OUP user menu

Fast Food for Slow Lorises: Is Low Metabolism Related to Secondary Compounds in High-Energy Plant Diet?

Frank Wiens , Annette Zitzmann , Nor Azman Hussein
DOI: http://dx.doi.org/10.1644/06-MAMM-A-007R1.1 790-798 First published online: 24 August 2006


The slow loris (Nycticebus coucang; Lorisidae) is a slow-moving, arboreal mammal with a very low metabolism relative to other eutherian species of its body mass. A slow pace of life has been causally linked to a low intake rate of usable energy due to a diet that is generally low in energy, is unpredictably periodically scarce, and contains high amounts of toxins or digestion inhibitors. To assess whether the slow loris faces any of these limitations, we studied its dietary habits in an area of West Malaysia (Manjung District, Perak State) by direct observations of radiocollared individuals and by fecal analysis. The diet was composed of 5 distinct types of food: floral nectar and nectar-producing parts, phloem sap, fruits, gum (another group of plant exudates), and arthropods. The largest proportion of feeding time was spent on phloem sap (34.9%), floral nectar and nectar-producing parts (31.7%), and fruits (22.5%). These foods should provide high amounts of easily digestible sugars, indicating that slow lorises did not face an energy-poor diet. Dietary habits were indistinguishable between rainy and dry seasons, even though most dry-season data were collected during periods of extreme drought induced by the 1997–1998 El Nino Southern-Oscillation event. However, many genera of food plants contain secondary compounds that are toxic or reduce digestibility. We suggest that low metabolism in slow lorises is at least partly related to the need to detoxify secondary compounds in high-energy plant diet.

Key words
  • gum-eating
  • nectar-eating
  • Nycticebus coucang
  • prosimian primate
  • sap-eating
  • tree-gouging

A slow pace of life in endothermic animals is characterized by low activity rates, low reproductive rates, low daily energy expenditure, and low basal metabolic rates (BMRs—Wikelski et al. 2003). Endotherms living a “slow life” may be under strong selection pressure to conserve energy, for example, if the acquisition of energy through their diet is exceedingly expensive (Mueller and Diamond 2001). However, other environmental factors such as predation risk can also potentially influence the pace of life (Ghalambor and Martin 2001). Specifically, a slow lifestyle has been interpreted as a mechanism for dealing with a diet that is either low in energy density (McNab 1986), becomes scarce during extended, but unpredictable, periods (Lovegrove 2000; McNab 1986), or contains high concentrations of toxic compounds or digestion inhibitors (Handy et al. 1999; McNab 1986; Sorensen et al. 2005). The 1st and 3rd of these limitations in energy supply have been cited as factors selecting for low metabolic rates in leaf-eating and ant- or termite-eating species, respectively (McNab 1978, 1984). Leaves, ants, and termites are supposed to yield few calories relative to the ingested volume (McNab 1978, 1984). Furthermore, they often contain high concentrations of chemical compounds with deleterious physiological effects on foragers (Deligne et al. 1981; Schmidt 1990; Swain 1979). The hypothesis that low energy requirements are advantageous to deal with unpredictable periodic food shortages may explain why species from geographic regions experiencing climatic extremes brought about by El Nino Southern-Oscillation (ENSO) events tend to have lower BMRs than species from other geographic regions (Lovegrove 2000). One of the regions where ENSO events often have marked climatic effects, causing failures of monsoon rains and therefore dramatic droughts, is southeast Asia (Chang 1997; Harger 1995; Stone et al. 1996). ENSO events are considered unpredictable because they occur in intervals between 2 and 10 years (Glantz 2001).

All members of the primate subfamily Lorisinae (following Schwartz et al. 1998), the lorises and pottos, are nocturnal and arboreal and have a peculiar mode of locomotion. Their movements are always smooth and deliberate and most of the time rather slow; they never jump (Ishida et al. 1992). Based on this slow lifestyle, we predicted low energy expenditure for lorisines. Indeed, all lorisines examined have BMRs lower than 60% of the predicted value (Hildwein 1972; Hildwein and Goffart 1975; Kleiber 1932; Müller 1979;Müller et al. 1985; Whittow et al. 1977). No estimates of daily energy expenditure are available on any lorisine species, but we expect it to be low as well (Ricklefs et al. 1996). Although the food habits of lorisines are still poorly known, Rasmussen and Nekaris (Nekaris 2000; Rasmussen 1986; Rasmussen and Nekaris 1998) suggest that the low metabolism and slow movements of lorisines are function-ally related to a diet containing a high amount of toxic insects.

We studied the slow loris (Nycticebus coucang; body mass 0.5–1.5 kg–Wiens and Zitzmann 2003), which has the lowest BMR of all lorisines, that is, 40% of the expected value (Müller 1979; Whittow et al. 1977). Only a few eutherians have similarly low relative BMRs, for example, sloths (Bradypus and CholoepusIrving et al. 1942; McNab 1978, 1984). The slow loris has a wide distribution in South and Southeast Asia, inhabiting tropical forests on the Asian mainland from Assam, south China to Vietnam, and further on the islands of Sumatra, Java, and Borneo up to the Phillipine Islands (Fooden 1991; Groves 1970;Lekagul and McNeely 1977; Napier and Napier 1967;Petter and Petter-Rousseaux 1979; Timm and Birney 1992). Various aspects of the biology of the slow loris have been explored in some detail by observing captive animals, but information from the field is extremely scarce. Before this study the only information available on the natural diet of the slow loris stemmed from a few chance observations by Elliot and Elliot (1967), Lim et al. (1971), and Medway (1978); an analysis of 4 stomach contents by Fooden (1967, 1976); direct observations of 1 radiocollared female and an analysis of 5 of its feces by Wiens (1995); as well as 21 observed feeding events plus the feces of 2 live captures by Barrett (1984). These data sets were probably not representative of natural diets of slow lorises, but suggested that slow lorises eat primarily fruit, and supplement their diet with invertebrates (including ants), vertebrates, leaf parts, and possibly plant gum, nectar, and bark.

We studied whether the slow loris indeed faces any of the 3 forms of limitation in energy supply proposed as main factors in the evolution of a slow lifestyle. To assess potential dietary energy constraints, we observed the foraging behavior of slow lorises in the wild and analyzed feces. We compared rainy- and dry-season data to determine whether unpredictability in food supply potentially affects slow loris lifestyle.

Materials and Methods

Study site.—We conducted field research within an 11-km2 strip of coastal land incorporating forested parts of the Segari Melintang, Tanjung Hantu, and Batu Undan Forest Reserve and adjacent more open areas around the village Labuan Bilis in Manjung District, Perak State, West Malaysia (4°18′N, 100°34′E). The 3 forest reserves are part of an isolated forest patch of approximately 3,000 ha that is bordered by plantation land (mainly oil palm [Elaeis guineensis] and mbber [Hevea brasilensis]), a river delta, and the sea. The relief of the study area is plain to gently undulating; elevation extends from sea level to 60 m above sea level. The area is crossed by a number of small, seasonal streams. Vegetation within the forest reserves is lowland dipterocarp forest and freshwater alluvial swamp forest. One of the hilly parts of the area where data were collected was covered with unlogged primary forest (Perak Virgin Jungle Reserve No. 1, totaling 408 ha). In other hilly parts and most flat parts of the study area selective logging conducted in earlier years has removed many of the taller, commercially valuable trees. As a result, the canopy in these parts is broken. In all forested parts the understory vegetation was widely dominated by the bertam palm (Eugeissona tristis). Vegetation in the parkland bordering the forest reserves is mainly secondary padang savanna (Whitmore 1984).

Annual average temperature was 26.7°C for 1968–1999, with minimum and maximum annual averages of 26.1°C and 27.6°C, respectively (recorded for 1998, during the study period). Annual rainfall average for 1951–1999 was 1,785 mm, with the highest precipitation occurring during October–December (northeast monsoon). The short dry season typically lasts from June to July (in litt., Malaysian Meterological Service). In 1997–1998 the ENSO phenomenon brought unusually long drought periods from January 1997 to June 1997 and February 1998 to April 1998 as well as heavier than usual rainfalls in May 1998 and December 1998 (Malaysian Meteorological Service, in litt.).

Trapping and observation.—Between May 1995 and July 1999, we captured 33 slow lorises either by hand (42 captures and recaptures), with wire-mesh live traps baited with banana and hung in trees (37 captures and recaptures), or in specially designed traps that were mounted to cover the inflorescences of the bertam palm, where slow lorises often fed (5 captures and recaptures). Upon capture, all animals except 2 infants weighing less than 200 g were immobilized with an injection of dissolved tiletamin and zolazepam (dosage: approximately 15 mg/kg body mass) and standard physical measurements were taken. Body mass (mean ± SD) of adult males was 737 ± 111 g (n = 8) and that of adult nonpregnant females was 637 ± 61 g (n = 11; the difference between males and females was significant; t-test: d.f. = 17, P = 0.023). We marked newly captured animals with transponders (Trovan, Euro I.D., Weilerswist, Germany) injected subcutaneously. We fitted 22 adult and subadult slow lorises with radiocollars (Biotrack Ltd., Wareham, United Kingdom) and tracked them on 327 nights for a total of about 600 h. We tracked 15 individuals intensively (>20 location records) for periods of between 1 and 11 months. Slow lorises in the study area became active around dusk and usually terminated their activities shortly before dawn (Wiens and Zitzmann 2003). In the following, we refer to the time between sunset and sunrise as “active time.” We conducted radiotracking by approaching on foot with a 4-element Yagi antenna and portable receiver until we actually saw the animal (“sighting,” n = 630) or until we identified the exact location of the animal (usually the tree they were staying in; n = 575). During each sighting of an animal we recorded the 1st behavior seen as an instantaneous observation (Altmann 1974). We recorded feeding when animal prey or plant material were being swallowed, chewed, or brought to the mouth (n = 141 instantaneous observations). In some cases where the animals could not be seen clearly because of vegetation and foliage obstructing the view, we scored feeding because movement pattern in combination with falling fruit or flowers indicated that the animal was doing so (n = 27 instantaneous observations). Sequences of behavior and observations not recorded as instantaneous observations were recorded ad libitum. Whenever an animal was feeding, we recorded the particular food item. Slow lorises that were regularly followed habituated quickly and could be observed without causing obvious disturbance sometimes at distances <5 m. We observed focal animals from the ground using a 4.5-V headlamp or flashlight and binoculars. Animals were handled in a humane manner following guidelines of the the American Society of Mammologists (Animal Care and Use Committee 1998). All methods were approved by the Economic Planning Unit at the Prime Minister's Department in Kuala Lumpur (permit UPE: 40/200/19 SJ. 179).

Dietary measures.—We determined diet from direct observation and analysis of feces. All observational data came from 15 intensively tracked lorises. We conducted quantitative analyses of observational data only on instantaneous observations. To avoid possible bias from irregular sampling intervals we included only instantaneous feeding observations on any 1 animal that were separated by >2 h. Two hours is the time required by a slow loris to cross the length of an average home range in traveling speed and observations separated by >2 h, therefore, were considered independent. The remaining 139 in-stantaneous feeding observations were used to determine percentage feeding times on different food items. We pooled all independent instantaneous feeding observations for each of the 15 individuals (total number per individual 9 ± 12 SD) and calculated an average feeding time per individual for each food type.

We collected a total of 118 complete fresh fecal pellets (combined dry mass: 136.4 g) of 25 captured animals (adults and subadults) from traps and from cages where we kept animals shortly before and after immobilization. Feces containing banana bait were not collected. We stored feces in 70% ethanol. Pellets from the same catch were treated as 1 sample. We included 47 fecal samples in the analysis (1–6 per animal). Fecal samples from the same loris were collected >5 days apart (average: 149 days) and, therefore, considered independent samples. Before analysis we softened feces by soaking in water, then dissected them. We examined component parts using a binocular microscope and grossly assigned them to 1 of the following categories: fruit part (whole seed, seed part, or fruit fiber), wood and bark, flower part (pollen grain, anther, petal, or receptacle), plant gum, or arthropods. Although for seeds and flower parts it was often obvious that they came from different species, we identified them to species only rarely. Larger remains from arthropods sometimes allowed identification to ordinal level. We expressed results of fecal analysis in terms of percentage of total number of fecal samples in which food items occurred. We recorded numbers of prey individuals per sample only for ants and lepidopteran larvae (by counting heads).

We measured sugar content (mean ± SD percent sucrose equivalent) of the most important single dietary item, floral nectar of the bertam palm (see “Results”) by determining the index of refraction of the nectar with a handheld refractometer (Bellingham & Stanley Ltd., Tunbridge Wells, United Kingdom) directly in the forest. We obtained samples from 33 flowers bearing nectar.

Data analysis.—For comparison between seasons, we treated months with less than 200 mm of rainfall as the dry season and the remainder as the rainy season. To compare feeding behavior between seasons we only used 8 individuals for which we obtained >5 independent sightings per season. We tested potential seasonal changes in diet composition among 6 individuals that inhabited the primary forest. We used only those individuals because our data were most complete for the primary forest. We applied Wilcoxon tests (Zar 1996) to compare foraging behavior and chi–square tests (Zar 1996) to determine differences in the occurrences of food remains in feces. All probabilities reported here are 2-tailed and statistical significance was accepted at the α = 0.05 level.


Foraging behavior.—Feeding accounted for an average ± SD of 20.5 ± 12.1% (n = 15) of the active time of slow lorises at Manjung. The diet consisted of 5 food types: plant sap; plant gum, which is a group of water-soluble exudates that seal wounds (Bearder and Martin 1980); floral nectar and flowers; fruits; and arthropods. Slow lorises spent most of their feeding time ingesting phloem sap and floral nectar or nectar-producing parts (Fig. 1). Slow lorises were not observed feeding on items other than those scored for quantitative analysis.

Fig. 1

—Proportion of feeding time slow lorises at Manjung District, Perak State, West Malaysia, spent on the intake of 5 different food types (values shown are median values). The sample sizes given are the number of slow lorises (nind.) and the number of feeding events (nevents).

When slow lorises ate sap they perforated the superficial layer of the cambium of trees or lianas (see Table 1 for species list) using their lower anterior teeth and lapped up the exposed sap. Slow lorises spent only a short time at any particular sap hole (<2 min) and then quickly moved a few meters within the same tree to gouge the next hole. Preferred trees were riddled with hundreds of small holes. On thinner twigs sometimes the bark was chewed off from larger areas.

View this table:
Table 1

Food plants of slow lorises at Manjung District, Perak State, West Malaysia, shown as instantaneous observations and identification in feces (as percentages of all observations or identifications).

Food type or speciesFamilyPlant typeInstantaneous feeding observations (%)Occurrence in feces (%)
Nectar and floral parts
Eugeissona tristisPalmaePalm41.021.3
Grewia panicultaTiliaceaeTree2.9?
Planchonella obovataSapotaceaeTree2.9
Ganua motleyanaSapotaceaeTree2.2
Buchanania arborescensAnacardiaceaeTree7.9
Mangifera griffithiiAnacardiaceaeTree5.8
Buchanania sessifoliaAnacardiaceaeTree3.6
Prunus polystachyaRosaceaeTree2.2
Ocrantomelon daoAnacardiaceaeTree0.7
Dacryodes rugosaBurseraceaeTree0.7
Neo-Uvaria foetidaAnnonaceaeTree0.7
Anacardium occidentaleAnacardiaceaeTree2.9?
Gluta curtisiiAnacardiaceaeTree0.7?
Ficus spp.MoraceaeTree6.5?
Diospyros kingiiEbenaceaeTree2.9?
Pometia pinnataSapindaceaeTree0.77
Ixonanthes icosandraAnnoniaceaeTree0.7?
Grewia laevionataTiliaceaeShrub0.77
Rhodomyrtus tomentosaMyrtaceaeShrub0.716.7c
Elaeis guineensisPalmaePalm0.72.1c
  • a As indicated by pieces of bark and wood.

  • b Identified from fiber.

  • c Identified from seeds.

Slow lorises collected gum mostly from sites where it had already exuded (because of previous injury) and solidified, and they used their lower anterior teeth as scoops. The quantity of available gum was usually large and slow lorises spent a substantial amount of time (≤10 min) at each site.

The single most frequently consumed food item was floral nectar of the bertam palm (Table 1). This stemless palm grows in dense clusters with fronds reaching up to 7 m in length. Its inflorescence is erect, 0.8–3.0 m tall, and consists of several hundred flowers (F. Wiens et al., in litt.). The amount of nectar available from 1 inflorescence was usually large. Lorises spent ≤30 min slowly climbing up and down on a single inflorescence to inspect several flowers and lap up nectar. Nectar contained 10.4 ± 4.3% sucrose equivalent (n = 33). Small and soft flowers of other plant species (Table 1) were swallowed complete.

Fecal analysis.—We could not directly trace nectar and plant sap in feces. However, 51.1% of fecal samples contained pieces of bark and wood. Because slow lorises are likely to ingest small fragments of fresh bark or wood each time they gouged holes or peeled off bark we considered the occurrence of these items as indicators of sap-eating. Flower parts indicative of nectar-eating were present in 44.7% of all feces. Chitinous remains of arthropods were present in 91.5% and parts of fruit in 70.2% of fecal samples. We found slimy translucent masses of reddish brown color that swelled up enormously after soaking in water in 55.3% of fecal samples. We assumed this to be mucilage from plant gum (Bearder and Martin 1980).

Seeds and seed parts found in feces belonged to 19 plant types. Single fecal pellets contained up to 4 types of seeds. However, only 2 seed types could be identified to the species level (Table 1).

Arthropods found in feces were overwhelmingly insects (Coleoptera, Orthoptera, Lepidoptera [larvae and imagi], and Hymenoptera), with the remainder being spiders. Ant remains (heads) were present in 40.4% of fecal samples. Only 1 fecal sample contained a substantial number of ants (23). All other samples contained fewer than 6 ant individuals. Remains from lepidopteran larvae were present in 12.8% of fecal samples with no sample containing more than 1. Feces did not contain any termite remains.

Comparisons between seasons.—The total feeding time, calculated as the proportion of time active, did not differ between rainy and dry season (Wilcoxon test: n = 8 lorises, Z = −1.400, P = 0.161). We did not find seasonal differences in the proportions of feeding time spent on the 5 different food types (Fig. 2). Similarly, we did not detect seasonal differences in the frequencies of occurrence of food remains in feces (chi-square test: χ = 1.225, d.f. = 4, P = 0.874).

Fig. 2

—Comparison of proportions of feeding time spent by slow lorises on the intake of 5 food types between rainy season and dry season in Manjung District, Perak State, West Malaysia. Data from the 2 seasons were paired data for the 6 individuals living in primary forest. Horizontal lines indicate median values, box indicates 25th to 75th percentiles, and error bars indicate 10th to 90th percentiles. None of the food type proportions differed significantly between seasons (Wilcoxon tests).


Our behavioral observations showed that slow lorises concentrated most of their feeding effort on the intake of phloem sap, followed by floral nectar and nectar-producing flower parts, and fruit. The fecal analysis confirmed those trends. This suggests that our data were not biased toward observing certain food types over others, and that fruit, floral nectar, and phloem sap indeed constituted a major part of slow loris diet. Sap- and gum-eating have been reported from many primate species (prosimians, callithrichids, and African cercopithecines), but only a few of them use sap or gum as a major source of nutrients. These species have been referred to as members of an exudate-feeding guild (Nash 1986; Sussman and Kinzey 1984). The slow loris, apparently, is another member of this guild. Most exudate-eating primates lick or scrape gum or sap from surfaces after previous insect infestation or breakage. Specialized gouging behavior to elicit sap or gum flow has so far been only found in 3 genera (Callithrix, Cebuella [Callitrichidae], and Phaner [Cheiroga-leidae]) and seems to be rare among vertebrates in general (Coimbra-Filho and Mittermeier 1976; Petter et al. 1971).

Dietary habits in the wild are only known for 2 other lorisine species, the similar-sized African potto (Perodicticus potto) and the smaller slender loris (Loris tardigradus) from South India and Sri Lanka. Stomach and cecum contents of 41 pottos contained on average 65% fruits, 21% gum, and 10% insects, mainly ants, with the remainder being leaves and fungi (Charles-Dominique 1977). Pottos have also been observed to feed on floral nectar (Grünmeier 1990) and are known to eat small vertebrates such as bats (Charles-Dominique 1977; Jones 1969) and birds (Charles-Dominique 1977), as well as geckos (Walker 1969). Information about diet of the slender loris in the wild stems from the direct observation of feeding behavior (Nekaris 2000; Fetter and Hladik 1970). They indicate that this species relies mainly on animal food, of which ants and termites form a large part. Other foods observed being consumed by slender lorises are molluscs, small vertebrates, and plant gum (Nekaris 2000). Recently, there is indirect evidence that the pygmy slow loris (Nycticebus pygmaeus), like the slow loris, is a tree-gouger and exudate-eater (Tan and Drake 2001).

It has been suggested that animals generally operate at an intensity close to maximum potential metabolism set by the rate of energy assimilation from food (McNab 1980). Hence, diet should profoundly influence the pace of life animals live. Species with permanent access to high-energy diet are expected to have a fast pace of life. Why do slow lorises that mostly eat nectar, sap, and fruit have a slow lifestyle? Three proximate explanations for a slow lifestyle have been put forward: food items are low in energy content; high-energy food is not available during extended, unpredictable periods; and secondary compounds in the diet reduce assimilation of energy from the diet.

Is slow loris diet energy-poor?—Fruit, floral nectar, and phloem sap provide high amounts of easily digestible monoand disaccharides (Baker 1975; Crafts 1961; Percival 1961; Waterman 1984; Zimmermann 1961), lipids, or both (Waterman 1984). Thus, we conclude that it is very unlikely that the slow pace of life in slow lorises is due to an energy-poor diet. Rather, slow-living slow lorises ingest a high-energy diet similar, for example, to that of fast-living sunbirds, honey-eaters, or nectarivorous bats (McNab 1983, 1988).

Gum also generally contains high concentrations of carbohydrates (Anderson et al. 1972; Anderson and Leon de Pinto 1985; Bearder and Martin 1980; Coimbra-Filho and Mittermeier 1976). For example, the gum of the cashew tree (Anacardium occidentale. Table 1) is 84% carbohydrate (cf. Coimbra-Filho and Mittermeier 1976). However, gum may not be a high-energy type of food. Gum, but not sap, nectar, and fruits, may often be largely indigestible for mammals that lack microbial fermentation because of the presence of 1-4-β-linkages between sugar residues in gum (Nash 1986; Waterman 1984). The slow loris is reported to lack a chambered site for microbial fermentation in its digestive tract (Osman-Hill 1953). Remains of gum found in feces indicate that slow lorises were not able to completely digest gum of some tree species.

Does high-energy diet come seasonally?—The 2nd form of limitation in energy supply suggested to be functionally related to a slow lifestyle in endotherms is that of an unpredictable and long period of shortage of food—and hence of energy supply. However, examination of our observational data showed that nectar of the bertam palm, the major food source for slow lorises in the primary forest, was available year-round. Nectar-producing bertam inflorescences were even available during the ENSO-induced periods of extreme drought in 1997 and 1998. This is consistent with other observations suggesting that bertam palms flower year-round at a fairly constant rate (Wong 1959). Bertam palm inflorescences attracted a variety of insects that slow lorises readily consumed. Sap and gum also are constantly available (Bearder and Martin 1980), although their chemical composition may fluctuate (Stewart et al. 1973). Only fruit may become scarce during certain times in evergreen tropical forests (Whitmore 1984). Yet, we did not find any evidence for seasonal differences in fruit availability in diets of slow lorises. Another way to detect longer-lasting periodic food shortages is by looking at the nutritional state of the animals themselves. The radiotracked individuals seemed in good physical condition during all captures; we observed a female slow loris successfully raising offspring during the 1997–1998 ENSO event. We conclude that these animals did not face extended periods of food shortage.

Is energy turnover compromised by plant secondary compounds?—A slow pace of life has further been linked to a diet containing high amounts of digestion-inhibiting compounds or toxins (McNab 1986; Mueller and Diamond 2001; Sorensen et al. 2005). Digestion-inhibitors can reduce energy-assimilation rates in the gut by binding with the substrate to be digested, inhibiting digestive enzymes, or being antimicrobial (Rhoades and Gates 1976; Robbins et al. 1991; Silverstein et al. 1996; Song et al. 2002). Toxins include all compounds that interfere with specific physiological processes within cells (Brattsten 1979; Waterman 1984). Insects are potential animal food items that often contain toxic and digestion-inhibiting substances. Such substances have been found in species from many different orders of insects and other arthropods (Teuscher and Lindequist 1994). Rasmussen and Nekaris (Nekaris 2000; Rasmussen 1986; Rasmussen and Nekaris 1998) explicitly mention ants, termites, and butterfly or moth larvae as “toxic groups” and suggested that the low BMRs of lorisines have evolved in relation to insectivory. Arthropods regularly appeared in the feces of slow lorises in small amounts. However, ants, termites, and lepidopteran larvae did not constitute a major proportion of these arthropods, contrary to what was found in the Calabar angwantibo (Arctocebus calabarensisCharles-Dominique 1977; Jewell and Oates 1969). Ants found in feces of lorises were mostly aggressive diurnal species of the genus Oecophylla that likely attacked the lorises and were subsequently ingested during grooming. The low occurrence of arthropod ingestion did not allow us to test whether such arthropods were toxic or repugnant.

Like animal matter, plant matter also can contain toxic or digestion-inhibiting compounds. Bark, sap, gum, and flowers of many plant species and families are known to contain secondary compounds with proven or suspected negative effects on mammals (Gartlan et al. 1980; Nash and Whitten 1989; Rosenthal and Janzen 1979; Waterman 1983, 1984; Wink 1999; Wrangham and Waterman 1981). Even floral nectars can contain toxic constituents (Adler 2000; Baker 1978). At least 7 important plant genera of which sap or gum was consumed by slow lorises during our study contain toxins or digestive deterrents (Anonymous 1996; Table 2). Some of these plants are known to be dangerous for humans. The sap of Gluta and Mangifera produce sores on the skin. The sap of some species of Mangifera have sometimes been used criminally to injure an enemy by causing vomiting and purging after intake. The bark of Gluta, dried, powdered, and given in water, kills humans. The gum of A. occidentale also is capable of blistering the skin, and, if taken internally, causes gastroenteritis with loss of control of the muscles and interrupted respiration (Burkill 1935). Thus, we hypothesize that slow lorises indeed did consume toxic or digestion-inhibiting secondary compounds along with their high-energy diet. However, a more direct proof of the ingestion of such compounds, for example, by the detection of biomarkers in excretions of slow lorises, has yet to be found.

View this table:
Table 2

Plants of which slow lorises consumed exudates and known secondary compounds in the plants (Anonymous 1996).

GenusParts eatenListantaneous feeding observations (%)Secondary compounds
ChisochetonSap7.2Steroids with furan- and pyran-rings
MangiferaSap5.8Steroids, spirolactones, furanones, farnesanolid, glucosyl-pentahydroxy-benzophenon, heptadecenyl-1,3-benzendiol, pentahydroxy-xanthenone, tannins (trigallic acid)
AnacardiumGum2.9Phenolics with long-chained carbohydrates, flavonoids
PrunusSap2.2Tannins, cyano-glycosides

Slow lorises may depend on plant parts containing toxic or digestion-inhibiting compounds not only for satisfying their need for energy, but also their need for nutrients such as nitrogen, calcium, magnesium, and potassium (Bearder and Martin 1980; Garber 1984; Hayashi et al. 2000).

A mechanism to deactivate toxins involves their conjugation with glucuronic acid, sulfate, glutathione, methyl groups, or acetyl groups (Clarke and Burchell 1994; Häussinger et al. 1988; Scheline 1991). The conjugates are then excreted via urine or bile. An upper limit to detoxification rate is apparently set by the supply of a cosubstrate for the conjugation (Foley et al. 1995; Illius and Jessop 1995). Thus, animals must balance intake rate of foreign compounds with nutrient intake rate (Provenza 1997). The main cosubstrate used in many mammals is glucuronic acid (Baldwin et al. 1980; Miners and MacKenzie 1991; Scheline 1991), which is derived from glucose. Excretion of glucuronic acid conjugates is thus often a drain on glucose reserves (Jessop and Illius 1997). Cork (1981) estimated that glucuronic acid excretion due to absorption of allelochemicals from Eucalyptus foliage might increase glucose requirements of koalas (Phascolarctos cinereus) by 20%. Euxanthic acid, a major component of the dye Indian yellow, is a sugar conjugate isolated from the urine of cows fed leaves of the mango tree (Mangifera indica), which is closely related to a common food tree for slow lorises (Morton 1987; Table 2). Woodrats (Neotoma stephensi and N. albigula) respond to energetic constraints imposed by plant secondary compounds by reducing energy expended on BMR and locomotor activity (Sorensen et al. 2005). High requirements for glucose, not only as a source of energy but also as a cosubstrate for detoxification, could explain why even slow lorises with their very low metabolic rate and rather slow locomotion might be compelled to rely on a sugar-rich diet. Although primate tree-gougers of the genera Callthrix and Cebuella also have lower than predicted BMRs, pointing to a generalized phenomenon (Genoud et al. 1997), less-pronounced reliance on plant exudates and their faster locomotion indicate less severe energetic constraints in these diurnal animals (Chivers 1998).

We suggest that slow lorises have high-energy diet available year-round. However, they may not assimilate more energy from their high-energy diet than they need to maintain their slow lifestyle. Their slow lifestyle may largely be determined by the need to detoxify plant secondary compounds in their diet.


This research was sponsored by the Deutsche Forschungsgemeinschaft (grant Pr 58/21 to H. Preuschoft), the Fazit Foundation (Graduate Student Fellowship to FW), the Universitätsverein Bay-reuth, Germany, and the Daimler-Benz AG. We would like to thank the current Director-General of the Forest Research Institute Malaysia (FRIM) Abdul Razak bin Mohd. Ali and his predecessor Mohd. Salleh bin Mohd. Noor for hosting our research in Malaysia. The Department of Wildlife and National Parks (PERHILITAN) helped in finding a suitable study site. We are also grateful to Lim Boo Liat, L. Ratnam, Chai Koh Shin, H. Preuschoft, and D. von Hoist for their continuous support. Herbarium material was identified by Saw Leng Guan and Mat Asri bin Ngah Sanah (FRIM). M. Wikelski kindly provided information on secondary compounds in food plants. Special thanks for their commitment go to our field assistants Tedung bin Uia and Tiee bin Sipang. We thank M. Wikelski, M. Hau, H. Rödel, and 2 anonymous reviewers for useful comments on the manuscript.


  • Associate Editor was Craig L. Frank.

Literature Cited

View Abstract