University of Zurich Posted at the Zurich Open Repository and Archive, University of Zurich

Home , Tylopoda

Clauss, M; Hofmann, R R; Streich, W J; Fickel, J; Hummel, J (2010). Convergence in the macroscopic anatomy of the reticulum in wild ruminant species of different feeding types and a new resulting hypothesis on reticular function. Journal of Zoology, 281(1):26-38. Postprint available at: http://www.zora.uzh.ch

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

Originally published at: Journal of Zoology 2010, 281(1):26-38. Winterthurerstr. 190 CH-8057 Zurich http://www.zora.uzh.ch

Year: 2010

Convergence in the macroscopic anatomy of the reticulum in wild ruminant species of different feeding types and a new resulting hypothesis on reticular function

Clauss, M; Hofmann, R R; Streich, W J; Fickel, J; Hummel, J

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

Originally published at: Journal of Zoology 2010, 281(1):26-38. 1

1 Convergence in the macroscopic anatomy of the reticulum in wild ruminant species of

2 different feeding types and a new resulting hypothesis on reticular function

4 Marcus Clauss1*, Reinold R. Hofmann2, W. Jürgen Streich3, Jörns Fickel3, Jürgen Hummel4

6 1Clinic for Zoo Animals, Exotic Pets and Wildlife, Vetsuisse Faculty, University of Zurich,

7 Winterthurerstr. 260, 8057 Zurich, Switzerland

8 2formerly at the Leibniz Institute for Zoo and Wildlife Research (IZW), Berlin, Germany

9 (retired); present address: Trompeterhaus, 15837 Baruth/Mark, Germany

10 3Leibniz Institute for Zoo and Wildlife Research (IZW), Alfred-Kowalke-Str. 17, 10315

11 Berlin, Germany

12 4Institute of Animal Science, Animal Nutrition Group, Endenicher Allee 15, University of

13 Bonn, 53115 Bonn, Germany

18 *correspondence to: Marcus Clauss, Clinic for Zoo Animals, Exotic Pets and Wildlife,

19 Winterthurerstr. 260, 8057 Zurich, Switzerland, phone: ++41 44 635 83 76, fax: ++41 44 635

20 89 01, email: [email protected]

22 Running head: Reticulum anatomy in ruminants

24 2

24 Summary

25 The reticulum is the second part of the ruminant forestomach, located between the rumen and

26 the omasum and characterised by a honeycomb-like internal mucosa. With its fluid contents,

27 it plays a decisive role in particle separation. Differences among species have been linked to

28 feeding style. We investigated whether reticulum size (absolute and in relation to rumen size)

29 and size of the crests that form the mucosal honeycomb pattern differ among over 60

30 ruminant species of various body sizes and feeding type, controlling for phylogeny. Linear

31 dimensions generally scaled allometrically, i.e. to body mass0.33. With or without controlling

32 for phylogeny, species that ingest a higher proportion of grass in their natural diet had both

33 significantly larger (higher) rumens and higher reticular mucosa crests, but neither reticulum

34 height nor reticulum width varied with feeding type. The height of the reticular mucosa crests

35 represents a dietary adaptation in ruminants. We suggest that the reticular honeycomb

36 structures do not separate particles by acting as traps (neither for small nor for large particles),

37 but that the structures reduce the lumen of the reticulum during contractions – at varying

38 degrees of completeness in the different feeding types. In browsing species with rumen

39 contents that may be less fluid and more viscous than those of the reticulum, incomplete

40 closure of the lumen may allow the reticulum to retain the fluid necessary for particle

41 separation. In grazing species, whose rumen contents are more stratified with a larger distinct

42 fluid pool, a more complete closure of the reticular lumen due to higher crests may be

43 beneficial as the reticulum can quickly re-fill with fluid rumen contents that contain pre-

44 sorted particles.

45 Key words: rumen, grazer, browser, rumination, forestomach, anatomy, physiology 3

46 Introduction

47 The process of rumination and the complicated forestomach of ruminants have

48 fascinated scientists for centuries (Peyer 1685, Haubner 1837). Apart from fermenting plant

49 material, which the ruminant forestomach shares with the foregut or hindgut of many other

50 herbivores (Stevens & Hume 1998), the forestomachs of ruminants and camelids have a

51 unique sorting function (Schwarm et al. 2008, Schwarm et al. 2009a) which ensures that large

52 ingesta particles are regurgitated and re-masticated (Fritz et al. 2009). This sorting mechanism

53 facilitates a high digestive efficiency (Clauss et al. 2009d) at comparatively high intakes in

54 ruminants, especially when compared to other foregut fermenters (Schwarm et al. 2009b).

55 The second section of the ruminant forestomach, the reticulum, has long been

56 recognised in domestic ruminants as the site of particle sorting (Reid 1985, Mathison et al.

57 1995, Okine et al. 1998). In wild ruminants, empirical evidence also points towards the

58 reticulum as the major site of particle sorting (Clauss et al. 2009a, Clauss et al. 2009b). The

59 bisphasic contractions of the reticulum in domestic ruminants lead to a re-jection of larger,

60 floating particles into the rumen, while the opening of the reticulo-omasal orifice allows the

61 passage of fluids and finer, denser particles; in the second contraction phase, the reticulum

62 contracts completely in cattle so that its lumen disappears, and the then empty reticulum

63 relaxes and re-fills again with contents from the ventral rumen with a high proportion of

64 fluids. An important precondition for the sorting function, apart from the precise timing of

65 contractions and opening of the reticulo-omasal orifice, is the correlation between density and

66 size of reticuloruminal contents – smaller particles are usually more dense, and are hence

67 passed on to the omasum by the reticulum (Sutherland 1988, Beaumont & Deswysen 1991,

68 Lechner-Doll, Kaske & Engelhardt 1991, Allen 1996, Hristov et al. 2003). This function of 4

69 the reticulum has been investigated in numerous studies and is, to date, undisputed (Okine,

70 Mathison & Hardin 1990, Kaske & Midasch 1997).

71 How the anatomy of the reticulum - especially the peculiar shape of its internal mucosa

72 which is arranged in net-like ridges (hence the name “reticulum”) or “honeycomb cells” (Fig.

73 1) – accomplishes its function is unclear. Historically, the reticular crests and honeycomb

74 cells have been suspected to play a role in water storage, to help grind coarse food during

75 contractions, or to be directly involved in the particle separation mechanism – either by acting

76 as traps that hold larger food particles and prevent their passage toward the omasum

77 (Hofmann & Schnorr 1982) or by acting as sedimentation traps that catch small dense

78 particles and direct them towards the omasum during contractions (Grau 1955, Reid 1985).

79 However, empirical data supporting either mechanism is lacking.

80 The height of the reticular crests and the respective depth of the honeycomb structures

81 vary greatly across species (Fig. 1 & 2). Würfel (1908) described respective differences

82 between cattle and sheep. Reticular crests are shallow in okapi (Okapia johnstoni), giraffe

83 (Giraffa camelopardalis) and deer, and higher in chamois (Rupicapra rupicapra), sheep and

84 cattle (Burne 1917, Neuville & Derscheid 1929). Grazing ruminants have in general higher

85 reticular crests than browsers, and more pronounced secondary, tertiary and even quaternary

86 crests (Hofmann 1969, Hofmann 1973, Langer 1988). Appropriate statistical treatments of

87 this observation, however, are lacking. Additionally, Hofmann (1973, 1989) suggested that

88 browsing ruminants have generally larger reticula than grazers – a suggestion that Langer

89 (1988) confirmed statistically using Hofmann’s (1973) data. The suggested functional reason

90 for this latter observation is that the whole reticulo-rumen might be involved in the particle

91 separation in grazers, in which rumen contents are prominently stratified, whereas in browsers 5

92 the sorting mechanism might be spatially limited to the reticulum only, which might therefore

93 need to be of a larger size (Clauss et al. 2009b, Clauss et al. 2009c, Hummel et al. 2009).

94 Here, we report comparative analyses of the reticulum size and the height of the

95 reticular crests from more than 60 ruminant species of different feeding types. We predicted

96 that

97 1. as linear measurements, the height and width of the reticulum, and the

98 height of the reticular crests, would scale allometrically to body mass (BM) with an

99 exponent of 0.33 (based on simple geometry, volumes should scale to BM1.00,

100 areas/surfaces should scale to BM0.67, and linear dimensions to BM0.33);

101 2. species with a higher proportion of grass in their natural diet (grazers)

102 would have a smaller reticulum, both in absolute terms and in relation to rumen size,

103 than similar-sized species with a lower proportion of grass in their natural diet

104 (browsers);

105 3. species with a higher proportion of grass in their natural diet would

106 have higher (taller) reticular crests than similar-sized species with a lower proportion

107 of grass in their natural diet.

108 In interpreting the results, we offer a potentially novel explanation of the functional

109 significance of reticular crests.

110

111 Methods

112 The core of the data originated from the second author and was supplemented with data

113 from other sources (Table 1). Additional animals were obtained over years from from hunts in

114 Germany, Austria, Switzerland, Finland, Canada, Botswana, Namibia and South Africa, and

115 from European zoological gardens; because studies that document and influence of diet on the 6

116 size of reticular crests are lacking, captive animals were also included (but marked in Table

117 1). The external measurements taken included the linear height of the rumen, the linear height

118 of the reticulum (from the cardia to the ventral border of the reticulum), the linear width of the

119 reticulum at its widest point, all measured in exenterated stomachs, and the maximal height of

120 the reticular crests. Body masses were from the original sources but represent estimates in

121 several cases. In the data reported here for the first time, body mass (BM) was always

122 determined either by weighing the whole animal prior to dissection or using the bilateral

123 masseter mass as a substitute measure according to Axmacher and Hofmann (1988).

124 Following Langer (1988), who calculated a ratio dividing reticulum height by the height of

125 the abdominal cavity (thereby effectively removing the influence of body mass), we

126 calculated the height ratio of reticulum:rumen to test whether, in relation to rumen size, the

127 reticulum differed with feeding type. To achieve a normal distribution of residuals, linear

128 measurements and the height ratio were ln-transformed. Because our expectation of an

129 allometric scaling of linear measurements with body mass was confirmed, we also expressed

130 linear measurements relative to BM0.33 (i.e., mm/kg0.33 or cm/kg0.33) for a visualisation of

131 results.

132 As in more recent evaluations of the influence of adaptation to the natural diet in

133 ruminants (Clauss, Kaiser & Hummel 2008), the percentage of grass in the natural diet

134 (%grass) was used to characterize species on a continuous scale. The bulk of the respective

135 data was taken from Van Wieren (1996) and from the data collection that formed the basis of

136 Owen-Smith (1997, data kindly provided by the author), which were supplemented by several

137 other publications (Table 1). Whenever seasonal data was available, the %grass used to

138 characterise a species represents the mean of the values from different seasons. This literature

139 data was collated from a variety of sources and methods, and does not represent the actual 7

140 diet ingested by the individuals measured in this study. Species were categorized as browsers,

141 intermediate feeders, or grazers according to Hofmann (1985, 1988, 1989, 1991, 2000) and

142 Hofmann et al. (1995).

143 As in similar studies (Pérez-Barberìa et al. 2004, Clauss et al. 2006a, 2008, Hofmann et

144 al. 2008, Clauss et al. 2009c), the limitations of our approach need to be stated, such as that

145 different measurements used to characterize a species do not derive from the same individual,

146 and most likely not even from individuals of the same population (in terms of geographical

147 location and time). This often applies to the body mass used (in our case, the estimated

148 masses from Hofmann 1973), and to the proportion of grass in the natural diet which may

149 already have shifted away, in the population from which the animals measured came from,

150 from the average value assumed for the species as a whole. Ideally, morphophysiological

151 measurements and dietary history information should be derived from the same specimens – a

152 postulation that can probably only be achieved for African species in which isotope analysis

153 in bone or teeth allows a quantification of the proportion of grass and browse ingested by a

154 specimen used to measure skeletal parameters (Codron et al. 2008). However, in large-scale

155 comparative analyses as this one, using species-related information from different sources is

156 usually acceptable.

157 Relationships among species were inferred from a phylogenetic tree based on the

158 complete mitochondrial cytochrome b gene. Respective DNA sequences were available from

159 GenBank (http://www.ncbi.nlm.nih.gov) for all ruminant species except for Gazella arabica,

160 G. dorcas, G. leptoceros, G. spekei and Ovis canadensis; these latter were excluded from the

161 phylogenetic analyses. Sequences were aligned using CLUSTALX (Thompson et al. 1997),

162 visually controlled and trimmed to identical lengths (1143 bp). To select the best-fitting

163 nucleotide substitution model for the data, a combination of the software packages PAUP* 8

164 (v.4.b10; Swofford 2002) and MODELTEST (v.3.7; Posada & Crandall 1998) was used.

165 Analysis was based on a hierarchical likelihood ratio test approach implemented in

166 MODELTEST. The model selected was the general time-reversible (GTR) model (Lanave et al.

167 1984, Tavaré 1986) with an allowance both for invariant sites (I) and a gamma (G)

168 distribution shape parameter (α) for among-site rate variation (GTR+I+G) (Rodriguez et al.

169 1990). The nucleotide substitution rate matrix for the GTR+I+G model was likewise

170 calculated using MODELTEST. Parameter values for the model selected were: -lnL =

171 19537.2988, I = 0.4342, and α = 0.8288 (8 gamma rate categories). The phylogenetic

172 reconstruction based on these parameters was then performed using the maximum likelihood

173 (ML) method implemented in TREEPUZZLE (v.5.2; Schmidt et al. 2002). Support for nodes

174 was assessed by a reliability percentage after 10000 quartet puzzling steps; only nodes with

175 more than 50% support were retained. The resulting tree is displayed in Figure 3. The basal

176 polytomy for familial relationships (Bovidae, Cervidae, Giraffidae and Antilocapridae) was

177 resolved assuming a soft polytomy (Purvis & Garland 1993). In order to meet the input

178 requirements for the phylogenetic analysis implemented in the COMPARE 4.6 program

179 (Martins 2004), we resolved the remaining polytomies to full tree dichotomy by introducing

180 extreme short branch lengths (l = 0.000001) at multifurcating nodes. Taxa grouping in the

181 bifurcating process followed the phylogenies proposed by Pitra et al. (2004) for Cervidae and

182 by Fernandez and Vrba (2005) for all other taxa.

183 The subjects of the comparative analyses were individual species, each characterized by

184 its respective measurements as described above. Statistical analyses were performed with and

185 without accounting for phylogeny to test for the validity of a general, functional hypothesis,

186 and then to discriminate between convergent adaptation and adaptation by descent. Data were

187 analysed by correlation and regression analysis. To include phylogenetic information, we 9

188 used the Phylogenetic Generalized Least-Squares approach (Martins & Hansen 1997, Rohlf

189 2001) in which a well-developed standard statistical method was extended to enable the

190 inclusion of interdependencies among species due to the evolutionary process. To test the

191 robustness of the results, the comparative analysis was performed for both a set of

192 phylogenetic trees involving branch lengths (tree 1) and another tree based only on the

193 phylogenetic topology (tree 2). As the results from these two methods had no relevant

194 differences in the results, only the tests using tree 1 are given here. The COMPARE 4.6

195 program (Martins 2004) was used for the phylogenetically controlled calculations. Other

196 statistical calculations were performed with the SPSS 12.0 software (SPSS, Chicago, IL). The

197 significance level was 0.05.

198

199

200 Results

201 Linear measurements were all significantly correlated to body mass, and 0.33 was

202 included in the 95 % confidence interval for the allometric exponent except for reticulum

203 width (Table 2). Rumen height and reticular crest height were both significantly correlated

204 with the percentage of grass in the natural diet, but reticulum height and width were not (Fig.

205 4, Table 2). Correspondingly, the relative rumen and reticular crest height (expressed per

206 BM0.33) showed an increase with %grass (Fig. 5a, d). In contrast, the relative reticulum height

207 (Fig. 5b) and width (not shown) showed no trend when plotted against %grass. The results of

208 the Phylogenetic Generalized Least-Squares approach confirmed these findings (Table 3).

209 The correlation between %grass and the logarithm of the ratio reticulum/rumen height

210 was negative (Fig. 5c: y = 0.65 – 0.001 (%grass); p[%grass]=0.033) with a very low r2 of 10

211 0.07. Using Phylogenetic Generalized Least-Squares, this correlation was even lower (r2 of

212 0.04) and no longer significant (p[%grass]=0.091).

213

214 Discussion

215 The results confirmed that linear measurements of the rumen and reticulum followed

216 expected allometric relationships of body mass0.33; they also indicate strong convergent

217 evolution of the internal muscosa of the reticulum with dietary niche in ruminants. In contrast

218 to the expectations, however, external measurements of the reticulum (‘reticulum size’) do not

219 vary systematically with the natural diet.

220 The finding that rumen size differs with diet (Fig. 4a, Tables 2 and 3) was not a primary

221 aim of this study but corroborates similar findings (Clauss, Kaiser & Hummel 2008). A larger

222 rumen in grazing ruminants is usually explained by the fact that grass requires longer

223 fermentation for optimal digestion (Hummel et al. 2006) and that grazers compensate the

224 intake-limiting effect of longer retention by larger rumen capacities. Interestingly, however,

225 this does not translate into an increased reticulum size in grazers (Fig. 4bc, 5b, Tables 2 and

226 3), and the combination of the two findings leads to the observation that the reticulum : rumen

227 ratio actually decreases with an increasing proportion of grass in the natural diet (Fig. 5c).

228 The effect, however, is small (low r2), and not significant after phylogenetic control. Thus, a

229 functional interpretation (Clauss et al. 2009b) appears unlikely. Reticulum size is more or less

230 constant in ruminants both across body sizes and the feeding types.

231 Investigations on the stratification of rumen contents (measured directly or indirectly)

232 and the particle size in different forestomach regions suggest that although the degree of

233 rumen contents stratification differs among ruminant feeding types, this difference does not

234 imply a corresponding differentiation of the reticular sorting mechanism (Clauss et al. 2001, 11

235 Tschuor & Clauss 2008, Clauss et al. 2009a, 2009b, 2009c). Based on faecal particle size data

236 from captive wild ruminants, Clauss et al. (2002) suggested that differences in rumen contents

237 stratification caused large browsers to have the larger particles measured in their faeces

238 compared to grazers. However, a study on aurochs (Bos primigenus taurus) and giraffe

239 (Giraffa camelopardalis) demonstrated that this difference only occurred in zoo animals but

240 not in the wild, indicating that the teeth of browsers may be less suited for captive diets

241 offered in captivity than are the teeth of grazers (Hummel et al. 2008a). In fact, dental designs

242 of browsers and grazers (Archer & Sanson 2002, Kaiser et al. 2010) and tooth wear of captive

243 browsers and grazers (Kaiser et al. 2009) differ systematically. The similar-sized reticula of

244 browsers and grazers appear to be equally efficient at sorting particles and ensuring a

245 selective retention of the larger, more buoyant particle fraction.

246 The conclusion that ruminants have an effective particle sorting mechanism irrespective

247 of feeding type fits the functional concept of the reticulum outlined in the Introduction, which

248 is largely independent of an assumed functional involvement of the reticular crests and

249 honeycomb structures themselves. If the crests had a crucial function in reticular sorting

250 mechanism, the smaller crests in browsing ruminants (Fig. 4d, 5d, Tables 2 and 3) should

251 reduce the sorting efficiency in this group, but this is not the case. A major prerequisite for the

252 separation mechanism as demonstrated in domestic ruminants (Sutherland 1988, Beaumont &

253 Deswysen 1991, Lechner-Doll, Kaske & Engelhardt 1991, Allen 1996, Hristov et al. 2003) is

254 a specific buoyancy behaviour of the ingested food particles – larger (still undigested un-

255 ruminated) particles float, and smaller (digested and ruminated) particles sediment. Although

256 Clauss et al. (2001) had suggested that in browsing animals this size separation according to

257 buoyancy may not occur, Clauss et al. (2009a,b) demonstrated that floating particles are

258 usually larger than sedimenting particles in both browsers and grazers. If the reticular 12

259 honeycomb structures and crests had an important function in either ‘trapping sediment’ or in

260 ‘catching larger particles’, both grazers and browsers should benefit equally from such a

261 mechanism. However, in browsers and grazers of different phylogenetic lineages, these crests

262 have converged to be shallow and prominent, respectively, suggesting that their major

263 function is not a direct mechanical involvement in the separation mechanism in the form of

264 ‘traps’.

265 High-density material in the reticulum will settle into the honeycomb structures, as in

266 slaughtered cattle (Grau 1955), as observed endoscopically in a goat whose reticulorumen had

267 been filled with grains and a transparent saline solution (Ehrlein 1979), and as demonstrated

268 in live sheep dosed with radio-opaque, highly dense iron sand (Reid 1985). While such

269 observations intuitively lead to assume an associated function, the density-dependent sorting

270 would work equally well if the reticular wall were smooth (as in some extreme browsers with

271 very shallow crests), because dense material would still be located at the bottom of the

272 reticulum and be transported towards the reticulo-omasal orifice during a reticular

273 contraction.

274 We propose that the mechanical function of the reticular crests is linked to the reticular

275 contractions and the associated reduction of the reticular lumen. A major prerequisite for the

276 particle sorting mechanism based on buoyancy characteristics of particles is a fluid medium in

277 which particles can actually float or sink. Therefore, invariably, the ingesta in the reticulum

278 has a comparatively high proportion of fluid (Clauss et al. 2009a, 2009b, Hummel et al.

279 2009). Particularly in predominantly grazing ruminants, this high proportion of fluid is

280 matched by a similar high proportion of fluid in the ingesta of the ventral rumen; in mixed

281 feeders and browsers, the ingesta in the ventral rumen may often be drier than that of the

282 reticulum (Clauss et al. 2009a, 2009b, Hummel et al. 2009). If fluid content of reticular and 13

283 ruminal ingesta differs, a contraction of the reticulum with complete closure of its lumen

284 might be undesirable, because when the reticulum subsequently relaxes, it will less likely re-

285 fill quickly with highly fluid ingesta necessary for the buoyancy-based separation mechanism.

286 In such a case, retaining a certain proportion of fluid in the reticulum by means of incomplete

287 reticular contractions may be more appropriate (Fig. 6a). Then ingesta entering the chamber

288 during relaxation can quickly be submitted to the next separation cycle. Therefore, the

289 reticular crests of browsers might be shallow to prevent complete closure of the reticular

290 lumen.

291 In grazers, with a more pronounced rumen contents stratification and a more distinct

292 fluid pool in the ventral rumen, the reticulum will likely be re-filled quickly with fluid after a

293 complete contraction. In such species, a complete contraction and re-filling with fluid ventral

294 rumen contents might even be beneficial, as the particle separation process due to buoyancy

295 and the ‘filter-bed effect’ (Faichney 2006) already occurs to some extent in the rumen (Clauss

296 et al. 2009a, 2009b, Hummel et al. 2009). In such species the reticulum could thus receive a

297 higher proportion of ‘pre-sorted’ material by complete contraction (Fig. 6b).

298 Overall, the reticular crests might not have a mechanical ‘trapping’ function but serve to

299 support reticular contractions and lumen reduction in two ways. First, the reticular mucosa is

300 only firmly anchored to the Tunica muscularis (the layer of smooth muscles covering the

301 whole reticulum) by means of elastic fibres at the very centre of the honeycomb structures

302 (Schels 1956); the rest of the reticular mucosa can be moved against the Tunica muscularis.

303 As the smooth muscles in the wall of the reticulum, the Tunica muscularis, contract, the

304 reticular crests are pressed together and help to reduce the lumen by forming a ‘secondary,

305 elevated inner surface’ in the reticulum (Hofmann 1969, pp. 81-82, 1973, p. 21, Hofmann &

306 Schnorr 1982, p. 31; Fig. 6) – which will be the higher the higher the crests are in the relaxed 14

307 state. Secondly, the primary and sometimes the secondary crests are equipped with a string of

308 smooth muscle at their luminal tip, the Muscularis mucosae, that serves to contract these

309 crests (Würfel 1908, Schels 1956, Hofmann 1973 p. 312, Hofmann & Schnorr 1982 pp. 61 &

310 63). Thus, the contraction of the whole reticulum is supported by a ‘pull’ from the net-like,

311 elevated muscle strings of the reticular crests – a mechanism that will have a stronger effect

312 the more prominent the crests are.

313 Our interpretation of the function of the reticular crests could be corroborated by other

314 comparative morphological data, for example by measurements of the thickness of the Tunica

315 muscularis of the reticulum in various ruminant species or by visualisation of reticular

316 contractions in different species. The interpretation is in line with our suggestion that as

317 ruminants evolved towards strict grazers, fluid throughput through their rumen (the first

318 forestomach compartment) increased (Clauss, Hummel & Streich 2006). This led to a more

319 pronounced rumen contents stratification (Clauss et al. 2009a, 2009b, 2009c, Hummel et al.

320 2009), possibly allowing a more efficient harvesting of the rumen microbial fauna (as

321 suggested by Hummel et al. 2008b). It also allowed the evolution of more prominent crests in

322 the reticulum (the second forestomach compartment) for more complete reticular contractions

323 to benefit from pre-sorting of particles in the stratified rumen contents (this study). But it also

324 necessitating larger omasa (the third forestomach compartment) to re-absorb the fluid to

325 prevent undue dilution of ingesta entering the sites of auto-enzymatic digestion (Clauss et al.

326 2006a). Together with the fact that a high relative fluid throughput at high food intake levels

327 sets ruminants apart from other foregut fermenters (Schwarm et al. 2009a), these findings

328 suggest that the consequences and potential benefits of high fluid throughput through the

329 forestomach of ruminants should be further investigated. Comparative studies of ruminant 15

330 anatomy are a prominent example of how the demonstration of convergent adaptations serves

331 to elucidate and generate new hypotheses on the link between form and function.

332

333 Acknowledgements

334 We thank the numerous hunters and zoological institutions that helped collecting

335 material, with special thanks to Kaarlo Nygrén. Norman Owen-Smith kindly provided the

336 data collection from his 1997 publication. MC thanks B. Schneider and J. Wick for support in

337 literature acquisition, J. Peter for the production of the schematic drawings, and two

338 anonymous reviewers and V. Hayssen for comments that improved the manuscript.

339

340 References

341 Allen, M. S. (1996) Physical constraints on voluntary intake of forages by ruminants. J. Anim. 342 Sci., 74, 3063-3075. 343 Andanje, S. A. & W. K. Ottichilo (1999) Population status and feeding habits of the 344 translocated subpopulation of Hunter's antelope or hirola (Beatragus hunteri) in Tsavo 345 East National Park, Kenya. Afr. J. Ecol., 37, 38 - 48. 346 Archer, D. & G. Sanson (2002) Form and function of the selenodont molar in southern 347 African ruminants in relation to their feeding habits. J. Zool., 257, 13-26. 348 Axmacher, H. & R. R. Hofmann (1988) Morphological characteristics of the masseter muscle 349 of 22 ruminant species. J. Zool., 215, 463-473. 350 Beaumont, R. & A. G. Deswysen (1991) Mélange et propulsion du contenu du réticulo- 351 rumen. Repr. Nutr. Dev., 31, 335-359. 352 Bigalke, R. C. (1972) Observations on the behavior and feeding habits of the springbok 353 (Antidorcas marsupialis). Zoologica Africana, 7, 333 - 359. 354 Branan, W. V., M. C. M. Werkhoven & R. L. Marchinton (1985) Food habits of brocket and 355 white - tailed deer in Suriname. J. Wildl. Manage., 49, 972 - 976. 356 Burne, R. H. (1917) Notes on some of the viscera of the okapi. Proceedings of the Zoological 357 Society of London, 1917, 187-208. 358 Campos-Arceiz, A., S. Takatsuki & B. Lhagvasuren (2004) Food overlap between Mongolian 359 gazelles and livestock in Omnogobi southern Mongolia. Ecological Research, 19, 360 455-460. 361 Clauss, M., M. Lechner-Doll, A. Behrend, K. Lason, D. Lang & W. J. Streich (2001) Particle 362 retention in the forestomach of a browsing ruminant, the roe deer (Capreolus 363 capreolus). Acta Theriol., 46, 103-107. 16

364 Clauss, M., M. Lechner-Doll & W. J. Streich (2002) Faecal particle size distribution in 365 captive wild ruminants: an approach to the browser/grazer-dichotomy from the other 366 end. Oecologia, 131, 343-349. 367 Clauss, M., J. Hummel, F. Vercammen & W. J. Streich (2005) Observations on the 368 macroscopic digestive anatomy of the Himalayan Tahr (Hemitragus jemlahicus). 369 Anat. Histol. Embryol., 34, 276-278. 370 Clauss, M., R. R. Hofmann, J. Hummel, J. Adamczewski, K. Nygren, C. Pitra, W. J. Streich 371 & S. Reese (2006a) The macroscopic anatomy of the omasum of free-ranging moose 372 (Alces alces) and muskoxen (Ovibos moschatus) and a comparison of the omasal 373 laminal surface area in 34 ruminant species. J. Zool., 270, 346-358. 374 Clauss, M., J. Hummel & W. J. Streich (2006) The dissociation of the fluid and particle phase 375 in the forestomach as a physiological characteristic of large grazing ruminants: an 376 evaluation of available, comparable ruminant passage data. Eur. J. Wildl. Res., 52, 88- 377 98. 378 Clauss, M., J. Hummel, J. Völlm, A. Lorenz & R. R. Hofmann, (2006b) The allocation of a 379 ruminant feeding type to the okapi (Okapia johnstoni) on the basis of morphological 380 parameters. In: Zoo animal nutrition III: 253-270. A. Fidgett, M. Clauss, K. 381 Eulenberger, J. Hatt, M, I. Hume, G. Janssens & J. Nijboer (Eds.). Filander Verlag, 382 Fürth, Germany. 383 Clauss, M., R. R. Hofmann, W. J. Streich, J. Fickel & J. Hummel (2008) Higher masseter 384 mass in grazing than in browsing ruminants. Oecologia, 157, 377–385. 385 Clauss, M., T. Kaiser & J. Hummel, (2008) The morphophysiological adaptations of browsing 386 and grazing mammals. In: The ecology of browsing and grazing: 47-88. I. J. Gordon & 387 H. H. T. Prins (Eds.). Springer, Heidelberg. 388 Clauss, M., J. Fritz, D. Bayer, J. Hummel, W. J. Streich, K.-H. Südekum & J. M. Hatt (2009a) 389 Physical characteristics of rumen contents in two small ruminants of different feeding 390 type, the mouflon (Ovis ammon musimon) and the roe deer (Capreolus capreolus). 391 Zoology, 112, 195-205. 392 Clauss, M., J. Fritz, D. Bayer, K. Nygren, S. Hammer, J. M. Hatt, K.-H. Südekum & J. 393 Hummel (2009b) Physical characteristics of rumen contents in four large ruminants of 394 different feeding type, the addax (Addax nasomaculatus), bison (Bison bison), red deer 395 (Cervus elaphus) and moose (Alces alces). Comp. Biochem. Physiol. A, 152, 398-406. 396 Clauss, M., R. R. Hofmann, J. Fickel, W. J. Streich & J. Hummel (2009c) The intraruminal 397 papillation gradient in wild ruminants of different feeding types: implications for 398 rumen physiology. J. Morphol., 270, 929–942. 399 Clauss, M., C. Nunn, J. Fritz & J. Hummel (2009d) Evidence for a tradeoff between retention 400 time and chewing efficiency in large mammalian herbivores. Comp. Biochem. Physiol. 401 A, 154, 376-382. 402 Clauss, M., S. Reese & K. Eulenberger, (2009) Macroscopic digestive anatomy of a captive 403 lowland anoa (Bubalus depressicornis). In: Zoo animal nutrition IV: 257-265. M. 404 Clauss, A. L. Fidgett, J. M. Hatt, T. Huisman, J. Hummel, G. Janssen, J. Nijboer & A. 405 Plowman (Eds.). Filander Verlag, Fürth. 406 Codron, D., J. S. Brink, L. Rossouw, M. Clauss, J. Codron, J. A. Lee-Thorp & M. 407 Sponheimer (2008) Functional differentiation of African grazing ruminants: an 408 example of specialized adaptations to very small changes in diet. Biol. J. Linn. Soc., 409 94, 755–764. 17

410 Dhungel, S. K. & B. W. O'Gara (1991) Ecology of the hog deer in Royal Chitwan National 411 Park, Nepal. Wildlife Monographs, 119, 1-40. 412 Du Plessis, S. S. (1972) Ecology of blesbok with special reference to productivity. Wildlife 413 Monographs, 30, 1-70. 414 Ehrlein, H. (1979) Motility of the forestomach in goats. Ann Rech Vet, 10, 173-175. 415 Eldridge, W. D., M. M. Macnamara & N. V. Pacheco, (1987) Activity patterns and habitat 416 utilization of pudus (Pudu puda) in south-central Chile. In: Biology and management 417 of the Cervidae: 352-369. C. M. Wemmer (Ed.). Smithsonian Inst. Press, Washington 418 DC. 419 Faichney, G. J., (2006) Digesta flow. In: Quantitative aspects of ruminant digestion and 420 metabolism: 49-86. J. Dijkstra, J. M. Forbes & J. France (Eds.). CAB International, 421 Wellingford, UK. 422 Fernandez, M. H. & E. S. Vrba (2005) A complete estimate of the phylogenetic relationships 423 in Ruminantia: a dated species-level supertree of the extant ruminants. Biol. Rev., 80, 424 269–302. 425 Ferreira, N. A. & R. C. Bigalke (1987) Food selection by grey rhebuck in the Orange Free 426 State. S. Afr. J. Wildl. Res., 17, 123 - 127. 427 Field, C. R. (1972) The food habits of wild ungulates in Uganda by analysis of stomach 428 contents. E. Afr. Wildl. J., 10, 17-42. 429 Fritz, J., J. Hummel, E. Kienzle, C. Arnold, C. Nunn & M. Clauss (2009) Comparative 430 chewing efficiency in mammalian herbivores. Oikos, 118, 1623-1632. 431 Gagnon, M. & A. E. Chew (2000) Dietary preferences in extant African Bovidae. J. 432 Mammal., 81, 490-511. 433 Geiger, G., R. R. Hofmann & R. König (1977) Vergleichend-anatomische Untersuchungen an 434 den Vormägen von Damwild (Cervus dama) und Muffelwild (Ovis ammon musimon). 435 Säugetierkundliche Mitteilungen, 25, 7-21. 436 Geist, V. (1999) Deer of the world. Their evolution, behaviour, and ecology. Swan Hill Press. 437 Grau, H. (1955) Zur Funktion der Vormägen, besonders des Netzmagens, der Wiederkäuer. 438 Berliner und Münchner Tierärztliche Wochenschrift, 15, 271-275. 439 Grobler, J. H. (1974) Aspects of the biology, population ecology and behaviour of the sable 440 (Hippotragus niger) in the Rhodes Matopos National Park, Rhodesia. Arnoldia 441 Rhodesia, 7, 1-36. 442 Hart, J. A. & T. B. Hart (1988) A summary report on the behaviour, ecology and conservation 443 of the okapi (Okapia johnstoni) in Zaire. Acta Zoologica et Pathologica Antverpiensa, 444 80, 19-28. 445 Haubner, G. C. (1837) Ueber die Magenverdauung der Wiederkäuer nach Versuchen, nebst 446 einer Prüfung der Flourens'schen Versuche über das Wiederkäuen. Anclam/Wien: 447 Verlag von W. Dietze/Gerold'sche Buchhandlung. 448 Heinichen, I. G. (1972) Preliminary notes on the suni (Neotragus moschatus) and the red 449 duiker (Cephalophus natalensis). Zoologica Africana, 7, 157 - 165. 450 Heptner, V. G., A. A. Nasimowitsch & A. G. Bannikov (1989) Mammals of the Sowiet Union 451 Vol. 1. Leiden, The Netherlands: E. J. Brill. 452 Hofmann, R. R. (1969) Zur Topographie und Morphologie des Wiederkäuermagens im 453 Hinblick auf seine Funktion (nach vergleichenden Untersuchungen an Material 454 ostafrikanischer Wildarten). Zentralblatt für Veterinärmedizin, Supplement 10, 1- 455 180. 18

456 Hofmann, R. R. & D. R. M. Stewart (1972) Grazer or browser: a classification based on the 457 stomach-structure and feeding habit of East African ruminants. Mammalia, 36, 226- 458 240. 459 Hofmann, R. R. (1973) The ruminant stomach. Nairobi: East African Literature Bureau. 460 Hofmann, R. R., G. Geiger & R. König (1976) Vergleichend-anatomische Untersuchungen an 461 der Vormagenschleimhaut von Rehwild (Capreolus capreolus) und Rotwild (Cervus 462 elaphus). Zeitschrift für Säugetierkunde, 41, 167-193. 463 Hofmann, R. R. & B. Schnorr (1982) Die funktionelle Morphologie des Wiederkäuer-Magens. 464 Stuttgart: Ferdinand Enke Verlag. 465 Hofmann, R. R. (1985) Digestive physiology of deer - their morphophysiological 466 specialisation and adaptation. Royal Society of New Zealand Bulletin, 22, 393-407. 467 Hofmann, R. R., (1988) Morphophysiological evolutionary adaptations of the ruminant 468 digestive system. In: Aspects of digestive physiology in ruminants: 1-20. A. Dobson & 469 M. J. Dobson (Eds.). Cornell University Press, Ithaca, NY. 470 Hofmann, R. R. (1989) Evolutionary steps of ecophysiological adaptation and diversification 471 of ruminants: a comparative view of their digestive system. Oecologia, 78, 443-457. 472 Hofmann, R. R., (1991) Endangered tropical herbivores - their nutritional requirements and 473 habitat demands. In: Recent advances on the nutrition of herbivores: 27-34. Y. W. Ho, 474 H. K. Wong, N. Abdullah & Z. A. Tajuddin (Eds.). Malaysia Society of Animal 475 Production, UPM Serdang. 476 Hofmann, R. R. & K. Nygren (1992) Morphophysiological specializations and adaptations of 477 the moose digestive system. Alces, Suppl. 1, 91-100. 478 Hofmann, R. R., M. H. Knight & J. D. Skinner (1995) On structural characteristics and 479 morphophysiological adaptation of the springbok (Antidorcas marsupialis) digestive 480 system. Transactions of the Royal Society of South Africa, 50, 125-142. 481 Hofmann, R. R. (2000) Functional and comparative digestive system anatomy of Arctic 482 ungulates. Rangifer, 20, 71-81. 483 Hofmann, R. R., W. J. Streich, J. Fickel, J. Hummel & M. Clauss (2008) Convergent 484 evolution in feeding types: salivary gland mass differences in wild ruminant species. J. 485 Morphol., 269, 240-257. 486 Hristov, A. N., S. Ahvenjärvi, T. A. McAllister & P. Huhtanen (2003) Composition and 487 digestive tract retention time of ruminal particles with functional specific gravity 488 greater or less than 1.02. J. Anim. Sci., 81, 2639-2648. 489 Hummel, J., K.-H. Südekum, W. J. Streich & M. Clauss (2006) Forage fermentation patterns 490 and their implications for herbivore ingesta retention times. Funct. Ecol., 20, 989- 491 1002. 492 Hummel, J., J. Fritz, E. Kienzle, E. P. Medici, S. Lang, W. Zimmermann, W. J. Streich & M. 493 Clauss (2008a) Differences in fecal particle size between free-ranging and captive 494 individuals of two browser species. Zoo Biol., 27, 70-77. 495 Hummel, J., P. Steuer, K.-H. Südekum, S. Hammer, C. Hammer, W. J. Streich & M. Clauss 496 (2008b) Fluid and particle retention in the digestive tract of the addax antelope (Addax 497 nasomaculatus) – adaptations of a grazing desert ruminant. Comp. Biochem. Physiol. 498 A, 149, 142-149. 499 Hummel, J., K.-H. Südekum, D. Bayer, S. Ortmann, J. M. Hatt, W. J. Streich & M. Clauss 500 (2009) Physical characteristics of reticuloruminal contents of cattle in relation to 501 forage type and time after feeding. J. Anim. Physiol. Anim. Nutr., 93, 209-220. 19

502 Kaiser, T. M., J. Brasch, J. C. Castell, E. Schulz & M. Clauss (2009) Dental wear patterns in 503 captive wild ruminant species differ from those of free-ranging conspecifics. Mamm. 504 Biol., 74, 425–437. 505 Kaiser, T. M., J. Fickel, W. J. Streich, J. Hummel & M. Clauss (2010) Enamel ridge 506 alignment in upper molars of ruminants in relation to their natural diet. J. Zool., 507 (submitted). 508 Kaske, M. & A. Midasch (1997) Effects of experimentally-impaired reticular contractions on 509 digesta passage in sheep. Br. J. Nutr., 78, 97-110. 510 Lanave, C., G. Preparata, C. Sacone & G. Serio (1984) A new method for calculating 511 evolutionary substitution rates. Journal of Molecular Evolution, 20, 86-93. 512 Langer, P. (1973) Vergleichend-anatomische Untersuchungen am Magen der Artiodactyla. II. 513 Untersuchungen am Magen der Tylopoda und Ruminantia. Gegenbaurs Morphol. Jb., 514 119, 633-695. 515 Langer, P. (1988) The mammalian herbivore stomach. Stuttgart/New York: Gustav Fischer 516 Verlag. 517 Lechner-Doll, M., M. Kaske & W. v. Engelhardt, (1991) Factors affecting the mean retention 518 time of particles in the forestomach of ruminants and camelids. In: Physiological 519 aspects of digestion and metabolism in ruminants: 455-482. T. Tsuda, Y. Sasaki & R. 520 Kawashima (Eds.). Academic Press, San Diego. 521 Lobao Tello, J. L. & R. G. van Gelder (1975) The natural history of nyala (Tragelaphus 522 angasi) in Mozambique. Bulletin of the American Museum of Natural History, 155, 523 319 - 386. 524 Loggers, C. (1991) Forage availability versus seasonal diets, as determined by fecal analysis, 525 of dorcas gazelles in Morocco. Mammalia, 55, 255-267. 526 Martins, E. (2004) COMPARE, version 4.6. Computer programs for the statistical analysis of 527 comparative data. Distributed by the author at http://compare.bio.indiana.edu/. 528 Department of Biology, Indiana University, Bloomington IN. 529 Martins, E. P. & T. F. Hansen (1997) Phylogenies and the comparative method: a general 530 approach to incorporating phylogenetic information into analysis of interspecific data. 531 Am. Nat., 149, 646-667. 532 Mathison, G. W., E. K. Okine, A. S. Vaage, M. Kaske & L. P. Milligan, (1995) Current 533 understanding of the contribution of the propulsive activities in the forestomach to the 534 flow of digesta. In: Ruminant physiology: Digestion, metabolism, growth and 535 reproduction: 23-41. W. von Engelhardt, S. Leonhard-Marek, G. Breves & D. 536 Giesecke (Eds.). Ferdinand Enke Verlag, Stuttgart. 537 Neuville, H. & J. M. Derscheid (1929) Recherches anatomiques sur l'okapi. IV. L'estomac. 538 Revue de Zoologie Africaine, 16, 373-419. 539 Okine, E. K., G. W. Mathison & R. T. Hardin (1990) Effects of changes in attributes of 540 reticular contraction on fecal particle sizes in cattle. Can. J. Anim. Sci., 70, 159-166. 541 Okine, E. K., G. W. Mathison, M. Kaske, J. J. Kenelly & R. J. Christopherson (1998) Current 542 understanding of the role of the reticulum and reticulo-omasal orifice in the control of 543 digesta passage from the ruminoreticulum of sheep and cattle. Can. J. Anim. Sci., 78, 544 15-21. 545 Owen, R. E. A. (1970) Some observations on the sitatunga in Kenya. E. Afr. Wildl. J., 8, 181- 546 195. 547 Owen-Smith, N. (1997) Distinctive features of the nutritional ecology of browsing versus 548 grazing ruminants. Zeitschrift für Säugetierkunde, 62 (Suppl. 2), 176-191. 20

549 Pérez-Barberìa, F. J., D. A. Elston, I. J. Gordon & A. W. Illius (2004) The evolution of 550 phylogenetic differences in the efficiency of digestion in ruminants. Proceedings of 551 the Royal Society of London B, 271, 1081-1090. 552 Peyer, J. C. (1685) Merycologia sive de Ruminantibus et Ruminatione Commentarius. Basle, 553 Switzerland: J. L. Koenig & J. Brandmyller. 554 Pfeiffer, C., (1993) Vergleichend-morphologische Untersuchungem am Verdauungstrakt des 555 chinesischen Muntjak (Muntiacus reevesii) einschliesslich saisonaler Veränderungen 556 an der Vormagenschleimhaut. 95. Giessen, Germany. 557 Pitra, C., J. Fickel, E. Meijaard & P. C. Groves (2004) Evolution and phylogeny of old world 558 deer. Mol Phylogenet Evol, 33, 880-895. 559 Posada, D. & K. A. Crandall (1998) Modeltest: testing the model of DNA substitution. 560 Bioinformatics, 14, 817-818. 561 Purvis, A. & T. Garland (1993) Polytomies in comparative analyses of continuous characters. 562 Systematical Biology, 42, 569-575. 563 Reid, C. S. W., (1985) The progress of solid feed residues through the rumino-reticulum: the 564 ins and outs of particles. In: Ruminant physiology: concepts and consequences: 79-84. 565 S. K. Baker, J. M. Gawthorne, J. B. Mackintosh & D. B. Purser (Eds.). University of 566 Western Australia Press, Perth, Australia. 567 Rodriguez, F., J. F. Oliver, A. Marin & J. R. Medina (1990) The general stochastic model of 568 nucleotide substitution. Journal of Theoretical Biology, 142, 485-501. 569 Rohlf, F. (2001) Comparative methods for the analysis of continuous variables: geometric 570 interpretations. Evolution, 55, 2143-2160. 571 Schaller, G. B. (1973) Observations on Himalayn Tahr (Hemitragus jemlahicus). Journal of 572 the Bombay Natural History Society, 70, 1-24. 573 Schels, H. (1956) Untersuchungen über Wandbau und Funktion des Netzmagens des Rindes. 574 München: Dissertation thesis. 575 Schmidt, H. A., K. Strimmer, M. Vingron & A. von Haeseler (2002) TREEPUZZLE: 576 maximum likelihood phylogenetic analysis using quartets and parallel computing. 577 Bioinformatics, 18, 502–504. 578 Schwarm, A., S. Ortmann, C. Wolf, W. J. Streich & M. Clauss (2008) Excretion patterns of 579 fluids and particle passage markers of different size in banteng (Bos javanicus) and 580 pygmy hippopotamus (Hexaprotodon liberiensis): two functionally different foregut 581 fermenters. Comp. Biochem. Physiol. A, 150, 32–39. 582 Schwarm, A., S. Ortmann, C. Wolf, W. J. Streich & M. Clauss (2009a) Passage marker 583 excretion in red kangaroo (Macropus rufus), collared peccary (Pecari tajacu) and 584 colobine monkeys (Colobus angolensis, C. polykomos, Trachypithecus johnii). J. Exp. 585 Zool., 311, 647-661. 586 Schwarm, A., S. Ortmann, C. Wolf, W. J. Streich & M. Clauss (2009b) More efficient 587 mastication allows increasing intake without compromising digestibility or 588 necessitating a larger gut: comparative feeding trials in banteng (Bos javanicus) and 589 pygmy hippopotamus (Hexaprotodon liberiensis). Comp. Biochem. Physiol. A, 152, 590 504-512. 591 Stafford, K. J. & Y. M. Stafford (1990) Stomach of the puku. Afr. J. Ecol., 28, 250-252. 592 Stafford, K. J. & Y. M. Stafford (1991) The stomach of the Kafue lechwe (Kobus leche 593 kafuensis). Anat. Histol. Embryol., 20, 299-310. 594 Stevens, C. E. & I. D. Hume (1998) Contributions of microbes in vertebrate gastrointestinal 595 tract to production and conservation of nutrients. Physiological Reviews, 78, 393-427. 21

596 Stubbe, C. (1971) Zur Ernährung des Muffelwildes (Ovis ammon musimon) in der DDR. 597 Beiträge zur Jagd- und Wildforschung, 7, 103-125. 598 Sutherland, T. M., (1988) Particle separation in the forestomach of sheep. In: Aspects of 599 digestive physiology in ruminants: 43-73. A. Dobson & M. J. Dobson (Eds.). Cornell 600 University Press, Ithaca, NY. 601 Swofford, D. L. (2002) PAUP*: Phylogenetic Analyses Using Parsimony (and Other 602 Methods), Version 4.0 Beta. Washington DC, Smithsonian Institition. 603 Tavaré, S. (1986) Some probabilistic and statistical problems in the analysis of DNA 604 sequences. Lectures on Mathematics in the Life Sciences, 17, 57–86. 605 Thomas, W. M. & S. M. Salis (2000) Diet of marsh deer (Blastocerus dichotomus) in the 606 Pantanal wetland, Brazil. Studies on Neotropical Fauna and Environment, 35, 165- 607 172. 608 Thompson, J. D., T. J. Gibson, F. Plewniak, F. Jeanmougin & D. G. Higgins (1997) The 609 CLUSTAL_X windows interface: flexible strategies for multiple sequence alignment 610 aided by quality analysis tools. Nucleic Acids Research, 15, 4876-4882. 611 Tschuor, A. & M. Clauss (2008) Investigations on the stratification of forestomach contents in 612 ruminants: an ultrasonographic approach. Eur. J. Wildl. Res., 54, 627-633. 613 Van Wieren, S. E., (1996) Browsers and grazers: foraging strategies in ruminants. In: 614 Digestive strategies in ruminants and nonruminants: 119-146. S. E. Van Wieren (Ed.). 615 Thesis Landbouw, University of Wageningen, The Netherlands. 616 Würfel, K. (1908) Vergleichende anatomische und histologische Untersuchungen über den 617 Bau des zweiten Magens und über die Muscularis mucosae des Vorderdarms der 618 Wiederkäuer (Rind und Schaf). Dresden: Dissertation thesis. 619 620 621 22

621 Table 1. Data used in this study (species sorted with increasing %grass in the natural diet) R Ret Ret Ret crest % Species FT BM Source Source height height width height grass kg cm cm cm mm Gerenuk Litocranius walleri BR 43.0 20.0 14.3 9.8 1.50 A 0.0 2 Suni Neotragus moschatus BR 6.2 12.8 7.2 4.9 1.50 A 0.0 6 Okapi Okapia johnstoni BR 205.5 45.5 18.8 23.0 1.25 L* 0.0 17 Giraffe Giraffa camelopardalis BR 750.0 75.0 32.0 18.8 2.00 A 0.2 2 Harvey’s duiker Cephalophus harveyi BR 16.0 18.0 11.8 10.3 1.20 A 1.0 3 Red duiker Cephalophus natalensis BR 8.3 16.0 9.0 4.5 B 1.0 3 Mazama Mazama americana BR 14.0 22.2 9.9 7.2 B 1.0 16 Moose Alces alces BR 331.8 63.0 28.5 2.55 BE 2.0 1 Pudu Pudu pudu BR 8.1 21.0 9.0 5.5 0.50 B* 3.0 22 Günther's dikdik Madoqua guentheri BR 4.1 10.0 7.4 5.0 1.00 A 5.0 10 Klipspringer Oreotragus oreotragus BR 11.4 16.5 10.8 7.8 1.80 A 5.0 1 Grey duiker Sylvicapra grimmia BR 14.0 15.3 10.0 7.8 1.20 A 5.0 2 Greater Kudu Tragelaphus strepsiceros BR 285.0 56.5 30.0 15.5 3.50 A 5.0 1 Vaal Rhebok Pelea capreolus BR 20.0 25.0 11.0 6.0 3.00 B 7.0 18 Roe deer Capreolus capreolus BR 16.6 25.0 11.0 0.60 BC 9.0 1 White-tailed deer Odocoileus virginianus BR 74.0 38.0 14.0 10.5 0.50 BH 9.0 1 Phillip's dikdik Madoqua saltiana BR 2.8 11.5 7.3 5.0 0.50 N* 10.0 3 Muntjak Muntiacus muntjak BR 13.8 18.5 8.7 7.3 0.75 G 10.0 1 Steenbok Raphicerus campestris IM 10.5 16.0 8.5 6.5 0.75 A 10.0 1 Lesser Kudu Tragelaphus imberbis BR 97.5 37.0 17.5 12.0 2.50 A 10.0 2 Bushbuck Tragelaphus scriptus BR 73.0 23.5 15.3 9.8 2.50 A 10.0 1 Kirk's dikdik Madoqua kirki BR 5.2 11.0 7.9 5.3 1.00 A 17.0 3 Nyala Tragelaphus angasi BR 95.0 44.5 19.5 11.0 1.75 B 20.0 19 Pelzen's gazelle Gazella dorcas pelzeni IM 17.3 22.7 12.2 7.3 1.50 N* 23.0 24 Swamp deer Blastocerus dichotomus IM 56.0 37.0 15.0 11.0 B* 24.0 21 Mongolian gaz. Procapra gutturosa IM 30.0 28.0 14.5 9.0 B 28.0 12 Springbok Antidorcas marsupialis IM 40.6 28.8 14.1 8.2 2.25 F 30.0 4 Grysbok Raphicerus melanotis IM 9.7 21.0 8.8 6.5 B 30.0 3 Reindeer Rangifer tarandus IM 130.0 59.0 30.0 16.0 B 36.0 1 Fallow deer Dama dama IM 93.0 38.0 15.0 10.0 1.00 BD 46.0 1 Red deer Cervus elaphus IM 190.0 56.0 26.0 1.50 N 47.0 1 Dama gazelle Nanger dama IM 32.0 30.0 11.0 7.0 1.30 N* 47.5 3 Oribi Ourebia ourebi GR 16.0 20.0 11.0 8.0 2.00 A 48.5 2 Hog deer Axis porcinus IM 54.4 34.2 14.5 10.5 B* 50.0 5 Anoa Bubalus depressicornis IM 53.2 40.0 19.5 10.0 6.00 M* 50.0 est Sika deer Cervus nippon IM 98.0 46.0 18.0 1.00 B 50.0 1 Grant's gazelle Gazella granti IM 50.0 28.0 17.0 10.3 2.50 A 50.0 1 Spekes gazelle Gazella spekei IM 10.0 24.0 10.0 7.0 1.00 N* 50.0 3 Goitered gazelle Gazella subgutturosa IM 13.1 25.0 10.0 7.0 1.00 N* 50.0 23 Chinese water deer Hydropotes inermis IM 12.0 20.3 9.3 5.1 B* 50.0 7 Soemmerring gaz. Nanger soemmerringii IM 25.0 31.0 12.0 7.0 2.50 N* 50.0 3 Eland Taurotragus oryx IM 465.0 42.0 37.0 21.0 2.25 A 50.0 3 Impala Aepyceros melampus IM 55.2 29.0 16.5 10.5 3.50 A 60.0 1 Alpine ibex Capra ibex IM 38.0 41.5 19.0 11.0 2.00 B 60.0 1 European bison Bison bonasus GR 575.0 78.0 41.0 12.5 10.50 B* 68.0 1 Sitatunga Tragelaphus spekii IM 55.0 37.0 18.0 11.0 B 68.0 20 Mouflon Ovis ammon musimon GR 32.0 25.0 11.0 7.0 2.00 BD 69.0 11 Axis deer Axis axis IM 57.0 34.0 14.0 8.5 B* 70.0 1 Laristan mouflon Ovis orientalis larist. GR 40.0 32.0 14.0 7.0 2.50 N* 70.0 est Chamois Rupicapra rupicapra IM 22.5 32.0 14.0 9.0 1.20 B 74.0 1 Blackbuck Antilope cervicapra GR 19.5 29.3 12.3 7.5 2.00 N* 75.0 1 Père David’s deer Elaphurus davidianus GR 192.0 56.0 22.0 14.0 B* 75.0 7 Thomson's gazelle Gazella thomsoni IM 21.0 22.0 12.5 7.5 1.25 A 75.0 3 Himalayan tahr Hemitragus jemlahicus IM 41.6 1.87 K* 75.0 8 Addax Addax nasomaculatus GR 83.8 51.5 25.3 14.5 6.90 N* 80.0 3 Banteng Bos javanicus GR 600.0 85.0 36.0 15.0 13.00 N* 80.0 est Waterbuck Kobus ellipsiprymnus GR 200.5 54.0 20.5 14.0 6.00 A 80.0 2 Barasingha Rucervus duvauceli IM 145.0 46.0 19.0 12.0 B* 80.0 1 Gemsbok Oryx gazella GR 181.5 41.0 22.0 14.0 5.00 A 82.0 2 Bison Bison bison GR 335.0 75.0 41.0 19.5 10.00 B 84.0 1 Hirola Beatragus hunteri GR 130.0 47.0 21.0 10.0 B 90.0 13 23

Blue wildebeest Connochaetes taurinus GR 182.0 49.0 27.5 16.0 5.00 A 90.0 1 African buffalo Syncerus caffer GR 598.7 66.0 43.0 23.0 12.00 A 90.0 1 Sable antelope Hippotragus niger GR 235.0 55.0 29.0 11.0 B 93.0 15 Puku Kobus vardoni GR 61.0 3.00 I 93.0 3 Uganda kob Kobus kob GR 79.0 35.0 14.5 9.5 6.00 A 95.0 9 Lechwe Kobus leche GR 72.0 36.0 17.0 9.0 4.56 BJ 95.0 3 Bohor Reedbuck Redunca redunca GR 45.0 25.0 11.5 8.8 3.00 A 95.0 3 Hartebeest Alcelaphus buselaphus GR 174.0 45.0 15.0 15.0 4.00 A 96.7 2 Mountain reedbuck Redunca fulvolufula GR 23.5 24.0 12.5 8.0 3.00 A 99.0 2 Tsessebe Damaliscus lunatus GR 119.0 44.0 18.0 12.5 4.00 A 99.3 2 Bontebok Damaliscus pygarus GR 72.5 47.0 20.0 13.0 5.00 B 100.0 14 622 623 Sources for morphological data: A (Hofmann 1973), B (Hofmann, unpubl.), C (Hofmann, 624 Geiger & König 1976), D (Geiger, Hofmann & König 1977), E (Hofmann & Nygren 625 1992), F (Hofmann, Knight & Skinner 1995), G (Pfeiffer 1993), H (Langer 1973), I 626 (Stafford & Stafford 1990), J (Stafford & Stafford 1991), K (Clauss et al. 2005), L 627 (Clauss et al. 2006b), M (Clauss, Reese & Eulenberger 2009), N (Clauss, unpubl.) 628 *data from captive specimens 629 630 Sources for %grass data: 1 (Van Wieren 1996), 2 (Owen-Smith 1997), 3 (Gagnon & Chew 631 2000), 4 (Bigalke 1972), 5 (Dhungel & O'Gara 1991), 6 (Heinichen 1972), 7 (Geist 632 1999), 8 (Schaller 1973), 9 (Field 1972), 10 (Hofmann & Stewart 1972), 11 (Stubbe 633 1971), 12 (Campos-Arceiz, Takatsuki & Lhagvasuren 2004), 13 (Andanje & Ottichilo 634 1999), 14 (Du Plessis 1972), 15 (Grobler 1974), 16 (Branan, Werkhoven & 635 Marchinton 1985), 17 (Hart & Hart 1988), 18 (Ferreira & Bigalke 1987), 19 (Lobao 636 Tello & van Gelder 1975), 20 (Owen 1970), 21 (Thomas & Salis 2000), 22 (Eldridge, 637 Macnamara & Pacheco 1987), 23 (Heptner, Nasimowitsch & Bannikov 1989), 24 638 (Loggers 1991), est (estimated) 639 640 641 642 24

642 Table 2. Regression analyses of linear anatomical measurements of the reticulum (data from 643 Table 1) according to the equation ln(linear measurement)=a + b*ln(body 644 mass)+c*(%grass) using the raw data Measurement R2 a (95% CI) p (a) b (95% CI) p (b) c (95% CI) p (c)

Rumen height 0.88 2.06 (1.93-2.20) <0.0001 0.34 (0.30-0.37) <0.0001 0.002 (0.000-0.003) 0.020

Reticulum height 0.89 1.44 (1.32-1.57) <0.0001 0.33 (0.29-0.36) <0.0001 0.000 (-0.001-0.002) 0.554

Reticulum width 0.82 1.21 (1.07-1.35) <0.0001 0.28 (0.24-0.32) <0.0001 -0.001 (-0.002-0.001) 0.267

Reticular crest height 0.61 -0.87 (-1.32- -0.41) <0.0001 0.30 (0.18-0.41) <0.0001 0.010 (0.006-0.015) <0.0001

645

646 Table 3. Regression analyses of linear anatomical measurements of the reticulum (data from

647 Table 1) according to the equation ln(linear measurement)=a + b*ln(body

648 mass)+c*(%grass) using the Phylogenetic Generalized Least-Squares approach

Measurement R2 b p (b) c p (c)

Rumen height 0.86 0.33 <0.001 0.002 0.024

Reticulum height 0.86 0.32 <0.001 0.000 0.470

Reticulum width 0.77 0.29 <0.001 -0.001 0.310

Reticular crest height 0.58 0.31 <0.001 0.010 <0.001

649

650 25

650

651 Figure 1. Examples for differences in the height of the crests and prominence of secondar and 652 tertiary crets of the reticular mucosa in (from left to right) giraffe (Giraffa 653 camelopardalis), suni (Neotragus moschatus), impala (Aepyceros melampus), and 654 waterbuck (Kobus elipsiprymnus). Mucosa samples photographed at various scales for 655 a visual comparison. From Hofmann (1973). 656

657 Figure 2. Example for differences in the height of the crests and prominence of secondar and 658 tertiary crets of the reticular mucosa in (from top to bottom) bushbuck (Tragelaphus 659 scriptus), wildebeest (Connochaetes taurinus), hartebeest (Alcelaphus buselaphus), 660 and zebu cattle (Bos primigenius indicus). Mucosa samples photographed at various 661 scales for a visual comparison. From Hofmann (1969). 662 26

662 663 Figure 3. Fifty % majority rule maximum likelihood tree (100,000 puzzling steps), depicting 664 the phylogenetic relationships among complete mitochondrial cytochrome b sequences from 665 66 ruminant taxa as used in the phylogenetically controlled statistics in this study (accession 666 codes from GenBank (http://www.ncbi.nlm.nih.gov). 667 27

667 a b

c d

668 669 Figure 4. Correlations between body mass and a) rumen height, b) reticulum height, c) 670 reticulum width, d) maximum height of the reticular crests in wild ruminants of 671 different feeding type. Data from Table 1. For statistics, see Table 2. 672 28

672 a b

c d

673 674 Figure 5. Correlation between the percentage of grass in the natural diet and a) the relative 675 rumen height, b) the relative reticulum height, c) the reticulum:rumen height ratio, d) 676 the relative maximum reticular crest height in wild ruminants of different feeding 677 type. Data from Table 1. For statistics, see Table 2. 678 29

678 a b

679 680 Figure 6. Schematic representation of the reticulum in relaxed and contracted state in a) a 681 typical ‘browser’ (with shallow reticular crests) and b) a typical ‘grazer’ (with high 682 reticular crests). Note the hypothetical difference in lumen closure between the two 683 feeding types. Drawings by Jeanne Peter.