Odontonema Nees.

— Odontonema consists of 29 species restricted to tropical and subtropical regions of the New World. Although Sarkar et al. (1980) reported n = 17 for O. bracteolatum (Jacq.) Kuntze (as Thyrsacanthus bracteotus Nees), chromosome counts of n = 21 have been reported for five other species of the genus (Takizawa 1957, as Thyrsacanthus rutilans; Daniel 2000; Daniel et al. 1990). This latter number is also common to relatives of Odontonema in Justicieae: Odontoneminae sens. str. (e.g., Chileranthemum, Oplonia and Pseuderanthemum, among genera from the New World; see above under Graptophyllum). Indeed, De (1966) noted both common chromosome numbers and karyotypic affinities between species of Pseuderanthemum and Odontonema. Here, we report approximate counts for O. callistachyum (Schltdl. & Cham.) Kuntze (Mexico and northern Central America) of n = ca. 63 and n = ca. 42–50 and for O. tubaeforme (Bertol.) Kuntze (Mexico and Central America) of n = ca. 20–30. There are no previous reports of a chromosome number for O. callistachyum; however, there is at least one previous report of n = 21 for O. tubaeforme (Daniel et al. 1990). Although both species are commonly cultivated, our study plants were wild-collected.

Given the likely basic number of x = 21 for Odontonema, and based on these approximate counts, the plants sampled for O. callistachyum would appear to be polyploids. Difficulties in obtaining an accurate count for these plants resulted from clumping of chromosomes, darkly staining cytoplasm, and/or meiotic irregularities (e.g., univalents, bivalents, and trivalents at metaphase I and anaphase I; lagging chromosomes). Similar meiotic irregularities are sometimes associated with hybrids, triploids, and/or sterility (e.g., Long 1966). Indeed, the approximate number for Daniel & Bartholomew 5013 is suggestive of triploidy. The very low percentage of non-staining pollen for both collections of O. callistachyum (Daniel 5368: number of grains = 377, 1% not staining; Daniel & Bartholomew 5013: number of grains = 97, 3% not staining) is similar to that for the sample of O. tubaeforme for which Daniel et al. (1990) reported n = 21 (Daniel & Bartholomew 4991: number of grains 205, 3% not staining). Thus, based on pollen stainability, the two collections of O. callistachyum with apparent meiotic irregularities would not appear to be sterile, and thus probably neither triploids nor interspecific hybrids.