Does Phylogenetic Distance Between Parental Genomes Govern the Success of Polyploids? Richard J.A. Buggs, 1 * Pamela S. Soltis, 2,3 Evgeny V. Mavrodiev, 1,2 V. Vaughan Symonds, 1 { and Douglas E. Soltis 1,3 1 Department of Botany, University of Florida, Gainesville, Florida 32611 2 Florida Museum of Natural History, University of Florida, Gainesville, Florida 32611 3 Genetics Institute, University of Florida, Gainesville, Florida 32610 ABSTRACT It has long been suggested that phylogenetic divergence between parental species determines the likelihood of their producing a successful polyploid, with closely related parents less likely to form a successful polyploid than more divergent parents. This suggestion has been based partly on observation of patterns of polyploid ancestry and partly by extrapolation from analyses of the processes that give rise to polyploids. Here we present a new survey of the patterns of the divergence between the parents of polyploids, based on node- based and clade-based analyses of phylogenetic trees. We also use the topology of the phylogenetic trees to inform a null expectation of the distance between parental species, assuming random crossing between all species pairs in a genus. We used molecular phylogenies now available for eight plant genera containing multiple polyploids whose parentage has been investigated: Tragopogon, Persicaria, Brassica, Leucaena, Spartina, Spiranthes, Nicotiana, and Glycine. We found that the phylogenetic distance between progenitors of polyploids did not differ significantly from the null expectation. In contrast, the distance between parents of diploid hybrids (both stable and unstable) in these genera were lower than would be expected with random crossing. We discuss how these findings may fit with recent progress, through genetic and genomic studies, in understanding the processes involved in polyploidization. INTRODUCTION During the past two de- cades, molecular tools have provided unprec- edented insights into the genetic and genomic consequences of polyploid speciation. Poly- ploids are now known to be more common, with more origins and with more dynamic genomes than previously appreciated (re- viewed in Liu and Wendel 2003, Osborn et al. 2003, Soltis et al. 2004b, Wendel and Doyle 2004, Rapp and Wendel 2005, Soltis et al. 2007, see also Kashkush et al. 2002, Bowers et al. 2003, He et al. 2003, Blanc and Wolfe 2004, Wang et al. 2004, Moore and Purugganan 2005, Buggs 2008). This recent progress in our understanding of polyploid evolution prompted us to re-visit a longstand- ing hypothesis regarding genetic factors gov- erning the formation and success of polyploid species. Early observations showed convinc- ingly that successful polyploid hybrids (allo- polyploids) could be derived from more distantly related parental species than could successful homoploid hybrids (Digby 1912, Clausen and Goodspeed 1925). Based on the views of polyploidy current in their day, several earlier researchers went further and suggested that a high level of differentiation between parental species is highly advanta- geous to polyploidization: they suggested that less-differentiated parental species (and also parents within the same species, producing autopolyploids) are less likely to produce successful polyploids than distantly-related parents. We review these ideas briefly below. Darlington (1937) formulated a general rule that there is an inverse relationship between the fertility of a diploid hybrid and that of a tetraploid hybrid (i.e., an allotetra- ploid) of the same parentage, so parental species whose hybrids exhibit high fertility are likely to produce an allopolyploid of high {Present address: Institute of Molecular Biosciences, Mas- sey University, Palmerston North, New Zealand. Received November 23, 2007; Accepted April 3, 2008. *email address: buggs@ufl.edu CASTANEA 73(2): 74–93. JUNE 2008 74