Crocus olivieri and Its Allies (Iridaceae)

Abstract

Summary. Crocus olivieri Gay belongs to a small group of closely allied species occurring in the Balkans and Asia Minor. A key to these and their subspecies is provided and the relationship between the species is discussed. The genus Crocus, although containing approximately one hundred specifically quite distinct taxa, is difficult to divide into clear infra-generic groups. There are however some aggregates of species, such as that to be discussed in this paper, which are definable. A study of the genus reveals the existence of several such groups or sections, and also a number of individual species which do not readily fit into any of these and which are best treated as isolated species with no close affinities. Since practically all the morphological characters used in distinguishing between species or groups of species are not reliable in every respect, it is necessary to adopt those features which seem to be most stable, have the most significance and are of practical use in identification. Thus, the genus can first be divided into two major components, which were called 'Divisions' by Maw (1886), these being based on the earlier three 'Divisions' of Herbert (I847). The critical factor here is the absence or presence of a membranous prophyll subtending the pedicel, at the apex of the young developing corm. This was referred to as the 'involucre' by Herbert and as the 'basal spathe' by Maw and by Bowles (1924). In most Crocus species the prophyll is hidden within the sheath formed by the true leaves and the outer scale-like cataphylls or sheathing leaves. Maw placed the species possessing a prophyll in his Division I, Involucrati, and those in which it is absent in Division 2, Nudiflori. Herbert had previously recognized a third Division, the Subnudi, to house two species which he considered variable in being either without a prophyll or having a poorly developed one. Maw discounted this and included these two species in the Involucrati. This seems to be justified by further studies of living material. Both Goldbach (1817) and Gay (1831) had previously used this method of dividing the genus, although at that time less than 26 species were known. For groups subordinate to these major divisions, both Herbert and Maw resorted to characters of the corm tunic to produce a number of 'Sections', depending on whether the tunic consisted of reticulate, parallel or interwoven fibres, was non fibrous and papery or coriaceous, or whether it split horizontally at the base into rings of tissue ('annuli'). These characters are of great value in defining a species or sometimes groups of species but have to be used with caution as infra-generic criteria. The differences between parallel and reticulate fibres for example, can become very slight and two obviously closely related species could be placed in different sections of the genus if the division between those sections is too rigid. J. G. Baker (1892: 76) devised an entirely different classification based on the degree of division of the style. Here again, although of the utmost