Molecular Ecology zyxwvutsr (1995) 4,29-38 zyxwvutsrqp A new zyxwv Holcus Holcus hypothesis for the origin of pentaploid from diploid Holcus lanatus L. and tetraploid mollis L. in France M. RICHARD, M.-F. JUBIER,t R. BAJON, P.-H. GOUYON* and B.LEJEUNEt Luboratoire de Biologie dgktale: populations et communauth, *laborntoire d‘Emlutwn et Systhtique des Vigktaux zy and thboratoire de zyxwvutsrqpon Biologie rnol&culaire vkgktale, Bat. 530., Universitk Paris zyxwvu Xl, Centre d’Orsay, F-91405 Orsay Cedex, France Abstract The Holm complex in France consists of zyxwv two species, Holm lanatus L. (Yorkshire fog; 2n zyxwvutsrqp = zyxwvut 2x = 14) and Holcus mollis L (Creeping soft-grass; 2n = 4w = 28) and an interspecific hybrid H,m. x H.1. (2n = 5x = 351, which is morphologically similar to Holcus mollis. A heterologous rDNA probe from wheat was used to detect the corresponding region in Holcus (s. 1.) genomic DNA-fragments, for six to eight plants from 13 populations located south-west of Paris. A restriction enzyme map of the ribosomal RNA genes (rDNA) in Holcus (s. 1.) was also constructed. The length polymorphism detected in the IGS region was used as a DNA fingerprint for the identification of different cytotypes and species of the Holm complex and for the typing and delimitation of individuals in populations. In the light of the results we reconsider the assumption that the pentaploid hybrid H.m. x H.Z. is purely clonal. New hypotheses concerning the origin of the pentaploid hybrid and its reproduction are proposed, and the consequences for genetic diversity in natural populations discussed. Keywords: Holcus, hybridization, polyploidy, clonal plants, rDNA Received 21 March 1994; revision received 8 June 1994; accepted 13 lune 1994 Introduction The evolutionary diversification of many plant families has been greatly influenced by a combination of p l y - ploidy and hybridization (Stebbins 1950; Grant 1981). As a result, several families contain closely related groups of species and their hybrid derivatives. Often such groups include numerous and diverse forms which intergrade with one another in morphology and ecology, but not to the same degree as that found in normal polytopic spe- cies. Such assemblages are known as species complexes (Darwin, 1859; Vavilov 1951; Harlan 1975; Pernb 1984). In other families the groups are well differentiated. A major challenge in plant evolutionary biology is to exam- ine the precise relationships among such closely related species in order to determine the processes generating the contemporary patterns of variation. In this context h e application of recently developed molecular tools is of ‘particular importance (Soltis & Soltis 1993). Correspondence: Murielle Richard. Tel. 33 1 69 417 284. Fax 33 1 69 417 353. An interesting example is the Holcus complex, which contains markedly differentiated forms that appear to be closely related to ecological conditions. This complex is composed of two species, the diploid H. lanutus L. (York- shire fog, 2n = Zr = 14) and the tetraploid H. mollis L. (Creeping soft-grass, 2n = 4x = 28) and an interspecific pentaploid hybrid, H.m. x H.I. (2n =5x=35), which is morphologically similar to Holcus mollis (Jones 1958). zy So far, the tetraploid and the pentaploid have been consid- ered as a single species H. mollis (s. 1.). The caespitose spe- cies H. lanatus reproduces mainly sexually’ (i.e. seed), whereas H. mollis, a rhizomatous species, has been de- scribed as a clonal plant with no alternative to vegetative spread (Ellstrand & Roose 1987). More precisely, Jones (1958) said that the pentaploid reproduces almost exclu- sively by vegetative means and that the tetraploid H. moltis, although fertile, is likely to reproduce mainly by rhizomes. In this study we address two questions related to the H. lanatus - H. mollis complex and the origin of the pentaploid ‘species’. First, is the pentaploid’s mode of re- production exclusively vegetative or does sexual repro-