EFFECTS OF INTRODUCED GREEN IGUANAS (IGUANA IGUANA) ON TROPICAL PLANT COMMUNITIES THROUGH SEED DISPERSAL AND GERMINATION

Despite the importance of dispersal and germination for plant life cycles and population dynamics, the effects of reptiles are often overlooked because herbivory is relatively rare in reptiles. Green Iguanas (Iguana iguana) enhance seed germination in some plant species in xeric habitats in its native range, but no studies have been conducted on introduced populations, such as in Puerto Rico. Because Green Iguanas can be abundant where they have been introduced, they have the potential to affect plant communities by dispersing and germinating seeds. In summer 2013, a total of 258 Green Iguana scat samples were collected in the Humacao Natural Reserve in southeastern Puerto Rico. An additional 53 scat samples were collected from captive Green Iguanas fed non-native C. papaya. To determine the percentage of seeds that germinated and the number of days to germination, seeds were extracted from scat and collected from fruit, and then planted under common garden conditions using four experimental treatments: 1) Digested seeds planted with feces, 2) Digested seeds planted without feces, 3) Undigested seeds planted with fruit, and 4) Undigested seeds planted without fruit. Four main species were identified from seeds in wild Green Iguana scat: native Anona glabra, Ficus sp., non-native Peltophorum pterocarpum, and Pterocarpus sp. Since multiple species in the genus Ficus and Pterocarpus were present at our study site and produce similar seeds, we could not identify seeds from these genera to the species level. Nonetheless, these seeds are either native Ficus citrifolia or non-native Ficus benjamina, and either native Pterocarpus officinalis or non-native Pterocarpus indica because these are the only species present at our study site. Seeds that passed through Green Iguanas exhibited reduced germination percentage in non-native P. pterocarpum, Pterocarpus sp., and non-native C. papaya seeds. In contrast to previous studies conducted in native habitats, Green Iguanas did not increase the germination percentage of any species in Puerto Rico, where Green Iguanas have been introduced. Passage through Green Iguanas reduced the days to germination of Ficus sp., non-native P. pterocarpum, and Pterocarpus sp., and increased the days to germination of non-native C. papaya. These results suggest the effect of Green Iguanas outside of their native range on germination percentage and days to germination depends on the species. Germination percentage and days to germination were both reduced for the dry seeds of P. pterocarpum and Pterocarpus sp. after passing through the Green Iguana gut. To assess seed dispersal potential by Green Iguanas, we collected GPS coordinates for scat samples and surrounding mature trees of the four main seed species found in scat samples (i.e., native Anona glabra, Ficus sp., non-native Peltophorum pterocarpum, and Pterocarpus sp.). Using these coordinates, we calculated the minimum distance between scat containing a specific seed species and the nearest tree of that species. Green Iguanas dispersed seeds throughout the habitats they used, but no trend or patterns was detected in dispersal of native and non-native plants, seed dispersal strategies, or types of seeds dispersed. Although minimum dispersal distances were relatively short for some species, mean distances were large enough for seeds of all species to be transported beyond the canopy of parent trees. Green Iguanas do not have consistent effects on seed germination among different plant species in introduced habitats, but because Green Iguanas have long retention time, defecate seeds that are relatively intact, and can move to dense forest and areas upstream where air and water seed dispersion cannot reach (e.g., A. glabra, P. pterocarpum and Pterocarpus sp.), Green Iguanas may be important seed dispersers in mesic habitats where they have been introduced. Further evaluation of Green Iguana effects on germination and dispersal are needed to determine how this species might influence specific species in plant communities outside of their native range.

percentage in non-native P. pterocarpum, Pterocarpus sp., and non-native C. papaya seeds. In contrast to previous studies conducted in native habitats, Green Iguanas did not increase the germination percentage of any species in Puerto Rico, where Green Iguanas have been introduced. Passage through Green Iguanas reduced the days to germination of Ficus sp., non-native P. pterocarpum, and Pterocarpus sp., and increased the days to germination of non-native C. papaya. These results suggest the effect of Green Iguanas outside of their native range on germination percentage and days to germination depends on the species. Germination percentage and days to germination were both reduced for the dry seeds of P. pterocarpum and Pterocarpus sp. after passing through the Green Iguana gut. To assess seed dispersal potential by Green Iguanas, we collected GPS coordinates for scat samples and surrounding mature trees of the four main seed species found in scat samples (i.e., native Anona glabra, Ficus sp., non-native Peltophorum pterocarpum, and Pterocarpus sp.). Using these coordinates, we calculated the minimum distance between scat containing a specific seed species and the nearest tree of that species. Green Iguanas dispersed seeds throughout the habitats they used, but no trend or patterns was detected in dispersal of native and non-native plants, seed dispersal strategies, or types of seeds dispersed.
Although minimum dispersal distances were relatively short for some species, mean distances were large enough for seeds of all species to be transported beyond the canopy of parent trees. Green Iguanas do not have consistent effects on seed germination among different plant species in introduced habitats, but because Green Iguanas have long retention time, defecate seeds that are relatively intact, and can move to dense forest and areas upstream where air and water seed dispersion cannot reach (e.g., A. glabra, P. pterocarpum and Pterocarpus sp.), Green Iguanas may be important seed dispersers in mesic habitats where they have been introduced.  Same letters above bars denote no significant difference among the four treatments. 37

INTRODUCTION
Mangrove forests cover eight percent of the world coastlines (Lugo et al. 1999). They are important ecologically as they build terrain, prevent erosion, protect against common tidal and cyclonic events, filter pollution, export organic matter to surrounding aquatic ecosystems, and are biodiversity hotspots (Lugo et al. 1999 (Rivero 1998). The Green Iguanas' native range extends from Mexico to mid-South America (Rand 1989 (Rand 1978, Iverson 1982, Troyer 1984, White 1985, van Marken 1992. They have low metabolic rates and high gut passage times compared to mammals and birds (van Marken 1992). Retention of microbes and nematodes in the intestinal tract allows Green Iguanas to degrade plant cell walls, achieving a digestibility as high as 54% of biomass consumed (Bjorndal 1979, Iverson 1980, Troyer 1984. Reptiles are also known to swallow large portions of food whole instead of chewing food items into small pieces (Bjorndal et al. 1990). This characteristic allows reptiles to process seeds without destroying them by chewing. Consumption of fruits, relatively long passage times, microbial activity in the intestinal tract, and the ability to swallow food whole or in large portions are characteristic of successful seed dispersers (Schupp 1993), suggesting that Green Iguanas may be performing this ecosystem function.
Seed dispersal and germination facilitation by Green Iguanas has been studied in the Results from these experiments examining only juveniles may not apply to adults since their intestinal tract may not be developed completely. Juvenile Green Iguanas need to consume adult feces to acquire intestinal microbes (Troyer 1982) and it is not clear from the publications of these studies whether this had occurred. Similarly, seed germination might be affected by feeding preference (Traveset 1998), such that feeding captive Green Iguanas only one plant species may not accurately reflect their natural diet. Finally, these germination experiments were conducted using Petri dishes under laboratory conditions rather than in a more representative common garden experiment using natural soil and field conditions (i.e., natural variation in temperature, humidity, and precipitation).
Seeds passing through the gut of Green Iguanas can exhibit enhanced germination, inhibited germination, or be unaffected (Traveset 1998). A review of the literature indicates that effects of digestion on seed germination are common: 50% of reptiles, mammals, and bird seeddispersers either enhance or inhibit seed germination percentage and germination rate (Traveset 1998). Seed germination percentage reflects how many seeds germinate and germination rate reflects how fast a seed germinates after passage. Green Iguanas can potentially enhance germination by removing pulp from the seeds, which can contain germination inhibitors or support potentially infectious fungi and bacteria (Traveset 1998). Microflora and pH changes in Green Iguanas' intestinal tract may also kill infesting parasitic larvae (Fragoso 1997). Another way germination can be enhanced is by the abrasive effect of teeth or abrasive effect of food or other items during the ingestion process. This physical abrasion can disrupt the seed coat allowing a more rapid absorption of water (imbibition) and nutrients. Similarly, chemicals in the digestive tract can disrupt the seed coat, also allowing more rapid imbibition and nutrient uptake.
A review of reptile effects on germination shows that reptiles enhanced germination percentage in 28% of studies, inhibited germination percentage in 16% of studies, and had no effect on germination percentage in 56% of studies (Traveset 1998). Furthermore, reptiles enhanced germination rate in 47% of studies, inhibited germination rate in 16% of studies, and had no effect on germination rate in 37% of studies (Traveset 1998 This study quantifies the effects of Green Iguana digestion on seed germination rate and germination percentage for several tropical plant species. Because Green Iguanas are the largest and most abundant vertebrate herbivore in Puerto Rican mangrove forests, their impact on seed germination may be critical for plant community structure. Dispersal can lead to a higher probability of establishment, survival, and germination because seeds are released from predation and interspecific competition, and can colonize new sites with better germination conditions (Janzen 1970, Connel 1978, Thompson and Wilson 1978, Clark and Clark 1984, Dirzo and Dominguez 1986, Andresen 2000. Enhanced germination can lead to faster absorption of nutrients, water, light, and predator avoidance (Sarukhan et al. 1984, Andersen 1999). If Green Iguanas feed on particular species and enhance their germination, then they may facilitate the spread of these species. Previous studies of Green Iguana diet in Puerto Rico have found plant material of both native and non-native plants in the stomachs of Green Iguanas (Govender et al. 2012). Furthermore, seeds of the highly invasive Brazilian Pepper (Schinus terebinthifolius) have been found in the stomachs of Green Iguanas living in mangrove forests (Govender et al. 2012). If Green Iguanas promote the germination of non-native species more than native species, then they could pose a severe threat to mangrove communities by facilitating the spread of invasive species.

Our study differs from previous studies in multiple ways. Previous studies focusing on
Green Iguana germination effects were conducted in xeric habitats within the species' native range. This may not be representative of relationships outside of the native range where vegetation may differ because of differences in climate and patterns of species introductions.
Second, previous studies used juvenile Green Iguanas, which could limit the effect on germination since their digestive tracts may not be fully developed. Furthermore, juvenile size limits the size of fruit that can be consumed, which may exclude some plant species. Third, previous studies held Green Iguanas in captivity and offered a limited selection of fruits; this may also influence results of germination experiments since food selection might affect digestive processes. Finally, previous studies used Petri dishes to germinate seeds using only two treatments: digested seeds without feces and undigested seeds without fruit residues. Restricting experiments to these two treatments may limit the interpretation of germination results because the effects of the presence of feces and fruit on seed germination are not taken into account.
Seeds with associated feces and fruits may be more representative of conditions in natural systems.
Recognizing these limitations, our study builds on the work of previous studies in several important ways. First, the study was conducted in mesic habitat in Puerto Rico, where Green Iguanas have been introduced. This broadens the scope for understanding germination effects of Green Iguanas and the possible ecological effects of widespread introduction of Green Iguanas to tropical habitats. Second, our study uses scat from adult Green Iguanas with completely developed digestive tracts, thus eliminating possible effects on germination related to undeveloped juvenile digestive tracts. Third, with the exception of one fruit species fed to captive Green Iguanas, all scat samples were collected from Green Iguanas that fed freely in the environment. Fourth, we used soil for our germination experiment and included four treatments: digested seeds with feces, digested seeds without feces, undigested seeds with fruit, and undigested seeds without fruits.
Results from previous studies using only two treatments, undigested seeds without fruit and digested seeds without feces, provide information on chemical and mechanical effects of the gut on seed germination, but not on the effect of seeds being separated from fruit pulp (Samuels & Levey 2005). Using only these two treatments may limit interpretation of the results because seeds removed from the fruit pulp are more likely to germinate, regardless of gut treatment (Samuels & Levey 2005). Removing the pulp from the fruit may enhance germination by reducing high osmotic pressure caused by sugar levels and eliminating germination inhibitors such as lipids, glycoalkaloids, coumarin, abscisic acid, hydrogen cyanide, ammonia, and lightblocking pigments (Eveneri 1949, Cipollini andLevey 1997). Including the treatment of undigested seeds with fruit provides information on the effect of seed passage and the effect of the seed being removed from the fruit pulp (Samuels & Levey 2005). Similarly, including the treatment of digested seeds with feces may help interpret if Green Iguanas are altering germination by providing nutrient-rich microenvironment in scat. The effect of scat on germination may be independent from either removing pulp or the effects of gut passage.
Including these four treatments is a robust way to evaluate the overall effect of Green Iguana gut passage on seed germination.  1973) and is divided in three major geographical areas, a central montane interior, "mogotes" or karst hills, and the sandy coastlines (Holdrige 1967). Our study site was located on the Natural

METHODOLOGY
Reserve of Humacao (NRH). This reserve is located on the southeastern part of the island (65.46⁰ North, 18.10⁰ West). The NRH is protected land and managed by the Puerto Rico Department of Natural Resource and the Environment (DRNA using the Spanish acronym).
NRH has a total area of 10.5 km 2 , with most of this area being covered by wetland environments ( Fig. 1). The closest meteorological station to NRH records annual precipitation of 1,000-2,000 ml and daily annual maxima and minima temperature of 30.7⁰C and 20.7⁰C, respectively (NOAA & DRNA 1986). Before becoming a natural reserve in 1986, the NRH was used for agriculture at the beginning of the 21 st century followed by urbanization starting about midcentury (DRNA 2009, Cowardyn et al. 1976). This transition from agricultural landscape to residential landscape is characteristic of most of Puerto Rico's forest. The NRH is characterized by six main habitat types: 1) coastal lagoons, 2) mangrove forest, 3) herbaceous swamps, 4) Pterocarpus forest, 5) secondary coastal forest, and 6) coastal grasslands. A total of 187 plant species, 188 vertebrate species, and 47 invertebrate species have been identified for the NRH (Negron et al. 1983, DRNA 1986, Vilella-Gray 1997, Lopez 2005, Cruz 2005  Both species are evergreen trees. One is native to Puerto Rico (P. officinalis) and the other (P. indicus) has been introduced and is native to Southeast Asia (Little et al. 1974). Both species can reach over 15 m in height and grow mainly in, but are not restricted to, coastal wetlands including freshwater and brackish swamps, the landward side of mangroves, and along stream banks (Little et al 1974;Weaver 1997). The number of seeds per frutescence differs between these species with P. officinalis having one dry, winged seed and P. indicus having two seeds per winged frutescence (Little et al 1974). Ficus sp. refers to two possible species present at our study site: Ficus citrifolia and F. benjamina. Because both species contain similar fleshy seeds, it is not possible to identify which species was found in Green Iguana scat samples. Both species are evergreen trees, one native to Puerto Rico (F. citrifolia) and the other non-native (F. benjamina) has been introduced and is native to Southeast Asia (Little et al. 1974;Bonstein and Patel 1992;Veneklaas et al. 2002). Both species can reach over 18 m in height and produce fleshy fruit with one seed. Yellow Flamboyant (Peltophorum pterocarpum) is an evergreen tree that is not native to Puerto Rico, but native to Southeast Asia (Salah et al 2005). Peltophorum pterocarpum can grow in different habitats ranging from mangrove tidal forests to tropical highlands (Mail-Hong et al. 2003). Peltophorum pterocarpum can reach 19 m in height and has dry winged frutescence with 1-2 seeds inside (Little et al. 1974).

Time Budget
We documented behavioral observations of 47 Green Iguanas over five days in January 2013. The amount of time spent in the following categories was scored: basking, movement within vegetation, feeding, reproduction, movement, intraspecific aggressive behavior, head bobbing, and swimming. Individuals were not observed for more than three hours each (<1 min to 175 min). In addition to behavior, we determined the sex of each Green Iguana and the species of tree it occupied, if possible. If Green Iguanas occupied vegetation, then we recorded if the plant had flowers, fruits, or both.

Vegetation Assessment
To help identify to species the seeds found in Green Iguana scat, we developed a seed catalog for the most common plant species at our field site. The Staff at the University of Puerto Rico Herbarium advised us on plant collection methods and provided the necessary materials.
Branches were cut from each species and placed between layers of cardboard, desiccant paper, and newspaper. During May 2013, we collected 50 plant species and prepared specimens for identification. Specimens were deposited at the University of Puerto Rico Herbarium and DNRA Herbarium staff will identify each specimen to species level. At the same time, we collected seeds from 17 species, which were peeled, cleaned, and placed in plastic bags. From these seeds, we developed a seed catalog, which was used as a reference to identify seeds found in Green Iguana scat. At the end of our field season, we recorded the locations using a GPS unit of all trees of the four focal species within our primary study area.

Sample Collection
Scat sample collection started in mid-June and continued through the end of July 2013.
We collected Green Iguana scat primarily along paths constructed for hiking and bike riding.
Two to four researchers covered the same area to minimize the risk of missing a sample.
Samples were found primarily on the ground, but also on anthropogenic structures, such as kiosks, machinery, and abandoned irrigation systems. When a sample was located, we recorded its location using a GPS unit. As we have seen Green Iguanas defecate one or more pellets at a time, we were careful to collect all adjacent pellets as a single sample to avoid pseudoreplication.
When considering whether multiple pellets were from the same sample or not, we evaluated proximity, moisture content, structure, and composition. Scat pellets from a single defecation event are likely close to each other, and similar in moisture content, color, texture, and plant material. Canopy structure above a scat sample was also examined because branch arrangement provided information on the potential trajectory for defecated feces. Using plastic gloves to handle the scat, each sample was placed individually in a plastic bag for transportation. When looking for samples in the field, we paid special attention to areas around Pterocarpus officinialis because we often found large concentrations of Green Iguana scat around this tree species. Pilot field observations indicated up to five Green Iguanas on a single P. officinialis tree.
Preliminary searches for scat under coconut trees fields had little success, so we did not search this area. To avoid collection of scat from other species, we avoided sampling in areas with feral cats and dogs because some scat might look similar to a single pellet of Green Iguana scat.
6. Dispersal We dissected Green Iguana scat less than one week after collection from the field or from captive lizards with the exception of samples S1-S17, which were dissected two weeks after collection. Scat sample dissection consisted of fracturing the scat sample with tweezers to extract the seeds. A dissecting microscope and a magnifying glass were used to locate seeds and separate them from the scat sample, although in some instances seeds were visible to the naked eye. No water was used to separate the scat samples because water could initiate the germination process prior to planting the seeds. Seeds separated from scat were placed in a separate plastic bag and labeled accordingly. We referenced our seed catalog to identify seeds. Some seeds found in the scat samples were not present in our seed catalog and remained unidentified.

Germination
We used seeds collected from the field or from captive lizards for the seed germination The treatments with fruit and feces were added to provide a more accurate representation of natural germination conditions because seeds ingested by Green Iguanas are covered in feces and undigested seeds are covered by fruit residue.
Plastic germination domes with 72 individual cells were used. Domes were given a letter and cells were numbered. Soil was placed on each individual cell initially and seeds were then covered with soil. Clean commercial soil from a nearby area was used for our experiment. For treatments that included seeds and fruit, 0.5 g of fruit was added for all species except the Pond Apple (Annona glabra). For A. glabra, individual unaltered seeds were used because this species produces discrete units of pulp with each seed. This better represents the natural conditions for this species compared to adding 0.5 g of fruit. For treatments that included seeds and feces, 0.5 g of feces was added. For treatments that included clean seeds, tissue paper was used to clean the seeds. No water was used to clean any of the seeds in any treatment. We identified seeds in scat samples from four main tree species: 1) Pond Apple (Annona glabra), 2) Ficus sp., 3) Yellow flamboyant (Peltophorum pterocarpum) and 4) Pterocarpus sp. As the multiple species of Ficus and Pterocarpus present at our study site produce similar seeds, we could not identify these seeds to the species level. Nonetheless, these seeds are either Ficus citrifolia or Ficus benjamina, and either Pterocarpus officinalis or Pterocarpus indica, as these species are present at our study site (DRNA 2009). We collected seeds for the control treatments in the field at NRH from multiple individuals of each tree species. Seeds of C. papaya for the control treatment of our captive lizards were purchased at a grocery store. All four treatments were not possible for each species. Only A. glabra, Ficus sp., and C. papaya received all four treatments. In the case of Peltophorum sp., we were not able to completely clean the feces from the seeds so the treatment consisting of digested seeds without feces was not possible. Additionally, the treatment of seeds with fruit was not possible because Peltophorum sp. have fruits that lack pulp. Pterocarpus sp.
seeds had to be cut on the sides to fit the germination cells. Seeds were watered daily using a pressure sprayer for a period of 60 days. Wells were checked daily for seed germination.
Successful germination was recorded when seedling was first visible at the soil surface. If a seed had germinated, it was recorded, identified, and marked to avoid over counting. To ensure similar environmental conditions for all seeds, germination domes were rotated daily.

IV. Statistical Analyses
For the germination experiment, we measured two responses, days to germination and germination percentage, for both the experimental (seeds consumed by Green Iguanas) and control (seeds found in the environment) treatments. The statistical program JMP Pro was used to conduct all statistical analysis. We used analysis of variance (ANOVA) to test for differences among the four treatments for days to germination. Days to germination for A. glabra, Ficus sp., and P. pterocarpum were log transformed to address unequal variances among groups. A post hoc Tukey's Honestly Significant Difference (HSD) test was used to determine differences among the four treatments when an overall significant effect was found. Following Crawley (2007), we use a General Linear Model (GLM) to test for differences among the four treatments for days to germination. Because germination percentage was calculated as a proportion, the results were bounded from 0-1. We used a GLM with a binomial distribution and a logit-link function. The binomial denominator incorporated in the model is a two-vector response that accounts for the number of germinated seeds versus the number of planted seeds. Establishing the binomial denominator is important because it takes into consideration sample sizes (Crawley 2007).

I. Time Budget
In a period of five days, we observed 47 Green Iguanas. Green Iguanas spent most of their time basking, followed by reproduction, movement within vegetation, other movement, intraspecific aggression, head bobbing, feeding, and swimming (Fig. 2). Green Iguanas most commonly selected white mangrove (Laguncularia racemosa) trees, followed by almond trees (Terminalia catappa), Pterocarpus sp., coconut palms (Coccos nucifera), flamboyant (Delonix regia), Albizia procera, and Thespesia populnea. Green Iguanas were also found on the ground and on unidentified trees (Fig. 3).

II. Scat Samples
We collected 258 Green Iguana scat samples ( Table 1). Out of these, almost half (n = 122) contained seeds. Green Iguana scat differed in the composition of seeds from various plant species. Most scat samples contained only one species of seed, followed by scat samples with two species of seeds, and the remaining scat samples had three species of seeds (Table 1). Seeds of five main species were found in Green Iguana scat; however, only four of these species could be identified: Annona glabra, Ficus sp., Pterocarpus sp., and Peltophorum pterocarpum (Table   1). Of these four species, A. glabra was the most abundant species found in Green Iguana scat samples, followed by Ficus sp., P. pterocarpum, and Pterocarpus sp. (Table 2). The unidentified abundant seed was present in 30% of the Green Iguanas scat samples. Numbers of seeds found on Green Iguana scat by species differed among plant species (Table 2). Of these four species, A.
glabra was the most abundant with almost half of the seeds, followed by Ficus sp., P.
pterocarpum, and Pterocarpus sp. (Table 2). No animal material was found in Green Iguana scat.

III. Germination
The effect of Green Iguanas on germination percentage and days to germination differed among the four focal plant species: Annona glabra, Ficus sp., Pterocarpus sp., and Peltophorum pterocarpum (Table 3). For A. glabra, there was no significant difference in the percentage of seeds that germinated among treatments (P = 0.77) (Fig. 4, Tables 3 & 4). Moreover, no significant difference was found among treatments for days to germination for A. glabra (F 3,19 = 2.01, P = 0.48) ( Table 6, Fig. 5). Thus, Green Iguana effects on seed germination of native A. glabra were neutral, meaning they did not enhance or inhibit germination in this species.
Germination percentage differed among treatments for C. papaya (P = 0.05; Table 4) with digested seeds with feces having higher germination percentage than other treatments (Fig.   6). Mean days to germination differed among treatments for C. papaya (F 3,275 = 4.8, P = 0.003; Tables 5 & 6) with digested seeds without feces taking longer than other treatments to germinate (Table 6, Fig. 7).
For Ficus sp., there was no significant difference among treatments for germination percentage (P = 0.73; Tables 3 & 4). In contrast, the mean days to germination differed among treatments for Ficus sp. (F 3,116 = 3.72, P = 0.01; Tables 5 & 6) where digested seeds with feces, digested seeds without feces, and undigested seeds without fruit had shorter days than undigested seeds with fruit (Table 6, Fig. 9).

IV. Dispersal
Minimum distance of Green Iguana scat (Fig. 14) containing seeds of the focal species to the nearest mature tree of the same species differed among the four species: Annona glabra, Ficus sp., Pterocarpus sp., and Peltophorum pterocarpum ( Table 7). Seeds of A. glabra were dispersed the farthest (Fig. 15), followed by Pterocarpus sp. (Fig. 16), then Ficus sp. (Fig. 17), and finally P. pterocarpum (Fig. 18). For A. glabra, most seeds were dispersed between 40-50 m from the nearest A. glabra tree (Fig. 19). In contrast, most seeds of Ficus sp. were not transported far, nearly 50% of seeds were found within 5 m of a mature tree (Fig. 20).

DISCUSSION
During January, Green Iguanas spent the vast majority of their time basking, and feeding made up less than 1% of their time budget (Fig. 2). (2) they may preferentially occupy dead mangroves to bask; or (3) they may be more easily observed when mangroves are dead (Lopez-Torres et al. 2011;Garcia-Quijano et al. 2011). If Green Iguanas spend little time feeding (Fig. 3) on mangrove and use other tree species to bask and feed (Fig. 4), then little evidence exists for negative impacts on mangroves in introduced habitats (but see Lopez-Torres et al. 2011;Carlo and Garcia-Quijano 2008). This is consistent with reports of Green Iguanas not being detrimental to mangrove communities in their native habitat (Henderson 1974, Lara-Lopez andGonzalez-Romero 2002).
Whether Green Iguanas are strict herbivores has been debated for many years (Swanson 1950;Rand et al. 1990;van Marken 1993;Townsend et al. 2005;Lopez-Torres et al.2011;Garcia-Quijano et al.2011;Govender et al. 2012). It has been reported that Green Iguanas consume carrion (Loftin and Tyson 1965), juveniles consume insects (Swanson 1950), and in recent years consumption of tree snails Drymaeus multilineatus and crabs Uca sp. has been reported (Townsend et al. 2005, Govender et al. 2012. Nonetheless, lack of stomach content information and the anecdotal nature of most observations questions the validity of these reports. Scat samples of Green Iguanas collected in our study revealed no evidence of animal material, suggesting that Green Iguanas are strict herbivores at this site in Puerto Rico. This is consistent with feeding studies in the native range (Rand et al. 1990;Lara-Lopez and Gonzalez-Romero 2002). Seeds found in Green Iguana scat samples belong to native and non-native tree species, including species with both fleshy and dried fruits. For most species, the physical appearance of seeds after ingestion and gut passage was different than that of undigested seeds. Seeds of A.
glabra that passed through the gut of Green Iguanas were darker and lacked pulp, whereas Ficus sp. seeds also lacked pulp, Pterocarpus sp. lacked wings, and P. pterocarpum seeds became softer, but remained inside the frutescence. Lack of pulp and loss of wings suggest mechanical processes (e.g., chewing and abrasive effects of gut passage), whereas changes in color and texture suggest chemical processes during gut passage. Thus, regardless of the effect on germination, Green Iguanas change the chemical and physical properties of seeds in some species in their introduced range by means of ingestion and gut passage.
At our study site, almost half of samples contained seeds (Table 1). In contrast, fewer than 7% of scat samples from within its native range contained seeds (Rand et al. 1990, Lara-Lopez andGonzalez-Romero 2002). Scat samples also differed in the number of plant species per scat sample between native and non-native sites. In their native range, scat of Green Iguanas contain one plant species 52% of the time, two species 23% of the time, three species 23% of the time, and four species 3% of the time (Rand et al. 1990). At our study site, scat contained one plant species 69% of the time, two species 27% of the time, and three species 3% of the time (Table 1). This suggests that Green Iguanas may have a more diverse diet in native compared to non-native sites. However, we were unable to identify all seeds to species in scat samples from our site, and Green Iguanas also eat other plant structures, such as leaves and flowers, which were not quantified in our study. Furthermore, we did not assess the availability of plant species, which would allow us to determine the extent to which Green Iguanas selectively feed on particular plant species. Nonetheless, we predict that Green Iguanas will feed on a wider variety of plant species in their introduced range because predation pressure is lower compared to their native range (Morales-Mavil et al. 2007). Reduced predation pressure could result in increased foraging times and distances from refugia.
Coastal wetlands throughout the neotropics exhibit associations with particular plant species, such as P. officinalis and A. glabra (Weaver 1997 pterocarpum is also associated with mangrove communities in Southeast Asia in areas with infrequent flooding; nonetheless, they are not mentioned as a tree species associated with mangroves in Puerto Rico (Weaver 1997). Given the species associated with coastal wetlands and also present at our study site, we found that Green Iguanas consumed fruits of three species: glabra (Lara-Lopez and Gonzalez-Romero 2002). Furthermore, observational studies in Puerto Rico documented Green Iguanas occupying A. glabra trees, but they were not observed feeding on this species (Figueiredo de Andrade et al. 2011). In the case of Ficus sp., vegetative material of this species was found in stomach contents of Green Iguanas in its native range, but no seeds were present (Rand et al. 1990;Lara-Lopez and Gonzalez-Romero 2002). These differences in consumption between the native and non-native range suggest that food selectivity may differ between habitats. Any shift in Green Iguana diet in non-native range could be associated with preferences for certain plant species, lack of competitors feeding on the same species, or changes in foraging patterns associated with lack of predation. The latter could be particularly true for A.
glabra because Green Iguanas likely need to handle this large fruit on the ground for longer periods of time to feed on it as compared to other species (J. Burgos-Rodriguez, pers. obs.). In native habitats, such exposure could increase risk of predation.
Green Iguanas may prefer A. glabra fruit compared to other species since seeds from A.
glabra represented almost have of all seeds found in scat ( . Germination effects of Green Iguanas in Puerto Rico did not consistently differ between native and non-native plant species. Germination percentage was not enhanced for any species, was neutral for native A. glabra (Fig. 4) and Ficus sp. (Fig. 8), but was inhibitory by reducing germination percentage in C. papaya (Fig. 6), non-native P. pterocarpus (Fig. 10), and Pterocarpus sp. (Fig. 11).
Germination was enhanced by reducing the number of days to germination in Ficus sp. (Fig. 9), non-native P. pterocarpum (Fig. 11), and Pterocarpus sp. (Fig. 13). However, the number of days to germination did not differ among treatments for native A. glabra (Fig. 5) and was inhibited by increasing days to germination for non-native C. papaya (Fig. 7). On the other hand, we found differences in germination between fleshy and dry fruits. Dry fruits of non-native P.
pterocarpum and Pterocarpus sp. exhibited the same effect on germination. When ingested these species showed both a decrease in the percentage of seeds that germinated and an increase in the number of days to germination. This suggests the type of seed may influence the effects of ingestion on germination. Seeds of non-native P. pterocarpum and Pterocarpus sp. are both dry, winged and indehiscent, meaning they cannot split open on their own (Little et al 1974). Green Iguanas may eliminate wings and facilitate the opening of such indehiscent seeds, thus enhancing germination. However, some seeds may be damaged by mechanical or chemical processes, and can be destroyed by digestion (Traveset 1998).
Seed germination can be altered by either removal of fruit pulp, providing nutrient-rich, moist microhabitats in scat, or by mechanical or chemical processes (Samuels and Levey 2005).
Our study showed that Green Iguanas affect this process in multiple ways. For native A. glabra, ingestion of seeds does not appear to effect germination. However, seed germination was overall very low for A. glabra (Table 3). This could be attributed to limitations in the experimental design related to watering patterns and the observation period. Although previous studies of A.
glabra show low germination in general, they watered seeds daily until the soil was saturated and observed germination for 120 days (Infante and Moreno-Casasola 2005). Our experimental design did not account for the tendency of A. glabra to germinate in flooded soils and the extended time needed for germination. For non-native C. papaya, germination percentage was lower for digested seeds with feces (Table 4, Fig. 6). On the other hand, C. papaya seeds digested with no feces took longer on average to germinate compared to seeds undigested with no fruit (Fig. 7, Table 4). We can conclude that the mechanical or chemical action of gut passage in Green Iguanas may slightly inhibit seed germination in C. papaya, but only when their feces is not present. For Ficus sp., the percentage of seeds germinating did not differ among treatments, thus we can conclude that ingestion of seeds is not important for this aspect of germination. On the other hand, undigested Ficus sp. seeds planted with fruit took longer to germination compared to undigested seeds with no fruit (Fig. 9, Table 7). This suggests the separation of seeds from fruit pulp enhances germination. This may be accomplished by ingestion by Green Iguanas, but this is confounded with the mechanical and chemical action of ingestion. Previous studies attempted to determine the effect of Green Iguanas on germination percentage and days of germination for Ficus sp. in native habitats, but juveniles Green Iguanas were unable to consume large Ficus sp. seeds (Benitez-Malvido et al. 2003). Such a comparison could be important because Ficus sp. have been categorized as keystone species for herbivores due to abundant year round fruit presence (Santinelo et al. 2007). Because P. pterocarpum and Pterocarpus sp. lack pulp and we could not clean fecal material from the seeds, we examined only undigested seeds with fruit and digested with feces. We were not able to determine whether Green Iguanas were important for removing pulp from seeds, have mechanical or chemical effects that alter germination, or whether scat microhabitat presence affects germination. As the effect of Green Iguanas on germination percentage and days to germination was inconsistent among plant species (e.g., C. papaya, P. pterocarpum, Ficus sp., and Pterocarpus sp.), it was difficult to summarize the overall effect on germination. Days to germination has been suggested as a better measure because seeds that germinate faster avoid predation and produce more vigorous seedlings with greater survival probabilities compared to late recruitment of conspecifics (Sarukhán et al. 1984;Andersen 1999). Also germination percentage could be inconsequential when plant species produce an overabundance of seeds. Green Iguanas could have an overall negative effect on germination of C. papaya since digested seeds with feces inhibit germination. On the other hand, Green Iguanas may have an overall positive effect on germination of P. pterocarpum and Pterocarpus sp. because digested seeds with feces germinate faster.
Minimum distances between mature trees and scat samples containing seeds of the same species indicate the patterns by which seeds are dispersed. Dispersal distance varied considerably among species. Although minimum dispersal distances may appear to be low for some species (e.g., 8 m for P. pterocarpum), distances were large enough that seeds of all species would be dispersed beyond the canopy of parent trees. Minimum distances showed no trend for dispersal of fleshy versus dry fruits. The potential for dispersal appears to be greater for A. glabra compared to other species in our study because of the large number of seeds of A. glabra per scat sample and the higher frequencies of relatively large dispersal distances for this species (Fig. 19; Tables 2 & 5). The Escape Hypothesis suggest that A. glabra will benefit from distant dispersal away from the source tree because proximity to the parent tree increases mortality due to density dependent effects such as increased predation, pathogen attacks, or seedling competition (Janzen 1970, Collins 1973 Frequency distributions for dispersal distances between scat and the nearest parental tree could not be constructed for P. pterocarpum and Pterocarpu sp. species because of low number of these seeds in Green Iguanas scat samples.
When considering our focal species, A. glabra is primarily dispersed by water, Ficus sp.
is dispersed by herbivores, and P. pterocarpum and Pterocarpus sp. are dispersed by air (Little et al. 1976;Weaver 1997). The ability of Green Iguanas to disperse these species may be beneficial not only for species dependent on animal-mediated dispersal like Ficus sp., but also for species dispersed by air and water. Although Green Iguanas do not have extensive home ranges (Dugan 1982;Morales-Mavil et al. 2007), their ability to move across the landscape creates the potential for seeds of these species to reach environments not otherwise easily accessed, such as upstream habitats, inland habitats that do not get flooded, and the interiors of dense forests where air dispersal may not be successful. Another aspect of Green Iguanas that could facilitate dispersal is the long retention time of seeds in the gut. Adult Green Iguanas have an average retention time of 5.5 days (Troyer 1984). Retention times of this length may allow Green Iguanas to move from feeding sites and disperse seeds away from the parent plants. Moreover, seeds ingested by Green Iguanas were defecated mostly intact, which is a good indicator of an effective disperser (Schuppe 1993).
To assess the effect of Green Iguanas on these plant species and their communities, it is important to integrate effects of germination percentage, days to germination, dispersal potential, and the life history. Green Iguanas feed on non-native species where they have been introduced in Puerto Rico (Govender et al. 2012). This study also demonstrates that Green Iguanas consume inconclusive, we suggest that Green Iguanas may have a positive effect on the ecosystem, in this case, by dispersing native A. glabra seeds. This is consistent with some studies that suggest dispersal to microhabitats is more important than the effect of gut passage on seeds (Rey and Alcántara 2000;Traveset et al. 2003).
Effects of Green Iguanas on the ecosystem through seed germination and dispersal may be influenced by fruit availability. Of our four focal species, P. pterocarpum and Ficus sp. fruit year round, and A. glabra and Pterocarpus sp. fruit between March-November (Little et al. 1976). Because P. pterocapum is non-native and fruits almost year round, Green Iguanas disperse them and enhance their seed germination, and their role in mangroves communities in Puerto Rico has not been studied, P. pterocarpum represents the highest potential risk of the species at our study site in Puerto Rico. Although Lugo (2009)