9510015 Jacobs The proposed research aims to define the statistical relationships between modern leaf form and climate in equatorial Africa, and from them, to estimate paleoclimate for middle to late Miocene (14 to 5 Ma) leaf assemblages from Kenya. To achieve these goals, herbarium sheets of species from 30 living plant communities across Africa will be videotaped and leaf images will be analyzed using Optimas. Information on leaf area, length, breadth, perimeter, and margin (entire or not, presence or absence of compound teeth, presence or absence of drip tip) will be recorded. Monthly climate data obtained from published sources, the U.S. Data Center, and the Ulited Nations Environmental Program Global Resource Information Database (GRID) will be mapped as contours on a base map of Africa using GSMAP and related utilities. Climate data for each plant site will be read from the map and interpolations will be checked by consulting local station records where possible. The leaf and climate data will be analyzed by appropriate statistical methods depending upon distributions and relationships among the variables. For example, multivariate methods such as correspondence analysis and canonical correspondence analysis, which are intended for nonlinear data, may be appropriate. The resulting statistical relationships will be calibrated. The calibration will be tested for accuracy by applying it to modern sites not included in the analysis. If climatic parameters of modern assemblages are estimated accurately, then the technique can be used to reconstruct climate for Miocene leaf assemblages. The middle to late Miocene is a significant time interval because the human family originated in Africa during that time. Further, modern atmospheric circulation patterns, most importantly the Asian monsoon, are thought to have originated in the Miocene (Quade et al., 1989; Prell et al., 1992). Monsoonal circulation is, in part, responsible for the distribution of modern plant communities, including savannas and grasslands. The origin of these biomes, perhaps coincident with the origin of the monsoons and the origin of hominids, remains a controversial subject (Retallack, 1992; Cerling et al., 1991; Kingston et al, 1994; Wheeler, 1991; Lovejoy, 1981). Similarly, hypotheses that explain disjunct distributions of modern forest species in East and West Africa suggest fragmentation of middle Miocene continuous forest as a causal factor (Lovett, 1993). The paleobotanical record for this time period offers a direct test for these hypotheses. In addition, the fossil record of most equatorial plant species is poor. The study of paleobotanical sites offers an opportunity to make contributions regarding phylogenetic relationships and biogeographic history of the living flora.
