Abstract:
This thesis is an attempt to provide an objective quantitative ecological interpretation to the trends and aggregations of the flora within representative portions of the so-called Nigerian Northern Guinea Savanna Zone, using ordination and classification. The aim of the investigation was to test the homogeneity of the plant distribution on which the delineation of the ecological zone had been established. My field surveys were preceded by preliminary stratifications of the environment into relatively homogeneous domains, partly based on geomorphological differences in the environment, using aerial photographic interpretations, and partly based on visible distinct forms of land use. 419 species were encountered in 714 quadrats. I used a standard agglomerative Cluster Analytic technique available in the Oxford University Computing Laboratory to sort plant communities from the floristic data matrices which I had already arranged into sets of synoptic matrices. The Cluster Analytic results were cross-checked with Rao’s Canonical Factor Analysis which I introduced into synecological investigations as an ordination method from the Statistical Packages for the Social Sciences of Nie, et al (1975). Fifty relatively distinct communities of different sizes and four sub-communities were detected by the two multivariate analytic methods. The communities represent relatively distinct floral assemblages. Nine primary factors controlling floral distribution, apart from the climate, were identified, largely by quantitative methods. I conclude that the ‘Northern Guinea Savanna’ has no ecological homogeneity apart from climate, and its use alone as a reason for delimiting the zone from the rest of the Country’s vegetation is inconsistent with the very bases of ecology. I suggest ‘cerrado’ as a comprehensive substitute for “the catachrestic word, savanna, which is applied to nearly every type of tropical vegetation from swamps to
ORDINATION OF NIGERIAN GUINEA SAVANNA PLANT COMMUNITIES
