In spite of extensive studies using chromosome number, floral anatomy,
pollen morphology, and serology, substantial disagreement still remai
ns in the systematics of the Berberidaceae. We employed two different
molecular systematic approaches, gene sequencing and restriction site
analysis, to test competing hypotheses about intergeneric relationship
s in the family. Both rbcL gene sequence and restriction site data are
congruent in recognizing four major chromosomal groups (x = 10, 8, 7,
and 5). Fragmentation of the Berberidaceae into smaller families, esp
ecially separation of Nandina as a distinct family was not supported.
The large woody genera, Berberis and Mahonia, are related to the monot
ypic herbaceous genus, Ranzania. Bongardia is nested within the x = 6
clade, and has a remote relationship to the Leontice group with which
the genus has been placed for long time in most previous classificatio
ns. Although there are some major incongruences, our molecular systema
tic studies support the morphological classification proposed by LOCON
TE & ESTES (1989).