Excretion Patterns of Fluid and Different Sized Particle Passage Markers In

Home , Pygmy hippopotamus

Schwarm, A; Ortmann, S; Wolf, C; Streich, W J; Clauss, M (2008). Excretion patterns of fluid and different sized particle passage markers in banteng (Bos javanicus) and pygmy hippopotamus (Hexaprotodon liberiensis): Two functionally different foregut fermenters. Comparative Biochemistry and Physiology - Part A, Molecular and Integrative Physiology, 150(1):32-39. Postprint available at: University of Zurich http://www.zora.uzh.ch Zurich Open Repository and Archive Posted at the Zurich Open Repository and Archive, University of Zurich. http://www.zora.uzh.ch

Winterthurerstr. 190 Originally published at: Comparative Biochemistry and Physiology - Part A, Molecular and Integrative Physiology 2008, 150(1):32-39. CH-8057 Zurich http://www.zora.uzh.ch

Year: 2008

Excretion patterns of fluid and different sized particle passage markers in banteng (Bos javanicus) and pygmy hippopotamus (Hexaprotodon liberiensis): Two functionally different foregut fermenters

Schwarm, A; Ortmann, S; Wolf, C; Streich, W J; Clauss, M

Posted at the Zurich Open Repository and Archive, University of Zurich. http://www.zora.uzh.ch

Originally published at: Comparative Biochemistry and Physiology - Part A, Molecular and Integrative Physiology 2008, 150(1):32-39. Excretion patterns of fluid and different sized particle passage markers in banteng (Bos javanicus) and pygmy hippopotamus (Hexaprotodon liberiensis): Two functionally different foregut fermenters

Abstract

Processing of ingesta particles plays a crucial role in the digestive physiology of herbivores. In the ruminant forestomach different sized particles are stratified into a small and a large particle fraction and only the latter is regurgitated and remasticated to smaller, easier-to-digest particles. In contrast, it has been suggested that in non-ruminating foregut fermenters, such as hippopotamuses, larger particles should be selectively excreted since they tend to be digested at a slower rate and hence can be considered intake-limiting bulk. In our study we determined the mean retention time (MRT) of fluids and different sized particles (2 mm and 10 mm) in six pygmy hippos (Hexaprotodon liberiensis) and six banteng (Bos javanicus) on a diet of fresh grass at two intake levels. We used cobalt ethylendiamintetraacetate (Co-EDTA) as fluid and chromium (Cr)-mordanted fibre (2 mm) and cerium (Ce)-mordanted fibre (10 mm) as particle markers, mixed in the food. Average total tract MRT for fluid, small and large particles at the high intake level was 32, 76 and 73 h in pygmy hippos and 25, 56 and 60 h in banteng, and at the low intake level 39, 109, and 105 h in pygmy hippos and 22, 51 and 58 h in banteng, respectively. In accordance with the prediction, large particles moved faster than, or as fast as the small particles, through the gut of pygmy hippos. In contrast, large particles were excreted slower than the small particles in the ruminant of this study, the banteng. Pygmy hippos had longer retention times than the banteng, which probably compensate for the less efficient particle size reduction. Although the results were not as distinct as expected, most likely due to the fact that ingestive mastication of the larger particle marker could not be prevented, they confirm our hypothesis of a functional difference in selective particle retention between ruminating and non-ruminating foregut fermenters. 1

Comparative Biochemistry and Physiology A 150: 32-39, 2008, doi: 10.1016/j.cbpa.2008.02.022

Excretion patterns of fluid and different sized particle passage markers in banteng (Bos javanicus) and pygmy hippopotamus (Hexaprotodon liberiensis): two functionally different foregut fermenters

Angela Schwarma,b*, Sylvia Ortmanna, Christian Wolfc, W. Jürgen Streicha, Marcus Claussd

aLeibniz Institute for Zoo and Wildlife Research (IZW) Berlin, Alfred-Kowalke-Str. 17,

10315 Berlin, Germany, [email protected] bFreie Universität Berlin, 14195 Berlin, Germany cHahn-Meitner-Institut (HMI), 14109 Berlin, Germany dClinic for Zoo Animals, Exotic Pets and Wildlife, Vetsuisse Faculty, University of Zurich,

Winterthurerstr. 260, 8057 Zurich, Switzerland, [email protected]

*corresponding author

Running head: Ingesta passage in two foregut fermenters

Abstract

105 h in pygmy hippos and 22, 51 and 58 h in banteng, respectively. In accordance with the prediction, large particles moved faster than, or as fast as the small particles, through the gut of pygmy hippos. In contrast, large particles were excreted slower than the small particles in the ruminant of this study, the banteng. Pygmy hippos had longer retention times than the banteng, which probably compensate for the less efficient particle size reduction. Although the results were not as distinct as expected, most likely due to the fact that ingestive mastication of the larger particle marker could not be prevented, they confirm our hypothesis of a functional difference in selective particle retention between ruminating and non- ruminating foregut fermenters.

Key words: Mean retention time; Forestomach; Foregut fermentation; Pygmy hippopotamus;

Cattle; Herbivore

Introduction

Among the variety of gastrointestinal tract designs that can be found in herbivorous mammals, several different foregut fermentation systems have evolved – e.g. in ruminants, camelids, sloths, peccaries, hippopotamuses, colobine monkeys, or macropod marsupials

(Langer 1988; Stevens and Hume 1998). Among the different foregut fermenters, two taxonomic groups, the camelids and the ruminants (taxonomic definition: Ruminantia, the

‘true’ ruminants), evolved a mechanism by which the ingested food is regurgitated and submitted to repeated mastication (i.e., rumination; the functional definition “ruminants” therefore includes camelids). Although a similar behaviour, termed ‘merycism’, has been observed repeatedly in macropods (Home 1814; Owen 1834; Moir et al. 1956; Calaby 1958;

Mollison 1960; Barker et al. 1963; Hendrichs 1965), the actual occurrence, circumstances and physiology of this behaviour have not been quantified; the seemingly low prevalence of it, as indicated by absence of reference to it in many experimental studies on macropod digestion, suggests that it is an occasional, facultative, but not obligatory strategy (Hume

1999).

Rumination has an important effect on the digestive physiology that sets functional ruminants apart: the particle size to which the ingesta is finally reduced to is distinctively smaller than in other similar-sized herbivores (Udén and Van Soest 1982; Grenet et al. 1984;

Okamoto 1997; comparing data from Clauss et al. 2002; Clauss et al. 2004a; Clauss et al.

2004b). Given the effect of particle size on the fermentation kinetics of plant material

(reviewed in Clauss and Hummel 2005), this smaller ingesta particle size represents an important digestive advantage. 4

A prerequisite for the efficient functioning of the rumination process is that, in the forestomach, those ingesta particles that need to be re-masticated are separated from those that do not require further mastication. In both camelids and true ruminants, this is achieved by a mechanism of stratification, whereby the ingesta particles separate according to their specific gravity into more buoyant and more sedimenting particles, with buoyant particles being the larger ones and sedimenting particles the smaller ones (Lechner-Doll et al. 1991).

The anatomical positions of the cardia (the opening of the oesophagus into the forestomach) and the orifice through which particles leave the main chamber of the forestomach (the

‘ostium reticulo-omasale’ in true ruminants / the ‘ostium ruminoreticularis’ in camelids,

Langer 1988), ensure that larger particles are regurgitated for rumination and only smaller ones leave the forestomach.

In parallel to what is known in ruminants, it has been suspected that larger particles are also excreted slower from the forestomach of non-ruminating foregut fermenters (Langer

1988; Foley et al. 1995). However, it appears that in non-ruminating foregut fermenters, the slower excretion of larger particles would not make as much sense as in ruminants, due to the fact that large particles tend to be digested at a slower rate and hence can be considered potentially intake-limiting bulk. In hindgut fermenters it has been found that larger particles are selectively excreted, presumably in order to clear the gut and to maintain high food intake

(Björnhag et al. 1984; Björnhag 1987; Björnhag 1989; Hume and Sakaguchi 1991; Cork et al.

1999). In parallel to the observations in hindgut fermenters, Clauss (2004) suggested that larger particles should be selectively excreted in nonruminant foregut fermenters and reviewed literature reports that supported such an interpretation for sloths. Similarly, Hume

(1999) speculated, based on data from Dellow (1982) and Forbes and Tribe (1970), that large particles might be excreted faster than small particles from the forestomach of macropods. In feeding trials with two common hippopotamuses (Hippopotamus amphibius) and two pygmy 5 hippopotamuses (Hexaprotodon liberiensis), Clauss et al. (2004a) observed that in one common and in one pygmy hippo large particles (2-10 mm) passed the gastrointestinal tract in parallel to small particles (<2 mm), whereas they were excreted faster in both other respective individuals. Those observations were, to our knowledge, the first direct experimental indication that large particles are not excreted slower than smaller particles from the forestomach of a nonruminant foregut fermenter. These findings are in contrast to domestic ruminants and camelids who selectively retain large particles (Lechner-Doll et al.

1990). However the sample size in the study of Clauss et al. (2004a) was small, the markers had not been simultaneously validated in a ruminant, and the particle marker length was not exactly defined. Therefore, we performed passage experiments in pygmy hippopotamuses and a wild “cattle-type” ruminant, the banteng (Bos javanicus), using particle markers of different length (2 mm and 10 mm) that originated from the same marker batch. We predicted that in the nonruminant foregut fermenter (pygmy hippo), larger particles are excreted faster than smaller particles or move together with the smaller particles, in contrast to the ruminant where the larger particles should be excreted slower than the smaller ones.

Materials and Methods

The trials were performed with six pygmy hippos and six banteng at the Zoological

Gardens of Berlin (ZGB) and Halle (ZGH) in summer 2005 and 2006. Body mass (BM) of the pygmy hippos were measured at the beginning and the end of each trial period, whereas

BM of the banteng were estimated by the keepers by visual judgement (height and width) and age and sex as reference parameters. Details of the animals are summarized in Table 1.

The animals were fed with fresh grass only, the staple diet during summer. The dry matter fraction of the offered grass at Berlin and Halle zoo contained on average 93 ± 3 and

92 ± 1% organic matter, 63 ± 2 and 60 ± 1% neutral detergent fibre (aNDFom), 34 ± 2 and 33 6

± 1% acid detergent fibre (ADFom) and 3 ± 1 and 5 ± 1% Lignin (sa), respectively (for fibre terminology see Udén et al. 2005). Since the grass diet was usually supplemented with fruits and vegetables in hippos and with sugar beet pulp in banteng, an adaptation period of 14 days was allowed to pass before the trial started. It was planned to study each animal in two trials on different intake levels (with a second adaptation period of 5 days in between) – ad libitum

(high intake, HI), and, subsequently, at approximately 75 % of the individual ad libitum intake (low intake, LI). Each trial lasted 7 days. Due to a shortage of grass, one pygmy hippo

(animal 6) and three banteng (animals 10-12) could only be assessed at one intake level (HI) and some animals had to be fed grass hay (soaked in water); this never exceeded one individual day per animal (animal 2: at day 1 after marker feeding; animal 4: day 6; animal 7: day 1; animal 8 and 9: day 2).

All animals were fed separately. Due to the husbandry techniques at the respective zoos, not all animals could be kept separately at all times, and feeding regimes differed.

Three banteng (animals 10-12) were kept together; two of these animals received a coloring agent in their food ration (animal 10: titanium dioxide 40 g/d; animal 11: brilliant blue food colour, Sensient Food Colors Germany GmbH, Geesthacht, 2 g/d; both fed twice daily), so that faeces could be ascribed to the individual animals.

Three pygmy hippos (animals 1-3) received food once daily, in the afternoon; the other pygmy hippos (animals 4-6) as well as the banteng received food twice daily, in the morning and in the afternoon. Food items offered and leftovers were quantified on a daily basis by weighing. During the day (approximately 08:00-18:00), the pygmy hippos were kept on land with no access to a water pool. During the night (approximately 18:00-08:00), the pygmy hippos had free access to a water pool, except for animal 1 on the high intake level.

Drinking water was provided at all times to all animals. 7

Markers for ingesta retention, cobalt ethylendiamintetraacetate (Co-EDTA; fluid marker), chromium (Cr)-mordanted fibre (2 mm particle marker) and cerium (Ce)-mordanted fibre (10 mm particle marker), were prepared according to Udén et al. (1980) and Heller et al.

(1986). Before mordanting, grass hay was dried (40°C) and ground through 2 and 10 mm square perforated screens (Retsch catalogue no. 03.647.0167 and -0024) with a retsch mill

(SM 2000, Retsch GmbH & CoKg, Haan, Germany). After removing dust and smaller than the desired particles by dry sieving (by hand), small as well as large particles were incubated with 33 g and 75 g mordant per 100 g particles (sodium dichromate dihydrate, Na2Cr2O7 *

2H2O, and cerium(III) chloride heptahydrate, CeCl3 * 7H2O), respectively. After mordanting and washing, the particles were dried at 65°C. The marker dose applied was 0.2 g/kg BM of each mordanted fibre marker, and 0.03 g/kg BM of Co-EDTA, respectively. Markers were fed in the afternoon shortly before the regular feeding. Co-EDTA was dissolved in little water, mixed with particle markers and offered at time zero (t0) in pygmy hippos with a portion of fruits or grass, and in banteng with soaked sugar beet pulp for better acceptance.

The markers were mostly consumed completely within 5 to 60 minutes; the middle of the recorded time period was used as t0 in subsequent calculations. After 60 minutes, any marker leftovers were removed if present, and the regular food was provided.

Individual defaecations were collected two days prior to and 7 days after marker application during day hours, and the exact time of defaecation was noted. Faeces voided into the water pool at night were not sampled; however, faeces voided at night on land were collected the next morning and treated as one defaecation unit, and an average time (between the last check the previous evening, approximately 18:00 h, and the first check in the morning, approximately 8:00 h) was ascribed. Defaecations were collected completely, cleaned from sand (in the case of the banteng), weighed, thoroughly mixed, and an aliquot 8

(200-400 g) was taken and stored frozen (-20°C). Representative samples of food offered and leftovers were taken and stored frozen as well.

Dry matter (DM) content of samples was determined by drying at 60°C to constant weight. Dried faecal samples were ground with a ‘Nossener mill’ (Gebrüder Jehmlich GmbH,

Nossen, Germany, 1 mm round perforated screen) and 0.3 g of each sample was subsequently wet microwave digested with 8 ml nitric acid (HNO3, 65%) and 2 ml hydrogen peroxide

(H2O2, 30%) (in duplicate with a MDS-2000, CEM GmbH, Kamp-Lintfort, Germany; 630 W,

2450 MHz). Digested faecal samples were analysed for Co, Cr and Ce concentration by inductive-coupled plasma mass spectroscopy (ELAN 6000, PerkinElmer Life and Analytical

Sciences, Milano, Italy). Faecal marker concentrations were corrected for the highest level of the respective elements determined in the faecal samples taken prior to marker feeding.

Mean retention time (MRT) in the total gastrointestinal tract (GIT) was calculated according to Thielemans et al. (1978): This method calculates the area under the excretion curve and defines MRT as the time that separates the total area under the excretion curve in two equal parts:

MRT = ∑(ti * dt * ci) / ∑(dt * ci)

With ti = time after marker application (h), dt = time interval represented by marker concentration (calculated as (((ti+1 – ti) + (ti - ti-1)) / 2), and ci = faecal marker concentration at time i (mg/kg DM). The middle of the sampling intervals was used as ti.

Fluid MRT in the reticulorumen of banteng (MRTfluidRR) and the forestomach of hippos (MRTfluidFRST) was calculated according to Grovum and Williams (1973); this calculation is based on the decrease of the faecal liquid marker concentration ci at time t

(mg/kg DM) with time after marker application ti (h) according to the equation: lnci = -k*ti + b. 9 with k = rate constant (h-1) and b = intercept. The reciprocal of k represents the fluid MRT in the RR and FRST (Hungate 1966). However, in some animals and trials the calculation of

MRTfluidRR and MRTfluidFRST yielded no difference to the MRTfluidGIT. In these animals fluid marker concentration did not decline to the baseline level before marker feeding, but declined in small steps at a slight elevated level from approximately 55 hours after marker feeding until the end of the trial, resulting in a smooth slope and thus a low rate constant.

Therefore, fluid marker concentration was corrected in all animals as follows: when marker concentration dropped below 1% of the peak concentration, fluid marker concentrations were set to zero (modified correction of Bruining and Bosch 1992).

MRTparticleRR or MRTparticleFRST is calculated as follows, based on the assumption that fluid and particles do not differ in passage characteristics distal to the RR (empirically confirmed by Grovum and Williams 1973; Kaske and Groth 1997; Mambrini and Peyraud 1997) or the

FRST (empirically confirmed by Dellow [1982] for kangaroos):

MRTparticleRR = MRTparticleGIT – (MRTfluidGIT – MRTfluidRR)

MRTparticleFRST = MRTparticleGIT – (MRTfluidGIT – MRTfluidFRST).

The “selectivity factor” – defined as the quotient of particle over fluid MRT, or as the quotient of large over small particle MRT (Lechner-Doll et al. 1990) – was calculated for both the total GIT and the RR/FRST.

The t-test (for equal or unequal variances, depending on the respective data) served for comparisons between species (Table 2 and 4). Repeated measurements ANOVA was used for comparisons within species (Table 3). Comparisons between the small particle results and the fluid or large particle results, respectively, were performed using linear contrasts. As an influence of body mass could not be excluded for comparisons within species, body mass was included into the RM-ANOVA as covariate. However, following Clauss et al. (2007a), an influence of body mass for comparisons between species was not assumed. The monotonous 10 association between pairs of variables was measured by calculating Spearman´s correlation coefficient (SCC). This analysis should be considered as explorative since both intake levels were combined. The significance level was generally set to α=0.05. All statistical calculations were performed with the SPSS 15 software (SPSS, Chicago, IL).

Results

The dry matter intake (DMI), as well as the different MRT measurements are recorded in Table 2. The relative DMI (mean ± SD) in pygmy hippos was 34 ± 13 g/kg0.75/d at the high intake level (HI, n=6) and 19 ± 12 g/kg0.75/d at the low intake level (LI, n=5). In banteng, the relative DMI was 51 ± 23 g/kg0.75/d (HI, n=6) and 47 ± 9 g/kg0.75/d (LI, n=3), respectively.

The mean MRTfluidGIT, MRT2 mm particleGIT and MRT10 mm particleGIT in pygmy hippos was 32

± 6 h, 76 ± 16 h and 73 ± 17 h at the high intake level and 39 ± 9 h, 109 ± 19 h and 105 ± 18 h at the low intake level. In banteng, the respective MRTs were 25 ± 3 h, 56 ± 3 h and 60 ± 4 h at HI and 22 ± 1 h, 51 ± 2 h and 58 ± 5 h at LI. Most MRT measurements tended to be longer in the hippos as compared to the bantengs, with significant differences mostly at the low intake level (Table 2).

Exemplary excretion curves from both species are displayed in Figures 1 and 2. The fluid excretion curves of both species were similar, with a steep increase and peak concentration reached after approximately 20 h. Particle excretion curves looked different between species, because the peak concentration was reached later in pygmy hippos, leading to two well-separated peaks for particle and fluid markers in this species. However, in both species the particle marker was almost excreted after 150 h. In hippos the excretion of larger particles appeared to start earlier than the excretion of smaller particles, whereas large particle excretion appeared comparatively delayed in the banteng. 11

Comparisons within species are documented in Table 3. In three models (banteng, high intake: GIT and RR, and pygmy hippo, low intake: FRST), a significant influence of body mass was found. As the number of contrasts is to be restricted at n-1=2 comparisons for n=3 levels of the repeated measurements (fluid, 2mm particles, 10mm particles), we compared small vs. large particles (the main topic of our investigation), as well as fluids vs. small particles, as these two marker systems are most commonly used in passage trials (cf.

Clauss et al. 2007a). Comparing fluids vs. small particles, MRT yielded significant differences within both species at the high intake level, both in the gastro-intestinal tract

(GIT), and in the forestomach (FRST) of the hippo and the reticulorumen (RR) of the banteng. This is also true for the pygmy hippos at the low intake level; for the three banteng measurements, a test could not be performed at the low intake level because of too few degrees of freedom. Comparing large vs. small particles, MRT in both the GIT and the FRST revealed significant differences in pygmy hippos at the high intake level, with large particles being excreted faster than the small ones. At the low intake level, we observed a comparable excretion pattern in the hippos, however the difference was not significant. In the banteng at the high intake level, the large particles were excreted significantly slower than the small particles.

Mean selectivity factors are displayed in Table 4. The selectivity factor for 10 mm particles as compared to 2 mm particles (GIT and RR/FRST) was significantly lower in hippos compared to banteng at both intake levels. The selectivity factors for 2 mm particles compared to fluids (GIT) were not associated with relative dry matter intake in pygmy hippos, in contrast to banteng (Fig. 3 legend). Similar patterns were noted for the comparison of 10 mm particles versus fluids (Fig. 3 legend). The selectivity factor for 10 mm particles compared to 2 mm particles (GIT) was also associated with relative DMI in banteng but not in hippos (Fig. 3). 12

Discussion

The limitations of passage studies have to be considered when interpreting the results.

In this study, calculating MRTfluidRR and MRTfluidFRST yielded sometimes no difference to the MRTfluidGIT, and fluid marker concentration needed to be corrected additionally to the background level substraction, for calculating fluid and particle MRT in the RR/FRST. Probably the number of samples for assessing the background level of the marker chemicals in the feces was too low and should receive more attention in the future.

In this study, we used hay particle markers, which were administered orally. The forage basis used for mordant marking – the hays – can have an effect on the retention times measured. For example, Cherney et al. (1991) found that the difference in MRT between 1 mm and 37 mm particles was bigger for stems (12h) than for leave blades (3 h) (Table 5).

Therefore, the best approach is to use markers from the same hay batch for all animals that are to be compared, as in this study. However, such sources of variation make comparisons of different studies to a certain extent problematic.

When investigating the effect of particle size on the retention of particles in the GIT, one experimental setup is to circumvent primary dental mastication of the particles, and insert the markers directly into the forestomach, by the use of a rumen or forestomach cannula.

Such an approach is common practice in domestic ruminants or camelids (e. g. Lechner-Doll et al. 1990), and forestomach cannula have also been used in forestomach-fermenting sloths

(Montgomery and Sunquist 1978). However, Wylie et al. (2000) found that masticated hay particles (obtained by an oesophagal cannula) inserted into ruminal digesta via a rumen cannula are excreted significantly slower than normally ingested hay or faecal small particles, suggesting that particles inserted via cannula may be positioned outside the normal intraruminal flow paths. Due to this effect, and because invasive surgery (cannulation of experimental animals) was not an option for this study, the particle markers were fed to the 13 animals, and hence could have been subjected to ingestive mastication. Although both the forestomach contents of pygmy hippos and the rumen contents of cattle consistently contain particles that are larger than the 10 mm large particle marker used in this study (M. Clauss, pers. obs.), it is likely that a certain proportion of the large particle marker was reduced in particle size during ingestive mastication. Therefore, it is to be expected that any difference between the two particle markers should be less distinct in this study as compared to what would have been measured had the markers been introduced directly into the forestomach.

These considerations are supported when our results for the banteng are compared with those obtained for domestic ruminants and camelids where the ingesta markers were introduced into the rumen or forestomach (Table 5).

For ruminants, the effect of ingestive mastication (and rumination) on the measurements of particle retention might vary with intake level. With decreasing intake mean fecal particle size decreases in cattle (Shaver et al. 1988) and sheep (Mudgal et al. 1982).

Correspondingly, Ulyatt et al. (1984) and Aitchison et al. (1986) found that in sheep chewing and rumination per unit feed increased at low intakes. The fact that the two banteng that ingested the lowest relative amount of DM (animals 11 and 12) also had the least distinct separation of retention parameters for the different particle sizes (small and large particles moved in parallel in these two animals) could be an indication that at low intake levels, differences in retention of different particle sizes might be more difficult to demonstrate or actually be negligible, due to the high rate of particle size reduction during ingestive mastication and rumination.

Despite these limitations, the results of this study support the hypothesis of a functional difference in the selective particle retention mechanism in the digestive tract of ruminating and non-ruminating foregut fermenters. In accordance with prediction of Clauss

(2004) and preliminary observations of Clauss et al. (2004a), the larger particles (10 mm) 14 were either excreted faster than (HI), or at the same time as (LI) the smaller particles (2 mm) in pygmy hippos. In contrast, the larger particles were excreted slower than the smaller particles in the ruminating banteng. For an assessment of the biological relevance of a 1 to 13 hours difference in small (2 mm) and large particle (10 mm) excretion, more information about differences in fermentation patterns between particles of different sizes would be required. Because particle size in general is one of the important factors of digestion kinetics

(see literature collation in Clauss and Hummel 2005), biological relevance of these findings can, however, be suspected.

The selectivity factors (2 mm particles/ fluid) in the total GIT of the pygmy hippos in this study (mean ± SD HI: 2.4 ± 0.3, LI: 2.9 ± 0.7) were similar to the values found for pygmy hippos feeding on grass (2.9 ± 0.2) by Clauss et al. (2004a). In comparison to that study, the marker excretion pattern showed the same distinct peaks for fluids and particles.

When compared to the banteng, the hippo had markedly longer fluid and small particle retention times in the GIT. It has been speculated that these particularly long retention times compensate for the less efficient particle size reduction in hippos (Foose 1982; Clauss et al.

2004a), but more comparative research is needed to quantify the according relationships.

Nonetheless, the direct comparison of pygmy hippos and bantengs adds another example for the general observation that among herbivores, particle mean retention time does not necessarily increase with increasing body mass (Clauss et al. 2007a). Compared to the bantengs, the retention time in pygmy hippos seemed to be more responsive to changes in food intake (Table 2). In hippos, the distinct acceleration of ingesta passage due to increased intake would limit the additional energy gained from eating more forage; this physiological characteristic of hippos, first stated by Clauss et al. (2007b), can explain the generally low food intake in hippos observed in this and other studies (Schwarm et al. 2006). 15

The results of the retention measurements in the banteng, which is considered a grazing ruminant (Clauss et al. 2006a; Hofmann et al. 2008), fit the pattern described earlier for wild ruminant species of different feeding types. The selectivity factor (2 mm particles/fluid) in the total GIT of the banteng was, with 2.2 ± 0.2 (HI) and 2.3 ± 0 (LI) within the range of 1.5 - 2.3 found for grazing ruminants (compared to 1.2 - 1.3 for the browsing and 1.4 - 1.6 for the intermediate feeding type) by Hummel et al. (2005). In the reticulorumen (RR) of the banteng, this selectivity factor was with 3.0 ± 0.1 (HI) and 3.0 ±

0.3 (LI) within the range of 1.9 - 3.8 described for grazing ruminants by Clauss and Lechner-

Doll (2001) and Hummel et al. (2005), in comparison to 1.8 - 2.2 for intermediate and 1.4 -

1.8 for browsing ruminants. When average data for MRTfluidRR and MRTparticleRR were plotted for each ruminant species for which data collected by comparative methods is available, the banteng measured in this study have even longer MRTfluidRR and longer

MRTparticleRR than cattle but still show the same pattern as other grazing species (Fig. 4).

Thus, the banteng is a another good example of a ruminant with a distinctive separation of fluid and particle passage (Clauss et al. 2006b). Since a fast fluid passage also occurs in other grazing ruminants such as water buffalo (Bubalus bubalis) (Bartocci et al. 1997), American bison (Bison bison) (Towne et al. 1988), or muskoxen (Ovibos moschatus) (Barboza et al.

2006) (all trials in which no comparable particle markers were used), the strategy of high fluid throughput (relative to particles) appears to be common in many grazing ruminants. A potential adaptive function could be that a constant supply of a low viscosity fluid phase is the prerogative for the physical mechanisms of flotation and sedimentation that result in the stratification of RR contents and its selective particle retention typical for large grazing species; this is in accord with the observation that grazers have larger omasa, a major function of which is water-reabsorption distal to the RR (Clauss et al. 2006b). 16

An interesting question arising from the results of this study is by which mechanism the different selective particle retentions are achieved in the different foregut fermenters. In ruminants, the interplay of particle density, particle size, and the anatomical outflow or regurgitation sites appears well established (Lechner-Doll et al. 1991). Applying the same principle to other forestomach designs, Clauss (2004) and Clauss et al. (2004a) speculated that the position of the outflow orifices in the forestomach of sloths and in hippopotamuses might favour a longer retention of the finer ingesta particles that also have a higher density.

In particular, the two blindsacs of the hippopotamus forestomach, with their ventrally- oriented dead ends, could act as ‘sedimentation traps’. This speculation is supported by the incidental finding of varying amounts of sand trapped in the blindsacs of captive pygmy hippopotamuses (M. Clauss, pers. obs. on three dissected specimens). More detailed investigations on the particle size and density distribution within the hippopotamus forestomach need to be performed to corroborate these speculations. Alternatively, it could be speculated that the ingesta in the hippopotamus forestomach is usually packed too densely, and is not oversaturated with water, so that a separation of particles according to their buoyancy characteristics (as in ruminants) simply cannot occur to a large extent.

To our knowlege, this is the first study that quantifies differences in particle retention mechanisms between functionally different foregut fermenters. The results point towards a particularly effective use of forage material per unit time in ruminants. It would be interesting to compare the results of this study to other nonruminant foregut-fermenting species by applying the same experimental set-up.

Acknowledgements

This study was supported by grants from the Freie Universität Berlin (NaFöG) to AS and the German Science Foundation (DFG) to SO (OR 86/1-1). We thank Ragnar Kühne, 17

Timm Spretke, Uwe Fritzmann, Kurt Goedicke, Conny Hofmann and their colleagues from the Zoological Gardens of Berlin and Halle, Germany, Heidrun Barleben and Doreen Pick from the IZW Berlin, and Olaf Dibowski, Stephan Kohlmüller and their colleagues from the

Landeslabor Brandenburg, Germany, for their engaged support. Thanks also to the referees, whose valuable comments greatly improved the manuscript.

Literature

Aitchison, E., Gill, M., Dhanoa, M., Osbourn, D., 1986. The effect of digestibility and forage species on the removal of digesta from the rumen and the voluntary intake of hay by sheep. British Journal of Nutrition 56, 463-476. Barboza, P.S., Peltier, T.C., Forster, R.J., 2006. Ruminal fermentation and fill change with season in an Arctic grazer: responses to hyperphagia and hypophagia in muskoxen (Ovibos moschatus). Physiological and Biochemical Zoology 79, 497-513. Barker, S., Brown, G.D., Calaby, J.H., 1963. Food regurgitation in the macropodidae. Australian Journal of Science 25, 430-432. Bartocci, S., Amici, A., Verna, M., Terramoccia, S., Martillotti, F., 1997. Solid and fluid passage rate in buffalo, cattle and sheep fed diets with different forage to concentrate ratios. Livestock Production Science 52, 201-208. Björnhag, G., Sperber, I., Holtenius, K., 1984. A separation mechanism in the large intestine of equines. Canadian Journal of Animal Science 64 (Suppl.), 89-90. Björnhag, G., 1987. Comparative aspects of digestion in the hindgut of mammals. The colonic separation mechanism. Deutsche Tierärztliche Wochenschrift 94, 33-36. Björnhag, G., 1989. Sufficient fermentation and rapid passage of digesta. A problem of adaptation in the hindgut. Acta Veterinaria Scandinavia 86 (Suppl.), 204-211. Bruining, M., Bosch, M.W., 1992. Ruminal passage rate as affected by CrNDF particle size. Animal Feed Science and Technology 37, 193-200. Calaby, J.H., 1958. Studies on marsupial nutrition. II. The rate of passage of food residues and digestibility of crude fibre and protein by the quokka. Australian Journal of Biological Sciences 11, 571-580. Cherney, D.J.R., Mertens, D.R., Moore, J.E., 1991. Fluid and particulate retention times in sheep as influenced by intake level and forage morphological composition. Journal of Animal Science 69, 413-422. Clauss, M., Lechner-Doll, M., 2001. Differences in selective reticulo-ruminal particle retention as a key factor in ruminant diversification. Oecologia 129, 321-327. Clauss, M., Lechner-Doll, M., Streich, W.J., 2002. Faecal particle size distribution in captive wild ruminants: an approach to the browser/grazer-dichotomy from the other end. Oecologia 131, 343-349. Clauss, M., 2004. The potential interplay of posture, digestive anatomy, ingesta density and gravity in mammalian herbivores, or why sloths do not rest hanging upside down. Mammal Review 34, 241-245. Clauss, M., Schwarm, A., Ortmann, S., Alber, D., Flach, E.J., Kühne, R., Hummel, J., Streich, W.J., Hofer, H., 2004a. Intake, ingesta retention, particle size distribution and 18

digestibility in the hippopotamidae. Comparative Biochemistry and Physiology A 139, 449-459. Clauss, M., Schwarm, A., Ortmann, S., Thierer, D., Hildebrandt, T., Hohmann, G., Flach, E.J., Kühne, R., Hofer, H., 2004b. Comparative analysis of physical and chemical characteristics of faeces from free-ranging and captive common hippopotami (Hippopotamus amphibius). Advances in Ethology 38, 26. Clauss, M., Hummel, J., 2005. The digestive performance of mammalian herbivores: why big may not be that much better. Mammal Review 35, 174-187. Clauss, M., Hofmann, R.R., Hummel, J., Adamczewski, J., Nygren, K., Pitra, C., Reese, S., 2006a. The macroscopic anatomy of the omasum of free-ranging moose (Alces alces) and muskoxen (Ovibos moschatus) and a comparison of the omasal laminal surface area in 34 ruminant species. Journal of Zoology London 270, 346-358. Clauss, M., Hummel, J., Streich, W.J., 2006b. The dissociation of the fluid and particle phase in the forestomach as a physiological characteristic of large grazing ruminants: an evaluation of available, comparable ruminant passage data. European Journal of Wildlife Research 52, 88-98. Clauss, M., Schwarm, A., Ortmann, S., Streich, W.J., Hummel, J., 2007a. A case of non- scaling in mammalian physiology? Body size, digestive capacity, food intake, and ingesta passage in mammalian herbivores. Comparative Biochemistry and Physiology A 148, 249-265. Clauss, M., Streich, W.J., Schwarm, A., Ortmann, S., Hummel, J., 2007b. The relationship of food intake and ingesta passage predicts feeding ecology in two different megaherbivore groups. Oikos 116, 209-216. Cork, S.J., Hume, I.D., Faicney, G.J., 1999. Digestive strategies of nonruminant herbivores: the role of the hindgut. In: Jung, H.J.G., Fahey, G.C. (Eds.), Nutritional ecology of herbivores, American Society of Animal Science, Savoy, IL, pp. 210-260 Dellow, D.W., 1982. Studies on the nutrition of macropodine marsupials. III. The flow of digesta through the stomach and intestine of macropodines and sheep. Australian Journal of Zoology 30, 399-406. Flores-Miyamoto, K., Clauss, M., Ortmann, S., Sainsbury, A.W., 2005. The nutrition of captive lowland anoa (Bubalus depressicornis): a study on ingesta passage, intake, digestibility, and a diet survey. Zoo Biology 24, 125-134. Foley, W.J., von Engelhardt, W., Chalres-Dominique, P., 1995. The passage of digesta, particle size, and in vitro fermentation rate in the three-toed sloth (Bradypus tridactylus). Journal of Zoology (London) 236, 681-696. Foose, T.J., 1982. Trophic strategies of ruminant versus nonruminant ungulates. PhD, University of Chicago, Chicago. Forbes, D.K., Tribe, D.E., 1970. The utilization of roughages by sheep and kangaroos. Australian Journal of Zoology 18, 247-256. Grenet, E., Martin-Rosset, W., Chenost, M., 1984. Compared size and structure of plant particles in the horse and sheep feces. Canadian Journal of Animal Science 64 (Suppl.), hoh345-346. Grovum, W.L., Williams, V.J., 1973. Rate of passage of digesta in sheep: 4. Passage of marker through the alimentary tract and the biological relevance of rate-constants derived from the changes in concentration of marker in faeces. British Journal of Nutrition 30, 313-329. Heller, R., Cercasov, V., von Engelhardt, W., 1986. Retention of fluid and particles in the digestive tract of the llama (Lama guanacoe f. glama). Comparative Biochemistry and Physiology A 83, 687-691. 19

Hendrichs, H., 1965. Vergleichende Untersuchung des Wiederkauverhaltens. Biologisches Zentralblatt 84, 681-751. Hofmann, R.R., Streich, W.J., Fickel, J., Hummel, J., Clauss, M., 2008. Convergent evolution in feeding types: salivary gland mass differences in wild ruminant species. Journal of Morphology 269, 240-257. Home, E., 1814. Lectures on comparative anatomy; in which are explained the preparations in the Hunterian collection. W. Blumer and Co, London. Hume, I.D., Sakaguchi, E., 1991. Patterns of digesta flow and digestion in foregut and hindgut fermenters. In: Tsuda, T., Saaski, Y., Kawashima, R. (Eds.), Physiological aspects of digestion and metabolism in ruminants, Academic Press, San Diego, pp. 427- 451 Hume, I.D., 1999. Marsupial nutrition. Cambridge University Press, Cambridge. Hummel, J., Clauss, M., Zimmermann, W., Johanson, K., Nørgaard, C., Pfeffer, E., 2005. Fluid and particle retention in captive okapi (Okapia johnstoni). Comparative Biochemistry and Physiology A 140, 436-444. Hummel, J., Steuer, P., Südekum, K.H., Hammer, S., Hammer, C., Streich, W.J., Clauss, M., 2007. Fluid and particle retention in the digestive tract of the addax antelope (Addax nasomaculatus) – adaptations of a grazing desert ruminant. Comparative Biochemistry and Physiology A, doi: 10.1016/j.cbpa.2007.11.001. Hungate, R.E., 1966. The rumen and its microbes. Academic Press, London. Kaske, M., von Engelhardt, W., 1990. The effect of size and density on mean retention time of particles in the gastrointestinal tract of sheep. British Journal of Nutrition 63, 457- 465. Kaske, M., Groth, A., 1997. Changes in factors affecting the rate of digesta passage through pregnancy and lactation in sheep fed on hay. Reproduction, Nutrition, Development 37, 573-588. Langer, P., 1988. The mammalian herbivore stomach. Gustav Fischer Verlag, Stuttgart/New York. Lechner-Doll, M., Rutagwenda, T., Schwartz, H.J., Schultka, W., von Engelhardt, W., 1990. Seasonal changes of ingesta mean retention time and forestomach fluid volume in indigenous camels, cattle, sheep and goats grazing in a thornbush savanna pasture in Kenya. Journal of Agricultural Science (Cambridge) 115, 409-420. Lechner-Doll, M., Kaske, M., von Engelhardt, W., 1991. Factors affecting the mean retention time of particles in the forestomach of ruminants and camelids. In: Tsuda, T., Sasaki, Y., Kawashima, R. (Eds.), Physiological aspects of digestion and metabolism in ruminants, Academic Press, San Diego, pp. 455-482 Lirette, A., Milligan, L.P., 1989. A quantitative model of reticulo-rumen particle degradation and passage. British Journal of Nutrition 62, 465-479. Mambrini, M., Peyraud, J.L., 1997. Retention time of feed particles and liquids in the stomachs and intestines of dairy cows. Direct measurement and calculation based on fecal collection. Reproduction, Nutrition, Development 37, 427-442. Moir, R.J., Somers, M., Waring, H., 1956. Studies on marsupial nutrition. I. Ruminant-like digestion in a herbivorous marsupial. Australian Journal of Biological Sciences 9, 293- 304. Mollison, B.C., 1960. Food regurgitation in Bennett’s wallaby and the scrub wallaby. CSIRO Wildlife Research 5, 87-88. Montgomery, G.G., Sunquist, M.E., 1978. Habitat selection and use by two-toed and three- toed sloths. In: Montgomery, G.G. (Ed.) The ecology of arboreal foliovores, Smithsonian Institution Press, Washington DC, pp. 329-359 20

Mudgal, V.D., Dixon, R.M., Kennedy, P.M., Milligan, L.P., 1982. Effect of two intake levels on retention times of liquid, particle and microbial markers in the rumen of sheep. Journal of Animal Science 54, 1051-1055. Okamoto, M., 1997. Comparison of particle size in the feces of various herbivores. Journal of the Rakuno Gakuen University (Natural Sciences) 22, 151-153. Owen, R., 1834. Notes on the anatomy of a new species of kangaroo. Proceedings of the Zoological Society of London Part II, 152. Prigge, E., Stuthers, B., Jacquemet, N., 1990. Influence of forage diets on ruminal particle size, passage of digesta, feed intake and digestibility by steers. Journal of Animal Science 68, 4352-4360. Schwarm, A., Ortmann, S., Hofer, H., Streich, W.J., Flach, E.J., Kühne, R., Hummel, J., Castell, J.C., Clauss, M., 2006. Digestion studies in captive Hippopomatidae: a group of large ungulates with an unusually low metabolic rate. Journal of Animal Physiology and Animal Nutrition 90, 300-308. Shaver, R., Nytes, A., Satter, L., Jorgensen, N., 1988. Influence of feed intake, forage physical form, and forage fiber content on particle size of masticated forage, ruminal digesta, and feces of dairy cows. Journal of Dairy Science 71, 1566-1572. Stevens, C.E., Hume, I.D., 1998. Contributions of microbes in vertebrate gastrointestinal tract to production and conservation of nutrients. Physiological Reviews 78, 393-427. Thielemans, M.-F., Francois, E., Bodart, C., Thewis, A., 1978. Mesure du transit gastrointestinal chez le porc à l´aide des radiolanthanides. Comparaison avec le mouton. Ann. Biol. anim. Bioch. Biophys. 18, 237-247. Towne, G., Nagaraja, T.G., Cochran, R.C., Harmon, D.L., Owensby, C.E., Kaufman, D.W., 1988. Comparisons of ruminal fermentation characteristics and microbial populations in bison and cattle. Applied and Environmental Microbiology 54, 2510-2514. Udén, P., Colucci, P.E., Van Soest, P.J., 1980. Investigation of chromium, cerium and cobalt as markers in digesta rates of passage studies. Journal of the Science of Food and Agriculture 31, 625-632. Udén, P., Van Soest, P.J., 1982. The determination of digesta particle size in some herbivores. Animal Feed Science and Technology 7, 35-44. Udén, P., Robinson, P., Wiseman, J., 2005. Use of detergent system terminology and criteria for submission of manuscripts on new, or revised, analytical methods as well as descriptive information on feed analysis and/or variability. Animal Feed Science and Technology 118, 181-186. Ulyatt, M.J., Waghorn, G.C., John, A., Reid, C.S.W., Monro, J., 1984. Effect of intake and feeding frequency on feeding-behaviour and quantitative aspects of digestion in sheep fed chaffed lucerne hay. Journal of Agricultural Science (Cambridge) 102, 645-657. Wylie, M.J., Ellis, W.C., Matis, J.H., Bailey, E.M., James, W.D., Beever, D.E., 2000. The flow of forage particles and solutes through segments of the digestive tracts of cattle. British Journal of Nutrition 83, 295-306.

Table 1. Details of the pygmy hippos (Hexaprotodon liberiensis) and banteng (Bos javanicus) studied at the Zoological Gardens of Berlin (ZGB) and Halle (ZGH)

Species Animal Born Sex BM Facility (kg) 1 1985 male 248 ZGB 2 1983 female 225 ZGB 3 1997 female 238 ZGB Pygmy hippo 4 1998 female 203 ZGH 5 1976 female 202 ZGH 6 2000 male 196 ZGH 7 2002 male 550a ZGB 8 2004 female 220a ZGB 9 2004 male 200a ZGB Banteng 10 1996 female 700a ZGB 11 2001 female 600a ZGB 12 1997 female 650a ZGB BM=body mass a Estimated. 22

Table 2. Dry matter intake (DMI, kg) and mean retention time (MRT, h) of fluid, 2 mm and

10 mm particles in the gastrointestinal tract (GIT) and forestomach/reticulorumen (FRST/RR) of pygmy hippos (Hexaprotodon liberiensis) and banteng (Bos javanicus) at the high (HI) and the low intake level (LI) and the P-values for interspecies comparison (t-test)

Species Animal Intake level DMI DMI MRT fluid MRT MRT 2 mm particles 10 mm particles GIT FRST/RR GIT FRST/RR GIT FRST/RR (kg) (g/kg0.75) (h) (h) (h) (h) (h) (h) 1 HI 1.1 18 31 10 79 59 79 58 1 LI 0.9 14 49 16 114 73 111 72 2 HI 2.5 43 25 11 75 62 73 60 2 LI 0.8 14 29 19 110 102 97 89 3 HI 3.0 50 28 18 66 55 63 53 Pygmy hippo 3 LI 2.4 40 34 21 77 64 77 64 4 HI 2.0 37 34 17 71 54 66 50 4 LI 1.0 19 36 31 124 119 124 119 5 HI 1.1 21 43 28 104 89 103 88 5 LI 0.4 7 49 29 120 100 114 94 6 HI 1.9 36 28 13 59 45 56 42 7 HI 6.8 60 24 16 58 52 67 61 7 LI 4.9 43 22 13 50 42 63 55 8 HI 4.5 79 23 16 54 49 59 54 8 LI 2.3 40 23 15 54 47 57 50 Banteng 9 HI 3.9 73 23 17 53 48 58 53 9 LI 3.0 56 22 16 50 44 54 48 10 HI 5.1 37 26 14 53 42 56 45 11 HI 3.4 28 30 16 61 48 61 48 12 HI 3.3 26 26 14 55 43 56 44 mean ± SD Pygmy hippo n=6 HI 1.9 ± 0.8 34 ± 13 32 ± 6 16 ± 7 76 ± 16 61 ± 15 73 ± 17 59 ± 16 Banteng n=6 HI 4.5 ± 1.3 51 ± 23 25 ± 3 16 ± 1 56 ± 3 47 ± 4 60 ± 4 51 ± 6 P 0.055 0.817 0.011 0.056 0.075 0.297 mean ± SD Pygmy hippo n=5 LI 1.1 ± 0.8 19 ± 12 39 ± 9 23 ± 6 109 ± 19 92 ± 23 105 ± 18 88 ± 21 Banteng n=3 LI 3.4 ± 1.3 47 ± 9 22 ± 1 15 ± 2 51 ± 2 44 ± 3 58 ± 5 51 ± 4 P 0.014 0.040 0.002 0.009 0.006 0.029

Table 3. Results of the Repeated Measurements ANOVA models (intraspecies comparison):

P-values for body mass and for linear contrasts (which refer to the 2 mm results as the reference category) of fluid, small (2 mm) and large particle (10 mm) mean retention times

(MRT) in the gastrointestinal tract (GIT) or forestomach (FRST)/reticulorumen (RR) of pygmy hippos (Hexaprotodon liberiensis) and banteng (Bos javanicus) at the high (HI) and of pygmy hippos at the low intake level (LI)

MRT Pygmy hippo Banteng Pygmy hippo HI HI LI n=6 n=6 n=5HI GIT FRST GIT FRST GIT n=6FRST BM 0.232 0.143 0.048 0.041 0.955 <0.001 Fluid vs. 2 mm <0.001 <0.001 <0.001 <0.001 0.002 <0.003 2 mm vs. 10 mm 0.022 0.006 0.034 0.034 0.143 0.186 particles particles

Table 4. Selectivity factor (SF) (mean ± SD) for 2 mm particles compared to fluid, 10 mm particles compared to fluid and 10 mm particles compared to 2 mm particles in the gastrointestinal tract (GIT) and the forestomach (FRST)/ reticulorumen (RR) of pygmy hippos (Hexaprotodon liberiensis) and banteng (Bos javanicus) at the high (HI) and the low intake level (LI) and the P-values for interspecies comparison (t-test)

Species Diet n SF 2 mm particles/ fluid SF 10 mm particles/ fluid SF 10 mm/ 2 mm particles GIT FRST/RR GIT FRST/RR GIT FRST/RR Pygmy hippo HI 6 2.42 ± 0.34 4.05 ± 0.92 2.34 ± 0.36 3.86 ± 0.73 0.97 ± 0.03 0.96 ± 0.03 Banteng HI 6 2.21 ± 0.17 3.03 ± 0.14 2.37 ± 0.30 3.28 ± 0.29 1.07 ± 0.06 1.08 ± 0.06 P 0.200 0.115 0.889 0.297 0.002 0.002 Pygmy hippo LI 5 2.86 ± 0.71 4.16 ± 0.92 2.73 ± 0.61 3.97 ± 0.76 0.96 ± 0.05 0.96 ± 0.05 Banteng LI 3 2.30 ± 0.04 3.04 ± 0.25 2.60 ± 0.23 3.52 ± 0.64 1.13 ± 0.11 1.15 ± 0.13 P 0.154 0.093 0.683 0.430 0.022 0.025

Table 5. Mean retention time (MRT, h) of fluid and different sized (hay-)particles in the gastrointestinal tract (GIT) and forestomach (FRST) of ruminants and pygmy hippopotamuses from the literature

Species Food Fluid Particles (mm) Administration Source

< 0.3 1 <= 2 3 5 10 20 37

GIT FRST GIT GIT FRST GIT FRST FRST FRST GIT FRST FRST GIT

Sheep (O. African pasture 19 38 55 rumen cannula 1 ammon) Cattle (B. African pasture 13 35 56 rumen cannula 1 taurus) Goat (Capra African pasture 14 30 47 rumen cannula 1 sp.) Camel (C. African pasture 11 27 53 rumen cannula 1 dromedarius) Sheep (O. orally (plastic hay 25 50 78 2 ammon) particles) Sheep (O. mixed hay 38 49 61 orally (stems) 3 ammon) Sheep (O. mixed hay 38 45 48 orally (leaf blades) 3 ammon) Cattle (B. rumen cannula grass hay 22 21 71 4 taurus) (nylon particles) Cattle (B. rumen cannula grass hay 55 69 5 taurus) (content mixed) Cattle (B. grass silage, 12 24 48 rumen cannula 6 taurus) concentrates Pygmy hippo fresh grass 31 19 89 77 74 63 orally 7 (H. liberiensis) Pygmy hippo fresh grass 35 16 91 72 88 69 orally this study (H. liberiensis) Banteng (B. fresh grass 24 11 54 43 59 48 orally this study javanicus) 1 (Lechner-Doll et al. 1990); 2 (Kaske and von Engelhardt 1990); 3 (Cherney et al. 1991); 4 (Prigge et al. 1990); 5 (Lirette and Milligan 1989); 6 (Bruining and Bosch 1992); 7 (Clauss et al. 2004a) 26

Figure Legends

Figure 1 Excretion pattern in a banteng (Bos javanicus) for fluids (marked with cobalt[Co]-

EDTA), small particles (2 mm; mordanted with chromium [Cr]), and large particles (10 mm; mordanted with cerium [Ce]). Marker concentrations adjusted to grant visual comparability.

Average night-time faecal samples are displayed by rhombic data points.

Figure 2 Excretion pattern in a pygmy hippopotamus (Hexaprotodon liberiensis) for fluids

(marked with cobalt[Co]-EDTA), small particles (2 mm; mordanted with chromium [Cr]), and large particles (10 mm; mordanted with cerium [Ce]). Marker concentrations adjusted to grant visual comparability. Average night-time faecal samples are displayed by rhombic data points.

Figure 3 “Selectivity factors” (SF; quotients of mean retention times of different ingesta phases in the GIT) for large particles (10 mm)/small particles (2 mm) in pygmy hippos

(Hexaprotodon liberiensis) and banteng (Bos javanicus) in relation to the level of dry matter intake (DMI, g/kg0.75/d)

Spearman correlation coefficients (SCC) of different SF (GIT) and relative DMI:

SF 2 mm particles/fluids in hippo: SCC=-0.36, P=0.277, in banteng: SCC=0.74, P=0.023,

SF 10 mm particles/fluids in hippo: SCC=-0.39, P=0.239, in banteng: SCC=0.72, P=0.029,

SF 10 mm/2 mm particles in hippo: SCC=0.07, P=0.833, in banteng: SCC=0.76, P=0.018.

These results are to be considered as explorative, because the same animals are included twice (with both high and low intake diet).

Figure 4 Relation of mean retention time of fluid and of chromium-mordanted hay particles

(≤2 mm) in the reticulorumen (MRTfluidRR and MRTparticleRR) for different feeding types. 27

Data collection from Clauss et al. (2006b) plus Flores-Miyamoto et al. (2005), Hummel et al.

(2008), and this study (means of all measurements). Browsers: AA Alces alces, CC

Capreolus capreolus, GC Giraffa camelopardalis, OJ Okapia johnstoni; Intermediate feeders: Bd Bubalus depressicornis, Ce Cervus elaphus, Chd Capra hircus f. dom., Ci Capra ibex; Grazers: an Addax nasomaculatus, bj Bos javanicus, btd Bos taurus f. dom., oad Ovis aries f. dom., oam Ovis ammon musimon. 28

700 Co mg/2kg DM Cr mg/kg DM 600 Ce mg/0.4kg DM

500

400 concentration

300 marker

200 Faecal 100

0 0 48 96 144 192 Time after marker application (h)

Figure 1

1200 Co mg/2kg DM Cr mg/kg DM 1000 Ce mg/0.5kg DM 800

concentration 600

marker 400

200 Faecal

0 0 48 96 144 192 Time after marker application (h)

Figure 2 29

1.4 Hippo Banteng

mm 1.2 mm/2 10 1.0 GIT SF

0.8 0 20 40 60 80 DMI (g/kg0.75/d)

Figure 3

Figure 4