The non-discrete nature of subspecies is evident from their definition as geographic segments of any given gonochoristic (bisexually reproducing) species differing from each other to a reasonably practical degree (e.g., at least 70-75%), but to less than totality. All subspecies are allopatric (either dichopatric [with non-contiguous ranges] or parapatric [with contiguous ranges], except for cases of circular overlap with sympatry); sympatry is conclusive evidence (except for cases of circular overlap) of allospecificity (separate specific status). Parapatric subspecies interbreed and exhibit intergradation in contact zones, but such taxa maintain the required level of distinction in one or more characters outside of those zones. Dichopatric populations are regarded as subspecies if they fail to exhibit full differentiation (i.e., exhibit overlap in variation of their differentiae up to 25-30%), even in the absence of contact (overlap exceeding 25-30% does not qualify for taxonomic recognition of either dichopatric populations or of parapatric populations outside of their zones of intergradation). Phenotypic adjustment to differing environmental conditions through natural selection is likely the primary factor in divergence of parapatric subspecies, and undoubtedly is involved in some dichopaffic subspecies. The founder effect and genetic drift are involved more in the latter than in the former.
In spite of the objective validity of subspecies, they have been seriously questioned with increasing vigor for over 40 years, beginning most prominently with Wilson and Brown (1953; but see Wilson 1994). More serious threats have come in recent years from proponents of the newly propounded Phylogenetic and Evolutionary Species Concepts.
The inherent non-discrete nature of subspecies, combined with their long, justifiable history of acceptance as an essential part of biosystematics, has perforce been a problem in the context of phylogenetic inference that logically deals with discrete entities. The impasse has led to unjustified rationalizations in the form of either outright rejection of subspecies or of forcing them into a higher rank (species) for which they are fundamentally not qualified as here (and commonly) defined.
Proponents of the Evolutionary Species Concept dispose of subspecies, for the most part, by disregarding them completely as a component of the biotic classification system (Frost 1995; Frost et al. 1992; Simpson 1961; Wiley 1981, 1992; Wilson and Brown 1953). Typically, in the application of that concept, subspecies are examined critically for their distinctiveness, and those that do not rigorously qualify in that context are not accepted taxonomically. By contrast, proponents of the Phylogenctic Species Concept have generally acted in concordance with the views espoused by Barton and Hewitt (1985), by elevating valid subspecies to species rank despite their non-discrete nature (Baum 1992; Cracraft 1992; Davis and Nixon 1992; Eldredge and Cracraft 1980, Nelson and Platnick 1981). Doing so leaves the term subspecies applicable only to minor geographic variation that usually would not be accepted as taxonomically significant, as, for example, clinal segments and pattern classes (e.g., Cracraft 1992:104). Were subspecies so relegated, we too would join Cracraft in rejecting them.
We hold that neither rationalization is justified or necessary. The rational policy is to continue to accept subspecies as a legitimate and essential part of biotic classification, but at the same time to exclude them from phylogenetic analyses that are logically limited to discrete, fully distinct entities. In that context, subspecies considerations cannot conflict in any way with species concepts and phylogenetic constructs based upon them. The subspecies category is already properly accommodated in the current edition of the International Code of Zoological Nomenclature, and its recognition, as in the past, is supported by biologists worldwide (Hawksworth et al. 1994).
The Rationale for Exclusion of Subspecies from the Phylogenetic Species Concept. — Current interpretation of phylogeny requires that, in phylogenctic classification, populations maintain evolutionary independence if they are to be recognized as taxonomically distinct units. Therefore, parapatric subspecies cannot be embraced in phylogenetic philosophy and practice because they have no phylogeny in the sense of a history of separate identity. On the contrary, dichopatric (= non-parapatric allopatric) populations may individually be regarded as full species, if sufficiently well differentiated (i.e., infallibly diagnosable, with fixed character differentiation), or, if not, as tertiary subspecies. Degrees of distinction (one or more characters) and the extent of intergradation differ widely in nature, and their interpretation in some cases inevitably results in subjective assessment of ranking that is equally inescapable in the application of any species concept. That fact is no argument against the validity of subspecies as nomenclatural entities, although it is a strong argument for excluding subspecies from phylogenctic analyses.
The Problems Created by Subspecific Exclusion from All Classification. — The emergence of a guiding phylogenetic (historical) philosophy in systematics was regarded by Frost et al. (1992:46) as “the biggest paradigm shift in systematics in the last 140 years.”
There is good reason for universal acceptance of the basic tenets of the Evolutionary or Phylogenetic species concepts in proper contexts. It is critical, however, to recognize that these contexts are not all-inclusive. The common perception of classification as conforming with and representing phylogenetic hypotheses must be superimposed upon the even more basic responsibility of reflecting biodiversity. Systematics and classification therefore have two functions, here referred to as taxonomic and phylogenetic. It is not sufficient that classification only reflect phylogeny; it must also reflect biodiversity by accounting for all “kinds” of animals, and subspecies as well as species constitute “kinds.”
The dual responsibility of classification to reflect both phylogeny and biodiversity has long been recognized (e.g., Simpson 1961:27-28). To limit the scope of these dual responsibilities of classification by exclusion of subspecies, through their distortion by elevation to species rank, or through abandonment, in order to reflect phylogeny at all classification levels, is unwarranted. Alternatively, having two classifications, one for each role (phylogenetic and biodiversal), would be chaotic.
Frost et al. (1992), in stating the case for Phylogenetic Classification, argued forcefully for abandonment of subspecies, and Frost later (1995) implemented his views in a novel way. So tempting has been that move toward simplicity that it has spawned widespread neglect of comprehensive studies of the geographic variation that often delineate subspecies. Thus two problems are constituted by subspecies in present contexts: their validity in the first place, and, if recognized at all, their integration within a phylogenetic system of classification.
Validity of Subspecies. — Despite arguments against admission of subspecies in taxonomy, that category has consistently proven useful and illuminating. Nevertheless, as pointed out by Frost et al. (1992), even such a staunch supporter as Mayr on occasion (1982) wrote deprecatingly about the nature and value of the subspecies category. In a more recent work, Mayr and Ashlock (1991:4 1) noted that “some of the best proofs of the occurrence of evolution have emerged from the study of polytypic species taxa. To convert the nominal species of all groups of animals into well- delimited polytypic species taxa is therefore one of the major tasks of taxonomy.” Nevertheless, they (ibid.:53) remarked that the “new systematics” of today recognizes “the subspecies as a category, not as an evolutionary unit.”
It is perfectly true that subspecies are not independent evolutionary units, as previously noted in the context of phylogenetic systematics. But, it is equally true that they are products of evolution and, where valid, are recognizably different taxonomic entities. To exclude valid taxonomic entities from classification misrepresents fact and neglects a level at which a major part of evolution is occurring at all times and at varying rates, although it is little studied, understood, and appreciated in most groups of animals. Such studies, however, “have provided the best available evidence for the process of allopatric speciation, the frequent origin of evolutionary novelties in peripherally isolated populations, and numerous intermediate stages in the evolutionary process, thus elucidating previously inexplicable discontinuities” (Mayr and Ashlock 1991:41).
The initial and most enduringly influential opponents of subspecies were Wilson and Brown (1953), who argued that “it is more informative to focus on the traits and not on the subspecies that might be concocted from them” (Wilson 1994:208). But, by 1994, Wilson (loc. cit.) later confessed that “Some populations can be defined clearly with sets of genetic traits that do change in a concordant, not a discordant manner. Furthermore, the subspecies category is often a convenient shorthand for alluding to important populations even when their genetic status is ambiguous.”
Indeed, perhaps the most powerful argument against the validity of subspecies is that intraspecific geographic variation commonly involves several or more characters varying independently either kaleidoscopically or discordantly, in part due to selection pressures differing with each environmental variable. It is nevertheless true that consistent (at the 70% level or better) recognizability of subspecies is evident within some species. Occasionally only one character qualifies, but usually a suite is involved. In multivariate contexts, subspecies are definable by the suite of characters that explains the largest adequate percentage of the variance of the data-set, thus minimizing the “noise” produced by the variance of other characters. Hence the spectre of an infinite number of subspecies is in reality a delusion. Blanket rejection of subspecies, even on the basis of deficient or defective analysis, denies their possible, if not proven, validity, under the criteria here stated, as consistently recognizable geographic entities. Altertively, elevation of them to full species rank denies their non-discrete nature. Neither alternative is justified.
Although defense of subspecies has been dismissed by some phylogeneticists as old-fashioned and outmoded, the fact remains that comprehensive and objective studies of geographic variation, especially in widely distributed (but also some narrowly distributed) species, often reveals the presence of distinctive geographic segments that eminently qualify for taxonomic, even if not phylogenetic, recognition (e.g., Grismer et al. 1994; Taylor et al. 1994). Rejection of subspecies on grounds of phylogenetic incompetence, or difficulty of analysis, or of burdensome museum administration, is spurious and fails to recognize fact. We view the current popularity of disregarding subspecies as an uncritical attitude.
Rejection of subspecies as non-genetic ecotypes unless proved to be genetic adaptations is unwarranted, inasmuch as dichopatric species of any closely knit species group are subject to the same uncertainty. Parsimony dictates acceptance that allopatric populations that are phenotypically distinguishable at a taxonomic level have their differences genetically based unless or until proved otherwise.
Rejection of the subspecies concept inevitably leads in many instances to unjustified elevation to species rank of taxa that should rank as subspecies. Careful consideration of rank alternatives for different allopatric populations is severely discouraged if one of those alternatives no longer exists.
Although the subspecies category has on occasion been abused in the past, the same is true of the species and every other nomenclatural category. Abuses particularly exist where populations are subject to predominantly discordant geographic variation or linear clinal variation, when segments are nevertheless recognized taxonomically. A comprehensive (rather than piecemeal) study of geographic variation within the species should precede the delineation and naming of subspecies, or should be the ultimate test of subspecies that have been erected piecemeal. The use of multivariate statistical procedures can provide approaches that are reasonably objective and not dependent on preconceptions about taxonomic membership. Nonetheless, the discriminatory power of such methods depends critically on the quality of the characters being analyzed and, in addition, the adopted standard for level of differentiation required for taxonomic recognition. Multivariate analyses (Thorpe 1987) are useful techniques for substantiation of subspecific validity, with revival of the now generally neglected 75% (or similar) rule (idem:7).
The conclusion most emphatically is that subspecies, even on strictly theoretical grounds, are a vital component of taxonomic practice, should be retained as an acceptable option where warranted, merit inclusion in classification, and are properly accommodated in current rules of nomenclature. They are of importance not only in systematics but also in other theoretical disciplines such as biogeography and evolution itself.
In addition, from a practical point of view and quite apart from theoretical validity, the subspecies category is vital to balanced conservational, medical, nutritional, and other applied biologies. In such considerations, the recognition of geographical segments, particularly of widely distributed species, is of fundamental importance because they may represent differentiated gene pools that can be substantially different in important ways (e.g., Glenn and Straight 1985, 1987). Although the focus of phylogenetic systematics is properly on the established directions of genetic change, it would be tragic if that focus were allowed to lead to the abandonment of studies of intraspecific geographic variation that are of such vital importance in the understanding, assessment, and conservation of biodiversity.
In that context, Ryder (1986) proposed recognition of “evolutionarily significant units” (ESU), alternatively “evolutionarily significant populations” (ESP), where they exist among the subspecies of any given species. That concept was endorsed and expanded by Vane-Wright et al. (1991) and Vogler and DeSalle (1993), although the terms ESU and ESP, as applied by these authors, are inappropriate and misleading, as noted by Cracraft (1991); in reality they were dealing with “conservationally” significant units (“CSU”). Nevertheless the focus by these authors upon subspecies and their relative merit for conservation was thoroughly justified. Unfortunately, Cracraft (op. cit.), observing that in his opinion all ESU’s are phylogenctic species, concluded that “Under this concept, subspecies have no special ontological [i.e., incipient species] status and can be abandoned.”
As an example of the importance of subspecies in conservational considerations, Sceloporus undulatus garmani, an arenicolous, terrestrial, cursorial, striped subspecies with reduced semeions (ventral color patches) appears to have become extinct, or nearly so, in parts of its range on the plains of eastern Colorado. On the contrary, another, very different, larger, cross-barred subspecies with well-developed semeions, S. u. erythrocheilus, remains common on the eastern slopes of the Rocky Mountain foothills, where it leads a scansorial life on trees and, mostly, rocks. The two subspecies occur within a few kilometers of each other in some areas, but no intergradation occurs and even gross sympatry is possible. Different conservational measures may be required for these two subspecies, but they would be difficult or impossible to prescribe or implement if no subspecies were recognized. Entire subspecies populations could become extinct if attention were directed to the species as a whole, which could be regarded as basically healthy. In cases such as this, elevating the constituent subspecies to species rank does not always provide a solution. For example, in the case cited, there is a complete, although stepped, continuity between the two taxa: S. u. garmani intergrades to the south with S. u. consobrinus, the latter with S. u. tristichus to the west, and that in turn with S. u. erythrocheilus to the north – a typical circular approximation of range, if not overlap.
Although the practical constraints of conservation efforts may require restriction of attention to the species rank, elevation of subspecies as here and as commonly defined to species rank would be a disservice to limited conservational efforts aimed at preservation only of major genetic resources.
In defending the concept of subspecies, i.e., maintaining them in classification much as in the past, we do not see them as of universal existence among organisms. In many groups they appear never to have evolved, no doubt for various reasons, including extreme vagility. We merely insist that in some groups they are readily evident, and that where this is the case, they merit taxonomic recognition, serving as they do as vehicles for the development of testable hypotheses regarding the causes, proximate and ultimate, for variation and its geographic stabilization.
Integration With Phylogenetic Classification. — The second problem posed by recognition of subspecies in the era of phylogenetic emphasis is perhaps of relatively minor concern, since it can be resolved by context and requires no operational modification of long-established custom. Where the context and goals are strictly phylogenetic, subspecies obviously should be disregarded. Where the context and goals are taxonomic to the lowest level, subspecies should be recognized, as they have been over the past few decades. They are entities already incorporated into taxonomic and nomenclatural procedure, as indicated by the present edition of the International Code of Zoological Nomenclature and its predecessors as early as the 1905 Regles. Biologists worldwide also reiterate their intent to maintain recognition of subspecies in the future (Hawksworth et al., 1994). The studies by Grismer et al. (1994) and Reichling (1994) demonstrate the compatibility of subspecies with phylogenetic systematics. There is no conflict, hence no need for exploration of alternative nomenclatural systems.
Intraspecific Taxonomic Derivation. — In the context of Phylogenetic Classification, as currently interpreted, “intraspecific phylogeny” approaches an oxymoron, and in effect has been widely so interpreted in recent discourses. Yet derivational patterns are commonly discernible among the subspecies of polytypic species, as noted by Thorpe (1987:3), who commented that intraspecific “patterns of geographic variation can [italics ours] be caused by both current ecological conditions and historical factors, i.e. phylogenesis.” The evolutionary phenomena culminating with at- tainment of species rank are also involved in patterns of subspeciation. However, we suggest that intraspecific evolutionary patterns evident among subspecies be designated as taxogenetic (taxon-forming), rather than phylogenctic (derivational), in order to avoid what might be regarded as an oxymoronic application of the latter term.
Indeed, Olmstead’s (1995) proposal for distinguishing “apospecies” (with one or more uniquely derived characters) and “plesiospecies” (with a unique combination of characters, none derived) can logically be extended to subspecies. Thus, “aposubspecies” (e.g., Sceloporus undulatus erythrocheilus with unique lip/throat reddish coloration; S. u. belli, with fused gular semeions) can be distinguished from “plesiosubspecies” (e.g., most other subspecies of S. undulatus).
Summary. — The nature and evolutionary implications of biodiversity is a central challenge of modern biology. To appreciate the full extent and significance of biodiversity, we assuredly need to recognize and have names for the discrete, unquestionably phylogenetically competent taxa belonging to the categories of species, genera, families, etc. We need also to recognize and have names for the non-discrete but genetically distinctive geographic segments of species, even if they are regarded as phylogenetically incompetent (i.e., derivationally non-committal). Both concerns are essential to analyses of diversity. It perforce follows that to pursue such analyses properly a comprehensive and unified system of nomenclature is essential. Our system must include all these groups, both discrete and non-discrete.
Subspecific nomenclature is more than a step in the study of natural variation. It is a tool for conveying information about major patterns of variation within species, flagging opportunities for causal analyses. A subspecies name draws attention to a geographic segment of a species that in some way is recognizably different, and beyond that, it provides information about diversity to a variety of disciplines from genetics and ecology to conservation biology. The study of intraspecific variation could go on without the use of subspecific nomenclature, but a tremendous lot of biotic diversity would be obscured and made almost inaccessible by the absence of names. Workers will fail to notice biological diversity without names.
BARTON, N. H., and G. M. HEWITT. 1985. Analysis of hybrid zones. Ann. Rev. Ecol. Syst. 16:113-148, figs. 1-4.
BAUM, D. 1992. Phylogenetic species concepts. Trends Ecol. Evol. 7:1-2.
CRACRAFT, J. 1991. Systematics, species concepts, and conservation biology. Abstracts Soc. Cons. Biol., 5th Ann. Mtg., Univ. Wisconsin-Madison: 79.
CRACRAFT, J. 1992. Species concepts and speciation analysis. In M. Ereshefsky (ed.), The Units of Evolution: Essays on the Nature of Species, pp. 93- 100. Massachusetts Inst. Tech., Cambridge. xviii, 405 pp.
DAVIS, J. I., and K. C. NIXON. 1992. Populations, genetic variation, and the delimitation of phylogenetic species. Syst. Zool. 41(4):421-435.
ELDREDGE, N., and J. CRACRAFT. 1980. Phylogenetic Patterns and the Evolutionary Process. Columbia Univ. Press, New York.
FROST, D. R. 1995. Foreword to the 1995 printing. In H. M. Smith, Handbook of Lizards (reprint), pp. xvii-xxv. Comstock, Ithaca, New York.
FROST, D. R., A. G. KLUGE, and D. M. HILLIS. 1992. Species in contemporary herpetology: comments on phylogenetic inference and taxonomy. Herpetol. Rev. 23(2):46-54.
GLENN, J. L., and R. C. STRAIGHT. 1985. Distributions of proteins similar to Mojave toxin among species of Crotalus and Sistrurus. Toxicon 23:28.
GLENN, J. L., and R. C. STRAIGHT. 1987. Variation in the venom of Crotalus lepidus klauberi. Toxicon 25:142.
GRISMER, L. L., H. OTA, and S. TANAKA. 1994. Phylogeny, classification and biogeography of Goniurosaurus kuraiwae (Squamata: Eublepharidae) from the Ryukyu Archipelago, Japan, with description of a new subspecies. Zool. Sci. 1 1:319-335.
HAWKSWORTH, D. L., J. McNEILL, P. H. A. SNEATH, R. P. TREHANE, and P. K. TUBBS. 1994. Towards a harmonized bionomenclature for life on earth. Biol. Intern., Special Issue (30):1-45.
E. MAYR. 1982. Of what use are subspecies? Auk 99:593-595.
E. MAYR and P. D. ASHLOCK. 199 1. Principles of Systematic Zoology. Second edition. McGraw-Hill, New York. xvi, 475 pp.
NELSON, G., and N. PLATNICK. 1981. Systematics and Biogeography: Cladistics and Vicariance. Columbia Univ. Press, New York. xiii + 567 pp.
OLMSTEAD, R. G. 1995. Species concepts and plesiomorphic species. Syst. Bot. 20(4):623-630.
REICHLING, S. B. 1994. The taxonomic status of the Louisiana pine snake (Pituophis melanoleucus ruthveni) and its relevance to the evolutionary species concept. J. Herpetol. 29(2):186-198.
RYDER, 0. A. 1986. Species conservation and systematics: the dilemma of subspecies. Trends Ecol. Evol. l(l):9-10. SIMPSON, G. G. 1961. Principles of Animal Taxonomy. Columbia Univ. Press, New York. xiii, 247 pp.
TAYLOR, H. L., C. BEYER, L. HARRIS, and H. PHAM. 1994. Subspecific relationships in the teiid lizard Cnemidophorus tigris in Southwestern Arizona. J. Herpetol. 28(2):247-253.
THORPE, R. S. 1987. Geographic variation: a synthesis of cause, data, pattern and congruence in relation to subspecies, multivariate analysis and phylogenesis. Boll. Zool. 54:3-1 1.
VANE-WRIGHT, R. I., C. J. HUMPHRIES, and P. H. WILLIAMS. 1991. What to protect? Systematics and the agony of choice. Biol. Cons. 55:235-254.
VOGLER, A. P., and R. DESALLE. 1993. Diagnosing units of conservation management. Cons. Biol. 8(2):354-363. WILEY, E. R. 1981. Phylogenetics: The Theory and Practice of Phylogenetic Systematics. Wiley, New York. xv, 439 pp.
WILEY, E. R., 1992. The evolutionary species concept reconsidered. In M. Ereshefsky (ed.), The Units of Evolution: Essays on the Nature of Species, pp. 79-92. Massachusetts Inst. Tech., Cambridge.
WILSON, E. 0. 1994. Naturalist. Island Press, Washington, D.C. xii, 380 pp.
WILSON, E. O., and W. L. BROWN. 1953. The subspecies concept and its taxonomic application. Syst. Zool. 2(3):97-111.