Society of Systematic Biologists

The Evolutionary Species Concept Reconsidered

Author(s): E. O. Wiley
Source: Systematic Zoology, Vol. 27, No. 1 (Mar., 1978), pp. 17-26
Published by: Taylor & Francis, Ltd. for the Society of Systematic Biologists
Stable URL: http://www.jstor.org/stable/2412809
Accessed: 07-02-2018 22:48 UTC

JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide
range of content in a trusted digital archive. We use information technology and tools to increase productivity and
facilitate new forms of scholarship. For more information about JSTOR, please contact [email protected].

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at
http://about.jstor.org/terms

Society of Systematic Biologists, Taylor & Francis, Ltd. are collaborating with JSTOR to
digitize, preserve and extend access to Systematic Zoology

This content downloaded from 199.17.62.177 on Wed, 07 Feb 2018 22:48:18 UTC
All use subject to http://about.jstor.org/terms
 THE EVOLUTIONARY SPECIES CONCEPT
RECONSIDERED

E. 0. WILEY

Abstract
Wiley, E. 0. (Division of Fishes, Museum of Natural History, University of Kansas, Law-
rence, KS 66045). 1978. Syst. Zool. 27:17-26.-The concept of species (as taxa) adopted by
an investigator will influence his perception of the processes by which species originate. The
concept adopted should have as universal applicability as current knowledge permits. Simp-
son's definition of a species is modified to state: a species is a lineage of ancestral descendant
populations which maintains its identity from other such lineages and which has its own
evolutionary tendencies and historical fate. This definition is defended as that which has
widest applicability given current knowledge of evolutionary processes. Four corollaries are
deduced and discussed relative to other species concepts: (1) all organisms, past and present,
belong to some evolutionary species; (2) reproductive isolation must be effective enough to
permit maintenance of identity from other contemporary lineages; (3) morphological distinc-
tiveness is not necessary; and (4) no presumed (hypothesized) single lineage may be subdi-
vided into a series of ancestral-descendant "species." The application of the evolutionary
species concept to allopatric demes and to asexual species is discussed and it is concluded
that the lack of evolutionary divergence forms the basis for grouping such populations into
single species. It is suggested that some ecological species definitions lead to under-esti-
mations of the rate of extinction due to interspecific competition because their logical frame-
work excludes unsuccessful species from being species. Finally, the implications of accepting
an evolutionary species concept to the field of phylogeny reconstruction are discussed. [Spe-
cies concepts; evolution; phylogeny reconstruction.]

In all probability more paper has been that investigator views the "facts" of na-
consumed on the questions of the nature ture. This is as true for the species con-
and definition of the species than any cept as it is for the concepts of relativity
other subject in evolutionary and system- theory and quantum mechanics. I submit
atic biology. In my opinion this is be- that the conceptual problems of species
cause the species is usually viewed as an definitions are no worse than the concep-
essential entity in the evolutionary pro- tual problems of many other words in sci-
cess. A species, in the phylogenetic ence.
sense, may be viewed as the largest ag- The extensive literature regarding the
gregate of individual organisms that histories of various species concepts has
evolve as a unit. There has been much been reviewed by Mayr (1957, 1963,
investigation of the process, but the ques- 1969), Simpson (1961), Dobzhansky
tions of what a species is and how species (1970), Grant (1971), Sokal (1973), and
originate are far from solution. Sokal Sneath and Sokal (1973), among others.
(1973) has suggested that there are two An anthology of papers on species con-
major problems, an adequate species def- cepts has been compiled by Slobodchi-
inition, and the origin of species. I agree, koff (1976; see Platnick, 1977, for signif-
but the two problems are not separate. icant omissions). Another historical
The definition of the word species will review is unwarranted and I refer the in-
be built on a species concept and the con- terested reader to these works and ref-
cept itself will profoundly affect the way erences therein. Instead, I will attempt
in which investigators view the origin of to resurrect and defend the species con-
the species they study. As Popper (1968), cept of Simpson (1961; see also Ghiselin,
Eldredge and Gould (1972), and others 1969; Grant, 1971) as best suited for deal-
have pointed out, the concepts and the- ing with the species and its origins.
ories an investigator holds influence how Any species concept "adopted" in light
17

This content downloaded from 199.17.62.177 on Wed, 07 Feb 2018 22:48:18 UTC
All use subject to http://about.jstor.org/terms
 18 SYSTEMATIC ZOOLOGY

of current knowledge must in my opin- for criticizing various species definitions

ion, fulfill several roles. (1) It must have because they are not "operational." Fur-
as universal validity as current knowl- thermore, there can be little justification
edge permits. This would be reflected in for applying pure Operationalism to
its applicability to species through time,either systematic or evolutionary philos-
to asexual as well as sexual species, to ophy. It should be recognized that the
plants as well as to animals. (2) It should critics of "non-operational" species con-
make possible the formulation of an hy- cepts (cf. Sokal and Crovello, 1970;
pothesis that a particular group of organ- Sneath and Sokal, 1973) have themselves
isms either comprises a species or does been unable to supply an "operational"
not. The hypothesis must be testable in definition. Perhaps this is because the
principle (cf. Hempel, 1965). (3) It must word "species" is a universal, and uni-
subsume within its logical framework all versals cannot be defined in an opera-
valid special case definitions of species. tional manner (Popper, 1968).
(4) It must be capable of dealing with
species as spatial, temporal, genetic, epi- DEFINITION OF TIlE EVOLUTIONARY

genetic, ecological, physiological, phe- SPECIES CONCEPT

netic, and behavioral entities. (5) It must I propose that a simple modification of
clearly specify what types of species Simpson's (1961:153) species definition
origins are possible and what types are fulfills the criteria discussed above. A
not possible. This final role is essential species is a single lineage of ancestral
to the testability in principle of the con- descendant populations of organisms
cept itself for if the concept permits which maintains its identity from other
everything and prohibits nothing, it is such lineages and whichi has its own evo-
scientifically useless (Popper, 1968). lutionary tendencies and historical fate I
This definition implies that (1) species
OPERATIONALISM AND SPECIES
can be thought of as individuals rather
CONCEPTS
than classes (cf. Ghiselin, 1966, 1974; see
Some biologists (for example, Ehrlich, discussion by Hull, 1976, and comment
1961; Sneath and Sokal, 1973) have ar- by Mayr, 1976), and (2) that species are
gued that a species definition must be historical temporal, and spatial entities.
"operational" within the context of the The definition is empirical in that it per-
philosophical school of Operationalism mits hypotheses to be derived from it
espoused by Bridgmann (1945) among
others. Operationalism has no doubt I Simpson's definition states: "An evolutionary
been beneficial in influencing scientists species is a lineage (an ancestral-descendant se-
quence of populations) evolving separately from
to clarify their concepts and definitions.
others and with its own unitary evolutionary role
However, Hull (1968) has argued that and tendencies." Both definitions stress unity (as
pure Operationalism is not possible in does Ghiselin's, 1974) and both imply that the spe-
science and that species concepts and cies is the most inclusive unit of evolution. My def-
definitions do not have to conform to inition
op- has the advantage of not implying that spe-
cies must change (evolve). If species go through a
erationalist philosophy to be scientific rapid (or even saltatory) phase of evolutionary
(see also Hempel, 1966). In my opinion, change and then settle down to a longer period of
Operationalism has been found wanting relative stasis (of. Eldredge and Gould, 1972), then
because its "operational" definitions "evolving" is not to be preferred over "maintaining
identity" as species may maintain their identity rel-
were either circular (Popper, 1968) or
ative to other species with or without changing.
failed to fulfill their set goal of separating,
"Identity" as used here connotes individual identity
by definition, the cognatively meaningful and does not connote either stasis or change in mor-
from the cognatively meaningless. If "op- phology. As an analogy: I have maintained my own
erational" definitions within an opera- identity through the years in spite of the fact that
my morphology has changed substantially from age
tionalist philosophy are flawed in these 6 to present (also see Vendler, 1975, for other dis-
ways, then there can be little justification cussions of individuality).

This content downloaded from 199.17.62.177 on Wed, 07 Feb 2018 22:48:18 UTC
All use subject to http://about.jstor.org/terms
 SPECIES CONCEPTS 19

which can be tested in fact or principle ily or order (I exclude from this discussion
(i.e., the concept of empiricism advocat- the category genus because our system of
ed by Hempel, 1965). For example, we nomenclature demands that every spe-
may frame the hypothesis, "these two cies belong to a genus). As Goldschmidt
populations (or groups of populations) (1940), Schindewolf (1950), L0vtrup
maintain separate identifiable evolution- (1974) and others have asserted, phylo-
ary lineages." Evidence used to test such genetic evolution has proceeded from the
an hypothesis can come from a variety of apex to the base. Therefore, natural phy-
sources depending on the nature of the logenetic classification must also proceed
organism and the genetic, phenetic, spa- from the apex to the base if it is to reflect
tial, temporal, ecological, biochemical, the pattern of evolution.
and/or behavioral evidence which is No genus could have originated before
available to test the question. Whether a the family in which it belongs originated.
group of organisms is or is not a species By this I do not mean that genera, fami-
then becomes an hypothesis to be tested. lies, or any other supraspecific taxa
evolve. What I mean is that the stem spe-
LOGICAL COROLLARIES OF THE
cies of a genus cannot originate before
DEFINITION the stem species of the family to which
Logical corollaries are those state- that genus is a member. This is a logical
ments or implications which follow di- corollary of the axiom of continuity of
rectly from an axiom, a definition, or an descent and simply a rejection of spon-
hypothesis. The evolutionary species taneous generation. It is logically derived
concept has several logical corollaries from the assertion that higher taxa origi-
which permit its evaluation. I do not sug- nate concurrently with their stem spe-
gest that the corollaries outlined below cies, i.e., that the family at the time of its
are the only corollaries possible. origin was composed of a single evolu-
Corollary 1.-All organisms, past and tionary species (Hennig, 1966; Wiley,
present, belong to some evolutionary 1977; for another point of view see Plat-
species. This corollary is logically evi- nick, 1976, 1977). For example, the an-
dent from the observation that every or- cestral species of the gar family Lepisos-
ganism belongs to a lineage including at teidae must have originated before either
least its parents (Vendler, 1975). To deny of the ancestral species of this family's
this, one would have to invoke sponta- included genera, Lepisosteus and Atrac-
neous generation. Evolutionary species tosteus. On the practical level of recon-
differ from supraspecific taxa in that evo- structing phylogenies, the synapomor-
lutionary species are lineages or continua phies of the family Lepisosteidae are
whereas higher taxa are groups of sepa- hypothesized to have originated before
rate lineages linked by past continua. the synapomorphies of the genus Lepi-
Thus, higher taxa are historical constructs sosteus. On the theoretical level, when I
whose sole basis for existence is depen- make such a statement I mean that the
dent on how accurately they document characters which I hypothesize as syn-
past continua whereas evolutionary spe- apomorphic for Lepisosteidae are those
cies are the extinct or living continua characters which are hypothesized to
themselves. In other words, evolutionary have arisen in the stem species of the
species make history, natural higher taxa family as autapomorphies and passed on
(monophyletic groups) are history, and to the descendants of that species (the
un-natural higher taxa (para- and poly- ancestral stem species of Lepisosteus and
phyletic groups) misrepresent history. Atractosteus). And, the stem species of
In terms of classification, as opposed to these genera later developed their own
pattern and process, all organisms belong sets of separate autapomorphies which
to a species, but not all species necessar- were passed on modified or unmodified
ily belong to a higher taxon such as a fam- to their descendants. I do not wish to im-

This content downloaded from 199.17.62.177 on Wed, 07 Feb 2018 22:48:18 UTC
All use subject to http://about.jstor.org/terms
 20 SYSTEMATIC ZOOLOGY

ply that any particular set of characters Corollary 3.-The evolutionary spe-
are "family characters" or that another set cies concept does not demand that there
are "generic characters." The synapo- be morphological or phenetic differences
morphies which characterize the stem between species, nor does it preclude
species of Lepisosteidae do not make this such differences. Therefore, any investi-
taxon a family, the significance of Lepi- gator may under- or over-estimate the
sosteidae lies in the fact that it is mono- true number of evolutionary species at
phyletic and not on its particular categor- one time plane or through several time
ical rank. When I look at a phylogram I planes when he or she bases that number
perceive all branches as evolutionary on morphological or phenetic difference.
species or higher taxa represented by The phenetic species concept (as dis-
their evolutionary stem species. Farris cussed, if not explicitly defined, by
(1976) has recently discussed the rela-Sneath and Sokal, 1973) is a special case
tionships between categorical rank, taxa,
of this corollary when phenetic differ-
and classification which I think makes ences do occur between separately
these relationships clear and which is evolving lineages. The corollary specifi-
compatible with both the above discus- cally carries the connotation that real
sion and the evolutionary species con- evolutionary lineages exist in nature out-
cept. side man's ability to perceive these lin-
Corollary 2.-Separate evolutionary eages (cf. the attitude taken by Dobzhan-
lineages (species) must be reproductive- sky, 1970; Huxley, 1940; Mayr, 1963,
ly isolated from one another to the extent 1969; Meglitsch, 1954; Ghiselin, 1974;
that this is required for maintaining their but not Darwin, 1859, or Gilmour, 1940;
separate identities, tendencies, and his- see Meglitsch, 1954 for a critique of Gil-
torical fates. This corollary covers the mour's attitude). The under- or over-es-
special case of the biological species def- timation of the actual number of evolu-
inition (Mayr, 1963; Dobzhansky, 1970). tionary species represents the
The biological species concept stresses observational error in the systems or
the community gene pool and reproduc- methods employed to differentiate spe-
tive isolation. Both are inherent in the cies. For example, it is possible that a fos-
evolutionary species concept (as Simp- sil assemblage includes two sibling spe-
son, 1961, concluded). And, in spite of cies with identical morphology in their
the objections of Sokal and Crovello preserved characters, but which operated
(1970) and of Ehrlich and Raven (1969), as separate lineages when living. A pa-
the biological species concept seems a leontologist would conclude, on the basis
testable special case definition covering of the evidence at hand, that only one lin-
the sympatric occurrence of sexually re- eage existed. Thus, he would under-es-
producing sister species (species that are timate the actual number of evolutionary
each other's closest relatives). Thus, species. The same investigator might in-
while the biological species concept may terpret a sexually dimorphic species as
be valid it is not inclusive (see Hull, two species, an over-estimation. A neon-
1971, for a detailed criticism of Sokal and tologist might recognize two evolution-
Crovello, 1970).2 ary species as a single species because
they hybridize at their zone of sympatry,
2 Rosen (pers. comm., and two manuscripts in
while a future neontologist might con-
press) has argued that the biological species con-
cept is simply invalid from the point of view of tax- clude that this documented hybridization
onomy and geographic analysis of monophyletic represented a sorting out and reinforce-
groups of taxa. This is because its premises are de- ment of the two species' identities (Dob-
signed to sacrifice evidence of cladistic relation-
zhansky, 1970) rather than a swamping of
ships in favor of real or imagined data on reproduc-
tive isolation when the two conflict. If so, one might
their identities. In practice, as Sokal and
ask if the concept should be used at all, at least by Crovello (1970), and others have pointed
taxonomists and biogeographers. out, the morphological or phenetic dif-

This content downloaded from 199.17.62.177 on Wed, 07 Feb 2018 22:48:18 UTC
All use subject to http://about.jstor.org/terms
 SPECIES CONCEPTS 21

ferences between populations will usu- If this were true, then Simpson's defi-
ally determine the number of species rec- nition and the modification presented
ognized. But this does not require here would be useless. However, evolu-
investigators to accept a phenetic species tion is composed of two geneological pro-
definition or a phenetic species concept cesses. One, the continuum, ties all of life
(see Hull, 1971, for discussion of this together. The other, punctuations of the
point). Rather, it requires that an inves- continuum, produces diversity, when ac-
tigator adopt the working hypothesis that companied by differentiation followed
sufficient morphological difference com- by divergence, by providing indepen-
bined with geographic distribution are dence of lineages. As Bonde (1975:294)
adequate evidence for separate evolu- has stated: "This continuum is subdivid-
tionary lineages. But what is "adequate" ed in a non-arbitrary way by the specia-
evidence? I suggest a clearly stated hy- tion process which delimits 'natural' spe-
pothesis with corroborating instances cies in the time dimension." Every
which are both relevant and parsimoni- punctuation of the continuum followed
ous. What is adequate will change with by divergence results in a single evolu-
increased knowledge of organisms stud- tionary lineage being split into two or
ied and experience in dealing with sim- more evolutionary lineages. There is no
ilar problems. If this sounds rather vague, doubt that one can run from man to pro-
I suggest that no concept or method can tist in one classificatory taxon, but, in my
guarantee a "correct" solution will follow opinion, that taxon would be Eucaryota,
from it. There is no guard against practic- not species Homo sapiens. There was a
ing "bad science" unless "bad" is de- genus Homo before there was a species
fined in relation to some authoritarian Homo sapiens, just as there was a class
standard. What is important is that the Vertebrata before any of the Recent ver-
hypothesis be open to testing. tebrates evolved. Thus, we tie together
Corollary 4.-No presumed separate, increasingly ever larger taxa on the basis
single, evolutionary lineage may be sub- of the continuum they are hypothesized
divided into a series of ancestral and des- to have shared in the past, and if we
cendant "species." This corollary is, per- adopt a truly natural classification, this
haps, self-evident and would hardly be classification will document past contin-
worth discussing if it were not for the ua, not bury their reality or existence.
rather common practice among some pa- It is the punctuation of continua which
leontologists of subdividing a single lin- is the first prerequisite that may lead
eage into a number of "species." Simp- to evolutionary diversity among species
son (1961) viewed this practice as at any one time. These punctuations are
necessary for escaping a kind of infinite for the most part what are termed allo-
regression in classification. He said patric speciation events (Mayr, 1963, for
(1961:165): "If you start at any point in discussion), but there is no conceptual
the sequence and follow the line back- reason to exclude sympatric speciation as
ward through time, there is no point a special case. In the case of the allopatric
where the definition would cease to ap- speciation model, one might expect that
ply. You never leave an uninterrupted, in most circumstances the ancestral spe-
separate, unitary lineage and therefore cies would become extinct at the time of
never leave the species with which you speciation. This is because it is improb-
started unless some other criterion of def- able that either of the two daughter spe-
inition can be brought in. If the fossil rec- cies would have the same identity, role,
ord were complete you could start with and historical fate as the ancestral spe-
man and run back to a protist still in the cies. Thus, in most cases the methodo-
species Homo sapiens. Such classifica- logical necessity of postulating extinction
tion is manifestly both useless and some- of ancestral species in phylogeny recon-
how wrong in principle." struction as advocated by Hennig (1966)

This content downloaded from 199.17.62.177 on Wed, 07 Feb 2018 22:48:18 UTC
All use subject to http://about.jstor.org/terms
 22 SYSTEMATIC ZOOLOGY

is biologically (as well as methodologi- species could stand the loss of geograph-
cally) sound. However, an implication of ically or ecologically unique gene com-
the evolutionary species concept would binations without its role and tendencies
seem to be that if the ancestral species being changed. The problem of determin-
can lose one or more constituent popu- ing the survival of a particular ancestral
lations without losing its historical iden- species in this situation is the practical
tity or tendencies, then it can survive testability of the situation. Engelmann
such a split. For example, there are her- and Wiley (1977) have argued that ances-
petologists who recognize phenetically tor identification based on morphological
different parthenogenic "clones" of uni- or stratigraphic evidence is invalid. If so,
parental salamanders and lizards as spe- it may not be possible at present to sepa-
cies (Darevsky, 1966; Wright, 1967; Uz- rate out those cases where peripheral
zell and Barry, 1971). Darevksy (1966) budding results in extinction of the ances-
recognized four uniparental species of tral species from those cases where it does
Lacerta from the Caucasus and suggested not. The point I would like to make is that
their derivation from L. saxicola. The both concepts are compatible whether or
uniparental whiptail lizard Cnemidopho- not we have the methodological tools to
rus opatae is said by Wright (1967) to be discriminate between them at the pres-
derived from the sexually reproducing ent time.
species C. tesselatus. Does this deriva-
ASEXUAL SPECIES AND ALLOPATRIC
tion, probably caused by a fortunate mu-
DEMES
tation in a single female, require that
C. tesselatus became extinct at the time Asexual species have been the bane of
this female's first brood hatched? I doubt all proposed species definitions that are
that the origin of C. opatae has any more not overtly typological. Dobzhansky
effect on the identity and evolutionary (1970) has termed the asexual species a
tendencies of C. tesselatus than did the pseudospecies. Yet, asexually reproduc-
death of any one other member of the ing lineages have existed and do exist.
population. One might argue that such Phenetic taxonomists have, with justifi-
"clones" are not species but variants of cation, pointed to the difficulties of ap-
a single species (see Zweifel, 1965). Yet, plying various biological species con-
they do represent independent lineages cepts to asexual species. Mayr (1963)
(albeit perhaps of short duration) and concluded that his biological species def-
thus do correspond to evolutionary spe- inition was restricted to sexually repro-
cies. Another example is the origin of a ducing species. Simpson (1961) suggest-
plant species by hybridization and sub- ed the evolutionary species concept
sequent polyploidy. This speciation covered asexual species. I suggest that
mechanism would seem to have little ef- asexual species can be accounted for un-
fect on either parental species (except the der the evolutionary species concept in
waste of some ova and pollen) unless the the same way we place allopatric demes
hybrid successfully outcompeted its par- into a single sexually reproducing spe-
ents and caused their extinction, which cies. We may ask, "why are the Siberian
would be possible only if the parental and North American populations of wol-
species survived the actual origin of the verines considered the same species?"
hybrid. It might even be reasonable to (Kurten and Rausch, 1959). One might
postulate that the allopatric separation of answer that we assume they would inter-
a peripheral isolate of a large population breed if they were brought into contact.
could occur without initially affecting This answer is impossible to evaluate be-
that large population. The survival of the cause it assumes some future event. Any-
same species through more than a few of thing could happen, but only what does
these "buds," however, would appear un- happen can be used to corroborate an hy-
likely because it is unlikely that any one pothesis. While such a statement might

This content downloaded from 199.17.62.177 on Wed, 07 Feb 2018 22:48:18 UTC
All use subject to http://about.jstor.org/terms
 SPECIES CONCEPTS 23

be considered a prediction, it cannot be asexual clones has the potential for split-
used to justify the deduction. One might ting into two or more separate indepen-
argue that there must be migration which dent evolutionary species, but it does not
keeps the populations from diverging follow that any significant divergence
morphologically. There is no evidence at must occur. Lack of differentiation is as
hand for this, just the a priori assumption valid a historical fate as differentation
that they must diverge unless some other and it may have real genetic, epigenetic,
factor (such as migration) is present. I ecological or other bases. We may hy-
suggest that Siberian and North Ameri- pothesize a lack of differentiation even if
can wolverines are considered the same we cannot distinguish between truly
species because we have no corroborat- undifferentiated populations and differ-
ing evidence that they have reached a entiated lineages whose nature has gone
point of divergence where we can de- undetected. Failure to distinguish be-
duce that they are following separate tween the two represents an inadequacy
evolutionary pathways. It is possible that of our systems of observation and method
1,000 or 10,000 years from now an inves- and not the inadequacy of the concept.
tigator will see differences and conclude In other words, we are faced with the
that wolverines are following two inde- same Type I and Type II errors that
pendent evolutionary paths. It is possible plague all systems of hypothesis testing,
that wolverines are already differentiated the chances of rejecting a true hypothesis
but we have examined the wrong char- and the chances of accepting a false hy-
acters for detecting that differentiation. I pothesis.
would suggest that whether wolverines
ECOLOGY AND THE EVOLUTIONARY
are now differentiated or will differen-
SPECIES
tiate in the future, the decisive factor in
this process will be the geographic event Van Valen (1976:233) recently offered
which separated them and thus set up a modification of Simpson's (1961) spe-
their potential independence. But at this cies definition: "A species is a lineage (or
time we have no corroboration that this closely related set of lineages) which oc-
particular geographic event will lead to cupies an adaptive zone minimally dif-
separate evolutionary paths and thus we ferent from that of any other lineage in
have no reason to recognize two evolu- its range and which evolves separately
tionary species. Meglitsch (1954) and from all lineages outside its range."
Simpson (1961) argued that an evolution- Although there is much in this defini-
ary species definition can be applied to tion and Van Valen's subsequent discus-
asexual organisms because the popula- sion that I can agree with, the definition
tions have retained the capacity to evolve has certain logical difficulties. Any spe-
as a unit if artificial (not man-made) bar- cies is a component of its environment
riers are removed, implying the assump- and must be at least minimally adapted
tion that uniparental lines are derived to the biotic and abiotic factors of that
from biparental lines (Dougherty, 1955; environment if it is to survive. Species
Pontecorvo, 1956; Stebbins, 1960). Per- cannot be divorced from their environ-
haps so, but like potential interbreeding, ment any more than they can be divorced
this calls for evidence not at hand. My from their gene pools or their morpholo-
reasoning above would also apply to pre- gies. But, species do not have to occupy
dominantly asexually reproducing spe- minimally different niches or adaptive
cies. It is not the evolution of asexual re- zones from other species within their
production which "permits" us to ranges to be considered species.3 It is
consider genetically separate clones as a
3I recognize that the term "adaptive zone" as
single species, but their lack of signifi-
used by Van Valen (1976) is not the exact equivalent
cant evolutionary divergence. Any spe- to "niche," but the distinction is irrelevant to this
cies composed of allopatric demes or discussion.

This content downloaded from 199.17.62.177 on Wed, 07 Feb 2018 22:48:18 UTC
All use subject to http://about.jstor.org/terms
 24 _ _ _ _ _ _ _SYSTEMATIC ZOOLOGY

A B C A C B C ed that if the Florida scrub jay were elim-

inated, then it should not be considered
a separate species, inspite of reproduc-
tive isolation, because the two jays are
ecological equivalents. I would suggest
that the elimination of one sister group
a. b. C by another through competition (and not
swamping of the gene pool) is corrobo-
ration of separate lineages and therefore
FIG. 1.- (a). One of four possible aspect-B state- corroboration of separate species rather
ments of the phylogenetic relationships between than the reverse. Both Van Valen's and
species A, B, and C. (b). An aspect-A statement with Pitelka's discussions have merit, but it is
'"species" B considered the direct ancestor of "spe- my opinion that such concepts lead to an
cies" C. (c). An aspect-A statement with species A
underestimation of the rates of extinction
considered the ancestor of species B and C.
due to interspecific competition.

THE EVOLUTIONARY SPECIES AND

possible for two species to share essen- PHYLOGENY RECONSTRUCTION

tially the same niche within the same Engelmann and Wiley (1977) argued
range. In the case where resources are that ancestor-descendant hypotheses
not limiting, two such species could co- based on morphological or stratigraphic
exist. In the case where resources are evidence were unscientific because they
limiting, one of the species could replace could never be corroborated or refuted
the other through interspecific competi- by valid tests. We addressed our discus-
tion from that portion of the range where sions largely to alternate hypotheses such
they are sympatric, or, entirely via ex- as those shown in Figs. la and lb. Fig-
tinction. Indeed, if interspecific compe- ure la represents what Nelson (1973)
tition causes at least some extinctions, it terms an aspect-B phylogeny in which
can work only where the niches of the only cladistic relationships are expressed
competing species are similar enough for and ancestral species remain hypotheti-
competition to occur or where one spe- cal. Figure lb represents Nelson's (1973)
cies' niche completely overlaps the oth- aspect-A phylogenetic hypothesis in
er's (MacAuthur and Levins, 1967). which, for this example, species B is hy-
This definition, applied to successful pothesized the direct ancestor of species
species, could be considered another C. If an investigator adopts the evolution-
special case definition of the evolution- ary species concept, then Fig. lb is log-
ary species concept as defined above. ically impossible because "species" B
But, given the original definition of Van and "species" C are part of the same uni-
Valen (1976) one might argue that a spe- tary evolutionary lineage and thus B be-
cies forced to extinction through inter- longs to the same species as C. Thus, the
specific competition was not a species at problem collapses into a two-taxon prob-
all. Interestingly, Pitelka (1951) presents lem whose relationships are already re-
exactly this argument in considering a solved since A must be the sister species
hypothetical scenario involving blue jay of BC given no other taxa. Of course, pop-
evolution. Given the hypothetical situa- ulation B could be considered ancestral
tion where the Florida scrub jay and the to population C, given that BC is an evo-
Texas jay were physiologically incapable lutionary species, since it is a logical de-
of interbreeding, could the Florida jay be duction that some individuals (sampled
considered a species if it should fail to or unsampled) from population B are the
survive when it meets its ecological populational ancestors of all of the indi-
equivalent and sister species, the Texas viduals of population C (see Engelmann
jay? (Pitelka, 1951:379). Pitelka suggest- and Wiley, 1977:2, for a discussion of

This content downloaded from 199.17.62.177 on Wed, 07 Feb 2018 22:48:18 UTC
All use subject to http://about.jstor.org/terms
 SPECIES CONCEPTS 25

population ancestors).4 I would conclude ancestor in question. Finally, with regard

that much of the discussion we presented to stratigraphic criteria, if the ancestor
was directed toward a problem which can survive the split which produced one
does not exist under the evolutionary or more descendant species, then strati-
species concept and which we would graphic criteria cannot serve to falsify or
not have discussed in the manner we did corroborate an ancestor descendant hy-
if we had adopted this concept. However, pothesis on the species level, since it is
I would maintain that our analysis of the conceivable that an ancestor could post-
problem is applicable in cases where it date its descendants in the fossil record.
is postulated that one species gave rise to But, if the ancestor cannot be established
two or more daughter species and itself on morphological grounds, then perhaps
became extinct (Fig. lc). I would also sub- that question is moot.
mit that in a situation where an ancestral
species is evolving phyletically (anagen- ACKNOWLEDGMENTS

ic evolution) so that the autapomorphies Earlier drafts of this paper were read by numer-
which it will leave to its descendants as ous reviewers whose comments are greatly appre-
synapomorphies are fixed over a period ciated. These include Drs. Richard Johnston, Rob-
ert Hoffmann, and Norman Slade and Messrs.
of time, then the problem, on a morpho-
Randall Moss, Gregory Pregill, and David Wiseman
logical level, will never be resolved. This (University of Kansas); Drs. George Engelmann,
is because when the ancestor splits from Eugene Gaffney, Norman Platnick, Donn Rosen,
its own ancestor it will presumably have and Bobb Schaeffer (American Museum of Natural
History); and Dr. Michael Ghiselin (University of
few, if any, of the derived characters
California at Berkeley). My special thanks to Dr.
which will be fixed in its own popula- David Hull (Univ. of Wisconsin-Milwaukee) and
tions at a later time (or times) and left to Dr. James S. Farris (State Univ. of New York, Stony
its descendants when it speciates and be- Brook) for their extensive and penetrating com-
comes extinct. Thus on a morphological ments which greatly strengthened the final draft.
Ms. Debora Bennett prepared the figure. The con-
level the analysis might result in a spe-
clusions drawn, and any mistakes in logic made, are
cies being the primitive sister group of my own responsibility.
its descendants, plus the ancestor of its
descendants, plus the third taxon in a tri- REFERENCES
chotomy with its descendants depending
on the rate of phyletic change and the BONDE, N. 1975. Origin of "higher groups": view-
points of phylogenetic systematics. Problemes
number of samples from various strati-
actuels de paleontologie-6volution des verte-
graphic levels available to the investiga- bres. Coll. Internat. C.N.R.S., no. 218:293-324.
tor. This is because at various times the BRIDGMANN, P. W. 1945. Some general principles
ancestor might have none, part, or all of of operational analysis. Psychological Rev. 52:246-
the shared derived characters of its des- 249.
DAREVSKY, I. S. 1966. Natural parthenogenesis in
cendants. Thus, even if we could identify
a polymorphic group of Caucasian rock lizards
one part of the lineage as the actual related to Lacerta saxicola Eversmann. J. Ohio
"ancestor," earlier parts of the same spe- Herpet. Soc. 5:115-152.
cies would not be identifiable as the DARWIN, C. 1859. On the origin of species by
means of natural selection, or the preservation of
favoured races in the struggle for life. John Mur-
4 Of course it is impossible to verify (in the ab- ray, London.
solute sense) that B is a sample of the ancestral DOBZHANSKY, T. 1970. Genetics of the evolution-
continuum leading to C. But this is, again, a fault ary process. Columbia University Press, New
of our system of observations and is common to all York.
observational sciences. It is analogous to the wol- DOUGHERTY, E. C. 1959. The origin of sexuality.
verine problem and to the problem of "verifying" Syst. Zool. 4:145-169.
that similar features are homologous. One could EHRLICH, P. 1961. Has the biological species con-
corroborate the hypothesis by showing that a series cept outlived its usefulness? Syst. Zool. 10:167-
of intermediate samples between B and C formed 176.
a statistically insignificant continuum (as a whole) EHRLICH, P., AND P. H. RAVEN. 1969. Differentia-
and could refute the hypothesis by showing the op- tion of populations. Science 165:1228-1231.
posite. ELDREDGE, N., AND S. J. GOULD. 1972. Punctuated

This content downloaded from 199.17.62.177 on Wed, 07 Feb 2018 22:48:18 UTC
All use subject to http://about.jstor.org/terms
 26 SYSTEMATIC ZOOLOGY

equilibria: an alternative to phyletic gradualism. MEGLITSCH, P. A. 1954. On the nature of the spe-
In Schopf, T.J.M. (ed.), Models in paleontology. cies. Syst. Zool. 3:49-65.
Freeman, Cooper, and Co., San Francisco, pp. NELSON, G. J. 1973. Classification as an expression
82-115. of phylogenetic relationships. Syst. Zool. 22:344-
ENGELMANN, G. F., AND E. 0. WILEY. 1977. The 359.
place of ancestor-descendant relationships in PITELKA, F. A. 1951. Speciation and ecological dis-
phylogeny reconstruction. Syst. Zool. 26:1-11. tribution in American jays of the genus Aphelo-
FARRIS, J. S. 1976. Phylogenetic classification of coma. Univ. California Publ. Zool. 50:195-464.
fossils with recent species. Syst. Zool. 25:271- PLATNICK, N. I. 1976. Are monotypic genera pos-
282. sible? Syst. Zool. 25:198-199.
GHISELIN, M. T. 1966. On psychologism in the log- PLATNICK, N. I. 1977. Monophyly and the origin
ic of taxonomic controversies. Syst. Zool. 15:207- of higher taxa: a reply to E. 0. Wiley. Syst. Zool.
215. 26:355-357.
GHISELIN, M. T. 1969. The triumph of the Dar- PONTECORVO, G. 1958. Present trends in genetic
winian method. University of California Press, analysis. Columbia University Press, New York.
Berkeley. POPPER, K. R. 1968. The logic of scientific discov-
GHISELIN, M. T. 1974. A radical solution to the ery. Harper Torch Books, New York.
species problem. Syst. Zool. 23:536-544. ROSEN, D. E. (in press). Vicariant patterns and his-
GILMOUR, J. S. L. 1940. Taxonomy and philosophy. torical explanation in biogeography. Syst. Zool.
In Huxley, J. S. (ed.), The new systematics. Clar- ROSEN, D. E. (in press). Fishes from the uplands
endon Press, London, pp. 461-468. and intermontane basins of Guatemala. Bull.
GOLDSCHMIDT, R. 1940. The material basis of evo- Amer. Mus. Nat. Hist.
lution. Yale University Press, New Haven. SCHINDEWOLF, 0. H. 1950. Grundfragen der pa-
GRANT, V. 1971. Plant speciation. Columbia Uni- laontologie. Schweizerbart'sche Verlagbuchhand-
versity Press, New York. lung, Stuttgart.
HEMPEL, C. G. 1965. Aspects of scientific expla- SIMPSON, G. G. 1961. Principles of animal taxon-
nation and other essays in the philosophy of sci- omy. Columbia University Press, New York.
ence. The Free Press, New York. SLOBODCHIKOFF, C. N. (ed.). 1976. Concepts of
HEMPEL, C. G. 1966. Philosophy of natural sci- species. University of California Press, Berkeley.
ence. Prentice-Hall, Inc., Englewood Cliffs, New SNEATH, P. H. A., AND R. R. SOKAL. 1973. Nu-
Jersey. merical taxonomy. W. H. Freeman, San Francis-
HENNIG, W. 1966. Phylogenetic systematics. Uni- co.

versity of Illinois Press, Urbana. SOKAL, R. R. 1973. The species problem reconsid-
HULL, D. L. 1968. The operational imperative- ered. Syst. Zool. 22:360-374.
sense and nonsense in operationalism. Syst. Zool. SOKAL, R. R., AND T. J. CROVELLO. 1970. The bi-
17:438-457. ological species concept: a critical evaluation.
HULL, D. L. 1971. Contemporary systematic phi- Amer. Nat. 104:127-153.
losophies. Ann. Rev. Ecol. Syst. 1:19-54. STEBBINS, R. C. 1960. The comparative evolution
HULL, D. L. 1976. Are species really individuals? of genetic systems. In Tax, S. (ed.), The evolution
Syst. Zool. 25:174-191. of life. University of Chicago Press, Chicago, pp.
HUXLEY, J. S. 1940. Towards the new systematics. 197-226.
In Huxley, J. S. (ed.), The new systematics. Clar- UZZELL, T., AND J. C. BARRY. 1971. Leposoma
endon Press, London, pp. 1-46. percarinatum, a unisexual species related to L.
KURTEN, B., AND R. RAUSCH. 1959. Biometric com- quianense; and Leposoma ioanna, a new species
parisons between North American and European from Pacific Coastal Colombia (Sauria: Teiidae.).
mammals. Acta Arctica, Fasc. IX. E. Munksqaard, Postilla 154:39 pp.
Copenhagen, pp. 1-44. VAN VALEN, L. 1976. Ecological species, multi-
LQVTRUP, S. 1974. Epigenetics, a treatise in the- species, and oaks. Taxon 25:233-239.
oretical biology. John Wiley and Sons, New York. VENDLER, Z. 1975. Possibility of possible worlds.
MACARTHUR, R. H., AND R. LEVINS. 1967. The Canadian J. Phil. 5:57-72.
limiting similarity of convergence and diver- WILEY, E. 0. 1977. Are monotypic genera para-
gence of coexisting species. Amer. Nat. 101:377- phyletic? A reply to Norman Platnick. Syst. Zool.
385. 26:352-355.
MAYR, E. 1957. Species concepts and definitions. WRIGHT, J. W. 1967. A new uniparental whiptail
In Mayr, E. (ed.), The species problem. Amer. lizard (Genus Cnemidophorus) from Sonora,
Assoc. Adv. Sci. Publ. 50:1-22. Mexico. Arizona Acad. Sci. 4:185-193.
MAYR, E. 1963. Animal species and evolution. ZWEIFEL, R. G. 1965. Variation in and distribution
Harvard University Press, Cambridge. of the unisexual lizard Cnemidophorus tessela-
MAYR, E. 1969. Principles of systematic zoology. tus. Amer. Mus. Novitates 2235:49 pp.
McGraw-Hill, New York.
MAYR, E. 1976. Is the species a class or an indi- Manuscript received April 1977
vidual? Syst. Zool. 25:192. Revised August 1977

This content downloaded from 199.17.62.177 on Wed, 07 Feb 2018 22:48:18 UTC
All use subject to http://about.jstor.org/terms