Search

The angiosperm family Rosaceae poses a number of noteworthy systematic problems as well as many questions concerning morphological and chromosomal evolution. Phylogenetic analysis of rbcL gene sequences was performed to address systematic and evolutionary problems of Rosaceae. Both rbcL sequence variation and the presence of duplicated sequences near the 3' end of rbcL were useful in determining phylogenetic relationships in this family. Analyses of rbcL sequences indicate that there are groups of genera within Rosaceae comparable to the subfamilies Maloideae, Amygdaloideae, and Rosoideae, although the composition of each group differs from traditional circumscriptions. According to analysis of rbcL data, Maloideae and Amygdaloideae each include additional taxa not normally associated with them. All members of Rosoideae with x = 9 are phylogenetically well separated from the x = 8 and 7 members of the subfamily. In addition, Spiraeoideae are not monophyletic but appear to consist of several distinct evolutionary lineages. The rbcL-based phylogenies suggest that chromosome numbers are more reliable indicators of some generic alliances than the more commonly used fruit types. Sequence data are also useful in determining the alliances of several problematic genera, suggesting that the capsular and follicular-fruited genera Vauquelinia, Lindleya, and Kageneckia (usually placed in Spiraeoideae) should be included in the pome-fruited subfamily Maloideae, and that Quillaja is not a member of Rosaceae. Molecular data are consistent with several suggestions for the ancestral chromosome numbers and fruit types of Rosaceae, but do not support any one hypothesis for either. This study also suggests that the subfamily Maloideae may have descended from spiraeoid ancestors and that the pome is derived from follicular or capsular fruit types.

Nuclear DNA content has been estimated using flow cytometry for 17 species and eight cultivars of Malus and for 44 species of 29 other genera within the Rosaceae. Compared to other angiosperms, diploid genome sizes vary little within the family Rosaceae and within the genus Malus. C-values of genera within the subfamilies Spiraeoideae and Rosoideae are among the smallest of flowering plants thus far reported. In general, the Maloideae have the largest diploid genomes of the family, consistent with their higher chromosome numbers and presumed polyploid origin. The Rosaceae, including such economically important plants as almond, apple, strawberry, and rose, is considered a natural group held together by similarities in floral structures. The four subfamilies are defined by fruit type (Robertson, 1974), and each includes polyploid series with fairly consistent chromosome base numbers: Spiraeoideae (x = 9); Amygdaloideae (x = 8); Rosoideae (x = 7, 8, and 9); and Maloideae (x = 17) (Sax, 1931, 1932). The subfamily Maloideae, with its relatively high base chromosome number, has been hypothesized to be either of autopolyploid (Darlington and Moffett, 1930) or allopolyploid origin (Sax, 1931, 1932; Stebbins, 1950). Although numerous chromosome numbers have been reported for Rosaceae, the amount of DNA per nucleus (C-value) has been reported for only 14 species of the family (Bennett and Smith, 1976, 1991; Bennett, Smith, and Heslop-Harrison, 1982; Arumuganathan and Earle, 1991b). Apart from the utility of genome size data in ongoing molecular studies in this important plant family, the amount and distribution of nuclear DNA content variation among related taxa may give insights into genomic evolution that underlies or parallels speciation (Raina and Narayan, 1984; Ohri and Khoshoo, 1986; Price, 1988). In this study, flow cytometry was used to estimate nuclear DNA contents of 28 genera from each of the four Rosaceae subfamilies. Compared to Feulgen densitometry or reassociation kinetics, flow cytometry is a rapid and reliable method for estimating C-values in plants (Galbraith et al., 1983; De Laat, Gohde, and Vogelzang, 1987; Raybum et al., 1989; Raybum, 1990; Michaelson et al., 199 la). We here report low levels of nuclear DNA variation within the Rosaceae and find that Spiraeoideae C-values are among the smallest of angiosperms. Relatively large C-values of Maloideae support the polyploid origin of the subfamily.

A B S T R A C T The genus Filipendula Mill. is generally separated from Spiraea L. in systematic keys on the basis of a single fruit character. In some taxonomic treatments of the Rosaceae, where subfamilies are used, this places the genera in separate subfamilies. Karyological studies can be useful in assaying the justifiability of such treatment and are needed because of serious discrepancies between previous reports of chromosome numbers and the recent textbook designation, on dubious grounds, of F. vulgaris as an example of a "permanent chromosome hybrid." The results given in this paper show that x = 7 in this genus (compared with x = 9 in Spiraea) and the reasons for rejecting previous counts of 2n = 15 for F. vulgaris are presented. "Permanent chromosome hybridity" for this species is also rejected. The possibility that a cytotype with 2n = 16 may exist in the northern part of the range of F. ulmaria cannot be completely discounted, but positive evidence is presented for 2n = 14 in this species (even though 2n = 16 has been reported most frequently recently). The basic number 7 for Filipendula is in agreement with the placing of this genus in the subfamily Rosoideae even though the hereditary peculiarities (apomixis and permanent chromosome hybridity) shown by some other members of this subfamily are apparently not now needed to explain the cytological situation in Filipendula.