Introduction

Piaggio et al. (2024) challenged our revision of the subspecies taxonomy of a genetically and geographically isolated population of Odocoileus virginianus (white-tailed deer) in Douglas County (DC), Oregon (O. v. douglasi; Douglas County white-tailed deer) because of our definition of a subspecies and alleged inappropriate application of cranial morphology and population genetic data. Our response will focus on justifying interpretations of those genetic data and detailing the analyses and environmental and ecological circumstances of our morphological data (Smith et al. 2003, 2024) in the context of the subspecies concept we endorse. Their opposition seems to be based on two underlying tenets: 1) recognition of subspecies should focus on a molecular lineage perspective of an organism's history (i.e., a delineation of geographically isolated molecular clades); and 2) allopatric populations isolated because of human-caused disturbance should be reconnected through repatriation rather than recognized as different taxa.

Rather than imposing clade structure as a prerequisite for infraspecific taxonomy, we chose to acknowledge geographically separated reproductive communities with distinguishing morphological features and corroborating genetic variation as subspecies (Pritchard et al. 2000, Patton and Conroy 2017, Diersing 2019). The comments offered by Piaggio et al. (2024) regarding variation in cranial morphology among the three populations dismissed the likelihood of unique selective pressures associated with local habitat and environmental conditions (Smith et al. 2024). Instead, Piaggio et al. (2024) argue that the differences are clinal or were influenced by introgression from O. hemionus along the Lower Columbia River (LCR) without evidence of any causal mechanism responsible for the variation in cranial morphology or any verification of the extent of hybridization during the period when the skulls were collected. They also disregarded the potential influence of significant hybridization between O. hemionus (mule deer) and O. virginianus along the LCR (Hopken et al. 2015) on genetic differentiation among O. virginianus populations as relevant to considering infraspecific taxonomy. Additionally, discounting the legitimacy of anthropomorphic disturbance during a period now recognized as the Anthropocene epoch (Crutzen 2002) within which the sixth mass extinction is occurring (Barnosky et al. 2011, Toussaint et al. 2021) is impractical and unwise. Besides, restoring a shared gene pool through restoring natural corridors (most of which are under private ownership) that have been substantially altered for over a century is unrealistic. Regardless, conservation goals should focus on maintaining local ecological and evolutionary processes rather than maintaining specific phenotypes (Moritz 1999, Anderson and Weir 2022) with less attention on the evolutionary continuum from populations to species (Coates et al. 2018).

Below, we discuss in greater detail apparent weaknesses of criticisms and alternative explanations presented by Piaggio et al. (2024). We provide evidence from our study and the literature in support of our research and conclusions that warranted the revision of O. virginianus subspecies taxonomy.

Subspecies Concept

We acknowledge the role of genetics and lineage in taxonomy, which is a paradigm advanced initially by Hennig in 1966 (cited in Patton and Conroy 2017). Today, much of the infraspecific literature has focused on delineating geographically isolated molecular clades (Patton and Conroy 2017). However, imposing clade structure as a necessary criterion for infraspecific taxonomy is contrary to Hennig's own philosophy and ignores the underlying genetic basis of morphological features (Patton and Conroy 2017). The goal of investigating the infraspecific taxonomy of mammals is to identify and to recognize all allopatric reproductive communities with distinguishing features as unique subspecies. This conceptual framework establishes a “nonhierarchical, nonreciprocal monophyletic definition for infraspecific taxa” (Patton and Conroy 2017:1019). Braby et al. (2012:699) defined subspecies as “groups that comprise evolving populations representing partially isolated lineages of a species that are allopatric, phenotypically distinct … and that these character differences are correlated with evolutionary independence according to population genetic structure.”

Formal recognition of subspecies should not be a strict application of a molecular-only perspective of an organism's history (Patton and Conroy 2017, Diersing 2019). Geographically separated, reproducing populations with distinguishing phenotypic characteristics should be acknowledged and recognized as subspecies (Braby et al. 2012, Patton and Conroy 2017). The population of O. v. douglasi is isolated geographically and genetically from O. virginianus along the LCR and northeastern Oregon (Piaggio and Hopken 2009, Hopken et al. 2015, Piaggio et al. 2024). O. v. douglasi is morphologically and genetically distinguishable from conspecifics along the LCR and northeastern Oregon. The emergence of unique haplotypes and genetic differentiation in the context of allopatric populations experiencing different environmentally mediated selective pressures indicate the LCR and DC populations of O. virginianus are evolving along separate trajectories and undergoing allopatric speciation (Anderson and Weir 2022).

Cranial Morphology

Smith et al. (2003) examined crania of adult O. virginianus from Idaho (n = 18), northwest Oregon and southwest Washington (n = 117; Gavin and May 1988), and DC, Oregon (n = 129; Smith 1981). Idaho skulls were museum specimens; skulls from northwest Oregon and southwest Washington were collected in the mid-1970s by T.A. Gavin; and W.P. Smith collected skulls in DC, Oregon in the late 1970s. Smith et al. (2024) reviewed all skull data from Smith et al. (2003) and analyzed mensural data from 97 complete adult skulls.

Piaggio et al. (2024) challenged both the results and interpretations of the findings, offering criticisms of the statistical analyses and offering alternative explanations. They questioned the significance of the variation presented in Figure 3 (principal component analysis [PCA] plot) of Smith et al. (2024) because of overlapping ellipses. We acknowledge a reduction in statistical power (which contributed to the extent of ellipse overlap) because the sample size was reduced from 264 specimens to the 97 complete adult skulls used in the PCA (Smith et al. 2024). Still, after controlling for differences related to sex or age among the complete skulls, the first axis explained 93.2% of the variation, distinguishing Douglas County (DC) specimens from those of Idaho and the Lower Columbia River (LCR) (P < 0.0001) by a combination of shorter basilar and nasal lengths, narrower braincase, and narrower nasals (Smith et al. 2024:Table 1). On the second axis, specimens from Idaho were distinguishable from those of the LCR and DC (P < 0.001) by having longer basilar lengths and broader braincases (Smith et al. 2024). Moreover, when considered in the context of the 11 significant cranial dimensions in Table 1 (Smith et al. 2024), the evidence of statistically significant variation in cranial morphology among the three O. virginianus populations is compelling.

Interestingly, Piaggio et al. (2024) argued that some of the cranial variability could be associated with introgression from O. hemionus, despite not presenting any mensural data from the skulls of hybrids or any evidence regarding the extent of hybridization that occurred during the 1970s when the skulls were collected from the LCR region. They also questioned if the variation in cranial morphology reported by Smith et al. (2024) represented taxonomically relevant differences and speculated without any empirical evidence that it was more likely clinal variation associated with differences in nutritional resources (Piaggio et al. 2024).

Similar geographical variation in cranial dimensions has been reported for several large mammal species (Smith 1991; Genov 1999; Molina and Molinari 1999; Wehausen and Ramey 1993, 2000; Kennedy et al. 2002). Some taxa show continuous variation in skull morphology corresponding to climatic or other environmental gradients (Kennedy et al. 2002, Heffelfinger and Heffelfinger 2023), whereas other taxa show abrupt dissimilarity associated with geographical isolation (Diersing 2019) and display substantial genetic dissimilarity among regional populations (Miller 1995, Patton and Conroy 2017). The primary consideration when interpreting cranial variation in the context of subspecific taxonomy is whether morphological variation is indicative of parallel genetic divergences (Patton and Conroy 2017, Diersing 2019).

The three populations of white-tailed deer in our study showed significant variation in cranial morphology that is associated with geographical and genetic isolation (Piaggio et al. 2024) and corroborating genetic differentiation (Smith et al. 2024). Although environmental gradients and differences in nutritional resources across locations could produce differences in physical attributes, phenotypic traits could just as likely be caused by underlying genetic differences. Such investigations into causal mechanisms of morphology were beyond the scope of Smith et al. (2024)'s study; moreover, Piaggio et al. (2024) did not provide any evidence to support their claim. Conversely, multiple studies have described substantial differences in climate, vegetation, and community ecology between the LCR region and southwestern Oregon (Franklin and Dyrness 1973; Gavin et al. 1984; Smith 1985a, 1985b), all of which have contributed to selective pressures unique to local environmental circumstances that were likely responsible for phenotypic attributes and corresponding genetic variation (Smith et al. 2024).

Finally, we acknowledge the expertise and years of experience that Piaggio et al. (2024) have with the life history and population genetics of cervids. Still, it is noteworthy to consider that the focal populations and surrounding landscapes of Smith et al. (2024)'s investigation received extensive ecological, morphological, population genetics, and taxonomic study by T.A. Gavin and W.P. Smith (Smith 1981, 1985a, 1985b, 1987, 1991; Gavin et al. 1984; Gavin and May 1988; Smith et al. 2003). Indeed, W.P. Smith coauthored the first version of the Odocoileus virginianus leucurus (Columbian white-tailed deer) Recovery Plan (Paullin et al. 1983) and served as a consultant to the Recovery Team for decades, contributing essential information regarding critical habitat that led to the delisting of the DC population. T.A. Gavin initiated the first population genetic study of O. v. leucurus to examine the taxonomy of O. v. leucurus (Gavin and May 1988). Clearly, their experience and knowledge from multiple years of intense field work and analyses provide unique insights regarding the LCR and DC populations.

Genetics

The following statement in Piaggio et al. (2024:308) is inaccurate and misleading: “Overall, the LC/ JBH and DCOR subpopulations of O. v. leucurus do not meet the subspecies definition of Smith et al. (2024:101), given they have more shared alleles between them than private ones that separate them.” Although we highlighted the unique haplotype and allele in the DC population, our definition of subspecies does not rely on the presence or absence of single haplotypes or alleles, or the number of shared alleles between populations. Smith et al. (2024:101) defined a subspecies as “isolated, evolving populations of a species that are allopatric and phenotypically distinct, in which quantifiable attribute differences are correlated with evolutionary independence as evidenced by population genetic structure (Braby et al. 2012).” We used genetic differentiation (F_ST) as evidence of evolutionary independence (Wright 1978).

According to Wright (1978), an F_ST value of 0.10 indicates moderate genetic divergence, whereas values between 0.15 and 0.25 represent greater genetic differentiation; F_ST > 0.25 represents significant genetic differentiation. The F_ST values among the three O. virginianus populations all were greater than the F_ST value between O. h. hemionus (mule deer) and O. h. columbianus (black-tailed deer) (F_ST = 0.10), which are two recognized regional subspecies of O. hemionus. Of particular significance regarding the relevance to subspecific status was the finding that O. v. leucurus and O. v. douglasi populations showed the greatest genetic divergence (F_ST = 0.31), which was markedly greater than the genetic differentiation between either of these populations and O. v. ochrourus (Northwestern whitetail) (0.15 and 0.19). Moreover, a plot of factorial correspondence analysis of deer genotypes revealed three distinct white-tailed deer groups separated on the first two axes with no shared individuals (Piaggio and Hopken 2009:Figure 7).

We acknowledge that F_ST can be greater as a result of low genetic diversity, but this does not necessarily diminish the significance of genetic divergence among the populations or its relevance to subspecies taxonomy (Cook and MacDonald 2001). Indeed, the characteristic low genetic diversity of the DC population confirms the absence of appreciable gene flow from other O. virginianus populations (O. v. leucurus and O. v. ochrourus). We also agree that genetic drift or inbreeding could reduce genetic diversity (Piaggio et al. 2024), and therefore increase genetic divergence from other white-tailed deer populations. However, there are numerous endemic mammals across the Alexander Archipelagos with low genetic diversity because of founder effects and genetic drift, but that has not diminished their recognition and acknowledgement as endemic subspecies (Cook and MacDonald 2001). Still, we share their concern regarding low genetic diversity, inbreeding, and the risk of extinction (Piaggio et al 2024). However, recognizing the population in DC as a new subspecies does not preclude introducing genetic diversity through translocating individuals from other populations, as was accomplished with Puma concolor coryi (Florida panther) (Johnson et al. 2010). Indeed, it seems much more practical and realistic than restoring natural corridors across western Oregon to increase the genetic diversity of the DC population and restore a shared gene pool (Piaggio et al. 2024).

Conclusion

We do not agree with the species concept, alternative explanations of possible mechanisms responsible for variation in cranial morphology, the perspective that genetic differentiation (F_ST) is an irrelevant indicator of evolutionary independence, or the perspective that human-caused disturbance has not been a legitimate, relevant geological influence on O. virginianus population distributions or genetics that were presented by Piaggio et al. (2024). Nonetheless, our publication is a working hypothesis, as are most scientific papers, about what constitutes a new subspecies of white-tailed deer, based on extensive analyses of abundant morphological and genetic data. It is available for examination, criticism, and debate. We will consider our work a success if it generates meaningful discussion; this is what science is all about. We appreciate all comments and points of view.