Figure 2. Species as reproductive communities emerging from the past (a, c, d) or as communities at this instant, and possibly projecting into the future (b). Reproductively incompatible genotypes shown by dark blue versus teal green colours. [...]
Abstract
Biologists and philosophers of science have been unable to fully resolve the decades-long controversy as to what kind of unit of living biodiversity should receive the valued label “species”: reproductive communities (among sexual organisms), genealogical groups, or clusters of organisms that share traits. Among these choices, which represent a spectrum from process to history to observable outcome (respectively), the latter (more operationalist) concepts are not viable. Species of sexual organisms must embody or imply cohesive and integrating processes such as interbreeding and shared ecological pressures if they are to have sufficient power to bear the burden we give them: to predict or explain traits across the genome and among organisms. This commitment to cohesive process is needed whether biologists use species as taxonomic containers to synthesize data, as minimal phylogenetic units, or as actors in evolutionary diversification. These varied uses can be satisfied via a concept of reproductive community, but not the strict Biological Species Concept (BSC). Its two drawbacks are a focus on the contemporary and a restriction to intrinsic factors. Current reproductive compatibility may predict future matings, but it does not explain well the traits and genes that living organisms already have. The organisms alive today were shaped by isolating factors of the past, not those of the present, to whatever extent those differ. The most broadly-useful species concept must therefore see species retrospectively, as reproductive communities of the past. As well, the BSC’s exclusion of extrinsic factors renders each of its units incomplete in explanation and synthesis. Reproductive communities in nature were isolated not just by intrinsic (genetic) differences, but also by purely extrinsic (e.g., geographic) factors. Such reproductive communities were and are real, natural entities whose integrated and self-reinforcing cohesive processes constrained genealogical descent and aligned the distribution of many traits. This Retrospective Reproductive Community Concept (RRCC), formalized mathematically in multispecies coalescent models, justifies the traditional practice of taxonomists using morphological data to seek the echoes of past reproductive cohesion. However, which reproductive communities naturally deserve to be ranked as species, and which as demes or populations, is a vexing question. There is no natural, discrete and broadly informative species rank that applies universally, or perhaps even usually. To whatever extent species rank is justified, it is as justified for asexual organisms as for sexuals. The presence or absence of sex is just one example of the variability biologists confront. Because cohesive processes vary among clades, a useful and broadly-applicable species concept cannot specify detailed cohesive mechanisms. Nor can it perfectly align the named species of taxonomy with units of evolution, because the latter are not structured to match taxonomy’s partition of boxes. Taxonomic species should approximate, but can only approximate, evolutionary units. Settling on retrospection, letting go of a natural meaning for species rank, and accepting taxonomy as approximation allow biology to turn to the far more daunting task: listening to the natural world to understand the many interacting processes that built distinction and identity, that shaped the reproductive communities emerging out of the past into the present day.