Regular issues
Special Issues

Caribbean Naturalist
    CANA Home
    Range and Scope
    Board of Editors
    Editorial Workflow
    Publication Charges

Other EH Journals
    Northeastern Naturalist
    Southeastern Naturalist
    Urban Naturalist
    Eastern Paleontologist
    Eastern Biologist
    Journal of the North Atlantic

EH Natural History Home

Ecology and Conservation of the Endangered Puerto Rican Endemic, Varronia bellonis (Boraginaceae)
Ana María Sánchez-Cuervo, Benjamin J. Crain, and Víctor José Vega-López

Caribbean Naturalist, No. 20 (2014): 1–19

Full-text pdf: complete with cover or manuscript only (Accessible only to subscribers. To subscribe click here.)


Site by Bennett Web & Design Co.
Caribbean Naturalist 1 A.M. Sánchez-Cuervo, B.J. Crain, and V.J. Vega-López 22001144 CARIBBEAN NATURALIST No. 20N:1o–. 1290 Ecology and Conservation of the Endangered Puerto Rican Endemic, Varronia bellonis (Boraginaceae) Ana María Sánchez-Cuervo1,2,*, Benjamin J. Crain2, and Víctor José Vega-López1,2 Abstract - Varronia bellonis (Serpentine Manjack) is an endangered shrub endemic to Puerto Rico that is in need of renewed management strategies. To address this issue, we examined V. bellonis’ distribution, habitat requirements, population dynamics, population structure, and reproductive phenology. Our ultimate goals were to uncover critical ecological characteristics of this species, evaluate the current recovery strategies, and help guide conservation management practices. We found that V. bellonis could grow in various habitats, yet its overall range has been reduced to a single site. Moreover, the population has recently declined by 42% and a deterministic growth model suggests the species will be extinct in 44 years. While the population did not have a biased sex ratio, the majority of individuals were small and incapable of reproduction. For those individuals capable of reproduction, their flowering, fruiting, and recruitment were seasonally synchronized and corresponded with changes in rainfall patterns. Floral visitors were few, seed dispersal distances were short, and recruit survival was low. Overall, our study reveals that V. bellonis has multiple ecological characteristics contributing to its rare status and limiting its recovery, but preventable anthropogenic disturbances are the leading causes of this species’ decline. Although V. bellonis is protected under the US Endangered Species Act, the initial recovery strategies and management practices have not been successful, and therefore we argue for additional proactive strategies that include in situ and ex situ measures such as increased protection, improved habitat management, active propagation and reintroduction programs, and continued research efforts. Introduction Varronia bellonis (Urb.) Britton (Serpentine Manjack; see Borhidi et al. 1988, Britton and Wilson 1925) is a rare and endemic plant from the island of Puerto Rico (USFWS 1995). Since its discovery in Maricao in 1899, V. bellonis has also been encountered in Sabana Grande, Arecibo, Yauco, San German, Utuado, and Ciales, but only rarely (e.g., Woodbury s.n., NY 967246, UPR 000234; Axelrod 8622 NY, UPRRP; Proctor 49502 SJ; Cedeño et al. 22 MAPR, UPRRP; Axelrod & Potter 9590 UPRRP; Trejo-Torres & Crespo 1524 UPRRP; Axelrod 8469 NY, UPRRP; Sustache 653 SJ). In 1995, the US Fish and Wildlife Service (USFWS) formally proposed that V. bellonis be listed as an Endangered plant species (USFWS 1995, 1996), and two years later a final rule was made to classify the species as such pursuant to the US Endangered Species Act of 1973 (V. bellonis is actually listed as Cordia bellonis Urb.; USFWS 1997). 1Department of Biology, University of Puerto Rico-Mayagüez , PO Box 9012, Mayagüez, PR 00681. 2Department of Biology, University of Puerto Rico-Río Piedras, PO Box 23360, San Juan, PR 00931. *Corresponding author - Manuscript Editor: James D. Ackerman Caribbean Naturalist A.M. Sánchez-Cuervo, B.J. Crain, and V.J. Vega-López 2014 No. 20 2 Despite the legal conservation status of V. bellonis and the fact that it occurs on protected lands, most of the original populations have been disrupted by human activities. In Maricao, 34 individuals were eliminated mainly due to roadside clearing and road reconstruction (Breckon and Kolterman 1994; USFWS 1997, 1999). Likewise, the population in the Río Abajo State Forest was threatened by highway construction in 1994–1995 (USFWS 1997, 1995, 1999). Consequently, 111 plants were removed and relocated to the Cambalache nursery where V. bellonis had previously been grown from seed (Breckon and Kolterman 1996, USFWS 1999). Nevertheless, only 61 individuals survived the transplant process (USFWS 1999). Consequently, as of 1997, only 99 known plants remained in the wild, and the population of V. bellonis appeared to be in rapid decline. A reintroduction program was implemented between 2002 and 2003 at four locations in the Río Abajo State Forest with the surviving plants from the Cambalache nursery (USFWS 1999). To our knowledge, no one has conducted a comprehensive survey of the species since that time. For those reasons, we studied the biological status of V. bellonis and the ecological factors that may be influencing its recovery. Specifically, we aimed to answer the following questions: (1) How are the geographic distribution and underlying habitat characteristics affecting the status of V. bellonis? (2) How does the population size and population structure of V. bellonis affect its recovery potential? (3) How is the reproductive biology of V. bellonis affecting the recovery of the species? The answers to these questions are needed to fill important data gaps in the formal recovery plan regarding the current distribution, the existing population size and structure, and the ecological limitations of the species. Once identified, this information can aid long-term conservation and management efforts. Methods Study species Varronia bellonis was originally described as Cordia bellonis in 1899 by Urban (1899) from specimens collected by Paul Sintenis at Monte Alegrillo, Maricao. In 1925, Nathaniel Britton transferred the species to Varronia. Plants are woody dioecious shrubs, 1–3 m tall with arching to erect branches (Fig. 1; Acevedo-Rodriguez 2005, Breckon and Kolterman 1993, Liogier 1995). Additional details on the reproductive biology and phenology of the species are unknown. At the time of its listing as an endangered species, the known population of V. bellonis consisted of only 210 individuals of undocumented size, life stage, or gender, which occurred in three forests (Table 1; USFWS 1997). Five individuals were recorded in the Susúa State Forest (USFWS 1997), 87 individuals were recorded at 17 locations in the Maricao State Forest (USFWS 1997), and an additional 118 individuals were recorded among 12 small locations in the Río Abajo State Forest (USFWS 1997). Together, the plants documented in these three State Forests represented the entire known population of V. bellonis (USFWS 1997). Data on the population’s structure or dynamics has yet to be collected. Varronia bellonis occurs between 230 and 820 m elevation on steep slopes or on saddles between limestone hills (USFWS 1999). Populations exist along the Caribbean Naturalist 3 A.M. Sánchez-Cuervo, B.J. Crain, and V.J. Vega-López 2014 No. 20 edges of roads, highways, and riversides in both open and vegetated areas (Acevedo- Rodriguez 2005, Breckon and Kolterman 1993, Liogier 1995). They are most frequently found on limestone- and serpentine-derived soils. Additional characteristics of V. bellonis’ habitat (i.e., site attributes and soil properties) remain largely undescribed. Field-Site Description To assess the status of V. bellonis, we performed surveys of the three known populations located in the Susúa, Río Abajo, and Maricao State Forests from 2004 to 2006 (Fig. 2 inset). Elevation ranges between 150 and 875 m in these forests (Ventosa-Febles et al. 2005), and the areas we surveyed were in mountainous terrain Figure 1. Varronia bellonis in the Maricao State Forest. (A) Seedlings, (B) adult plant, (C) female flowers, (D) male flowers, (E) nocturnal floral visitor (Noctuidae), and (F) mature fruits. Caribbean Naturalist A.M. Sánchez-Cuervo, B.J. Crain, and V.J. Vega-López 2014 No. 20 4 characterized by sheer ravines and intermittent streams. A diversity of substrates can be found among these forests including limestone and serpentine formations (DNER 1976, USFWS 1999); the Maricao forest soils are primarily serpentine (75–100%; DNER 1976). Average annual rainfall is between 1339 to 2606 mm, with a relatively dry season occurring from December through March. The average temperature ranges from 21.7 to 25.5 °C (Silander et al. 1986, USFWS 1999). Three life zones occur within these reserves: subtropical wet forest, subtropical moist forest, and subtropical lower montane wet forest (Ewel and Whitmore 1973, USFWS 1999). In total, 845 species of vascular plants have been reported in the forests including 278 tree species and 123 Puerto Rican endemics (DNR 1976). Twelve of these species are restricted to serpentine soils, and six are found only in the Maricao State Forest (Cedeño 1997). Overall, these forests provide unique habitat for a number of distinctive species in addition to V. bellonis and may warrant special attention from conservation stakeholders. Data collection and analysis We conducted a primary survey from March 2004 to March 2006 that initially consisted of multiple visits to each location in the Maricao State Forest (monthly surveys), the Río Abajo State Forest (10 surveys), and the Susúa State Forest (3 surveys) where Breckon and Kolterman documented occurrences of V. bellonis during previous surveys (Breckon and Kolterman 1993, 1994, 1996; G.J. Breckon and D.A. Kolterman, University of Puerto Rico-Mayagüez, Mayagüez, PR, pers. comm.). When the species was encountered, we then searched for additional plants within Figure 2. Distribution of Varronia bellonis patches in the Maricao State Forest with an insert map showing the locations of the Río Abajo, Maricao, and Susúa State Forests in Puerto Rico. Surveys were conducted in each of these forests to document all populations of the species. Caribbean Naturalist 5 A.M. Sánchez-Cuervo, B.J. Crain, and V.J. Vega-López 2014 No. 20 a 50-m radius of each individual we located. If additional plants were discovered during this process, we expanded the search to all areas within 50 m of these new individuals. This procedure was repeated until we discovered no additional individuals at the site. Furthermore, we reviewed collection records from herbaria at the University of Puerto Rico’s Botanical Garden in San Juan (UPR), the University of Puerto Rico-Río Piedras (UPRRP), the University of Puerto Rico-Mayagüez (MAPR), and the Department of Natural and Environmental Resources in San Juan (SJ), and from the New York and Missouri Botanical Garden databases (MBG 2013, NYBG 2013) to find added distribution data. We used this information to search other areas opportunistically as time permitted. This strategy increased the probability that we surveyed areas that were likely to support V. bellonis populations. During our surveys, we used a Garmin 76™ Global Positioning System (GPS) to record the geographic coordinates of each population we observed and we tagged all individuals. We used the Geospatial Conservation Assessment Tool (GeoCAT) to evaluate the range dynamics of the species (Bachman and Moat 2012; available online at To identify underlying habitat parameters associated with V. bellonis, we gathered the following data next to each plant: slope, aspect, soil depth, litter depth, pH, and canopy cover (open or closed). At each site, we used an inclinometer to measure slope (%), a compass to measure aspect (°), a standard meter stick to measure depths (cm) of soil layers, and the Soil Tester (Kel Instrument Co., Inc., Wyckoff, NJ, USA) to record soil pH. We categorized canopy cover as open (≤50% cover) or closed (>50% cover) based on visual estimates. All measurements of habitat parameters were taken in the Maricao and Río Abajo Forests over two consecutive weekends in June of 2004. Using the measured habitat parameters, we performed three principal component analyses (PCA) to evaluate differences in environmental characteristics related to: (1) the distribution of females and males, (2) the distribution of reproductive and non-reproductive individuals, and (3) the different sites (Maricao vs. Río Abajo). We developed a main matrix containing data for all individuals and their respective habitat parameters. The main matrix was relativized by each column (i.e., each environmental variable). We also included three secondary matrices that separated all individuals by sex, reproductive state, and site of occurrence. All PCAs were based on correlation coefficients with a Euclidean distance measure. When visual separation was observed in the PCA plots, we used a multi-response permutation procedure (MRPP) based on Euclidean distance to test for significant differences between occupied habitats. All ordination analyses were performed using PC-ORD v. 5.0 (McCune and Mefford 2002). To analyze population trends, we collected demographic data for V. bellonis at each site where it occurred. We recorded the total population size (including seedlings) during our primary survey from 2004–2006 and revisited each population to reassess its status during a follow-up survey in 2012. We used a deterministic model, λ = (Nt+x/Nt)(1/x) (where λ = 1.0 indicates a stable population; Morris and Doak 2002), to calculate an annual population growth rate (λ) for the entire Caribbean Naturalist A.M. Sánchez-Cuervo, B.J. Crain, and V.J. Vega-López 2014 No. 20 6 population of V. bellonis. Additionally, we calculated a separate growth rate for only the non-seedling individuals (i.e., juveniles and adults) and another for only the adult individuals because seedling and juvenile numbers tend to fluctuate more than numbers of mature plants. To assess population structure based on sex ratios, we documented the gender of all reproductive individuals. As V. bellonis is a dioecious species, a chi-square analysis was performed to determine if the population had a biased sex ratio, which could influence reproductive performance within the population (Sinclair et al. 2012). To assess population structure based on size classes, we measured the size of all plants. For seedlings (i.e., plants less than 0.5 cm in diameter), we measured overall height. For non-seedling plants, we recorded the basal diameter of each individual at a height of 5 cm above ground. This method was chosen because overall plant height is not a reliable measure of plant size due to the fact that older apical branches are frequently broken off. We created a histogram to classify the frequency distribution of the basal diameters of all individuals. We calculated the skewness of the histogram as the third central moment (g1) of the distribution and tested the significance of the skew statistic with a two-tailed t-test (Sokal and Rohlf 2012). Observations on reproductive biology and phenology were conducted monthly from 2004 to 2006 in the Maricao Forest. To identify relationships between plant size and reproductive capacity, we performed a two-tailed unpaired t-test analysis to detect differences between the basal diameters of reproductive individuals and non-reproductive individuals. We performed a second two-tailed unpaired t-test to detect differences between the basal diameters of reproductive males and females. Additionally, we recorded the timing of four reproductive phases: budding, flowering, fruit set, and fruit maturation and determined the proportion of the population that successfully produced these reproductive structures. During the entire flowering season (from August to November 2004), we also performed weekly observations of floral visitation at 10 reproductive plants (five males and five females). Observation periods were from 0600 hours to 1000 hours, 1400 hours to 1600 hours, and 1800 hours to 2000 hours on 12 days for a total of 96 hours of field observations. Using taxonomic keys in the field, we identified all floral visitors to at least the family level and counted the number of flowers visited by each individual. Lastly, we recorded the timing of seed dispersal and seedling recruitment. All new seedlings were marked and monitored for survival. We used data from the National Oceanic and Atmospheric Administration’s Maricao 2 SSW weather station located at 18º09'0.4" N, 66º59'20" W (SERCC 2007) to look for associations between the number of individuals with reproductive structures (buds, flowers, immature fruits, and mature fruits) and total monthly precipitation. All analyses were performed with R statistical software version 2.12.2 (R Development Core Team 2012). Results Distribution and habitat characteristics During our primary surveys from 2004 to 2006, we encountered V. bellonis populations at only two sites, the Maricao State Forest and the Río Abajo State Forest. Caribbean Naturalist 7 A.M. Sánchez-Cuervo, B.J. Crain, and V.J. Vega-López 2014 No. 20 Although the species had been observed in the Susúa State Forest in 1992 (USFWS 1997), we were unable to locate any plants there. Our surveys of the surrounding areas in the Susúa forest did not reveal any previously undocumented individuals or populations. In our 2012 follow-up survey, we were also unable to locate any plants in the Río Abajo State Forest, leaving only a single remaining population. Accordingly, we estimate that V. bellonis’ area of occupancy has been reduced by ~66% and its area of extent has been reduced by ~90% based on our analysis of these measures. Habitat and substrate characteristics varied widely where V. bellonis was found. All populations we located were in the subtropical moist forest (Maricao) and subtropical wet forest (Río Abajo) life zones (Ewel and Whitmore 1973). All plants were found in secondary forests, at the edges of roads, and along paths in reserves and recreational areas. Occupied sites had slopes ranging from 0° to 90° (mean = 40.7, SD = 25.3) and aspects ranging from 20° to 355°. Virtually all individuals occurred on serpentine soils, which is a prominent substrate in the Maricao Forest where the majority of plants were located. Soil depth ranged from 0 to 50 cm (mean = 18.4, SD = 10.4), and litter depth ranged from 1 to 16 cm (mean = 6.1, SD = 2.8). Soil pH ranged from 5.6 to 7.2 (mean = 6.3, SD = 0.3). Most individuals (75%) occurred at sites with sparse (i.e., open) canopy cover. The two PCAs evaluating differences in environmental characteristics related to the distribution of males and females and the distribution of reproductive and nonreproductive individuals showed that the two primary axes explained 59.7% and 62.6% of the variance, respectively (Fig. 3A, B). Visually, we did not notice any separation in environmental characteristics between males and females or reproductive and non-reproductive individuals. Nevertheless, differences between the environmental characteristics at each site (Maricao vs. Río Abajo) were apparent in the third PCA (Fig. 3C). There was significant separation in the model comparing these two groups, as confirmed by the MRPP results (A = 0.041, P < 0.05), and the two primary axes explained 59.1% of the variance. The separation in the model was driven principally by soil depth and litter depth, which were all greater in the Río Abajo State Forest sites. Population size and structure We documented 226 plants during our primary surveys from 2004 to 2006 (Table 1). At the four locations in the Río Abajo State Forest where 61 individuals were re-introduced, we found that only 34 individuals remained. We did not find any new individuals in the Río Abajo State Forest. In the Maricao State Forest, we found 34 of the 87 individuals that were previously documented in the area, but the remaining 53 could not be located (Fig. 2; Breckon and Kolterman 1993). We also found 158 individuals (the majority of which were seedlings and non-reproductive plants) in this forest that were not documented during previous surveys (Table 1, Fig. 2). During our follow up surveys in October 2012 (Table 1), none of the 34 individuals we documented during our previous survey at the reintroduction sites in the Río Caribbean Naturalist A.M. Sánchez-Cuervo, B.J. Crain, and V.J. Vega-López 2014 No. 20 8 Figure 3. Ordination of a principal components analyses comparing the distributions of (A) females and males, (B) reproductive and non-reproductive individuals, and (C) sites of occurrence according to the six environmental characteristics. Caribbean Naturalist 9 A.M. Sánchez-Cuervo, B.J. Crain, and V.J. Vega-López 2014 No. 20 Abajo State Forest remained. In the Maricao State Forest, we observed only 122 of the 192 individuals documented in 2006. Together, these results indicate that the total population of V. bellonis has exhibited a λ of 0.90 over the 6-year period from 2006–2012 and it has experienced a reduction of ~46% over that time. Consequently, by these measures we estimate that the total population of V. bellonis will reach quasi-extinction (n < 2 plants) within 44 years. Considering only the non-seedling plants, in 2012 we observed 69 of the 118 individuals found in 2006. Accordingly, the non-seedling population has exhibited a λ of 0.91 and has experienced a reduction of ~42% from 2006–2012. Based on this rate of change, we estimate that the nonseedling population will reach quasi-extinction within 40 years. Considering only the mature adult population, in 2012 we were able to locate 40 of the 52 individuals found in 2006. Thus, this subset of the population has exhibited a λ of 0.95 a nd has experienced a reduction of ~24% from 2006–2012. Thus, the adult population is estimated to reach quasi-extinction in 69 years. The total population of V. bellonis had relatively few reproductive plants in 2006 (Table 1). In the Río Abajo State Forest, reproductive plants made up only 2.9% of the population, whereas the Maricao Forest population had 26.5% reproductive plants. In 2012, the total population consisted of 32.8% reproductive plants. The sex ratio did not differ from 1:1 in 2006 (χ2 = 1.59, df = 1, P = 0.2) or in 2012 (χ2 = 1.04, df = 1, P = 0.3). Seedlings ranged in height from 2 to 50 cm. Basal diameter for all non-seedling individuals was 0.5–5 cm (mean = 1.8, SD = 1.01). The frequency distribution of Table 1. Number of individuals of Varronia bellonis distributed in three State Forests in Puerto Rico. Numbers in parenthesis represent new records in the 2004–2006 main survey. For details of the 1993–1994 survey, see USFWS (1999). Survey (year)/Forest Males Females Non-reproductive Seedlings Total 1993–1994 Susúa 1 0 4 0 5 Río Abajo 0 0 118* 0 118 Maricao 0 0 87* 0 87 Total 1 0 209 0 210 2004–2006 Susúa 0 0 0 0 0 Río Abajo 0 1 33 0 34 Maricao 16 (14) 13 (8) 5 (28) (108) 192 Total 30 22 66 108 226 2012 Susúa 0 0 0 0 0 Río Abajo 0 0 0 0 0 Maricao 24 16 29 53 122 Total 24 16 29 53 122 *No information on gender was collected during this survey. In the 2004–2006 main survey, we characterized the gender of 29 individuals (16 males and 13 females) that were identified but not sexed in the 1993–1994 survey. Caribbean Naturalist A.M. Sánchez-Cuervo, B.J. Crain, and V.J. Vega-López 2014 No. 20 10 non-seedlings (Fig. 4) was moderately right skewed (t = 3.82, P = 0.001) toward smaller basal diameters to form a reverse “J” shaped curve as described by Smith et al. (1997). Reproduction and phenology Reproduction by V. bellonis was essentially limited to larger plants. Reproductive individuals had greater basal diameter (mean = 2.2 cm, SD = 0.14) than non-reproductive ones (mean = 1.1 cm, SD = 0.08). The variances in basal diameter between reproductive and non-reproductive individuals were different (F = 0.3, P = 0.00004), and a t-test assuming unequal variances between the two groups confirmed a difference in diameters (t = -6.9, P < 0.0001), with reproductive individuals being larger. In contrast, variances in basal diameter between reproductive males and females were not significantly different (F = 1.8, P = 0.07), and a t-test assuming equal variances indicated that there was no significant difference in basal diameter (t = -0.25, P = 0.8) between males (mean = 2.2 cm, SD = 0.2) and females (mean = 2.3 cm, SD = 0.1). Thus, both male and female plants needed to reach relatively large sizes before they became reproductive. We monitored the flowering and fruiting phenologies of 51 reproductive individuals between 2004 and 2006 (Fig. 5). In 2004, the onset of the first flower buds was in August, and budding continued until February on individuals that had Figure 4. Frequency distribution of basal diameters in Varronia bellonis in the Maricao State Forest. Caribbean Naturalist 11 A.M. Sánchez-Cuervo, B.J. Crain, and V.J. Vega-López 2014 No. 20 already begun flowering and/or fruiting, but buds forming after November were aborted before blooming (Fig. 5A). Anthesis ran from August to November (Figs. 1C and D, 5A). We observed 29 flowers being visited by moths (Noctuidae; Fig. 1E), 21 were visited by Apis mellifera L. (Honey Bee; Apidae), and six were visited by hoverflies (Syrphidae). In general, visits from Honey Bees and hoverflies were short and characterized by fast and erratic movements, but moths made complete landings on the flowers and sat and foraged for longer per iods. Figure 5. Synchronous phenology of Varronia bellonis reproductive phases in the Maricao State Forest: (A) buds and flowers and (B) immature and mature fruits, compared to average monthly precipitation (mm) between March 2004 to February 2006. Caribbean Naturalist A.M. Sánchez-Cuervo, B.J. Crain, and V.J. Vega-López 2014 No. 20 12 We detected the first immature (i.e., green) fruits in September 2004 and continued to observe them until February 2005 (Fig. 5B). At the beginning of November 2004, we observed the first mature fruits, as indicated by their transition from green to red (Figs. 1F, 5B); 66.7% of females had produced mature fruits at that time. Mature fruits were observed until February 2005, but only on 14.3% of the female population since most fruits had fallen by then. In our second period of observations (from August 2005 to February 2006), reproductive phenology was very similar to the previous year but it occurred in fewer individuals (25.0% of the population as compared to 26.5% in the 2004–2005 season; Fig. 5A, B). Reproductive phenology appeared to follow annual precipitation patterns (based on average monthly rainfall) during our surveys (Fig. 5A, B). In general, buds and flowers formed in the wet season, and immature and mature fruits developed in the dry season. During the fruiting season, we did not detect any seed dispersal by animals. Instead, we observed that the majority of mature fruits fell directly to the ground where they usually decomposed. After the 2004–2005 fruiting period, we recorded the recruitment of 49 new seedlings (Fig. 1A). Most seedlings (94%) were found directly under the crowns of mature females, while a few (6%) were found some distance away (less than 15 m). These new recruits exhibited high mortality rates; 20 seedlings (~40%) had already perished by May 2005. Similar mortality rates were observed in 2006. Discussion Despite federal protection, implementation of a formal recovery plan, and concerted efforts of conservation managers, V. bellonis is becoming increasingly rare. It is clear there are several geographic, demographic, and biological attributes of V. bellonis that act as potential barriers to its recovery. The data presented in this analysis should improve the ability of conservation managers to counteract these limiting factors and to implement a suitable recovery strategy. Distribution and habitat characteristics Our surveys reveal that there has been a significant contraction in V. bellonis’ range size over the past decade. The loss of the populations in the Río Abajo State Forest and in the Susúa State Forest (which were presumably extirpated) eliminated the northernmost and southernmost portions of the species’ range. With only one population remaining, the likelihood that stochastic events or anthropogenic disturbances will lead to the extinction of this species is greatly increased (Gaston 1994, Purvis et al. 2000). Moreover, there is a positive relationship between a species’ range size and its abundance (Gaston 1994). Range contractions can have indirect consequences at the population level (e.g., disruption of metapopulation dynamics, increased inbreeding, and Allee effects), which can lead to species extinction (Schemske et al. 1994). Accordingly, the restricted range of V. bellonis is a potentially serious limitation to its recovery via several mechanisms, and the conservation strategy for V. bellonis should include specific plans for increasing its range size. Caribbean Naturalist 13 A.M. Sánchez-Cuervo, B.J. Crain, and V.J. Vega-López 2014 No. 20 Although the overall geographic range of V. bellonis is restricted and fragmented, we found that V. bellonis does not appear to have overly restrictive habitat requirements. Still, our results show that V. bellonis is associated with certain habitat characteristics, and the loss of the reintroduced individuals in Río Abajo is a strong indication that these characteristics need to be taken into consideration when attempting to propagate and reintroduce the species (Godefroid et al. 2011). Soil depth and litter depth were the main factors that differentiated the habitats in Maricao and Río Abajo, and canopy cover affects each of these variables (e.g., Breshears et al. 1997). Because the surviving 34 re-introduced individuals in the Río Abajo forest were located in 2006 under a dense canopy on comparatively deep soils with substantial leaf-litter cover, it is possible that they did not survive to 2012 due to these particular habitat characteristics. While loss of critical habitat is a limiting factor for any species, these plants are apparently capable of surviving in disturbed areas such as secondary forests, road edges, and river margins. Thus, suitable habitats may be more widely available for reintroduction efforts. Additionally, active maintenance of secondary successional sites to mimic V. bellonis’ natural habitat is a potential management option. Our findings on the habitat characteristics associated with V. bellonis should provide a useful guide to insure that optimal sites are chosen and maintained during future reintroductions. Population size and structure Despite the discovery of new individuals, the population of V. bellonis continues to become smaller. Our 2012 survey indicates that the population has experienced a decline of ~42% since 1994. Seedling mortality appeared to be a significant hindrance to overall population growth; however, mortality among juvenile and adult populations also contributed to the overall population decline. With such low population numbers, several factors such as demographic stochasticity, catastrophes, or reduced fecundity can contribute to the loss of the species (Morris and Doak 2002, Schemske et al. 1994). For that reason, a suitable recovery strategy must include plans to increase survival rates among the existing populations in the Maricao State Forest. To identify the best methods to achieve this goal, continued monitoring is needed to develop matrix models that can fully assess the population dynamics of this species and to help identify the most critical life stages and vital rates. The demographic structure of V. bellonis and its potential effects on population growth is also a potential cause for concern. Although the sex ratio of V. bellonis was statistically indistinguishable from 1:1, the sex of the 28 non-reproductive individuals in the Maricao State Forest needs to be established to determine if the ratio remains unbiased. This is an important consideration since a biased sex ratio could reduce the effective population size and create a genetic bottleneck that may limit the recovery of the species (Schemske et al. 1994, Sinclair et al. 2012). Therefore, the recovery strategy for V. bellonis should include a population genetic study to determine if genetic bottlenecking is occurring as a consequence of low effective population size. In turn, this information can be used to guide the placement Caribbean Naturalist A.M. Sánchez-Cuervo, B.J. Crain, and V.J. Vega-López 2014 No. 20 14 of individuals of either sex during reintroduction events in an effort to facilitate reproduction and genetic mixing. Another concern related to the demographic structure of the V. bellonis population is that it is composed mostly of seedlings (41.4%) and non-reproductive plants (21.8%). Varronia bellonis seedlings have high mortality rates after recruitment (February–March), and consequently, seedling mortality could result in population growth rates smaller than those estimated here. Demographic models have shown that populations can grow more slowly than their stable growth rate if immature individuals are overrepresented in the population (Stott et al. 2012). Conversely, high proportions of immature plants in the population reflect high fecundity, which may have a greater impact on population persistence than adult survival. Therefore, long-term demographic analyses using stage-based matrix models to assess asymptotic and transient population dynamics should be the next step for estimating the effects of population structure by determining which size classes have the largest impact on the population growth rate (Schemske et al. 1994). Our results on the size-class distribution in the population and reproductive capacity should make this type of analysis a more feasible component of a formal recovery strategy as additional surveys are completed. Reproduction and phenology The reproductive biology of V. bellonis suggests that several factors may be restricting its recovery. First, reproduction was limited to larger plants that comprised a small proportion of the population. In view of that, it may take multiple years for a large portion of individuals to become reproductively active and to contribute to population growth. Therefore, efforts should be made to stimulate plant growth through plant- and site-maintenance efforts so that individuals may reach reproductive sizes as quickly as possible. Propagation methods commonly used for stimulating plant growth in nurseries, e.g., watering, fertilizing, and pruning, have proven effective for bolstering plant populations during recovery efforts for several plant species and need to be tested on cultivated populations of V. bellonis (Godefroid et al. 2011, Khurana and Singh 2001). Although our study did not take into account actual fruit set, our impressions are that the number of fruits produced was very low as compared to the number of flowers produced. This condition could be a limiting factor in V. bellonis. Like other related dioecious species such as Cordia collococca L. (Red Manjack) and C. panamensis L. Riley (Hairy Lay-Lay) (Opler et al. 1975), V. bellonis flowers are visited by only a few generalist insects. Also, there is a potential risk of pollen limitation because of the large distance (mean = 200 m) between male and female plants and the low number of flowers visited (De Jong et al. 2005, Pascarella 1996). Therefore, pollination studies are needed to determine whether V. bellonis is pollen or resource limited. The possibility of pollen limitation also suggests that proactively increasing fruit production is a potential avenue to stimulate population growth. Reproductive phenology must also be considered in the recovery plans for V. bellonis because the species is seasonally synchronous and correlated with Caribbean Naturalist 15 A.M. Sánchez-Cuervo, B.J. Crain, and V.J. Vega-López 2014 No. 20 precipitation patterns. Other factors (e.g., temperature, solar radiation) could act as added flowering and fruiting stimuli (Numata et al. 2003, van Schaik et al. 1993), and it is clear that climate change should be considered in the formal recovery strategy for V. bellonis because of the changes in temperature and precipitation patterns that are expected in Puerto Rico (Scatena 1998). These predicted changes could disrupt several aspects of the species reproductive phenology and limit its recovery. Recommendations for species recovery Considering the ongoing natural and anthropogenic threats to this species, increased efforts from conservation biologists and managers are essential to prevent the extinction of V. bellonis. Specifically, research and management priorities should focus on four actions: (1) Improving protection and enforcement. Safeguarding the remaining population will directly help to maintain the species range size and bolster population growth. Since poorly regulated path maintenance, routine roadside clearing, and road expansion in Maricao were among the main factors leading to the decline of V. bellonis’ population, increased protection of this population should prevent additional losses from avoidable anthropogenic disturbances. Special attention should be given to individuals along roads and trails because roadside vegetation management (e.g., brush cutting) could readily extinguish these populations. Barrier fences have been constructed on multiple occasions to stop damage to endangered plant populations (e.g., Jusaitis 2005, Loope et al. 2009, Maschinski et al. 1997), and this strategy should be considered to avoid further population declines triggered by trail and road maintenance (Godefroid et al. 2011). (2) Improving site maintenance. Varronia bellonis was associated with certain habitat characteristics, and maintenance of the areas immediately surrounding populations could help to bolster growth and survival rates (Godefroid et al. 2011). Clearing or thinning neighboring plants could be a useful approach to reduce resource competition and improve plant vigor (Capo-Arteaga and Newton 1991, Godefroid et al. 2011, Lewis and Tanner 2000). However, responses to these treatments can be species specific (Capo-Arteaga and Newton 1991, Ewing 2002, Lewis and Tanner 2000), and therefore these strategies need to be experimentally tested to optimize maintenance protocols. The recovery strategy for V. bellonis should include plans to test these management options to determine if they will enhance plant growth and survival. (3) Improving propagation and reintroduction efforts. An effective propagation and reintroduction program for V. bellonis will directly contribute to range expansion and increase genetic variation via maintenance of small isolated populations (Leppig and White 2006). Reintroductions can also help buffer against stochastic events that could eliminate the last remaining population. However, successful reintroduction programs are dependent on a number of factors (Godefroid et al. 2011), and appropriate consideration must be given to seed/seedling sources and availability, seed collection and storage techniques, propagation methods and rearing duration, and the number of individuals to be reintroduced (Godefroid et al. 2011, Khurana and Singh Caribbean Naturalist A.M. Sánchez-Cuervo, B.J. Crain, and V.J. Vega-López 2014 No. 20 16 2001). Limited seed production restricts the opportunity for managers to propagate the species, and for that reason, the recovery plan should include strategies for collecting and handling seeds to insure their viability for propagation (Khurana and Singh 2001). A population genetic study needs to be conducted so that a genetically diverse set of seeds and propagules can be used during reintroduction efforts to improve genetic mixing in the population (Godefroid et al. 2011). Once new propagules are available (juveniles and/or adult plants), optimizing reintroduction sites to maximize growing conditions and survival rates, e.g., choosing locations with sparse canopy cover and appropriate soils, should be a primary consideration. Locations within Maricao as well as in Río Abajo, Susúa, and Ciales should provide suitable habitats for reintroduction since much of the species’ original range was in these locations. Additionally, mature male and female individuals (if available) should initially be planted in close proximity to each other to increase the potential for natural reproduction. As part of the formal recovery strategy, these practices should stimulate the recovery of V. bellonis through population augmentation and range-expansion opportunities. (4) Continue monitoring efforts. The final recommendation for V. bellonis’ recovery strategy, continued monitoring, is an important link between each of the previous recommendations. Monitoring efforts are needed not only where the species is known to occur in Maricao, but also in areas where the species occurred in the past like Río Abajo, Susúa, and Ciales in case new individuals or populations may have reestablished themselves in these formerly occupied sites. Since environmental conditions are inherently dynamic, ongoing monitoring is critical for updating management strategies in response to potential changes. Continued observations are critical for detecting long-term demographic trends, for observing the effects of management treatments and decisions, and for re-evaluating the conservation status of the species. Therefore, the commitment to continual monitoring of V. bellonis should be a part of its recovery strategy so that management efforts can efficiently and effectively adapt to changing scenarios. In sum, V. bellonis could become extinct in the near future without enhanced protection from government authorities and local stakeholders. Upgrading the status of V. bellonis to critically endangered should be given serious consideration and an updated recovery plan should be created based on the issues highlighted by this study. Acknowledgments We thank Adrián Muñiz, Oficial de Manejo del Bosque Estatal de Maricao for helping with the logistics of this project. Additional thanks to Dr. Duane Kolterman and Dr. Gary Breckon for their advice and comments during our research. We are also grateful for the comments on the manuscript provided by Seth Rifkin. Literature Cited Acevedo-Rodriguez, P. 2005. Vines and climbing plants of Puerto Rico and the Virgin Islands. Contributions from the US National Herbarium 51:1–483. Bachman, S., and J. Moat. 2012. GeoCAT: An open source tool for rapid Red List assessments. BGjournal 9:1–3. Caribbean Naturalist 17 A.M. Sánchez-Cuervo, B.J. Crain, and V.J. Vega-López 2014 No. 20 Borhidi, A., E. Gondár, and Z.S. Orosz-Kovács. 1988. The reconsideration of the genus Cordia. Acta Botanica Hungarica 34:375–423. Breckon, G.J., and D.A. Kolterman. 1993. Cordia bellonis Urban (Boraginaceae). Final report under Cooperative Agreement No. 1416-0004-92-970 between US Department of the Interior, Fish and Wildlife Service, and University of Puerto Rico, Mayagüez Campus. Mayagüez, PR. 14 pp. Breckon, G.J., and D.A. Kolterman. 1994. Cordia bellonis Urban (Boraginaceae). Final report under Cooperative Agreement No. 1448-0004-93-973 between US Department of the Interior, Fish and Wildlife Service, and University of Puerto Rico, Mayagüez Campus. Mayagüez, PR. 4 pp. Breckon, G.J., and D.A. Kolterman. 1996. Cordia bellonis Urban (Boraginaceae). Final report under Cooperative Agreement No. 1448-0004-94-9113 between US Department of the Interior, Fish and Wildlife Service, and University of Puerto Rico, Mayagüez Campus. Mayagüez, PR. 5 pp. Breshears, D.D., P.M. Rich, F.J. Barnes, and K. Campbell. 1997. Overstory-imposed heterogeneity in solar radiation and soil moisture in a semiarid woodland. Ecological Applications 7:1201–1215. Britton, N.L., and P. Wilson. 1925. Scientific Survey of Porto Rico and the Virgin Islands, Botany of Porto Rico and the Virgin Islands, Myrtales to Lycopodiales. New York Academy of Sciences, New York, NY, USA. Capo-Arteaga, M., and M. Newton. 1991. Survival and growth of five species of Pinus seedlings after different approaches to competition control:“Bridging” studies between Oregon and Mexico. New forests 5: 219–238. Cedeño, J.A. 1997. Vascular flora of the Río Maricao watershed, Maricao Commonwealth Forest, Maricao, Puerto Rico. M.Sc. Thesis. University of Puerto Rico, Mayagüez, PR. 158 pp. De Jong, T.J., J.C. Batenburg, and P.G.L. Klinkhamer. 2005. Distance-dependent pollen limitation of seed set in some insect-pollinated dioecious plants. Acta Oecologica 28:331–335. Department of Natural Resources (DNR). 1976. The master plan for the Commonwealth forests of Puerto Rico: Area of planning and evaluation. San Juan, PR. 259 pp. Ewel, J.J., and J.L. Whitmore. 1973. The ecological life zones of Puerto Rico and the US Virgin Islands. Forest Service research paper ITF-18, Institute of Tropical Forestry, San Juan, PR. 72 pp. Ewing, K. 2002. Effects of initial site treatments on early growth and three-year survival of Idaho fescue. Restoration Ecology 10:282–288. Gaston, K.J. 1994. Geographic range sizes and trajectories to extinction. Biodiversity Letters 2:163–170. Godefroid, S., C. Piazza, G. Rossi, S. Buord, A.-D. Stevens, R. Aguraiuja, C. Cowell, C.W. Weekley, G. Vogg, and J.M. Iriondo. 2011. How successful are plant species reintroductions? Biological Conservation 144:672–682. Jusaitis, M. 2005. Translocation trials confirm specific factors affecting the establishment of three endangered plant species. Ecological Management and Restoration 6:61–67. Khurana, E., and J. Singh. 2001. Ecology of seed and seedling growth for conservation and restoration of tropical dry forest: A review. Environmental Conservation 28:39–52. Leppig, G., and J.W. White. 2006. Conservation of peripheral plant populations in California. Madrono 53:264–274. Caribbean Naturalist A.M. Sánchez-Cuervo, B.J. Crain, and V.J. Vega-López 2014 No. 20 18 Lewis, S.L., and E.V. Tanner. 2000. Effects of above- and belowground competition on growth and survival of rain forest tree seedlings. Ecology 81:2525–2538. Liogier, A. 1995. Descriptive flora of Puerto Rico and adjacent Islands: Spermatophyta– Dicotyledoneae, Melastomataceae to Lentibulariaceae. University of Puerto Rico, Río Piedras, PR. Loope, L., F. Starr, and K. Starr. 2009. Protecting endangered plant species from displacement by invasive plants on Maui, Hawaii. Weed Technology 18:1472–1474. Maschinski, J., R. Frye, and S. Rutman. 1997. Demography and population viability of an endangered plant species before and after protection from trampling. Conservation Biology 11:990–999. Missouri Botanical Garden (MBG). 2013. Tropicos. St. Louis, MO, USA. Available online at Accessed 15 February 2004. McCune, B., and M. Mefford. 2002. PC-ORD Multivariate Analysis of Ecological Data MjM Software, Gleneden Beach, OR, USA. Morris, W., and D. Doak. 2002. Quantitative Conservation Biology: Theory and Practice of Population Viability Analysis. Sinauer Associates, Sunderland, MA, USA. Numata, S., M. Yasuda, T. Okuda, N. Kachi, and N.S.M. Noor. 2003. Temporal and spatial patterns of mass flowerings on the Malay Peninsula. American Journal of Botany 90:1025–1031. The New York Botanical Garden (NYBG). 2013. C.V. Starr Virtual Herbarium. Bronx, NY, USA. Available online at Accessed 30 March 2004. Opler, P.A., H.G. Baker, and G.W. Frankie. 1975. Reproductive biology of some Costa Rican Cordia species (Boraginaceae). Biotropica 7:234–247. Pascarella, J. 1996. Reproductive ecology of Picramnia pentrandra (Picramniaceae) in South Florida. Caribbean Journal of Science 32:99–104. Purvis, A., J.L. Gittleman, G. Cowlishaw, and G.M. Mace. 2000. Predicting extinction risk in declining species. Proceedings of the Royal Society of London. Series B: Biological Sciences 267:1947–1952. R Development Core Team. 2012. R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. Scatena, F.N. 1998. An assessment of climate change in the Luquillo mountains of Puerto Rico. Pp. 193–198, In R.I. Segarra-García (Ed.). Third International Symposium on Water Resources. American Water Resources Association, San Juan, PR. Schemske, D.W., B.C. Husband, M.H. Ruckelshaus, C. Goodwillie, I.M. Parker, and J.G. Bishop. 1994. Evaluating approaches to the conservation of rare and endangered plants. Ecology 75:584–606. Silander, S., H. Gil de Rubio, M. Miranda, and M. Vazquez. 1986. Los bosques de Puerto Rico. Pp. 210–236, In J.L. Vivaldi (Ed.). Compendio Enciclopedico de los Recursos Naturales de Puerto Rico. Department of Natural Resources, San Juan, PR. Sinclair, J.P., J. Emlen, and D.C. Freeman. 2012. Biased sex ratios in plants: Theory and trends. Botanical Review 78:63–86. Smith, D.M., B.C. Larson, M.J. Kelty, and P.M.S. Ashton. 1997. The Practice of Silviculture: Applied Forest Ecology. John Wiley and Sons, Inc., New York, NY, USA. 537 pp. Sokal, R., and J. Rohlf. 2012. Biometry: The Principles and Practice of Statistics in Biological Research, 4th Edition. W.H. Freeman and Company, New York, NY, USA. 937 pp. Southeast Regional Climate Center (SERCC). 2007. Historical climate summaries for Puerto Rico and the US Virgin Islands. University of North Carolina at Chapel Hill, NC, USA. Available online at Accessed 25 January 2013. Caribbean Naturalist 19 A.M. Sánchez-Cuervo, B.J. Crain, and V.J. Vega-López 2014 No. 20 Stott, I., D.J. Hodgson, and S. Townley. 2012. Popdemo: An R package for population demography using projection matrix analysis. Methods in Ecology and Evolution 3:797–802. United States Fish and Wildlife Service (USFWS). 1995. Endangered and threatened wildlife and plants; proposed endangered status for the plant Cordia bellonis. 50 CFR Part 17, RIN 1018–AD48, Proposed Rules. Federal Register 60:50176–50180. Washington DC, USA. 5 pp. USFWS. 1996. Endangered and threatened wildlife and plants; review of plant and animal taxa that are candidates for listing as endangered or threatened species, Part 3. 50 CFR Part 17, Proposed Rule. Endangered and Threatened Species, Plant and Animal Taxa. Federal Register 61:7595–7613. Washington DC, USA. 19 pp. USFWS. 1997. Endangered and threatened wildlife and plants; determination of endangered status for the plant Cordia bellonis, Part 5. 50 CFR Part 17, Endangered and Threatened Wildlife and Plants; Final Rules, RIN 1018–AD48, Federal Register 62:1643–1647. Washington DC, USA. 5 pp. USFWS. 1999. United States Fish and Wildlife Service. Recovery plan for Cordia bellonis. Atlanta, GA, USA. 18 pp. Urban, I. 1899. Species novae, Praesertim portoricences. Pp. 291–481, In I. Urban (Ed.). Symbolae Antillanae seu Fundamenta Florae Indiae Occidentalis I. Fratres Borntraeger, Berlin, Germany. van Schaik, C.P., J.W. Terborgh, and S.J. Wright. 1993. The phenology of tropical forests: Adaptive significance and consequences for primary consumers. Annual Review of Ecology and Systematics 24:353–377. Ventosa-Febles, E., M. Camacho, J. Chabert, J. Sustache, and D. Davila. 2005. Puerto Rico critial wildlife areas. Department of Natural and Environmental Resources, Bureau of Fish and Wildlife, Terrestrial Resources Division, San Juan, PR. 385 pp.