitat (natural and urbanized) and sex (female and male) and the interaction effect of habitat and sex on each of the following variables: body mass, SVL, BCI, GSI, and number of lung pentastomes. The sex-specific variables of ovary mass, testes mass, number of follicles, and fat scores were compared between urbanized and natural habitats using one-way ANOVAs separately for each sex. Because these same variables were also mass-related we applied one-way ANCOVAs with body mass as a covariate to determine the effect of adjusting for mass differences in snakes between the two habitat types. All ANOVAs were performed using Type III sequential sums of squares due to the unbalanced nature of our data. We applied Levene’s tests to examine the assumption of equality of within group variances. Levene’s tests indicated that the group variances for raw data were heteroscedastic for most measured variables. To equalize variances, continuous variables were log-transformed and discrete count variables were square-root transformed prior to all ANOVAs and ANCOVAs. We applied Pearson product-moment correlations to examine the relationship between the number of buildings as a continuous variable within the estimated activity range of each snake and SVL, body mass, BCI, gonad mass, GSI, fat score, number of follicles, and number of lung pentastomes. Correlations were performed separately for females and males. To examine the frequency of prey in the stomach of sampled snakes, we used a 2 × 2 × 2 Chi-square test for mutual independence. All statistical analyses followed Sokal and Rohlf (2012) or Zar (2010) and were conducted in R ver. 4.1.2 (R Core Team, 2021). Figures were made using the ggplot2 and ggbreak packages in R (Wickham, 2016; Xu et al., 2021).
RESULTS
We necropsied a total of 394 B. constrictor (226 females, 155 males, and 13 juvenile snakes unidentifiable to sex) that were collected between December 2016 and March 2023. The smallest female to possess ovaries greater than 1 g mass with clearly discernable follicles indicative of the onset of reproductive maturation was 88.5 cm SVL. Because we only conducted gross anatomic assessments, we cannot be sure of the developmental or endocrine state of females at this size; however, all females below 88.5 cm SVL had small, undeveloped ovaries less than 1 g. In females greater than 88.5 cm SVL, we consistently found larger ovaries that were rich with Graffian follicles consistent with reproductive development (Bertocchi et al., 2021). The smallest male Boa to possess measurable and well-developed testes greater than 1 g was 75.0 cm SVL. Again, while we cannot be certain of the developmental or endocrine state of testes above or below that size, all male snakes smaller than 75.0 cm SVL had small and visibly undeveloped testes. These results suggest 88.5 cm SVL as the minimum size at maturation for female Boas and 75.0 cm SVL as the minimum size at maturation for male Boas on Aruba. These sizes at which we observed gonadal maturation in both female and male snakes was much smaller than sizes at maturity that have been observed for other populations of Boa constrictor (163 cm and 173 cm SVL for female Boa constrictor occidentalis from two different populations and 146 cm and 149 cm SVL for males from the two populations as noted for females; Cardoza and Chiaraviglio, 2011).
Of the 394 B. constrictor examined, 198 had sufficiently precise geographic information for subsequent spatial analyses. Of these, 24 (9 females, 15 males) were from natural areas having no human structures within a 250 m radius and 174 (115 females, 57 males) were from urbanized areas having at least one human structure within the estimated 250 m activity range radius (Table 1). The proportion of males and females differed between samples from natural and urbanized areas (Table 1). Females occurred in a higher proportion than males in urbanized areas (Z2 = 2.71, P = 0.007) and in a lower proportion than males in natural areas (Z2 = 2.84, P = 0.009). The sex ratio of snakes sampled from urbanized areas was not equal (χ21 = 18.88, P < 0.0001), but was actually 2:1 biased toward females (Table 1).
Two-way ANOVAs comparing males and females in natural and urbanized habitats for mass, SVL, BCI, GSI, and number of lung pentastomes showed differences only for mass and SVL (Table 2). The interaction effect in the analyses of both mass and SVL suggested that males and females show important differences between natural and urbanized habitats. Both the mass and SVL of females exhibited larger increases in urbanized habitats than they did in males (Fig. 2A, B; Table 3).
Citation: Ichthyology & Herpetology 113, 4; 10.1643/h2024099
Because of mass-related and sex-specific nature of ovary mass, testes mass, number of follicles, and fat score, these variables were compared between natural and urbanized habitats separately for males and females using both one-way ANOVAs and ANCOVAs adjusting for body mass (Table 4). In females, the unadjusted ovary mass and number of follicles were both greater in urbanized habitats than in natural habitats (Fig. 2C, Table 3); however, adjusted ovary mass and follicle number did not differ between habitats. This suggests that the difference in both characteristics was a reflection of the greater mass of the female snakes collected from urbanized habitats (Fig. 2A). None of the other variables examined differed between urbanized and natural habitats (Table 4).
Correlation analyses indicated few significant relationships between the variables examined and the number of buildings found within activity ranges of B. constrictor (Table 5). Only male mass showed a rather strong positive relationship (r73 = 0.39) with the number of buildings in their activity range. In addition, the gonad mass of males and the BCI of both males and females also generally increased with the number of buildings (Table 5).
A total of 34 sampled B. constrictor had identifiable prey items in their stomachs, and only three of these snakes were from natural habitats (1 male and 2 females). Most snakes containing prey (31) were collected from urbanized habitat (5 males and 26 females). A Chi-square test for mutual independence examining prey occurrence by sex and habitat (2 × 2 × 2) indicated the existence of a three-way interaction (χ24 = 14.85, P = 0.005). A comparison of the frequency of prey occurrence in males and females illustrated that a higher than expected frequency of females and a lower than expected frequency of males contained prey in urbanized habitats, while there was a lower than expected frequency of females and a higher than expected frequency of males that did not contain prey in natural habitats (Fig. 3). The three snakes from natural habitats had all eaten small native lizards (Ameiva bifrontata and Cnemidophorus arubensis) or, in one case, a native passerine bird (Mimus gilvus); however, ten snakes from urbanized habitats (32%) had eaten large, non-native prey including five chickens (Gallus gallus domesticus), three rats (Rattus rattus), and one pigeon (Columba livia domestica).
Citation: Ichthyology & Herpetology 113, 4; 10.1643/h2024099
DISCUSSION
Establishment of invasive species requires that they have high reproductive rates and competitive advantages over native species (Sakai et al., 2001). On Aruba, we found significant effects of urbanization on reproductive condition and size of invasive Boa constrictor. In particular, female B. constrictor benefited morphologically, and likely physiologically, in urbanized habitats. Females from urbanized habitats grew larger in mass and length, had larger ovaries, and produced more follicles relative to those found in natural habitats. Consequently, they were potentially capable of producing more numerous offspring. Our sampling found that female Boas with very large ovaries and/or ovarian follicles (i.e., those containing well-developed secondary oocytes) on Aruba were exclusively in urbanized habitats (Fig. 2C). While we were not able to directly measure fitness of females from urbanized habitats, the gonadal differences of females between urban and natural habitats have direct fitness consequences, as ovarian and follicular development are necessary steps to oocyte (“egg”) and thus embryo/offspring production. In prior research of Boas on Aruba, Boas were found to commonly feed on larger non-native prey which are often associated with humans, including chickens, rats, pigeons, rabbits, cats, dogs, large iguanas, and even small goats (Quick et al., 2005; Reinert et al., 2021). These prey, most of which are non-native, are mostly prevalent in urbanized habitats and subsidize the diet of the population of B. constrictor, which was reinforced by our data of recent prey consumption analyzed by habitat type. A similar relationship is responsible for the successful invasion of Boiga irregularis on the island of Guam (McCoid, 1999; Rodda et al., 1999b). In addition to prey, sources of water are more prevalent in urbanized areas, which may also contribute to the increased growth and improved physiological conditions. These water sources include rain catchment cisterns, catchment reservoirs, animal watering systems, and even leaking pipes and hoses. Natural areas on Aruba are comprised of very open dry thorn-scrub and desert habitat containing no natural permanent water sources and receive an average of only 409 mm of rainfall annually (Reinert et al., 2021).
Male B. constrictor do not show the same marked increases in reproductive characteristics with respect to urbanized habitat. However, it should be noted that there was a strong positive correlation between the number of buildings and the mass of male snakes. Because male B. constrictor reach sexual maturity at smaller sizes, male snakes may not be as severely limited by prey in natural habitats. If females eat at similar frequency in both urbanized and natural habitats, then prey size and quality must be considered to be an important driver of female B. constrictor attaining reproductive maturity only in urbanized areas.
Our findings suggest a likely mechanism for a source–sink dynamic between urban and natural habitats by which a large invasive predator may only persist in natural areas due to an urban source population that likely receives a resource subsidy from humans. The naturally xeric Aruban landscape may not provide enough prey biomass to support reproduction by the large-bodied B. constrictor. The most frequent source of natural mortality for the small endemic Aruba Island Rattlesnake (Crotalus durissus unicolor) was starvation (Reinert et al., 2008), further indicating the biological effect of limited prey resources for snakes in the natural desert habitat of Aruba. Given that an endemic, small-bodied snake frequently succumbs to starvation, it is likely that a much larger and non-native snake would experience either similarly high mortality or have severe reproductive limitations in natural habitats (especially for large females that need sufficient resources to grow embryos to full term). Furthermore, the male-biased sex ratio in natural areas may indicate that large-bodied reproductive females with higher energetic demands (Van Dyke and Beaupre, 2011; Lima-Santos et al., 2021) experience increased mortality in the natural habitats or migrate away from such habitats.
Additional demographic and spatial data are necessary to understand dispersal of B. constrictor between urbanized and natural habitats to determine if these two habitats are demographic sources and sinks, respectively. Do female snakes die in natural areas as they outgrow their resource base or do they disperse to urban areas in pursuit of better resources? If small adult female snakes disperse into urban areas, then these urban habitats may act as spillback sources to maintain a continuously growing population across all habitat types.
In another large-bodied invasive snake study system, the Burmese Python (Python molurus) caused a crash of endothermic prey after they invaded the Everglades in southern Florida (Dorcas et al., 2012). Something similar likely occurred on Aruba as B. constrictor rapidly expanded across the island consuming native birds and mammals (Reinert et al., 2021). Thus, current food resources and condition for Boas on Aruba are different than when they first spread across the island. However, in the insular system of Aruba, large native prey species richness is much lower than in southern Florida, where pythons continue to expand their range (Bartoszek et al., 2021). These island–mainland differences will likely continue to drive different trajectories for two different large-bodied constricting invasive snakes.
Rodda and Tyrrell (2008) found a correlation between traits that make reptile species good pets and those that allow for success in urban areas, and the first Boas introduced onto Aruba were likely released pets (Bushar et al., 2015). Factors influencing evolution in urban environments include novel selection pressures due in impervious surfaces, urban heat island effect, and distinct community compositions (Johnson and Munshi-South, 2017), and this model may be particularly well suited to Boas on Aruba. On Aruba, the large source population of B. constrictor around urban resource subsidies may demographically buoy this invasive species, allowing it to persist on what would otherwise be an inherently hostile island for colonization.
Liu et al. (2020) found support for the hypothesis that environmental matching between invasive and native ranges is key for successful animal invasion. The association between human structures and Boa activity likely drives habitat conservatism between native and invasive Boa ranges, as the environment on Aruba near human dwellings tends to have more complex three-dimensional habitat structure, large trees, anthropogenic water sources, and an abundance of larger-bodied prey, likely matching native habitat for this tropical forest species. In the absence of these anthropogenic changes to the environment, natural habitats on Aruba may have been too different from native forested habitat of B. constrictor, thus preventing successful establishment of this species. We have found no historic evidence of prior populations of B. constrictor on Aruba, Bonaire, or Curaçao, despite Aruba being only 27 km from the South American mainland, where B. constrictor is native. It is also worth noting that occasional specimens of B. constrictor have been reported on the neighboring (and environmentally similar) island of Curaçao, but an invasive population similar to Aruba has not become established there (van Buurt, 2005, 2006). Restated, our data support that, while sampling bias may contribute to our results, urbanized areas on Aruba may act as demographic spillovers for Boas on Aruba; in the urbanized areas, large reproductive animals likely produce an abundance of offspring that spill into natural areas that would otherwise be demographically nonviable.
The following multiple lines of evidence suggest that our lack of samples of females without large or mature ovarian follicles in the natural areas is not due to limited sampling or a site by sex detection bias within the natural landscapes. First, female reproductive Boas are the largest individuals. In this study, females were 1 kg heavier and 25 cm SVL longer than males. Reproduction in these snakes is size dependent, and females begin to mature at a larger size than males (88.5 cm SVL and 78.0 cm SVL, respectively). Given small females and reproductive adults were frequently detected in natural areas, we would expect to also find the larger females if they were present. Second, the female-biased sex ratio in samples from urbanized areas suggests that the detection rate of females may be actually higher than males on Aruba. Lastly, we did dissect female Boas that were above the minimum size for gonadal maturation from natural areas; however, the entire sample of females from natural habitats had very small regressive ovaries both as measured by total mass and GSI. Qualitatively, these ovaries were too small to produce viable oocytes, which in captive B. constrictor are 4 cm in diameter at ovulation (Bertocchi et al., 2021). While we do not have precise comparisons of sampling effort between urbanized and natural habitats, natural habitats on Aruba are principally contained with the Arikok National Park, which is traversed daily by humans and includes approximately 30 full-time rangers that conduct regular surveys for Boas. Boas that would be detected by any rangers or visitors to the park were treated in the same way as described for all other Boas in this study, and would have been included in this dataset. Additionally, our research group has other long-term research in natural areas on Aruba and included approximately 12 weeks of field work in these areas; all Boa constrictor encountered by our research team while in the field were also surrendered to the authorities and also included in this dataset. Therefore, while we cannot effectively quantify Boa survey efforts in these habitats, the absence of adult reproductive females appears to be a biological effect of their general absence in these areas compared to urban areas in which they are frequently encountered.
Boas display a wide range of adaptability as evidenced by the native insular B. imperator in Belize, where island Boas adapted by reducing body size and clutch size (Boback, 2006). Also, sexual size dimorphism was evident in Belizean mainland B. imperator, but not present in the insular populations (Boback, 2006). Apparently, large body size in females is an evolutionary constraint to insular Boa populations. Selection pressure on female body size may cause females in insular populations to get progressively smaller or at least reach reproduction at smaller sizes. Despite direct predation pressure from B. constrictor, Aruban Whiptail Lizards (Cnemidophorus arubensis), an important prey species in natural habitats (Quick et al., 2005), likely experienced an ecological release from the presence of B. constrictor that reduced native lizard predators and competitors (Goessling et al., 2015). Surprisingly, this small endemic lizard has increased in population density following the successful invasion of B. constrictor, despite Boas preying upon them. Consequently, while large-bodied prey species have likely been reduced in natural habitats on Aruba, the natural desert ecosystem does retain a very dense prey base of small-bodied lizards (Reinert et al., 2021). Boas continue to persist around humans and the evolutionary potential remains for this species to persist with the strong resource subsidy from urban environments; as this represents a likely subsidy that creates a continuous source population, future research may warrant examining adaptation of Boas to the more limiting natural landscapes on Aruba to test how reproductive characteristics in females may respond to such environmental limitations.
DATA ACCESSIBILITY
Unless an alternative copyright or statement noting that a figure is reprinted from a previous source is noted in a figure caption, the published images and illustrations in this article are licensed by the American Society of Ichthyologists and Herpetologists for use if the use includes a citation to the original source (American Society of Ichthyologists and Herpetologists, the DOI of the Ichthyology & Herpetology article, and any individual image credits listed in the figure caption) in accordance with the Creative Commons Attribution CC BY License.
AI STATEMENT
The authors declare that no AI-assisted technologies were used in the design and generation of this article and its figures.
A map of Aruba showing all areas considered to be urban shaded in yellow in our analyses of Boa constrictor habitat urbanization. Areas within 250 m of a human building were considered urban, which totaled approximately 14,000 ha of the approximately 19,500 ha island, or approximately 72% of Aruba.
Plots of mass (A), snout–vent length (B), and gonad mass (C) of male and female Boa constrictor from natural and urbanized habitats on Aruba. Lines represent mean values, boxes represent standard error, and points represent individual snakes.
Observed and expected frequency of B. constrictor with prey in the stomach from natural and urbanized habitats in Aruba. Arrows indicate categories with significant divergence between observed and expectancy frequencies of prey consumed.
Contributor Notes
Associate Editor: J. M. Davenport.