If species are individuals, it is because they are lineages, and lineages are a kind of individual. But lineages are processes extended over time, not things, and not all processes qualify as individuals. For a process to be an individual it must have some significant integration and coherence and the most plausible candidate for providing this is sexual reproduction.
Sexual reproduction, however, is not found in most species. Asexual reproduction suggests that a new lineage might be said to start with each individual and certainly with any significant genetic novelty; certainly such lineages should not be identified with species. This raises two kinds of question. First, what kinds of lineages can we or should we trace as evolving processes in the great mass of evolutionary situations that do not involve evolving individual-like lineages? Standard tree of life representations suggest a structure of distinct and coherent lineages, but we know that this is quite misleading. A more realistic picture of prokaryotic evolution allows space for a number of evolutionary processes, including notably the kinds of super-individual multi-species processes recently discussed by Ford Doolittle.
Second, are there evolutionary processes that are unique to those species that form individual lineages? Is there an evolutionary advantage to the emergence of coherent lineages? Many of the evolutionary processes that have been most discussed in recent evolutionary theory—niche construction, cultural, symbolic and epigenetic inheritance, maternal effects—may well be limited to integrated sexual species or subsets of these. I shall explore the question what kind of lineages do make possible these various processes.
A moral that should emerge from the discussion is that the general species as individuals thesis, and the generalization of some of the evolutionary processes characteristic of such species, while important ideas in their contexts, are excellent illustrations of the dangers of taking large charismatic metazoans as paradigms biological individuals. A more defensible view is that there are many different kinds of lineages, some of these constitute species, and only a subset of this latter group qualify as species individuals.