A Reanalysis of Interspecific Association in an Old Pasture

Recently, there have been a number of references to perceiving the vegetation in a community from the 'plant's-eye-view' (Turkington 1975; Harper 1977, 1978; Turkington & Harper 1979a, b; Fowler & Antonovics 1981). It is interesting to ask questions about communities from the plant's-eye-view rather than from the usual anthropomorphic viewpoint, since the fitness of an individual plant is most influenced by those plants that make demands on the same environmental resources-primarily immediate neighbours. Investigations of interspecific association and community diversity have been dominated by the measurement of species presence-absence, cover, frequency etc. in quadrats (see review by Janson & Vegelius 1981), but this scale is often too large and fails to document the finer-scale relationships that are immediately relevant to the individual plant's experience (see Mack & Harper 1977). It can be argued that the biologically most meaningful analysis of vegetation would impose no anthropocentric scale, but would ask of the plants themselves how they sensed the composition of the community; this is to ask for a 'plant'seye-view'. Various plotless sampling methods have been devised that permit questions to be asked concerning the nature of the community from the plant's-eye-view. Yarranton (1966) developed such a technique and applied it to a bryophyte-lichen community. This method was modified and used in pastures by Turkington (1975) and Turkington & Harper (1979a, b); the method, called the 'contact sampling' method, was also used by Turkington, Cavers & Aarssen (1977), Aarssen, Turkington & Cavers (1979), Turkington & Cavers (1979), and Saulei (1981). A modified form of the method has been used by Aarssen (1983). Further, Stowe & Wade (1979) developed the speciesjuxtapositions method and species-regions method for an old field in Illinois, and Fowler & Antonovics (1981) the neighbouring point analysis for a mown field in North Carolina. Ratcliffe (1983) has applied Stowe & Wade's (1979) methods in the alpine zone of southern British Columbia, Canada. All of these methods focus attention on the way in which an individual plant experiences its environment of neighbours and they are sensitive and convenient for detecting fine-scale interspecific association. De Jong, Aarssen & Turkington (1983) and Meagher & Burdick (1980) have drawn attention to the frequent erroneous statistical treatment of data collected by various plotless methods. However, it would be unfortunate if pertinent questions in plant ecology could not be addressed adequately for lack of appropriate and available analyses. When the contact-sampling method is used to collect data, the usual method of