OIKOS 102: 253–262, 2003 Ruminant diversification as an adaptation to the physicomechanical characteristics of forage. A reevaluation of an old debate and a new hypothesis Marcus Clauss, Matthias Lechner-Doll and W. Ju ¨rgen Streich Clauss, M., Lechner-Doll, M. and Streich, W. J. 2003. Ruminant diversification as an adaptation to the physicomechanical characteristics of forage. A reevaluation of an old debate and a new hypothesis. – Oikos 102: 253–262. The validity of Hofmann’s classification of ruminants into browsers/‘‘concentrate selectors’’, intermediate feeders and grazers/‘‘grass and roughage eaters’’ and of his consecutive physiological postulates has repeatedly been questioned. In contrast to former concepts, which all focused on the chemical characteristics of the respective forages, we propose a new hypothesis on the main driving force of ruminant diversification, namely the physicomechanical characteristics of the respective for- ages. In contrast to browse, grass tends to stratify and form a ‘‘fibrous raft’’ in the reticulorumen. The significantly more capacious forestomachs of grazers, and the significantly thicker rumen pillars (indicating the strength of reticulorumen muscle equipment) of their forestomachs, are interpreted as particular adaptations to this forage characteristic. With these parameters, we present, for the first time, two single morphological measurements that allow the statistical reconstruction of Hofmann’s classification. A small forestomach capacity and the lack of strong reticulorumen muscles in browsers would explain the observed exclusiveness with which browsers avoid grass forage under natural conditions, which we confirmed using two datasets on the composition of the natural diet. Both rumen pillar thickness and relative forestomach capacity were significantly correlated to the grass content of the natural diet, respectively. Our functional interpretation was also supported by a stepwise regression analysis with the proportion of grass in the natural diet as dependent variable and the rumen pillar thickness, the relative forestomach capacity, and the body weight as independent variables, which revealed significant equations. M. Clauss, Inst. of Animal Physiology, Physiological Chemistry and Animal Nutrition, Veterinaerstr. 13, DE-80539 Muenchen, Germany (clauss@tiph.etmed.uni - muenchen.de).– M. Lechner -Doll and W. J. Streich, Inst. of Zoo Biology and Wildlife Research (IZW) Berlin, Germany. Ruminants have been classified according to their feed- ing habits for a long time (Talbot and Talbot 1962, Lamprey 1963, Van Zyl 1965, Jarman 1974, Hansen et al. 1985, Bodmer 1990, Gagnon and Chew 2000) and differences in ecology and behaviour have been noted. However, it was the pioneering work of Hofmann (1969, 1973, 1988, 1989) that brought the concept of different feeding types to the high level of attention it is still receiving today. Hofmann observed morphological differences between the feeding types – the ‘‘concen- trate selectors’’ 1 /browsers (BR), intermediate feeders (IM) and grazers (GR) – and hypothesized consequent physiological differences. There are several limitations to Hofmann’s presenta- tions. Hofmann (1973) did not record the body weights for most individuals he investigated but gave mostly 1 The use of the term ‘concentrate selectors’ has been criticised by Hoppe (1984), Robbins et al. (1995), Owen-Smith (1997) and Clauss et al. (2002b); we prefer the term ‘‘browser’’, even when referring to Hofmann’s morphological classification. Accepted 19 February 2003 Copyright © OIKOS 2003 ISSN 0030-1299 OIKOS 102:2 (2003) 253