The Limits of Delimitation

Those of us who aren’t so well-versed in speciation research--or perhaps even biologists from a few decades ago who were in their day but are just now waking up from some kind of cryogenic chamber--may be puzzled by the question of why relatively easy access to full genome data has not cleanly solved the problem of how to define species. We live in a modern world where genomes of numerous species have been fully mapped out, but determining what constitutes enough genomic divergence to define a species unit remains a surprisingly nuanced issue. Even a complete understanding of a species’ genome may not always be a skeleton’s key. A recent review article by Singhal et al., titled “A Genomic Perspective on Species Delimitation,” provides a succinct summary of many of those lingering difficulties.

I remember a few years ago being amused by a (now paywalled) episode of the Big Biology podcast with entomologist Michael Sharkey, about a paper he produced that used genetic barcoding techniques to identify over 400 new species of South American wasps in one fell swoop. Despite the waves of backlash he received from concerned taxonomists claiming a simple “cut and paste” barcoding method provided nowhere near enough information to accurately identify “real” species, Sharkey maintained in a fairly convincing way that in the face of our ever-growing ecological crisis where so many unidentified species are in danger of extinction, his methodology is the only feasible way to immortalize a good share before it’s too late. Hymenoptera, the most diverse order on earth, contains over 150,000 identified species, with an estimated 1 million more remaining unidentified. Sharkey argued that without high-volume identification methods, an inestimable number of these species will go extinct before being recognized by science. He goes on to mention that his article contained supplementary material pointing to the museum specimens he collected, as well as detailed methodology that anyone can follow to develop his work. It is an interesting philosophy that highlights the diverse assortment of practical challenges biologists face these days, but it clashes somewhat with the gist of this review article.

Singhal et al. begin with a nice summary of genomic delimitation techniques, followed by several brief case studies, one of which focused on redpolls. They highlight research that led to the collapse of three existing redpoll species into one (and ruthlessly took a hard-earned Hoary Redpoll from my life list, damn them!). In this case, RADseq and Multispecies Coalescent data revealed no significant difference between the three putative species, yet populations showed clear differentiation in certain genomic regions that were associated with the subtle morphological differences that previously formed the basis of their species status. “Whole-genome resequencing data” revealed the culprit: a large inversion in an area of the genome coding for plumage and beak morphology. Inversions tend to be relatively immune to recombination, so it’s natural that even in a freely breeding population, an inverted chunk of genes could persist as a separately diverging region. Without a careful understanding of these and related structural factors, such divergence can be interpreted as meeting an arbitrary threshold of difference that could justify species demarcation, but with no effect on reproductive isolation, this would be a mistake.

I took this picture of one of the hundreds of Common Redpolls that were eating seeds in this winter field in Vermont a few years ago. Birders had reported one or two Hoary Redpolls in this large flock, so I spent a few hours sifting through trying to convince myself that one was paler with a smaller bill. I finally did and even though my lifer was ripped away from me a few months later when the species were lumped, it was not a bad way to spend a morning! 

Genomic data is significantly more accurate than single-locus barcoding techniques, and even more accurate than purely morphological approaches. It is becoming cheaper and more effective to use, but that doesn’t mean it offers a perfect solution. As the redpoll example illustrates, and as Singhal et al. go on to hammer home, it is still only a proxy for understanding what is really happening at the ecological level. The researchers recommend thinking of genomic data not as a neutral and objective sequence of nucleotides, but as one (albeit important) factor in a broader context of ecological and phenotypic relationships. Sequence divergence is not always enough to characterize real world reproductive barriers. As a quick aside, they also briefly highlight the utility of relating bacterial symbiont divergence to host species divergence, a fascinating topic which reminded me of this very cool paper from Marcella Baiz et al. on wood warblers (Baiz 2022).

Looking back at De Quieroz’s seminal 2007 “Species Concepts and Species Delimitation” paper, it bears asking, how much progress have we made on this front? At the time, De Quieroz tried to reassure us that a lot of the acrimony in the field stemmed from confusing “species delimitation” with “species conceptualization.” Maybe it’s wise to leave these controversial questions to more experienced researchers than myself, but perhaps we can agree that by defining species simply as a “separately evolving metapopulation lineage,” De Quieroz mostly solved the species conceptualization side of the problem (or at least neatly cleaned it up), giving us the freedom to toil entirely over the delimitation side. As genomic techniques have exponentially developed since 2007, I think it’s fair to say researchers have made huge leaps on this more technical rather than philosophical side, though as Singhal et al. illuminates, there are still many subtleties and pitfalls to watch out for.

 

Works Cited:

Singhal, Sonal, et al. “A Genomic Perspective on Species Delimitation.” Annual Review of Ecology Evolution and Systematics, 3 Sept. 2025, https://doi.org/10.1146/annurev-ecolsys-102723-055311.

“Hymenoptera - an Overview | ScienceDirect Topics.” Www.sciencedirect.com, www.sciencedirect.com/topics/agricultural-and-biological-sciences/hymenoptera.

Baiz, Marcella D, et al. “Gut Microbiome Composition Better Reflects Host Phylogeny than Diet Diversity in Breeding Wood‐Warblers.” Molecular Ecology, vol. 32, no. 2, 20 Nov. 2022, pp. 518–536, https://doi.org/10.1111/mec.16762.

De Queiroz, Kevin. “Species Concepts and Species Delimitation.” Systematic Biology, vol. 56, no. 6, 1 Dec. 2007, pp. 879–886, https://doi.org/10.1080/10635150701701083. Accessed 13 Oct. 2019.