Origins and Evolution of Subfam. Maloideae (Rosaceae)

Abstract

Subfam. Maloideae (x = 17) has been held to be tetraploid since 1931 while later, Sax elaborated by Stebbins, specifically implicated subfam. Spiraeoideae (x = 9) and Amygdaloideae (x = 8) in an allotetraploid ancestry. The allotetraploid theory has gained considerable credibility with many workers on Rosaceae but, although support for it is not unanimous, and it is sometimes ignored, there has never been any attempt at refutation. To date, the theory has depended mainly on chromosome constitution, and, in a rather general way, on the morphological characteristics of the respective taxa. In this paper we review the allotetraploid hypothesis and generally verify it by an extensive character analysis of the Maloideae, Amygdaloideae, and Spiraeoideae. Since no specifically rosoid or quillajoid characters exist in the Maloideae, no candidate near-ancestral genera in subfam. Rosoideae or tribe Quillajeae can be pinpointed. The autapomorphic fruit of the Maloideae-the pome-is derivable from spiraeoid or proto-amygdaloid characters. In a parsimony analysis of a 96 taxa x 36 variable data matrix the phenetic genera were, for the most part, returned as clades, but their placement was to a significant extent, peculiar. The consistency index was extremely low. Furthermore, the addition of only three OTUs resulted in major relocations of some of the genera. We concluded that, in this type of situation, parsimony analysis is not robust, and sought other explanations for the relationships suggested. We favour the idea that several clades of Maloideae originated independently in a highly reticulate system existing shortly after the original allotetraploid cross. This notion of polychotomous early evolution alone, we feel, makes sense of the general correspondence of phenetic genera to cladistic genera, remarkably low consistency index, great instability of the cladogram with respect to changing only a few OTU's, and the manifestly great ability of most maloidean genera to hybridize. Thus, the precise phylogeny of the Maloideae must remain quite unknown until a very different kind of investigation has been successfully conducted. The present paper is one of a series by the authors reviewing generic limits and examining evolutionary pathways in the subfam. Maloideae. We develop the Sax-Stebbins hypothesis on the allotetraploid origin of the Maloideae and seek to define more clearly the characteristics of the original maloid in order to inform the search for evolutionary pathways. The Maloideae are an important group of woody plants containing close to 1000 species and 20-30 genera dependent on circumscription (Phipps et al. 1990). As one of the four subfamilies of the Rosaceae, which is often considered quite pivotal in dicotyledonous evolution, Maloideae have received a great deal of taxonomic and revisionary attention since the time of their first formal recognition by de Jussieu in 1791. However, little of this attention has been explicitly cladistic. Subfamily Maloideae constitutes a natural group of plants, with synapomorphic pome fruit, base chromosome number of x = 17 and an autapomorphic suite of rust parasites (Savile 1979). The widespread occurrence of gameto-