The taxonomic affinities of the plants formerly-referred to Oncidium section Oncidium, (Orchidaceae) in the Caribbean Islands have been studied with respect to morphological, cytogenetical, biochemical and geographic data. The data was-examined employing numerical taxonomic methods. The study involved 57 taxa (28 taxa of the plants belonging to the orchid group studied and 29 related-species)-and a total of 63 attributes. The literature and all available herbarium specimens-concerning the plants studied were viewed critically. This resulted in the clarification of some problems of identification and taxonomy. The plants studied were reclassified into, the*genera Braasiella, Hispaniella, Olgasis and Tolumnia. For all but one species formerly referred to Oncidium section Oncidium, living material was obtained through purchase, donation or through personal collecting in the Caribbean Islands. Living plants were examined in respect to morphological and ultrastructural characters. For all but three taxa, the chromosome number (2n) was determined. The data obtained was compared with the data formerly reported. For nine taxa, the diploid chromosome numbers are reported for the first time. Field studies were performed in selected areas of Jamaica, Haiti and the Dominican Republic to determine some aspects of the pollination ecology, ecology and distribution of some of the taxa. -3- Leaf flavonoid fingerprinting was performed on material obtained from all 57 taxa studied, using two dimensional paper chromatography. All data obtained was collated using numerical taxonomic methods. It became apparent during this study that a number of previously described taxa needed clarification as far as their identity is concerned. Furthermore, it became apparent that the species formerly referred to, oncidium sect. Oncidium, differ from all other groups of the genus Oncidium sensu Garay and stacy, as well as from all genera within the Oncidiinae, in such a way as to warrant their separation as autonomous genera. These have been published by Rafinesque (1837), Braem (1980) and Braem et al. (1984). This work also shows the difficulty of gaining total objectivity in a systematic study. The main problems arise from the selection of attributes, and the difficulties experienced in the coding of the data.