Forty selections, including 37 cultivars of Hamamelis spp., were evaluated for genetic similarities using randomly amplified polymorphic DNA (RAPD) markers. Cluster analysis identified seven groups, which included three groups of H. ×intermedia cultivars, two groups of H. vernalis, and one group each of H. mollis and H. japonica. Three H. ×intermedia cultivars, `Arnold Promise', `Westerstede', and `Carmine Red', did not group closely with the other 20 cultivars of H. ×intermedia. Selections of the North American species H. vernalis were quite distinct from the Asiatic selections. However, data are presented that suggest hybridization exist between Asiatic Hamamelis spp. and H. vernalis. Genetic similarities between known half-sib families provides evidence that the cultivar pairs `Arnold Promise'—`Winter Beauty' and `Carmine Red'—`Hiltingbury' are, themselves, not likely half-sibs.
Robert D. Marquard, Eric P. Davis and Emily L. Stowe
Robert D. Marquard, Charlotte R. Chan, Eric P. Davis and Emily L. Stowe
Numerous isozyme systems were found to be polymorphic in witch hazel (Hamamelis spp.). However, aconitase (ACO), malate dehydrogenase, phospho-glucose isomerase (PGI), and phosphoglucomutase were most useful for identification of cultivars. From these enzyme systems, three genes were identified that control patterns of ACO (2) and PGI (1). Isozymes can be used to help verify cultivars and their simple inheritance could be useful to validate hybrids and gene flow between plants. DNA was readily extracted from young leaf tissue after grinding in liquid nitrogen and extraction in warm CTAB. DNA was amenable to amplification using polymerase chain-reaction technology. Primers (400) were screened to identify polymorphic RAPD bands. Ultimately, 19 primers were used to generate 68 RAPD markers that were reproducible. Cultivars were scored for presence or absence of the 68 markers. Genetic similarities were calculated using a Nei coefficient and clustering was conducted for more than 40 cultivars using a UPGMA program. Arbitrarily, the cultivars were assigned to seven groupings after cluster analysis. The seven classes gave one group each of H. japonica and mollis; two groups of H. vernalis; and three groups of H. ×intermedia. Clustering allowed some interpretation about relatedness among cultivars and genetic similarity data helped assign some cultivars to a particular taxa that were previously in question.