types for a complete analysis of phenotypic variation (Scharloo, 1991), (ii) the possibility that the fitness contributed by performing one task depends upon the ability to perform other tasks simultaneously (i.e., epistatic fitness-contributions; Ewens, 1979; Franklin and Lewontin, 1970; Lewontin, 1974), (iii) the likelihood that the dynamics of a walk depend upon the walk's point of origin on the landscape in addition to intrinsic genetic or developmental barriers to phenotypic transformations, and (iv) the requirement to treat biological realistic temporal and spatial variations in the topology of fitness-landscapes, as well as (v) the complex interactions among a panoply of physical and biological variables that collectively define fitness (Kauffman, 1993; Gould, 1980; Levin, 1978).

Nonetheless, the image of the fitness-landscape continues to inspire questions about the tempo and mode of evolution—for example, what is the relation between the number of fitness peaks and the number of functional tasks that an organism must simultaneously perform to grow, survive, and reproduce? Although there is no a priori reason to assume that the number and location of phenotypic optima depend upon the number of tasks an organism must perform, there are good reasons to believe that manifold functional obligations author "course-grained" landscapes with many phenotypic optima. For example, engineering theory shows that the number of equally efficient designs for an artifact generally is proportional to both the number and the complexity of the tasks than an artifact must perform (Meredith et al., 1973) because the efficiency with which each of many tasks is performed must be relaxed due to unavoidable conflicting design specifications for individual tasks (Gill et al., 1981), and, as the number of tasks increases, the number of configurations that achieve equivalent or nearly equivalent performance levels increases (Brent, 1973). If such relationships hold true for organisms, these relations may account for the morphological and anatomical diversity seen among even closely related species. Indeed, the sharp logical distinction between "optima" and "maxima," on the one hand, and the observation that a multi-task artifact may assume diverse appearances, on the other, suggest the hypothesis that the imposition of manifold obligations increases the number of equally fit phenotypes.

Although the problematic analogy between engineered and biological systems speaks to the topology of fitness-landscapes, it sheds no light on questions related to the dynamics of walks—for example, what is the relation between the number of tasks that an organism must perform and the magnitude of the morphological transformations between neighboring variants required to reach fitness optima? Are walks confined to nearest-neighbor variants or are they free to reach compara-



The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement