tively distant morphologies? At some level, the number and magnitude of phenotypic transformations comprising a walk must depend upon both the location of fitness peaks and the extent to which the fitness of neighboring variants are correlated. However, it is evident also that the extent to which a walk proceeds depends upon the ability of an organism to alter its phenotype. Although the developmental repertoire of most organisms permits some latitude in external shape and internal structure, walks undoubtedly are governed by genetic or developmental mechanisms that establish barriers to transformations among neighboring variants on the landscape (Alberch, 1980, 1981, 1989; Odell et al., 1981; Oster et al., 1980). Thus, morphological transformations among phenotypes are not equiprobable, and walks cannot be governed exclusively by the topology of the fitness-landscape.

The extent to which walks are genetically or developmentally unfettered is a matter of relative rather than absolute degree because it undoubtedly varies among organisms and changes over evolutionary time. For example, the developmental "plasticity" of plants appears extremely high in comparison with most animals (Schmid, 1993; Sultan, 1987; Van Tienderen, 1990). By the same token, certain periods of evolutionary time are characterized by exceptionally high rates of phenotypic innovation as, for example, the colonization of the terrestrial landscape by the first vascular plants (Figure 1). Indeed, one is left with the impression that the walks of plants, in general, and those of early tracheophytes, in particular, feature phenotypic transformations sufficient to achieve many, perhaps most, of the morphological optima widely scattered over their fitness-landscapes.

An added advantage to dealing with plants is that their fitness calibrates closely with the operation of physical laws and processes governing the exchange of mass and energy between the plant body and the external environment, which have remained constant over evolutionary time (Gates, 1965; Brent, 1973; Alberch, 1980, 1981, 1989; Oster et al. , 1980; Gill et al., 1981; Odell et al., 1981; Nobel, 1983; Sultan, 1987; Van Tienderen, 1990; Niklas, 1992; Schmid, 1993). Thus, the broad outlines of the fitness-landscape for plants likely have remained comparatively constant. If so, then plants may be the ideal venue for examining the relation between the topology of fitness-landscapes and the dynamics of more or less unrestricted walks.

The assumption that walks are unimpeded over stable fitness-landscapes greatly simplifies attempts to explore the relation between