Contributions to the cytotaxonomy of the Commelinaceae*

Further material of Gibasis geniculata (Jacq.) Rohw. (syn. Tradescantia geniculata Jacq.) and other Gibasis species collected in the field has been studied. The existence of several species in our earlier experimental material is confirmed. These include G. geniculata itself (2n= 32 or 48 small chromosomes), G. oaxacana D. R. Hunt (2n= 16 small chromosomes) and G. schiedeana (Kunth) D. R. Hunt, which has two chromosome forms, 2n= 10 and 2n= 16, both with large chromosomes. These forms are diploid and tetraploid based on x= 5 and x= 4 respectively and show a Robertsonian relationship with each other. The cytology of tetraploid (2n= 20) G. karwinskyana is confirmed and that of a diploid form (2n=10) described. The recently described G. consobrina D. R. Hunt (2n= 20) is shown to be cytologically comparable with G. karwinskyana, but to differ in significant details. Next, the cytology of G. pulchella (Kunth) Rafin., the type species of Gibasis, is described and also that of the allied G. matudae D. R. Hunt. Both G. pulchella (2n= 10, 15) and G. matudae (2n= 10) show interchange heterozygosity, with complete rings of ten chromosomes at meiosis in some plants of G. pulchella. Preliminary comments are made on the cytology of G. aguensis (Standl. & Steyerm.) Rohw. (2n= 10), another ally of G. pulchella, and on the G. linearis group, where the basic number appears to be x= 6 but two karyotype patterns have been found. A discussion of chromosome architecture in Gibasis in relation to the taxonomy of the genus concludes the paper.