A biosystematic study of the Digitaria leucites complex in North America (Poaceae)

Abstract

Anatomical, morphological and cytological studies were made of the Digitaria leucites complex. Materials studied were herbarium specimens from various herbaria and the additional collection of D. runyonii Hitchc. and D. texana Hitchc. from throughout their geographical area. Based on these studies ten species are recognized: D. albicoma Swallen, D. distans (Chase) Fern., D. ekmanii Hitchc., D. leucites (Trin.) Henr., D. pauciflora Hitchc., D. runyonii, D. simpsonii (Vasey) Fern., D. subculva Hitchc., D. texana, and D. simpsonii "Curtiss". The last taxon is a new entity first encountered in this study. First glume and the spikelet vestitures were examined using scanning electron microscopy. The investigation revealed considerable variations in size and shape of the first glumes and the distribution of macrohairs, bicellular microhairs, prickles and papillae on the second glumes and the lower lemmas of the taxa studied. However, these characters were of diagnostic value for D. pauciflora, D. simpsonii, D. subculva and D. simpsonii "Curtiss". The leaf epidermes and anatomy studied provided some characters which could be used in distinguishing the species of D. leucites complex. Keys for identification of the species based on leaf epidermis and leaf anatomy are given. The diagnostic characters were the number of epidermal cells at the base of macrohairs, size and shape of apex of distal cells of microhairs, presence of the intercostal prickles or hooks, shape of the keel, bundle sheath extension nature and the distribution of adaxial sclerenchyma strands associated with the bundles and the maximum lateral cell count. Chromosome numbers of D. runyonii and D. texana were both found to be 2n=54. The correlation of pollen size and pollen viability with chromosome number are presently unknown in this complex. However, it is speculated that the rare and endemic species D. pauciflora and D. subculva had high male and female sterility. Morphological data were analysed using principal component analysis, cluster analysis, and canonical discriminant analysis. A total of twenty-six characters were used in these analyses. Important discriminating characters found in each analysis were given. The spikelets size played an important role in separating most taxa. High intraspecific variability was found in D. leucites...