Introduction
Seed dispersal involves horizontal and vertical movements of seeds away from the parent plant (Chambers and MacMahon, Reference Chambers and MacMahon1994). It is a key process in the life history of plants that allows them to occupy new environments (Howe and Smallwood, Reference Howe and Smallwood1982) and to escape the attack of pathogens, mortality by predation and seedling competition under the parent plant (Janzen, Reference Janzen1970). At a larger scale, long-distance dispersal may also be crucial for plants, driving species metapopulation structure and dynamics, gene flow, colonization opportunities and range expansion, and becomes particularly relevant in response to climate change and habitat fragmentation (Cain et al., Reference Cain, Milligan and Strand2000). Different abiotic (e.g. wind and water) and biotic agents (animals) are involved in fruit and seed dispersal until seeds reach the place where they will germinate and a new plant will establish (van der Pijl, Reference van der Pijl1982). Many plants have developed mutualistic relationships with animals to assure seed dispersal, in some cases offering attractive and nutritious fruits as rewards to endozoochorous dispersers (Jordano, Reference Jordano and Fenner2000; Jordano et al., Reference Jordano, Forget, Lambert, Böhning-Gaese, Traveset and Wright2011).
In temperate communities, more than 60% of tree species are dispersed by vertebrates, and it has been estimated that at least 50% and often 75% or more of the tree species in tropical forests have fleshy fruits adapted for animal dispersal (Howe and Smallwood, Reference Howe and Smallwood1982). Endozoochory involves fruit consumption by animals and seed dispersal through defecation; therefore, seed retention time in the digestive tract and distance travelled by animals during said time will affect seed dispersal distances (Nathan and Muller-Landau, Reference Nathan and Muller-Landau2000). Added to this, seeds can be scarified during the digestive process, increasing germination speed but with the risk of also increasing seed mortality if seed coats weaken too much. However, final seedling establishment depends on predation or subsequent movements of seeds from faeces, and on availability of suitable sites for plant establishment (Janzen et al., Reference Janzen, Demment and Robertson1985).
Most seed dispersal studies over the last decades have taken the approach based on the seed dispersal effectiveness (SDE) framework (Schupp, Reference Schupp1993; Schupp et al., Reference Schupp, Jordano and Gómez2010). Disperser effectiveness was defined as the contribution a disperser makes to plant fitness, ideally measured as the number of reproductive adults recruited through the activity of a dispersal agent (Schupp, Reference Schupp1993). Later, it was proposed that SDE is a broader and more comprehensive concept, in that, it not only considers the effectiveness of dispersal that a single animal species provides but also the whole dispersal effectiveness that a plant receives from the assemblage of dispersers (Schupp et al., Reference Schupp, Jordano and Gómez2010). In addition, it is important not only to consider the effectiveness of seed disperser animals in plant fitness and population dynamics but also to highlight the benefits that animals obtain from seed dispersal mutualism, in the form of effective food resource provisioning (Schupp et al., Reference Schupp, Jordano and Gómez2017). According to the SDE model applied to endozoochorous seed dispersal systems, the effectiveness of animals as seed dispersers depend on two components, the quantity of dispersal, or number of seeds dispersed, and the quality of dispersal, which depends on the quality of the treatment in the digestive tract and the quality of seed deposition, which in turn determines the probability that a dispersed seed produces a new recruit (Schupp, Reference Schupp1993; Schupp et al., Reference Schupp, Jordano and Gómez2010).
The quality of seed treatment in the digestive tract depends on the animal's morphological and physiological characteristics, such as digestive system, body mass and diet (Traveset, Reference Traveset1998). During the passage, through the mechanical action of digestion, a seed's coat or endocarp can be altered by chewing. When the seed is consumed with other hard food items, including other seeds, it is likely to suffer mechanical breakdown (Samuels and Levey, Reference Samuels and Levey2005). Mechanical alteration of the seed coat or endocarp may facilitate imbibition at deposition sites, thus promoting germination. Alternatively, the gut's mechanical action may also destroy seeds (Samuels and Levey, Reference Samuels and Levey2005). Through the chemical action of the digestive tract, fluids may alter the seed coat or endocarp, thus affecting germination patterns. Digestive enzymes and stomach acids serve to break foods down, including seeds. This chemical treatment may also be affected by symbiotic bacteria and protozoa, which can digest structural cell wall polymers during fermentation (Traveset, Reference Traveset1998; Traveset et al., Reference Traveset, Robertson, Rodríguez-Pérez, Dennis, Green, Schupp and Westcott2007).
Previous studies have shown that seed treatment in the digestive tract of animals has a more significant effect on seeds of trees from temperate than from tropical regions (Traveset, Reference Traveset1998; Traveset and Verdú, Reference Traveset, Verdú, Levey, Silva and Galetti2002). An example in the temperate Monte region could be the key tree species Prosopis flexuosa DC. (Fabaceae, Mimosoideae) (Álvarez and Villagra, Reference Álvarez and Villagra2009). The fruit of P. flexuosa is a modified indehiscent pod with a thin epicarp, a mesocarp that can be fleshy, sugary or fibrous and several endocarp segments (Burkart, Reference Burkart1976). Seed germination is hindered by physical dormancy and seed scarification by different agents increases germination (Catalán and Balzarini, Reference Catalán and Balzarini1992; Peláez et al., Reference Peláez, Bóo and Elia1992; Peinetti et al., Reference Peinetti, Pereyra, Kin and Sosa1993; Campos and Ojeda, Reference Campos and Ojeda1997; Ortega-Baes et al., Reference Ortega-Baes, de Viana and Suhring2002; Campos et al., Reference Campos, Peco, Campos, Malo, Giannoni and Suárez2008). Only a few days after reaching the soil, P. flexuosa fruits are removed by animals with different functional roles in the dispersal of the seeds. On the one hand, some opportunistic frugivores disperse seeds by endozoochory, such as medium-sized native mammals (e.g. Dolichotis patagonum, Lagostomus maximus and Lycalopex gymnocercus), large native herbivores (e.g. Lama guanicoe and Rhea americana) and non-native animals (e.g. Lepus europaeus, Sus scrofa, Equus asinus and Bos taurus; Campos and Ojeda, Reference Campos and Ojeda1997; Campos et al., Reference Campos, Peco, Campos, Malo, Giannoni and Suárez2008, Reference Campos, Campos, Mongeaud, Borghi, de los Ríos and Giannoni2011, Reference Campos, Velez, Miguel, Papú and Cona2018; Campos and Velez, Reference Campos and Velez2015). On the other hand, small rodents (e.g. Graomys griseoflavus, Akodon dolores, Eligmodontia typus, Calomys musculinus and Microcavia maenas) practice food hoarding, making both larder-hoards and scatter-hoards with different effects on seed survival (Campos et al., Reference Campos, Giannoni, Taraborelli and Borghi2007, Reference Campos, Campos, Giannoni, Rodríguez, Albanese and Cona2017; Giannoni et al., Reference Giannoni, Campos, Andino, Ramos-Castilla, Orofino, Borghi, de los Ríos and Campos2013). In this relationship among P. flexuosa and its assemblage of dispersers, previous studies have shown that quantitative aspects of SDE directly related to animal visits and fruit removal are affected by changes in habitat heterogeneity under different land uses (Campos et al., Reference Campos, Campos, Miguel and Cona2016; Bessega et al., Reference Bessega, Pometti, Campos, Saidman and Vilardi2017; Tabeni et al., Reference Tabeni, Miguel, Campos and Cona2017; Miguel et al., Reference Miguel, Cona and Campos2017, Reference Miguel, Tabeni, Cona and Campos2018a,Reference Miguel, Jordano, Tabeni and Camposb). Although the development of the SDE model for P. flexuosa began years ago, using an approach that combines field and laboratory experiments, the aim of the present study is to provide data to fill in some gaps in the model. To do this, we focused on the experimental assessment of seed treatment quality in the digestive tract of some native animals. As was pointed out by Picard et al. (Reference Picard, Papaix, Gosselin, Picot, Bideau and Baltzinger2015), studies of this kind remain rare because they rely on individual monitoring in controlled conditions, with heavy logistic constraints and associated with small sample sizes, mainly when wild species are considered. In particular, it remains important to establish whether variations occur among different species of animals consuming the same fruits and whether such variations could be attributed to differences among animal species.
We tested differences among five disperser species covering a variety of taxa, diet, digestive system and body size: Dolichotis patagonum (Mammalia, Rodentia; herbivore; hindgut fermenter; ≈8 kg), Lycalopex gymnocercus (Mammalia, Carnivora; carnivore and opportunistic frugivore; monogastric hindgut fermenter; ≈6 kg), Rhea americana (Birds, Rheiform; herbivore; hindgut fermenter; ≈30 kg), Chelonoides chilensis (Reptilia, Testudines; herbivore; hindgut fermenter; ≈1.5 kg) and Lama guanicoe (Mammalia, Artiodactyla; herbivore; foregut fermenter; ≈90 kg). The animal species were assumed to have different retention times and digestive treatment effects because of their variation in taxa, digestive system and body mass (Illius and Gordon, Reference Illius, Gordon and Highes1993). Our objective was to analyse differences among species in the quality of seed treatment in the digestive tracts considering: (1) mean retention time of ingested seeds, (2) recovery of viable seeds, (3) seed germination in comparison with seeds collected from trees and (4) germination of seeds after two different periods of retention in the gut. We tested three predictions:
(1) Because L. guanicoe is a foregut fermenter with a digestive process that includes rumination (‘ruminant-like’; Hume, Reference Hume1989) and tortoises, such as C. chilensis, have low metabolic rate and food intake (Bjorndal, Reference Bjorndal, Mackie and White1997), the mean seed retention time will be longer in these species compared to the others.
(2) Because of the long mean retention time, total recovery and germination will be lower for seeds ingested by L. guanicoe and C. chilensis than for those consumed by the other species.
(3) Germination will be higher for seeds recovered during the first period of passage through the digestive tracts of all species.
Materials and methods
Feeding and seed recovery experiments
The fruits of P. flexuosa were collected from 20 adult trees randomly selected in Ischigualasto Provincial Park (29°55′S, 68°05′W, San Juan Province) during the fruiting seasons (January to March) of 2014–2017. Fruits were stored in paper bags in a refrigerator at 5°C until the start of experiments, following the methodology proposed by Cony (Reference Cony1993) for preserving seeds. The feeding experiments were conducted with captive animals from three institutions: the Ecological Reserve and Wildlife Rehabilitation Centre (Mendoza), the Mendoza Zoo and the Centre for Rehabilitation of Wild Fauna, Environmental Education and Recreation (San Juan). Animals were kept and handled according to the Animal Care and Use Committee of the American Society of Mammalogists (Sikes et al., Reference Sikes and Gannon2011) and conforming to the legal requirements of the Agencies of Renewable Natural Resources of Mendoza and San Juan (No. 444/2016). Experiments were performed from November 2016 to December 2018 with captive adult individuals of D. patagonum (N = 8), L. gymnocercus (N = 3), L. guanicoe (N = 6), R. americana (N = 5) and C. chilensis (N = 6). Each animal was isolated in a clean enclosure and fed on its usual food and with drinking water provided ad libitum; therefore, no adaptation period was needed before the experiments. The selected individuals showed no signs of stress during the trials because they had been kept in captivity since an early age and were used to human presence. They were active and in good health conditions throughout the study. During the first day of the trial, and for once, we offered them fruits containing controlled amounts of seed (between 30 and 300 seeds depending on the species) mixed with their usual food to facilitate a good acclimation to the experimental condition. We selected apparently undamaged fruits, discarding by external observation the ones with holes in the epicarp produced by the emergence of adult insects (Velez et al., Reference Velez, Chacoff and Campos2018). We quantified the number of seeds provided, assuming a single seed in each pod segment. We began the experiments in the morning, ensuring that each animal ingested as many seeds as possible (Table 1). We interrupted the first-day feeding session when it appeared that all seeds had been ingested or when the animal refused to eat, and the remaining seeds were removed. The days following ingestion of the seeds, each animal had free access to freshwater and received its usual food every day. The faeces of each individual were collected daily and stored in labelled paper bags until processing. To ensure that we covered retention time until reaching a plateau in the cumulative proportion of retrieved seeds when it was logistically possible, we collected all fresh faeces for 2 weeks, except for tortoise faeces that were collected for 35 d. To recover the seeds, each faecal sample was thoroughly washed under running water in a sieve (71 μm). Seeds were dried at room temperature and stored in the laboratory until the germination experiment.
Table 1. Animal species that ingested P. flexuosa seeds
Experimental details are shown. Letters indicate differences in mean retention time (days) among species.
Germination and viability of seeds
Germination and viability were measured for seeds ingested by animals. In July 2018 and May 2019, germination tests were performed in incubators (Precision GCA Corporation, Scientific Model 818, Chicago, Illinois, USA) in the dark, and at a constant temperature of 30°C (Cony and Trione, Reference Cony and Trione1996). Seeds were placed in sterile, plastic Petri dishes (9 cm diameter) containing filter paper disks resting on a thin layer of cotton, all materials having previously been sterilized. Ingested seeds were previously immersed in a solution of sodium hypochlorite (2%) for 10 min to remove fungi and superficial bacteria (Sauer and Burroughs, Reference Sauer and Burroughs1986). Dishes were initially moistened with a suspension of Captan fungicide at 2 g l−1, being thereafter moistened as needed with sterile water. Apparently, healthy seeds recovered from animal faeces were cultivated, with sample sizes determined by seed availability. Seeds recovered from each individual by deposition day were kept in separate Petri dishes containing a maximum of 25 seeds. Dishes were randomly repositioned every week to avoid a chamber position effect. Germination, identified as visible radicle protrusion, was recorded every day for 30 d. In order to know the viability of the seeds offered to animals considering the condition and provenance of seeds, we carried out germination tests with seeds collected from 20 trees. Collected seeds were mixed in a pool. We used visually healthy seeds with endocarps and made 25 replicates of 20 seeds. It should be noted that there is a 32% pre-dispersal loss of P. flexuosa seeds due to abortion and insect predation. In many cases, insects die during development and do not emerge, then endocarps remain intact and seed loss cannot be externally determined (Velez et al., Reference Velez, Chacoff and Campos2018).
To establish whether seeds that had failed to germinate in the previous experiments were viable, non-germinated seeds from trees and ingested by animals were subjected to the standard treatment with a 2,3,5-triphenyltetrazolium chloride test (Pili-Sevilla, Reference Pili-Sevilla1987), which detects seed viability by staining the embryo tissue pink/red. Then, those seeds that germinated and those that were positive in the viability test were considered viable seeds.
Data analysis
Seed recovery was calculated as the proportion of viable seeds daily removed from faeces to the viable seeds consumed by individuals. Because animals were fed on whole fruits containing viable and non-viable seeds, the pre-ingestion loss was estimated using the seeds collected from trees.
Mean retention time is the time that the average digested particle remains in the digestive tract (Warner, Reference Warner1981) or the time for a 50% seed recovery. Expressing the results in terms of mean retention time is a convenient method for comparing the curves of cumulative seed recovery over time along their entire lengths (Castle, Reference Castle1956). The mean retention time of seeds ingested by each individual was computed as the weighted sum of seeds recovered per day related to the total number of seeds recovered (Gardener et al., Reference Gardener, McIvor and Jansen1993; Razanamandranto et al., Reference Razanamandranto, Tigabu, Neya and Odén2004). Differences among species in mean retention time were analysed using a Kruskal–Wallis rank-sum test with an a posteriori pairwise Wilcox test.
In order to analyse differences in the recovery of seeds ingested by species, we used a generalized linear mixed model (GLMM) with a binomial error distribution and a logit link function because there was no overdispersion. The total number of seeds recovered from faeces was used as a numerator in the response variable, and the number of seeds ingested by individuals as a denominator. Individuals were considered a random effect nested within species.
Differences in germination were analysed between seeds recovered from faeces and seeds collected from trees using a GLMM with a binomial error distribution and a logit link function. The total number of germinated seeds was used as a numerator in the response variable. The number of seeds recovered from faeces of individuals or the number of viable seeds in the replicates of seed collected from trees was used as denominators.
In order to evaluate the effect of retention time in the digestive tract on seed germination, we estimated for every species the time needed for an 80% seed recovery; using this information, we grouped the data into two periods: the first considering the days required to reach 80% seed recovery (P1) and the latter including the days until the end of the experiments (P2). We chose this percentage of recovery because values showed low variations among individuals of each species. We used a GLMM with a binomial error distribution and a logit link function to analyse the effect of periods (P1 and P2) on the percentages of germinated seeds. For this model, the number of germinated seeds during every period was used as a numerator in the response variable and the total number of germinated seeds as a denominator. Individuals nested within species were considered as a random effect. All analyses and graphs were performed in R version 3.6.1 (Team RC, Reference Team2016).
Results
A mean of 65% (SE = 4.5) of the seeds obtained from trees were viable, and it represents the viability of seeds ingested by animals in our experiment. Similar results were found by Velez et al. (Reference Velez, Chacoff and Campos2018). Cumulative curves of seed recovery over time for each species showed the same general sigmoid shape. After an initial rise, the curves rose sharply and gradually flattened in the final stages of seed recovery. It is important to note that the seed recovery for C. chilensis lasted 30 d (Supplementary Fig. S1). The mean retention time of ingested seeds varied among the studied species. Chelonoidis chilensis showed the longest mean retention time, followed by L. guanicoe, L. gymnocercus, R. americana and D. patagonum in the decreasing order. Only between C. chilensis and D. patagonum, there was a statistically significant difference (Kruskal–Wallis chi-squared = 20.50, df = 4, P-value < 0.001; Table 1).
Total recovery of seeds after passage through digestive tracts was around 50% of the viable ingested seeds for most animal species, except for L. guanicoe (Fig. 1 and Table 2). The model fitted to analyse the percentages of seeds recovered from faeces of different animals showed that the seeds ingested by L. guanicoe were recovered to a lower percentage than those ingested by the other species (Table 2).
Fig. 1. Total percentages of recovered seeds of P. flexuosa after passing the digestive tracts of animal species. Letters show significant differences among species. The line in the box indicates the median value of the data. The upper and lower hinges of the box indicate the 75th and 25th percentiles of the dataset, respectively. The ends of the vertical lines indicate the minimum and maximum data values.
Table 2. Percentages of total recovered seeds after ingestion by animals, and total germination of ingested seeds and seeds collected from trees
Percentages of germinated seeds after the two periods of retention in the digestive tracts of animal species. Mean percentages (±SE) are shown. Values of recovered seeds were recalculated considering the viability obtained for ingested seeds (65%). Letters indicate significant differences among seed sources.
a Significant differences between periods.
The model fitted to test whether the percentages of germinated seeds differ among sources (recovered from faeces and collected from trees) showed that germination was significantly low for seeds ingested by R. americana and L. gymnocercus, and for seeds collected from trees (Tables 2 and 3). Total seed germination was over 50% for seeds consumed by L. guanicoe, D. patagonum and C. chilensis (Fig. 2).
Fig. 2. Total percentages of germination of P. flexuosa seeds recovered after passing the digestive tracts of animal species. Letters show significant differences among species. The line in the box indicates the median value of the data. The upper and lower hinges of the box indicate the 75th and 25th percentiles of the dataset, respectively. The ends of the vertical lines indicate the minimum and maximum data values.
Table 3. Results of generalized linear mixed models with a binomial error structure testing whether the proportion of P. flexuosa seeds recovered from faeces differs among species, the proportion of germinated seeds differs among sources (faeces and trees), and between periods of retention (P1 and P2) in the digestive systems of animals
The germination analysis of seeds recovered after two periods of retention in the digestive tract showed a lower germination for seeds recovered during P2, that is, the seeds that remained in the digestive tracts after 80% of the ingested seeds had been recovered (Tables 2 and 3).
Discussion
The different ways in which seeds are treated in the digestive systems of endozoochorous animals is one of the components of dispersal quality (Schupp, Reference Schupp1993; Schupp et al., Reference Schupp, Jordano and Gómez2010). This study demonstrated some important differences in the quality of the treatment that P. flexuosa seeds receive when they pass through the digestive tract of native animals. The mean retention time of seeds does not seem to be the most important factor in total seed recovery and germination because the species with the longest retention time (C. chilensis) was not the one causing the greatest loss of seeds. In general, after a mean retention time of 5.5 d (SE = 2.72, range 1–16), it was possible to recover nearly half the viable seeds ingested, except for seeds consumed by L. guanicoe. Of the recovered seeds, more than 50% germinated during a month after recovery, except seeds ingested by R. americana and L. gymnocercus, which germinated in low proportions similarly as seeds collected from trees. For all animal species, germination was lower for seeds excreted during P2.
Some limitations are recognized due to working with animals in captivity because of alterations in the animals’ activity patterns imposed by confined spaces, and shifts in their diets relative to those of animals in the wild (Picard et al., Reference Picard, Papaix, Gosselin, Picot, Bideau and Baltzinger2015). We observed that retrieved seeds were sometimes consumed again by individuals, a behaviour often seen in captive animals such as R. americana (Renison et al., Reference Renison, Valladares and Martella2010), and common in some cecum fermenters such as D. patagonum (Hume and Warner, Reference Hume, Warner, Ruckebusch and Thivend1980). Beyond its limitations, this type of study allows us to estimate the effect of different digestive tract treatments on a known number of seeds consumed in a controlled moment.
Since the digestive systems differ greatly by taxon, both morphologically and physiologically, we can expect differences in gut retention times with effects on recovery and germination of ingested seeds (Traveset et al., Reference Traveset, Robertson, Rodríguez-Pérez, Dennis, Green, Schupp and Westcott2007). A rather long retention time with a prolonged exposure to digestive fluids may result in the removal of protective coats with an increase in germination speed or damage of the embryo (Gardener et al., Reference Gardener, McIvor and Jansen1993; Traveset and Verdú, Reference Traveset, Verdú, Levey, Silva and Galetti2002). Because of that, we expected to find that seeds ingested by L. guanicoe and C. chilensis would have longer mean retention time, lower seed recovery and lower germination of recovered seeds due to embryo damage by mastication and long exposure to digestive fluids.
Our prediction was partially justified. Results showed that C. chilensis had the longest mean retention time of ingested seeds, and the rodent D. patagonum had the shortest, whereas the other species had intermediate values. For C. chilensis, the mean retention time for ingested seeds of P. flexuosa was 16 d, similar to the 5–15 d reported in experiments with P. nigra seeds (Varela and Bucher, Reference Varela and Bucher2002). For P. nigra, seed recovery was almost 90% (Varela and Bucher, Reference Varela and Bucher2002), whereas recovery was 56% for P. flexuosa seeds. Despite the long retention time of seeds in the digestive tract of tortoises, their keratinized beak, instead of teeth, allows swallowing whole fruits, with seeds appearing in faeces with little physical change (Varela and Bucher, Reference Varela and Bucher2002).
Although the mean retention time of P. flexuosa seeds ingested by L. guanicoe was not longer than in the other studied species (4 d), and falls within the range of mean retention time recorded for markers ingested by L. lama (2–10 d depending on particle size; Clemens and Stevens, Reference Clemens and Stevens1980), the lowest percentages of viable seed were recovered after L. guanicoe's digestion (12%). A previous study that analysed faeces of L. guanicoe, collected in the field, documented large percentages of damaged seeds of P. torquata, P. chilensis and P. flexuosa (20–45%; Campos et al., Reference Campos, Peco, Campos, Malo, Giannoni and Suárez2008). Despite the low seed recovery, the present study showed that 62% of the seeds were able to germinate, similar to the proportion of germinable seeds of P. chilensis (70%) and P. flexuosa (50%) found in faeces collected from the field (Campos et al., Reference Campos, Peco, Campos, Malo, Giannoni and Suárez2008). From captivity experiments and field studies, we can observe a low recovery of viable P. flexuosa seeds with approximately half of them germinating after passage through the digestive tract of L. guanicoe. This effect on seed survival can be explained by the digestive process of this species, which includes rumination and the action of a forestomach with three gastric compartments for microbial fermentation, all resulting in high digestibility of plant material (Esteban and Thompson, Reference Esteban and Thompson1988).
Despite the marked difference in retention times between C. chilensis and D. patagonum, both the total percentage of seeds recovered (56 and 54%, respectively) and the germination of seeds (52 and 55%) were similar. Then, these similarities observed between both dispersers suggest that their difference in retention time is not the most important factor affecting seed recovery and germination. For many species, the time that seeds remain in the digestive tract of dispersers does not have an effect on germination, either positive or negative. Other factors, such as the type of food ingested along with fruits (with variable acidity, water content, etc.), may even be more relevant in determining the extent to which seeds are mechanically or chemically abraded (Traveset, Reference Traveset1998). Dolichotis patagonum is a small hindgut fermenter or a cecum fermenter, with a cecum as the primary site of microbial activity (Hume and Warner, Reference Hume, Warner, Ruckebusch and Thivend1980). In captivity, mean retention time of fluid and particles in the gastrointestinal tract of D. patagonum was 27 h (Sakaguchi et al., Reference Sakaguchi, Nippashi and Endoh1992), with high digestive efficiency (Kufner and Durañona, Reference Kufner and Durañona1991). From faeces of D. patagonum collected in the wild, 65% of P. flexuosa seeds were viable and between 12 and 25% were able to germinate (Campos and Ojeda, Reference Campos and Ojeda1997; Campos et al., Reference Campos, Peco, Campos, Malo, Giannoni and Suárez2008), although, from these studies, we do not know how long those seeds remained in the digestive tract or how many seeds were originally ingested.
For the remaining two species, R. americana and L. gymnocercus, our results showed intermediate values of retention time, a higher seed recovery after ingestion than for L. guanicoe, and the lowest percentages of germination among all studied species. Rhea americana is a large herbivorous bird that eats fruits and seeds of a wide range of sizes (Pratolongo et al., Reference Pratolongo, Quintana, Malvárez and Cagnoni2003; Renison et al., Reference Renison, Valladares and Martella2010). Rhea species are characterized by a short colon with particularly enormous paired caeca as the main site for microbial fermentation (Stevens and Hume, Reference Stevens and Hume1995). They have efficient digestibility, with a mean digesta retention time for small particles (8 mm) of 7–19 h (Frei et al., Reference Frei, Ortmann, Reutlinger, Kreuzer, Hatt and Clauss2015), although our results showed a mean retention time of 3 d for P. flexuosa seeds (5 mm). We recovered 51% of the ingested seeds, of which 21% were able to germinate. A previous study performed with captive individuals showed that less than 10% of P. nigra and P. alba seeds were recovered within 36 h following ingestion, and germinability of the recovered seeds was 5% for P. alba and 20% for P. nigra (Renison et al., Reference Renison, Valladares and Martella2010). The effect of digestion on seed germination was similar to that found in our study, although the low recovery percentage reported by Renison et al. (Reference Renison, Valladares and Martella2010) might suggest that in this work the observation time was not long enough. It has been proposed that an excessive mechanical grinding in a bird's gizzard may be detrimental for seeds (Traveset et al., Reference Traveset, Robertson, Rodríguez-Pérez, Dennis, Green, Schupp and Westcott2007). Even though R. americana did not appear to scarify seeds and increase germination percentage as others species do. We found no strong effect on viability because seed retrieval after passage through its gut was similar to that obtained for the other species, except for L. guanicoe, a species quite similar to R. americana in body size, but different in being a foregut fermenter.
The quality of treatment of L. gymnocercus, a monogastric (or ‘single-stomached’) animal was similar. As most carnivores, it has a relatively short and simple gastrointestinal tract; the glandular stomach is a unilateral dilatation of the digestive tract and a distinct hindgut is absent (Stevens and Hume, Reference Stevens and Hume1998). The mean retention time of P. flexuosa seeds was almost 3 d, longer than the mean retention time found in previous studies for Schinus molle fruits ingested by L. culpaeus (19 h; Silva et al., Reference Silva, Jacksic and Bozinovic2005), and for Vachellia (Acacia) aroma seeds ingested by L. gymnocercus and Cerdocyon thous (14 h; Varela and Bucher, Reference Varela and Bucher2006). It is important to note that alteration of an animal's activity patterns imposed by space limitation is a good reason to be careful about gut retention times obtained in captive carnivores (González-Varo et al., Reference González-Varo, López-Bao and Guitián2013). In one of the studies, authors had to discard 20 individuals because they exhibited stress symptoms, which can affect gastrointestinal motility by accelerating seed evacuation (Varela and Bucher, Reference Varela and Bucher2006). Several studies reported seed dispersal of different plants by species of Lycalopex (e.g. Pseudalopex and Dusicyon), mainly comparing the viability and germination of seeds obtained from faeces collected in the field and seeds from trees (Bustamante et al., Reference Bustamante, Simonetti and Mella1992; Castro et al., Reference Castro, Silva, Meserve, Gutiérrez, Contreras and Jaksic1994; León-Lobos and Kalin-Arroyo, Reference León-Lobos and Kalin-Arroyo1994; Campos and Ojeda, Reference Campos and Ojeda1997; Varela and Bucher, Reference Varela and Bucher2006; Dellafiore, Reference Dellafiore2018; Maldonado et al., Reference Maldonado, Loayza, Garcia and Pacheco2018). Only two studies carried out experiments in captivity, both of them with L. gymnocercus. Dellafiore (Reference Dellafiore2018) reported a 100% seed recovery and a 42% germination of Pyracantha atalantoides seeds after 24 h of retention in the gut. Varela and Bucher (Reference Varela and Bucher2006), after 36 h, recovered 52% of viable seeds of Acacia aroma ingested by animals and achieved 36% germination. After a 3-d mean retention time for P. flexuosa seeds, our results showed a 43% recovery of the viable ingested seeds, of which only 8% were able to germinate, similar to the results for germination of P. flexuosa seeds obtained from faeces collected in the field (Campos and Ojeda, Reference Campos and Ojeda1997). Other studies have also reported a low effect of digestion on germinability of seeds of Lithrea caustica and P. alba, and it has been proposed that retention time in the fox's digestive tract is too short to break the physical dormancy of seeds (León-Lobos and Kalin-Arroyo, Reference León-Lobos and Kalin-Arroyo1994; Maldonado et al., Reference Maldonado, Loayza, Garcia and Pacheco2018).
The variety of seed treatments offered by endozoochorous dispersers to P. flexuosa represents part of the qualitative component of SDE. Seeds contained in an indehiscent fruit need a mechanism to be released. Decomposition of the pericarp in dryland soil can take some years owing to fruit hardness and slowness of decomposition processes. During that time, seeds are exposed to predation by arthropods and small-sized rodents (Ortega-Baes et al., Reference Ortega-Baes, de Viana and Saravia2001; Giannoni et al., Reference Giannoni, Campos, Andino, Ramos-Castilla, Orofino, Borghi, de los Ríos and Campos2013; Velez et al., Reference Velez, Chacoff and Campos2018). In this scenario, fruit ingestion by endozoochorous species represents for seeds a way by which they are released from fruit avoiding predators under the parent tree, with some costs to seed recovery after the passage through the animals’ digestive systems. In the studied assemblage, the highest cost was for seeds ingested by L. guanicoe.
Seed germination does not seem to be significantly favoured when passing through digestive tracts, compared to the germination obtained in previous laboratory studies using healthy seeds that have been scarified by mechanical or chemical means, which is optimal for breaking dormancy and maximizing germination under laboratory conditions (e.g. Catalán and Macchiavelli, Reference Catalán and Macchiavelli1991; Catalán and Balzarini, Reference Catalán and Balzarini1992; Cony and Trione, Reference Cony and Trione1996; Ortega-Baes et al., Reference Ortega-Baes, de Viana and Suhring2002). Nevertheless, in the assemblage of seed dispersers, the passage by the digestive system of some species partially released seeds from the physical dormancy imposed by hard impermeable seed coats and increased germination percentages (seeds ingested by L. guanicoe, D. patagonum and C. chilnesis), and other dispersers (R. americana and L. gymnocercus) kept germination similar to that of seeds from trees receiving no scarification treatment (i.e. seeds from trees not released from endocarps). In the last case, seeds must wait for a subsequent environmental scarification by soil abrasion or extreme temperature cycles. It has been proposed that the passage through the digestive tract of animals produces asynchrony because endozoochory changes the timing of germination. In habitats characterized by environmental unpredictability, germination asynchrony might result in an overall fitness benefit for the plant that allows waiting for good conditions for seedling survival (Fedriani and Delibes, Reference Fedriani and Delibes2009). Thus, the benefits of germinating faster depend on the context.
In a global study, it has been proposed that mammals have a positive effect as enhancers of seed germination, although some groups, such as rodents, cause low seed germination and high mortality after ingestion. Particularly for Fabaceae seeds, mammals do not have a significant effect on germination (except in the case of Primates; Torres et al., Reference Torres, Castaño and Carranza-Quiceno2020). In contrast to this general study, we found that the quality of seed treatment by D. patagonum, the rodent of our study, was similar in its high seed recovery to L. gymnocercus, and similar in its high seed germination to L. guanicoe. Furthermore, we consider that mammals effectively contribute to dispersal of P. flexuosa seeds.
Because field studies of P. flexuosa report high quantities of viable seeds in faeces (e.g. Campos and Ojeda, Reference Campos and Ojeda1997; Campos et al., Reference Campos, Peco, Campos, Malo, Giannoni and Suárez2008), the numbers obtained in captivity experiments imply that a large number of seeds must have been consumed to compensate for the recovery of approximately half the seeds, with half of them having germination success after gut passage. In order to complete the SDE model for P. flexuosa, future research would be needed to compare the faeces of different dispersers as germination substrate, and to determine the conditions for seed germination and seedling establishment in the microhabitats where seeds are dispersed. In addition, data from animal movement patterns (home range, habitat use) would provide information about the potential seed dispersal distance by animal species.
Supplementary material
To view supplementary material for this article, please visit: https://doi.org/10.1017/S096025852000032X.
Acknowledgements
We thank the staff of the Ecological Reserve and Wildlife Rehabilitation Centre (Mendoza), the Mendoza Zoo and the Centre for Rehabilitation of Wild Fauna, Environmental Education and Recreation (San Juan) for their willingness and help with captivity experiments. We also thank Juan Villarroel, Nicolás Carlos and Sofía Campagna for their help during the captivity experiments and the laboratory work. This study was supported by the Consejo Nacional de Investigaciones Científicas y Técnicas (CONICET, PUE IADIZA), the Agencia Nacional de Promoción Científica y Tecnológica (PICT 2017-2154) and the Presidencia de la Nación Argentina and Gobierno de San Juan (Project Native Forest). N. Horak assisted us in drafting the English version.
Author contributions
C.M.C., M.I.C. and F.M.C. conceived the ideas and designed methodology; L.R., N.M., M.I.C., C.S., C.M.C., C. de los R. and F.M.C. collected and analysed the data; C.M.C. and F.M.C. acquired funding for the project; C.M.C. led the writing of the manuscript. All authors contributed critically to the drafts and gave final approval for publication.
