Several stages can be identified in Darwin's effort to formulate natural selection. The first stage corresponded, roughly speaking, to the period up to 1844. It was characterized by Darwin's attempt to base his model of geographic speciation on an individualistic dynamics, with species understood as reproductively isolated populations. Toward the end of this period, Darwin's ignorance of the laws of variations and heredity led him to adopt varieties and species as the units of variations. This had the extremely important effect (...) of stimulating him to consider the process of speciation as involving populations. At the end of this period, Darwin also began to regard adaptation as being exclusively toward places in the economy of nature. Thus he faced the problem of integrating the process of natural selection with the process of speciation. Individual variants were the units that fueled the first process, whereas varieties produced new speices. There was no link between adaptation and speciation, except whatever could be supplied by a quasi-historical, developmental idea of optimizing the amount of life.In the second stage, I contend, Darwin's reading of Milne-Edwards crystallized his previous insights into a coherent whole. Milne-Edwards' comments on the advantage of functional specialization could readily be understood in terms of the advantage accruing to the individual, relative to other members of its species, from occupying a different niche. Milne-Edwards' discussion of the division of labor suggested that organisms which moved into unoccupied niches would enjoy reduced competition, and hence a differential advantage in survival and reproduction; thus they would induce the species to do likewise. Rather than base his explanation on an analogy with the artificial economy, Darwin chose the principle of the optimalization of the amount of life per unit area as the overall explanatory principle. The difficulties connected with integrating different levels of description were therefore circumvented, insofar as the problem of diversity and speciation was concerned. Although natural selection considered individuals as the units of selection, and the units of variations were varieties and species, the dynamics of the process understood in terms of natural selection, competition, division of labor and niches could give a plausible account of how individual advantage could be transferred to the species, and how diversity resulted from this mechanism. The problem of the different levels of descriptions was confined to how the properties of variations in individuals (in particular, the frequency of variations and their transmission) were responsible for the assumed variability characteristic of varieties and species. This problem Darwin never solved.A third stage occurred in 1858 with the amalgamation of the tree-of-life vizualization of the process of speciation. Speciation, geographic distribution, and systematics were all then embedded in a conceptual matrix with vast explanatory powers. (shrink)