Keywords

4.1 Introduction

Understanding a species’ food requirements and resource acquisition strategies is fundamental to understanding its ecology. Based on their preference for the two major plant types, monocots (grasses and sedges) and dicots (forbs, shrubs, and trees), large herbivores have historically been classified as either (a) grazers (i.e., bulk and roughage feeders), (b) browsers (selectors of concentrated dicotyledonous herbage), or (c) intermediate or mixed feeders that consume either grass or browse at different seasons or in different areas (Hofmann and Stewart 1972; Hofmann 1973, 1989; Duncan and Poppi 2008). Grass and browse represent very different food resources and their consumption poses different constraints for herbivores (Duncan and Poppi 2008). Browse species, when compared to grasses, have higher levels of soluble cell content and nitrogen that are beneficial for large herbivores, but at the same time have higher levels of lignin and secondary metabolites that are detrimental to large herbivores (Demment and Van Soest 1985; Gordon and Illius 1994; Gordon 2003; Duncan and Poppi 2008). These fundamental differences between grasses and browse have led to different adaptations in animals specializing on one or the other plant type, with implications for all aspects of their ecology and life history (Hofmann and Stewart 1972; Hofmann 1973, 1989; Gordon 2003; Duncan and Poppi 2008).

Based on differences found in the relationships between stomach structure and feeding habits of East African ruminants, Hofmann et al. proposed that a species’ digestive trait syndrome was the primary factor deciding forage selection and feeding habits of large herbivores (Hofmann and Stewart 1972; Hofmann 1973, 1989). However, subsequent analyses that statistically accounted for differences in body mass among species found limited evidence for morphological and anatomical differences between large herbivores belonging to different feeding categories (Gordon and Illius 1994; Gordon 2003). Thus, while additional studies have found evidence to support Hofmann’s proposition (Clauss et al. 2003, 2008, 2010; Hofmann et al. 2008), it is now widely accepted that body mass plays a critical role in the resource ecology of large herbivores (Prins and Olff 1998).

Hofmann and Stewart (1972) originally proposed their foraging categories based on differences between only ruminant species, which was not surprising considering that 92 % of the ~260 large herbivore species worldwide are ruminants. All ruminants belong to the order Artiodactyla (even-toed ungulates) and the remaining 8 % of large herbivores are hindgut fermenters that are confined to the mammalian orders Perissodactyla (odd-toed ungulates) and Proboscidea (elephants ). Ruminants are foregut fermenters , i.e., they derive the bulk of their energy from the forage they consume before passage through the gut, primarily by the action of symbiotic microbes on plant material during a fermentation process in the rumen , or a similar morphological structure as in hippos or camelids (Van Soest 1994). In hindgut fermentation, though energy is first derived when food passes through the stomach, additional energy is derived from fermentation of food in the hindgut, primarily the caecum .

The dominance of the large herbivore guild by ruminants is the result of a ruminant radiation that began ~10 Ma during the Miocene, which coincided with the expansion and diversification of grasslands worldwide (Perez-Barberia et al. 2001; Prins and Gordon 2008; Bouchenak-Khelladi et al. 2009). The first large herbivores date back to the early Cenozoic (~55 Ma), a time when grasslands were not common, which suggests that browsing was the more primitive of the two major large herbivore diets (Bodmer and Ward 2006). However, when grasslands began expanding across the globe around 10 Ma, especially in the tropics, the first specialized grazers —which were primarily ruminants—emerged (Cerling et al. 1993, 1997; Bouchenak-Khelladi et al. 2009). An increasing speciation of ruminants continued to match the expansion of grasslands worldwide, which in turn was being fuelled by the emergence of an alternate to the prevailing C3 photosynthetic pathway in plants—the C4 photosynthetic pathway, which enabled plants to function at lower CO2 levels (Janis 2008; Damuth and Janis 2011).

Although ruminants dominate the large herbivore guild and many ruminants are grazers, the majority of the smallest ruminants primarily subsist as browsers and/or frugivores . Ruminants, in contrast to hindgut fermenters, are morpho-physiologically limited in their daily intake due to their longer digestive retention times and passage rates (Beekman and Prins 1989; Clauss et al. 2003, 2010). The longer retention times help maximize the energy extracted by the microbial community in the gut from low quality forage, often grasses. These longer retention times, however, are better tolerated by larger rather than smaller bodied species. The innate capacity for larger bodied animals to tolerate forage of lower quality is a function of the mismatch between the nonlinear relationship of metabolic requirements (MR) and body mass (BM), MR = 70 BM0.75, and the near-linear relationship of gut capacity (GC) and body mass, log GC = 1.03 log BM−0.94 (Demment and Van Soest 1985). Therefore, compared to larger species, ruminants of smaller body mass are more dependent on higher quality forage, and to satisfy their nutritional requirements, smaller herbivores forage on browse and fruits that are of higher quality more than they forage on grasses (Bell 1970, 1971; Jarman 1974).

Large herbivores also have to contend with differences in the physical demands of feeding on grasses versus browse. Grasses in general are more fibrous, i.e., they have a greater proportion of cell wall (cellulose and hemicellulose) to cell content, and have higher levels of abrasive components like silica (Kaiser et al. 2009). Therefore when compared to browsers , grazers tend to have bigger masseter muscles, larger and deeper angles of the jaw, broader muzzles, and longer masseteric fossa on the skull (Gordon and Illius 1988; Clauss et al. 2008; Hofmann et al. 2008; Janis et al. 2010).

Building on what we know from studies of large herbivores in Europe , North America , and Africa, this chapter explores patterns in the feeding preferences of different large herbivore species in South and Southeast Asia (SSEA). We classified species as grazers if they consumed primarily graminoids year-round, browsers if they consumed primarily non-graminoids year-round, and mixed feeders if they either switched between grazing and browsing or consumed both browse and grasses. The wide range in body mass of species in the large herbivore assemblage in SSEA , which is similar to the body mass range observed in the large herbivore assemblage in Africa, allows us to analyze broad trends in foraging preferences as related to body mass and digestive syndromes .

One significant difference between the herbivore assemblages found in SSEA and Africa is that while there is only one Cervid (deer) species within the assemblage of 94 large herbivore species found in Africa (the Atlas deer Cervus elaphas barbarus found in the Atlas mountains in Africa’s extreme Northwest), ~30 % of large herbivore species that inhabit SSEA are Cervid species. It is, however, unclear how Bovids and Cervids differ in their digestive physiology and therefore, species of both groups were grouped together for our analysis (Langvatn and Hanley 1993; Hobbs et al. 1996; Mysterud 2000; Seagle 2003; Hewison et al. 2009; Beguin et al. 2011; Redjadj et al. 2014). Although pig species (Suidae) are sometimes included in analyses related to large herbivore forage selection, they are indeed omnivores and were therefore excluded from this study (Rerat 1978). Excluding the 13 suid species, we were left with 70 large herbivore species to analyze: six belonging to the order Perissodactyla , one belonging to the order Proboscidea (the Asian elephant ), and the remaining 63 species (~90 %) are ruminants and belong to the order Artiodactyla.

4.2 Forage Preference—Body Mass , Roughage , and Vegetation Quality

Given the paucity of data on forage preferences of many large herbivore species in SSEA, we pooled data and classified species in related groups, primarily by genus (Table 4.1). In general, forage selection by large herbivores is guided by the body mass of a species, its anatomical and morphological specializations, and the quantity, quality (mainly nitrogen and phosphorus), and the composition (level of roughage) of plant material (Bell 1971; Jarman 1974; Gordon and Illius 1996). In addition, the ratio of the abundance of grasses and browse at local scales influences the composition of a large herbivore assemblage found in a given area (Gordon 2003; Prins and Van Langevelde 2008; Owen-Smith 2014). In the savannas of Africa, for example, where browse is generally patchily distributed among vast grasslands, grazers dominate the large herbivore assemblage (Mcnaughton and Georgiadis 1986). In contrast, in tropical Asia where grasslands are patchily distributed within forested landscapes, mixed feeders should dominate the large herbivore assemblage. As expected, the majority of large herbivore species in SSEA are mixed feeders (Fig. 4.1). However, concentrate feeders (browsers ) dominate the smallest size categories, while bulk feeders , predominantly grazers, dominate the largest size categories (>500 kg, Fig. 4.1).

Table 4.1 Feeding preferences of groups of large herbivore species in Asia listed by order, family, and genus
Full size table
Fig. 4.1
figure 1

Proportions of 70 large herbivore species found in South and Southeast Asia that are grazers (often bulk foragers), mixed feeders (that utilize both grasses and browse species), and browsers (often concentrate feeders ) plotted in relation to body mass . The line plots the percentage of species that are ruminants in each body mass class. Numbers within bars denote the number of species of grazers, mixed feeders , or browsers in each body mass class. Although the 13 pig species found in the region are in the overall list of large herbivores for the region (see Chapter 1), they are indeed omnivores and were therefore excluded from this analysis

Full size image

Studies of large herbivore assemblages in other regions have reported that browsers are found across the entire body size range of herbivores (Sponheimer et al. 2003). This is true for the assemblage in SSEA as well—ruminant browsers dominate the lowest body mass classes (Fig. 4.1), while the browsers in the largest body mass class are nonruminants (see Sect. 4.3 for more details). In SSEA, the smallest species are almost exclusively concentrate feeders and primarily frugivores (Fig. 4.1, Table 4.1) and typically occupy forested habitats . For example, the small muntjac and mouse deer species are selective with everything they feed on, selecting leaves and not stems when grazing, and selecting leaves, buds, and flowers when browsing (Barrette 1977). Mouse deer , classified as browsers, are probably the most selective of large herbivore species in SSEA, and represent a distinct frugi-folivore category (Bodmer 1990; Kingdon et al. 2013). They are similar in body mass to the smallest ruminant species found in Africa, the dik-diks, which are known to be concentrate feeders (Manser and Brotherton 1995; Dehority and Odenyo 2003; Kingdon et al. 2013). The tendency for smaller ruminants to be browsers has been previously noted (Case 1979; Gordon and Illius 1994; Clauss et al. 2003), and the increased reliance on high-quality fruits in the smallest bodied species is potentially a consequence of their relatively greater metabolic demands (Bodmer 1990).

The mesoherbivore body mass class (50–500 kg) was comprised primarily of mixed feeders (Fig. 4.1), including both Bovid and Cervid species. Although Fig. 4.1 depicts a low percentage of browsers in the intermediate body mass range, this can be deceiving as many of the mixed feeders in the intermediate body mass categories browse extensively, just not exclusively, like for example, the ubiquitous sambar Rusa unicolor (Padmalal et al. 2003; Ahrestani et al. 2012). However, other Cervids such as the chital Axis axis appear to adopt a mixed-feeding strategy. For example, chital Axis axis in South India were found to graze during the wet season, but then switched to browse during the middle-late dry season (Ahrestani et al. 2012)—the bottleneck period when rainfall is practically nonexistent, graminoid growth is dormant, and grasses have their highest levels of roughage and lowest levels of nutrients (Scoones 1995; Drescher et al. 2006). At present, however, it is unclear which factors regulate the extent to which different mesoherbivore species in SSEA rely on browse relative to graze, and how this changes with body mass , season, and across environmental and vegetation gradients.

While concentrate feeding is confined to the smaller animals, bulk feeding, primarily of grasses, is confined to larger bovid species that are ruminants and equid species that are hindgut fermenters (Fig. 4.1). In general, grazing is more common among species with larger body mass , and in Africa’s large herbivore assemblage , grazing ruminants are larger than browsing ruminants (Bell 1971; Case 1979; Bodmer 1990; Van Wieren 1996; Perez-Barberia et al. 2001). It has been argued that the tendency for larger ruminants to be grazers rather than browsers is not a consequence of physiological or digestive limitations (Gordon and Illius 1994), but rather a result of limitations imposed by forage availability and abundance (Van Soest 1994; Clauss et al. 2003). Given their large metabolic requirements, it is easier for large-bodied species to meet their daily forage requirements by consuming grasses that are often more abundantly available than browse (Clauss et al. 2003). However, when browse is available in abundance, browsing ruminants can achieve large body sizes as in the case of moose Alces alces and giraffe Giraffa camelopardalis (Clauss et al. 2003).

The majority of large-bodied species in SSEA are wild Bovini species, i.e., species belonging to the genus Bos, which are similar to domestic cattle and are primarily grazers. Because of their low metabolic requirement to gut capacity ratio, these large Bovini species are able to tolerate high levels of roughage . However, wild Bovini species that inhabit forests, like the gaur and banteng , have been recorded browsing on multiple plant species (Schaller 1967). Gaur have also been observed eating the bark of trees during the dry season (Pasha et al. 2002), which could be related to the lack of nutrition in tropical deciduous forests during the dry season (Ahrestani et al. 2011). The many observations of Bovini species migrating to low-lying regions near rivers and drainage lines during the dry season seem to suggest that these wetter habitats —that typically retain a herbaceous layer in the dry season—act as key foraging habitats for species like the gaur in the dry season (Conry 1989; Ahrestani and Karanth 2014; Prins and Van Oeveren 2014).

4.3 Digestive Physiology Trait Syndromes: Hindgut versus Foregut

The distribution of foregut and hindgut fermenters across the body mass gradient of the large herbivore assemblage in SSEA (2–3070 kg) is illustrated in Fig. 4.1. There is a complete absence of hindgut fermenters in the lower body mass classes and a complete absence of ruminants in the megaherbivore class (>1000 kg). In general, the global megaherbivore guild is dominated by hindgut fermenters except for the giraffe Giraffa camelopardalis and hippopotami Hippopotamus sp. and it has been argued that the foregut digestive trait syndrome is not the most efficient for species with body mass >1000 kg (Demment and Van Soest 1985; Clauss et al. 2003).

Compared to ruminants, hindgut fermenters can tolerate forage of lower quality primarily because of lower retention times and faster passage rates in their digestive trait syndromes (Van Soest 1994). Their higher faster rates also allow hindgut fermenters to achieve larger body sizes than ruminants. All four megaherbivores in Asia (the Asian elephant and the Sumatran , Javan and Indian rhinoceros) are hindgut fermenters . The Asian elephant is a mixed feeder that consumes both browse and grass throughout the year, or follows a foraging regime that switches between dry season browsing and wet season grazing (Sukumar 1990, 1992). Despite being a mixed and bulk forager—a foraging strategy that is probably necessary to satisfy the metabolic requirements of over three tonnes of body mass —it is possible that the Asian elephant is selective in what it seasonally eats in some areas, similar to what the African elephant does in areas where nutrient levels vary significantly among forage species (Woolley et al. 2009; Owen-Smith and Chafota 2012; Pretorius et al. 2012).

Amongst the rhinocerotids, the Indian rhinoceros Rhinoceros unicornis (the Indian rhino ) “is a semi- hypsodont , and is able to graze”, while the closely related R. sondaicus (the Javan rhino ) and the more distantly related Dicerorhinus sumatrensis (the Sumatran rhino ) are brachydont browsers (Prothero et al. 1989). The two lineages (Rhinoceros and Dicerorhinus) have been present in SSEA since the Pliocene, and R. sondaicus appears to be a more recent species than R. unicornis (Prothero et al. 1989; De Iongh et al. 2005, also see Chap. 2). The contrasting feeding strategies of the Asian rhinoceros species bear similarities to their African counterparts. The two rhino species in Africa, though possessing similar hindgut digestive trait syndromes, have contrasting diets ; the white rhino Ceratotherium simum is more or less an exclusive grazer and the black rhino Diceros bicornis is more or less an exclusive browser (Owen-Smith 1992; Luske et al. 2009; Van Lieverloo et al. 2009). A similar scenario was also present in the North American Micocene ungulate assemblage (Prothero et al. 1989), which had one grazing rhino species sympatric with one browsing rhino species.

Only three hindgut fermenter species below 500 kg inhabit SSEA: two Equus spp. (Asiatic and the Tibetan wild ass ) that are both grazers , and the Asian tapir that is considered a mixed feeder. Both Equus species inhabit dry habitats characterised by sparse and patchily distributed forage of poor quality, conditions that are better handled by hindgut rather than foregut fermentation trait syndromes (Clauss et al. 2003). In contrast to the dry habitats of the Equus, the Asian tapir, inhabits mainly moist forests and often rainforests (Clements et al. 2012; Table 4.1). In general, except for its frugivorous habits (Campos-Arceiz et al. 2012), little else is known about the diet of Asian tapir. The other three tapir species in the world are all found in Central and South America , where they often inhabit forests (Fragoso 1997; Galetti et al. 2001). In general, forests do not support high levels of graminoid production and the morphology of the tapir’s snout suggests an evolutionary adaptation to frugivory and browsing (Janis 1984; Milewski and Dierenfeld 2013). Much of the research on the tapir species in Central and South America has focused on frugivory , which has shown that the diets of these species in forests, lowland river basins, and marshlands are extremely diverse and include more than 100 plant species (Salas and Fuller 1996; Lizcano and Cavelier 2000; Downer 2001; Tobler 2002; Chalukian et al. 2013).

4.4 Body Size and Habitat Associations

Of the ruminants, members of the Tragulidae and Moschidae are restricted to the smallest size classes, in contrast to the Cervids and Bovids that have larger body mass ranges (Fig. 4.2). Cervids in SSEA do not appear to get as large as Bovids (Fig. 4.2) and in both Cervids and Bovids, the greatest number of species are found in the smaller body mass classes (Fig. 4.2). The largest Cervid in Asia —the Kashmir stag , Cervus elaphus hanglu, a species that shares its genus and species lineage with the only deer species found in Africa, the barbary stag Cervus elaphus barbarus—has a body mass comparable to the body mass of the Equus spp. in Asia; all other Cervid species in Asia are smaller. The nonruminants of SSEA are much larger (>150 kg) with 4 of the 7 nonruminants falling within the megaherbivore category.

Fig. 4.2
figure 2

The species of four families of the order Artiodactyla (M Moschidae (5 species), T Tragulidae (8 species), C Cervidae (20 species), B Bovidae (30 species)) found in South and Southeast Asia distributed by body mass

Full size image

Although large herbivores in SSEA occupy a range of different habitats from arid grasslands to rainforests, there are discernible habitat associations within different groups. Tragulids and Moschids , which are predominantly browsers and frugivores , are restricted to forested habitats. Cervids and Bovids occur in a wide array of different habitat types, although the habitat ranges of Cervids tend to be more restricted than those of Bovides. In general, Cervids are nearly never found in extreme dry habitats, presumably because of their greater dependence on moisture when compared to Bovids. Among nonruminants, the Equids are restricted to arid habitats, the single Tapirid to forests and river basins, the Rhinocerotids to forests and grasslands, while the Asian elephant occurs in a range of habitats from tall grasslands to savannas to forests.

The only ruminant found in the dry and hot landscape that the Indian wild ass Equus hemionus inhabits, is the chinkara Gazella benetti. The chinkara, <20 kg, is smaller by an order of magnitude when compared to the wild ass, ~200 kg, which is a non-ruminant. Constraints imposed by the sparse and patchily distributed forage in these hot and dry habitats seem to allow only the smallest ruminants to persist in these inhospitable habitats. In contrast to the low number of species inhabiting extremely hot and dry habitats, the kiang Equus kiang, which inhabits dry and cold high altitude terrain that supports only sparse and patchily distributed forage in conditions of extreme seasonality, is sympatric with multiple ruminant species: mountain goat and sheep species that are generally mixed feeders , and the larger yak, a Bovini species that is considered a grazer . Cold and dry habitats in SSEA, therefore, support a more diverse ruminant assemblage when compared to hot and dry habitats.

4.5 Conclusions and Future Directions

The foraging preferences exhibited by large herbivore species in SSEA, especially at the extreme ends of the body mass gradient, suggest that body mass plays a role in shaping the forage preference of these species (Fig. 4.1). Additionally, the absence of ruminants from the megaherbivore class in this assemblage supports the proposition that the foregut trait syndrome does not commonly support body mass >1000 kg (Van Soest 1994; Clauss et al. 2003). Existing data suggest that Cervid species in Asia are predominantly mixed feeders, some throughout the year and others seasonally, and that the level of concentrate feeding by deer species increases with decreasing body mass (Table 4.1). The majority of Bovid species <200 kg are mixed feeders too, and the largest Bovids (wild cattle >500 kg) are roughage feeders of primarily grasses (Table 4.1).

From Table 4.1, we see that there is an acute lack of data on the foraging ecology of the majority of large herbivore species in SSEA. The framework—based on interacting causal factors of body mass, digestive trait syndromes, and habitat heterogeneity—that helped understand patterns observed in the foraging by large herbivore assemblages in Africa was established nearly half a century ago (Bell 1970, 1971; Jarman 1974). That framework, however, has yet to be fully tested on a large herbivore assemblage in SSEA. In general, classifying large herbivores either as grazers, browsers , or mixed feeders helps us understand their foraging ecology. However, the seasonal variation in the levels of available nutrients , roughage, and quantity in plant material ultimately determines what a large herbivore consumes over space and time, and forage selection is directed by a complex set of drivers that does not allow for a distinct optimal prediction (Hanley 1982). In the context of prevailing global change , understanding these mechanisms has got a lot more challenging as bush encroachment further modifies grasslands (Mitchard and Flintrop 2013) and increased levels of CO2 can potentially change the distribution of C3 and C4 grasses and woody vegetation worldwide (Chamaille-Jammes and Bond 2010). Finally, though natural history descriptions began in SSEA more than a century ago, it is imperative that ecological research increases in the region as SSEA seriously lags behind in understanding the fundamentals of its natural heritage.