Keywords

22.1 Introduction

Habitat fragmentation and its consequences are currently the main threat to primate conservation (Cowlishaw and Dunbar 2000), and to biodiversity worldwide (Laurance and Peres 2006). This phenomenon, caused primarily by human population growth, leads to archipelagos of habitats, which in turn implies reduction and isolation of wildlife populations (Malcolm 1997; Robinson 1996). The increased perimeter to area ratio of fragments also causes easier human exploitation of forests (e.g. logging, hunting), and a decrease of suitable habitat for species unable to persist in edge environments (Laurance et al. 2002). Populations living in fragments are exposed to genetic flow reduction that causes genetic erosion and makes them more vulnerable (Lande and Barrowclough 1987; Spielman et al. 2004).

In order to better estimate ecological viability and to develop effective management actions, it is essential to identify habitat features that determine species persistence in fragmented habitats. Island biogeography factors such as fragment size, age, and isolation have been often identified as strong predictors of species persistence (Whittaker and Fernandez-Palacios 2007). However, flexible species able to reduce or expand home ranges, move through the matrix, and/or adjust their diet are likely to persist in areas no longer identified as suitable by models (Estrada and Coates-Estrada 1996; Lovejoy et al. 1986; Onderdonk and Chapman 2000). In primates, for example, fragment size and isolation have not been always found to be good predictors of species presence/absence . In fact, while in Madagascar these factors seem to govern primate persistence (Ganzhorn et al. 2000), the same is not true for other areas (Onderdonk and Chapman 2000). Primates density as well, which is strongly affected by habitat degradation (Bowers and Matter 1997), does not always relate to obvious fragment features (Peres 1997).

The howler monkeys, genus Alouatta, which includes 14 species (International Union for Conservation of Nature (IUCN) 2012), are generally considered very flexible (Lovejoy et al. 1986; Bicca-Marques 2003; Garber et al. 2006). These monkeys often occur in fragments where other primates (e.g. Ateles sp.) cannot persist (Estrada and Coates-Estrada 1996; Gilbert 2003; Arroyo-Rodrìguez and Dias 2010). In particular, howler monkeys’ persistence in fragmented habitats seems to be based on their ability to adapt diet to food availability (Onderdonk and Chapman 2000; Bicca-Marques 2003; Rivera and Calmè 2006), increase leaf intake (Rodrìguez-Luna et al. 2003; Asensio et al. 2007), consume secondary or exotic plant species (Onderdonk and Chapman 2000), and also epiphytic or parasitic plants (Rodrìguez-Luna et al. 2003; Asensio et al. 2007). However, they are also very selective with regard to food resources and the monthly diversity of their diet is generally low, with only few plant species regularly included (Estrada et al. 1999). Food selection is based on nutritional contents, vegetative stage, and secondary compounds (Glander 1982), and it is vital in fragments where preferred food resources are scarce or scattered (Chiarello 2003; Juan et al. 2000). Howler monkeys are also able to rearrange their ranging pattern (Estrada and Coates-Estrada 1996; Onderdonk and Chapman 2000), and their time budget (Juan et al. 2000; Silver and Marsh 2003; Cristòbal-Azkarate and Arroyo-Rodrìguez 2007). Finally, in small or degraded fragments they can even use a fission–fusion feeding strategy to decrease intra-group feeding competition (Leighton and Leighton 1982; Asensio et al. 2007).

The ecology of Alouatta has been comprehensively studied, although the factors predicting their persistence in different forest fragments are not clear yet (Arroyo-Rodrìguez and Dias 2010). Individual density may increase at the initial stages of the fragmentation process due to predator disappearance (Estrada et al. 2002; Rosales-Meda et al. 2007), but this is also dependent on the quality and availability of food resources (Nagy and Milton 1979). In some areas, highly productive secondary vegetation seems to favour high-density populations, although this may cause resource depletion in the long term, leading to possible local extinctions in small fragments (Lovejoy et al. 1986; Arroyo-Rodrìguez and Dias 2010).

This study aims at investigating the consequences of high population density on the activity budget, ranging pattern, and nutritional ecology of the mantled howler monkey (Alouatta palliata) population inhabiting two forest fragments at La Suerte Biological Field Station, north-eastern Costa Rica. The two forest fragments are connected by a narrow corridor but differ in terms of size and forest history. A. palliata individual density in the smaller of those fragments is very high compared to the other area (150 vs. 57 ind/km2; Adams et al., unpublished) and to the species average (Campbell et al. 2011). Longitudinal demographic data do not show a population decline over the last decade (Pruetz and Leasor 2002; Adams et al., unpublished). We collected eco-ethological data and food samples on two groups of howler monkeys in each forest fragment over three summers (July–August) from 2009 to 2011. We hypothesize that if the small fragment is overcrowded, this would cause high competition and resource depletion, altering the time budget, and decreasing the quality of the diet of howler monkeys compared to animals living in the non-crowded forest. To test this hypothesis, we analysed behavioural, nutritional, and ranging data.

22.2 Methods

22.2.1 Study Site

The study site was located in north-eastern Costa Rica, at La Suerte Biological Field Station (10° 26′ 15″ N; 86° 43′ 30″ W; Fig. 22.1). The site is 50 m above mean sea level and is covered by a lowland rainforest, with annual rainfall around 3900 mm and mean temperature of 24 °C (Bezanson et al. 2008). It includes two forest fragments, which are embedded in a matrix of abandoned fields, grazed pastures, and coconut plantations. La Suerte River connects both fragments, functioning as a corridor of riparian forest. The smaller forest (SF) fragment is about 20 ha and is one of the remaining stands of primary vegetation in the area. The larger forest (LF) fragment, to the north, covers 250 ha of secondary forest that was exploited for logging and grazing until the late seventies (Renee Molina, personal communication). Three primate species inhabit the area: mantled howler monkey (A. palliata), white-headed capuchin (Cebus capucinus), and Geoffroy’s spider monkey (Ateles geoffroyi). According to Adams et al. (unpublished), the two fragments are structurally similar but the SF has greater tree species diversity and larger crown volume of emergent trees. Howler monkeys’ density is considerably higher in the SF, 150 ind/km2 in 1999 and 2010, than in the LF, 30 ind/km2 in 1999 and 57 ind/km2 in 2010 (Adams et al., unpublished; Pruetz and Leasor, 2002).

Fig. 22.1
figure 1

Study site

Full size image

22.2.2 Study Species

A. palliata is the second largest neotropical primate (Zingeser 1973), and its distribution encompasses all the Central American forests (Estrada and Coates-Estrada 1994). This monkey is diurnal, and largely arboreal, exploiting the higher canopy layers (Silva and Da1981), where it feeds mainly on leaves and fruits (less on flowers, buds, petioles, etc.; Milton 1980). Despite the high-fibre diet, it does not possess anatomical or physiological specializations to maximize cellulose digestion (Chivers and Hladik 1980). On average, A. palliata groups are composed of 3–4 males, 7–10 females, and the number of offspring and subadults equals to females number (Carpenter 1934). Its activity pattern is characterised by long resting period interrupted essentially by feeding/foraging bouts (Smith 1977).

22.2.3 Behavioural, Nutritional, and Ranging Data

Behavioural data were collected via a 5-min focal instantaneous sampling (Altmann 1974) during July and August 2009, 2010, and 2011. Two groups were studied in each forest fragment, and each group was observed between 3 and 4 consecutive days per season from dawn to dusk. A total of 480 observation hours were collected. The ethogram included moving (moving on the same tree and travelling), feeding (all the actions related to food processing), resting, and other (social behaviours, self-grooming, scratching, stretching, drinking, defecating, and urinating). Only adult individuals were observed and used to estimate daily activity budget. Data collection also included posture, food item, feeding (or resting) tree’s code and global positioning system (GPS) coordinates. Feeding trees were marked and identified to species level on a following day with the aid of a local botanist. Samples of food items (leaves or fruits) eaten during behavioural observations were also collected from the same tree where the monkeys were observed eating. GPS coordinates were recorded every 30 min and analysed via RANGES 8 to estimate the home range size (via Kernel analysis; Worton 1989) and the daily path length.

22.2.4 Biochemical Analyses

Food samples were dried in the sun at La Suerte, and later analysed for their nutritional content at the laboratory of the Department of Veterinary Science, University of Pisa, and at the Department of Animal Ecology and Conservation, University of Hamburg following the methods outlined in Donati et al. (2011). Dry matter was used to determine the amount of the following nutritional compounds: crude protein (CP), crude fat (CF), neutral detergent fibre (NDF), acid detergent fibre (ADF), ashes (Donati et al. 2007). Nonstructural carbohydrates (NSCs) were assessed with the formula: NSC = 100-(%CP + %CF + %NDF + %ashes)(Milton 2008; Norconk et al. 2009). Metabolizable energy (ME) was estimated according to Conklin and Wrangham (1994) as follows: ME (kcal/100 g d.m.) = (4× CP×0.89) + (4×NDF×0.41) + (4×NSC) + (9×CF). Dried food samples collected in 2009 were also analysed with atomic absorption spectroscopy in order to get the concentration of mineral elements, which included macroelements (calcium, phosphorus, magnesium, sodium) and microelements (iron, copper, manganese, zinc; Williams-Guillèn 2003).

22.2.5 Data Analyses

Daily percentages of feeding, moving, and resting as well as daily path lengths of each group of monkeys were averaged per fragment and compared using the Mann–Whitney nonparametric U test. The diet diversity of each group was evaluated with the Shannon’s diversity index and averaged per fragment. The comparisons of the nutritional contents of the food items eaten in each fragment were also undertaken via the U test. For the comparisons, we showed both the difference between the food samples per se and the difference between the weighted diets using the proportion of feeding time as coefficient.

22.3 Results

22.3.1 Activity budget

At the La Suerte Biological Station, resting was the most represented activity in howler monkeys, consisting in 65.67 % ± 1.49 (n = 33 days) of the daily activity budget, while feeding and moving were the other main activities, 12.27 ± 1.71 and 18.52 ± 1.80, respectively. There was no significant difference in the monkey time budget between the two forest fragments (Table 22.1).

Table 22.1 Activity budget: Comparison of the main activities between small forest (SF) groups and large forest (LF) groups
Full size table

22.3.2 Ranging Pattern

The mean home range of SF groups was considerably smaller than the areas used by the monkeys in the LF (Fig. 22.2a). As for daily path length (Fig. 22.2b), SF groups moved significantly less than LF groups (U = 30.0, p = 0.02, n = 11 days SF, n = 13 days LF).

Fig. 22.2
figure 2

Ranging patterns: Comparison among mean home ranges (a) and mean daily path length (b) of the howler monkey groups in the two forest fragments

Full size image

22.3.3 Nutritional quality of the diet

We collected a total of 28 different food samples during behavioural observations; 12 from the SF and 16 from the LF. The Shannon diversity index (H’) for the diet of the two groups did not differ appreciably (SF groups: 0.90; LF groups: 0.87).

Despite food samples collected in the SF contained substantially lower values of fibre fractions (Table 22.2), statistical analyses did not show any significant difference. The weighted diet (Table 22.3) did not show any significant difference as well between the dietary regime of the monkeys in the two forest fragments.

Table 22.2 Nutritional comparison of food samples from plant species eaten by Alouatta palliata collected in the small (SF) and large forest (LF) fragment
Full size table
Table 22.3 Comparison of the weighted diet between small forest (SF) groups and large forest (LF) groups
Full size table

The comparison between the two fragments in terms of mineral content of food items showed that in SF samples calcium values were almost double than those from the LF, as supported by a strong statistical tendency, while phosphorus and magnesium showed similar values in both fragments (Table 22.4). Overall food samples of howler monkeys contained very low amounts of sodium. Microelements’ contents were similar between the samples of the two fragments, except for manganese, which was significantly higher in the LF samples (Table 22.4).

Table 22.4 Mineral composition: comparison between the food samples from plant species eaten by Alouatta palliata collected in the small (SF) and large forest (LF) fragment
Full size table

22.4 Discussion

Overall, the howler monkeys’ groups living in the two forest fragments of La Suerte did not differ in terms of activity and diet, while their ranging pattern was found to be substantially dissimilar. The activity budget of the groups in La Suerte is not only similar between the two fragments but also to the values reported in more extensive behavioural studies on this species (Carpenter 1934; Milton 1980; Williams-Guillèn 2003). Inactivity was prevailing, with no differences among the groups. Regarding the other two main activities, feeding and moving, there were no significant differences as well. Milton (1998) suggested that high values of inactivity in Alouatta are the consequence of large quantity of leaves passing through the intestinal tract and slowing down digestion due to cellulosic material. Other studies do not support this hypothesis and indicate that inactivity is rather a phylogenetic trait of the genus, independent from the diet (Pavelka and Knopff 2004). In La Suerte, inactivity was not correlated to obvious forest/fragment features or monkey density, in line with previous studies (Cristobal-Azkarate and Arroyo-Rodrìguez 2007).

In contrast to time budget, ranging patterns showed large differences between the howler monkey groups. In particular, the monkeys in the LF had larger home ranges and travelled further. This is in agreement with previous reports showing that ranging size is affected by fragment size (Cristobal-Azkarate and Arroyo-Rodrìguez 2007). It is possible that in the high-density SF howler monkey groups moved less to avoid interactions with conspecific groups. In contrast, in the LF the howler groups did not simply move further but their size was smaller than in the SF, as a consequence of a more obvious fission–fusion strategy (Occhibove 2010). This behaviour is typically used to reduce intra-group competition when resources are limited (Arroyo-Rodrìguez and Dias 2010). Thus, the behavioural evidence seems to indicate, contrary to our expectation, that in the LF preferred food resources may have been more limited. The structure of the two forest fragments may help to explain this finding. In fact, vegetation plots indicate that the LF had low floristic diversity and smaller average food tree size compared to the SF, which may make the former area less suitable for the howler monkeys and may limit its carrying capacity (Adams et al., unpublished).

According to Chapman et al. (2002), folivorous primate biomass is strongly related to the protein-to-fibre ratio of the leaves used as food. Although the biological relevance of these measures is a matter of debate (DeGabriel et al. 2013; Wallis et al. 2012), the index has been used to evaluate food quality of herbivorous, since nitrogen is considered a limiting nutrient in many terrestrial ecosystems (White 1993). In fact, monkeys often prefer leaves with higher protein and low fibre ratios (Chapman et al. 2002; Ganzhorn 2002; Hanya and Bernard 2012; Milton 1979) also because fibres require microbial fermentation and contain components (e.g. cellulose and hemicellulose) that are only partially digestible (McNab 2002). In La Suerte, nutritional data indicated that NDF and ADF were higher in food samples from LF groups, with a clear statistical trend. Nevertheless, CP/NDF and CP/ADF ratios were not significantly different between the food samples from the two fragments. The monkey-weighted diet echoed the results from food samples with even slighter nutritional differences, suggesting that animal feeding has further compensated for the variation in nutrient availability between the two fragments. Thus, overall the monkeys living in the two fragments did not show major difference in terms of macronutrient intake. It is important to mention, however, that analyses based on feeding time can yield biased results and they do not always correlate with actual intake, due to the differences in food size, density, and processing time (Rothman et al. 2011). However, feeding time has been shown to provide acceptable proxies of actual food intake in case of leaves’ consumption (Kurland and Gulin 1987).

We used the National Research Council (NRC) (2003) tables for nonhuman primates to examine the nutritional composition of the food samples in relation to the monkeys’ nutritional requirements. NDF and ADF percentages of all the samples were higher than the minimum suggested to keep gastrointestinal health (10–30 % and 5–15 %, respectively). The proportion of protein, in all samples, was in line with the lower limit of the range suggested by the NRC (2003). Milton (1998) suggested that leaves eaten by howlers have to contain at least 11 % of CP (on dry matter) to cover their needs. Our food samples were all above this value.

Although primates prioritize proteins as source of energy (Felton et al. 2009b; Rothman et al. 2011), overall they do not require high proteinaceous diets (Oftedal 1992). Mature leaves, basically the only food category consumed in both fragments in La Suerte during our study period, were found to be the main source of energy and to contain more sugars than young leaves (Occhibove 2010). Mature leaves, however, also contained higher fibre and lower protein concentrations than young leaves, thus making in principle this food less palatable (Behie and Pavelka 2012). Behie and Pavelka (2012) propose that food selection in folivorous primates may be driven by energy intake rather than protein intake maximization, since the protein amount is above the minimum requirements in both mature and young leaves. This idea is supported by similar results in a gorilla population at Bwindi National Park, Uganda (Ganas et al. 2009). Our samples of leaves from La Suerte were, in fact, well above the minimum protein requirements and contained comparable sugars in both fragments. This seems to suggest that in the high-density area, howler monkeys do not feed on fallback food resources, despite the predicted potential high inter and intra-group competition.

The food mineral content was very similar between the two forest fragments, except for calcium and manganese, more represented in the SF and LF samples, respectively. It is well documented that minerals are fundamental to the health (Robbins 1993) and mineral availability has been proposed as a limiting factor to population growth of frugivorous primates (Rode et al. 2006). However, little is known regarding the role of minerals in primate diet selection (Felton et al. 2009a). In fact, few authors support the idea that minerals influence primate diet selection, partly because minerals can be obtained from nonfood sources (e.g. salt licks, soil, eggshells). Moreover, some studies indicate that mineral intake of wild primates eating natural diets exceeds established requirements of humans and deficiencies seem unlikely (Milton 2003; Rothman et al. 2006). Our preliminary results from La Suerte indicated that dietary calcium may show large differences even between contiguous areas. Previous work suggested that some mature leaves were consumed more than young leaves as important sources of minerals , especially of calcium, iron, and manganese (Silver et al. 2000, 2003). Felton et al. (2009b) also found that spider monkeys (Ateles chamek) ate large amounts of figs probably because they contained high concentrations of minerals, especially available calcium (O’Brien et al. 1998; Silver et al. 2000), which is crucial for maintenance and reproduction (Robbins 1993). Data on a larger sample of food collected over a longer period of time are necessary to properly test the potential role of mineral availability in La Suerte.

Looking at the NRC tables (2003) for mineral requirements in primates, the leaves from the SF showed an amount of calcium far higher than the recommended value, while in the LF samples the values were comparable. Similarly to Williams-Guillèn (2003), phosphorus concentration was found to be half of the value indicated in the NRC (2003). In both fragments, magnesium and iron were higher than the recommended value of the NRC (2003), suggesting that howlers’ requirements were met for these elements. According to previous studies on the nutritional ecology of wild A. palliata (Williams-Guillèn 2003) and to the nutritional content of the most common food of other herbivorous (e.g. forage), sodium concentration was very low, that is four to five times lower than the suggested values by the NRC (2003). This seems to be usual for folivorous primates and it is not clear what the implications of it are. In contrast, manganese concentration in both fragments was much higher than the NRC (2003) recommended value, especially in the LF where it exceeded by around 20 times the proposed concentration. This amount may cause neurologic toxicity (Burton and Guilarte 2009). Also Williams-Guillèn (2003) in Nicaraguan A. palliata found higher amounts of manganese than the NRC (2003), but we found much higher values in La Suerte. In fact, hyper-abundance of some elements rather than deficiency may also play a role in determining the distribution of these monkeys and more research needs to be focussed on mineral elements.

In conclusion, in both forest fragments of the La Suerte Biological Station, howler monkeys seem able to reach their nutritional requirements, adjusting their feeding strategy and ranging pattern to the size and resource availability in each area. No sign of major diet quality drop due to monkey high density and possible overcrowding were identified in this study. This demonstrates the excellent ability of A. palliata to cope with different environments. Nevertheless, the long-term resource depletion due to overcrowding in the SF may still represent a risk (Estrada et al. 1999; Pruetz and Leasor 2002). Therefore, it is necessary to improve the connection between the fragments in order to minimize this risk and maximize gene flow between the two small populations. It is also important to note, however, that our data covers only 2 months per year, and this may lead to biased conclusions, since food availability varies seasonally and supra-annually (Cant 1980). More detailed investigations over a full year are needed to get a clear picture of the demographic and ecological constraints for A. palliata in La Suerte.