Plant collections, field and herbarium studies were used to evaluate classifications of Impatiens. Accepted interpretations of the perianth are questioned; and a program of developmental studies on several categories of inflorescences proposed. The fruit has been incorrectly described. It is of two types, differing in manner of dehiscence. Chromosome numbers are an inadequate taxonomic character because of parallel cytoevolution. Karyotypes were prepared, but it was found that homologues were not distinguishable. A technique for Hy-banding was developed, and homologues in plants of I. leschenaultii Wall. identified. It is proposed that Hy-banding can identify marker chromosomes which will be a more efficient character than complete karyotypes. The classifications of Impatiens are discussed, and it is concluded that one founded on flower-pollinator relationships is the best, although convergence is a problem. The genera of the Balsaminaceae are reviewed. It is concluded that these show only slight deviations from Impatiens, and are rejected.

No description available.