Edited by: William J. McShea, Smithsonian Conservation Biology Institute (SI), United States
Reviewed by: Andrew Hope, Kansas State University, United States; Peter Moyle, Center for Watershed Sciences, University of California, Davis, United States
*Correspondence: Robert M. Zink,
This article was submitted to Animal Conservation, a section of the journal Frontiers in Conservation Science
This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.
More than 170 subspecies are listed as threatened or endangered under the US Endangered Species Act. Most of these subspecies were described decades ago on the basis of geographical variation in morphology using relatively primitive taxonomic methods. The US Fish and Wildlife Service defaults to subspecies descriptions by taxonomists working with specific groups of organisms, but there is no single definition of subspecies across plants and animals. Valid tests today usually entail molecular analyses of variation within and among populations, although there is no reason that behavioral, ecological or molecular characters could not be used, and include tests for significant differences between samples of the putative endangered subspecies and its nearest geographic relatives. We evaluated data gathered since subspecies listed under the ESA were described finding about one-third are valid (distinct evolutionary taxa), one-third are not, and one-third have not been tested. Therefore, it should not be assumed that because a subspecies occurs in a checklist, it is taxonomically valid. If the US Fish and Wildlife Service intends to continue listing subspecies, we suggest that they convene taxonomic experts representing various groups of organisms to provide a minimal set of criteria for a subspecies to be listed under the ESA.
The legislative basis for much of the conservation effort in the United States is the Endangered Species Act (ESA), passed in 1973 and modified in 1978, 1982, and 1988. Given its title, one might expect it to apply only to species, but in fact it also can be used to list subspecies and distinct population segments (vertebrates only) as threatened or endangered. Listing decisions usually come about when either a U.S. citizen or organization, or the Fish and Wildlife Service itself, determines that the population size of one of these taxonomic entities places it in danger of extinction (endangered) or in almost as much peril (threatened). The species or subspecies is placed on a list of candidate species, and the Fish and Wildlife Service is directed to use the best available scientific or commercial data in making a ruling as to whether the taxon merits listing as endangered or threatened. In this paper we examine the taxonomic category or rank of subspecies. In particular, we determine whether modern tests of subspecies limits have confirmed the validity of listed subspecies, most of which were described more than a half-century ago, as described by National Academy Member John C. Avise in the opening quote.
Systematists, taxonomists, and evolutionary biologists have struggled to define the term
A subspecies is a formal taxonomic category that is specified by three Latin names: the genus name, the species name, and the subspecies name. Definitions of subspecies range from whatever a taxonomist says is valid to multi-character genetic and morphological assessments (
Rigorous descriptions of subspecies has not historically been the status quo. Consider the Rio Grande subspecies (
There are examples in which subspecies correspond to genetically or morphologically defined units that have experienced evolutionarily independent histories and therefore qualify for listing under the ESA. For example, the spotted owl (
Molecular methods have revolutionized tests of subspecies and their evolutionary independence (
Unlike morphological characters, molecular characters used to date are often considered “selectively neutral”—that is, not influenced unduly by natural selection—and hence the only reason for congruent geographic patterns is that they reflect a common underlying evolutionary history.
New molecular methods, often-called next-gen, have resulted in the possibility of surveying thousands to millions of loci, often in the form of single nucleotide polymorphisms. The El Segundo blue butterfly (
Thus, the next-generation sequencing methods need to be interpreted with caution so as not to confuse sampling and genetic gaps (see below) and so as not to cherry-pick SNPs that favor one hypothesis over another. Given examination of enough SNPs, it would be likely to find some in only one or a few populations, making it seem like support for their distinctiveness. That is, one might exclude characters that suggest a different pattern, whereas overall differentiation should be assessed across all characters (e.g., SNPs). That is, conflicting characters should be a part of the analysis so as not to bias the result to a preconceived conclusion. In addition, next-gen methods do not guarantee similar findings from different labs, as in the case of the California gnatcatcher (
Costs of preservation vary widely within and among different groups of organisms (
The coastal California gnatcatcher (
Most ESA-listed subspecies were described before 1950 (137 of 175), and 150 (86%) were described before 1966 (see
Distribution of years in which subspecies listed under the endangered species act were described. Arrow
The molecular methods used evolved from relatively crude assessment of distinguishing alleles at protein-coding loci (allozyme electrophoresis) to studies involving thousands of base-pairs at the DNA level. Most molecular examinations (n = 92) of subspecies limits used mtDNA (n = 67), and some were combinations of mtDNA and microsatellites (n = 19) or mtDNA and nuclear DNA (n = 14). Evaluations of listed subspecies vary widely in their sampling size, from a single individual to over 100 samples. Given the variation in the areal extent of listed subspecies’ distributions, a diversity in sampling size is not surprising; however, the relative percentage of the distribution covered by sampling also varies widely. In several cases researchers were able to include only a single population represented by one individual, thus making inferences of population distinctiveness difficult.
We examined 165 listed subspecies to determine how many were supported by modern analyses (see
Review of Subspecies Listed under the US Endangered Species Act.
Genus | Species | Subspecies | Valid? | Reference |
---|---|---|---|---|
|
||||
|
|
|
yes (island) |
|
|
|
|
yes (island) |
|
|
|
|
no |
|
|
|
|
no |
|
|
|
|
no |
|
|
|
|
no data | |
|
|
|
no |
|
|
|
|
yes |
|
|
|
|
no |
|
|
|
|
no (only haplotype frequencies differ) |
|
|
|
|
equivocal |
|
|
|
|
yes |
|
|
|
|
not tested | |
|
|
|
prob. (endemic to Hawaii, no test of other ssp.)* | |
|
|
|
prob. (endemic to Guam, no test of other ssp.) |
|
|
|
|
not tested | |
|
|
|
not tested* | |
|
|
|
not tested | |
|
|
|
yes, historically |
|
|
|
|
no (+/−)—equivocal |
|
|
|
|
not tested | |
|
|
|
not tested* | |
|
|
|
no |
|
|
|
|
not tested | |
|
|
|
prob. species* | |
|
|
|
no |
|
|
|
|
no |
|
|
|
|
no | |
|
|
|
no |
|
|
|
|
yes (prob. undersplit) |
|
|
|
|
no |
|
|
|
|
yes |
|
|
|
|
yes | |
|
|
|
no | |
|
|
|
yes |
|
|
|
|
yes |
|
|
||||
|
|
|
no (no comparisons) | |
|
|
|
yes |
|
|
|
|
yes |
|
|
|
|
yes |
|
|
|
|
no |
|
|
|
|
no |
|
|
|
|
not tested | |
|
|
|
not tested | |
|
|
|
not tested | |
|
|
|
not tested | |
|
|
|
not tested | |
|
|
|
once, not anymore |
|
|
|
|
yes |
|
|
|
|
no, prob. introgressed with rainbow trout* | |
|
|
|
yes | |
|
|
|
not tested | |
|
|
|
not tested | |
|
|
|
not tested | |
|
|
|
not tested | |
|
||||
|
|
|
not tested | |
|
|
|
not tested, prob. introgressed |
|
|
|
|
no |
|
|
|
|
no* |
|
|
|
|
no* |
|
|
|
|
no |
|
|
|
|
not tested | |
|
|
|
prob., intergrades* |
|
|
|
|
no | |
|
|
|
no* | |
|
|
|
no (genomics) |
|
|
|
|
not tested* | |
|
||||
|
|
|
not tested |
|
|
|
|
yes/no—equivocal |
|
|
|
|
no molecular data |
|
|
|
|
no* |
|
|
|
|
not studied* |
|
|
|
|
not tested (within ssp. analysis) | |
|
|
|
not tested (within ssp. analysis) | |
|
|
|
no |
|
|
|
|
prob. extinct* |
|
|
|
|
yes (no bootstrap support)* |
|
|
|
|
not tested | |
|
|
|
yes |
|
|
|
|
not tested |
|
|
|
|
not tested |
|
|
|
|
not tested* |
|
|
|
|
yes (morph), not tested | not tested |
|
|
|
not tested, may be distinct | |
|
|
|
not tested | |
|
|
|
yes (mtDNA COI sequences) |
|
|
|
|
not tested | |
|
|
|
not tested* | |
|
|
|
not tested | |
|
|
|
not tested | |
|
|
|
not tested | |
|
|
|
not tested | |
|
|
|
yes (AFLP)* |
|
|
|
|
not tested* | |
|
|
|
prob., small n—equivocal |
|
|
|
|
not tested* | |
|
|
|
yes |
|
|
|
|
no |
|
|
|
|
not tested | |
|
|
|
not tested | |
|
|
|
n = 1, so not tested* | |
|
|
|
n = 1, so not tested* | |
|
|
|
not tested | |
|
|
|
not tested |
|
|
|
|
not tested |
|
|
|
|
not tested |
|
|
|
|
no |
|
|
|
|
not tested | |
|
||||
|
|
|
yes, network unrooted |
|
|
|
|
yes()?—equivocal |
|
|
|
|
no* |
|
|
|
|
no |
|
|
|
|
no |
|
|
|
|
no (yes if ssp. boundary changed) |
|
|
|
|
no |
|
|
|
|
no |
|
|
|
|
no |
|
|
|
|
no |
|
|
|
|
no (island) |
|
|
|
|
yes (translocations) |
|
|
|
|
yes |
|
|
|
|
yes (RMZ unpubl. analysis of mtDNA sequences in GenBank) | |
|
|
|
no |
|
|
|
|
no |
|
|
|
|
no, nested in |
|
|
|
|
yes, should be species* |
|
|
|
|
not tested* | |
|
|
|
no* |
|
|
|
|
no | |
|
|
|
yes |
|
|
|
|
not tested |
|
|
|
|
no |
|
|
|
|
yes (1 mismatched haplotype) |
|
|
|
|
no |
|
|
|
|
yes* (some overlap) | |
|
|
|
* | |
|
|
|
yes* | |
|
|
|
not tested* | |
|
|
|
yes* (some overlap) | |
|
|
|
yes* (some overlap) |
|
|
|
|
yes* (some overlap) | |
|
|
|
yes (island) |
|
|
|
|
once, not anymore | |
|
|
|
no | |
|
|
|
North American likely diff. from South American |
|
|
|
|
yes, but includes all NA | |
|
|
|
yes |
|
|
|
|
no or not tested | |
|
|
|
||
|
|
|
yes |
|
|
|
|
yes |
|
|
|
|
no |
|
|
|
|
no | |
|
|
|
yes |
|
|
|
|
yes |
|
|
|
|
no |
|
|
|
|
not tested? | |
|
|
|
not tested? | |
|
|
|
not tested? | |
|
|
|
not tested? | |
|
|
|
no |
|
|
|
|
no |
|
|
|
|
not tested* | |
|
|
|
equivocal |
|
|
|
|
yes |
|
|
|
|
no |
|
*The Integrated Taxonomic Information System considers the subspecies invalid. (In most cases, the basis for an invalid conclusion is a change in the scientific name of the described subspecies.)
This list does not include other subspecies already considered invalid by the Integrated Taxonomic Information System: Aerodramus vanikorensis bartschi, Bufo hemiophrys baxteri, Drymarchon corais couperi, Epicrates monensis monensis, Epioblasma torulosa gubernaculum, Epioblasma torulosa rangiana, Epioblasma obliquata obliquata, Epioblasma torulosa torulosa, Epioblasma obliquata perobliqua, Epioblasma florentina florentina, Gasterosteus aculeatus williamsoni, Lasiurus cinereus semotus, Oxyloma haydeni kanabensis.
Distribution of results of evaluation of listed vertebrate subspecies. “Not tested” means samples from listed subspecies were not compared with samples from adjacent subspecies, there were too few samples, or samples were not examined.
We found that the ITIS classification departs from our summary. More than 40 of the 51 subspecies (78%) that were not supported by our evaluation were considered valid on the ITIS website.
The molecular methods used to test subspecies have evolved greatly over the past few decades, owing to a large increase in resolving power. With the new potential to describe genomes of individuals, some issues should be recognized. First, if sampling is not evenly spaced, sampling gaps will give the illusion of discrete taxonomic boundaries (
It is possible for a subspecies that is not evolutionary distinct (a category that includes many subspecies) to be ecologically important—perhaps important enough to merit listing. Examples might include keystone species such as large carnivores: the Florida panther, for instance. However, providing quantitative data of ecological importance might be as large a task as documenting taxonomic distinctiveness. The lack of consistency among subspecies definitions used in ESA listings is a major failing of taxonomists. To further the use of taxonomic work in conservation decisions, this failing ought to be addressed.
We understand that some will have the view that all subspecies proposed for listing should be accepted at face value because they might be valid but not protected because of no current tests of their validity, hence, their loss would be lamentable (a Pascal’s Wager argument). We argue here that given the high cost of subspecies preservation and the fact that roughly 50% of subspecies tested are supported by modern methods, it should be unacceptable to list a subspecies under the ESA without modern analyses confirming
The original contributions presented in the study are included in the article/
RZ designed project, conducted analyses, wrote the draft. LK gathered raw data and edited the manuscript. Both authors contributed to the article and approved the submitted version.
We thank M. Cronin for valuable suggestions on the manuscript. The project received support from the Center for Growth and Opportunity at Utah State University.
The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.
All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.
The Supplementary Material for this article can be found online at:
ITIS: Integrated Taxonomic Information System (home page), accessed May 31, 2022,