![Thumbnail: [Volume 79, 1951 page 387]](/journals/rsnz/images/rsnz_79/thumbnails/rsnz_79_03_0573_0387_ac_02.gif)
Origin and Evolution
Thelymitra appears to have originated in the south and to have moved northwards into Tasmania and Australia at a comparatively early date. In contrast to Pterostylis (cf. Trans. R.S.N.Z., 77, 1949, 234) there seems to have been no northward movement into New Zealand. The local species can all be traced back, directly or indirectly, to an Australian ancestry, and were probably windborne originally across the Tasman. Until the Australian jordanons are dealt with as a whole, coherently described and adequately illustrated, it is impossible to define the phylogeny of half the New Zealand species. In a generic revision Thelymitra would be divided into two subgenera, and these again into a number of superspecies in the sense of Mayr. (Systematics and the origin of species, New York, 1942). Such delimitation is not advisable in a paper of this nature, so the subgenera have been provisionally called sections and the superspecies called groups, in anticipation of an eventual revision of the whole genus.
Section (i)—in which the column-wings consist of two primary lateral lobes, large, more or less erect, usually simple, but occasionally fimbriate.
![Thumbnail: [Volume 79, 1951 page 388]](/journals/rsnz/images/rsnz_79/thumbnails/rsnz_79_03_0574_0388_ac_02.gif)
Section (ii)—in which the primary lobes have been welded together into a large midlobe which usually carries two secondary lobes, either ciliate or variously crenulate.
Section (i) appears to be the more primitive and often has a complicated crest in the sinus of the lobes. Section (ii) has possibly originated from (i) by the gradual expansion of this crest, filling the space between the primary lobes and giving rise to the tubercles and calli which are to be found on the midlobe of the more primitive members of the second section. Several species (Th. pulchella the only one in New Zealand) show intermediate characters between the two sections. In section (ii) the erect, non-ciliate form of the secondary lobes to be found in group (d) is probably more primitive than the elaborately ciliate condition of groups (e), (f) and (g). The erect midlobe of group (f) is probably more primitive than the culcullate condition of group (g). The non-ciliate secondary lobes in T. sanscilia are retrograde rather than primitive and the species appears to be derived from a ciliate ancestor. Thelymitra has affinities and possibly a common ancestry with Calochilus, but very far back. In its simple, regular perianth and divided column Thelymitra shows decidedly primitive characters.