A Retrospective View of Old-field Succession after 35 Years

In 1947 I began my doctoral studies to determine the causes of early old- field succession in the North Carolina Piedmont. I present here a review of the research that led me to undertake those studies and a synopsis of the research on patterns and mechanisms of old-field succession which followed. Although a few species are ubiq- uitous among old fields, the patterns and rate of species replacement vary considerably among and within regions. In few areas are stages as distinct or progress so rapid as in North Carolina. I expect the mechanisms are also varied and suggest that factors such as site history, farming practices and seed availability have been underrated. In general, much of the variation in old-field composition appears to be related to varia- tions among species in germination requirements and life histories. PRE- 1947 RESEARCH In the autumn of 1947, as a graduate student at Duke University, I began working on the causes of old-field succession in North Carolina. The sequence of species in- vading and dominating abandoned farmland in North Carolina had been determined by Crafton and Wells (1934) and Oosting (1942) with little disagreement. Crabgrass (Digitaria sanguinalis (L.) Scop.) usually occupies fields in the late summer and autumn following the last cultivation. Horseweed (Erigeron canadensis L.) dominates fields the 1st year of abandonment, followed in turn by aster (Aster pilosus Willd.) and broomsedge (Andropogon virginicus L.) in the 2nd and 3rd years. Broomsedge fields are invaded by pines of several species, which usually form a complete canopy in 10 or 15 years. After one generation pines are replaced by hardwoods. Dr. Oosting admitted (pers. comm.) that although he knew the sequence of species in early old-field succession, he did not know what caused these distinct and fast changes and challenged me to find the causes. Clements (1916) had proposed that pioneer species change the environment and make it more suitable for species that invade by increasing the humus in the soil, pro- viding shade and improving water and mineral conditions. This idea was widespread and generally accepted until the 1940s as the major cause of succession. Crafton and Wells (1934) found that one of the causes of old-field succession is competition for water. Coile (1940) and Oosting and Kramer (1946) had investigated the role of water and light in the replacement of pines by hardwoods. At about the same time the idea that one plant may produce some chemicals which are inhibitory or toxic to other plants (allelopathy) had been advanced by Davis (1928), Went (1942), Bonner and Galston (1944) and others. Such toxins could be the cause of one species replacing another in the old-field sequence. Perhaps horseweed produces a toxin that delays the invasion of aster, or aster produces a toxin that helps eliminate horseweed from the old- field succession in North Carolina. Drew (1942) worked out the sequence of species on abandoned farm fields in Missouri and speculated on the causes of variations from field to field. He concluded that the only event that correlated with the differences was what the last crop had been. From my research (Keever, 1950) I concluded that the early dominants are largely determined by the capability of their seeds to germinate at the time of year a suitable habitat is available. After the germination of seeds, the entrance and exit of each