Pollen studies, especially those concerned with size frequency, have received more attention fronm the paleobotanist than from the systematic botanist. This interest is due to the limited paleoecological information that can be gained by knowing only the genus of the fossil pollen. Generic identifications have permitted some inferences about major conditions, but more refined evaluations depend on specific identifications of pollen. And, generally, finding qualitative similarity of pollen within a genus (Erdtman 1952), workers have turned to pollen size and number of apertures as a key to the identification of species. Specific identifications of fossil pine pollen have been attempted by correlating size-frequency data with that of extant species (Cain 1940; Cain and Caill 1948). Generalizations based on studies of this kind have led to some rather profound paleoecological interpretations (Whitehead 1962b). In recent years, however, many questions have been raised as to the reliability of size-frequency studies that are beset by problems in sampling, treating, and mounting modern and fossil pollen (see Buell 1946; Faegri 1960; Whitehead 1962b). Assuming that the sampling problems can be overcome, studies oln the pollen of modern hickory and other genera well represented in Tertiary and Pleistocene sediments could prove valuable to the paleobotanist and systematic botanist. Both Whitehead (1960; 1962a) anid Stachurska (1961) have surveyed the morphology of the pollen of Juglandaceae, but they did not attempt a comprehensive statistical analysis of Carya. This study of pollen size in Carya could serve as a standard for comparison with fossil grains inasmuch as all extant species are included, except Carya cathayensis Sarg., for which staminate flowers are not knlown. Results of this work already have proved of value in supportingo earlier findings on the correlation between stomatal size and the ploidal level (Stone 1961) and have been partially successful in delimiting species or groups of species. Selection of pollen samples, the technique of preparing the pollen for observation, and to a lesser extent the method of measuring the grains are all factors which must be controlled and standardized. Proper selection of pollen samples requires adequate coverage of geographic areas and environmental situations in which a species occurs, as well as consideration of phenotypic variation in pollen size within a single tree. In this study a large sampling of trees from the southeastern United States is included. All told, however, 16 U. S. states, northern Mexico, and southeastern Asia are represented. Mimeographed sheets with the collection data of annotated specimens are available on request. I wish to thank Dr. Delzie Demaree and Professor Wayne Manning for their help in securing some of the rare species and Dr. Donald Whitehead for offering helpful suggestions in preparation of the manuscript. Appreciation for the loan of pollen material is acknowledged to curators of the following herbaria: British Museum (Natural History), Museum National D'Histoire Naturelle (Paris), Florida State University, University of Florida, Harvard University, Louisiana State University, Missouri Botanical Garden, University of North Carolina, University
Read full abstract