The two subspecies of Ulex mn$mus L. (subsp. eumpms and subsp. latebracteatus) are easily differentiated by the shape and size of their bracteoles, and by their chromosome number. However, in the north-west of the Iberian Peninsula, the areas of both subspecies are not clearly delimited, the shape of the bracteoles is intermediate and the width is close to the diagnostic limit. Thus, allocation of some specimens to either subspecies is diffcult. Recently, a controversy has arisen concerning the intrapopulational uniformity of the chromosome number in U. mmpaezu from west France and north Spain, and the usefulness of ploidy level and chromosome number as clear-cut indicators of the two subspecies. Thus, the merits of recognizing the subspecies as independent are re-analysed here. Our results for north-west Spain show that the chromosome number is uniform and itable within populations and is clearly related to the morphological characters. Correlation and discriminant analyses show that bracteoles and stomata sizes are highly correlated and are good predictors of ploidy level. Furthermore, the results support the maintenance of two subspecies in U. wmpam, which can be delimited by a combination of morphological characters, chromosome number and geographical distribution. 0 1997 The Linnean Society of London