Similarity of the Flora of the Erie Islands: Implications for Conservation Biology

Article excerpt

Using a compiled list of the vascular flora for 21 islands in the western basin of Lake Erie, Preston's resemblance measure and Connor and Simberloff's Null Hypothesis I were applied to native and alien species subsets in order to examine how these findings fit within the equilibrium theory. Based on Preston's measure, it was found that the similarity of species between island pairs was more likely to fit with MacArthur and Wilson's equilibrium theory when native species were considered than when alien species were considered. Native species have restricted distributions -- especially those found on the smaller islands -- and appear to be less randomly distributed about the islands. This study also found that a large component of the alien flora is widespread in its distribution, and appears to be randomly distributed among most of the islands.

Key words: similarity measures, equilibrium theory, Lake Erie Archipelago, flora, conservation biology, island biogeography

Introduction

There are a number of numerical techniques common to the school of island biogeography as defined by MacArthur and Wilson (1967). The basic techniques center on correlation and regression analyses involving species numbers, area, and other variables, the selection of which depend upon what the investigator is attempting to test (e.g., McNeil and Cody 1978; McCord 1980; Quinn et al. 1987). Similarity analysis (Crovello 1981) is less commonly applied by members of the MacArthur and Wilson school, but is a common method of biogeography in general. In this study, the similarity of the floras of the islands located in the western basin of Lake Erie is examined using two biogeographic-specific similarity measures: Preston's (1962) resemblance measure, and Connor and Simberloff's (1978)`Null Hypothesis I.'

The islands, located in the western basin of Lake Erie -- with Ohio State lying to the south and the Province of Ontario lying to the north -- provide an excellent opportunity to study a wide variety of biogeographic theories because so much is known about their natural history (Langlois 1954; Downhower 1988). Several aspects of the MacArthur and Wilson model of island biogeography have already been applied to the islands' flora (Klinkenberg 1983; 1988; 1992), and the results of that work provide some insights into the species-area relations and the distributions of native and alien species. Those results show, for example, that the effect on species numbers by factors such as the level of disturbance and the perimeter of each island were different for different floristic components (i.e., native versus alien species), and that collecting effort -- as represented by the distance of each island from the biological research station located on Gibraltar Island -- played some role in the observed numbers of species found on each island. However, because the previous species-area analyses only considered species totals and ignored both the similarity of the species found on each island and the inter-island differences, the following analyses were undertaken using Preston's and Connor and Simberloff's similarity measures.

Preston's `resemblance equation' -- or similarity measure, if the complement of the resemblance measure is taken -- was developed specifically by Preston (1962) to provide a quantitative statement of similarity that explicitly takes area into account, and also takes into account the interchange of species assumed in the equilibrium model of island biogeography. Possibly because it is a transcendental equation and, therefore, cannot be computed as easily as can the simple matching coefficient (Sneath and Sokal 1973), for example, Preston's similarity measure has rarely been applied. Nonetheless, its qualities are such that it deserves a closer look.

Preston's similarity measure was derived from the power model (Preston 1962), and was specifically designed to indicate whether or not two areas were in equilibrium with each other. …