New insights into the polyploid complex Cenchrus ciliaris L. (Poaceae) show its capacity for gene flow and recombination processes despite its apomictic nature Amina Kharrat-Souissi A,E , Alex Baumel B , Franck Torre B , Marianick Juin B , Sonja Siljak-Yakovlev C , Anne Roig D and Mohamed Chaieb A A Faculté des Sciences, Département de Biologie, Laboratoire de Biologie et d’Ecophysiologie des végétaux en milieu aride, Université de Sfax, Tunisie. B Institut Méditerranéen d’Ecologie et de Paléoécologie, UMR-CNRS/IRD 6116 IMEP – Bâtiment Villemin, Europole de l’Arbois – BP 80, 13545 Aix-en-Provence Cedex 04, France. C Université Paris-Sud UMR 8079, Département de Biodiversité Systématique & Évolution, Bâtiment 36091405 Orsay Cedex, France. D Centre INRA PACA, Domaine de Saint Paul URFM, Unité Écologie des Forêts Méditerranéennes, 84914 Avignon, France. E Corresponding author. Email: kharratsouissi@yahoo.fr Abstract. Cenchrus ciliaris L. is a C 4 perennial grass of arid lands which is under the focus of different ecological issues such as response to desertification, quality of forage grass and impacts of invasions. Here, molecular and morphological analyses of the genetic diversity of several Tunisian provenances of C. ciliaris were performed to better understand the phenotypic polymorphism of this agamospermous and polyploid grass. Ten phenotypic traits associated with productivity were measured in a common garden environment. Amplified Fragment Length Polymorphism (AFLP) markers were developed to investigate the structure of genetic diversity among and within provenances and between the three ploidy levels. Heritable phenotypic traits showed considerable differences within provenances. Surprisingly, AFLP markers revealed the existence of genotypic variations between individuals of the same sibship and a high G/N value (0.55). A neighbour-joining tree based on AFLP markers revealed three major groups; tetraploid, pentaploid and a mix of pentaploid and hexaploids. These groups do not correspond completely to the geographical origin of samples. The results underline the possibility of sexual reproduction, recombination and gene flow within and between populations of C. ciliaris. In respect with the well known dynamic nature of polyploid genomes, these results should have strong consequences for the future management of this grass for both conservation and invasion issues. Introduction The severe process of degradation undergone by many arid and semiarid ecosystems, commonly known as desertification, constitutes one of the main environmental problems occurring in the 21st century (Kassas 1995; Whitford 2002; Reynolds et al. 2007). The main causes of desertification in arid countries are heavy livestock grazing, severe climatic conditions and the ecological fragility inherent to many plant communities (Kassas 1995; Yates et al. 2000; Li et al. 2008). The fact that perennial herbaceous plants are the most severely affected by the above factors is accelerating the process of desertification (Valone et al. 2002; Jauffret and Lavorel 2003). The C 4 perennial Cenchrus ciliaris L. (Poaceae) is the predominant grass in many arid lands. Despite its large distribution in Tunisia from semiarid lands to the border of the Sahara desert, overgrazing has decimated the populations of C. ciliaris, and it now contributes little to natural plant communities. By contrast, within protected areas (such as the National Parks of Bouhedma and Sidi Toui in Tunisia) this species is still abundant and vigorous (Bousquet 1992; Chaieb et al. 1992). Paradoxically, since C. ciliaris is an excellent pasture grass, it has been introduced worldwide to improve range lands for cattle rearing or for erosion control (Brits et al. 2003). Following its introduction C. ciliaris has invaded extensive areas of arid and semiarid lands in Australia affecting fire regimes (Jackson 2004; Miller et al. 2010). In a changing world the evolutionary response will contribute to either persistence or expansion of native and invasive species (Parmesan 2006). For these reasons, several studies have focussed on the polymorphism of C. ciliaris, for ecological restoration purposes (M’Seddi et al. 2002; Visser et al. 2008; Kharrat-Souissi et al. 2010), for improving the grass forage quality for livestock (Arshadullah et al. 2011), or the resistance to drought (Mansoor et al. 2002; Arshad et al. 2007) and in the context of plant invasion success (Gutierrez- Ozuna et al. 2009). A key step towards understanding the CSIRO PUBLISHING www.publish.csiro.au/journals/ajb Australian Journal of Botany, 2011, 59, 543–553 Ó CSIRO 2011 10.1071/BT10312 0067-1924/11/060543