Okay so, I'm slowly becoming embroiled in the seemingly eternal debate over species concepts, and to make matters worse, the ones on microbial species. For much of the past century, bacterial species were considered by many to be a fiction, given their perceived clonal reproduction. Others grouped them, somewhat arbitrarily, by morphological and metabolic characters. More recently, genetic similarity has been used to carve boundaries between prokaryotic species. However, in what has been almost a 180 degree turn from the clonal objection, over the past two decades, many researchers have pointed to widespread horizontal gene transfer to either refute the validity of bacterial and archaeal species or to posit some form of ecotypic species concept. I question these conclusions, as well as those that dismiss the validity of phylogenetic inference for prokaryotes and the Tree of Life. I'll leave the latter discussion, although closely related, for another time.
In terms of prokaryotic species concepts, my impression is that Carl Woese was on the right track with his concept of progenotes. In prokaryotes, hgt appears to be widespread, but we know that it declines with increasing phylogenetic distance. Based on Woese' progenote idea, perhaps sexual reproduction is merely one end of a continuum? At the other end, maybe we find a progenote "mass" characterized by totally promiscuous hgt. As differentiation occurred, hgt would have become more restricted within lineages. But, hgt does provide major selective advantages (as does sex), so perhaps in one lineage, hgt became "institutionalized" in the form of gamete production, recombination and sexual reproduction... And keep in mind that in eukaryotes, interspecific hybridization is also more widespread than historically believed, and also declines with phylogenetic distance.
So, what this might imply for microbial "species" is not that they don't exist, but, given greater "intraspecific" genetic variation, are broader than eukaryotic species. That "reproductive isolation" may be fully applicable. This concept might not be so useful from a clinical point of view, where more pragmatic differentiation is needed. What do folks think?