cu

The Species Problem

A Symposium Presented at the Atlanta Meet-
ing of the American Association for the
Advancement of Science, Decemrer 28-29, 1955

Edited by

ERNST MAYR

Publication No. 50 of the

AMERICAN ASSOCIATION FOR THE ADVANCEMENT OF SCIENCE

Washington, D.C., 1957

© 1957

The American Association for the

Advancement of Science

Library of Congress Catalog
Card Number 57-11321

• ^ c

PREFACE

Few biological problems have remained as consistently chal-
lenging through the past two centuries as the species problem.
Time after time attempts were made to cut the Gordian knot and
declare the species problem solved either by asserting dogmat-
ically that species did not exist or by defining, equally dogmat-
ically, the precise characteristics of species. Alas, these pseudo-
solutions were obviously unsatisfactory. One might ask: "Why
not simply ignore the species problem?" This also has been tried,
but the consequences were confusion and chaos. The species is a
biological phenomenon that cannot be ignored. Whatever else the
species might be, there is no question that it is one of the primary
levels of integration in many branches of biology, as in syste-
matics (including that of microorganisms), genetics, and ecology,
but also in physiology and in the study of behavior. Every living
organism is a member of a species, and the attributes of these
organisms can often best be interpreted in terms of this relation-
ship. This is particularly true in comparative studies.

The continued interest in the species problem thus requires no
apology. Indeed the interest in the species is perhaps greater now
than it has been at any other time during the last hundred years.
One reason for this is the increase in autecological investigations,
which include a joint study of the physiological and ecological
properties of species that has resulted in an active contact be-
tween physiology and ecology, with the species the area of over-
lapping interest. A second reason is the current tendency in popu-
lation genetics to study the interaction, rather than the action, of
genes, and thus to lead to a study of gene pools of which the
species is the largest. A third is the emergence of a rejuvenated
new systematics that devotes much of its attention to the intimate
study of the population structure of species, a new school that is
strongly represented in the southeastern United States. It is not
surprising therefore that the species problem received more votes
than any other subject when Dr. J. G. Carlson, Chairman of Sec-

iii

iv PREFACE

tion F (Zoology) of the American Association for the Advance-
ment of Science, together with the officers of the Association of
Southeastern Biologists, asked for suggestions on topics for a sym-
posium at the 1955 meeting of the AAAS in Atlanta. These two
organizations therefore decided to sponsor a symposium on the
species problem and invited me to serve as chairman. The Asso-
ciation's Section G (Botany), the American Society of Parasitolo-
gists, and the Society of Systematic Zoology agreed to cosponsor
the symposium. The strong backing of the symposium by these
organizations greatly contributed to its success.

When selecting participants in the symposium one of my con-
siderations was to avoid duplication of similar earlier symposia.
Verne Grant, John Imbrie, and Hampton L. Carson, who had not
recently spoken on the species problem, were invited to discuss
the species problem in plants, in paleontology, and in genetics.
Their fresh approach was supplemented by treatment of groups
of organisms or topics that had not before been related to the
species problem: John Langdon Brooks agreed to speak on fresh-
water organisms, T. M. Sonneborn on protozoans, C. Ladd Prosser
on the physiological aspects of the species problem, and John A.
Moore on the species problem as seen by an embryologist.

The papers are published essentially as given at Atlanta, with
one exception. Dr. Sonneborn, when preparing his contribution,
realized the need of a comprehensive survey of the species prob-
lem in the protozoans and particularly in the ciliates with their
interesting and versatile modes of reproduction. He found thai
the nature of their population structure, whether they are in-
breeders or outbreeders, is responsible for the specific aspects of
much of their reproductive specialization. To gather the material
with which to substantiate this new interpretation necessitated a
review of the widely scattered literature and could not be brought
to completion until well after the Atlanta meeting. Dr. Sonne-
born's masterly synthesis, a fundamental contribution to the pro-
tozoan literature, was not available for the chairman's summation.

It would be too much to expect this symposium to solve the
species problem, yet it made a solid contribution toward its solu-

PREFACE v

tion. It broadened the base on which to discuss the problem by
utilizing new organisms. It led to a delimitation of the areas of
general agreement among biologists. It presented a clear state-
ment of the various species concepts and frankly stated and
enumerated difficulties in their application to different types of
natural populations. Finally, it illuminated certain aspects of the
ageless species problem that had been neglected previously, and
it attempted a statement of still controversial issues. From these
papers it should be evident that the species problem is still one
of the important issues in biology.

Ernst Mayr

CONTRIBUTORS

John Langdon Brooks, Osborn Zoological Laboratory, Yale Uni-
versity, New Haven, Connecticut

Hampton L. Carson, Department of Zoology, Washington Uni-
versity, St. Louis, Missouri

Verne Grant, Rancho Santa Ana Botanic Garden and Claremont
Graduate School, Claremont, California

John Imbrie, Department of Geology, Columbia University, New
York, New York

Ernst Mayr, Museum of Comparative Zoology, Harvard Univer-
sity, Cambridge, Massachusetts

John A. Moore, Department of Zoology, Barnard College and
Columbia University, New York, New York

C. Ladd Prosser, Physiology Department, University of Illinois,
Urbana, Illinois

T. M. Sonneborn, Department of Zoology, Indiana University,
Bloomington, Indiana

Vll

CONTENTS

Species Concepts and Definitions

Ernst Mayr 1

The Species as a Field for Gene Recombination

Hampton L. Carson 23

The Plant Species in Theory and Practice

Verne Grant 39

The Species Problem in Freshwater Animals

John Langdon Brooks 81

The Species Problem with Fossil Animals

John Imbrie 125

Breeding Systems, Reproductive Methods, and Species
Problems in Protozoa

T. M. Sonneborn 155

An Embryologist's View of the Species Concept

John A. Moore 325

The Species Problem from the Viewpoint of a Physiologist

C. Ladd Prosser 339

Difficulties and Importance of the Biological Species

Ernst Mayr 371

Index 389

73475

IX

SPECIES CONCEPTS AND DEFINITIONS

1 e

ERNST MAYR: museum of comparative zoolog%

HARVARD COLLEGE, CAMBRIDGE, MASSACHUSETTS

The importance of one fact of nature is being recognized to an
ever increasing extent: that the living world is comprised of more
or less distinct entities which we call species. Why are species so
important? Not just because they exist in huge numbers, and be-
cause each species, when properly studied, turns out to be differ-
ent from every other, morphologically and in many other respects.
Species are important because they represent an important level
of integration in living nature. This recognition is fundamental to
pure biology, no less than to all subdivisions of applied biology.
An inventory of the species of animals and plants of the world
is the base line of further research in biology. Whether he realizes
it or not, every biologist — even he who works on the molecular
level — works with species or parts of species and his findings may
be influenced decisively by the choice of a particular species.
The communication of his results will depend on the correct iden-
tification of the species involved, and thus, on its taxonomy.

Yet, when I was first approached by the Chairman of the Di-
vision of Zoology of the American Association for the Advance-
ment of Science to organize a symposium on the species problem
I was, to put it mildly, hesitant. Much discussion of this subject
in recent years suggested that there was perhaps no need for
such a symposium. Ensuing correspondence, however, convinced
me otherwise, and certain publications showed clearly that fur-
ther thinking on this subject is welcome, if not necessary. The
species problem continues to be one of the most disputed sub-
jects in biology, in spite of the intense preoccupation with it
during the past two hundred years. The recent publications by
Spurway, Burma, and Arkell attest this. This symposium can be

1

2 CONCEPTS AND DEFINITIONS

considered a success if it throws light on some of the disputed
questions or even if it does nothing more than lead to a more
precise phrasing of the basic points of disagreement.

One way of laying a foundation for such an investigation is to
recall some of its history. Who was the first to realize that there
is a species problem and what was his proposed solution? What
were the subsequent developments? Time does not permit a
thorough coverage of the field, but even a glance at the high
lights is revealing. If we open a history of biology, the two names
mentioned most prominently under the heading of "Species" will
be Linnaeus and Darwin. Linnaeus will be cited as the champion
of two characteristics of the species, their constancy and their
sharp delimitation (their "objectivity" ) . One of the minor trag-
edies in the history of biology has been the assumption during
the hundred and fifty years after Linnaeus that constancy and
clear definition of species are strictly correlated and that one
must make a choice of either believing in evolution (the "incon-
stancy" of species) and then having to deny the existence of
species except as purely subjective, arbitrary figments of the
imagination, or, as most early naturalists have done, believing in
the sharp delimitation of species but thinking that this necessi-
tated denying evolution. We shall leave the conflict at this point
and merely anticipate the finding made more than a hundred
years after Linnaeus that there is no conflict between the fact
of evolution and the fact of the clear delimitation of species in a
local fauna or flora.

The insistence of Linnaeus on the reality, objectivity, and con-
stancy of species is of great importance in the history of biology
for three reasons. First, it meant the end of the belief in spon-
taneous generation as far as higher organisms are concerned, a
belief which at that time was still widespread. Lord Bacon and
nearly all leading writers of the pre-Linnaean period, except Ray,
believed in the transmutation of species and the Linnaean con-
ception "of the reality and fixity of species perhaps marks a neces-
sary stage in the progress of scientific inquiry." (See Poulton,
1903, pp. lxxxiv-lxxxvii for further references on the sub-
ject). "Until about 1750 almost no one believed that species were

E. MAYR 3

stable. Linnaeus had to show that species were not erratic and
ephemeral units before organic evolution as we know it could
have any meaning." (Conway Zirkle in litt.) The idea that the
seed of one plant could occasionally produce an individual of
another species was so widespread that it died only slowly. We
all know that it raised its ugly head once more during the past
ten years. In spite of Redi's and Spallanzani's experiments spon-
taneous generation was still used in 1851 by the philosopher
Schopenhauer as an explanation for the origin of higher cate-
gories. Linnaeus thus did for the higher organisms what Pasteur
did one hundred years later for the lower.

A second reason why his emphasis was important is that it
took the species out of the speculations of the philosophers who
approached the species problem in the spirit of metaphysics and
stated, for instance, that "only individuals exist. The species of a
naturalist is nothing but an illusion" (Robinet, 1768). We shall
return later to the point why species are more than merely an
aggregate of individuals.

A third reason why the insistence on the sharp delimitation of
species in the writings of Linnaeus is of historical importance is
that it strengthened the viewpoint of the local naturalist and
established the basis for an observational and experimental study
of species in local faunas and floras, of which Darwin took full
advantage.

Linnaeus was too experienced a botanist to be blind to the
evidence of evolutionary change. Greene (1912) gathered nu-
merous citations from his writings which clearly document Lin-
naeus' belief in the common descent of certain species, and
Ramsbottom ( 1938 ) and Sirks ( 1952 ) have traced how Linnaeus
expressed himself more and more freely on the subject, as his
prestige grew. Paradoxically, Linnaeus did more, perhaps, to lay
a solid foundation for subsequent evolutionary studies by em-
phasizing the constancy and objectivity of species than if he,
like Darwin, had emphasized the opposite.

Darwin looked at the species from a viewpoint almost directly
opposite to that of Linnaeus. As a traveler naturalist and par-
ticularly because of his studies of domesticated plants and ani-

4 CONCEPTS AND DEFINITIONS

mals he was impressed by the fluidity of the species border and
the subjectivity of their delimitation. The views of both Linnaeus
and Darwin underwent a change during the life of each. With
Linnaeus the statements on the constancy of species became less
and less dogmatic through the years. In Darwin, as the idea of
evolution became firmly fixed in his mind, so grew his conviction
that this should make it impossible to delimit species. He finally
regarded species as something purely arbitrary and subjective. "I
look at the term species as one arbitrarily given for the sake of
convenience to a set of individuals closely resembling each other,
and that it does not essentially differ from the term variety which
is given to less distinct and more fluctuating forms. . . . The
amount of difference is one very important criterion in settling
whether two forms should be ranked as species or variety." And
finally he came to the conclusion that "In determining whether
a form should be ranked as a species or a variety, the opinion of
naturalists having sound judgment and wide experience seems
the only guide to follow" (Darwin, 1859). Having thus elimi-
nated the species as a concrete unit of nature, Darwin had also
neatly eliminated the problem of the multiplication of species.
This explains why he made no effort in his classical work to
solve the problem of speciation.

The seventy-five years following the publication of the Origin
of Species (1859) saw biologists rather clearly divided into two
camps, which we might call, in a somewhat oversimplified man-
ner, the followers of Darwin and those of Linnaeus. The followers
of Darwin, which included the plant breeders, geneticists, and
other experimental biologists minimized the "reality" or objectiv-
ity of species and considered individuals to be the essential units
of evolution. Characteristic for this frame of mind is a symposium
held in the early Mcndelian days, which endorsed unanimously
the supremacy of the individual and the nonexistence of species.
Statements made at this symposium (Bcssey, 1908) include the
following: "Nature produces individuals and nothing more. . . .
Species have no actual existence in nature. They are mental con-
cepts and nothing more. . . . Species have been invented in order
that we may refer to great numbers of individuals collectively."

E. MAYR 5

Taxonomists, one of the speakers claimed, did not merely name
the species found in nature but actually "made" them. "In
making a species the guiding principle must be that it shall be
recognizable from its diagnosis." A leftover from this period is
the statement of a recent author: "Distinct species must be sep-
arable on the basis of ordinary preserved material."

It is a curious paradox in the history of biology that the redis-
covery of the Mendelian laws resulted in an even more unrealistic
species concept among the experimentalists than had existed
previously. They either let species saltate merrily from one to
another, as did Bateson and DeVries, defining species merely
as morphologically different individuals, or they denied the ex-
istence of species altogether except as inter/grading populations.
Whether these early Mendelians considered species as continuous
or discontinuous units, they all agreed in their arbitrariness and
artificiality. There is an astonishing absence of any effort in this
school to study species in nature, to study natural populations.

A study of natural populations had become the prevailing pre-
occupation in an entirely independent conceptual stream,- that
of the naturalists, which ultimately traces back to Linnaeus. The
viewpoint of the naturalist was particularly well expressed by
Jordan (1905), who stated "The units of which the fauna of a
region is composed are separated from each other by gaps which,
at a given place, are not bridged by anything. This is a fact
which can be checked by any observer. Indeed, the activity of
a local naturalist begins with the searching out of these units
which with Linnaeus we call species." (For a more detailed
discussion see Mayr, 1955. ) Although this was the prevailing
viewpoint among taxonomists, it was completely ignored by the
general biologists by whom, as a result of Darwin's theory, "Spe-
cies were mostly regarded merely as arbitrary divisions of the
continuous and ever changing series of individuals found in na-
ture ... of course, active taxonomists did not overlook the
existence of sharply and distinctly delimited species in nature —
but as the existence of those distinct units disagreed with the
prevailing theories, it was mentioned as little as possible" (Du
Rietz, 1930). The two streams of thought are still recognizable

6 CONCEPTS AND DEFINITIONS

today even though most geneticists, under the leadership of
Dobzhansky, Huxley, Ford, and others, have swung into Jordan's
camp. The principal opponents of the concept of objectively
delimitable species are today found among philosophers and
paleontologists. Publications maintaining this viewpoint are those
of Gregg (1950), Burma (1949, 1954), Yapp (1951), and Arkell
( 1956 ) . These are only the most recent titles in a vast literature,
some of which is cited in the bibliography.

The point which is perhaps most impressive when one studies
these voluminous publications is the amount of disagreement that
has existed and still exists. The number of possible antitheses
that have been established in this field may be characterized by
such alternate views, to mention only a few, as follows:

Subjective versus objective;

Scientific versus purely practical;

Degree of difference versus degree of distinctness;

Consisting of individuals versus consisting of populations;

Only one kind of species versus many kinds of species;

To be defined morphologically versus to be defined biologically.

To give a well-documented history of the stated controversies
would fill a book. As interesting as this chapter in the history
of human thought is, the detailed presentation of the gropings
and errors of former generations would add little to the task
before us. Let us concentrate therefore on the gradual emergence
of the ideas which we, today, consider as central and essential.
Three aspects are stressed in most modern discussions of species,
that (1) they are based on distinctness rather than on difference
and are therefore to be defined biologically rather than morpho-
logically, (2) they consist of populations, rather than of uncon-
nected individuals, a point particularly important for the solution
<>l the problem of speciation, (3) they are more succinctly de-
fined by isolation from non-conspecific populations than by the
relation of conspecific individuals to each other. The crucial
species criterion is thus not the fertility of individuals, but rather

E. MAYR 7

the reproductive isolation of populations. Let us try to trace the
emergence of these and related concepts.

It is not surprising that species were considered merely "cate-
gories of thought" by many writers in periods so strongly
dominated by idealistic philosophy as were the eighteenth and
nineteenth centuries. Thoughts as that expressed in the above
quoted statement of Robinet were echoed by Agassiz, Mivart, and
particularly among those paleontologists who considered their
task merely the classification of "objects" ( = fossil specimens ) .
In opposition to this, an increasingly strong school developed
which considered species as "definable," "objective," "real." Lin-
naeus was, of course, the original standard bearer of this school
to which also belonged Cuvier, De Candolle, and many taxon-
omists in the first half of the nineteenth century. They supported
their case sometimes by purely morphological arguments such
as Godron ( 1853 ) who stated : "c est un fait incontestable que
toutes les especes animales et vegetables se separent les unes des
autres par de caracteres absolues et tranchees." Others used a
more biological argument, as I will discuss below.

What is unexpected for this pre-Darwinian period, however,
is the frequency with which "common descent" is included in
species definitions. When such an emphatically anti-evolutionary
author as v. Baer (1828) defines the species as "the sum of the
individuals that are united by common descent," it becomes
evident that he does not refer to evolution. What is really meant
is more apparent from Ray's species definition (1686) or a state-
ment by the Swedish botanist Oeder ( 1764 ) that it characterizes
species "dass sie aus ihres gleichen entsprungen seien und wieder
ihres gleichen erzeugen." Expressions like "community of origin"
or "individus descendants des parents communs" (Cuvier) are
frequent in the literature. These are actually attempts at recon-
ciling a typological species concept ( with its stress of constancy )
with the observed morphological variation. Constancy was a
property of species taken very seriously not only by Linnaeus
and his followers but curiously enough also by Lamarck and by
Darwin himself: "The power of remaining constant for a good

8 CONCEPTS AND DEFINITIONS

long period I look at as the essence of a species" (letter to
Hooker, Oct. 22, 1864). Such constancy in time was the strongest

gument in favor of a morphological species concept, but it
could be proved only by the comparison of individuals of differ-
ent generations. Different morphological "types" that are no more
different than mother and daughter or father and son can safely
be considered as conspecific. They are "of the same blood." It is
obvious that this early stress of descent was essentially the
consequence of a morphological species concept. Yet this con-
sideration of descent eventually led to a genetic species definition.

Virtually all early species definitions regarded species only
as aggregates of individuals, unconnected except by descent, as
is evident not only from the writings of Robinet, Buffon, and
Lamarck, but also of much more recent authors (e.g., Britton,
1908; Bessey, 1908). The realization that these individuals are
held together by a supraindividualistic bond, that they form
populations, came only slowly. Illiger ( 1800 ) spoke of species
as a community of individuals which produce fertile offspring.
Brauer (1885) spoke of the "natural tie of blood relationship"
through which the "individuals of a species are held together,"
and which "is not a creation of the human mind ... if species
were not objective, it would be incomprehensible that even the
most similar species mix only exceptionally and the more distant
species never." Plate (1914) was apparently the first to state ex-
plicitly the nature of this bond: "The members of a species arc
tied together by the fact that they recognize each other as be-
longing together and reproduce only with each other. The sys-
tematic category of the species is therefore entirely independent
of the existence of Man." Finally, in the language of current
population genetics this community becomes the "co-adapted
gene pool," again stressing the integration of the members of
the population rather than the aggregation of individuals (a
viewpoint which is of course valid only for sexually reproducing
organisms).

The growth of thinking in terms of populations went hand in
hand with a growing realization that species were less a matter
of difference than of distinctness. "Species" in its earlier typolog-

E. MAYR 9

ical version meant merely "kind of." This, as far as inanimate
objects are concerned, is measured in terms of difference. But
one cannot apply this same standard to "kinds of" organisms,
because there are various biological "kinds." Males and females
may be two very different "kinds" of animals. Jack may be a
very different "kind" of a person from Bill, yet neither "kind"
is a species. Realization of the special aspects of biological varia-
tion has led to a restriction in the application of the term species
to a very particular "kind," namely the kind that would inter-
breed with each other. The first three authors found * by me who
state this clearly are Voigt (1817), "Man nennt Spezies . . . was
sich fruchtbar mit einander gattet, fortpflanzt"; Oken (1830),
"Was sich scharet und paaret, soil zu einer Art gerechnet wer-
den"; and Gloger ( 1833 ) , "What under natural conditions regu-
larly pairs, always belongs to one species." (He stated that by
stressing "regularly" he wanted to eliminate the complications
due to occasional hybridization.) Gloger later (1856) gave a
different, but similar definition: "A species is what belongs to-
gether either by descent or for the sake of reproduction." It is
interesting how completely all these definitions omit any refer-
ence to morphological criteria. They are obviously inapplicable
to asexually reproducing organisms.

This is an exceedingly short outline of some of the trends in
the development of a modern species concept. More extensive
treatments can be found in the publications of Geoffrey St. Hi-
laire (1859), Besnard (1864), de Quatrefages (1892), Bachmann
(1905), Plate (1914), Uhlmann (1923), Du Rietz (1930), Kuhn
(1948), and other authors cited in the bibliography. Several
conclusions are self-evident. One is that biological or so-called
modern species criteria were already used by authors who pub-
lished more than one hundred years ago, long before Darwin.
Another is that a steady clarification is evident, yet that there is
still much uncertainty and widespread divergence of opinion
on many aspects of the species problem. It is rather surprising
that not more agreement has been reached during the past two

* Still earlier statements can no doubt be found in the extensive litera-
ture, particularly on hybridization.

ft- * A - *

10 CONCEPTS AND DEFINITIONS

hundred years in which these questions have been tossed back
and forth. This certainly cannot be due to lack of trying, for an
immense amount of time and thought has been devoted to the
subject during this period. One has a feeling that there is a hid-
den reason for so much disagreement. One has the impression
that the students of species are like the three blind men who
described the elephant respectively as a rope, a column, or a
giant snake when touching its tail, its legs, and its trunk.

Perhaps the disagreement is due to the fact that there is more
than one kind of species and that we need a different definition
for each of these species. Many attempts have been made during
the last hundred years to distinguish these several kinds of
species, among the most recent being those of Valentine ( 1949 )
and Cain (1953). Camp and Gillis (1943) recognized no less
than twelve different kinds of species. Yet, a given species in
nature might fit into several of their categories, and in view of
tli is overlap no one has adopted either this elaborate classifica-
tion or any of the simpler schemes proposed before or after-
wards.

Species Concepts

An entirely different approach to the species problem stresses
the kaleidoscopic nature of any species and attempts to determine
how many different aspects a species has. Depending on the
choice of criteria, it leads to a variety of "species concepts" or
"species definitions." At one time I listed five species concepts,
which I called the practical, morphological, genetic, sterility,
and biological (Mayr, 1942). Meglitsch (1954) distinguishes
three concepts, the phenotypic, genetic, and phylogenetic, a
somewhat more natural arrangement. Two facts emerge from
these and other classifications. One is that there is more than
one species concept and that it is I utile to search for the species
concept. The second is that there are at least two levels of con-
cepts. Such terms as "practical," "sterility," "genetic" signify con-
crete aspects of species which lead to what one might call "ap-
plied species concepts. They specify criteria which can be
applied readily to determine the status of discontinuities found

E. MAYR 11

in nature. Yet they are secondary, derived concepts, based on
underlying philosophical concepts, which might also be called
primary or theoretical concepts. I believe that the analysis of
the species problem would be considerably advanced, if we could
penetrate through such empirical terms as phenotypic, morpho-
logical, genetic, phylogenetic, or biological, to the underlying
philosophical concepts. A deep, and perhaps widening gulf has
existed in recent decades between philosophy and empirical
biology. It seems that the species problem is a topic where pro-
ductive collaboration between the two fields is possible.

An analysis of published species concepts and species defini-
tions indicates that all of them are based on three theoretical
concepts, neither more nor less. An understanding of these three
philosophical concepts is a prerequisite for all attempts at a
practical species definition. And all species criteria or species
definitions used by the taxonomist in his practical work trace
back ultimately to these basic concepts.

The Typological Species Concept. This is the simplest and
most widely held species concept. Here it merely means "kind
of." There are languages, as for instance German, where the term
for "kind" (Art) is also used for "species." A species in this con-
cept is "a different thing." This concept is very useful in many
branches of science and it is still used by the mineralogist who
speaks of "species of minerals" (Niggli, 1949) or the physicist
who speaks of "nuclear species." This simple concept of every-
day life was incorporated in a more sophisticated manner in the
philosophy of Plato. Here, however, the word eidos (species, in
its Latin translation) acquired a double meaning that survives
in the two modern words "species" and "idea" both of which are
derived from it. According to Plato's thinking objects are merely
manifestations, "shadows," of the eidos. By transfer, the indi-
viduals of a species, being merely shadows of the same type, do
not stand in any special relation to each other, as far as a typol-
ogist is concerned. Naturalists of the "idealistic" school endeavor
to penetrate through all the modifications and variations of a
species in order to find the "typical" or "essential" attributes.
Typological thinking finds it easy to reconcile the observed

12 CONCEPTS AND DEFINITIONS

variability of the individuals of a species with the dogma of the
constancy of species because the variability does not affect the
essence of the eidos, which is absolute and constant. Since the
eidos is an abstraction derived from individual sense impressions,
and a product of the human mind, according to this school, its
members feel justified in regarding a species "a figment of the
imagination," an idea. Variation, under this concept, is merely an
imperfect manifestation of the idea implicit in each species. If the
degree of variation is too great to be ascribed to the imperfections
of our sense organs, more than one eidos must be involved. Thus
species status is determined by degrees of morphological differ-
ence. The two aspects of the typological species concept, subjec-
tivity and definition by degree of difference, therefore depend on
each other and are logical correlates.

The application of the typological species concept to practical
taxonomy results in the morphologically defined species, "degree
of morphological difference" is the criterion of species status.
Species are defined on the basis of their observable morphological
differences. This concept has been carried to the extreme where
mathematical formulas were proposed (Ginsburg, 1938) that
would permit an unequivocal answer to the question whether or
not a population is a different species.

Most systematists found this typological-morphological concept
inadequate and have rejected it. Its defenders, however, claim
that all taxonomists, when classifying the diversity of nature into
species, follow the typological method and distinguish "arche-
types." At first sight there seems an element of truth in this
assertion. When assigning specimens either to one species or to
another, the taxonomist bases his decision on a mental image of
these species that is the result of past experience with the stated
species. The utilization of morphological criteria is valuable and
productive in the taxonomic practice. To assume, however, that
this validates the typological species concept overlooks a number
of important considerations. To begin with, the mental construct
of the "type" is subject to continuous revision under the impact
of new information. If it is found that two archetypes represent
nothing more than two "kinds" within a biological species, they

E. MAYR 13

are merged into a single one. It was pointed out above that
males and females are often exceedingly different "kinds" of ani-
mals. Even more different are in many animals the larval stages,
or in plants sporophyte and gametophyte, or in polymorph popu-
lations the various genotypes. A strictly morphological-typological
concept is inadequate to cope with such intraspecific variation.
It is equally incapable of coping with another difficulty, namely
an absence of visible morphological differences between natural
populations which are nevertheless distinct and reproductively
isolated, and therefore to be considered species. The frequent
occurrence of such "cryptic species" or "sibling species" in nature
has been substantiated by various genetic, physiological, or
ecological methods. They form another decisive argument against
defining species on a primarily morphological basis. Any attempt
in these two situations to define species "by degree of difference"
is doomed to failure. Degree of difference can be specified only
by a purely arbitrary decision.

More profound than these two essentially practical considera-
tions is the fact that the typological species concept treats species
merely as random aggregates of individuals which have the "es-
sential properties" of the "type" of the species and "agree with
the diagnosis." This static concept ignores the fact that species
are not merely classes of objects but are composed of natural
populations which are integrated by an internal organization and
that this organization (based on genetic, ethological, and eco-
logical properties) gives the populations a structure which goes
far beyond that of mere aggregates of individuals. Even a house
is more than a mere aggregate of bricks or a forest an aggregate
of trees. In a species an even greater supraindividualistic cohesion
and organization is produced by a number of factors. Species are
a reproductive community. The individuals of a species of higher
animals recognize each other as potential mates and seek each
other for the purpose of reproduction. A multitude of devices
insures intraspecific reproduction in all organisms. The species
is an ecological unit which, regardless of the individuals of which
it is composed, interacts as a unit with other species in the same
environment. The species, finally, is a genetic unit consisting of a

14 CONCEPTS AND DEFINITIONS

large intercommunicating gene pool whereas each individual
is only a temporary vessel holding a small portion of this gene

ol for a short period of time. These three properties make the
species transcend a purely typological interpretation or the con-
cept of a "class of objects."

The very fact that a species is a gene pool, with numerous
devices facilitating genie intercommunication within and genie
separation from without, is responsible for the morphological
distinctness of species as a byproduct of their biological unique-
ness. The empirical observation that a certain amount of mor-
phological difference between two populations is normally corre-
lated with a given amount of genetic difference is undoubtedly
correct. Yet, it must be kept in mind at all times that the biolog-
ical distinctness is primary and the morphological difference sec-
ondary. As long as this is clearly understood, it is legitimate and
indeed very helpful to utilize morphological criteria. This caution
has been exercised, consciously or unconsciously, by nearly all
proponents of the morphological species concept. As pointed out
by Simpson (1951) and Meglitsch (1954), they invariably
abandon the morphological concept when it comes in conflict
with biological data. This was true for Linnaeus himself and for
his followers to the present day.

The typological species concept has a certain amount of opera-
tional usefulness when applied to inanimate objects. Ignoring
the population structure of species, however, and incapable of
coping with the facts of biological variation, it has proved
singularly inadequate as a conceptual basis in taxonomy. Much
of the criticism directed against the taxonomic method was pro-
voked by the application of the typological concept by taxono-
ii lists themselves or by other biologists who mistakenly con-
sidered it the basis of taxonomy.

The Second Species Concept. This is sometimes called the
aondimensional species concept and has no generally accepted
designation. The essence of this concept is the relationship of
two coexisting natural populations in a nondiinensional system,
that is, at a single locality at the same time (sympatric and syn-
chronous). This is the species concept of the local naturalist. It

E. MAYR 15

was introduced into the biological literature by the English
naturalist John Ray and confirmed by the Swedish naturalist
Linnaeus. It is based not on difference but on distinction, and
this distinction in turn is characterized bv a definite mutual re-
lationship, namely that of reproductive isolation. The word
"species" is here best defined in combination with the word
"different." The relationship of two "different species" can be
objectively defined as reproductive isolation. We have, thus, an
objective yardstick for this species concept, something that is ab-
sent in all others. Philosophers have objected to the use of the
terms "objective" or "real" for species, and it may be more
neutral to use the terms arbitrary or nonarbitrary (Simpson,
1951). Presence or absence of interbreeding of two populations
in a nondimensional system is a completely nonarbitrary criterion.
Since the nondimensional species concept is based on a relation-
ship, the word species is here equivalent to words like, let us say,
the word brother, which also has a meaning only with respect
to a second phenomenon. An individual is a brother only with
respect to someone else. Being a brother is not an inherent prop-
erty as hardness is a property of a stone. Describing a presence
or absence relationship makes this species concept nonarbitrary.

This species concept seems so self-evident to every naturalist
that it is only rarely put in words. That the species is more than
an aggregate of individuals, held together by a biological bond,
has long been realized, as was pointed out in the historical survey
above. The interbreeding within the species is more conspicuous,
and it was thus more often emphasized than is the reproductive
isolation against other species. Eimer, as early as 1889 (p. 16)
defined species as "groups of individuals which are so modified
that successful interbreeding [with other groups] is no longer
possible." The first author, however, who stated the nondimen-
sional species concept in its full extent and implication was
Jordan (1905).

In spite of its theoretical superiority, the nondimensional
species has a number of serious drawbacks (which will be dis-
cussed later), particularly its limitation to sexually reproducing
species and to such without the dimensions of space and time.

16 CONCEPTS AND DEFINITIONS

Yet, as a basic, nonarbitrary yardstick, this is the species concept
on which we have to fall back whenever we encounter a border-
line situation.

The Third Species Concept. This is a concept of an entirely
different kind, it is the concept of the polytypic or multidimen-
sional species. In contradistinction to the other two concepts, of
which one is based on a degree of difference, the second one on
the completeness of a discontinuity, this concept is a collective
one. It considers species as groups of populations, namely such
groups as interbreed with each other, actually or potentially.
Thus this species concept is a concept of the same sort as the
higher categories, genus, family, or order. Like all collective cate-
gories it faces the difficulty, if not impossibility, of clear demarca-
tion against other similar groupings. What this species gains in
actuality by the extension of the nondimensional situations in
space and time, it loses in objectivity. As unfortunate as this is,
it is inevitable since the natural populations, encountered by the
biologist, are distributed in space and time and cannot be di-
vorced from these dimensions. Thus, this species concept likewise
has its good and its bad points.

Species Definitions

All our reasoning in discussions of "the species" can be traced
back to the stated three primary concepts. As concepts, of course,
they cannot be observed directly, and we refer to certain ob-
served phenomena in nature as "species," because they conform
in their attributes to one of these concepts or to a mixture of
several concepts. From these primary concepts, just discussed,
we come thus to secondary concepts, based on particular aspects
of species. We have already mentioned the so-called morpho-
logical species concept, which, in most cases, is merely an ap-
plied typological concept, using morphological criteria. The case
of the so-called genetic species concept shows that all three of
the basic concepts can be expressed, on this level, in genetic
terms. Some geneticists, for instance, subscribed to the typolog-
ical concept and defined species by the degree of genetic differ-
ence as did Lotsy or DcVrics; others stressed the genetic basis

E. MAYR 17

of the isolating mechanisms between species thereby endorsing
the nondimensional species concept; still others finally empha-
sized the gene flow among interbreeding populations in a multi-
dimensional system, thus adopting the multidimensional collective
species concept. All three groups of geneticists thought they
were dealing with a uniquely "genetic species concept," yet they
were merely observing secondary manifestations of the primary
concepts.

It is evident from the analysis of the morphological and ge-
netic species concepts, that such derived concepts are attempts
to deal directly with the discontinuities in nature. In the past,
almost every taxonomist worked with his own personal yardstick
based on a highly individual mixture of elements from the three
basic concepts. As a consequence one taxonomist might call
species every polymorph variant, a second one every morpholog-
ically different population, and a third one every geographically
isolated population. Such lack of standards, which is still largely
characteristic for the taxonomic literature, has been utterly con-
fusing to taxonomists and other biologists alike. It has therefore
been the endeavor of many specialists within recent decades to
find a standard yardstick, on which there could be general agree-
ment. A historical study of species definitions indicates clearly
a trend toward acceptance of a synthetic species definition, often
referred to as "biological species" definition. It is essentially
based on the nondimensional ("reproductive gap") and the
multidimensional ("gene flow") species concepts. Nearly all
species definitions proposed within the last fifty years incorporate
some elements of these two concepts. This is evident from the
species definitions of Jordan (Mayr, 1955), Stresemann (1919),
and Rensch (1929). Du Rietz (1930) called the species "a
syngameon . . . separated from all others by . . . sexual isola-
tion." Dobzhansky (1935) was apparently the first geneticist to
define species in the terms customary among naturalists and
taxonomists, namely interbreeding and reproductive isolation;
other recent definitions are variants of the same theme. Mayr
(1940) defined species as "groups of actually or potentially in-
terbreeding natural populations which are reproductively iso-

!

: • « •> * a -- •

18 CONCEPTS AND DEFINITIONS

lated from other such groups." Simpson (1943) gave the defini-
tion "a genetic species is a group of organisms so constituted and
so situated in nature that a hereditary character of any one of
these organisms may be transmitted to a descendent of any
other," and Dobzhansky ( 1950 ) defined the species as "the
largest and most inclusive . . . reproductive community of sexual
and cross-fertilizing individuals which share in a common gene
pool."

It might be useful to mention some qualifications which are
often included in species definitions but needlessly so. Anything
that is equally true for categories above and below species rank
should be omitted, since there is no sense burdening a species
definition with features which do not help discrimination be-
tween species and infraspecific populations.

1. Species characters are adaptive. This component of Wal-
lace's ( 1889 ) species definition was correctly rejected by Jordan
(1896). Adaptiveness is not diagnostic for species characters
and not even necessarily true. Not every detail of the phenotype
needs to be adaptive as long as the phenotype as a whole is
adaptive and as long as the genotype itself is the result of selec-
tion.

2. Species are evolved and evolving. Again this is true for
the entire organic world from the individual to the highest cate-
gories and adds nothing to the species definition.

3. Species differ genetically. This is only the morphological
species concept expressed in genetic terms. It does not permit
discriminating species from infraspecific populations or from
individuals.

4. Species differ ecologically. This qualification is unneces-
sary and misleading for the same reasons as the genetic one.
Ecological differences exist for all ecotypes within species and
in general for all geographical isolates. Conspecific populations
are sometimes more different ecologically than are good species.

A yardstick, such as the biological species concept, is not auto-
matic. To apply it properly requires skill and experience. This
is particularly true in the recognition of situations where it cannot
be applied, lor one reason or another, and where the worker has

E. MAYR 19

to fall back on the criterion of "degree of difference." It will be
one of the tasks of this symposium to investigate to what extent
the diversity of animal and plant life permits application of the
standard yardstick of the biological species concept. When can
it not be applied and for what reasons? What types of difficulties
are there? And finally, are there perhaps other basic concepts
in addition to the stated ones? By approaching the species prob-
lem with new questions and new material, perhaps this sympo-
sium can make a contribution to its solution.

REFERENCES

Arkell, W. J. 1956. The species concept in paleontology. Systematics

Assoc. Publ. No. 2, pp. 97-99.
Bachmann, H. 1905. Der Speziesbegriff. Verhandl. schweiz. natur-

forsch. Ges., 87, 161-208.
Baer, K. E. von. 1828. Entwickelungs-Geschichte der Thiere. Konigs-

berg.
Besnard, A. F. 1864. Altes und Neues zur Lehre iiber die organische

Art (Spezies). Ahhandl. zool. mineral. Ver. Regensburg, 9, 1-72.
Bessey, C. E. 1908. The taxonomic aspect of the species question. Am.

Naturalist, 42, 218-24.
Brauer, F. 1885. Systematisch-zoologische Studien. Sitzber. Akacl.

Wiss. Wien, 91 (Abt. 1), 237-413.
Britton, N. L. 1908. The taxonomic aspect of the species question.

Am, Naturalist, 42, 225-42.
Burma, B. H. 1949a. The species concept: A semantic review. Evolu-
tion, 3, 369-70.
Burma, B. H. 1949b. The species concept: Postscriptum. Evolution,

3, 372-73.
Burma, B. H. 1954. Reality, existence, and classification: A discussion

of the species problem. Madrono, 7, 193-209.
Cain, A. J. 1953. Geography, ecology and coexistence in relation to

the biological definition of the species. Evolution, 7, 76-83.
Camp, W. H., and C. L. Gillis. 1943. The structure and origin of spe-
cies. Brittonia, 4, 323-85.
Darwin, C. 1859. On the Origin of the Species by Means of Natural

Selection. London.

20 CONCEPTS AND DEFINITIONS

Dobzhansky, T. 1935. A critique of the species concept in biology.

Phil. Sci., 2, 344-55.
Dobzhansky, T. 1950. Mendelian populations and their evolution.

Am. Naturalist, 84, 401-18.
Doederlein, L. 1902. Uber die Beziehungen nahe verwandter "Thier-

formen" zu einander. Z. Morphol. AnthropoL, 26, 23-51.
Dougherty, E. C. 1955. Comparative evolution and the origin of sex-
uality. Systematic ZooL, 4, 145-69.
Du Rietz, G. E. 1930. The fundamental units of botanical taxonomy.

Svensk. Bot. Tidsskr., 24, 333-428.
Eimer, G. H. T. 1889. Artbihhing und Vcrwandtschaft bei Schmetter-

lingen. Jena, Vol. II, p. 16.
Geoffroy Saint Hilaire, I. 1859. Histoire naturelle generate des regnes

organiques, 2, 437, Paris.
Ginsburg, I. 1938. Arithmetical definition of the species, subspecies

and race concept, with a proposal for a modified nomenclature.

Zoologica, 23, 253-86.
Gloger, C. L. 1833. Das Abandern der Vogel durch Einjluss des

Klimas. Breslau.
Gloger, C. L. 1856. Ueber den Begriff von "Art" ("Species") und was

in dieselbe hinein gehort. /. Ornithol., 4, 260-70.
Godron, D. A. 1853. De Vespece et des races dan les ctrcs organises et

specialment dc Vanite de Vespece humaine. Paris, 2 vols.
Greene, E. L. 1912. Linnaeus as an evolutionist. In Carolus Linnaeus,

pp. 73-91. C. Sower & Co., Philadelphia, Pa.
Gregg, J. R. 1950. Taxonomy, language and reality. Am. Naturalist,

84, 419-35.
Huxley, J. 1942. Evolution, the Modern Synthesis. Allen and Unwin,

London.
Illiger, J. C. W. 1800. Versuch ciner syslnnafischcn vollstandigcn

Tcrminologie fur das Thicrrcich und Pflanzenreich. Helmstedt.
Jordan, K. 1896. On mechanical selection and other problems. Novii.

ZooL, 3, 426-525.
Jordan, K. 1905. Der Gegensatz zwischen geographischer und nicht-

geographischer Variation. Z. wiss. ZooL, 83, 151-210.
Kuhn, E. 1948. Der Artbegriff in dor Palaontologie. Eclogue Geolog.

Helv., 41, 389-421.
Lorkovicz, Z. 1953. Spezifische, semispezifische und rassische Dif-

ferenzierung bei Erebia tyndarus Esp. Had. Acad. Yougoslave, 294,

315-58.

E. MAYR 21

Mayr, E. 1940. Speciation phenomena in birds. Am. Naturalist, 74,

249-78.
Mayr, E. 1942. Systcmatics and the Origin of Species. Columbia Uni-
versity Press, New York, N.Y.
Mayr, E. 1949. The species concept: Semantics versus semantics.

Evolution, 3, 371-72.
Mayr, E. 1951. Concepts of classification and nomenclature in higher

organisms and microorganisms. Ann. N.Y. Acad. Sci., 56, 391-97.
Mayr, E. 1955. Karl Jordan's contribution to current concepts in sys-

tematics and evolution. Trans. Roy. Entomol. Soc. London, pp. 45-66.
Mayr, E. 1956. Geographical character gradients and climatic adapta-
tion. Evolution, 10, 105-8.
Mayr, E., and C. Rosen. 1956. Geographic variation and hybridization

in populations of Bahama snails ( Cerion ) . Am. Museum Novit., 1806,

1-48.
Mayr, E., E. G. Linsley, and R. L. Usinger. 1953. Methods and Prin-
ciples of Systematic Zoology. McGraw-Hill Book Co., New York,

N.Y.
Meglitsch, P. A. 1954. On the nature of the species. Systematic Zool.,

3, 49-65.
Niggli, P. 1949. Probleme der Naturwissenschaften (Der Begriff der

Art in der Mineralogie). Basel.
Plate, L. 1914. Prinzipien der Systematik mit besonderer Berucksich-

tigung des Systems der Tiere. In Die Kultur der Gegenwart, III

(iv, 4), pp. 92-164.
Poulton, E. B. 1903. What is a species? Proc. Entomol. Soc. London

for 1903, pp. Ixxvii-cxvi.
Quatrefages, A. de. 1892. Darwin et les precurseurs frangais. Paris.
Ramsbottom, J. 1938. Linnaeus and the species concept. Proc. Lin-

nean Soc. London. (150 session), pp. 192-219.
Ray, J. 1686. Historia Plantarum, p. 40.
Rensch, B. 1929. Das Prinzip geographischer Rassenkreise und das

Problem der Artbildung. Borntrager Verl., Berlin.
Schopenhauer, A. 1851. Parcrga und Paralipomena: Heine philoso-

phische Schriften. Vol. 2, pp. 121-22. Berlin.
Simpson, G. G. 1943. Criteria for genera, species, and subspecies in

zoology and paleozoology. Ann. N.Y. Acad. Sci., 44, 145-78.
Simpson, G. G. 1951. The species concept. Evolution, 5, 285-98.
Sirks, M. J. 1952. Variability in the concept of species. Acta Biothe-

oretica, 10, 11-22.

22 CONCEPTS AND DEFINITIONS

Spring, A. F. 1838. Ueber die naturhistorischen Begriffe von Gattung,

Art und Abart und iiber die Ursochen der Abartungen in den or-

ganischen Reichen. Leipzig.
Spurway, H. 1955. The sub-human capacities for species recognition

and their correlation with reproductive isolation. Acta XI Congr.

Intern. Orn., Basel, 1954, pp. 340-49.
Stresemann, E. 1919. Uber die europaischen Baumlaufer. Verliandl.

Orn. Ges. Bay cm, 14, 39-74.
Sylvester-Bradley, P. C. 1956. The species concept in paleontology.

Introduction. Systematics Assoc. Publ. No. 2.
Thomas, G. 1956. The species concept in paleontology. Systematics

Assoc. Publ. No. 2, pp. 17-31.
Uhlmann, E. 1923. Entwicklungsgedanke und Artbegriff in ihrer

geschichtlichen Entstehung und sachlichen Beziehung. Jena. Z.

Naturw., 59, 1-114.
Valentine, D. H. 1949. The units of experimental taxonomy. Ada

Biotheoretica, 9, 75-88.
Voigt, F. S. 1817. Grundziige einer Naturgcschichte ah Geschichte der

Entstehung und weiteren Ausbildung der Natiirkbrper. Frankfurt

a.M.
Wallace, A. R. 1889. Darwinism: An Exposition of the Theory of Nat-
ural Selection, ivith Some of Its Applications. London.
Yapp, W. B. 1951. Definitions in biology. Nature, 167, 160.

THE SPECIES AS A FIELD FOR
GENE RECOMBINATION

HAMPTON L. CARSON: department of zoology,

WASHINGTON UNIVERSITY, ST. LOUIS, MISSOURI

The impact of genetics has reached into every corner of biol-
ogy. Nowhere has its influence been stronger than in the related
fields of taxonomy, speciation, and evolution. In fact, around the
fresh insight and new techniques provided by genetics has grown
a new science, which might be called experimental evolution.
Basically, its substance is derived from population genetics.

Huxley's The New Systematics (1940) chronicled the effects
of genetic discoveries on taxonomy, ecology, paleontology and
geographical distribution. At the same time the groundwork
was laid for the revitalization of work on speciation. The same
period, that is, just before and during the early years of World
War II, witnessed the appearance of an extraordinary series of
broad works on evolution. Thus the books of Dobzhansky ( 1937),
Mayr (1942), Huxley (1942), and Simpson (1944) came in
rapid succession. What is remarkable about these works is not
their differences of opinion but rather their essential agreement
on basic principles. A modern synthetic theory of evolutionary
cause had apparently been crystallized.

These key syntheses drew their freshness of approach largely
from the attention given to genetic phenomena; their function
was to integrate widely in evolutionary thought the major prin-
ciples established by the more strictly genetic and mathematical
fundamentals elaborated by the earlier works of Fisher ( 1930 ) ,
Haldane (1932), and Wright (1932). It takes time for a synthe-
sis of this enormity, even in its barest outlines, to be evaluated
and digested and thus serve as a guide to future biological work.
One could hardly expect immediate wide dissemination of such

23

24 GENE RECOMBINATION

a major generalization, let alone the years of contemplation and
testing which precede general acceptance of any such idea. In
the present instance, an initial delay was imposed by World War
II. Only now is it just beginning to be possible for the new gen-
eration of evolutionists to start critical evaluation of what appears
to have been the most spectacular era in basic biological thought
about evolution since the years following the publication of the
Origin of Species.

The most immediate effect of the syntheses mentioned above
was to stimulate a rash of investigations designed to answer
new questions about the species and especially about its popula-
tions. These new questions arise primarily from purely theo-
retical consideration of the consequences of Mendelian inheri-
tance in interbreeding communities of organisms. A geneticist,
in most instances, can work only with organisms that will inter-
breed freely so as to produce abundant offspring. His work is
thus of necessity usually at or below the species level. It will
therefore be a foregone conclusion that the evolutionarily in-
clined geneticist will have developed his own perspective — his
own particular way of looking at the species. The geneticist who
is oriented toward populations is really attempting the experi-
mental study of heredity on a level above that of the individual.
The present effort will be a somewhat personal picture — one
geneticist's view of the accomplishments of this approach in the
last fifteen years and the effects that it has had on the species
concept.

Genotype, Phenotype, and Wild Type

The term genotype is sometimes used in a strictly operational
way in experimental genetics to refer to the genie condition
under observation at a single locus or sometimes at a relatively
small number of loci (e.g., Aa or Aa Bb Cc Dd). In the sense
used in this paper, however, the word refers to the entire genetic
constitution or the sum total of the genes of an individual organ-
ism. This point requires emphasis because, as will be pointed out,
from the evolutionary point of view, it is unrealistic to select a

H. L. CARSON 25

few genes for special attention. We are being forced more and
more to emphasize the total genotype.

The geneticist is continually trying to see in behind the facade
of the phenotype to the genotypic basis that is the hereditary
endowment of an individual. The genotype is truly a property
of an individual organism, a kind of personal information which
crosses the narrow hereditary bridge from one generation to the
next. Sexual reproduction generates diverse genotypes, and each
genotype confers individuality. It is because of this individuality
of the genotype that the suffix "type" in the word conveys an
erroneous idea. The geneticist is wholly unable to set up types
with properties of the types of the taxonomist or, in fact, to work
with types of any sort. Indeed, the major emphasis of this ar-
ticle arises from the fact that one of the most important corol-
laries of our modern understanding of sexual reproduction is
that types are absolutely foreign, in fact diametrically opposed,
to the whole concept.

The term "wild type" has run into similar difficulties. It, again,
was originally used in an operational sense to refer to an allelic
condition at a single locus on a chromosome. An unjustified as-
sumption of uniformity from individual to individual is made
when the term wild type is extended to cover many loci not
under direct experimental observation. Thus, in some of the ear-
lier studies of genetics there was a tendency for natural popula-
tions of organisms to be interpreted as if the individuals com-
prising them were essentially homozygous for "wild-type" genes,
with only a few scattered recessive mutant alleles here and there
at various loci. This "classical" hypothesis is now giving way to
what Dobzhansky (1955) refers to as the "balance" hypothesis.
This latter designation refers to the fact that recent studies of
Mendelian populations under natural conditions have revealed
that a fantastic amount of genetic variability is concealed in
recessive form in the individuals of these populations.

The mechanisms which maintain a high level of genie variabil-
ity in natural populations are only just now becoming clear. It
is not within the scope of this short paper to review them, and

26 GENE RECOMBINATION

the reader is referred to the review of Dobzhansky (1955). The
very existence of this extensive variability, however, is the real
focal point in the unique view of the species that the geneticist
is beginning to take. The statement may safely be made that in
all probability, no two individuals in a sexual population are
genotypically identical. The geneticist thus sees, in every popula-
tion, arrays of individually different genotypes. He sees the
group as a whole, as a statistically varying array. The newer
methods of population genetics are attempting to deal directly
with the laws governing the behavior and fate of genotypes in
populations.

The Species as a Gene Pool

Those who have been bold enough to advance genetical defini-
tions of species are on safest ground when they deal with species
which can be shown to be sexually reproducing, cross-fertilizing
arrays of individuals. When a species has such a breeding struc-
ture, it is tempting to view the hereditary material possessed by
the species as a whole as a gigantic pool of genes. Reproducing
individuals may be viewed as throwing random assortments of
their genes, packaged in gametes, into the pool. The individuals
of the next generation are thus genetically chance recombinants
drawn from this gene pool. Dobzhansky (1950) has utilized
Wright's term "Mendelian population" for the reproductive com-
munity of individuals who share in a common gene pool.

Such a formulation as the above should serve only to direct
thinking along the general path followed by the population ge-
neticist. If it is taken more seriously than this, it is likely to be
misleading. Simple concepts of gene frequency and the conse-
quent related predictions of the Hardy- Weinberg law may be
demonstrated in a very simple model of a gene pool. A hundred
poker chips, 30 red and 70 white can be pooled in a fishbowl. The
two colors represent the two alternative alleles at one locus.
Drawing out chips in pairs thus represents zygote formation for
a single locus; the homozygotes will be red-red and white-white
and the heterozygotes red-white. The relationship of gametic

H. L. CARSON 27

frequency to zygotic frequency, predicted by the Hardy-Wein-
berg law, may thus be directly observed.

This model may be elaborated somewhat; for instance, one
may intuitively grasp some of the enormous statistical conse-
quences of Mendelian heredity in sexual reproduction by drawing
zygotes from two or more fishbowls simultaneously. One may
attempt to diagram by this method the long-range effects of
selection on certain zygotic types. The elimination of a deleterious
gene, for instance, may be observed by extending selection over
a number of generations. In short, it is possible through such a
model to gain a crude measure of realization of how evolutionary
changes in gene frequency can occur through the media of mu-
tation, recombination and differential reproduction.

Such models, however, have serious drawbacks. In many
ways, for instance, it is unsatisfactory to verbalize a mathemat-
ical concept like the above. The naturalist will also object to the
pictorial analogy, even for emphasis of a point of view, of the
species population to chips in a fishbowl. It is so crude as to make
it necessary to point out immediately the important ways in
which species populations do not behave like poker chips in a
fishbowl. The employment of the word "pool" in the expression
above, is nevertheless tremendously important to stress the
point that groups of individual organisms really possess a cor-
porate genotype, transcending that of the individuals. No other
word in common usage appears to come closer than "pool" in
expressing the nature of the phenomena involved. At this junc-
ture it seems unwise to invent a new one; rather it is better to
tell what a species pool is really like.

The Local Population

The term "gene pool" has also been used in a number of
senses. Thus, as is done very widely, one may refer to the total
gene resources, both past and present, of a species as its gene
pool. Whenever the term "gene flow" is employed, as is often
glibly done when neither actual genes nor actual flow are under
experimental observation, this broad view of the gene pool in

28 GENE RECOMBINATION

both space and time is implied. This concept is most useful for
the student of the species who has an essentially historical ap-
proach. He attempts to interpret the mosaic of present-day popu-
lations of a single species in terms of the action of evolutionary
directive factors on the totality of gene resources which have
been available in its ancestral populations.

In yet another sense, the parameter of time may be omitted
and the contemporary genetic material of a species may be
viewed as constituting a pool of genes. Now in this case, the
emphasis is on the total genie material currently found through-
out the geographical range of the species. This gene pool is thus
the actual legacy willed by selection to the present-day series of
generations which a single human investigator may observe. This
type of pool is the province of the animal and plant breeder, the
eugenicist, and the evolutionist who may be tempted to predict
what may happen next.

Neither of the above "pools," however, is the really crucial one
from the point of view of the population geneticist. His attention
is focused on a gene pool which can be defined much more rig-
orously and which may be studied with precise observational
and experimental methods. This is the gene pool found in a
single contemporary local breeding population, made up of ran-
domly, or almost randomly, interbreeding individuals. Wright has
encouraged the use of Gilmour and Gregor's ( 1939 ) term deme
to refer to this microgeographical unit, momentarily suspended in
time, which lies at the base of the hierarchal array of species
populations.

Because breeding populations of species are entities distributed
in space, the individual genotypes formed in local populations
in any one generation are never drawn at random from the
entire gene pool of the species. Despite' dispersal and outcrossing
mechanisms, organisms tend to breed with those that are close
to them, in a physical sense, in their environment. This gives
reality to the deme and makes it the unit which most closely re-
sembles the fishbowl model.

Of central importance is the fact that the local population is
the only part of the larger gene pool which is active from the

H. L. CARSON 29

evolutionary point of view in any one generation, that is, at any
one time. Although transitory, the local population is always the
unit in which any new character first makes its appearance. It is
the only place where gene recombination can be effective as a
directive factor in evolution. Indeed, the local population, at any
one time level, always represents the point of contact between
the hereditary material of a species and its environment. It is at
this point that new genotypes are generated and this is where
they meet the immediate acceptance or rejection of the omnipo-
tent environment.

The local population is not equivalent to any of the intraspe-
cific designations of the taxonomist, such a subspecies, local race,
or variety. The local population is not only suspended in time,
but it is usually much smaller than any of these. Furthermore,
even a local race may have a population structure which divides
it genetically into a series of partially isolated interbreeding
populations. A local population can, for example, be a hybrid
swarm; in such a case we are observing the release of an enor-
mous amount of variability through the recombination of genomes
which have been separated historically. The local population may
indeed have any one of quite a variety of past histories, which
will affect the amount of genetic recombination which occurs.
The fate of the genetic variability and of the continuing com-
position of the gene pool nonetheless is determined by the
inexorable laws of population genetics and the directive factors
of evolution working at a local level.

Recombination in Populations

Genetics is really the study of sexual reproduction. An enor-
mous amount of research on meiosis and chromosome cycles,
perhaps the bulk of it, was engaged in when only the vague
outlines of the evolutionary and genetic meanings of these chro-
mosome behaviors were clear. The details of sexual phenomena
in protozoa, for example, have been filled in with great elaborate-
ness as an adjunct to the ancient problems of taxonomic rela-
tionships and phylogenetic interpretations. The chromosome cy-
cles that one observes in the life history of an organism, how-

30 GENE RECOMBINATION

ever, are not merely historical relics to be studied with the
techniques and attitudes of comparative anatomy and phylogeny
— they are rather dynamic genetic systems, concerned in most
cases with the recombination of genes. They should be studied
for what they are — basic mechanisms which determine the ge-
netic makeup of the individuals which comprise species popula-
tions.

One of the major consequences of the particulate nature of the
hereditary material is the potentiality which automatically exists
for recombination. The sexual mode of reproduction, as had been
brought out so brilliantly in the work of Darlington ( 1939 ) and
White (1954), is biologically meaningful only when viewed
basically as a process of recombination of hereditary particles.
The analytical methodology of the whole science of genetics is
rooted in recombination; in fact, perhaps the most satisfactory
definition of a gene is a hereditary entity which acts as a unit
in recombination.

Given a certain amount of hereditary diversity provided by
gene mutation, the sexual process, through its essential feature
of recombination, can generate an enormous number of diverse
genotypes. These genotypes, or recombinants, are exposed to
the action of natural selection in local populations. In the above
context, recombination occupies a focal position in modern
theory of evolutionary cause. For this reason, the devices whereby
sexual reproduction accomplishes recombination deserve close
scrutiny.

Recombination in sexual reproduction is a function of three
processes. These are: crossing over between homologous chro-
mosomes, random segregation of chromatids from bivalent chro-
mosomes, and chance recombination of gametes. Thus the great-
est amount of recombination will result in an organism which
has intensive and extensive crossing over, high chromosome num-
ber and a high degree of interbreeding between relatively
unrelated individuals (outcrossing). Conversely, if exchanges
between homologous chromosomes are few, if the chromosome
number is small and if a considerable degree of inbreeding oc-

H. L. CARSON 31

curs, the amount of recombination that can be produced may be
relatively much smaller.

As natural selection operates with gene combinations, it fol-
lows that where the hereditary material of the species is faced
with a new environmental challenge, the success with which the
species meets this challenge depends to a very great extent on
its ability to form new gene combinations. In other words, the
possession of a high potential for recombination carries with it
a correspondingly high potential for rapid evolution, whereas
general inability to recombine the hereditary material implies
evolutionary stagnation and greatly restricts ability to alter the
composition of the gene pool.

If one wants to interpret the recent past history of a species or
to predict how rapid and far-reaching an evolutionary change
might occur in it, it is of the greatest importance to know the
specific system by which it reproduces. If the system is a sexual
one, then some knowledge of the potential recombination which
might be accomplished in a local population of this species is
imperative. Thus, the geneticist views the species population in
part with a sort of x-ray vision — looking within to such matters
as details of meiosis — as well as assessing the importance of the
more extrinsic factors which affect recombination, such as the
mating system.

An important point of emphasis is that estimates of the ability
to undergo recombination express only a potential ability — a
potential with which the species might respond should the con-
ditions arise which demand evolutionary aggression. If the con-
ditions do not arise, the status quo is likely to be maintained.
The potential may be there but never actually realized. The
possession of a high recombination potential by a species is no
more an automatic guarantee of further progress than is a high
mutation rate.

To return to the point of view of the local population. As a
single generation is bridged by a population, the increment of
change possibly depends entirely on the wealth of gene com-
binations which are formed in the local pools. The widespread

32 GENE RECOMBINATION

geographical nature of most species means that only a small por-
tion of the total pool of genes in the species is drawn into com-
bination at any one time. Thus it is that the geneticist finds
himself forced, like the members of all other disciplines which
deal with species, to consider both geographical and historical
matters.

The Species in Space and Time

As we have seen previously, the local population occupies a
really pivotal position for the genetic nature of the species. Its
composition depends not only on the past history of the local
population itself but also on the history of the populations near
it, and from which it may have inherited genes. Thus the pa-
rameters of time and space in the study of local populations are
so interrelated that it is often difficult to separate their effects.
The words "gene flow" are sometimes used to describe the process
whereby the local population obtains influx of hereditary ma-
terial from the outside. This is, in most cases, an extremely slow,
gradual process, usually extending over many generations.

In a larger sense, then, the gene pool of the species is dis-
tributed in time as well as in space. The local population where
the actual recombination trials are taking place may owe its
contemporary composition to very complex interactions with
the populations in its immediate ancestry. In certain cases, cyto-
genetic studies enable us to prove that this past history involved
polyploidy or hybridization or both. In other cases, it is possible
to interpret the present genetic peculiarities of a species popula-
tion as being traceable to recent strong isolation of a local popu-
lation. In any event, the net result is an alteration of the content
of the gene pool and thus both qualitatively and quantitatively
of the raw materials with which recombination operates. Thus,
for instance, a newly formed hybrid swarm, such as one that
has resulted from the breakdown of ecological barriers by hu-
man disturbance of the habitat, may be looked upon primarily as
a process resulting in the sudden release of recombinations con-
sequent upon the original wide outcross. Hybridization between
species with high chromosome numbers which maintain high

H. L. CARSON 33

frequencies of distributed crossovers is, in fact, the condition
which would appear to be able to release the greatest amount
of hereditary variability in the shortest possible time.

How Much Recombination?

The amount of recombination permitted in a given species by
the genetic system of that species is best looked upon as an adap-
tive property which is under the control of natural selection.
Too much recombination is inadaptive because it tends to break
up adaptive complexes of genes which have been welded to-
gether by natural selection. This reduces the immediate fitness of
the organism. In order to survive, a species must maintain high
immediate fitness and it is not surprising that devices limiting or
preventing recombination are common in species populations.
Not infrequently, as will be brought out below, the process ap-
pears to have been carried to its logical conclusion and recom-
bination is inhibited completely. At this point, sexual reproduc-
tion ceases to have significance as a biologically important proc-
ess.

Organisms which have completely discarded sexual reproduc-
tion, and reproduce exclusively by asexual means are definitely
in the minority in nature. Thus it appears that too little recom-
bination jeopardizes the ability of the species to meet drastic
changes in environmental conditions over very long periods of
time (see Thoday, 1953). Most organisms which have survived
to the present day display a balance between these two forces.
There is neither too much recombination on the one hand, which
tends to break up adaptive complexes, or too little, which leads
to evolutionary rigidity, specialization, inability to change, and
eventually extinction in the face of an environmental challenge.

Open, Restricted, and Closed Recombination Systems

If we refer to the sum total of all the devices leading toward
recombination as the "recombination system," we may attempt to
assess the capacity of a given species to generate recombinations
in a typical contemporary local population. The effectiveness of
the recombination system obviously varies enormously from one

34 GENE RECOMBINATION

species to the next. Such variation is almost certain to be con-
tinuous in such a way that hard and fast categories could not con-
ceivably be applied. Thus the attempt which follows is designed
merely to give a very rough outline for this formulation of the
problem. The point should again be stressed that it is in almost
every case necessary to speak in terms of potential rather than
actual recombination. Obviously, too, recombination can be effec-
tive only where there are differences which can be recombined.
Precise methods do not as yet exist for estimating the total
amount of recombinable heterozygosity in a species or a popula-
tion. Heterozygosity has its ultimate origin in the mutation proc-
ess, but its existence in any given population will depend on
many different factors such as possible recent wide outcrossing,
including species hybridization, a recent or past high rate of
mutation, or recent drastic fluctuations in population size.

A recombination system wherein a high degree of recombina-
tion is permitted may be referred to as open. The human species,
for example, displays all the ingredients which may be associated
with such a system. The chromosome number (2n = 48) is high,
so that a substantial amount of recombination occurs by the seg-
regation of whole chromosomes alone, quite apart from crossing
over. The number of chromosomally different gametes, which can
potentially be formed by an individual which carries one pair of
allelic differences per pair of homologous chromosomes is given
by the simple formula 2", where n is the number of chromosome
pairs concerned. In the present example, the number is 224, more
than 16 million, produced by a single individual. Perhaps even
more important, however, there is evidence that crossing over
between homologues is extensive and unhampered throughout
the genome. This system is superimposed on the foregoing one.
If the assumption is made that a moderate number of allelic
differences exist per chromosome pair, a fantastically high re-
combination potential is provided on an individual basis.

Even under rather close local inbreeding systems, the purely
chromosomal devices mentioned above have high potential
effectiveness. But superimposed on these internal mechanics are
the intricate and far-reaching influences of the choice of mates,

H. L. CARSON 35

the mating system, which is the external and ecologically affected
phase of genetic recombination. Through it, chance recombina-
tion of gametes is effected. In human populations, the mating
systems are intricate in the extreme. Suffice it only to say here
that we are witnessing in a few generations the rapid recombina-
tion of large segments of the human gene pool which were once
separate.

Flies of the genus Drosophila may serve as an example of a
recombination system which is much more restricted than the
above example ( Patterson and Stone, 1952 ) . The highest chromo-
some number in the group is 2n = 12, and most species have suf-
fered a diminution in whole chromosome recombination potential
by further reduction of the effective chromosome number by
centromere fusions. Secondly, this genus of flies, together with a
large segment of the higher Diptera have no crossing over in the
male sex. This condition, despite the distributed crossing over
in females, reduces the effective recombination due to crossing
over by one-half compared to organisms which have dis-
tributed crossovers in both sexes. In certain species, the accu-
mulation of inverted sections in the chromosomes in natural
populations results essentially in the isolation of blocks of the
germ plasm from effective genetic recombination. This system is
particularly evident and important in many endemic species of
Drosophila. The outcrossing potential of Drosophila is also rela-
tively low. Many species crosses can be made in the laboratory,
but, with rather insignificant exceptions, do not occur under
natural conditions.

Drosophila as a genus shows an interesting variety of recombi-
nation systems. In some species, inversions are absent, or nearly
so, from natural populations, whereas in others inversions are
abundant. Even within an individual species, there are some
populations which are highly heterozygous for inversions and
others which are essentially homozygous. For instance, Carson
(1954, 1955) has devised a means of comparing quantitatively
the amount of genetic recombination which occurs in different
populations of Drosophila rohusta of the eastern United States.
Marginal populations of this species display a much greater po-

36 GENE RECOMBINATION

tential for genetic recombination than do populations from the
central part of the range. Marginal populations thus appear to
have a recombination system that is especially well suited to
evolutionary aggression and pioneering.

In any organism in which reproduction is solely by asexual
means, the recombination system may be considered closed. The
cases most easily interpreted in this regard are those in which the
asexual method of reproduction has clearly come to be irrevo-
cably and completely substituted for the sexual method. Such a
situation exists, for example, in certain of the obligate apomicts
of the genus Crepis (Stebbins, 1950). Adaptation, as it becomes
increasingly perfect, may be increasingly interfered with by
recombination in such a way that selection will tend to bypass
and then gradually eliminate entirely the very process that makes
adaptation possible. Organisms which have eliminated sexual
reproduction may be looked upon as representing a closed sys-
tem, a dead end in evolution.

To the student of genetic recombination and its importance
in evolution it would appear likely that all organisms which repro-
duce exclusively by asexual means have acquired this property
secondarily and originally reproduced by sexual reproduction. It
is perhaps not generally realized in biology that recent findings
in population genetics have thrown new light on the basic
biological meaning of sex. It may well be that many of the older
concepts on the origin of sexual reproduction will have to be
reexamined in detail.

Conclusion and Summary

The thesis is advanced that the most meaningful approach to
the sexually reproducing species that the geneticist may make
is by studying and describing it as a system of recombining genes.
Such a system is often spoken of as a "gene pool. The breeding
structure ol the species is such that there is an almost continuous
compartmentation ol tin's larger gene pool into smaller and
smaller units ol recombination, with varying degrees ol isolation
from one another. At the bottom of the ladder and in a focal
position is the local population, which at any one lime level is

H. L. CARSON 37

the functional gene pool from which recombinations are being
drawn. These combinations, encapsulated as individual organ-
isms, are confronted by natural selection at the local level. In a
broader sense, however, the gene pool includes all pools of the
species that are separated geographically and temporally from
one another.

The pattern of sexual reproduction existing in a particular
local population constitutes its recombination system at that
point in time and space. This may be a relatively open or closed
system, depending on the chromosomal mechanisms and mating
system that prevail. Thus a species may have extensive genetic
recombinations available for selection in any one generation or
series of successive generations, or the potentiality for recombina-
tion may be partially closed or even locked up and unavailable.
These characteristics will be of very great importance in under-
standing ( 1 ) the past history of a species, ( 2 ) its present variabil-
ity, and (3) its potentiality for future evolutionary change.

REFERENCES

Carson, H. L. 1954. Variation in genetic recombination in natural
populations. /. Cellular Comp. Physiol, 45 (suppl. 2), 221-36.

Carson, H. L. 1955. The genetic characteristics of marginal popula-
tions of Drosophila. Cold Spring Harbor Symposia Quant. Biol, 20,
276-87.

Darlington, C. D. 1939. The Evolution of Genetic Systems. The Uni-
versity Press, Cambridge, England.

Dobzhansky, T. 1937. Genetics and the Origin of Species. Columbia
University Press, New York, N.Y.

Dobzhansky, T. 1950. Mendelian populations and their evolution.
Amer. Naturalist, 84, 401-18.

Dobzhansky, T. 1955. A review of some fundamental concepts and
problems of population genetics. Cold Spring Harbor Symposia
Quant. Biol, 20, 1-15.

Fisher, R. A. 1930. The Genetical Theory of Natural Selection. The
Clarendon Press, Oxford, England.

Gilmour, J. S. L., and J. W. Gregor. 1939. Demes: A suggested new
terminology. Nature, 144, 333.

38 GENE RECOMBINATION

Haldane, J. B. S. 1932. The Causes of Evolution. Harper & Bros., New
York, N.Y.

Huxley, J. 1940. The New Systematics. The Clarendon Press, Oxford,
England.

Huxley, J. 1942. Evolution, the Modern Synthesis. Allen and Unwin,
London.

Mayr, E. 1942. Systematics and the Origin of Species. Columbia Uni-
versity Press, New York, N.Y.

Patterson, J. T., and W. S. Stone. 1952. Evolution in the Genus Dro-
sophila. The Macmillan Company, New York, N.Y.

Simpson, G. G. 1944. Tempo and Mode in Evolution, Columbia Uni-
versity Press, New York, N.Y.

Stebbins, G. L., Jr. 1950. Variation and Evolution in Plants. Columbia
University Press, New York, N.Y.

Thoday, J. M., 1953. The components of fitness. Syni])osia Soc. Exptl.
Biol, 7, 96-113.

White, M. J. D. 1954. Animal Cytology and Evolution. The University
Press, Cambridge, England.

Wright, S. 1932. The roles of mutation, inbreeding, crossbreeding and
selection in evolution. Proc. 6th Intern. Congr. Genetics, 1, 356-66.

THE PLANT SPECIES

IN THEORY AND PRACTICE

VERNE GRANT: rancho santa ana botanic garden and

CLAREMONT GRADUATE SCHOOL, CLAREMONT, CALIFORNIA

Biological Units

A philosophical tendency in modern biology, which has been
in process of formulation for a century or more, is the view that
the world of life is composed of a series of units organized into
successively higher levels of complexity. The elementary bio-
logical units, gene, chromosome, cell, individual, deme (or local
population), society, species, biotic community, represent differ-
ent levels of organization of biological materials. All these units,
in spite of numerous differences, share three common properties:
an internal integration, the power of reproduction, and the ability
to become modified in time. It would be correct to describe them
all as integrated wholes, as reproductive units, and as entities
capable of development.

The process of intellectual criticism has sharpened the concept
of each one of the biological units. The progress of thinking fol-
lows a pattern. The unit in question is discovered and its auton-
omy and unity are frequently overemphasized by the earlier
investigators in the field. With the growth of knowledge a reac-
tion sets in. Some biologist in a later period is certain to maintain
that the entity has no well-marked boundaries and hence has no
definable unity, or even that it has no objective existence at all.
This heresy forces a reexamination of the concept hitherto held
and of the evidence upon which it was based. A controversy
over "the problem of the gene" or "the concept of the association"
ensues, which may eventually culminate in a deeper understand-
ing of the nature of the particular unit.

One of the most common adjustments which has to be made

39

40 THE PLANT SPECIES

in the development of a unit concept in biology is the recogni-
tion that the unit does not enjoy an independent and uncondi-
tioned existence but that it forms a constituent part of a higher
organization. Neither the extreme position that the unit is auton-
omous (and therefore objectively defined), nor the opposite
position that it is merely an integral part of a more complex unit
(which alone is objectively demarcated), is likely to prove cor-
rect in the final synthesis. The truth is more likely to lie in the
middle ground that the elementary unit possesses an objective
existence which, however, is subject to certain limitations due to
its integration into a higher system.

Thus in nineteenth century biology the adherents of the cell
theory were forced to retreat from their original position under
the pressure of the knowledge that the organism as a whole con-
stitutes a physiological unit and not a mere aggregation of cells.
The compromise solution embodied in the now generally ac-
cepted organismal theory did not, however, negate, but merely
qualified, the unitary nature of the cell. The individual in turn
is now recognized as an elementary component of a more com-
plex unit, the breeding population, in cross-fertilizing organisms.
The facts of modern population genetics which make this con-
clusion inescapable can be accepted without denying the exist-
ence of the individual as a subordinate but real unit in its own
right. The acceptance of these facts, moreover, is fatal to the
opposite extreme view, still held in some quarters, that the in-
dividual is the ultimate biological unit.

A second adjustment which is frequently necessary in the
formulation of a satisfactory concept of a biological unit is the
redrawing of the boundaries. A unit as originally defined may
require subdivision in the light of subsequent evidence. The
concept of the gene seems to be going through this process at
the present time. There is accumulating evidence that the
Mendelian gene, which is the unit of crossing over, is not equiva-
lent to the unit of biochemical activity in the chromosomes. The
theoretical difficulty of delimiting individuals in clonally repro-
ducing organisms has led to the recognition of two units in
place of one: the individual, as defined by physiological auton-

V. GRANT 41

omy, and the clone, defined by descent from a single zygote.
Similarly, various authors have been forced to recognize different
kinds of species.

The task of finding objective criteria for the definition of the
biological units, that is, criteria for marking off their boundaries
which can be applied by different observers with equivalent
results, has in general proved easier in the case of the microscopic
units than with the macroscopic ones. Thus the chromosome and
the cell are clearly definable units; there is every reason to expect
that the recombination gene and the functional gene can be
satisfactorily defined; but an inordinate amount of controversy
has been devoted to the species and the biotic community. As
between the latter two entities, moreover, it is apparent that
biologists are much closer at the present time to an understand-
ing of the species question than they are to finding a basis for
the definition of the community. This largest and most complex
of the biological units presents the greatest difficulties of defini-
tion and delimitation of them all.

Evolution of the Species Concept

The original species concept was that of primitive man. His
idea of species was based on the observation of discrete units in
the local fauna and flora. He had a specific name for each kind
of animal and plant to which he paid any attention; these names
corresponded rather well to objectively existing populations in
nature. Essentially the same species concept was adopted by the
local naturalist. Thus John Ray in 1686 defined the species as the
unit which breeds true within its own limits (cf. Darlington,
1940).

The primitive species concept was not equipped for coping
with the phenomenon of geographical variation, which only be-
came apparent when the stage of man's activities expanded from
the restricted home territory to an area equal in extent to the
distribution range of many species. A species which was clear-cut
in the local community might become the center of a "species
problem" when a series of collections from different localities
was assembled and compared. To accommodate the expanded

42 THE PLANT SPECIES

knowledge of individual species gained since the nineteenth
century an expanded concept was required. Criteria had to be
sought that would lead to the demarcation of species not only
within a local community but also on the wider stage of the
world fauna and flora. The result of this search, conducted by
numerous students during the past seventy years, is the bio-
logical species concept, which defines species on the basis of
reproductive isolation or the presence of barriers to gene ex-
change.

The period between primitive man and the modern evolution-
ists, that is, the period during which geographical variation and
the intergradation of morphological characters were known, but
the idea of reproductive isolating mechanisms had not yet been
formulated to explain the limits of this variation, was marked by
an interregnum in the species concept. The Linnaean or typologi-
cal species concept, which emphasized morphological difference
rather than discontinuity as the criterion of species and recog-
nized species universally throughout the plant and animal king-
doms, prevailed for two centuries during this interregnum, and
it has inevitably left its stamp on modern systematics.

In its original conception the Linnaean species was a fixed and
immutable entity. In the nineteenth century, with the increase
in knowledge concerning natural variation and the rise of the
theory of evolution, the fixed category became regarded more
often as an arbitrarily determined stage above the level of the
variety. The morphologically defined species, which seemed to
be an objective reality in the period of Linnaeus and Cuvier,
became frankly recognized as a subjective concept in the nine-
teenth and early twentieth centuries.

Critique of the Species Concept

The problem of defining the species in plants has been com-
plicated by the fact that different botanists have applied the
name of species to entities ranging in magnitude from homo-
zygous biotypes (Lotsy) to species groups (Engler). A common
practice among botanists is to recognize as a species any popu-
lation exhibiting distinctive morphological characters combined

V. GRANT 43

with a definite geographical range. The result, of course, is that
many named species are equivalent to the geographical races of
a polytypic species. There are also a few botanists who adopt the
so-called biological species concept.

The opinion held by many botanists today that the species is
undefinable is in part a heritage from the interregnum of the
typological species concept. This agnosticism is written into the
International Code of Botanical Nomenclature which through
four editions has studiously refrained from defining the species.
If one is operating within the framework of a typological species
concept, this attitude is certainly justified.

The species which is defined on the basis of a certain number
and degree of morphological differences is indeed a subjectively
defined entity. No criterion of the amount of morphological dif-
ference between two forms that marks them as species, rather
than as subspecies, sections, or genera, etc., and that is capable
of application generally has ever been proposed. It follows that
different observers confronted with the same range of variation
in a series of specimens will frequently be unable to agree on
the taxonomic disposition of the case, in so far as their decisions
are predicated on a typological species concept. That such differ-
ences of opinion are commonplace is evident to anyone who has
compared different taxonomic treatments of the same group pre-
pared by different but equally well qualified authors.

It is generally agreed by the adherents of both the typological
and the biological species concepts that the species of the typo-
logical concept can be defined only as that entity which a com-
petent systematist regards as a species. This conclusion consti-
tutes the strongest argument against retaining the typological
species concept in taxonomy, now that an objectively defined
concept is available to take its place. Unfortunately, the advo-
cates of the subjectivity of the species as a unit usually ignore
the existence of the biological definition. Thus Mason writes that,
"I have seen no putative definition of a taxonomic category so
worded as to be incapable of application either to the next higher
or the next lower category of the taxonomic structure. That
which is a species to one taxonomist may be a subspecies to

44 THE PLANT SPECIES

another . . ." (Mason, 1950). The presence of reproductive iso-
lating mechanisms is a feature of species, but not of subspecies;
the smallest population possessing such reproductive barriers is
the species, and not the polytypic section or genus.

The difficulties of the typological species concept have led
some students to adopt, not the biological, but still another, nomi-
nalistic concept of the species. The proponents of this concept
stipulate that the species is (are) an "empty category," "mental
units rather than biological units," "highly abstract fictions," an
"abstract category in the taxonomic structure" which "has no
foundation in reality and obviously cannot be objectively de-
fined" (Mason, 1950; Davidson, 1954; Burma, 1954).

The nominalistic concept is logically sound although scientifi-
cally barren. If all that is implied by the word "species" is the
category in the taxonomic hierarchy corresponding to the specific
name in the binomial system of nomenclature, there can be no
disagreement as to its artificiality. In this sense of the word, the
unicorn, if properly described according to the Rules of Nomen-
clature, would be as real a species as the horse.

According to Mason again (1950), "The wisdom of past expe-
rience has dictated that the taxonomist purposely refrain from
defining these categories [species, genus, family, order, etc.] in
any way that will impose restrictions on the freedom with which
he may express the interrelationships that he construes to exist."
With regard to the genus, family, and order the absence of an
objective definition is not so much a matter of wisdom as of
necessity. With regard to the species the attempt to preserve a
wide elasticity of usage, besides being unnecessary, is also un-
wise, since it leads to a chaotic condition in taxonomic work
which could be avoided. Every branch of natural science is
sooner or later compelled to define its basic units in objective
terms. It would be remarkable if plant taxonomy were to prove
unique in this respect. In the meantime, as so aptly pointed out
by Camp (1951), the adherence to a subjective criterion of its
basic unit is responsible for the fact that plant taxonomy, in many
of its branches, is not yet a science but an art.

The opinion that the species is not a definite entity has been

V. GRANT 45

arrived at in three different ways. We have seen that the typo-
logical species definition leads to the conclusion that the species
is a subjective entity. The nominalistic species concept frankly
eschews the search for an objective criterion of species. It remains
to note that the biological species concept itself denies the exist-
ence of definite species units in certain plant groups. In fact, a
large part of the species problem, as it currently exists in botany,
arises from the attempt to apply the species concept to groups
in which biological species do not exist. The view of many bota-
nists that the species is undefinable is, in such groups, justified by
the facts.

If one wishes to use the term species everywhere throughout
the plant and animal kingdoms, one must adopt the typological
species concept, since morphological similarity and convenience
in classification are the only criteria that can be applied univer-
sally. The general usage of the species category is a traditional
practice in taxonomy. It has even become a dogma, sanctioned
by a pronouncement in the Code of Nomenclature that all plants
belong to species. Neither the species concept of primitive man,
nor that of the naturalist, nor that of Linnaeus, however, was
based on a consideration of organisms representing all the types
of life cycle now known to modern biology. The historical grounds
for applying the species concept to all forms of life are conse-
quently not compelling.

The issue between alternative concepts of the species can and
should be decided on the basis of the respective merits of the
contending viewpoints rather than on purely historical considera-
tions. The chief advantage of the typological species concept is
its universality; this meritorious feature is offset by the corre-
sponding disadvantage of subjectivity. The biological concept,
on the other hand, enables us to define the species objectively in
the numerous and dominant class of sexual organisms, but it is
not applicable outside such organisms.

To the present author, the privilege of applying a subjective
species definition to all organisms does not represent a very im-
portant gain; whereas the price that must be paid for this priv-
ilege, namely the giving up of an objectively defined species con-

46 THE PLANT SPECIES

cept applicable to the majority of higher organisms, does repre-
sent a very important loss. If in any given group of organisms
we do not find an evolutionary development into reproductively
isolated populations, we can, after all, fall back upon a morpho-
logical criterion and call the various forms by some other name
than species, such as taxon or binom. Biologists who reserve the
category of species for morphologically similar groupings of in-
dividuals, on the other hand, have not left any alternative desig-
nation for the reproductively isolated system of breeding pop-
ulations.

The reproductively isolated population, whether it is called a
species or some other name, undoubtedly represents a major
biological unit, which must be investigated and discussed. This
task will be facilitated if the unit in question can simply be re-
ferred to as "the species" or, if any qualification is needed, as
"the biological species." For purely operational purposes, if for
no other reason, the adoption of the biological definition of the
species is justified in discussions of the origin, nature, and be-
havior of reproductively isolated population systems.

The substitution of a biological for a morphological definition
of the species does not provide a panacea for the solution of every
complex situation in nature. The naive assumption that it should
do so, more often tacitly implied than explicitly stated, is respon-
sible for the disillusionment in the biological species concept felt
by some biologists whenever they run into difficulties of inter-
pretation. Such difficulties should be expected, considering the
wide diversity of life cycles, breeding systems, and population
structures existing among sexual organisms. All that can be
claimed for the biological species concept is that it provides a
framework that works better than any other yet proposed for the
analysis of species problems.

The remainder of this paper will be devoted to a consideration
of certain problems involved in the recognition and demarcation
of biological species in the higher plants. Tin's inquiry will give
attention to the difficulties encountered in setting the limits of
species, the causes of those difficulties, and the biological signifi-
cance of the causes.

V. GRANT 47

Criteria of Reproductive Isolation

Relatively few groups of organisms have been studied from
enough points of view with sufficient thoroughness to reveal their
whole pattern of species-separating mechanisms. In most groups
the presence and distribution of isolating mechanisms will have
to be inferred from indirect evidence of various sorts. Among the
most important of these are the finding of discontinuities in the
variation pattern, marginal overlapping of distribution areas, com-
plete sympatry, or hybrid sterility in the laboratory or experi-
mental garden.

If these criteria of reproductive isolation are consistent among
themselves in any given group, there will be no species problem.
It may happen, however, that one or more criteria will fail to
coincide with the rest. Under these circumstances an automatic
application of any single criterion in drawing species boundaries
might easily lead to a false picture of the natural species. The in-
consistency indicates that a critical appraisal of the criteria them-
selves is always needed.

A reproductively isolated cross-fertilizing population should
reveal its genetic separation from other such populations by a
gap in the pattern of variation. Some plant species in every genus
do show the expected morphological discontinuity, but others for
various reasons do not. If applied uncritically, the criterion of
discontinuity would often result in treating as species populations
belonging to other categories. In a polytypic species with a dis-
junct distribution, the subspecies may be separated by morpho-
logical gaps, i.e., Gilia leptantha. The morphologically discrete
units in Agoseris seem to be the species groups; in Ceanothus
they frequently correspond to the sections; in the California oaks
one of the discrete entities is a subgenus; and in Allophyllum it
is the entire genus, which has, in fact, until recently been treated
as a single species.

The overlapping range of variation of three biologically well
isolated species of Gilia is shown in Fig. 1. Each symbol repre-
sents a population. The coordinates give the measurements of the
two most important diagnostic characters for these three species,

50

45

40

X

X

35

30

o

o

25

o

X

20

o

X

X

X
X

15

E

3

10

A

AX
A
A
X
A

Ax

AX

X
X

A

A

A
X

X

AX

A
X

X
XX
X
X
X

*
X

Length

13
of

16
corolla, mm.

21

Fig. 1. Overlapping range of variation of three intersterile species of
Leafj stemmed Gilia with respect to two important diagnostic characters.
Each dot represents the average characteristics of a single local population.
• (.'. capitata abrotanifolia, X G. achiUcacfoJia, A Q. angelensis. From

Evolution, 7 (1953)

48

V. GRANT 49

the ordinate the number of flowers in the inflorescence and the ab-
scissa the length of the corolla. The absence of a discontinuity
in the variation patterns of the different species is apparent.
Similar results are obtained when other characters are used.

Relatively few biologists subscribe any longer to the old idea
that species can be defined by the sterility of hybrids. There are
too many cases of good species which can produce fertile hybrids
under laboratory or garden conditions although they remain
amply distinct in nature. Furthermore, all degrees of sterility
exist so that the distinction between "intersterile" and "inter-
fertile" is frequently an arbitrary one. The existence of intra-
specific sterility phenomena also makes any general application
of the sterility criterion of species hazardous. Nevertheless, there
can be no doubt that sterility barriers comprise some of the most
effective mechanisms which separate species, and their presence
must be weighed carefully in any attempt to devise a classifica-
tion of species which accords with the realities of nature.

The most essential property of a species is its ability to main-
tain its distinctive features, or rather the gene patterns which
determine those features, from generation to generation in a
community of organisms. Sympatry is the final test of species
status. Two sympatric populations will normally be regarded as
belonging to separate species. It by no means follows as an auto-
matic rule, however, that related allopatric populations are neces-
sarily conspecific.

We have several pairs of related allopatric species in Gilia.
Gilia tricolor and G. angelensis, for example, belong to the same
species group and occupy similar ecological habitats in contigu-
ous geographical areas ( Grant, 1952 ) . It was at first assumed as a
working hypothesis that they were members of the same poly-
typic species. This hypothesis had to be abandoned when it was
learned that the two forms are completely intersterile in the
experimental garden and probably possess different genomes as
well. The intersterility of G. tricolor and G. angelensis stands in
marked contrast to the absence of any known intraspecific steril-
ity within either entity. The facts are very similar in the case of
G. millefoliata and G. laciniata; G. splendens and G. leptalea; and

50 THE PLANT SPECIES

G. sinuata and G. crassifolia. Examples such as these should warn
us against automatically combining related allopatric forms in
the same species. The category of superspecies, proposed by Mayr
( 1942 ) , is available for groups of geographically isolated but
nevertheless valid species.

It is no more justifiable in certain cases to treat sympatric pop-
ulations as distinct species than to reduce all allopatric popula-
tions automatically to the status of subspecies. We now know
that the terminal races of a polytypic species may overlap in
range in one part of the distribution area and yet be connected
by more or less continuous intergradation in other parts of the
area. When the population system is considered as a whole, it is
commonly found impossible to split it naturally into more than
one species.

How frequently the condition of sympatry arises within the
limits of a single polytypic species is a moot point. Goldschmidt
( 1952 ) is "disappointed" that "only a few examples, quoted over
and over again, exist to demonstrate what should be a most fre-
quent occurrence." Why we should expect overlapping rings of
races to be a commonplace phenomenon is not clear; but if fresh
examples are desired, it may be recorded that three have turned
up in the last few years in the family Polemoniaceae. Thus Gilia
capitata, the largely allopatric G. tenuiflora-latiflora group, and
Ipomopsis aggregata all exhibit terminal overlap of the extreme
subspecies. The evolutionary significance of this situation cannot
be considered here. As far as the species problem is concerned,
it indicates that marginal sympatry is not, in itself alone, a re-
liable species criterion.

Magnitude of the Species Problem in Plants

A biological species exhibits three characteristics (Mayr, 1948,
1949). It possesses its own distinctive morphological characters,
which arc separated from those of other species by a prominent
liiutus in the variation pattern. It possesses its own particular
ecological requirements, which reduce the competition with
other sympatric forms. It lias a combination of reproductive iso-

V. GRANT 51

lating mechanisms, which prevent or greatly inhibit gene ex-
change with other species.

These characteristics are clearly displayed in most groups of
vertebrate animals. The same facile generalization can not be
made for the higher plants. The circumstance that the biological
species concept has been almost entirely a product of work in
systematic zoology and preeminently in vertebrate systematics
is a reflection of this fact. The only botanist who contributed in
an important way to its development in the early period was Du
Rietz (1930). Botanists as a whole have accepted the biological
species concept much less generally than have the zoologists.

Zoologists are inclined to believe that the difficulties of the
species concept in plants are confined to a few exceptional cases,
i.e., Fisher (1954, p. 87). To the botanist, on the other hand, it
frequently appears as though the clear-cut species is the excep-
tion and the taxonomically critical group is the rule. So charac-
teristic is this impression that one plant taxonomist even felt
obliged recently to apologize for publishing an excellent mono-
graph of a large genus in which there were no outstanding diffi-
culties of species delimitation. The true state of affairs in higher
plants should be expressed in quantitative terms. This large task
can only be commenced here.

In the California flora 32 of the families of seed plants have
four or more genera and 25 or more species in all. Every one of
these 32 large or middle-sized families contains critical groups in
which the species are not clearly defined. In addition many
smaller families in this flora also present real problems of species
delimitation, i.e., the Typhaceae, Iridaceae, Juncaceae, Fagaceae,
Salicaceae, Nyctaginaceae, Apocynaceae, Loasaceae, Lobeliaceae.
The species concept thus runs into practical difficulties in all the
large families, all the medium-sized families, and many of the
small families in the California flora.

Similarly, there is a species problem of varying proportions in
nearly all the large angiospermous genera of the north temperate
zone. To list them all, from Achillea to Zinnia, would be a tire-
some task. If by "good species" is understood those entities which

52

THE PLANT SPECIES

are set apart from all their congeners by a prominent disconti-
nuity in the variation pattern so that botanists have long been
in agreement as to their validity, it is much easier to cite the few
large genera in which all the species are good than the numerous
genera in which at least some of the species are poorly behaved.
Plant genera, like plant families, which have attained any con-

Table I. The Frequency of Good Species in Some Genera of Seed Plants

No.

Good

No.

good

species,

Genus

Region

Family

species

species

%

Trees

Pinus

Pacific slope

Pinaceae

18

12

67

Quercus

Pacific slope

Fagaceae

17

7

41

Shrubs

Ceanothus

California

Rhamnaceae

43?

0

0

Opuntia

Southern Cali-

fornia

Cactaceae

17

0

35

Diplacus

Pacific slope

Scrophulariaceae

7

3

43

Perennial herbs

Aquilegia

Northern hemi-

sphere

Ranunculaceae

5?

2

40

Asdepias

North America

Asclepiadaceae

108

108

100

Crepis

Western North

America

Compositae

10

3

30

Annual herbs

Clarkia

Western hemi-

sphere

Onagraceae

33

4

12

Mentzelia

California

Loasaceae

13

9

G!)

Gilia

Western hemi-

sphere

Polemoniaceae

52

21

40

siderable size and yet present a complete assemblage of well-
defined species are exceptional.

It would be of interest to have some definite data on the rela-
tive frequency of good and poor species in various plant genera.
Table I shows the proportion of good species in eleven genera
representing as many plant families and all the major life forms.
All these genera have been more or less carefully studied by
modern methods. They were selected primarily for that reason
and secondarily in order to illustrate a wide range of evolutionary

V. GRANT 53

patterns. The standard of reference is in each case the total num-
ber of species for a given region according to some recent revi-
sional study or Munz' A California Flora (in press).

As Table I shows, the number and percentage frequency of
discrete species varies from genus to genus within wide limits.
Of the genera represented only Asclepias apparently is without
problematical species for the area considered, and only Ceano-
thus is without any good species in the area considered; but there
may be critical species of Asclepias in Mexico and Central Amer-
ica, and, conversely, Ceanothus seems to have about two good
species in eastern North America. All the genera thus contain
some good species as well as some poorly defined ones.

The frequency of clearly demarcated species varies from group
to group within a single phylad. Thus within the family Pole-
moniaceae the two small genera, Allophyllum and Langloisia,
both consisting of herbaceous annual plants, may be contrasted.
There are four species of Langloisia, all of which are clearly
marked; not one of the five species of Allophyllum, on the other
hand, has well-defined boundaries. The 40 per cent of good
species estimated for the genus Gilia in this same family is an
average figure which covers extreme fluctuations from one infra-
generic group to the next. Of the species composing the sec-
tions Saltugilia and Giliastrum, for example, 60 per cent are good,
whereas the sections Gilia and Arachnion contain only 20 and 14
per cent good species.

Of interest to the zoologist will be the conclusion that critical
species and species groups are the rule rather than the exception
in higher plants. Of interest to those botanists who regard the
species as a completely subjective concept is the demonstration
that at least some discrete units, generally recognized as species,
are to be found in almost every genus.

Causes of the Species Problem

The difficulties encountered in the practical application of a
biological definition of the species in many groups of higher
plants may be referred to several causes.

One factor, which has long been acknowledged to contribute

54 THE PLANT SPECIES

to the imperfect delineation of many species, is the process of
gradual speciation. We find species in all stages of divergence
from the ancestral forms. Some examples of borderline cases
between races and species, as manifested in superspecies com-
posed of very different and discontinuous allopatric forms, in
overlapping rings of races, and in species groups exhibiting mar-
ginal sympatry, have been discussed by Mayr (1942).

Whether to regard a pair of populations in transition from
races to species as one or the other entity may at times require
an arbitrary decision. Such arbitrary decisions have to be made
here and there in most, if not all, of the genera listed in Table I.
This difficulty of interpretation is, however, in these genera and
elsewhere in higher plants, minor in comparison with the diffi-
culties arising from other causes yet to be considered.

A second contributing factor to the species problem is the oc-
currence of sibling species. Species which are virtually indistin-
guishable morphologically are not uncommon in the higher plants.
Every specialist can cite examples from his own group.

Cryptic species in plants are frequently associated with poly-
ploidy. Two diploid species, AA and BB, combine to form a new
allotetraploid species, AABB. The new population does not
merely fill the morphological gap between the original diploids,
but proceeds to segregate variations in the direction of each
parental species. The result is that until cytotaxonomic and cyto-
genetic work is carried out the true biological species will not
be recognized. Instead there will stand some artificial and ad-
mittedly unsatisfactory classification. In the hands of a conserva-
tive taxonomist the entire assemblage may be treated as one spe-
cies, whereas a morphologically minded taxonomist is likely to
recognize two species consisting respectively of AA plus certain
variants of AABB, and BB plus other variants of AABB. As poly-
ploidy continues to the higher levels of hexaploidy, octoploidy,
and so on, the number of species involved in a morphologically
indistinguishable complex of course increases.

Some of the most familiar plants comprise clusters of sibling
species. The sagebrush or Artemisia tridentata group, for exam-
ple, is a complex consisting of an unknown number of biological

V. GRANT 55

species. Being the dominant plant of the Great Basin and a char-
acteristic subdominant in the mountains and high plateaus of
western North America, the sagebrush has inevitably received
much attention from stockmen, foresters, range specialists, botan-
ists, and other naturalists. Nevertheless, despite this general atten-
tion, as well as two taxonomic revisions, one in 1923 based on
field and herbarium studies, and the other in 1953 utilizing
chromosome counts in addition to the classical methods, a classifi-
cation of the sagebrushes which accurately represents the natu-
rally occurring biological units still eludes us. The recent studies
of Ward ( 1953 ) have at least defined the problem and pointed
the way toward its eventual solution. There is reason to hope that
the continuing investigations of Ward, Beetle, and others will
result ultimately in the discovery and delimitation of all the con-
stituent species in the Artemisia tridentata complex. The pros-
pects that botanists will ever be able to determine correctly
many of these species, without field observations, cytological
observations, or both, are not, however, encouraging.

Among the genera listed in Table I, sibling species associated
with polyploidy contribute to the taxonomic difficulties in Gilia,
Clarkia, and Mentzelia.

A third cause of obscure species lines in plants is natural hy-
bridization including introgression. If two or more species begin
to hybridize, and if the derivative forms succeed in establishing
themselves as natural populations, the original discontinuities
between the species may become utterly blurred. This is the
cause of the overlapping ranges of variation of three species of
Gilia shown in Fig. 1. It is the most important cause of taxonomic
difficulties in the Leafy-stemmed Gilias and the Cobwebby Gilias,
the two most critical sections of the genus. The boundary lines
between four species of Cobwebby Gilia are obscured by past
hybridization of each species with two or more of the other four
to such an extent that it is not apparent whether there are four
species or one ( Fig. 2 ) . Interspecific hybridization is the cause of
the fraction of poorly defined species in Pinus, Quercus, Ceano-
thus, Diplacus, and Aquilegia, listed in Table I, and in numerous
other genera of higher plants.

56 THE PLANT SPECIES

The fourth difficulty of the species which will be mentioned
here is the reversion in many plant groups to asexual reproduc-
tion. The biological species is a community of cross-fertilizing
individuals linked together by bonds of mating and isolated re-
productively from other species by barriers to mating. Both the
internal organization and the external boundaries of species are
defined in terms of gene exchange. When sexuality is absent, the
biological species concept breaks down. Since the individual
members of asexual assemblages do not share in a common gene
pool, the group is not integrated into true species. The taxonomic
difficulties in apomictic groups like Crepis sect. Psilochaenia and
the clonally reproducing Opuntki, to cite only two examples listed
in Table I, are due to the development of swarms of variants,
including various hybrid types, which propagate themselves
asexually.

Hermaphroditism, which prevails in the flowering plants, opens
up the possibility of another deviation from sexuality and cross
fertilization as it is known in the higher animals. Many her-
maphroditic plants set seeds by a mixture of outcrossing and
selfing, and many others, particularly among the annual herbs,
are autogamous or automatically self-pollinating. It might be
supposed that these deviations would contribute to partial or
complete breakdown of species integrity in partially or com-
pletely self-pollinating plants. Only under certain circumstances
(as in the autogamous but highly heterozygous Oenothera bien-
nis group) does this appear to be true. Regular self-pollination,
although it modifies the intraspecific variation pattern in im-
portant respects, does not prevent the integration of the separate
populations into species units in most autogamous groups.

The explanation of the compatibility of self-pollination with
species integration seems to be that autogamy is rarely if ever
absolute. The self-pollination, although predominating for a
while, is supplemented by enough rare outcrossing to tie the
individuals together into populations and the populations into
species.

To summarize, natural hybridization ranks high as a source of
a species problem in the higher plants. Sibling species and asexual

V. GRANT 57

reproduction, but not self-pollination, are important contributing
factors. A part of the species problems is due to the continuity of
the speciation process. As far as the higher plants are concerned,
however, the latter factor seems to be a minor source of difficulty
in comparison with the other three causes mentioned.

Biological Species Concept Reconsidered
in the Light of Its Difficulties

Four factors which contribute to the difficulty in application
of the biological species concept have been considered. These
factors are: (1) sibling species; (2) the gradual formation of
species from races; (3) hybridization; and (4) asexual reproduc-
tion. Each factor has a somewhat different bearing on the ques-
tion of the objective reality of the species as a biological unit.

Sibling Species. The difficulties of the species concept arising
from the existence of sibling species are practical, not theoretical,
difficulties. The distinction between sibling species and normal
species is relative. Most sibling species when studied long enough
and hard enough are found to be separable morphologically. Ex-
ternal diagnostic characters were eventually discovered even for
the example par excellence of a pair of cryptic species, Drosophila
pseudoobscura and D. persimilis.

The naming of sibling species in plants as a result of biosystem-
atic studies has provoked a certain discontent among many
herbarium curators and floristic taxonomists. The biosystematist
does indeed have a responsibility to determine and annotate as
many large herbarium collections as is feasible. He also has a
responsibility to science not to suppress his findings merely in
order to facilitate the task of herbarium filing. Two courses are
open to the curator or general taxonomist who is confronted with
undetermined material of sibling species. Let him determine it
as to species group rather than as to species. Or let him scrutinize
the material with more than average thoroughness.

If the taxonomic procedures in different phyla are compared,
as Dobzhansky and Epling (1944) pointed out, the amount of
effort considered necessary for making a determination is not a
standard quantity. Dipterists are in the habit of killing and pin-

58 THE PLANT SPECIES

ning a specimen before identifying it. Bacteriologists, on the
other hand, are accustomed to carrying out physiological tests
with live specimens. In the case of sibling species in Drosophila
or Anopheles, etc., the dipterist may have to emulate the bac-
teriologist by employing physiological tests.

In a similar way phanerogamic taxonomists usually identify
species by characters visible with the naked eye, a hand lens or
a stereoscopic microscope, whereas cryptogamic botanists have
long been habituated to the use of a compound microscope in
taxonomic work. The size of pollen grains and guard cells in flow-
ering plants is a microscopic character which varies proportion-
ately with the level of ploidy in some ( but not all ) polyploid com-
plexes. Where sibling species differ in ploidy level this character
may prove diagnostic. Preliminary studies carried out by the
author and an advanced botany class on herbarium material in
both the Artemisia tridentata and A. vulgaris groups have re-
vealed significant differences in pollen grain and stomatal size
in each one of these polyploid complexes, some of which can be
correlated with known chromosome numbers. Species recognition
in these and other critical groups might be facilitated by meas-
urements of cell size. Of course the chromosome number itself
is a morphological character. Nothing but convention stands in
the way of using characters which require a magnification greater
than 25 X or 50 X in the keys to difficult genera of higher plants.

In the last analysis, if the search for workable macroscopic or
microscopic characters has been unsuccessful, it may prove neces-
sary to give up the attempt to distinguish the specimens of cer-
tain sibling species in the herbarium. There will be no objections
on the part of the biosystematists who make the segregations in
obscure groups to the filing of their indistinguishable species in
the same folder. The species still stand, however, as objectively
real biological units, whether the taxonomist can recognize them
on visible characters or not. A system of classification of the bio-
logical species must be judged, not on the basis of convenience,
but according to whether it represents accurately or inaccurately
the realities in nature.

Gradual Formation of Species. Contrary to the belief held by

V. GRANT 59

Darwin, Wallace, and their early followers, the objective exist-
ence of species is not compromised by the fact that they come
into being by gradual processes of evolution. All the biological
units are capable of orderly growth and development in space and
time. The duplication of the chromosome strands appears to be
achieved in stages, but no one has thought to disqualify the
chromosome as an organizational unit because of its gradual
mode of reproduction. Similarly, the origin of species in most
cases is a process which seems to be completed by stages. During
a transitional stage the diverging population is neither one nor
two species but is an incompletely separated pair. The circum-
stance that the taxonomic hierarchy provides no category for taxa
caught midway between races and species will inevitably obscure
the biological process nomenclaturally, but should not be allowed
to obscure it descriptively.

The above difficulty is probably in the majority of cases more
theoretical than real. Although the evolution of species from races
may be continuous, evolution crosses boundary lines. The bound-
ary line between race and species, as defined by Dobzhansky
( 1935 ) , is that stage of the evolutionary process when an actually
or potentially interbreeding population becomes segregated into
two or more reproductively isolated populations. Compared with
the duration of life of an average species, the crossing of this
boundary line is probably a fairly rapid process. Most populations
at any given instant will consequently be found to exist either
in the stage of races or in the stage of species, and not in an inter-
mediate stage, in so far as their status is affected solely by the
gradualness of speciation.

Natural Hybridization. The difficulties of the species concept
which are brought about by natural interspecific hybridization
are both theoretical and practical. The difficulty is not apparent
and is tacitly overlooked by taxonomists when the hybridization
is limited to the rare formation of a few hybrid individuals. It
cannot be overlooked in either theory or practice when the hy-
brid derivatives establish themselves as an intermediate popula-
tion in nature. A mere trickle of genes across an interspecific bar-
rier, which is manifested in the production of an occasional

60 THE PLANT SPECIES

hybrid, is a very different situation from a mass flow of genes
between species by way of a bridging population of hybrid
origin. In the one case the hybridizing entities are still good
species in the taxonomic sense and only slightly contaminated in
the biological sense; in the second case they have probably
become geographical subspecies; and between these extreme con-
ditions there is every intermediate stage in the merging of spe-
cies.

If it should prove useful to designate the merging of separate
species into a single species by a special term, the process might
well be called secondary speciation. Secondary speciation is that
stage of the evolutionary process at which two or more repro-
ductively isolated populations become combined into one inter-
breeding population. The process is a gradual one and, as with
primary speciation, presents a formal difficulty during the period
of transition.

A concrete example of secondary speciation documented by
fossil records is the progressive fusion of Pinus muricata and P.
remorata since Pleistocene time (Mason, 1949). These two
closed-cone pines were distinct and sympatric on the Santa Bar-
bara mainland of California in the Pleistocene. Since that time a
single variable population with the combined characteristics of
both P. muricata and P. remorata has developed in this area. In
Pleistocene time Pinus remorata was uniform and distinct on
Santa Cruz Island thirty miles off the Santa Barbara coast. Dur-
ing this period P. remorata was probably the only species present
on the island. Today both species exist on the island, P. muricata
probably having arrived in the Late Pleistocene or Post-Pleisto-
cene. Some of the recent populations of P. remorata are no longer
uniform and sharply distinct from P. muricata but, on the con-
trary, are variable and introgressive witli characteristics which
indicate that they arc merging with P. muricata.

If the merging of species proceeds at a rapid rate and goes to
completion, the number of populations which cannot be assigned
to the alternative categories of species or subspecies will lie rela-
tively small. Enough cases are actually known, however, of spe-
cies groups in an intermediate stage of secondary speciation as to

V. GRANT 61

give rise to serious difficulties of species delimitation. Further-
more, so far as the botanist can judge from the study of numerous
cases of natural hybridization, the process of secondary specia-
tion does not necessarily always go to completion, but may re-
main more or less indefinitely in an undefined halfway stage.

Asexual Reproduction. The reversion to asexual reproduc-
tion, as numerous authors have pointed out, spells the end of
species in the biological sense of the word. The smallest inte-
grated unit above the individual is the clone or biotype; the
smallest well-defined taxonomic unit may be a huge polymorphic
complex, an agamic or clonal complex.

For purposes of classification the taxonomist must name and
describe selected morphological types in asexual groups. Whether
these aggregations of individuals should be called "species" or
not is a matter of taste. One tradition favors a general and hence
indiscriminate use of the category of species in all groups of or-
ganisms. Some botanists would qualify this usage in asexual
groups by employing the terms agamospecies or taxonomic spe-
cies. The term agameon has also been suggested (Camp and
Gilly, 1943).

The present author prefers the neutral term, binom, suggested
by Camp ( 1951 ) . The grounds for this preference are that it
saves the term species for the isolated population system and
avoids the confusion of concepts inherent in the application of
the same word to two basically dissimilar phenomena.

In summary, the existence of the species as an objective bio-
logical unit is not impaired by morphological indistinctness or
by the continuity of the evolutionary process. The loss of sex-
uality, on the other hand, removes the very foundations on which
the species exists as a type of breeding population. As a result,
biological species do not exist in asexual groups.

The consequences of natural hybridization for the biological
species concept cannot be stated so categorically. The discrete-
ness of the species unit is a relative matter in a hybridizing com-
plex. An attempt to evaluate the significance of natural hybridiza-
tion for the objective reality of species will be made in the next
section.

(32 THE PLANT SPECIES

Significance of Hybridization for the Species Concept

The bearing of natural hybridization on the species concept
may be illustrated by three concrete examples selected from
Gilia. All these examples involve diploid sexual organisms with a
normal chromosome cycle and a rate of natural cross fertilization
which varies from high to low but never reaches exclusive in-
breeding. The integrity of the species as a biological unit is
slightly impaired in the case of the first two examples but it
breaks down in a more serious way in the third example. It is
this third case that proves the most instructive for the question
of the objective reality of species limits in a hybrid complex.

The first case is that of Gilia capitata, a widely distributed spe-
cies on the Pacific slope of North America, and Gilia millefoliata,
a species of the coastal sand dunes. The two species come into
contact in a number of places along the Pacific coast. Morpho-
logically, they are amply distinct, G. capitata being an erect plant
with globose flowering heads and G. millefoliata a prostrate
spreading plant with spotted flowers borne in small clusters. They
are also well isolated by sterility barriers and can be crossed only
with difficulty in the experimental garden. The hybrids then are
highly sterile and are characterized by a low degree of chromo-
some pairing. On Cape Mendocino in northern California a local
breakdown in the complex of reproductive isolating mechanisms
has led to a limited amount of natural hybridization. The bound-
aries of the species are, however, quite clearly defined even here,
which suggests that the populations are maintaining themselves
as two distinct species.

The second case concerns Gilia capitata itself (Grant, 1950).
This is a polytypic species composed of eight subspecies which
align themselves morphologically and ecologically into three
main groups. These major subdivisions of the species in all proba-
bility represent a trio of formerly distinct species. Vestiges of the
original reproductive isolation between the ancestral species still
persist in a few areas ol syinpatric overlap. The extreme popula-
tions are connected by a complete series of intergradations. Sinee
the intermediate populations occupy geologically more recent

V. GRANT 63

habitats than the extremes, since they are highly variable from
colony to colony whereas the extreme forms are quite uniform,
and since the variability of the intergrades shows the character
correlations which are expected when a hybrid segregates for
two or more multifactorial characters, it is judged that the inter-
mediates are historically more recent than the extremes and are
of hybrid origin between them. If Gilia capitata can be regarded
as a single species, therefore, it must be one of secondary origin
from the union of previously isolated populations.

The most serious difficulties of classification are encountered
in the Gilia tenuiflora-latiflora group (Grant and Grant, 1956).
From the standpoint of intergradation and gene exchange the
whole group, comprising Gilia tenuifora, G. latijiora, G. diegen-
sis, G. leptantha, and G. cana, could be regarded as a single
species. That it is in reality more than this, however, is indicated
by various facts.

The major constituents of the complex are more different from
one another than any conspecific elements known in the genus.
Each major constituent is comparable to other large species in
Gilia and is in fact polytypic with varying numbers of well-
marked subspecies. The reasoning involved in this argument does
not presuppose a morphological species definition but rather a
certain general correlation within a genus, or in this case even
within a section, between morphological differentiation and bio-
logical species formation (Mayr, 1942; Rollins, 1952).

The treatment of Gilia teniiifloia, G. latiflora, and G. cana as a
single species would obscure rather than accurately portray their
relationships. Each species named has other genetically close rel-
atives which are without question specifically distinct, whereas
a pronounced reduction in fertility associated with a lowered
degree of chromosome pairing is found in the hybrids between
the major elements. Finally, a number of sympatric occurrences
are known between the major elements. All these facts are diffi-
cult to reconcile with the idea that the major entities in the Gilia
tenuiflora complex are the members of a single species in any
normal sense of the word.

The existing evolutionary pattern in the Gilia tenuijlora group

64 THE PLANT SPECIES

can be explained on the hypothesis that G. tenuiflora, G. latiflora,
and G. cana went through a stage of barely completed primary
speciation which was followed by extensive hybridization in
various combinations (Fig. 2). The hybridization has gone far
enough to prevent the original extreme populations from persist-
ing as clear-cut species, but has not fully completed the process
of secondary speciation. The major constituents thus possess some
of the characteristics of subspecies in so far as they replace one

Fig. 2. The syngameon of Gilia tenuiflora-latiflora, consisting of four
semispecies linked by natural hybridization in nearly every possible com-
bination. The abbreviations stand for Gilia tenuiflora, G. cana, G. latiflora,
and G. leptantha.

another geographically, intergrade in the zones of contact, and
retain a certain degree of interfertility; they also possess some
characteristics of species, as manifested in marginal sympatric
contacts and in a degree of morphological and genetic divergence
which is comparable to that of species elsewhere in the phylad.
The situation just described is not confined to Gilia. It is
known also in Finns, Quercus, and Aquilegia, to cite other exam-
ples from among those listed in Table I, and in numerous other
genera.

V. GRANT 65

Davidson (1954) has pointed out, "If we wish to argue that
species have an objective reality, we may select examples such
as Ginkgo. If we wish to argue that species do not exist in nature
at all, we may select such examples as the willows." The species
of willow are notoriously confused by natural hybridization,
whereas this factor can scarcely affect Ginkgo biloba, which is
the sole living member of its phylad.

The frequent occurrence of the willow pattern of evolution,
as contrasted with the ginkgo pattern, is perhaps reflected in the
species concept held by so many of the nineteenth century plant
hybridizers. Their practical experience impressed upon many of
these early students, Naudin, Nageli, Darwin, Mendel, Kerner,
De Vries, and others, as it has also on many modern botanists,
an explicit disbelief in the notion that the species as a unit is
qualitatively distinct from the race.

The failure of conformity of the evolutionary developments in
certain groups of plants to the biological species concept in its
present formulation suggests the need for a reexamination of our
definitions. It sometimes proves necessary to split a biological
unit into two or more units. As previously mentioned, the incon-
sistencies of the "concept of the individual" when applied to
clones forced a recognition of both individuals and clones; and a
splitting of the gene into functional and recombinational genes
is apparently in the offing. Similarly, the evidence derived from
evolutionary studies of higher plants indicates the existence of
a unit of interbreeding higher than the species. That unit is the
sum total of the species linked together by frequent or occasional
hybridization.

Categories supplementary to the species have been proposed
by several authors. The coenospecies of Turesson (1922) and the
commiscuum and comparium of Danser (1929), being based on
the fertility relationships of species in the experimental garden,
are artificial categories of use only for the discussion of a special
aspect of natural populations. The coenogamodeme and syngamo-
deme of Gilmour and Heslop-Harrison (1954) are synonymous
with coenospecies and comparium respectively. The term micton
of Camp and Gilly (1943), although corresponding to a natural

66

THE PLANT SPECIES

©

ro

CJ

c

o

o

<9

<u

Q.

F

CO

<D

o

1

o

o>

E

o>

c

a.

>N

in

CO

*

3>

<u

o.
CO

g
o

&

CO

a
E

c CO
o>

u

■i)
o.
CO

o

TJ
c
o
o
u>
CO

a

c

O
CD

a

t/2

CD

3

v.

C

o
o

be

c

O

■XI
1/3

X

CD

a

o
o

o

£

C3
bC

o

o

C3

c

CD

£

a

o

CD

■

>

CD

o

T3

•t— <

o

CD
■XI

-t-J

CD

o

(3

a

CD

'>

O

ffl

>.

XI

^^

Pu

<D
>

CD

i--1

CO

CD

F

hii

tn

-4-J

(5

CD

V. GRANT 67

unit, is by definition a particular type of species, a species namely
of hybrid origin, and is not applicable to an assemblage of hy-
bridizing species. The commonly used term, species group, on
the other hand, connotes an assemblage of related species but
carries no implication of natural hybridization between them,
and the same is true of the superspecies of Mayr (1942).

The homogamic complex proposed by Grant (1953) for a
hybrid complex, the derivative forms of which are sexual, meioti-
cally normal, and entirely or predominantly diploid, is not en-
tirely suitable for our present purpose either, since this concept
carries the connotation of past as well as present hybridization
on an extensive scale. The homogamic complex is a type of phylo-
genetic reticulum which, at any given instant of time, may or
may not correspond to the group of interbreeding species. The
former is thus a more inclusive category than the latter, for it
may contain species which were formerly united by hybrid con-
nections but no longer actually hybridize.

In order to find a convenient means of designating the hybrid-
izing species complex we must go back to the term syngameon
of Lotsy (1925, 1931). The syngameon was defined by Lotsy as
a "habitually interbreeding community" or "pairing community."
Thus stated, the syngameon is nothing more than the breeding
population or perhaps the biological species. The fact that Lotsy
used the term species in a very narrow sense for groups of like
individuals lends strength to the interpretation that his syn-
gameon was equivalent to what we would now call species. By
the context and the choice of illustrative examples, particularly
the European Betulas, however, it is perfectly clear that Lotsy
conceived the syngameon in a broader sense than the biological
species. His syngameon, as applied to concrete cases, was a clus-
ter of hybridizing species. We are justified, therefore, in taking
up this old, generally familiar, and etymologically appropriate
term for the interbreeding unit higher than the species.

The syngameon, as redefined according to present-day con-
cepts, is the sum total of species or semispecies linked by fre-
quent or occasional hybridization in nature; a hybridizing group
of species; the most inclusive interbreeding population. A dia-

68 THE PLANT SPECIES

grammatic illustration of the changing composition of a syn-
gameon within a homogamic complex is given in Fig. 3.

Reference to the fertility or sterility of the hybrid types is pur-
posely omitted from the foregoing definition, contrary to earlier
efforts in the era of Lotsy, Turesson, Danser, etc., as not being
of major importance for the development of a syngameon.

In a syngameon the original biological species either may per-
sist as discrete entities, as exemplified by Gilia capitata and G.
millefoliata in the syngameon to which they belong, or come to
occupy a definitely subordinate position to the total complex. In
the latter case the species as a biological unit is found in varying
degrees of dissolution. When species-like entities are discernible
on the basis of partial reproductive isolation externally contrast-
ing with free gene exchange internally, these constituent units
can be designated, following Sibley ( 1954 ) , as semispecies. ( This
term was used earlier by Mayr (1940) in a somewhat different
sense.) The syngameon of Gilia tenuiflora-latiflora, for example,
consists of four semispecies as shown in Fig. 2. In some syn-
gameons, finally, the original species have become more or less
completely swamped by hybridization, as has happened with
staminea and capitata in the development of the present Gilia
capitata.

The species, semispecies, or remnants of species are organized
into a higher unit by frequent or infrequent gene exchange across
presently or formerly existing reproductive barriers. The most
inclusive interbreeding population is, consequently, not the spe-
cies but the syngameon. Depending on the degree of integration
attained in a syngameon, its constituent species may be preserved
as such or partially or completely lost. The syngameon is in any
case a biological unit in its own right.

The Botanical versus the Zoological Species Problem

If comparisons are made between taxonomically well studied
groups of seed plants and vertebrate animals, the boundaries of
species are found to be, on the average, less well marked in the
plants than in the animals. The causal factors of the species prob-
lem can accordingly be expected to exhibit a contrasting degree

V. GRANT 69

of influence in plants and animals. A comparative approach to
the species problem may thus throw some light on its causes in
higher plants. Such an approach may even shed light on the
ultimate causes behind the immediate causes.

Since the factor of gradual speciation is common to both ani-
mals and plants, it cannot be cited as a cause of the proportion-
ately greater species problem in plants unless certain additional
assumptions are made. If the rate of speciation is higher in plants
than in animals, and if this speciation is still continuing in plants
but has tapered off or ceased in many animal groups, the propor-
tion of poorly defined species due to incomplete divergence will
be greater in plants.

The first assumption can be safely granted. There are about
six times as many species of angiosperms as vertebrates in the
world (ca. 250,000 to 40,000), even though the angiosperms orig-
inated later in geological history than the vertebrates. Whether
this speciation is continuing into the recent period more actively
in plants than in animals, however, is difficult to judge with the
evidence available. The Age of Man has probably retarded evo-
lution in the larger land mammals and speeded evolution up in
weedy and cultivated plants, but the contrast between higher
plants and animals with respect to the clarity of species is not
confined to these particular groups. All that we can safely con-
clude is that the process of gradual species formation will result
in a larger number, but not necessarily a higher proportion, of
poorly defined species in higher plants as compared with verte-
brate animals.

Since, so far as we can judge, incomplete speciation is not a
major contributing factor to the species problem in plants, any
differences that do exist between plants and animals with regard
to the relative frequency of incompletely formed species can
probably be ignored for the purpose of the present analysis.

A second cause of the species problem, the occurrence of sib-
ling species, does show marked differences in frequency in plants
and animals. Sibling species are of course well known in animals,
being fairly common in the Diptera and occurring also in the
vertebrates. In the latter they are apparently not common. In the

70 THE PLANT SPECIES

higher plants, by contrast, cryptic species are not at all uncom-
mon; examples can be found in nearly every large family
and in a majority of large genera. When we recognize the
fact that cryptic species are a common byproduct of polyploidy,
and that polyploidy is a commonplace mode of speciation in
plants but an exceptional phenomenon in animals, we can readily
account for the observed differences between plants and animals
as to the frequency of sibling species.

Similarly, asexual reproduction, which is extremely rare in the
vertebrates and uncommon in the insects, is a normal means of
propagation in higher plants. Vast polytypic assemblages of
asexual microspecies, such as the agamic complex in Crepis sect.
Psilochaenia or the clonal complex of the Opuntia phacacantha
group, have no known counterpart in the higher animals, al-
though they represent a standard pattern of evolution in plants.

A prominent difference in the mode of reproduction between
higher plants and higher animals is the almost universal dioecism
of the latter and the very frequent hermaphroditism of the
former. Many hermaphroditic plants, moreover, reproduce par-
tially or predominantly by self-pollination. As already noted,
however, the species problem does not seem to be materially
increased in plants over that in animals by this difference in mode
of reproduction alone. Apparently a very low rate of natural cross
pollination in many autogamous plants is consistent with an in-
tegration of the members into species units.

Natural hybridization seems to play a rather different role in
the evolution of plant and animal species. Natural hybrids have
of course been recorded in various animal groups, as exemplified
by Cockrum's (1952) check list for birds. Attention lias also been
paid to the effects of natural hybridization on the variation pat-
tern and on speciation in several groups of animals, as for in-
stance birds (Mayr and Gilliard, 1952; Sibley, 1954; Miller, 1955),
mammals (McCarley, 1954; Rudd, 1955), amphibia (Blair, 1941,
1955; Volpe, 1952), 'fishes (Hubbs, 1955), butterflies (Hovanitz,
1919), ;iud snails (Ilubendiek, 1951), to cite only a few repre-
sentative studies. In the fishes this hybridization may be a fairly
common and significant process. The consensus among zoologists,

V. GRANT 71

however, as expressed by Mayr (1942), Blair (1951), Hubendick
(1951), Dobzhansky (1953), Fisher (1954), and other authors,
is that hybridization is a rare and abnormal event which usually
ends with the production of a few interesting but inconsequential
freaks.

No similar categorical conclusion with regard to the plant king-
dom can be seriously considered by botanists. That natural hy-
bridization has evolutionary consequences of considerable im-
portance was suggested in an early period by Naudin (1863),
Kerner (1891), Lotsy (1916), and others. Botanical studies, in
numerous groups of higher plants, utilizing more refined tech-
niques developed especially by Anderson, have amply confirmed
and extended the conclusion reached by the pioneer hybridolo-
gists that hybridization has profoundly affected the course of
plant evolution (cf. Anderson, 1953; Grant, 1953; Anderson and
Stebbins, 1954). This conclusion seems to be as true for the lower
vascular plants as for the seed plants (Manton, 1950; Wagner,
1954).

A consideration of both the zoological evidence and the bo-
tanical evidence leads us to a tentative general conclusion. The
zoological evidence now available, accepted at its face value, sug-
gests that natural interspecific hybridization may have been a
small or even a negligible factor in the evolution of many higher
animals. The botanical evidence proves that natural interspecific
hybridization has been an important evolutionarv process in the
plant kingdom.

The direct effects of natural hybridization on the species prob-
lem in plants are plain enough. These effects range in magnitude
from a slight blurring of species distinctions to the development
of a huge intricate syngameon in which the original species are
more or less lost as distinct entities. It remains to point out that
natural hybridization is indirectly involved in two other sources
of a species problem in plants.

Sibling species, as previously noted, are a byproduct of poly-
ploidy; polyploids outside the experimental garden are in the
great majority of cases allopolyploids; and allopolyploids are the
polyploid derivatives of interspecific hybrids. Asexual reproduc-

72 THE PLANT SPECIES

tion, similarly, affects the species problem most markedly where
it is associated with hybridity, as is the case in both the agamic
complex and the clonal complex. Even self-pollination, which
under normal conditions is compatible with good species divi-
sions, can if associated with hybridity and other peculiarities of
the genetic system cause a disintegration of true species, as illus-
trated by the heterogamic complex of the Oenothera hookeri-
biennis group.

Natural hybridization is thus directly or indirectly involved in
nearly all the major sources of a species problem in plants. The
conclusion is inescapable that this process is the most important
single factor in the breakdown and indistinctness of plant species.
The conclusion that plant species are more affected by hybridiza-
tion than animal species agrees well with the observed differences
in the average clarity of species boundaries in the two kingdoms.

Physio-Genetic Nature of the Species

Among evolutionists there has been some speculation as to the
reasons for the generally higher level of reproductive isolation in
animals than in plants. Ethological isolating mechanisms prob-
ably account for a part of the difference. Specific behavior pat-
terns which stimulate conspecific matings but inhibit interspecific
matin gs constitute an important element in the complement of
species-separating barriers in most animals. In higher plants such
isolating mechanisms, while not entirely wanting, are certainly
represented in a much weakened form. The sterility barriers also
seem to be more nearly absolute in many animal groups than in
plants. This may be a consequence of the apparent prevalence
of genie sterility in animals and of chromosomal sterility in plants
(Dobzhansky, 1951), as well as of the greater absolute fecundity
of long-lived plants (Stebbins, 1950).

Explanations such as these, however, merely touch the surface
of the problem. If animal species are in general less contaminated
by hybridization than plant species, the ultimate cause must be
sought not in the strength of the isolating mechanisms but in
the evolutionary forces which build up those isolating mecha-
nisms. A possible explanation may be that natural selection favors

V. GRANT 73

a certain amount of transgression of species barriers in plants,
whereas in animals a more rigid isolation of species is of selective
value. This hypothesis is consistent with other underlying differ-
ences between plants and animals.

The most basic difference between plants and animals is in the
degree of complexity of the individual organism. The develop-
ment of the relatively simple plant body is commensurate with
an open system of growth by which individual parts are built up
in series. The animal body is a far more complex and delicately
balanced system which must develop as a whole without dis-
ruption of the internal organization.

Although very little is actually known about this question, it
is probable that the genetic determinants of growth and develop-
ment correspond in complexity to the degree of integration of
the individual organism. Thus the factorial basis of individual
character differences between races and species of plants is not
infinitely complex but can often be resolved into a fairly small
number of genes (Clausen, 1951; Baker, 1951; Burton, 1951;
Grant, 1950). The more simple characters in animals, such as
coat color or blood types in mammals (Castle, 1954; Race, 1950),
can likewise be explained in terms of a few genes. By contrast,
such fundamental characters of animals as the sex differences are
determined by exceedingly complex systems of genes which defy
an exact factorial analysis.

The genie complement of any species is the product of a long
period of natural selection. The injurious effects of an unrestricted
influx of foreign genes from other species are the probable basis
of ad hoc isolating mechanisms (Dobzhansky, 1951). A species
possessing a highly integrated and finely balanced gene system
will, moreover, suffer relatively worse effects from hybridization
than one with a simpler and more loosely integrated set of ge-
netic determinants. In the case of the latter type of species, the
disadvantageous effects of interspecific gene exchange may even
be outweighed by the advantages resulting from an increase in
variability.

The ability of a species to tolerate the presence of foreign genes
is thus a function of the degree of complexity and integration

r

^', • _..*.

74 THE PLANT SPECIES

attained by its system of genie determinants. The chances of ob-
taining a favorable recombination of the genotypes of two species
will vary from nil to high, depending upon the amount of internal
coordination in each specific genotype. A perfectly functioning
bicycle can be built up out of the parts of several different makes
of bicycle, but a watch containing the cogwheels of two or more
different makes will probably not run accurately and may not
run at all.

It is highly unlikely that the partial to complete breakdown of
the species as a biological unit could continue on a wide scale
in many plant genera without the sanction of natural selection.
In order to explain this result we must show first that the relaxa-
tion of isolating barriers is not necessarily disadvantageous to a
population, and secondly that the resulting gene flow may confer
positive advantages. The hypothesis that a simple physiological-
morphological organization in plants, reflecting a correspondingly
simple genie pattern, permits a certain freedom in the exchange
of genes between species which is not enjoyed by the vastly more
complex animals, satisfies the first condition. The increase in
variability resulting from the pooling of the mutations of not one,
but two or several species populations is the advantageous aspect
of hybridization demanded by the second condition. Together
these conditions are capable of explaining an important differ-
ence between plants and animals as to the nature and behavior
of their species.

Conclusions

According to tin1 biological species concept the species is a
population set apart from the rest of the living world by repro-
ductive isolating mechanisms. Its boundaries, therefore, should
be marked by a prominent gap in the variation pattern. Good
discrete species are in fact found in all major plant groups. These
well-defined species, however, constitute only a fraction of the
populations in their respective phylads. The botanist finds prob-
lematical species intermixed in varying proportions with good
species in most of the families and genera. This is in marked eon-

V. GRANT 75

trast to the situation which the zoologist finds in higher animals
in which clearly circumscribed species are the rule and poorly
defined ones the exception.

A survey of some typical plant genera in which systematic
difficulties are encountered shows that the species problem is due
to a variety of causes. ( 1 ) Where species originate from races by
a continuous and gradual process of evolution, some populations
at any given instant will be in a halfway stage between race and
species. (2) Some populations which must be acknowledged as
species on the basis of reproductive isolation do not happen to
possess easily visible distinguishing characters. (3) The isolation
of species may not be complete with the result that some inter-
mediate populations of hybrid origin may arise. (4) Plants often
revert to asexual means of propagation in which case the very
basis of the organization of populations into species, namely in-
terbreeding and exchange of genes internally and reproductive
isolation externally, is lost. The most important single cause of a
species problem in plants is natural hybridization.

How does the biological species concept stand up in the light
of the various difficulties of application listed above? The objec-
tive reality of the species as a biological unit is not impaired by
the gradual formation of species or by the existence of cryptic
species. The loss of sexuality, on the other hand, simply means
that true biological species do not exist in certain plant groups.
The effects of natural hybridization on the objective reality of
the species, finally, cannot be summarized in categorical terms,
since the discreteness of the species unit is a relative matter in a
hybridizing complex.

Inasmuch as the species is defined as an interbreeding popula-
tion isolated from other populations by reproductive barriers,
natural hybridization is a factor opposed to the integrity of the
species. If the hybridization is very limited in extent, its effects
may be overlooked in taxonomic practice. But if the hybridiza-
tion continues on an extensive enough scale, the result may be
the breakdown of the original species. Between these extreme
conditions there is a wide range of intermediate stages in species
disintegration due to hybridization. Where the process of hybridi-

76 THE PLANT SPECIES

zation runs the full course to secondary speciation, the original
extreme populations undergo a progressive change in status from
species through semispecies to subspecies.

Many existing plant groups, as exemplified by certain sections
of the willows, oaks, birches, columbines, Ceanothiis, and Gilia,
are aggregations of hybridizing semispecies. The major elements
composing these vast polymorphic assemblages are not good
species in the usual sense because they interbreed more or less
freely with one another. It is equally inappropriate to regard them
as merely races and the whole assemblage as a single species,
because the extreme forms show partial reproductive isolation
from one another and have attained a degree of morphological
differentiation equivalent to that of good species in other
branches of the phylad. The solution suggested for this difficulty
of the species concept is to recognize the whole assemblage as a
unit of interbreeding higher than the species and to designate it
by the term syngameon of Lotsy. The syngameon is defined as a
group of hybridizing species or semispecies.

The "splitting of the species" is not without precedents in biol-
ogy. The conceptual difficulties of defining the individual in plant
groups which propagate asexually forced a recognition of two
classes of units: individual, as defined on the basis of physiolog-
ical autonomy, and clone, the sum total of the individuals de-
scended from a single zygote. Physiological geneticists are cur-
rently splitting the gene into two distinct units, having found that
the unit of biochemical activity in the chromosome is not the
same as the unit of crossing over. Similarly in higher plants it is
necessary to recognize the existence of both the species and the
syngameon or complex of species linked together by hybridiza-
tion.

The final question as to why the syngameon has developed as
a common evolutionary unit in higher plants and not in higher
animals is tantamount to the question why plant species have
been so much more affected by hybridization than animal species.
A comparison of higher plants and animals with respect to the
degree of organization of the genotype may help to explain these
differences. The great complexity and internal coordination of

V. GRANT 77

the animal body probably reflects an equally complex and deli-
cately balanced ensemble of polygenes, whereas the far more
simply organized plant body is the resultant of a simpler and
more loosely integrated gene system. Foreign genes will conse-
quently be much less likely to fit harmoniously into the highly
integrated genotype of an animal than into the more loosely
organized genotype of a plant. Natural selection will accordingly
eliminate the products of interspecific hybridization, or prevent
the hybridization from occurring by the erection of breeding
barriers, to a greater extent in animals than in plants.

Acknowledgments. The manuscript was critically reviewed
by Dr. Jay M. Savage, of the University of Southern Cali-
fornia, to whom the author is indebted for numerous helpful
suggestions. Mr. Howard Latimer of the Rancho Santa Ana
Botanic Garden also made constructive suggestions with regard
to one section of the manuscript. The problems of classification
in certain difficult western American genera of plants were dis-
cussed with Dr. Philip A. Munz of the Rancho Santa Ana Bo-
tanic Garden. The interest and advice of these colleagues is
gratefully acknowledged. Finally, the deliberations in Atlanta
brought out several new points, which were taken into considera-
tion in the preparation of the final manuscript.

The work on the Cobwebby Gilias, incorporated into the dis-
cussion of the effects of hybridization on species distinctness,
was accomplished with the aid of a research grant from the Na-
tional Science Foundation, which is likewise thankfully ac-
knowledged.

REFERENCES

Anderson, E. 1953. Introgressive hybridization. Biol. Revs. Cambridge
Phil. Soc, 28, 280-307^

Anderson, E., and G. L. Stebbins, Jr. 1954. Hybridization as an evolu-
tionary stimulus. Evolution, 8, 378-88.

Baker, H. G. 1951. Hybridization and natural gene-flow between
higher plants. Biol. Revs. Cambridge Phil. Soc., 26, 302-37.

78 THE PLANT SPECIES

Blair, A. P. 1941. Variation, isolation mechanisms, and hybridization
in certain toads. Genetics, 26, 398-417.

Blair, A. P. 1955. Distribution, variation, and hybridization in a relict
toad (Bufo microscaphus) in southwestern Utah. Am. Museum
Novitates, No. 1722, 1-38.

Blair, W. F. 1951. Interbreeding of natural populations of vertebrates.
Am. Naturalist, 85, 9-30.

Burma, B. H. 1954. Beality, existence, and classification: a discussion
of the species problem. Madrono, 12, 193-209.

Burton, G. W. 1951. Quantitative inheritance in pearl millet (Pennise-
tum glaucum). Agron. J., 43, 409-17.

Camp, W. H. 1951. Biosystematy. Brittonia, 7, 113-27.

Camp, W. H., and C. L. Gilly. 1943. The structure and origin of spe-
cies. Brittonia, 4, 323-85.

Castle, W. E. 1954. Coat color inheritance in horses and in other mam-
mals. Genetics, 39, 35-44.

Clausen, J. 1951. Stages in the Evolution of Plant Species. Cornell
University Press, Ithaca, New York.

Cockrum, E. L. 1952. A check-list and bibliography of hybrid birds in
North America north of Mexico. Wilson Bull., 64, 140-59.

Danser, B. H. 1929. Uber die Begriffe Komparium, Kommiskuum und
Konvivium und liber die Entstehungsweise der Konvivien. Genetica.
11, 399-450.

Darlington, C. D. 1940. Taxonomic species and genetic systems. In The
New Sysf( matics. Oxford University Press, Oxford, England.

Davidson, J. F. 1954. A dephlogisticated species concept. Madrono,
12,246-51.

Dobzhansky, T. 1935. A critique of the species concept in biology.
Philosophy of Science, 2, 344-55.

Dobzhansky, T. 1951. Genetics and the Origin oj Species. 3rd ed. Co-
lumbia University Press, New York, N.Y.

Dobzhansky, T. 1953. Natural hybrids of two species of Arctostaphylos
in the Yosemite region of California. Heredity, 7, 73-79.

Dobzhansky, T., and C. Epling. 1944. Contributions to the genetics,
taxonomy, and ecology of Drosophila pseudoobscura and its rela-
tives. Carnegie Inst. Wash. Puhl. No. 544.

l)u Rietz, G. E. 1930. The fundamental units of biological taxonomy.
Svensk Botan. Tidskr., 24, 333-428.

Fisher, II. A. 1954. Retrospect of the criticisms of the theory of natural

V. GRANT 79

selection. In Evolution as a Process, Allen and Unwin, London.

Gilmour, J. S. L., and J. Heslop-Harrison. 1954. The deme terminology
and the units of micro-evolutionary change. Genetica, 27, 147-61.

Goldschmidt, R. B. 1952. Evolution, as viewed by one geneticist. Am.
Scientist, 40, 84-98.

Grant, A., and V. Grant. 1956. Genetic and taxonomic studies in Gilia.
VIII. The Cobwebby Gilias. El Aliso, 3, 203-87.

Grant, V. 1950. Genetic and taxonomic studies in Gilia. I. Gilia capi-
tata. El Aliso, 2, 239-316.

Grant, V. 1952. Genetic and taxonomic studies in Gilia. III. The Gilia
tricolor complex. El Aliso, 2, 375-88.

Grant, V. 1953. The role of hybridization in the evolution of the Leafy-
stemmed Gilias. Evolution, 7, 51-64.

Hovanitz, W. 1949. Increased variability in populations following nat-
ural hybridization. In Genetics, Paleontology, and Evolution. Prince-
ton University Press, Princeton, New Jersey.

Hubbs, C. L. 1955. Natural hybridization in fishes. Systematic Zool.,
4, 1-20.

Hubendick, B. 1951. Recent Lymnaeidae. Their variation, morphology,
taxonomy, nomenclature, and distribution. Kgl. Svenska Veten-
skapsakad. Handl, 3, 1-223.

Kerner, A. 1891. Pflanzenleben. 2 vols. Leipzig and Vienna.

Lotsy, J. P. 1916. Evolution by Means of Hybridization. The Hague.

Lotsy, J. P. 1925. Species or Linneon. Genetica, 7, 487-506.

Lotsy, J. P. 1931. On the species of the taxonomist in its relation to
evolution. Genetica, 13, 1-16.

Manton, I. 1950. Problems of Cytology and Evolution in the Pteridoph-
yta. Cambridge University Press, Cambridge, England.

Mason, H. L. 1949. Evidence for the genetic submergence of Pinus
rcmorata. In Genetics, Paleontology, and Evolution. Princeton Uni-
versity Press, Princeton, New Jersey.

Mason, H. L. 1950. Taxonomy, systematic botany and biosystematics.
Madrono, 10, 193-208.

Mayr, E. 1940. Speciation phenomena in birds. Am. Naturalist, 74,
249-78.

Mayr, E. 1942. Systematics and the Origin of Species. Columbia Uni-
versity Press, New York City, N.Y.

Mayr, E. 1948. The bearing of the new systematics on genetical prob-
lems. The nature of species. Adv. in Genetics, 2, 205-37.

80 THE PLANT SPECIES

Mayr, E. 1949. Speciation and selection. Proc. Am. Phil. Soc, 93, 514-

19.
Mayr, E., and E. T. Gilliard. 1952. Altitudinal hybridization in New

Guinea honeyeaters. Condor, 54, 325-37.
McCarley, W. H. 1954. Natural hybridization in the Peromyscus leu-

copus species group of mice. Evolution, 8, 314-23.
Miller, A. H. 1955. A hybrid woodpecker and its significance in specia-
tion in the genus Dendrocopos. Evolution, 9, 317-21.
Naudin, Ch. 1863. Nouvelles recherches sur l'hybridite dans les vege-

taux. Ann, sci. nat. (Paris) (Botan.), 19, 180-203.
Race, R. R. 1950. The eight blood group systems and their inheritance.

Cold Spring Harbor Symposia Quant. Biol, 15, 207-20.
Rollins, R. 1952. Taxonomy today and tomorrow. Rhodora, 54, 1-19.
Rudd, R. L. 1955. Variation and hybridization in shrews. Systematic

Zool, 4, 21-34.
Sibley, C. G. 1954. Hybridization in the red-eyed towhees of Mexico.

Evolution, 8, 252-90.
Stebbins, G. L., Jr. 1950. Variation and Evolution in Plants. Columbia

University Press, New York City, N.Y.
Turesson, G. 1922. The genotypical response of the plant species to

its habitat. Hcreditas, 3, 211-350.
Volpe, E. P. 1952. Physiological evidence for natural hybridization of

Bufo americanits and Bufo fowled. Evolution, 6, 393-406.
Wagner, W. H. 1954. Reticulate evolution in Appalachian Aspleniums.

Evolution, 8, 103-18.
Ward, G. H. 1953. Artemisia, section Scriphidium, in North America.

A Cytotaxonomic study. Contribs. Dudley Herbarium, 4, 155-205.

THE SPECIES PROBLEM IN
FRESHWATER ANIMALS

JOHN LANGDON BROOKS: yale university, new haven,

CONNECTICUT

Most of the problems encountered in the systematics of fresh-
water animals are of the same nature as those met in the system-
atics of terrestrial or marine animals. The only justification for
considering these problems in freshwater animals is the possibil-
ity that some of the complicating factors may operate with a
peculiar intensity or in unusual combinations in the freshwater
fauna. Whereas a comprehensive survey of the systematics of
the diverse animal stocks that comprise this freshwater fauna
might reveal the information we seek, such a survey would be
lengthy and our labors might not find a proportionate reward.
What I propose to do in this paper is to select from biologically
different, but equally successful, freshwater stocks two genera
the systematics of which have proved difficult to comprehend.
By analyzing the "species problem" in each, it will be possible
to see whether the sources of the systematic difficulty in the two
groups are at all similar. If any similarities are evident, it can
tentatively be concluded that these represent evolutionary
processes working with peculiar intensity in fresh waters.

Generalizing from selected, extreme examples must be done
with proper caution both in the selection of the cases and in the
force and extent of the generalizations. The two genera selected
are Daphnia, small planktonic crustaceans, and Coregonus, the
true whitefishes. These are two of the roughly half-dozen sys-
tematically most troublesome genera of common freshwater
animals. Daphnia was chosen because that is the group with
which I am most familiar, and for this reason the greater part
of our attention will be devoted to it. Much of the information

81

82 FRESHWATER ANIMALS

about this genus is based upon investigations described more
fully in a monograph on the systematics of North American
Daphnia (Brooks, 1956). Coregonus was selected as the second
group because it represents a very different kind of organism ( one
especially different from Daphnia in its reproductive behavior),
its systematics has proved particularly thorny, and, possibly
most important, it has been the subject of a penetrating study
by Gunnar Svardson (1949-1953), of the Institute of Freshwater
Research at Drottningholm, Sweden. Whereas the genera chosen
represent a reasonable biological diversity, they are both groups
which flourish especially in the glaciated areas of North Amer-
ica and Eurasia. Thus both groups are alike in having been
subjected to the disruptive influences of the Pleistocene glacia-
tions. However, as most of the bodies of standing fresh water
are now in areas which were glaciated, the rich fauna which has
developed in these waters will have, in greater or lesser degree,
shared the same history. Any assessment of the role of these
Pleistocene events in the evolution of even these two genera is
well outside the scope of this study.

The Species Problem in Daphnia

Daphnia are small (% to 4 mm. long), planktonic Cladocera
common in lakes and small ponds. They feed primarily on small
algae which they filter from the water by means of combs on
some of their thoracic appendages, which are enclosed within a
bivalve carapace. The genus is worldwide in distribution (as
many genera of Cladocera tend to be). The ranges of some spe-
cies are rather limited, embracing but a portion of a continent,
whereas those of others are broad, extending over several conti-
nents.

A recent study (Brooks, 1956) indicates that there arc fifteen
species living in North America, out of a total of about thirty
well-characterized species. However, it is probable that there
are about fifty extant species. "While the distributions even of
these thirty known species arc not so completely known as one
might wish, a rough estimate, on the assumed basis of forty to
filty species, for the other major land areas would be: South

J. L. BROOKS 83

America, 7; Eurasia, 25; Africa, 15; Australia, 3. Although the
subsequent discussion of the systematic problems in Daphnia is
based primarily on a study of the genus as it occurs in North
America, it is almost certain that Eurasia and Africa present
problems of equal or greater magnitude, and probably of a sim-
ilar nature.

An indication of the extent of the difficulties of Daphnia sys-
tematics can be found in the varied taxonomic treatment that
the members of this genus have received. During the past seventy
years several hundred names have been proposed for Daphnia
populations, and these have been trinomial and quadrinomial
combinations, as well as the usual binomials. For example, in
Birge's widely used key to North American Cladocera (in Ward
and Whipple, 1918), six binomials, seven trinomials, and four
quadrinomials are used to indicate the fourteen entities that he
recognized. However, these polynomials, although usually based
on collections from one or a few localities, were usually consid-
ered to designate "morphological variants" which might occur
anywhere, rather than geographically limited variants. Thus the
trinomials, which are common in the literature on Daphnia
systematics, seldom referred to a geographical subspecies, in the
maimer now common in the systematics of, for example, birds.

When Wagler, about 1936, attempted a revision of the genus
he was faced with this accumulated taxonomic chaos. His solu-
tion, seemingly a reaction against the prevailing system, was to
classify all Daphnia into eleven species all indicated by bino-
mials. The simplicity of this system is in large part due to the
fact that several of these "species" are aggregates of what, in
reality, are clearly distinct species. This, however, was not an
example of the lumping as opposed to the splitting of groups of
closely related species, because many of the aggregated species
are quite dissimilar.

Two of these aggregates are Daphnia longispina and Daphnia
pulex. Since many phenotypic characteristics of each are var-
iable, both seasonally and geographically, all the species could
be put into "present" or "absent" categories on the basis of only
one relatively constant morphological feature. This was the mid-

84 FRESHWATER ANIMALS

die pecten of the postabdominal claw. Either the teeth of this
pecten were much larger than those of the proximal and distal
pectens, the "pulex" criterion, or they were the same size as
those of the other pectens, the "longispina" criterion (see Fig.
8). However, classifying species on the basis of a single charac-
ter, while often convenient, frequently, if not usually, fails to
indicate the phylogenetic relationships between these species.
It has become increasingly apparent that Wagler's simplified
classification of Daphnia does not avoid this pitfall.

We can possibly appreciate the results of these taxonomic
philosophies, which we might call the "polynomial" and the
"simplified," where they have both been used to classify the same
assemblage of forms. In 1950 Kiser applied the polynomial phi-
losophy to the North American Daphnia. His classification named
51 entities. These comprised four species, 14 "subspecies," and
33 "forms." His Daphnia pulex consisted of seven subspecies,
three of which have one, three, and five forms, respectively. His
Daphnia longispina is composed of six subspecies, four of which
have four, six, six, and eight forms, respectively. Pennak (1953)
divides the North American Daphnia into the same four species.
He considers that Daphnia pulex and Daphnia longispina are
each highly variable species, but that little is gained by naming
the more distinctive forms of each. However, the systems of
classification of both Pennak and Kiser are basically the same,
because each places primary emphasis on variation in a single
character, the relative size of the teeth in the middle comb of
the postabdominal claw. Detailed study has revealed that there
are six North American species that would be considered Daph-
nia pulex in either of the previously mentioned classifications
(D. pulex Leydig, D. middendorffiana Fischer, D. sch0dleri Sars,
D. catawba Coker, D. parvula Fordyce, D. retrocurva Forbes);
and seven that would have been considered D. longispina (D.
ambigua Scourfield, D. longiremis Sars, D. rosea Sars, D. laevis
Birge emend. Brooks, D. duhia Hcrrick emend. Brooks, D. gale-
ata mendotae Birge, D. thorata Forbes). Interestingly, D. longis-
pina O. F. Miiller, originally described in Europe, docs not occur
in North America. Examination of Figs. 8 and 9 will reveal some

J. L. BROOKS 85

of the distinctive characters of ten of the thirteen species by
which the subgenus Daphnia is represented in North America.

Our main task is to discover the reasons why the characteriza-
tion and identification of these common freshwater animals has
presented so much difficulty. But before proceeding to a con-
sideration of these sources of difficulty, we should consider some
of the basic biological peculiarities of the animals being studied.

The average life span of a Daphnia is between a few weeks
and several months, depending on the temperature. At the tem-
perature (about 20° C.) common in ponds and the upper waters
of lakes in summer, the female of most species lays her first brood
of eggs about a week after she is born. Subsequent broods of
eggs are laid at two- or three-day intervals. Although it is diffi-
cult to assess the life spans under natural conditions, probably
two or three weeks would be a reasonable estimate of the most
frequent life span at these temperatures (Brooks, 1946). These
same species grow much more slowly at the low temperatures
(4° C. or less) characteristic of temperate lakes (dimictic lakes
in the terminology of Hutchinson and Loffler, 1956) during
winter. A few of the individuals born in the fall survive until the
waters warm in the spring. Then they produce a few broods of
eggs and die. Although it is clear that the few females which do
survive have a life span of three or four months, it is difficult
to assess the average life span at these temperature under nat-
ural conditions.

A Daphnia population living in a large permanent body of
water may have as many as ten or fifteen generations per year
(see below for the basis of this estimate). At the other extreme
are the temporary-pond populations, which may have only one
or two generations per year. The species maintains itself for
the remainder of the year in the form of dormant early embryos
(so-called resting eggs) within a resistant case, the ephippium.

The populations of Daphnia consist almost entirely of females
(the exceptional presence of males is discussed later) that are
able to maintain themselves because the eggs produced under
propitious environmental conditions are diploid, and they begin
cleavage soon after they leave the oviduct. The diploidy of this

86

FRESHWATER ANIMALS

succession of individuals is apparently maintained by a failure
of the second reduction division during maturation of the oocyte
nucleus. The eggs so produced are often referred to as "partheno-
genetie" eggs, and although in some exceptional situations (Ed-
mondson, 1955) this designation loses its utility, it is sufficient
for present purposes. This derivation of one generation of females

Fig. 1. Diagrammatic representation of the modes of reproduction in
Daplinia. Usually reproduction is uniparental and all the individuals arc
females. Occasionally, however, the females produce eggs requiring fer-
tilization, and more or less simultaneously some of the parthenogenetie eggs
develop into males instead of females. Females bearing ephippia and
"sexual" eggs, and males usually appear just after population maxima. Eggs
(actually early embryos) shed in the ephippium are in a dormant state and
they can often withstand drying and freezing. If conditions improve, the
female can resume parthenogenetie egg production after she has shed an
ephippium. Males are short-lived.

from the previous by diploid parthenogenesis is represented dia-
grammatically in Fig. 1 by the two females in the right half of
the diagram. As long as the food supply remains adequate, the
female generations will succeed one another parthenogenetically
lor an apparently indefinite1 period (Banta, 1939). Shortly after
each molt, the parthenogenetie eggs are extruded into the space
between the body proper and the dorsal part of the bivalve cara-
pace. The) are prevented from falling out of this space, referred

J. L. BROOKS 87

to as the brood pouch or brood chamber, by the finger-like pro-
jections on the posterior part of the abdomen (see Fig. 1). The
number of eggs is influenced by the size and age of the female,
but for females of a given size, the level of the food supply is
the principal determinant of clutch size. While very large in-
dividuals may produce twenty, or thirty, or even up to fifty eggs
per brood, the smaller adults under the "usual" range of condi-
tions may have from none to five eggs per brood (Slobodkin,
1954; Brooks, 1946). These eggs develop directly into free-swim-
ming individuals which escape from the brood chamber during
the next molting of the parent. It should probably be noted
here that Daphnia embryos develop at the expense of the ma-
terial stored in the egg; there is no nourishment otherwise derived
from the mother while the embryos are in the brood chamber.

Although this diploid parthenogenesis is repeated as long as
nutritive conditions remain adequate, a sudden drop in food
intake can, in some of the females, provoke a change in the
reproductive physiology. This change involves the cytoplasm of
the developing oocytes as well as the meiotic events of the nu-
cleus and the structure of the developing carapace. The cyto-
plasm of the oocytes becomes dark brown and opaque, as op-
posed to the light-colored (green-gray-blue), hyaline appearance
characteristic of parthogenetic eggs. The nucleus undergoes both
reduction divisions so that the resulting egg is haploid. These
eggs must be fertilized before they will develop. The cells of
the dorsal part of the carapace change from flattened and color-
less to columnar and pigmented, and form the ephippium (the
saddle-shaped area apparent on one female in Fig. 1 ) . The ephip-
pium clasps the two fertilized eggs which have been extruded
into the brood chamber. The eggs undergo partial development
while in the brood chamber, but soon reach a quiescent stage. In
this dormant state the eggs can withstand freezing or drying. At
the time of the next molt the ephippium, clasping its eggs, is
shed with the rest of the exoskeleton. It may sink to the bottom
or be caught in the surface film. Its ability to stay in the surface
film is a consequence of the hydrofuge nature of its cuticle.

After the female has shed her old exoskeleton with its ephip-

88 FRESHWATER ANIMALS

pium, she may, if nutritive conditions continue to be poor, have
no eggs in the brood chamber (no oocytes having matured
while she was carrying the ephippium). If conditions should
have improved, however, it is possible for her to resume quickly
the production of parthenogenetic eggs.

The males necessary for the fertilization of the haploid ephip-
pial eggs develop from diploid parthenogenetic eggs. Males usu-
ally appear both in nature and in laboratory culture shortly after
a population maximum. The exact nature of the environmental
stimulus for the development of males instead of females is not
clear, nor is the mechanism by which this is accomplished (see
Banta, 1939, for a summary of investigations into this problem).
Shortly before the time that a population maximum is reached,
the amount of food per individual usually falls rapidly. The ap-
parent environmental determinant of ephippial egg formation
(diminution of food supply) thus initiates the slower process of
ephippial egg formation just before males begin to develop (at
the time of population maxima ) .

There can be little doubt of the selective value of uniparental
reproduction to Cladocera, and especially to Daphnia in which
it is especially prominent. When food is adequate, every feeding
member of the population is an egg-producer — making a most
efficient transformation of food into eggs. This enormous repro-
ductive potential, however, carries with it the danger of over-
shooting the population size and density that the food supply
can support. Possibly the chief significance of the episodes of
sexuality from the point of view of population dynamics is that
they quickly reduce both the reproductive rate and the feeding
rate of the population. The diversion of some diploid partheno-
genetic eggs into males reduces the number of egg producers
developing from oocytes already formed. The oocytes beginning
to form in other females follow a path of development which, if
fertilization ensues, produces an embryo with a dormant period
of varying duration. Thus a part of the population will exist for
a time as nonfeeding, nonreproducing individuals. These mech-
anisms permit a temporary reduction in the size of the feeding

J. L. BROOKS 89

population, so that there is less chance that death by starvation
will reduce it more drastically.

The general reproductive pattern of a population determines
its genetic system even as the pattern is genetically controlled.
The range of genetic variation among the offspring produced
by diploid parthenogenesis (of the type with which we are con-
cerned) will very nearly be coextensive with that of the parents.
Although mutation and crossing over will probably produce
some increased variation, this increase in the amount and kind
of variation can be expected to be considerably less than that
which would occur after sexual reproduction. (See Banta, 1939,
for a summary of work by Banta and co-workers on variation
following the two types of reproduction.)

The relative frequency of uniparental and biparental inheri-
tance differs according to the environment which the population
inhabits. By and large Daphnia are most successful in two types
of environments: either large, permanent bodies of water, or
small ponds, which become regularly (or irregularly) uninhabit-
able because of the disappearance of the water or the formation of
ice. A species is usually better adapted to one type or the
other, although some can maintain themselves in both. In many
lake-dwelling populations, reproduction is almost exclusively
uniparental. Sexual forms occur only at infrequent intervals and
then comprise only a very small percentage of the population.
The offspring produced sexually probably have, in most cases,
little effect upon the genetic composition of the population be-
cause of the relative unhatchability of ephippial eggs after long
periods of uniparental reproduction. Banta (1939) believes that
this reduced viability of sexually produced offspring is due to
the deleterious recessive mutations that had accumulated during
the periods when the population was reproducing by diploid
parthenogenesis. However, it is possible that a single recombinant
might be more successful than many of the other clonal genotypes
which constitute the population. Such an individual could give
rise by parthenogenesis to a clone of individuals with similar
constitutions. This clone, if truly superior in that particular

90 FRESHWATER ANIMALS

body of water, might slowly replace the other clones which com-
prise the population.

In a population of a pond-inhabiting species the ability to
produce a relatively large number of viable resting eggs (em-
bryos) would have considerable selective value. For most
species sexual reproduction is a necessary prelude to the forma-
tion of the resting stage. [It is of interest to note in passing, how-
ever, that the mechanism for producing resting eggs partheno-
genetically has become firmly established at least once, and it
has probably arisen several times (Edmondson, 1955; Brooks,
1956).] In these pond populations it is to be expected that nat-
ural selection will favor the development and retention of high
sensitivity to the environmental conditions which trigger the
processes leading to the formation of males and to the production
of haploid eggs. A rapid and extensive conversion of the popula-
tion from uniparental to biparental reproduction would thus re-
duce the demands on a dwindling food supply as well as produce
the (usually) necessary resting stages.

After this consideration of the reproductive peculiarities of
Daphnia we are now in a position to appreciate more fully the
factors that appear to be responsible for the difficulties encoun-
tered in the systematica of this genus. These are:

1. The wide range of phenotypic variation due to the sensitiv-
ity of the developmental processes to environmental conditions.

2. The frequent coexistence of flourishing populations of sev-
eral species in the same body of water.

3. The introgression of genetic material between these coexist-
ing species.

Phenotypic Variation. The phenotypic variation in Daphnia
is readily discernible as it changes the body shape. Figure 2
indicates some of the variants to be found in a pond-dwelling
species (Daphnia sch0dleri Sars). The form of any crustacean is
structurally determined by the form of its exoskeleton. The differ-
ences between A, B, and C of Fig. 2 are due to the different ex-
tent to which the exoskeleton of head, carapace, and shell spine
have grown. In the species illustrated here, as well as in all the
other members of the genus with which we are concerned in this

J. L. BROOKS

B

91

Fig. 2. Phenotypic variation in Daphnia. These three specimens of
Daphnia sch0dleri Sars exemplify the variation in body form that charac-
terizes most species of Daphnia. The principal differences are in the length
of the shell spine and the extension of the exoskeleton along the head mar-
gins. In lateral view the extent of the crest, as this lamellar extension is
called, can be judged by the distance between the head margin and such
structures as the optic vesicle and the origins of the antennal muscles.
Dorsal views of specimens A and C indicate lamellar nature of the crest.

92 FRESHWATER ANIMALS

paper, the differences in the relative growth of the exoskeleton
of these body parts are especially noticeable because they alter the
outline of the body when the animal is viewed laterally, as it
usually is. When the extreme forms (A and C) are viewed in the
dorsal aspect, the chief difference is in the extension of a very
thin lamella which develops over the midline of the head and
carapace. It is this thin lamella, composed of two closely op-
pressed sheets of exoskeleton separated by a thin film of haemo-
lymph, which is responsible for the most striking shape differ-
ences between the variant individuals. The mid-dorsal extension
of the head margin is called a "crest" when it covers a relatively
broad section of the anterior and dorsal portions of the head
margin, and a "helmet" (see Figs. 3, 4) if narrower at the base
and taller. The bases of the antennule muscles are included in
the drawings to indicate the position of the body margin (in
lateral view) if the crest or helmet were not present. As might
be expected, the relatively excessive exoskeletal growth of B and
C as compared with A is also expressed in the larger and more
numerous spinules along the dorsal line of the carapace and on
the ventral margins of the valves, and in the relatively greater
length of the shell spine. Although in the species depicted here
the most exuberant form ( C ) is larger than the most conservative
(A), this is usually not true; the individuals with the most exag-
gerated shapes often are smaller. It is therefore not a question of
increasing disproportion with increasing size.

This wide range of phenotypic variation occurs in generations
that have developed parthenogenetically. Various laboratory ex-
periments on Daphnia (see especially Brooks, 1946, 1948) indi-
cate that a clone formed from any individual picked at random
from a natural population can express the same range of pheno-
typic variation that the natural population exhibits under similar
environmental conditions. Although various genetic mechanisms
are known in diverse organisms by which recombinations are pos-
sible even when the orthodox meiotic process is absent, there is
no evidence that they play a role in the determination of the
phenotypic variation in Daphnia. Tin's does not mean to imply
that some such mechanism may not be effective in increasing

J. L. BROOKS 93

the genetic diversity during the periods when the population is
reproducing by diploid eggs.

As the determinants of this phenotypic variation are not ge-
netic, they must be environmental. The simplest hypothesis is
that the size of the crest or helmet is dependent on the environ-
mental conditions during the growth of the individual concerned.
In order to study this relationship it is desirable to have a popula-
tion showing an extreme range of phenotypic variation in a
natural situation in which the population's environment is as
homogeneous at any one time as is possible.

A fortunate combination of these desiderata occurred in Ban-
tam Lake, Connecticut. This is a large, shallow lake with three
basins, the largest and deepest of which is about a mile in diam-
eter and seven meters deep. In such a body of water, the water
is likely to be nearly the same temperature from top to bottom
because the winds are able to mix the water throughout most of
the season when there is no ice cover. As there was evidence to
indicate that temperature plays a major role as a determinant of
the phenotype, such a lake is to be preferred over one in which
the warmer upper strata are underlain in midsummer by strata
of cool or cold water. Although the behavior of most of the spe-
cies that can develop helmets is such that they remain, day and
night, in the upper strata, the possibility that a portion of any
population might have penetrated into the cold, lower waters
would complicate the analysis of the environmental effects. In
the year with which we are concerned the principal species of
Daplinia was D. retrocurva Forbes. This species exhibits prob-
ably the most extreme variation of any North American species
and one of the most striking in the genus and, indeed, in the
Cladocera.

The population of D. retrocurva in Bantam Lake during 1945
was sampled at approximately two-week intervals from April 1,
one week after the ice melted, until August 10. By the middle of
August the population had so dwindled in size that sampling
yielded only a few specimens. (However, it must be remembered
that even a population so sparse as to yield but a few specimens
to this standard sampling procedure could amount to a total size

94

FRESHWATER ANIMALS

CL-J89

HL-147

CL-755

HL- 510

Fig. 3. Sec legend with Fijf. 4.

J. L. BROOKS

95

HL=234

7th GENERATION

Figs. 3 and 4. Seasonal phenotypic variation in population of Daphnia
retrocurva Forbes in Bantam Lake, Conn., 1945. Series of camera lucida
drawings in each horizontal row depict characteristic shape of each size
group on date indicated. Head length (HL) and carpace length (CL) of
each given in microns. Diagonal lines trace development of newborn indi-
viduals of each sampling date. All have same magnification.

96 FRESHWATER ANIMALS

of something of the order of a million individuals.) The sam-
pling was done so that all depths of the lake were equally repre-
sented at all dates and so that each series of samples was quanti-
tatively equivalent. In equivalent columns of water there were
three specimens of retrocuwa on April 1. By the end of that
month there were over three thousand individuals, and the
number remained between three and five thousand until the mid-
dle of June. By the end of June this had dropped to about one
thousand, and on July 21 and August 10 there were less than
two hundred retrocuwa in the standard column of water. These
population densities are of interest in relation to the phenotypic
variation.

The appearance of the individuals that occurred at the dates
of sampling is indicated in Figs. 3 and 4. The proportions of all
undistorted specimens in a sample (up to a maximum of about
one hundred) were measured, and the figures drawn in each
horizontal row depict the average proportions (of head vs. cara-
pace) for each size grouping that occurred. By studying the
distribution of size and proportions in this temporal series of
population samples, it was possible to estimate whether or not
the majority of the newborn in the population sample represented
a new generation or merely successive broods of the same gen-
eration. For example, on April 15 the newborn individuals were
almost certainly the offspring of the over-wintering females ( seen
as the largest females on both April 1 and April 15 ) . The female
with a carapace length of 865 microns (CL = 865) represents
the most numerous size group. These are the individuals which
were the first-born of the year, and which were approaching
maturity. However, the old, large over-wintering females were
the only ones producing eggs, and these in large numbers. It
was not until the end of April that the second generation pre-
dominated among the newborn. Although such an analysis lie-
comes less clear-cut later in the season, it is reasonable to esti-
mate that there were seven generations during the four-and-a-
half-month period of observation. On this basis one can make the
even rougher estimate that there are probably about ten or
twelve generations of Daphnia retrocuwa per year in tin's lake.

J. L. BROOKS 97

The diagonal lines in Figs. 3 and 4 trace the history of the new-
born of each sample.

With this brief description of the population in mind, we can
proceed with a consideration of the correlation between the
characteristics of the phenotypic variants and the environmental
conditions under which these variants developed. Measurements
of the relative size of head and carapace of the individuals of
the population provide insight into the relative developmental
rate of the helmet as compared with that of the rest of the body,
of which carapace length is taken as a characteristic dimension.
Furthermore, the ratio of head length to carapace length in the
newborn provides an index to the rate of relative growth prior
to birth. Field and laboratory studies indicated that the relative
length of the head at birth is controlled by the temperature
during development. However, the rate of relative growth of
the helmet after birth when the animal is free-swimming is not
directly correlated with the temperature. It appears that both
high temperature (above 18-20° C. for many species) and turbu-
lence of the water are necessary conditions for the development
of helmets of the maximum size of which a population is capable
(Brooks, 1946, 1947). The details of the interrelation of these
major determinants of the relative growth rate of the helmet have
not been worked out, but a promising hypothesis is that both
factors ( temperature and turbulence ) have their effect by raising
the metabolic rate.

Although laboratory experiments to determine the importance
of turbulence in species other than Daphnia galeata have not as
yet been done, there is a considerable body of observational evi-
dence indicating that large helmets develop in those lake-dwelling
species that live in the upper waters of a lake. In the majority of
lakes in the temperate zone this will mean that these populations
during the winter months will live in water that is less than 4° C,
often near 0° C. However, in midsummer these water strata are
commonly above 18-20° C. except in very large lakes, in which,
incidentally, large helmets usually are not found. These animals
thus have to swim, feed, grow, and reproduce in a very wide
temperature range. The exuberant form which develops at high

98

FRESHWATER ANIMALS

emperatures is an almost inevitable result of the action of a
genotype which has to be able to maintain activities at slightly
above 0° C.

In the past it has usually been assumed that the ability to
produce these extreme helmets has been retained in the genetic
makeup because of selective advantage conferred by the pos-
session of such a bizarre body shape. Although this is not the
place to develop the inadequacies of this group of hypotheses,

largest helmet
commonly occurring

form characteristic
of entire range

DAPHNIA PARVULA

Fig. 5. Geographical variation in the shape of the helmets <>i maximum
size in Daphnia retrocurva Forbes. The three specimens with large1 helmets
typify the helmet shapes developed in the three regions indicated. The
specimen at the top is a reminder that the vast majority of specimens
throughout the range of the species have helmets smaller than the ones
depicted. For comparison the distribution in North America and the charac-
teristic invariant shape of D. parvula Fordyce are indicated. D. parvula is
the only species closely related to D. retrocurva.

J. L. BROOKS 99

several kinds of information should possibly be considered here
that are not consistent with the hypothesis of advantage con-
ferred by the helmet. The first such evidence is presented in Fig.
5, which indicates the range of Daphnia retrocurva and some of
the forms of the helmet. The animal at the top is a reminder that
most of the individuals in any population have a small helmet,
smaller than the one depicted. In Bantam Lake, for example, the
population size when the helmets were large was less than one-
tenth that when the helmets were unspectacular (cf. Figs. 3 and
4). The other three specimens of retrocurva indicate the geo-
graphical variation in the extreme form of the helmet. One of
the elaborate hypotheses (Woltereck, 1930) for the functional
significance of the helmet maintains that a retrocurved helmet, of
which the specimen from Lake Mendota is probably the most
extreme in the genus, has a different effect from one which is
elongated in the body axis. Many specimens of D. retrocurva
have tall helmets that are not retrocurved. The specimen drawn
as characteristic of the Canadian and West Coast areas has a
more retrocurved helmet than many in these areas. The special
pleading necessary to account for this cline and for the fact that
most of the helmeted forms are intermediate in shape between
the two extremes, with ( supposedly ) quite different effects, tends
to cast doubt on the various hypotheses of this type so far pro-
posed.

However, rather different and more telling evidence is avail-
able which makes it improbable that the ability to produce a
helmet is maintained in the genotype because of selective ad-
vantage conferred by the helmet itself. This comes from the
consideration of species such as Daphnia longiremis, which usu-
ally has no more than the slight crest indicated on the topmost
of the three specimens drawn in Fig. 6. Throughout the indicated
range in North America, and in northern Eurasia as well, longi-
remis occurs in populations of such round-headed individuals.
On rare occasions, however, populations occur in which helmeted
individuals are common. The places where such populations have
been found in North America are indicated on the map. Two
types of helmets occur, and the differences are probably geneti-

100

FRESHWATER ANIMALS

Fig. 6. Phenotypic variation in Daphnia longiremis Sars. D. longiremis
is distributed throughout northern North America and northern Eurasia.
Throughout this range most populations are composed of individuals re-
sembling the topmost specimen. However, a few scattered North Ameri-
can populations are known in which individuals with large helmets have
appeared. These helmets are of the two forms indicated. The environ-
mental circumstances in which the two populations (one in Wisconsin, one
in Saskatchewan) produced retrocurved helmets arc indicated diagram-
matically in Fig. 7.

cally determined. These extreme helmets occur in widely scat-
tered localities, yet the broad type from northern Alaska is very
like the broad type from Wisconsin. Furthermore, we know that
the population in Saskatchewan, which during one particular
season contained individuals with retrocurved helmets, in other
years consisted of individuals with not! dug but a small crest

J. L. BROOKS

101

<_> CM"

<a

at _« ^2 rs or:

o ,

&5

e y

o

C r-l

a; £

71 TJ 'r uj
<D X£ > CS

tc

c

3

' 3

C/5 f- r— » — ^ ^

VJ _J ^_J ^T*.

S-l

O

O 4-<

2-

0 D

<D

>

3 2 U

s

5

-a r5 d

S
-

?> — bO h,

"Sb «« i .2 £ .a ^,

o o _, > cS

"t: C <~*

CD o

2" CO! £ ^

<D

c
o

■c "~ « "£ i *

3 r-4. > rt l_l

o

SC/3

0) <D

o

3 lis g

SO) -M

„ 3 iu X c\|

0 J2 T3 — ■ CO
4H-^ru ^ 05

O 03 3 CD 3 r— I

CD ^

-3

Q

\ ^

o

p

fz*j=~^

5 'P

K -t->

CD O

M

5 i< u v mh kt
C « n* j- ,r: fl >

rt C at 71 nr *j

w

&H

0

_. W3 O

E 3

1> ^5 f^s rt

r> ^S 3 -Q

102 FRESHWATER ANIMALS

(Brooks, 1956). It is difficult to believe that a character mani-
fest in such sporadic fashion could be of such adaptive sig-
nificance that the ability to produce this character would be
carried latent in the germplasm of a species with an extensive
distribution over two continents.

We fortunately know something of the physical properties of
the water in two of the lakes in which populations with retro-
curved helmets occurred. We are thus in a position to compare
the environmental characteristics at the time of helmet develop-
ment with the more usual conditions under which helmets are
not produced. This comparison is made in diagrammatic fashion
in Fig. 7. Under a specimen representing the helmet character-
istic of each population is depicted half of a vertical section
through the lake in which the population lives, and a graph
showing the temperature at different depths. The section at the
left is characteristic of the majority of lakes of the temperate
zone in midsummer. The graph indicates that an upper stratum
of water is warm, about 20° C; below this there is a region
where the temperature decreases rapidly with increasing depth.
Most of the bottom waters are slightly above the temperature of
maximum density (4° C). The vertical hatching on the lower
part of the section indicates that D. longiremis usually lives in the
cool bottom waters. D. retrocurva, on the other hand, would live
in the warm, turbulent upper layers. Such populations, living in
cool waters in summer as well as in winter, show little seasonal
variation, and at most develop the small crest indicated.

The center section represents the situation in Muskellunge
Lake, Wisconsin, on August 26, 1931. This was reported, but
differently interpreted, by Woltereck (1932). Here Daphnia
longiremis with tall, slightly retrocurved helmets occurred, and
as we might expect, the population was living in warmer (and
more turbulent) water than the species usually does. The bot-
tom waters were devoid of oxygen, and the population could
maintain itself only in the middle strata of water. The diagram
at the right represents the physical conditions in Lake Waskesiu,
Saskatchewan, during midsummer of 1932 and the helmeted
longiremis that developed therein. The thermal conditions in

J. L. BROOKS 103

Waskesiu in this and other years, has been extensively studied
by Rawson ( 1936 ) . The diagrammatic representation is based
upon his data and the specimens examined were collected by
him at this time. Waskesiu is a large, relatively shallow lake
which usually becomes stratified during midsummer, although
the bottom temperatures are usually higher than those consid-
ered typical in the diagram at the left. In 1932, however, excep-
tionally strong winds caused the stratification to break down
and mixed the surface waters to greater and greater depths, until
finally the water temperature at all depths was very nearly the
same (around 18° C). Under these conditions longiremis could
live throughout the entire depth of the lake. Specimens were
found living in the upper five meters! They bore retrocurved
helmets, as drawn.

Examination of these two instances of helmeted longiremis
indicates that this extreme body shape was evoked when the
populations were forced to live in warm (and turbulent) water.
It also makes it seem unlikely that the genetic ability to develop
this phenotype under rare and abnormal circumstances is pre-
served in the genotype of a widespread population because of
any advantage conferred by this extreme phenotype during these
exceptional circumstances.

It seems rather more likely that the variations in body shape
in D. longiremis, as in the rest of the genus, are a consequence
of maintaining a genetic constitution that will allow any individ-
ual to develop and function under any of the large range of
environmental conditions which the species can tolerate. The
morphogenetic processes must permit development at tempera-
tures slightly above 0° C. Morphogenesis controlled by the same
genotype must also produce an effective organism at a tempera-
ture twenty degrees higher. That there are differences in the
shape of the organisms produced by these same processes func-
tioning at such different temperatures is not surprising.

Although the morphological changes in Daphnia are especially
pronounced and have considerably complicated the problem of
species determination, such extreme variation occurs in many
freshwater organisms. From our investigation of the nature of

104 FRESHWATER ANIMALS

this phenotypic variation, we should expect to find it whenever
short-lived animals have to develop and remain active under a
wide range of thermal conditions. The plankters of freshwater
lakes of the temperate zone might be expected to exhibit this
variation to a pronounced degree, and they do. Short-lived plank-
tonic Cladocera (Bosmina, Daphnia), rotifers, and copepods, all
show striking morphologic variation, whereas related species and
genera which, for example, are benthic or longer-lived show
much less pronounced variation.

Species Associations. The confusion which the wide intra-
specific range of body shapes causes for the investigator seeking
to delimit the species is confounded by the frequent coexistence
of large populations of several species of Daphnia in the same
body of water. Although the associations in lakes tend to be
larger, even small, temporary ponds commonly have two species
living at the same time. In addition to the fact that several
species may occur in a lake at the same time, different species
may comprise the dominant associations in different years. Vari-
ous evidence indicates that such shifts are possible because a
species may exist in a lake in a population so small as to be
undetected except with very extensive sampling and careful
examination of the material collected (Brooks, MS). However,
under favorable conditions this population may attain sufficient
size to be readily collected.

Figure 8 indicates the array of species that can be found living
in New England lakes. In some parts of North America even
larger arrays of species can be found, but this will suffice as an
illustration of the problem. Collections in the upper waters of any
New England lake may disclose populations of any one or more
of the first five species living in the upper waters. In the bottom
waters in midsummer usually cither D. ambigua or D. longiremis
will be found, with D. ambigua more common in the lakes of
southern New England. There is no indication of preferential
associations. In a survey of 107 lakes in southern and central
Maine, it was found that the seven species occurred with tlio
approximate percentages indicated in Fig. 8 (Brooks, 1956, MS).
The most frequent associations were between the species most

J. L. BROOKS 105

successful in the area; that is, chance appeared largely to deter-
mine the species to be found living together. The largest number
of species yet found simultaneously in a lake in North America
(or the world) are the six that lived in Aziscoos Lake, Maine.
The one species, parvula, which was not found in Aziscoos at
that time was one of the least common in the Maine lakes studied.

CATAWPA

PARVULA RETROCURVA

OUBIA

GALEATA

MENOOTAE

AMBIGUA

LONGIREMIS

ABREPTOR,
FEMALE

ABREPTOR,
MALE

ANTENNULE,
MALE

POSTABDOMINAL
CLAW

FREQUENCY IN 107
MAINE LAKES

LARGEST ASSOCIATION
(AZISCOOS LAKE)

UPPER WATERS

BOTTOM WATERS

90

10

40

33

33

10

25

Fig. 8. Species of Daphnia commonly occurring in New England lakes.
Most lakes have more than one species common at any time. The largest
association found was in Aziscoos Lake, Maine, which had the six species
indicated. Characteristic morphological and ecological traits of each species
are indicated.

Although the coexistence of this many closely related species
raises many interesting theoretical questions (which will be
discussed elsewhere), our immediate interest is in the effect on
the problems of species recognition. The confusion which the
presence of as many as four or six species in a lake presents for

106 FRESHWATER ANIMALS

the unwary investigator is great, because each population may
exhibit considerable variation. Furthermore, in the late spring
and early summer, when the Daphnia are likely to be most
numerous, their shapes tend to be least distinctive (cf. Figs. 3,

4).

There is, however, a further possible complication to species
recognition which may be the result of the continued coexistence
of several species. This is the oft-noted tendency for two popula-
tions of different species inhabiting a lake to show some resem-
blance to each other (see, for example, some of Woltereck's pa-
pers; Kiser, 1950). Thus, each of the two common species in
Moosehead Lake, Maine, Daphnia dubia and Daphnia galeata
mendotae, tend to differ slightly in body shape from their char-
acteristic forms (cf. Fig. 8) — the helmet of galeata may be more
slender, that of dubia broader usual. Other structures may show
some intermediacy, yet the specific identity of each is not in
doubt. This phenomenon has been interpreted by Woltereck as
being a direct effect of the peculiar features of the water of the
lake in question on the species living in it. A more reasonable
hypothesis is that the genes of the two species are somehow ad-
mixed in the individuals of such populations. One possibility is
that these populations have developed from a chance hybrid
between the two species, at one of the infrequent periods of
sexual reproduction. A hybrid individual genetically very like
one of its parents might be able to compete successfully with its
parents. Of course, only one viable female hybrid is necessary
to start a clonal population. If its genotype were superior in the
peculiar environment of that lake, especially at some critical time,
it could conceivably replace the parental species most like itself.

Hybridization. Tims we are introduced to the third aspect of
the species problem in Daphnia. This is the occurrence of popu-
lations partaking of the characteristics of two species. The follow-
ing facts about these intermediate populations are noteworthy.

1. They occur more commonly in pond-dwelling species than
in lake species. This is consistent with the fact that periods of
sexual reproduction are more frequent in the pond species and

J. L. BROOKS 107

affect a greater percentage of the population. Thus the chances
for forming hybrids is considerably greater.

2. The hybrid populations occur in regions where both puta-
tive parent species live.

3. The hybrids themselves are usually chimaera-like in that
in some features they are quite like one parent, in some quite like
the other. In this they are very like the only experimentally pro-
duced hybrids between species of Daphnia, those obtained from

MIDDENOORFIANA

PULEX

SCH0DLERI

ABREPTOR,
FEMALE

ABREPTOR,
MALE

ANTENNULE,
MALE

Fig. 9. Three similar species of pond-dwelling Daphnia common in
northern and western North America. Although the postabdominal claws
(on abreptor) of females are similar (all of pulex type), other morpho-
logical differences in female are apparent (head shape, length of shell spine,
shape of carapace). Structure of males indicates distinctiveness of each
species more clearly than that of female. Male of middendorffiana very rare.
Many Daphnia populations in regions where these species coexist appear to
be hybrids between various pairs of these species. (See Fig. 10.)

108

FRESHWATER ANIMALS

MIDDENDORFIANA

Fig. 10. Structure and distribution of Daphnia middendorffiana Fischer
and D. pulex Leydig and of populations that appear to be hybrids between
them. In a broad belt between the northern region where middendorffiana is
one of the few species of Daphnia and the southern region where pulex is
common, neither of these species is found in its "pure" form. In this belt
there are many populations, the individuals of which have some character-
istics of both middendorffiana and pulex.

J. L. BROOKS 109

a cross between European Daphnia pulex and Daphnia obtusa
performed by Agar and reported by Scorn field (1940).

The species most frequently involved in these presumably
hybrid populations, called, for the sake of brevity, simply "hy-
brid," are three pond-dwelling species of the northern and west-
ern parts of North America. These are Daphnia middendorffiana
Fischer, Daphnia pulex Leydig, and Daphnia sch0dleri Sars. Al-
though each of these is variable, the females drawn in Fig. 9 can
be considered characteristic. The postabdominal claw in each
is similar and variable, so that it is not difficult to understand
that all three have in the past usually been assigned to Daphnia
pulex when the possession of such a claw was taken as the sole
species criterion. The abreptor and antennule of the male
are distinct in all three, and when males occur, the populations
can be readily determined. Both pulex and sch0dleri range widely
over the western parts of this continent (pulex actually over most
of it), and hybrid populations between them occur scattered
throughout the large area of overlap. However, the best illustra-
tion of such hybrid populations concerns those between pulex
and middendoiffiiana. As can be seen in Fig. 10, Daphnia mid-
dendorffiana is most common in the very northern parts of North
America, and is one of the chief zooplankters in the numerous
tundra ponds and lakes of that region. Daphnia pulex extends
fairly far to the north but apparently cannot compete success-
fully with middendorffiana in the extreme north. Between the
areas in which these two species are common there is a broad
belt, from central Alaska eastward at least to Hudson Bay, where
most of the pulex-like Daphnia exhibit a combination of pulex
and middendorffiana characteristics. These have been labeled
"pulex with middendorffiana introgressed" in Fig. 10, merely to
indicate a guess as to their probable nature and origin. Although
it is not indicated on Fig. 10, Daphnia middendorffiana occurs in
scattered populations south as far as central California and
Wyoming. Scattered populations in this region of western United
States resemble the hybrids common to the region indicated
on the map.

This complex of middendorffiana, pulex, and sch0dleri, the

110 FRESHWATER ANIMALS

females of which are sometimes difficult to distinguish, together
with scattered populations combining characteristics of any two
in somewhat chimaera-like fashion, has posed a considerable
systematic problem, probably the most difficult one encountered
in the North American assemblage.

Populations intermediate between other species also occur, but
they are largely confined to regions such as the Pacific North-
west where northern and southern species assemblages have
come into contact (Brooks, 1956). The distribution and proper-
ties of these populations bear the same resemblance to the puta-
tive parents as did those in the case just discussed.

On the basis of this kind of evidence it seems justified to
propose that introgressive hybridization is a factor that has com-
plicated the systematics of Daphnia. As yet, however, neither
experimental crosses nor analyses of the sort that Anderson
(1949) has developed for plants have been done with these
Daphnia populations. Because hybridization has been viewed
with much skepticism, often justified, comparable data on its
importance in the systematics of most other groups of animals
are wanting. Hubbs (1955) has provided an excellent summary
of hybridization in North American fishes, which are the most
adequately investigated group of animals from that point of
view. We shall again mention the possibility of introgressive
hybridization in the discussion of the species problem in Core-
gonus.

The Species Problem in Coregonus

Coregonus, the genus of the true whitefishes, is one of the three
genera that comprise the whitefishes. These genera are con-
sidered by Berg (1940) to constitute one of the two subfamilies
of the Salmonidae. Members of the genus Coregonus live in the
rivers (many species tolerate brackish water) and lakes of north-
ern Eurasia and North America. These whitefish are exceedingly
numerous in the fresh waters of the very northern parts of these
two continents.

The classification of these fish has aroused interest, possibly
because some provide excellent food. Aside from the inherent

J. L. BROOKS 111

difficulties in their classification, considered below, it has been
difficult to grasp the problem in its entirety because the bulk of
the whitefish populations live north of the regions heavily popu-
lated by man. The numerous and diverse attempts following the
standard practices of ichthyology have resulted in cumbersome
systems employing many tri- and quadrinomials (see, for ex-
ample, Berg, 1948, and Hubbs and Lagler, 1947). The chief
source of difficulty has been the diversity of the morphological
and physiological characteristics of these Coregonus populations.
Not only does each major drainage system have its peculiar
forms, but also several forms may occur in any one lake. It was
not until the investigations by Gunnar Svardson into the biology
of the populations living in various Swedish waters that we have
had any insight into the nature of the systematic relationships
within this genus. The remainder of the discussion will be con-
cerned with the factors contributing to the difficulty of recog-
nizing species of Coregonus, as these factors have been uncov-
ered by Svardson in a series of papers (1949-1953).

Unfortunately Svardson has not yet attempted to name the
species that his investigations have disclosed, nor has he at-
tempted to unravel the chaotic taxonomy of the genus. His
general feeling, however, is that the genus comprises two major
species groups (superspecies in the sense of Mayr). Svardson
calls these two superspecies the lavaretus group (after Corego-
nus lavaretns Linn.) and the albula group (after C. albula Linn.).
Not only are the species within each group nearly indistinguish-
able but even the two superspecies are also very similar. The
best character distinguishing the superspecies apparently is that
the lower jaw in species of the albula group is always longer than
the upper. The lavaretus group, on which all of Svardson's work
was done, has its greatest number of species in Europe and
western Russia, where the albula group is represented by a much
smaller number of species. In eastern Asia and in North America,
however, it appears that the albula group has formed many
species, whereas there are relatively few lavaret us-\ike species.
Walters ( 1955 ) pointed out that the glaciated portions of Eurasia
and North America have far more Coregonus species than do the

112 FRESHWATER ANIMALS

unglaciated portions. However, too little is known of the species
and their distribution to permit meaningful speculation about
the historical zoogeography of this genus.

Svardson has found that the chief source of difficulty in the
recognition of Coregonus species is the extraordinary phenotypic
variability of each genetic stock (species). The coexistence of
several species in a lake (see Fig. HA), a circumstance which
often complicates the problem for the systematise has actually
provided Svardson with the major clues to the identity of the
species. In these ways both Coregonus and Daphnia present
somewhat similar problems and, as in Daphnia, there are evi-
dence of the formation of hybrids and some resemblances be-
tween the species of major drainage systems which suggest
introgressive hybridization.

Phenotypic Variation. The primary reason for the failure of
orthodox taxonomic methods in Coregonus systematics has been
that the phenotypic characteristics are not constant for, and
therefore not indicative of, a given genotype. Quite the contrary
is true; the phenotype is highly variable, depending upon the
environmental conditions under which the genetic stock devel-
oped. These plastic characteristics include meristic ones (number
of vertebrae, fin rays, scales), size and proportions, and also
physiological ones ( growth rate, onset of sexual maturity, length
of life). The allometric growth relationships are complex because
the proportions within a stock differ not only in fish of different
size (length) but also in fish of the same length but of different
age (Svardson, 1949, 1950). The number of gill rakers is the only
phenotypic characteristic that is apparently unaffected by the
environmental circumstances during development. This is the
only phenotypic manifestation indicative of the genotype, and,
therefore, of taxonomic utility. However, the number of gill rakers
is under close selective control and may vary somewhat between
allopatric populations of a species. Svardson was able to modify
the g\\\ raker count in a population introduced into a formerly
Coregonus-hee lake by imposing an artificial selection for this
I actor at the time of introduction. It is little wonder that the

B

Fig. 11. Mature whitefish of various species of the Coregonns lavaretus
species group inhabiting Swedish lakes. A, spawners of two different species
inhabiting the same lake. B and C indicate the phenotypic differences be-
tween populations of the same genotype that have developed in different
environments. Lower drawing in B indicates phenotype of spawning fish
of a certain population. When fry from this population are transplanted to
a different lake they develop into mature fish resembling the upper speci-
men. Upper figure in C represents a mature fish of a different population.
When mature fish of this population were transplanted to another lake, they
grew to size of the lower specimen after two summers in the different en-
vironment. All figures from Svardson (1949); B after Olofsson (1934); C
after Runnstrom (1944).

113

114 FRESHWATER ANIMALS

systematics of this large number of very similar species has been
confused!

Svardson (1952) was able to demonstrate the environmentally
determined plasticity of all these phenotypic characteristics by
transplanting stocks with well-known characteristics to lakes that
lacked Coregonus populations. Different and unrelated aspects of
the environment may affect the phenotypic expression of certain
of the above-noted plastic features. For example, the number of
scales along the body, a feature that has frequently been used
to distinguish subspecies and even species, was shown by Svard-
son to be affected both by temperature during development and
by the state of nutrition of the female parent. The eggs produced
by well-fed populations are larger than those of poorly fed ones,
and more scales are initially laid down in the larger fry that
develop from the larger eggs.

The extraordinary plasticity of these fish can be seen in com-
parisons of specimens from such transplanted stocks with cor-
responding ones from the original populations. In Fig. 11B the
lower specimen is the normal size of a spawning whitefish from
Lake Lomsjon. When this stock was transplanted to Lake Oxvatt-
net, the normal size at spawning was that of the upper specimen.
Some stocks retain this plasticity into maturity, as is demon-
strated in Figure 11C. Spawning whitefish of Lake Naekten attain
the size indicated in the upper specimen. When adult fish from
this stock were transplanted to Lake Algsjo, they had, in two
summers, achieved the size of the lower specimen. Further dif-
ferences in growth, size, and length of life between populations
of the same species in different lakes can be seen by comparing
the graphs of Fig. 12.

The studies on the environmental plasticity of these fish are
summarized by Svardson (1953, p. 165) with these words:

. . . species of whitefish, which as far as can be judged must be the
same form from lake to lake, can have different forms in one lake or
another. The same species can have a long life or die young, grov
rapidly or have extremely poor growth, live in large or small popula-
tions, spawn in running or still water at different depths and at ex-
tremely varying times. Some variation is already found in the habits ol

J. L. BROOKS 115

the species in one and the same lake, but when the same species is
compared from different localities the variation is very considerable.

The intra-lacustrine variation mentioned in the last sentence of
the quotation refers primarily to differences between fish of dif-
ferent year classes.

Species Associations. Many lakes are known to have several
populations of Coregonus in them, and Svardson records two
lakes each of which has five species of the lavaretus group. The
populations that we now observe coexisting in some lakes are
known to have done so for long periods of time ( of the order of
a century or two) because fisheries records of this age mention
the local names for the different kinds of whitefish. It seems rea-
sonable to assume that many, if not most, of the Coregonus spe-
cies associations are of equal, and probably greater, stability.
However, the stability of such associations can be destroyed, as
has been indicated (Svardson, 1949) by an intriguing case, the
details of which are but sketchily known. In this instance a Lap-
land lake was known to have had two species of Coregonus which
maintained their integrity by having temporally separated spawn-
ing periods. When a third stock was introduced (apparently in
an attempt to improve the fishing), the result was the creation
of the large, panmictic, highly variable population that exists
there today. It is supposed that the introduced stock had a
spawning season that overlapped the hitherto separated seasons
of the other two.

In his 1953 paper Svardson presents data concerning the
species of Coregonus found associated in three lakes of as many
different river systems in central Sweden. These rivers all flow
southeastward into the Baltic Sea from the mountains that run
down the middle of the Scandinavian peninsula. From south to
north the lakes are Lake Idsjon, Lake Storsjon, and the Hornavan-
Uddjaur series of lakes. Figure 12 indicates the species of Core-
gonus in each lake and some of the characteristics of each popula-
tion. Svardson considers the populations indicated in these
graphs to be species, but he did not assign them scientific names.
Their common names are therefore used. As indicated in the
previous section, the phenotypic characteristics of different year

116

FRESHWATER ANIMALS

600

^ LARGE WHITEFISH
^^ (21)

500 _

400_

300 _

/ObLUE WHITEFISH

y> (36)

200 _

/

I00_

//.......RIVER WHITEFISH LAKE

,f' (26) IDSJON

0

r 2 4 6 8 10 12 14 16 18 20 22
1 1 1 1 1 1 1 1 1 1 1

600 _

^ LARGE WF. (22)

5 500 _

5

s'

5 400 _

./

H 300-
O

2

RIVER WF..^/
(28) .,y/

-1 200 _j

,-'^/> PLANKTON WF. (38)
/>j£-BHJE WFS* (34.3S)

<

h- 100 —

o

#f LAKE
// ST0RSJ0N

0

' 2 4 6 8 10 12 14 16 18 20 22 24
1 1 1 1 1 1 1 1 1 1 1 1

500

^- LARGE WF. (20)

400 _

/V'ASP" (45)

300 _

200 _
100-

/ BLUE WFS* (34, 39)
/ *^- — PLANKTON WF (38)

h'' LAKE
A HORNAVAN

0

/ 2 4 6 8 10 12 14

r i i i i i i i

AGE IN YEARS

Fig. 12.

J. L. BROOKS 117

classes of a population are somewhat different, and the data
given here can be considered normal for these stocks. It can be
seen from Fig. 12 that Lake Idsjon has three species of Core-
gonus, whereas each of the other two lakes has five, although
the graphs for the last two present the data for only four species
each. The reason for this discrepancy is that subsequent to the
publication of his 1953 paper Svardson discovered (reported in
Runnstrom, 1953) that the "blue whitefish" of several Swedish
lakes were in reality two closely similar species. Lakes Storsjon
and Hornavan are two of the lakes discovered to have two species
of "blue whitefish." Populations of these two can be distinguished
by their gill raker counts; one has an average of 33-35 (considered
34 in Fig. 12) and the other 38-39 (considered 39).

For populations of closely similar biparental organisms to
maintain themselves within the same ecosystem it is necessary
( 1 ) that each be able to find sufficient food in the face of actual
or potential competition from the others, and (2) that each be
able to maintain its genetic integrity in the face of possible admix-
ture with the other population. As the Coregonus populations
of these lakes have been able to maintain themselves, it is clear
that they must be able to meet the above conditions. Svardson's
conclusion that each such population belongs to a species differ-
ent from the others with which it continues to coexist is entirely
reasonable.

The manner in which these populations have been able to meet
the two conditions for continued coexistence is not known in
detail, although we know more of the breeding habits than of
the food requirements. We can appreciate the ways in which

Fig. 12. Characteristics of sympatric Coregonus populations of three
Swedish lakes. Common names are used because scientific ones for these
populations have not been established. The only certain species criterion
is the number of gill rakers. The average number of rakers on the first left
gill arch of each population is given in parentheses. The two numbers and
the asterisk after the blue whitefish of Storsjon and Hornavan indicate that
two very similar species of "blue whitefish" occur in these lakes. Com-
parisons of characteristics of same species in different lakes indicate lability
of such phenotypic characteristics as size at given age and longevity. Slight
differences in average number of gill rakers (in part may be due to sam-
pling errors) are in part, apparently, genetically determined.

118 FRESHWATER ANIMALS

the coexisting species maintain their genetic integrity by con-
sidering the spawning behavior of the species of the Hornavan
lake system (Svardson, 1953). The pertinent information is
summarized in Table I. These differences in time of spawning
and types of spawning ground apparently suffice to permit the
populations to reproduce with relatively little intermixing.

However, the reproductive isolation between the populations
is by no means complete. Ripe "asp" have been taken on the
spawning grounds of the great whitefish in December, even
though their normal breeding time is a month and a half earlier.

Table I. Spawning Behavior of Coregonus Species of Hornavan Lake System

Species Spawning

(local name) period Spawning place

"Asp" November In running water of rivers between lakes

and in mouths of influent rivers

Large whitefish 1 )ecember Shallow water above stony bottom; widely

distributed in lakes

Blue whitefish" December Deeper water over stony bottom; widely

(two species) January distributed except in northern part of

Lake Hornavan

Plankton whitefish February Restricted in feeding and spawning to deep

waters of Lake Hornavan; spawns on
bottom at relatively great depth (30
100 m.)

" The spawning habits of the two species known by this name have not been
different iated.

The local fishermen have long recognized a special kind of
whitefish that they call "albask," which Svardson thinks are
most likely hybrids between the "asp" and the great whitefish.
Hybridization between late spawners among the bine whitefish
and early plankton whitefish in areas where their spawning
grounds meet is to be expected, and Svardson finds some evi-
dence that would seem to substantiate this possibility.

Intermediates and Hybridization. Individuals that appear in-
termediate between various of the species (like the "albask just
mentioned I occur with a low frequency. However, only a portion
of these can be established as hybrids; the only hybrids readily

J. L. BROOKS 119

detectible are those between species with widely divergent gill-
raker counts. Many of these aberrant individuals owe their inter-
mediate appearance to the phenomenon of "shoal-trapping." This
term refers to the circumstance in which an individual becomes
separated from the shoal of similar-sized fish of its own species,
with which it would normally swim, and joins a shoal of a differ-
ent species. A shoal-trapped fish is thus exposed to the range of
environmental conditions proper to the alien species comprising
the shoal. Because of the extreme sensitivity of the develop-
mental processes in the whitefish to environmental conditions,
the shoal-trapped individual develops a phenotype determined
genetically by one species, but so modified by the environment
characteristic of the alien species that it strongly resembles the
species that it has inadvertently adopted. It is only when the
species involved have very different numbers of gillrakers that
the genetic stock of the shoal-trapped individual can be identi-
fied. Furthermore, it is only when the trapping has occurred
fairly late in the life history that the foreign fish is sufficiently
different from its shoal-mates to attract the attention of an ob-
server. If the trapping occurs in the fry stage, the phenotype of
the trapped fish will be little different from that of its adopted
species, so little different that it will very likely escape the atten-
tion even of a trained observer.

Not only are shoal-trapped fish difficult to distinguish from
hybrids, but they are likely to increase the production of hybrids
because such a trapped fish is likely to spawn along with the rest
of the shoal. The resultant mixture of gametes is likely to produce
hybrids, if we can judge from the high yield in cross fertilizations
between species of the Coregonus hvaretus group. Although Fi
hybrids are readily produced in artificial crosses, it has not yet
been established whether F2 hybrids and backcrosses are pro-
duced with proportionate facility. The detected natural hybrids
are rare and could be almost entirely of the first generation. The
progeny from a backcross would probably be extremely difficult
to detect in nature. However, if they did occur, even with a very
low frequency, they might account for the introgression of
genetic material from one species to another. Svardson remarks

FRESHWATER ANIMALS

that the various species of a river system often exhibit similari-
es which, aside from mutual introgressions, are difficult to
understand. Further investigations on hybridization and intro-
gression apparently are in progress at Drottningholm.

Summary

1. The systematics of two "difficult" genera of freshwater ani-
mals, the crustacean Daphnia and the whitefish Coregonus, have
been examined in an attempt to discover any common sources
of the species problem in these groups that might be considered
as peculiar, at least in their intensity, to freshwater animals.

2. The extraordinarily wide range of phenotypic variants that
may develop from the same genotype emerges as the chief
source of the species problem in both Daphnia and Coregonus.
For the systematist this means that the phenotypic character-
istics that are taken as indicators of the genotype, i.e., the species,
must be chosen with great care. From an evolutionary point of
view, this wide range of phenotypic variation is a reflection of
the ability of the genotype to prosecute successful development
over the wide range of environmental conditions occurring in
the fresh waters that these animals inhabit. It is probably not
fortuitous that animals with this ability are among the more
successful inhabitants of these inland waters (both Daphnia and
Coregonus and the respectively related groups, Bosmina and the
chars, for example). For Daphnia the extreme environmental
conditions succeed each other cyclically, with the seasons, so
that the entire range of phenotypic diversity is seen in the an-
nual succession of generations of these short-lived animals. In
Coregonus the extraordinary phenotypic diversity, while in part
expressed in different year classes of the same population, is
principally expressed by allopatric populations. The conditions
of life are sufficiently different from one lake to another so that
nearly all the phenotypic characteristics of a species are different
from lake to lake.

3. A second source of difficulty in the recognition of species in
both genera has been that in many bodies of fresh water several
highly variable species may coexist. In North America as many

J. L. BROOKS 121

as six species of Daphnia are known to occur together in the
same lake, while five species of Coregonus of the same species
group ( superspecies ) have been reported from Sweden (the area
studied by Svardson).

4. The coexistence of several closely related species in the
same body of water provides the possibility for hybrid formation
and introgression. There is evidence for both of these phenomena
in both Daphnia and Coregonus, evidence consistent with the
different reproductive biology of the two groups. The hybrid
individuals in Coregonus and hybrid populations in Daphnia are
an additional source of taxonomic difficulty. In neither group has
the phenomenon of introgression been carefully investigated,
although in both groups it stands as the most likely explanation
for certain general similarities between the species of a given
region.

5. These factors, extreme phenotypic variation, coexistence of
several closely related species, and occasional hybridization ( and
introgression?), although by no means peculiar to freshwater
organisms, do appear to be largely responsible for the species
problem in two very different groups of freshwater animals.
Whether the sources of difficulty common to the systematics of
Daphnia and Coregonus will prove to be common to other "diffi-
cult" groups of freshwater animals will only be revealed by
further investigation.

REFERENCES

Anderson, E. 1949. Introgressive Hybridization. John Wiley & Sons,

New York, N.Y.
Banta, A. M. 1939. Studies on the physiology, genetics, and evolution

of some Cladocera. Paper No. 39. Dept. of Genetics. Carnegie Inst.

Wash. Publ. No. 513.
Berg, L. S. 1940. Classification of fishes, both recent and fossil. [In

Russian.] Trav. inst. zool. acad. sci. U.R.S.S., 5. No. 2. Reproduced,

with English translation by Edwards Bros, Inc., Ann Arbor, Mich.,

1947.

122 FRESHWATER ANIMALS

Berg, L. S. 1948. The freshwater fishes of U.S.S.R. and the surround-
ing countries. [In Russian.] Part I. Moscow.

Rirge, E. A. 1918. The water fleas (Cladocera). In Freshwater Biol-
ogy. John Wiley & Sons, New York, N.Y.

Brooks, J. L. 1946. Cyclomorphosis in Daphnia. I. An analysis of
D. retrocurva and D. galeata. Ecol. Monographs, 16 (4), 410-47.

Brooks, J. L. 1947. Turbulence as an environmental determinant of
relative growth in Daphnia. Proc. Natl. Acad. Sci. U.S., 33 (5),
141-48.

Brooks, J. L. 1956. The systematica of North American Daphnia. Mem.
Conn. Acad. Arts Sci.

Brooks, J. L. (MS) The vertical distribution and associations of Daph-
nia species in New England lakes.

Edmondson, W. T. 1955. The seasonal life history of Daphnia in an
arctic lake. Ecology, 38, 439-55.

Hubbs, C. L. 1955. Hybridization between fish species in nature. Sys-
tematic Zool, 4(1), 1-20.

Hubbs, C. L., and K. F. Lagler. 1947. Fishes of the Great Lakes Re-
gion. The Cranbrook Press, Mich.

Hutchinson, G. E., and H. Loffler. 1956. The thermal classification of
lakes. Proc. Natl. Acad. Sci. U.S., 42 (2), S4-S6.

k'iser, R. W. ,L950. A revision of the North American species of the
Cladoceran genus Daphnia. Edwards Bros., Ann Arbor, Mich.

Pennak, R. W. 1953. Fresh-water Invertebrates of the United States.
Ronald Press, New York, N.Y.

Rawson, D. S. 1936. Physical and chemical studies in lakes of the
Prince Albert Park, Saskatchewan. /. Biol. Board Can., 2(3), 227-84.

Scourfield, D. j. 1942. The "Pulex" forms of Daphnia and their sep-
aration into two distinct series, represented by D. pulex (De Geer)
and D. ohlusa kurz. Ami. and Mag. Nat. Hist., 9 ( ser. 11), No. 51,
202-18.

Slobodkin, L. B. 1954. Population dynamics in Daphnia obtusa. Kurz.
Ecol. Monographs, 24 (1), 69-88.

irdson, G. 1949. The coregonid problem. I. Some general aspects of
the problem. Rc})t. Inst. Freshwater Research Drottningholm, 29,
9-101.

Svardson, G. 1950. The coregonid problem. II. Morphology of two
coregonid species in different environments. Rept. Inst. Freshwater
Research Drottningholm, 31, 151-62.

Svardson, G. 1951. The coregonid problem. III. Whitefish from the

J. L. BROOKS 123

Baltic successfully introduced into fresh waters in north of Sweden.
Rcpt. Inst. Freshwater Research Drottningholm, 32, 79-25.

Svardson, G. 1952. The coregonid problem. IV. The significance of
scales and gillrakers. Rept. Inst. Freshwater Research Drottning-
holm, 33, 204-32.

Svardson, G. 1953. The coregonid problem. V. Sympatric whitefish
species of the Lakes Idsjon, Storsjon and Hornavan. Rept. Inst.
Freshwater Research Drottningholm, 34, 141-66.

Wagler, E. 1936. Die Systematik und geographische Verbreitung des
Genus Daphnia O. F. Muller mit besonderer. Beriicksichtigung der
sudafrikanischen Arten. Arch. Hydrobiol., 30, 550-56.

Walters, V. 1955. Fishes of western arctic America and eastern arctic
Siberia. Bull. Am. Museum Nat. Hist., 106, Article 5.

Woltereck, B. 1930. Alte und neue Beobachtungen iiber die geogra-
phische und die zonare Verteilung der helmlosen und helmtragen-
den Biotypen von Daphnia. Intern. Rev. Hydrobiol., 24, 358-80.

Woltereck, B. 1932. Baces, associations and stratification of pelagic
daphnids in some lakes of Wisconsin and other regions of the
United States and Canada. Trans. Wisconsin Acad. Sci., 27, 487-522.

THE SPECIES PROBLEM WITH
FOSSIL ANIMALS

JOHN IMBRIE: COLUMBIA UNIVERSITY, NEW YORK, n.y.

In spite of the extended attention this problem has received,
the nature of fossil species remains one of the most controversial
topics in paleontology. That this should be so in a decade that
has witnessed the publication of numerous synoptic studies on
the origin of species may surprise some students unfamiliar with
the materials and problems of paleontology. But two key ques-
tions are still being asked: What is a fossil species? How can
fossil species be recognized?

In defining fossil species it would be possible to ignore living
organisms completely and to frame definitions strictly in terms
of fossils. In fact, such an unbiological approach to paleontolog-
ical taxonomy is almost, but not quite, forced upon us in dealing
with some groups such as the conodonts whose biological func-
tions are obscure. But most attempts at a definition of fossil spe-
cies begin with a review of species concepts held by students
of living, sexual organisms, and from this foundation construct
a theoretical model of fossil species. The concept of fossil species
held by most paleontologists is largely an inference, an inference
based both on the observed structure of living species and on
a theoretical model of the evolutionary mechanism.

Paleontologists by no means agree on what a fossil species is.
Burma (1954), for example, has given thoughtful expression to
the thesis that species do not have an objective reality, a view
that is rejected by Simpson (1951) and others. In a symposium
on paleontological species, Eagar (1956) holds that for certain
groups of fossil clams a workable species concept must be
typological. Other students, including Newell (1956) and Syl-
vester-Bradley (1956), maintain that the concept of interbreed-

125

THE SPECIES PROBLEM WITH FOSSIL ANIMALS

5 populations is a necessary prerequisite for the definition and
lelineation of fossil species. Disagreement exists not only on the
nature of species but also on the inevitable question of their
proper scope. Thus Burma (1948) would distinguish as species
the smallest statistically recognizable grouping of populations,
whereas Imbrie ( 1956) argues that consistent application of such
a criterion would lead to useless multiplication of species names.
Since all the problems alluded to in the preceding paragraph
have their exact counterparts in neontology, it would be pointless
to make of this paper a compendium of paleontological disagree-
ment. Instead, we shall focus on those aspects of the species
problem which are unique, or are at least uniquely developed,
in dealing with fossil materials.

Species Concepts

In dealing with sexual organisms, whether fossil or living, two
fundamentally different species concepts can be employed. The
typological concept defines species as a group of individuals
essentially indistinguishable from some specimen selected as a
standard of reference. The biological species concept, on the
other hand, considers the species to be made up of one or more
variable, interbreeding populations. Both of these concepts serve
as theoretical bases for taxonomic work in paleontology today.

Different criteria may be emphasized in delimiting biological
species. For many students the criterion of interbreeding is de-
cisive, and species so defined may be referred to as genetic
species. In order to emphasize the fact that most modern and all
fossil species are distinguished more on the basis of morpholog)
than on breeding habits, some taxonomists recognize a morpho-
logical species concept as distinct from the genetical concept. But
this is really a trivial distinction. Morphological data are never
really considered by themselves; their interpretation is always
colored to some degree by prevailing theories on population
structure. Moreover, descriptions of species, like descriptions of
molecules and genes, arc inferences drawn from various sorts of
data, including observations on geographic distribution and ecol-
ogy as well as morphology and genetics.

J. IMBRIE 127

It is widely recognized that many anomalies arise when the
genetical definition of species as actually or potentially inter-
breeding groups of organisms is strictly applied. Asexnally repro-
ducing organisms are left out of account, for example, as are the
many instances on record in which local or temporary break-
downs occur in the genetical barrier between sympatric popula-
tions which on other evidence are classed as good species.
Simpson ( 1951 ) argues that such inconsistencies occur only be-
cause the interbreeding criterion is taken as the final test of
specific identity. He proposed that an evolutionary criterion be
substituted for the genetical so that the species is defined as a
segment of a phyletic lineage "evolving independently of others,
with its own separate and unitary evolutionary role and ten-
dencies." Clearly, this leaves wide latitude for judgment on the
part of the taxonomist delineating a species; and when taken
out of context Simpson's definition appears to be a different
verbalization of the famous dictum that a species is "a community
or number of related communities whose distinctive morpholog-
ical characters are in the opinion of a competent systematist suffi-
ciently definite to entitle it or them to a specific name" (Regan,
1926 ) . Taken in context, however, Simpson's definition differs by
insisting that a combination of morphological, biogeographical,
associational, ecological, and genetical data be used to assess the
evolutionary discreteness of a population or group of populations
under study. By the nature of the evolutionary process we cannot
eliminate arbitrary taxonomic judgments. Species-making will
remain a practical art as well as a scientific discipline.

Typological Species in Paleontology. The typological con-
cept of species is employed today either implicitly or explicitly
by a considerable number of paleontologists. To illustrate this
point we shall turn our attention to the work of a group of stu-
dents who have contributed a great deal to our knowledge of
British non-marine Carboniferous clams, notably Trueman, Weir,
Leitch, and Eagar (for a good summary see Weir, 1950). Like
their modern relatives the unionids, the shells described by these
workers (Anthraconaia, Carbonicola, etc.) display an extraordi-
nary amount of intrapopulation variation ( Fig. 1 ) . Now the mere

THE SPECIES PROBLEM WITH FOSSIL ANIMALS

fact of variation, however excessive, would not of course be
used as a justification for a typological approach. But the diffi-
culty in this case is enormously compounded by a combination
of stratigraphic, biologic, and practical economic circumstances.
In the first place, abundant collections of these shells have been
obtained from widely distributed localities and very numerous

POT CLAY COAL-SIX INCH MINE NON-MARINE SHELLS

^-^3*-C2*-€

c_.

"x

/

-■&

/

r^jb —

/" | \
/ . — lr- \

/ t

/

/ f

\ /

locol.tj 2, DABWEN Hill

locol.r^ 3 WRIGHTINGTON

Fig. 1. Variation diagrams of two local populations of Carbonicola ?
from a thin shell bed immediately above the Pot Clay and Six-Inch Mine
Coals of Yorkshire and Lancashire. These shells are referred to C. ? lenisul-
cata Trueman and C. ? aff. belhtla Bolton. Distribution diagrams, inset in
the upper left-hand corner of each pictograph, show the numerical strength
of variation trends. Black circles distinguish the figured variants, and white
circles show the disposition of the remaining shells, the position of each
one being controlled by its resemblance to one or more of the figured vari-
ants (Eagar, 1952b).

stratigraphic horizons in the British coal measures. The result is
an embarrassment of riches. To make the problem more complex,
many pairs of unit taxonomic characters vary independently, or
at least show low total correlation. From this it follows that
Statistical characterization of populations by means of simple
univariate or even bivariate distribution clusters will usually be
unsatisfactory. By means of multivariate regression analyses
Leitch (1940) has shown that it is possible objectively to identify

J. IMBRIE 129

variation norms and to place taxonomic discrimination on a quan-
titative basis. But such computations are far too laborious to en-
able the large number of population samples at hand to be proc-
essed. Another complicating factor is the labyrinthine course of
evolution followed by these British clams: with one exception,
rectilinear evolutionary trends have not been identified.

Faced with these complexities, Trueman and his co-workers
have evolved an ingenious and workable taxonomic procedure.
Each population sample is analyzed into variants which are illus-
trated and arranged in a systematic (and subjective) manner as
in Fig. 1. Subsequent analyses of similar populations can then
be abbreviated by means of appropriately placed dots on dis-
tribution diagrams. One or more modal or characteristic variants
in any population may then be described as a species. As a con-
sequence, an assemblage of shells from a single horizon and
locality may be described as a group of species, even though
there is every evidence that variation between the named species
is continuous.

The students who employ the system described above recog-
nize that their use of the term species violates the modern bio-
logical species concept, but the claim is made that this system
of nomenclature has proved to be the only practical one for
their peculiar biostratigraphic problems (Eagar, 1952a). Other
workers have argued that the same practical results could be
obtained within the framework of the biological species concept
(Sylvester-Bradley, 1952) and without burdening international
nomenclature with names of purely local application (Newell,
1956).

Biological Species in Paleontology. A majority of articulate
paleontologists working today employ a biological species con-
cept. Although this consensus reflects prevailing neontological
views, the paleontologists concept of species, which must take
into account important segments of geological time, is inevitably
more complex than the corresponding concept of zoologists. In
order to clarify this point it will be helpful to examine the sim-
plest possible phylogenetic model which includes the time di-
mension. Such a model, representing a small fragment of a

130

THE SPECIES PROBLEM WITH FOSSIL ANIMALS

phylogenetic tree, is shown in Fig. 2. The branches of this tree
represent ancestral-descendent population sequences replacing
one another through time as they undergo morphological and
genetic divergence. For the sake of simplicity, it is assumed that
the organisms involved are sexual, biparental, and that complica-
tions due to partial genetic barriers and Rassenkreise do not exist

12

13

II

rf^t

Divergence-

Fig. 2. A simplified theoretical phylogeny. At the indicated stratigraphic
horizons (A, B, C, D) transient species (1-13) are discrete evolutionary
units with no morphological or genetic overlap. In rare instances, where
the record of a line of descent is nearly continuous, segments of evolving
lineages may be designated as successional species. (After Newell, 1956.)

at levels A, B, C, and D. All the individuals living at time A arc
capable of interbreeding. By time B speciation has occurred so
that populations 2 and 3 arc no longer capable of genetic inter-
change. Five similar branchings are recorded in the diagram.

Given the simplified conditions of this model, contemporaneous
organisms arc always distributed among discrete, reproductively
isolated groups. For reasons discussed below, palcontological
samples almost always consist of the remains of individuals who
lived dining a fraction of geologic time so short that evolutionary

J. IMBRIE 131

morphological change within the sampled time interval cannot
be demonstrated. Tims species recognized on time planes A, B,
and C by the paleontologist represent stages in an evolutionary
continuum and are thus analogous to neontological species on
plane D. In some rather rare instances the paleontologist can
assemble material representing time planes so closely spaced that
modal shifts in an evolving succession of populations can be
demonstrated. Such circumstances are relatively rare, however,
and the paleontologist usually deals with widely spaced cross sec-
tions of evolving lineages. The term transient species will be used
in this paper to identify such cross sections, i.e., reproductively
isolated groups of individuals living during a single instant of
geological time. Normally the paleontologist's cross section is
limited geographically, and when this is the case, the transient
species of the paleontologist corresponds to the nondimensional
species of the neontologist (Mayr et at, 1953). In an increasing
number of instances, however, paleontologists have been able to
document taxonomically important infraspecific geographic dif-
ferences, i.e., polytypic transient species.

For practical reasons the neontologist can rarely undertake the
field sampling and breeding experiments required to document
statistically significant seasonal shifts in gene frequencies. Hence
the paleontologist of necessity and the neontologist by default
describe transient species in the majority of cases now on record.

In favorable circumstances the paleontologist can obtain ma-
terials which enable him to document gradual morphological
changes in a succession of fossil populations. In Fig. 2, for ex-
ample, collections taken at a number of horizons might define the
course of evolution between transient species 4 and 9. Quite
naturally, paleontologists use the term species for such segments
of phyletic lineages. Lineage segments of this sort can be called
successional species to distinguish them from the static groupings
referred to above as transient species. A number of theoretical
taxonomic and nomenclatural problems which arise in defining
the scope of successional species will be considered below in
connection with a discussion of Micraster.

132

THE SPECIES PROBLEM WITH FOSSIL ANIMALS

TYPOLOGICAL
SPECIES

BIOLOGICAL SPECIES

BASI S

MORPHOTYPE

LIVING POPULATION

FOSSIL POPULATION

DISTRIBUTION
IN TIME

VARIOUS

CONTEMPORANEOUS

ESSENTIALLY
CONTEMPORANEOUS

NON-
CONTEMPORANEOUS

TYPE

OF

SPECIES

TYPOLOGICAL
SPECIES

TRANSIENT
SPECIES

TRANSIENT
SPECIES

SUCCESSIONAL
SPECIES

1 '

U^J

/•» /

7 /'

PL/—]

<

UJ
1-

cr

GEN ETICAL

X

X
X
X
X

X

X
X
X

X

X
X
X
X
X

MORPHOLOGICAL

ASSOCIATIONAL

ECOLOGICAL

GEOGRAPHICAL

BIOSTRATIGRAPHICAL

Fig. 3. Tabular summary of some characteristics of typological, biologi-
cal, transient, and successional species. Operational taxonomic criteria asso-
ciated with each of these concepts are indicated.

Figure 3 summarizes aspects of the species concept which
have been emphasized in the preceding discussion.

Problems in Applying Biological Species Concepts to Fossils

Transient Species. In the foregoing discussion attention has
been focused on the theoretical model of the biological species
in use today by a majority of paleontologists. We shall now exam-
ine some of the practical problems which arise when this model
is applied to real paleontological data. Most of these problems
arise from three sources: the inadequacy of morphological data,
the prevalence of biased frequency distributions, and the incom-
pleteness of the available fossil record.

1. Inadequacy of inorphological data. The most obvious
shortcoming of paleontological data is that fossils normally pre-
serve only hard skeletal parts. To a student of soft-bodied living
animals (say Euglena) this difficulty might appear insuperable.
Indeed, some taxonomists assume for this reason alone that clas-

J. IMBRIE

133

sifications of living and extinct species are on an entirely differ-
ent level. For the vast majority of genera commonly found as
fossils, however, this judgment is exaggerated or untrue. There
can be, of course, no conflict between the classification of living
and extinct Euglena: like hosts of other soft-bodied forms this

-.23-.22-.2l -.20 -.!■» -.18 -.17 -.16 -.15 .14 -.13 -.12 -.11 -.I0-.C -J38-.0? -.0* -.03-.0* -.03 -.02 -.01 0 .01 .02 .03 £4 .05 .06 .07 .08 M .10

c

SKull length.

'

1

/

/
/
/
/

f —

\
\

f

I

i

Ci

' /

I
I

A»—

i

Zygomatic width

C'

/
/
/
/

{

At-

1
1

w;dth at molars

-! '

1

C' ■

ler

9+n

\
\
\

Ah

1

Upper cheeK teeth,

-pi

'

ng+h

s

/
/

1

Lower cheeK teeth

C>—

Fig. 4. Ratio diagram showing comparisons of certain osteological char-
acters in two species of the modern marten Mustek, and two modern sub-
species of Mustela sibirica. The latter show a strong overlap of measured
characters, whereas the two species show little overlap in the same charac-
ters. A, Mustela sibirica fontanierii; B, M. s. davidiana; C, M. altaica ka-
thiah. Abcissal scale represents for each character the difference between
the logarithm of any given value and the logarithm of the mean value of
the population selected as reference standard (in this case, population A).
Thus the horizontal distance between any two points represents the ratio
of either one to the other. Three points connected by a horizontal line are
plotted for a given character in each sample to represent the greatest, least,
and mean dimensions. The several means for each character in a popula-
tion are connected by dashed or solid lines. For discussion of ratio dia-
grams, see Simpson (1941). Mammal data from Allen (1938) (Colbert
and Hooijer, 1953).

genus leaves essentially no fossil record. Furthermore, it is gener-
ally true that specific discrimination in groups of animals com-
monly found as fossils is ( or can be ) based on skeletal morphol-
ogy. Consider, for example, the shell-bearing foraminifera, radio-
larians, corals, ectoproct bryozoans, brachiopods, snails, clams,

134

THE SPECIES PROBLEM WITH FOSSIL ANIMALS

echinoids, and mammals. For such groups differences in the
classification of fossil and living organisms at the species level
are rarely attributable to the inadequacy of morphological data.
Students of some of the groups just listed may argue that
many important specific taxonomic characters are not preserved
in fossils — pelage characteristics in mammals, for example. This
view, however, violates one of the basic tenets of taxonomy: that
the nature of a unit biological character is not so important in

— - i 1 1 1 1 1 1 1 1 « 1 r~'

- I 1 1 I I T ■

— I 1 I-1

Skull length ' A ,-,

B-

Bd" — ;

1
1

i
i

i
i

1

\
\
\
1
I

Rildte length j

1

\
\
\

j
i
i
/

1
1
1
1

1

/yoomatic wcHh \

i

— -L- —

««*

■ -i

i

Mastoid wid+h

. v.. ::::"-?."""

, v-

— i

— ^ ,

Upper molars, length

--

s

1

lower molars, length

Fig. 5. Ratio diagram comparing range of size variation in the fossil
bamboo rat (Rhizomys sinensis troglodytes) and three modern subspecies of
R. sinensis. A, Rhizomys sinensis troglodytes; Bd, R. s. clacicli; Bv, R. s.
lestitus; Bw, R. s. wardi. Diagram constructed in the same manner as
Fig. 4 (Colbert and Hooijer, 1953).

taxonomy as its statistical distribution within and between
populations (Simpson, 1943). This point is so fundamental to
an understanding of the paleontologist's approach to classification
on the species level that it will be illustrated with three examples
drawn from diverse sorts of animals.

In a study of Chinese Pleistocene mammals, Colbert and
Hooijer (1953) analyzed skeletal data on several species of
modern Asiatic mammals and made comparisons with similar
data on fossils. A ratio diagram (Fig. 4) based on skeletal meas-

J. 1MBRIE

I :}5

urements of two species of the modern marten Mustela indicated
that the amount of morphological overlap between two sub-
species is much greater than that between two species, showing
that osteological data alone reflect a classification (Allen, 1938)
based on geographic, eeologic, and pelage data. With this in

100 r

95

90

85

80

75

v
I

70

25

C/L

30

35

40

45

50

Fig. 6. Diagram showing estimated morphological ranges of seven pop-
ulations of the arcid pelecypod Anadara. H, maximum distance between
two lines parallel to the hinge; L, shell length; C, maximum depth of one
valve. Ellipses are statistical estimates of ranges to include 95% of the total
population ranges. Populations from Kojima Bay, Ariake Bay, and Dutch
Borneo are recent. Other populations are fossil. Samples from Dutch Borneo
and Minato silt (1) are assigned to Anadara granosa granosa; from Kojima
and Ariake Bays to Anadara granosa bisenensis; and those from the remain-
ing areas to Andara obessa (Kotaka, 1953).

136

THE SPECIES PROBLEM WITH EOSS1L ANIMALS

i • i

! : i
■ • i

«6-a

£ i §~ 5 "5

"$ §~~ s $ a "~ § ? 5 ^~

O

J. IMBRIE

137

H3

rt

o a

a o

o -^

tuaui|}»d$ jo jaqiuriN

bo

2 "3

QJ3

THE SPECIES PROBLEM WITH FOSSIL ANIMALS

mind, data on a collection of the Pleistocene bamboo rat
Rhizomys were compared with data on three subspecies of the
modern Rhizomys sinensis (Fig. 5). The extensive mutual overlap
among the four samples in the six measured skeletal features was
then used as a basis for considering the fossil a subspecies, R.
sinensis troglodytes.

A similar approach by Kotaka (1953) on recent and fossil
populations of the arcid pelecypod Anadara is illustrated in Fig.
6. Having concluded that the population in question could best
be distinguished on the basis of two shell indices ( H/L and
C/L), Kotaka calculated and graphed the 0.95 elliptical contour
of each bivariate density distribution. Note that although four
fossil and three recent populations are included in this study,
each receives the same taxonomic treatment; and there is every
reason to suppose that Kotaka's conclusions apply equally to fossil
and modern forms.

An illustration of the taxonomic importance of intrapopulation
variation patterns is furnished by Westoll's (1950) study of a
large sample of the Permian terebratuloid brachiopod DieJasma
elongata (Fig. 7). Frequency polygons of the thickness-length
ratio reveal a distinct bimodality in larger size classes, although
in all other respects shell characters display unimodal distribution
patterns. This bimodality is interpreted as sexual dimorphism.
Westoll's brachiopods and Kotaka's clams illustrate the same
point: sound taxonomic inferences can be made solely on the
basis of the distribution pattern displayed by shell features whose
functional significance is incompletely known.

One troublesome limitation of paleontological data lies in the
difficulty (usually the impossibility) ol distinguishing phenotypic
from genotypic variation. This problem is particularly evident to
the taxonomist analyzing statistically significant morphological
differences among a small number of samples whose stratigraphic
relationships arc poorly known. If suitable collections arc avail-
able for study, however, the taxonomic (but not the genetic)
problem may disappear. A case in point is McKerrow's (1953)
detailed study ol contemporaneous brachiopod communities sam-
pled from many localities along the outcrop of a thin stratigraphic

J. IMBRIE

139

unit known as the Fuller's Earth Rock. Although the genus
Ornithella displays a wide range of morphological variation at
each locality and a considerable overlap among geographically
separated populations, careful study reveals that certain modal
variants tend to characterize different localities, as indicated

Ornithella

(Q)

Atorfh

^ot/r/j

Wotfon/tAyr/s

(b)

/♦iv/A

■Souf/t

Rhync/7one//o/cJs

(C)

Fig. 8. Diagram of some lateral changes in brachiopod communities in
the Jurassic Fuller's Earth Rock, (a) Changes in the outline of Ornithella,
simplified and somewhat exaggerated; (/;) changes in the relative abun-
dance and in the anterior commissure of Wattonitkyris; (c) changes in the
relative abundance of the rhynchonelloids (McKerrow, 1953).

schematically in Fig. 8. Some morphological features in fact tend
to be symmetrically disposed about Mendips. Are we to interpret
these differences as the result primarily of genotypic or pheno-
typic variation? Even though this question must remain unan-
swered, the taxonomic expression of these data is in no way
affected. In either case the entire assemblage will be termed a
( transient ) species.

140

THE SPECIES PROBLEM WITH FOSSIL ANIMALS

Now suppose that McKerrow had had available only two geo-
graphically isolated but distinctive local populations. In this case,
taxonomic disposition of these collections would be far more sub-
jective. This serves to emphasize the point that the chief taxo-
nomic difficulties which arise in dealing with paleontological

10

>^/\

I mm

1.9 mm

Fig. 9. Graph showing size and frequency distribution of growth stages
represented in a population of the Silurian ostracod Bexjrichia jonesi from
Gotland (Spjeldnaes, 1951).

materials are due not so much to the limitations inherent in post-
mortem examination of skeletal anatomy, but rather to the incom-
pleteness of the available fossil record.

Biased frequency distribution. The size-frequency distribu-
tion of a few carefully studied paleontological samples seems to
reflect fairly accurately the size distribution of the original living

J. IMBRIE

141

population. For example, in a large collection of the Silurian
ostracod Beyrichia jonesi analyzed by Spjeldnaes (1951) there is
a progressive depletion in frequency at least from the fourth to
the eleventh instars (Fig. 9). This pattern has been taken as evi-

>
o

o

UJ

o

30

20

10-

Micraster

N = 348

27 33 39 45 51 57 63

Strophodonta

au

N =

4(

20

-

10

.

0

u

C \

Crurithyris
N * 511

J]

B

V

3j0 4.0 5.0 6.0 7.0 ao

Strophudonta

295 375 455 535

N = 49

■ nj

Jo \

6 8 10 12 14

SIZE IN MM

Fig. 10. Size-frequency distributions of four typical samples of fossil
invertebrates. N, number of measured specimens; A, measurements of great-
est width in a sample of the echinoid Micr aster coranguimim from the Cre-
taceous of Northfleet, England. Data from Kermack (1954). B, measure-
ments of maximum width in a sample of the brachiopod Crurithyris plano-
convexa from the Dry Shale of Kansas. Data from Tasch (1953). C, meas-
urements of maximum width of the brachiopod Strophodonta sp. from a
single locality in the Gravel Point formation, Michigan. Original data. D,
measurements of maximum length in a sample of Strophodonta extenuata
ferronensis from a locality in the upper Ferron Point formation, Michigan.
Original data.

dence that the collection approximates a random sample of the
original population, and on this assumption Kurten ( 1953 ) has
calculated life tables giving mortality rates at each growth stage
as well as other standard parameters of population dynamics.

THE SPECIES PROBLEM WITH FOSSIL ANIMALS

Samples of the sort just described are by no means common,
however. Four size-frequency distributions typical of fossil ma-
rine invertebrates are given in Fig. 10. Such distributions tend to
be unimodal and either symmetrical or slightly skewed; many
can be approximated by the normal distribution. In most in-
stances, the observed distributions are strongly biased and reflect
primarily a complex of ecologic, geologic, and operational factors
which at best bear only obliquely on the taxonomic problem
(Boucot, 1953; Kermack, 1954; Imbrie, 1955).

In dealing with groups of animals (mammals, for example)
where practical and objective morphological criteria are available
for the identification of definite growth stages, the existence of
biased size distributions may cause little concern to the taxono-
mist. In dealing with fossil groups lacking criteria of this sort,
however, particularly in groups displaying strongly allometric
growth patterns, taxonomic judgments may be considerably af-
fected. Once recognized, this problem can be quite easily solved
by characterizing sampled populations in terms of growth pat-
terns rather than growth stages ( Kermack, 1954; Parkinson, 1954;
Imbrie, 1956).

3. Incompleteness of the available fossil record. The most
serious limitation of paleontological data is the sparsity of fossils.
It is of course true that the total number of collecting localities
which have yielded good fossils from every system younger than
Precambrian is very large, and it is also true that future work will
bring forth an unknown but certainly very large amount of new
material from localities and horizons now unrepresented in exist-
ing collections. Nevertheless, from general theoretical considera-
tions on the nature of sedimentation and diagenesis, and from
practical experience in portions of the geological column which
have been thoroughly examined for fossils, most paleontologists
and stratigraphers would predict that no amount of future field
work will ever fill a majority of existing phyletic gaps between
transient species.

At least five factors are important in accounting lor the incom-
pleteness of the available fossil record, not including lack of ade-

J. IMBRIE 143

quate field work: nondeposition, erosion, migration, nonpreserva-
tion, and inaccessibility. Together, these factors account for the
prevalence of transient species.

Since the publication of the classic paper by Barrell (1917),
stratigraphers have generally admitted that very few local sedi-

B

»»"'». l.^•'.■^'.'.l'^•.TT'■^^.,•*:^^f^/?^^T^T^**•*••'T^'.M''.^■ ' ■'... "^"

Gap

VT.-. ■ .' ■'.'.' ■i,'.'Ty7rrT...... \. I

Gap

/ \

\

'•'.--' .'■'. '•'.'.'.'.-•'.■.•■'•■•■ •*>->>w^rT^'-'-'---J--^V. '•'■'. '!'.'.- '•.'.-• •

1 ■■ {' '*-• •'

N.

Gap

E

Trend

Fig. 11. Diagram to illustrate how gaps in the stratigraphic record re-
sult in a meager fossil record of a continuously evolving lineage. Gradual
evolution of a lineage is indicated by a series of shifting normal frequency
curves. Stippled areas represent preserved portions of the stratigraphic rec-
ord. Solid frequency curves represent four transient species collected from
layers A-D. Gaps in the fossil record, indicated by dashed curves, cannot
be eliminated by collecting in this area. (After Newell, 1956.)

mentary sequences contain a direct, continuous record of the total
time span represented in a given stratigraphic sequence. This is
because many fossiliferous strata were formed in water so shallow
that the permanent accumulation of a sedimentary layer (with
entombed fossils) was a rare event, possible only when the sur-
face of sedimentation was below the depth to which waves and

144

THE SPECIES PROBLEM WITH FOSSIL ANIMALS

currents were effective in moving sediment. Figure 11 illustrates
the formation of sedimentary and paleontological hiatuses accord-
ing to this principle.

Mere deposition of sedimentary layers does not safeguard their
preservation since large volumes of such rocks are destroyed by
uplift and erosion. Even if the sediments escape this fate, it does
not guarantee that even a small sample of skeleton-bearing or-
ganisms present in the original living community will become
collectable fossils. In order to be preserved a skeleton must escape

Region A

Region B

t

E

, i ■ ,.ii 1 1 1 1 .

Space

Fig. 12. The effect of facies migration on the fossil record of evolving
lineages. Well-defined transient species occnr at horizons 1 and 3 in region
A. Different transient species occur at horizons 2 and 4 in region B. Even
if deposition and preservation have been continuous in this area, inaccessi-
bility of the rocks lying between regions A and B makes the available record
of this lineage incomplete. (After Newell, 1956.)

chemical, biochemical, and mechanical destruction by agencies
acting before, during, and after burial.

Figure 12 illustrates the combined effect of migration and in-
accessibility in reducing the number of collectable fossils. Stip-
pled and unstippled areas on this diagram represent three con-
trasting environments (as well as the geologic record of those
environments, or facies) which have continuously occupied the
area. Three environments, each with its associated organic com-
munity, are represented at any given instant. Under normal cir-
cumstances geological conditions responsible for the localization
oi physical environments change, and the associated communities
sln'll accordingly. Now consider the effect of this migration on the

J. IMBRIE

145

fossil record of an evolving succession of populations. Four tran-
sient species belonging to one lineage are symbolized by numbers
1 through 4. Under ideal conditions, fossils representing the en-
tire lineage could be studied in outcrops of the appropriate facies.
In practice, however, the geological column can be studied only
in isolated, accessible areas. Hence a paleontologist examining
fossils from region A records two transient species 1 and 3, while
in region B his observations are limited to transients 2 and 4. If

TALLAHATTA LOWER WECHES UPPER WECHES

Fig. 13. Some of the characteristic features of four species of the
Eocene oyster Cubitostrea. Heavy stippling represents auricles. Formation
names indicate the stratigraphic position of individual species, as follows:
Tallahatta, C. perplicata; Lower Weehes, C. lisbonensis; Upper Weches, C.
smithvillensis; Cook Mountain, C. sellaeformis (Stenzel, 1949).

morphological gaps between studied transients are too large,
sound phylogenetic interpretations may be impossible.

One of the best documented examples of transient species re-
cording portions of an evolving lineage is Stenzel's ( 1949 ) study
of the oyster Cubitostrea. Four clear-cut species have been recog-
nized, each restricted to a different stratigraphic level in the Eo-
cene of the Gulf Coast. C. perplicata, the oldest known species in
this line, is characterized by small size, strong ribbing, thin shell,
absence of auricles, lack of arching, as well as other characters

o^-.

_d

FEET

YAZOO CLAY

00 .

*a.

SB

~3

,--

_

MOODYS 8R. N

w/:»y.w]

INVASION OF SEA

Ef^bE

** \KtGlONAL UlbCONI-UHMM If" ""

1500 -

^5^:

EXTENSIVE

_

REGRESSION

*»— *

LU

YEGUA

OF SEA

~

2

. -

LU

FORMATION

O

o

LU

ii-^i-i:^

SYMBOLS USED

z:-f^f^

1000 -

to

LU

_l
Q
Q

2
O

r-
O

Glouconiie
Colcoreous Shale
Sholes

ill

z

X.

o

I

s

-z.

Sonds

»—

_^

\~

- — _- ~ __-

X

o

rrr^= " \"

LU

Q.

--■ "

-I

<3

3

r^_-_^-_—

O

o

;=:- .z_—

(/)

tr

ICDE. ^LL Z

o

COOK MOUNTAIN

~ :^=r —

LU

FORMATION

^_-j^^

Cubifostreo selloeformis
/ (Conrod)

z

— _ -- - -^

500 -

a:

L^ 2 ,= _

o

.-»H

CO

INVASION OF SEA

<

STONE CITY

— -^r =

o

FORMATION

SPARTA SAND

- '-• -

REGRESSION

OF SEA

THERRILL SHALE

_-= —

^ C smithvillensis (Horns)

WECHES GLAUCON

Tf/rffrif

\^*C lisbonensis (Horns)

TALLAHATTA

YT^C. perplicota (Doll)

1 1

Fig. 14. Stratigraphic ranges of the chain of species comprising the
Cubitostrea sellaeformis lineage in the Gulf Coastal Plain (Stenzel, 1949).

146

J. IMBRIE 147

(Fig. 13). C sellaeformis, the youngest known species, is charac-
terized by large size, absence of ribs, thick shell, strongly devel-
oped auricles, and conspicuous arching. Gaps between these
extremes are partly bridged by C. lisbonensis and C. smithvillen-
sis.

This assemblage of species is viewed as the fossil record of a
continuously evolving stock leading to C. sellaefonnis. But the
point to be emphasized here is that each of these species repre-
sents only a stage in this evolution and is a transient, not a suc-
cessional, species. The stratigraphic range of each species is indi-
cated in Fig. 14. Although Cubitostrea is represented almost
continuously in beds ranging from the upper Tallahatta through
the Weches formations, the morphological and stratigraphic
ranges of the three species do not overlap. Similarly, transitional
forms linking C. smithvillensis and C. sellaeformis are unknown.
From independent stratigraphic evidence it is clear that the dis-
continuities just discussed are to be explained in terms of non-
deposition, erosion, and migration associated with advances and
retreats of the Gulf of Mexico.

Successional Species. All the difficulties discussed above in
connection with transient species, with the single exception of
the incompleteness of the fossil record, apply with equal force to
successional species. In addition there is the theoretical taxo-
nomic problem of selecting criteria for the subdivision of contin-
uous lineages. From what has been said it will be clear that this
problem is encountered far more often in the literature than in
the laboratory. More extensive field work will undoubtedly bring
forth more examples in the near future, but it is unlikely that the
problem of successional species will ever be a burden to the stu-
dent of fossils.*

One of the earliest and still the best example of successional
speciation is the fossil spatangoid sea urchin Micraster originally
described by Rowe ( 1899 ) . A summary of Micraster evolution

* Successional species may, however, prove to be common in cores taken
from deep ocean basins where deposition and preservation may be essen-
tially continuous. Here lie fascinating and nearly untouched areas for taxo-
nomic, evolutionary, and paleoecologic research.

THE SPECIES PROBLEM WITH FOSSIL ANIMALS

CORTES
1B0VIS

EVOLUTION
OF

MICRASTER

Fig. 15. Evolution of the spatangoid echinoid genus Micraster as re-
corded in six zones of the Cretaceous of southern England. Some of the
characteristic features of the five designated successional species are in-
dicated by ventral and lateral profiles. I, Zone of Cyclothyris cuvieri; 2,
/.one oi Terebratulina lata; 3, Zone of 1 1 olastcr planus; 4, Zone of Micraster
cortestudinarium; 5, Zone of Micraster coranguinum; 6, Zone of Marsupites
testudinarius. Constructed chiefly from data given In Kermack (1954).

J. IMBRIE 149

will serve as a convenient focus for a theoretical discussion of the
taxonomy of successional species. This review is based largely on
Kermack's (1954) critical restudy of Rowe's materials.

From many horizons in the chalky limestones of the English
Cretaceous large numbers of well-preserved Micraster have been
collected. Most of the limestones are soft enough so that simple
preparation techniques make it possible to free the tests from the
enclosing matrix. Owing to the complexity of the echinoid test, a
large number of unit taxonomic characters can be distinguished
and interpreted functionally by analogy with living spatangoids.
From many points of view, then, these fossils make ideal materi-
als for evolutionary and taxonomic studies.

Some of the outstanding features of the fossil record of Micras-
ter, as interpreted by Kermack, are shown in Fig. 15. The oldest
known members of the genus ( M. leskei ) possess small tests with
the widest portion of the test well ahead of midlength. The
mouth, lacking a distinct labium, is plainly visible in ventral as-
pect. The anterior ambital notch is shallow and the subanal fasci-
ole small. Closely spaced collections of the main line of evolution
from this form record a progressive shift in the average values of
the skeletal characters just mentioned, and in other characters as
well, giving rise to M. coranguinum. In this large species the
widest portion of the test is approximately at midlength, the
mouth is farther forward, and a strong labrum obscures the
mouth opening in ventral aspect. The anterior ambital notch is
deep and the subanal fasciole large. It should be emphasized that
distinctions between successive population samples can be made
only by comparison of the average values of overlapping fre-
quency distributions.

Two side branches of the main M. leskei-coranguinum line can
be distinguished. The first, M. corbovis, is a large form with a
number of characteristic features not discussed here. The second
branch, M. senonensis, differs from its contemporary M. corangui-
num principally in its complete or nearly complete lack of a sub-
anal fasciole. Frequency distributions of this character in a single
population studied by Kermack show two nonoverlapping clus-

THE SPECIES PROBLEM WITH FOSSIL ANIMALS

ters. In addition, M. senonensis tends to be relatively taller than
M. coranguinum.

Studies of living spatangoid echinoids make possible a reason-
able interpretation of the data presented above in terms of en-
vironmental adaptations. According to Kermack's view, the
evolution of the main M. leskei-corangninum stock is to be inter-
preted as progressively better adaptation to a burrowing habit.
This view is based primarily on known functions of the labrum
and subanal fasciole. In burrowing urchins, the labrum increases
the efficiency of the mouth as an organ of ingestion; and dense
cilia associated with the subanal fasciole produce strong posteri-
orly directed water currents which aid in discharging water from
the burrow.

The M. senonensis stock, on the other hand, seems to have
evolved adaptations for living on rather than in the bottom. For
this mode of life a subanal fasciole is unnecessary. The domelike
form of the test, moreover, might facilitate cleaning of the sur-
face; at least it would not hinder movement on the bottom.

We now come to the problem of expressing the facts and
inferences about Micraster taxonomically. At least three possible
criteria can be used for subdivision of such a continuum. First,
boundaries between species (or subspecies) might be placed at
those points in the phylogenetic tree where branching takes
place. Assuming that the phylogeny has been adequately docu-
mented, this criterion has at least the merit of objectivity. But
units so delineated may differ very greatly in scope. In the case
of Micraster — assuming a relatively constant morphological evo-
lutionary rate — the oldest lineage segment would be much
smaller in morphological scope than the youngest.

Stratigraphic boundaries offer another possible basis for the
delimitation of taxonomic units. Micraster, for example, might
be split into eleven species with taxonomic boundaries placed
at the limits of the six stratigraphic zones. Although to a biologist
such a course might appear rather arbitrary, to a stratigrapher
units so defined might be quite useful.

From a strictly biological point of view probably the most
satisfactory basis for subdividing an unbranched lineage is mor-

J. IMBRIE 151

phology. For example, taxonomic boundaries can be placed in
such a way that the total morphological difference between
corresponding parts of successive species is of the same order
as that between contemporary species. Application of this cri-
terion of course involves a subjective evaluation of total mor-
phological (and inferred genetical) differences.

The Micraster classification now in use and illustrated in Fig.
15 is based on a combination of the three criteria discussed
above, with emphasis on the morphology. Note that the inter-
pretation of the M. leskei-corangiiinum lineage as the main line
of Micraster evolution is reflected in the failure to subdivide
M. leskei and M. cortestudinarium.

Summary

A review of recent literature in paleontology indicates that
the nature of fossil species remains a controversial matter, al-
though the biological species concept based on variable, inter-
breeding populations has largely replaced the typological con-
cept based on morphotypes. Data on morphology, association,
biogeography, paleoecology, and biostratigraphy are used by the
paleontologist to delineate fossil species. Most species actually
defined by neontologists and paleontologists are transient species,
i.e., essentially contemporaneous aggregates of interbreeding pop-
ulations. In rare instances paleontologists describe segments of
phyletic lineages. Groupings of this sort are called successional
species.

The principal source of difficulty in applying the biological
species concept to fossils is the incompleteness of the available
fossil record. This lack is attributed to a combination of factors,
including nondeposition, erosion, migration, nonpreservation, and
inaccessibility. Taxonomic problems also arise from the limita-
tions inherent in morphological data and the prevalence of biased
frequency distributions. In dealing with successional species the
problem arises as to the selection of criteria for subdividing con-
tinuous lineages. Taxonomic boundaries may be designated on
the basis of phyletic branching, stratigraphy, or morphology.

152 THE SPECIES PROBLEM WITH FOSSIL ANIMALS

REFERENCES

Allen, G. M. 1938. The mammals of China and Mongolia. Natural

History of Central Asia, Vol. 11, Pt. 1, American Museum of Natural

History, New York, pp. 1-620.
Barrell, J. 1917. Rhythms and the measurement of geologic time. Bull

Geol. Soc. Amer., 28, 745-904.
Boucot, A. J. 1953. Life and death assemblages among fossils. Am. J.

ScL, 251, 25-40.
Burma, B. H. 1948. Studies in quantitative paleontology: I. Some

aspects of the theory and practice of quantitative invertebrate

paleontology. /. Paleontol., 12, 725-61.
Burma, B. H. 1954. Reality, existence, and classification: a discussion

of the species problem. Madrono, 12, 193-224.
Colbert, E. H., and D. A. Hooijer. 1953. Pleistocene mammals from

the limestone fissures of Szechwan, China. Bull. Am. Museum Nat.

Hist., 102, 1-134.
Eagar, R. M. C. 1952a. Some problems in invertebrate paleontology.

Proc. Leeds Phil. Lit. Soc., 6, Pt. 1, 50-53.
Eagar, R. M. C. 1952b. Variation with respect to petrological differ-
ences in a thin band of Upper Carboniferous non-marine lamelli-

branchs. Liverpool and Manchester Geol. J., 1, Pt. 2, 161-90.
Eagar, R. M. C. 1956. Naming carboniferous non-marine Iamelli-

branchs. In P. C. Sylvester-Bradley (1956).
Imbrie, J. 1955. Biofacies analysis. Geol. Soc. Amer., Spec. Papers,

No. 62, 449-64.
Imbrie, J. 1956. Biometrical methods in the study of invertebrate fos-
sils. Bull. Am. Museum Nat. Hist., 108, 211-52.
Kcrmack, K. A. 1954. A biometrical study of Micraster coranguinum

and M. (Isomicrastcr) senonensis. Trans. Hoi/. Soc. (London),

B237, 375-428.
Kotaka, T. 1953. Variation of Japanese Anadara granosa. Trans. Proc.

Palaeontol. Soc. Japan, n.s., No. 10, 31-36.
Kurten, B. 1953. On the variation and population dynamics of fossil

and recent mammal populations. Ada Zool. Fennica, 76, 1-122.
Leitch, J). 1940. A statistical investigation of the Anthracomyas of the
basal Similis-Pulchra Zone in Scotland. Quart, j. Geol Soc. London,
96, 13-37.

J. IMBRIE 153

Mayr, E., E. G. Linsley, and R. L. Usinger. 1953. Methods and Prin-
ciples of Systematic Zoology. McGraw-Hill Book Co., New York,
pp. 1-328.

McKerrow, W. S. 1953. Variation in the Terebratnlacea of the Fuller's
Earth Rock. Quart. J. Geol. Soc. London, 91, 97-122.

Newell, N. D. 1956. Fossil populations. In P. C. Sylvester-Bradley
(1956).

Parkinson, D. 1954. Quantitative studies of brachiopods from the
Lower Carboniferous reef limestones of England. I. Schizoplioria
resupinata (Martin). /. Paleontol, 28, 367-81.

Regan, C. T. 1926. Organic evolution. Rept. Brit. Assoc. Southampton,
1925.

Rowe, A. W. 1899. An analysis of the genus Micraster, as determined
by rigid zonal collecting from the Zone of Rhynchonella Cuvieri to
that of Micraster coranguinum. Quart. }. Geol. Soc. London, 55,
494-546.

Simpson, G. G. 1941. Large Pleistocene felines of North America.
Am. Museum Nat. Hist., No. 1136, 1-27.

Simpson, G. G. 1943. Criteria for genera, species, and subspecies in
zoology and paleozoology. Ann. N.Y. Acad. Sci,. 44, 145-78.

Simpson, G. G. 1951. The species concept. Evolution, 5, 285-98.

Spjeldnaes, N. 1951. Ontogeny of Beyrichia jonesi Boll. J. Paleontol,
25, 745-55.

Stenzel, H. B. 1949. Successional speciation in paleontology: the case
of the oysters of the scUaeformis stock. Evolution, 3, 34-50.

Sylvester-Bradley, P. C. 1952. In R. M. C. Eagar (1952a, pp. 52-53).

Sylvester-Bradley, P. C, ed. 1956. The species concept in paleontology.
Systematics Assoc. Publ. No. 2.

Tasch, P. 1953. Causes and paleoecological significance of dwarfed

fossil marine invertebrates. /. Paleontol, 27, 356-444.
Weir, J. 1950. Recent studies of shells of the coal measures, Sci. Prog-
ress, 38, 445-58.
Westoll, T. S. 1950. Some aspects of growth studies in fossils. Proc.
Roy. Soc. (London), B137, 490-509.

BREEDING SYSTEMS, REPRODUCTIVE
METHODS, AND SPECIES
PROBLEMS IN PROTOZOA*

T. M. SONNEBORN:f Indiana university,

BLOOMINGTON, INDIANA

The so-called modern biological species concept defines a species
as a group of actually or potentially interbreeding populations.
As the other papers in this symposium attest, this concept is
widely accepted as the only valid species concept. Yet it is of
admittedly limited applicability. In obligatorily self-fertilizing
organisms, this concept of the species contracts to include but a
single individual. In exclusively asexual organisms, the concept
admits of no species at all. On this concept, perhaps half or more
of the species recognized by students of the Protozoa are invalid
and, in principle, cannot be modified so as to become valid. In
like manner, practically every phylum of Invertebrates and all
major plant groups contain numerous organisms in which species
do not exist at all or in which each individual is a separate spe-
cies. This greatly limits the applicability of the modern biological

Note: Superior numbers in text refer to Addenda, page 314.

* Contribution #618 from the Department of Zoology, Indiana Uni-
versity.

f The work of the author and his collaborators, much of which has not
previously been published, was supported by grants from Indiana Uni-
versity, the Rockefeller Foundation, the American Cancer Society, and the
Atomic Energy Commission (Contract No. AT ( 11-1) -235 ) . The following
people have critically read part or all of this manuscript: Mary L. Austin,
G. H. Beale, T. T. Chen, Ruth V. Dippell, A. C. Giese, L. Gilman, E. Han-
son, Mrs. Margaret Johnson, R. F. Kimball, E. Mayr, D. L. Nanney, Ursula
Philip, J. R. Preer, and R. Siegel. I am immensely indebted to these experts
for checking the accuracy of my statements, for generously providing me
with unpublished information and permission to cite it, and for numerous
criticisms and suggestions.

155

156 PROTOZOA

species concept. It has been maintained, not without some reason,
that there are more species of Protozoa than of all other animals
together, for each species of the higher animals is believed to
be sole host to at least one species of parasitic Protozoa. If half
or more of the Protozoa, and many other organisms as well, are
in principle outside the domain of the modern biological species
concept, it is applicable to only a minority of all organisms.

Although this is perhaps the greatest objection to adopting
this species concept as the only valid one, there are also others.
In some organisms, what is now recognized as a single species
on morphological grounds would be broken up into fantastically
large numbers of "biological" species. Moreover, very commonly
these "biological" species would be to an appalling degree un-
recognizable and unidentifiable by routine taxonomic procedures
or by any other procedures that could reasonably be expected
of working biologists.

The Protozoa illustrate on a lavish scale these difficulties and
limitations of the modern biological species concept. There is
need for a broader species concept which will embrace the whole
of biology. But it should not sacrifice unnecessarily the very real
values of the so-called modern biological species concept. The
present paper will review the situation in the Protozoa and will
inquire as to how there might be developed for obligatory in-
breeders and asexual organisms a concept which embraces for
them a level of biological organization comparable to the one
which is embraced for outbreeders by the modern biological
species concept. Both concepts would, it seems, have to be special
cases of a more general concept applicable to all organisms. How-
ever, as will appear, such conceptual ordering does not in itself
settle the question of whether the unit groups of individuals and
populations designated by these concepts should always be called
species and assigned specific names. This is a separate1 matter and
is separately considered.

Until relatively recently, much of what now appears to be
significant information about the Protozoa in relation to these
problems was unknown. This is particularly true of the Ciliates
which, because of the extensive literature now available on them,

T. M. SONNEBORN 157

receive greatest attention. The pertinent modern literature is in
great need of survey from a broad unitary point of view. That
is attempted here in considerable detail. The basic facts about
the six genera of Ciliates that have been investigated along
modern lines are given species by species. This sort of work is
being pursued more extensively now than ever before. The pres-
ent review therefore will doubtless soon be outdated by new dis-
coveries.1 Nevertheless, it may serve a useful purpose not only
in summarizing the situation as it now stands, but also in point-
ing up the gaps in knowledge and in suggesting a central idea to
guide further investigation.

This central idea is that the system of breeding — inbreeding
or outbreeding to various degrees — which an organism normally
follows is closely correlated with a considerable number of its
major biological features; that these features have significance
primarily with respect to their consequences for the breeding
system; and that herein lies the key to the species and evolution-
ary problems in the Protozoa. Attempts are therefore made to
recognize, for each organism dealt with, those biological features
that bear upon the breeding system and species problems and to
assess their consequences. As will abundantly appear, although
enough is known to validate this point of view in a general way,
the attempt to apply it in detail to the organisms that have been
studied reveals a great dearth of basic facts. One of the chief
purposes of this paper is to call attention to the kinds of facts
that are needed, in the hope that investigators will be stimu-
lated to supply them. As a stimulus to such efforts, the paper
attempts to portray the depth of understanding of fundamental
problems and the unification of a field to which such efforts
surely would lead.

Before proceeding to the modern work on particular species
of Ciliates, which began in 1937, a few comments on some earlier
observations concerning the breeding systems of Ciliates will
serve to introduce it. Breeding in Ciliates has long been known
to take the form of conjugation. The animals unite in pairs, un-
dergo meiosis, fertilize each other, and separate; then both ex-
con jugants of each pair multiply by repeated fissions. Maupas

158 PROTOZOA

( 1888, 1889 ) discovered that a declining food supply was usually
a necessary, but not sufficient, condition to bring about conjuga-
tion. He further maintained that outbreeding was the rule, partly
because he obtained fruitful conjugation only when he had put
into his aquaria collections from different natural sources. When
conjugation took the form of inbreeding, it led only to death of
the exconjugants.

These observations and conclusions of Maupas were contro-
verted by the studies of Jennings (1910) on Paramecium cauda-
tum and P. aurelia. He cultivated in the same container strains
of the same species that could be distinguished by the size of the
animals. When conjugation occurred in these cultures, each
strain appeared to conjugate only among its own members, even
when both strains were conjugating at the same time. Moreover,
working with isolated single individuals of a single strain, he
found that conjugation could occur among individuals produced
in the course of a few fissions from a single ancestor and that
such exconjugants produced normal viable clones. A number of
successive closely inbred generations were followed with the
same result. On the basis of these observations, he denied the
validity of Maupas' generalization and held that conjugation in
Paramecium could occur only between individuals of the same
strain and that within a strain any individual could probably
mate and mate fruitfully with any other.

Early workers on breeding systems thus came to diametrically
opposed observations and conclusions. According to Maupas, the
Ciliates are obligatory outbreeders; according to Jennings, they
are obligatory inbreeders. As will appear, the observations of
both investigators were entirely correct. The differences are in
the Ciliates themselves. Some are outbreeders; some are inbreed-
ers. On the question of who can and does mate with whom, and
with what results, we shall have much to say below.

A few later workers (De Garis, 1935a,b; Midler, 1932) claimed
to have succeeded in obtaining crosses between different strains
and even between different species of Paramecium. Although
the methods employed by De Garis were not unobjectionable,
the fact that diverse strains of these species can be crossed has

T. M. SONNEBORN 159

been amply confirmed. Objections can also be raised against the
evidence for species crosses, but this needs to be reinvestigated
with the better methods now available. In any case, no viable
hybrids between species were claimed by these workers and
that, from our point of view, is perhaps the main thing.

The modern work on Ciliates began with the discovery of mat-
ing types in P. aurelia (Sonneborn, 1937). As originally used, the
term mating types refers to physiological differences between
individuals that mark them off into mating classes. True conju-
gation does not ordinarily occur between individuals of the same
physiological class or mating type; it occurs between individuals
of complementary physiological classes or mating types. This is
the sense in which the term has come to be used by students
of Protozoa and Fungi. But it has a somewhat different
meaning in Euplotes, as will be explained later, and the possi-
bility that conjugation can occur between individuals of the
same mating type in certain clones of some species or in certain
stages of the life cycle has not yet been completely excluded.
These special cases will be dealt with as they arise.

Commonly, but not invariably, the mating type of an individ-
ual is inherited by its asexually produced descendants. This
makes it possible to use clones rather than individual animals in
the study of breeding systems. Such studies quickly exposed
fundamental relations in P. aurelia (Sonneborn, 1938) and in
P. bursaria (Jennings, 1938) that have subsequently been widely
confirmed in sufficiently studied Ciliates in which mating types
have been found. If many collections are made from different
natural sources (ponds, lakes, streams, rivers, and the like) and
a strain from each source is established in the laboratory, the
general features of their breeding relations are ordinarily found
to be those described in the following paragraphs.

Clones of the same strain are either of the same or of comple-
mentary mating types. When complementary, they conjugate
promptly and abundantly with each other under appropriate
conditions. When of the same mating type, they cannot be made
to conjugate with each other. Every clone of every strain has a
definitely assignable mating type. When mating types of any one

160 PROTOZOA

rain are matched against the mating types of any other strain,
two very different results may be obtained. On the one hand,
the two strains may contain clones manifesting identical sets of
mating types. When this is so, the same rules of mating apply to
combinations of clones from different strains that apply to com-
binations of clones from the same strain : conjugation occurs if the
clones combined are of complementary mating types, not if they
are of the same mating type.

On the other hand, two strains may manifest quite different
sets of mating types. Then no conjugation will occur when any
clone of the one strain is combined with any clone of the other
strain. Thus, conjugation does not result from mere difference
of mating type, but from complementariness of mating type.
There are multiple sets of complementary mating types within a
single species. Some strains have one set and these interbreed
freely. Other strains have another set, and these also interbreed
freely. But strains with the first set of types are sexually isolated
from strains with the second set of types. When many strains
of a species are examined, a number of such groups of strains is
discovered. As a rule, each is sexuallv isolated from all the
others. The mating types of one set usually will have nothing to
do with mating types of the other sets.

From these results, it at once became evident that P. aurelia
and P. bursaria each consisted of a number of "biological spe-
cies." Both Jennings and Sonneborn were well aware of this and
considered the problem in their first papers of 1938, which re-
ported the basic facts. However, they were then unable to de-
scribe these biological species so that they could be readily dis-
tinguished and identified. They therefore agreed, for the time
being, to refrain from assigning specific names. Instead, they
designated the sexually isolated groups of strains within a spe-
cies as varieties, assigning to each variety a number. This usage
has persisted and spread, to the annoyance of some geneticists
and doubtless to the relief of protozoologists. Tn this paper, I
shall, among other things, inquire exhaustively — in view of sub-
sequent additions to knowledge — into the questions of whether
the time has come to recognize the varieties as species and of

T. M. SONNEBORN 161

whether they should be given species names. Meanwhile, I shall
continue to refer to them as varieties and will reserve the term
species for the species now recognized by taxonomists.

In the next eight sections, I shall deal with six genera of Cili-
ates that have been examined with respect to mating types and
varieties: Paramecium, Tetrahymena, Colpidiam, Euplotes, Sty-
lonychia, and Oxytricha. We begin with Paramecium, the first
genus to be studied in this way and the one about which most
is known. Pertinent information is available on six species of
Paramecium — a good deal on P. aurelia, P. multimicronucleatum,
P. bursaria and P. caudatum; less on P. calkinsi and P. trichium.
The first four will be dealt with in extenso and in order in the
immediately following sections; the other two will be only briefly
mentioned in the last section on Ciliates, along with the other
lesser studied Ciliates ( Stylonychia, Oxytricha, and Colpidium),
after sections that take up work on the more fully known Ciliates,
Tetrahymena, and Euplotes. The sections on Ciliates will be fol-
lowed by a section on obligatory inbreeding and asexually repro-
ducing Protozoa, chiefly Flagellates and Rhizopods. The last two
sections set forth the major conclusions and give a summary of
the whole paper.

Before proceeding to the four intensively studied species of
Paramecium, it should be emphasized that the six species of
Paramecium to be discussed are, with one possible exception
presently to be mentioned, readily recognizable and sharply set
off from one another by combinations of a few morphological
features: mainly, body shape and the number and structure of
the micronuclei. Three species have the familiar cigar-shaped
body: P. caudatum, P. aurelia, and P. multimicronucleatum.
P. caudatum has a single, compact, relatively large micronucleus.
P. aurelia has two smaller micronuclei which are vesicular in
structure. P. multimicronucleatum has the same kind2 of micro-
nuclei as P. aurelia, but their number varies from 2 to 7. As will
appear, I doubt whether a clear distinction2 can be made between
P. aurelia and P. multimicronucleatum. Pending clarification of
this difficulty, I shall treat the organisms of both "species" as if
they were all P. aurelia. The other three species have bodies

162 PROTOZOA

shaped like a Dutch wooden shoe viewed from the side: P. bur-
saria, P. calkinsi, and P. trichium. P. calkinsi has two vesicular
micronuclei. P. bursaria has a single, relatively large, compact
micronucleus and is unique in being the only species that is
normally colored; it carries a symbiotic, unicellular, green Alga.
P. trichium also has a single, compact micronucleus; the possibil-
ity that it is related to P. bursaria in the same way as P. aurelia
is related to P. multimicronucleatum needs to be investigated.
The sizes of the animals are also said to be characteristic for
each of these six species. The figures given by Wichterman
( 1953 ) are 70 to 90 v. for P. trichium, 85 to 150 y. for P. bursaria,
120 to 170 p for P. aurelia, 120 p for P. calkinsi, 170 to 290 |& for
P. caudatum, and 180 to 310 [a for P. multimicronucleatum.2
Although these figures are the best available, they are not entirely
satisfactory. Size varies greatly with nutritive conditions, stages
of the fission cycle, and stages of the life cycle. No comparative
study of the six species which properly takes such variables into
consideration is available. On the whole, however, it seems rea-
sonably certain that on the average P. trichium is the smallest
species, that P. caudatum and P. multimicronucleatum are the
largest, and that the others are intermediate.

Paramecium aurelia: Specificity of Mating Types;
Isolation of Varieties

In this section will be discussed not only the organisms that
ordinarily go by the name of P. aurelia, but also those which are
designated as P. multimicronucleatum and intermediate forms'-'
which I have recently found. It is difficult if not impossible to
draw a sharp line where one species ends and the other begins
(Sonneborn and Dippcll, 1956). 2 The older name, P. aurelia, will
therefore be tentatively used for all these organisms.

Tin's section begins with an enumeration of the varieties of
P. aurelia and an account of the primary basis of their recogni-
tion, mating type specificity. It proceeds to a consideration of
sexual and genetic isolation of the varieties, and then to their
geographical distribution. Lastly is discussed the question of

T. M. SONNEBORN 163

whether each variety may be considered to have a potentially
common gene pool.

The next section enumerates the differences that distinguish
the varieties and inquires whether these differences suffice, for
practical purposes, in the recognition and identification of the
varieties. At this point, the question of whether the varieties
should be given specific names is considered. Then follows a
section that deals with evolution in P. aurelia at various levels
and attempts to relate the biological differences among groups
of varieties, among varieties of a group, and among strains of a
variety to the systems of breeding, the various degrees of in-
breeding and outbreeding. This key section ends with a con-
sideration of the relation of systems of breeding to the species
problem.

Mating Types and Varieties. The primary basis for recog-
nizing the existence of distinct varieties in Ciliates is the spec-
ificity of the mating types, i.e., the specific selectivity of the
breeding relations. On this basis, 16 varieties of P. aurelia can
now be recognized. Varieties 1 to 7 were discovered by Sonne-
born (1938, 1942b), variety 8 by Sonneborn and Dippell (1946),
variety 9 by Beale and Schneller (1954), varieties 10 to 15 by
Sonneborn (1956a; unpubl.). What is here referred to as variety
16 was described under the name of P. midtimicronucleatum by
Giese (1941).3

The number of mating types per variety is remarkably uniform.
As a rule, there are two and only two interbreeding mating
types in a variety. Until recently, only one type was known in
variety 7; but only one strain of that variety had been found.
Sonneborn (1956a) found the missing mating type, along with
the one previously known, in each of several new strains of
variety 7. Only variety 16 remains as the exception to the rule.
According to Giese (1941), it has four mating types, each of
which can mate with the other three. Were it not for this unique
feature, there would be reason to suppose that Giese's variety is
identical with Sonneborn's variety 15.3

Mating relations between varieties follow a usual rule, to which

164

PROTOZOA

S

e

go

.£

'■+2

0>

.—

--

S3

>

CO

O

.P

H

o

%—

o

~-

CO

O

o

o

o

o

(>1

H

CO

p

b£

o

• -—I

_P

+2

o

'-2

o3

03

o

s

«

oC

"?.

P

-1-3

O

0>

rj

P

>

CO

;„

P

o

.2

'■+3

hH

c3

—

0/

•_

I I

«s

0!
CO

O

a

03
-♦J

GO
>>

CO

03
-P
H

X

X

X

X

X

X

X

X

X

I I

I I

I I

I I

I I

I I

I I I

I + I

+
+

I + I I II I + I + I o

I I

I I

I I I I

I _|_ i o o

+
+

I + I o

o o

+
+

I I

I I

+

I -H I + I c
+
+

c c o

+

I I I I

O o o o

I I

+

I -H I I

I I + I c °
+
+

I "H I o o o o c

+

+

I I

I I

+

a

a

r.

f*l

7Z

u

oS

i>

I I

I I

I I

I I

+ I

+

+

o o

I I

I I

+ I

+
+

o o o

o o o o c a o o

+

O 3 O O C O

OO CO OO' OO CO

o o o o o o

c c o c c

o c o

oo oo o o oo

o

HH ** ^

oo oc oo OO CO

CO CO OO OO oo

M -c>>ixxx

H X XX XX XX XX

■O I- 00 o -f

41 rt -°

o c —

d fe 2
£§-8

com
a) c c

o c'5

o— s

t. oj o

,»_> a

u c ,
-3-2 '
g*

o « .

^o.2
2 », s

o„C t-
£ >> <D

.= 22

Us J

"o.2.2

S 03 >,

■- ~ a
O-S-S

.25

■Sg

= "2 e
"m-2.2

HI

— B'B
■So*

si a c

"^ ?

<u C
^ C3 +
bC_.
.5 *>

|||

<Z - -

~ c
(1) t-

a> ^ C
fig*3.

- i E -

a u z
° *-= Z

>0-§*5

cs c a> =

r = 3

C _ _!.■-.

"r + e
.2 + §

fl ^ o

M.d 0
cS--"CO
S . u

w c F .
2.2 c-

OS;"

-•=..5 J:

.81. S

>, s + e

to o ! «

•■> +

u -

o

T. M. SONNEBORN 165

there are, however, a number of exceptions. The rule is that two
mating types belonging to different varieties do not conjugate
with each other. Indeed, they do not even react to each other
in any observed way. Exceptions to this rule are shown only by
certain combinations of mating types belonging to the eight
varieties 1, 3, 4, 5, 7, 8, 10, and 14, as shown in Table I. These
exceptions are of two kinds. First, certain combinations of mating
types of different varieties react sexually to each other and form
conjugating pairs. Second, other combinations give a weaker sex-
ual reaction, adhering to each other briefly, but they never pro-
ceed to form conjugating pairs. These intervarietal reactions are
often difficult to obtain except when the cultures are in optimal
condition for conjugation. It is therefore possible that some com-
binations which have never been observed to react might do so
and that others which have been seen to react only weakly
might yield some conjugant pairs.

The system of intervarietal mating reactions leads clearly to
the conclusion that, at least for the eight varieties listed in
Table I, the two mating types of one variety are related to the
two in other varieties. In the two pairs of varieties 1 and 5, and
4 and 8, the odd-numbered mating type of each variety mates
with the even-numbered mating type of the other : both I and
IX mate with II and with X; and both VII and XV mate with
VIII and with XVI. Hence, similarities exist between mating
types I and IX, between II and X, between VII and XV, and
between VIII and XVI. Types V and VI can be brought into the
system of similarities because V mates with both II and VI; and
II mates with both I and V. Hence I is similar to V and II is
similar to VI. The groups of similarities are enlarged by the facts
that V mates with II, VI, and XVI; and XVI mates with V, VII,
and XV. Hence V, VII, and XV form a similar set; and II, VI,
and XVI form another similar set. By adding the other reactions
in Table I, it can easily be shown that the odd-numbered mating
types in these eight varieties form one similar group which may
be designated the minus group; while the even-numbered mat-
ing types form another similar group, which may be called the
plus group (Sonnebom and Dippell, 1946). Thus, each of these

166 PROTOZOA

varieties has a plus and a minus mating type. The plus types of
the different varieties are comparable, and the minus types of the
different varieties are also comparable. Extension of this system
to the other eight varieties which are not listed in Table I can-
not be done on the basis of the same sort of evidence because
these other varieties have never been observed to yield inter-
varietal mating reactions. General considerations and other less
direct evidences, however, make it seem likely that the mating
types of those varieties which have only two mating types (i.e.,
all except variety 16 )3 also are in each case plus and minus in the
same sense.

In spite of the system of cross reactions and of intervarietal
mating, the uniqueness of each mating type and therefore of
each of the eight varieties in the table cannot be doubted. This
is shown in two ways by the mating reactions themselves. First,
the mating reaction is less intense when the plus type of one
variety is mixed with the minus type of another. Even in mix-
tures involving the varieties 4 and 8, which show the strongest
intervarietal mating reactions, one combination of odd and even
types never yields as high a percentage of pairing as a compar-
able combination within a variety. The other combination of odd
and even can yield as high a percentage of pairing as do the two
types of one variety; but the conditions for such an optimal
reaction seem to be stricter than those required within a variety.
Hence, as a rule this intervarietal reaction is also somewhat re-
duced. All other intervarietal combinations yield considerably
weaker reactions, usually very much weaker. Second, correspond-
ing types of different varieties commonly differ in the intensities
of their reactions with any type to which both react. Thus type
XVI of variety 8 fully conjugates with type V of variety 3, but
type VIII — the corresponding type of the closely related variety
4 — does not even react with type V. Indeed, Tabic I shows that
each mating type gives a set of reactions that is distinctly dif-
ferent from every other one. For example, type I of variety 1
reacts strongly with II, moderate!) with X, weakly with XVI,
and not at all with VI, while the corresponding type V of
variet) 3 reacts weakly with II, strongly witli VI, not at all

T. M. SONNEBORN 167

with X, and moderately with XVI. The uniquely defined system
of reactions of each mating type, including of course the total
absence of reactions between mating types of certain different
varieties, was the original, and remains the most reliable, basis
for distinguishing the varieties.

Isolation of the Varieties. The preceding account has passed
lightly over the important fact that conjugation can occur be-
tween certain different varieties. Since varieties 1, 3, 5, and 7
can interbreed, do they not have a common gene pool? Should
they not all be grouped together as one biological species? More-
over, since variety 4 conjugates with variety 8 and the latter con-
jugates with variety 3, should not varieties 4 and 8 also be mem-
bers of this species? These questions place squarely before us the
problem of what the common gene pools in P. aurelia really are.
To this problem, which includes the questions just asked, but is
broader than they, we now turn.

First of all, there seems clearly to be no gene flow from cer-
tain varieties to certain other varieties. The seven varieties not
listed in Table I have never been observed to give the slightest
sexual reaction to animals of any variety other than their own.
This may be due to inadequate study so far as the varieties which
are newly discovered are concerned. However, others, like varie-
ties 2 and 6, have been long known, and it is practically certain
that neither of these reacts with the other or with any of the
other long known first eight varieties; gene flow between each
of these and the other six must be considered as nonexistent.

Secondly, even when the bar to interbreeding is let down, as
among most of the varieties listed in Table I, other mechanisms
interfere with gene flow among them. The only open question is
whether the interference is completely effective or whether it is
not quite completely effective. Crosses between varieties 4 and
8 almost always end in death after a few fissions ( Sonneborn and
Dippell, 1946; Melvin, 1949). Levine (1953) thoroughly studied
the small proportion of Fl survivors. Although he could not prove
their hybrid nature, about a dozen among over 300 Fl clones
examined could have been true hybrids. Later generations, even
by backcrosses, from these rare survivors were, however, as

168 PROTOZOA

mviable as the Fl generation. Thus, although a complete bar
i gene flow was neither demonstrated nor controverted, intro-
gression of genes from one variety into the other must be either
very rare or nonexistent.

The other four possible crosses between varieties, 1 by 3, 1
by 5, 1 by 7, and 3 by 8, regularly yield a viable Fl. The F2
generations and backcrosses prove, however, to be completely or
nearly completely nonviable. Butzel (1953) found abnormalities
even in the Fl of the cross between varieties 1 and 7. Like
Levine, he was unable to prove that recombinants occur in the
F2 or backcrosses. I have obtained a little evidence, not fully
satisfactory, that the rare survivors in the F2 or later generations
may sometimes be recombinants. Some clones behave as if they
were the plus (or minus) mating type of both parental varieties,
but a descendant of such a clone after a later fertilization may
behave as if it had a mating type of only one of the parental
varieties. Whether this is due to change of dominance relations in
a heterozygote or to recombination is not clear. Further, descend-
ants of the hybrids have been maintained for a year or more.
They reproduced slowly, but mated avidly when given a chance.
It seems unlikely that life could have been maintained so long
without recurrent fertilizations, although it is possible, in view
of some of our recent work (Sonneborn, unpubl. ), that these
may have survived by repeated regenerations of functional nuclei
from the original hybrid macronucleus instead of developing
macronuclei from the fertilization nucleus after each fertilization.

Thus, knowledge of the consequences of intervarietal crosses,
in the five combinations in which they can occur, is still not
sufficient to give a definite answer to the question of whether
gene flow can occur among those varieties. On the other hand,
the difficulties in analyzing the hybrids are due to the low viabil-
ity of the later generations and to the abnormality of the survi-
vors. This shows that, if genes flow between varieties, the flow
can be at most but a trickle and iu great danger of drying up.
In short, the genes of every variety are virtually isolated from the
genes of every other variety. Before the significance of this ge-
netic isolation of the varieties can be properly evaluated, it is nee-

T. M. SONNEBORN 169

essary to inquire how extended each variety is in space and
whether potential gene flow exists within a variety.

Distribution of the Varieties. About 260 collections of P.
aurelia from various parts of the world have been identified as
to variety in my laboratory and in the laboratories of Giese
(variety 16) and Beale. Most of these collections, about 200,
were made in the United States. Of the rest, most were collected
in Europe (chiefly by Beale), South America, and Japan. A few
were collected in India (5), Australia (4), Madagascar (1),
Puerto Rico ( 1 ) , and Hawaii ( 1 ) . Although the sampling is
spotty, unbalanced, and grossly inadequate, a few important
relations are clear (Beale, 1954; Sonneborn, 1956b).

At least three of these varieties, 1, 2, and 6, are probably dis-
tributed around the world. Varieties 1 and 2 have been found
in Europe, North and South America, and Japan; variety 1 has
also been found in Australia and Hawaii. Variety 6 has been
found in the United States, in Puerto Rico, and in southern India.
It is rare in the United States, except in Florida; but all five col-
lections from the other two localities belong to this variety. In
addition to these three, variety 4 is also of wide distribution,
having been found in North and South America, Australia, and
Japan, but not yet in Europe.

The other twelve varieties seem to be much less widely dis-
tributed. Variety 9 has been found only in Europe, variety 12
only in Madagascar, the other ten varieties only in the United
States. These statements need some qualification. Varieties 12,
13, 15 and 16 are large paramecia; until recently we have paid
little attention to paramecia of that size on the assumption that
they were not P. aurelia.'-* So we have only one collection of large
paramecia from outside the United States, and at present that
is our sole strain of variety 12. It would be premature to suggest
that 12 does not occur in the United States or that 13, 15, and
16 do not occur on other continents. However, even when these
three varieties of large paramecia are excluded, this leaves the
seven smaller varieties 3, 5, 7, 8, 10, 11, and 14 that have been
found only in the United States. If other parts of the world were
studied on a similar scale, it would not be surprising to find com-

] TO PROTOZOA

parable numbers of varieties restricted to each continent. On
the other hand, such a search might well reveal that some of the
varieties found so far only in North America occur also else-
where. Yet it is not to be assumed that we have already found
all the varieties on this continent. Within the past year, three
new small animal varieties (10, 11, and 14) have been found
here; we have only begun to search for the large animal varieties;
the far north (Canada and Alaska) has not been sampled at all;
and we still have some large animal strains from the United
States that we have not yet been able to identify or even to
get to mate at all.

Without going into the quantitative details, which are given
in Table II for the small animal varieties, it is clear that tempera-
ture is a major factor influencing the range of a variety on the
continents where it occurs. Variety 6 has the highest temperature
preference, being the only variety found in the hottest collecting

Table II. Distribution by Rough Temperature Zones of the Smaller Animal

Varieties of Paramecium <turdiaa

Total

Variety

number

of

■(.I lections

Zone

2

3

9

5

1

4

8

11

7

14

10

G <

Cold

33

31

19

8

6

0

0

3

0

0

0

0

36

Cool

30

23

4

15

20

G

1

0

0

0

0

1

110

Moderate

26

3

0

10

45

10

0

0

0

0

0

0

31

Warm

12

0

0

0

18

18

29

6

6

4

2

4

49

Hot

0

0

0

0

0

0

0

0

0

0

0

100

5

Total no. of

collections

59

37

11

23

47

21

15

4

3

2

1

8

231

" Varieties 12, 13, and 15 are the larger animal varieties; these are excluded
because they have been little searched for. The numbers in the body of the table
(except the bottom row and last, column) are the percentages of all collections
from a zone which belong to each variety. Cold zone: northernmost U.S.A.,
Norway, and Scotland; cool /one: England, northern France, Germany, and a
strip across the U.S.A. just south of the most northern states; moderate zone:
hi (Honshu), southeast, Australia, Switzerland, northern Italy, and the
middle region of the I'.S.A. (north-central California to Virginia and the Caro-
linas); warm zone: southernmost U.S.A. (New Mexico to Florida), Peru, north
and central Chile; hot zone: Puerto Rico and southern India.

T. M. SONNEBORN 171

zone. Varieties 7, 8, 10, and 14 come next. With three exceptions,
my strains of these varieties come from Florida; the exceptions
are from Mississippi (variety 14), New Mexico (variety 8), and
Maryland (variety 8). Variety 4 ranges from cool to warm
climates, but it is absent from the coldest and hottest collecting
sources; it is equally prevalent in moderate and warm regions,
rarer in cool regions. Variety 1 has been found in all zones except
the hottest, but its peak frequency is in the moderate zone.
Variety 2 has the same range as variety 1, but its frequency de-
creases with rising temperature. Varieties 3 and 5 have never
been found either in warm or hot regions; variety 3 has its peak
in the cold zone, variety 5 in the cool zone. Variety 9 has a dis-
tribution in Europe somewhat like that of variety 3 in North
America.

Much less is known about the distribution of the larger animal
varieties. That is why they were not included in Table II. Giese's
(1941, 1957) survey of variety 16 is the most extensive. He
obtained 16 collections of this variety from various regions of the
United States extending from Oregon and New Hampshire in the
north to Texas in the south. From this it would seem that variety
16 has a wide temperature range. I have now four collections of
variety 15 from the state of Washington in the northwest to
Florida in the southeast. It too appears to have a wide tempera-
ture range similar to that of variety 16.3 Of variety 13 I have
four collections, all from southern United States (New Mexico
and Florida), suggesting a restricted warm climate localization.
Only the one strain of variety 12 from Madagascar is known, but
no other large paramecia outside North America have been
examined.

So far as present knowledge goes, it would appear that some
varieties, both small and large animal varieties of P. aurelia, have
wide temperature ranges while others — again both small and
large animal varieties — have narrow temperature ranges. Inde-
pendently of this, of the small animal varieties, some are re-
stricted to one continent and others are distributed around the
world in the appropriate temperature zones. Whether this is also
true of the large animal varieties remains unknown.

172 PROTOZOA

Do the Strains of a Variety Form a Potentially Common Gene
Pool? The distribution of each variety over a greater or lesser
area raises the question of whether it is composed of subdivisions
between which genes cannot flow or whether each variety, no
matter how widely distributed, constitutes a potentially common
gene pool. Laboratory analysis shows that normal vigorous re-
combinants can always be obtained in the F2 generation after
crossing any two normal strains, i.e., samples of different local
populations, of the same variety. This happens regardless of
whether the two strains come from different continents or from
neighboring ponds or streams. It has indeed become standard
laboratory practice to make use of the genie diversity of different
strains by introducing certain genes of one strain into the genetic
background of another strain through a series of selected back-
crosses. There thus can be no doubt that the genes of one variety
form a potentially common pool.

That is the major fact, but there are also two other related
facts of considerable significance in this connection. First, when
strains of the same variety are crossed, at least in some varieties,
the F2 and the first backcrosses invariably yield some nonviable
clones. The proportion of nonviable clones depends upon which
strains have been crossed. In variety 4, the frequency of non-
viable clones obtained in the autogamous F2 generation after
crossing certain strains may be as low as 15%; it may be as high
as 98% after crossing certain other strains; and various combina-
tions of strains yield percentages that cover the range between
these extremes (Sonneborn, unpublished; Dippell and Hanson,
unpublished). The same strain crosses regularly give approxi-
mately the same results. The magnitude of the mortality seems
to vary independently of the geographical proximity or remote-
ness of the source of strains; but this needs to be investigated
more fully.

Certain strain crosses within a variety thus yield about the
same percentage of F2 mortality as intervarietal crosses. One
wonders therefore whether gene flow is possible between such
extreme pairs of strains. Margolin ( 1956a, 1)) lias obtained nor-
mal, \ iable recombinations between strains 32 and 172 of variety

T. M. SONNEBORN 173

4, which yield 85% mortality in the F2, and also between strains
172 and 51, which yield almost the highest F2 mortality observed,
about 96%.

Thus, with respect to F2 mortality alone, there appears to be
a virtually uninterrupted series of gradations from crosses be-
tween varieties down to crosses between different clones of the
same strain. The former approach 100% F2 mortality; the latter
usually show 0% F2 mortality, at least in some varieties. But the
series is not continuous when other features are taken into
account. The survivors in the F2 or backcrosses show no im-
provement in viability at subsequent autogamies or backcrosses
following intervarietal crosses; but the mortality decreases to
zero at the second autogamy or after several backcrosses follow-
ing interstrain crosses within a variety. The survivors in the
later generations are not by any means of normal vigor after
intervarietal crosses; they are quite normal in vigor after inter-
strain crosses within a variety. All intervarietal crosses that can
be made give essentially the same low survival in the F2 and
backcrosses; it is impossible for the genes of two varieties to be
recombined in normal vigorous combinations. Although different
strains of the same variety sometimes yield as low F2 survival
as do diverse varieties, their genes can always be recombined in
such combinations. The sharp break between varieties is thus
never matched by the break between strains of the same variety.
Among the latter, gene flow is readily possible; among the former,
it is nonexistent or almost so.

Paramecium aurelia: Differences among Varieties
and Problem of Identification

The preceding section showed that each of the 16 varieties of
P. aurelia has a potentially common gene pool which is effec-
tively cut off from the gene pool of every other variety. Each
variety therefore qualifies as a species according to the modern
biological species concept. Two questions now arise. ( 1 ) Should
the term "variety" be replaced by "species" in referring to them
henceforth? (2) Should each former variety be baptized with a
different specific name, while the old "species" P. aurelia becomes

174 PROTOZOA

defunct as a species name and is raised to subgeneric status? The
first question could be answered in the affirmative without neces-
sarily giving an affirmative answer to the second question. There
might be good grounds for objecting to this mass baptism, in
spite of concluding that the varieties are nevertheless species. To
this extent the two questions are independent. On the other
hand, if each variety should be given a specific name, then this
answer to the second question automatically answers the first
question. The second question is therefore considered first and
is alone dealt with in this section.

The decision as to whether the varieties should be given spe-
cific names depends, in my opinion — and I believe there would
be general agreement on this — upon the feasibility of identifica-
tion of the varieties by means available to working biologists,
means that do not present unreasonable demands. Hence, before
a decision can be reached, the known differences among the
varieties must be summarized in sufficient detail to make clear
whether they could reasonably be utilized in the practical job
of identification. At the same time other probable differences, not
yet known or not sufficiently known, will be considered in order
to answer the related question of whether other useful differences
might be available for the same purpose.

At this point the reader must be prepared to abandon some
common, but grossly erroneous, ideas about the Protozoa. Many
of them, including Paramecium are far from simple, either in
structure or in life processes. Nor are all those which go by a
single name pretty much alike. On the contrary, when one at-
tends closely to those going under a single name, such as
P. aiirelia, studying in various ways samples of populations col-
lected from far and wide around the world, one marvels at the
complexity and variety of their lives. This section, which por-
trays this situation, will doubtless overwhelm the reader in
much the same way as the investigator who devotes his life to
studying these creatures is overwhelmed by their seemingly end-
less diversities which he slowly and with great labor discovers.
So, reader, you can scarcely be more overwhelmed than I am.
Indeed, not so much so. For I know, as you cannot, how pitifully

T. M. SONNEBORN 175

small a fraction we have learned about the complexity, variety
and marvels of this "simple, primitive, stereotyped animal." Yet
there is no escaping the task of trying to comprehend the situa-
tion if we are to attempt to reduce it to significant order in rela-
tion to species problems. I shall therefore take up the various
differences among the varieties in this section and shall try to
reduce them to significant order in the next section.

Mating Type Specificity. The definite and sharp distinctions
among the mating types uniquely mark off each variety, as
described in the preceding section. No further comment is
needed here. Toward the end of this section, the usefulness of
these distinctions in identification will be considered.

Distribution. The distributions of the varieties in the major
land masses and according to temperature were stated in the
preceding section. The usefulness of this knowledge in relation to
identification is limited of course by the inadequacy of the sam-
pling which underlies our present knowledge. With due allow-
ance for this, the source of a culture still has suggestive value
in identification. For example, a sample from northern Europe
in all probability would be found to belong to variety 1, 2, or 9,
and there is no use considering other possibilities until these have
been tested. Such considerations save much time in identifica-
tion, but seldom if ever could a strain be reliably identified solely
on the basis of its natural source.

Temperature in Relation to Survival and Fission Rate. The
characteristic distribution of each variety with respect to tem-
perature suggests the existence of varietal differences in tem-
perature tolerance. Very little is known about this. Sonneborn
and Dippell (unpublished data of 1942) found that strain P of
variety 1 from Maryland grew well at 36.4°, but the strains of
the other four varieties (2, 3, 4, and 7) tested did not survive at
this temperature very long. Of the latter, variety 3 (strain 58
from Indiana) and variety 7 (strain 40 from Florida) grew well
at 34°; but variety 2 (strain 50 from Oregon) and variety 4
(strains 29 from Maryland and 51 from Indiana) did not survive
at 34°. These results indicate that predictions based on geograph-
ical distribution cannot be relied upon. For varieties 3 and 7

176 PROTOZOA

showed a similar temperature tolerance, although the latter has
been found only in Florida whereas the former never has been
found in warm climates and is most abundant in cold climates.
However, this particular strain of variety 3 came from near the
southern limit of its range, and there may be strain differences
in temperature tolerance within a variety. Indeed, some strains
of variety 4 grow well at 35° (Margolin, 1956b), while others
die at that temperature. Preer (unpublished) has even found
temperature mutants within a strain of variety 2. The mutant
cannot survive at 31° while the wild type can. Of the tolerance
for low temperatures even less is known. We have some indica-
tions that variety 4 strains become abnormal if long grown at
10°. There is much need for a systematic study of temperature
tolerance. It may prove to be diagnostic for varieties with limited
natural range of climates, but more variable in varieties with
wide natural temperature distributions.

The relation of temperature to fission rate may be studied in
several ways: comparing different varieties at a standard tem-
perature, comparing the maximal fission rate at whatever tem-
perature yields it, comparing the temperatures which yield the
maximal fission rate, and comparing the entire curve relating tem-
perature to fission rate over the whole viable range of tempera-
tures. The latter method is the most revealing. Sonneborn and
Dippell (unpublished data of 1942) did this for the representa-
tives of varieties 1, 2, 3, 4, and 7 mentioned above. Up to about
23°, they found only minor differences in fission rate among these
strains; but at higher temperatures the differences became
strongly accentuated and each variety gave a distinctive curve.
The question of the existence of strain differences within a variety
was not satisfactorily analyzed in their work. For the strains
examined, the maximal fission rates attained were in part related
to the maximum temperature tolerated, for fission rate increases
with temperature almost up to that point, but even at lower tem-
peratures differences appeared. Thus both at 28.5° and 30.5°,
the order of the varieties in fission rates, from high to low, was
4, 1, 7, 3, 2, covering a range (at 28.5°) from 5.8 down to 4.2
fissions per day. Varietal differences in fission rate thus unques-

T. M. SONNEBORN 177

tionably exist, but much more needs to be known before this
could become of much value in identification.

Morphological, Anatomical, and Cytological Differences. The
most obvious morphological features are of course size and shape.
Unfortunately, however, these vary with cultural conditions,
stage of the division cycle, stage of the life cycle from one fer-
tilization to the next, and so on. Nevertheless, under comparable
conditions there are some striking differences, particularly in
length and breadth, among the varieties. Some unpublished ob-
servations by Sonneborn and Dippell, made in 1942, may be com-
pared with unpublished observations by Elizabeth Powelson
made 14 years later on different strains of the same varieties.
Only the mean lengths of overfed vegetative animals are given
here :

Variety l Variety 2 Variety 4
Sonneborn and Dippell (1942) 139 M 145 m 112 n

Powelson (1956) 131 m 149 m 112 m

This remarkable agreement inspires much confidence in the
existence of very real size differences among the varieties, so
that others which were made at only one time may be added:
Variety 3, 123 n; Variety 7, 140 n; Variety 12, 204 fi. The meas-
urements on variety 12 are by Powelson, those on varieties 3 and
7 by Sonneborn and Dippell. Measurements on other varieties
are not available, but long familiarity with the material permits
me to estimate that the characteristic size in varieties 8, 10, and
14 is essentially the same as in variety 4, i.e., about 112 n. Like-
wise, the animals of varieties 13 and 15 and, judging by Giese's
report, 16, are very large, surely 170 m or more. The other varieties
are intermediate and roughly in the range 125 to 150 ^. This
leaves open the question of strain differences within a variety.
Such differences doubtless exist, at least in some varieties, but on
the whole they probably never blur the distinctness of the three
major varietal size classes listed above. The latter provide an
easy means of at once placing an unknown strain in one of three
groups of varieties. No measurements are necessary to do that,
for the differences are considerable to a practiced eye.

178 PROTOZOA

A cvtological feature of taxonomic importance is the number
of mieronuclei. As mentioned earlier, the number is supposed to
be two in P. aiirelia and two to seven, but usually four, in P. mul-
timicronucleatum. Because we include both here as P. aurelia,
varietal differences in this respect might be expected.2 A serious
difficulty, however, is that much variation in micronuclear num-
ber can occur in a single strain, as has been repeatedly observed
by a number of investigators. Some of this variation seems to be
correlated with the age of the clone under observation. Observa-
tions of number of mieronuclei in clones of unknown age are in-
secure grounds for identification. Thus, the clones of large
paramecia which we have thus far obtained had mostly two or
no mieronuclei; but further study, as set forth below, showed
clearly that young clones of the same variety normally have four
mieronuclei. Experience of this sort leads to the conclusion that
micronuclear number in this material becomes a reliable diagnos-
tic trait only when used for young clones and only then when a
considerable sample of clones is available to indicate the usual
condition.

The young stage, which is in my experience most critical, is
the stage shortly before the first fission after fertilization. At this
time there are present the differentiated mieronuclei and macro-
nuclear anlagen that develop from the products of the first few
divisions of the syncaryon. In varieties 13 and 15 the normal
condition is to have four mieronuclei and four macronuclear an-
lagen. Giese does not give the number for variety 16, but from
other information he provides I conclude that this variety will
also be found normally to develop four nuclei of each sort.3 All
other varieties normally form two nuclei of each sort. The persist-
ent qualification with the word 'normally" is necessary, for vari-
ations occur in all varieties. But the normal condition is found
in the overwhelming majority of animals at this stage.

Other cvtological and anatomical features are not as yet suf-
ficiently known to be discussed. Among these, the greatest need
is for stud) of the silver line system and chromosome number
and morphology. The former is under investigation by Powelson

T. M. SONNEBORN 179

(1956). At present, only variety 4 has been carefully studied in
respect to chromosome variation (Dippell, 1954). I have reason
to suspect that very different results might be obtained with
varieties 15 and 16, and perhaps certain others.

In sum, three groups of varieties may be readily distinguished
on the basis of size, and the largest varieties may be further sub-
divided on the basis of the number of nuclei just after fertiliza-
tion. Variety 12 regularly has two micronuclei and two macro-
nuclear anlagen at this stage whereas the other large animal
varieties regularly have four of each. All the smaller varieties
regularly have two of each.2,3

Killers and Sensitives. Some strains of P. aurelia liberate a
poison to which they are resistant, but which kills other strains
of paramecia. Only certain varieties of P. aurelia include strains
of killers. They have been reported in variety 2 (Sonneborn,
1939) and variety 4 (Sonneborn, 1943). A different kind of killer,
which liberates no poison but kills by prolonged contact during
mating, was discovered in variety 8 (Siegel, 1954). I have re-
cently found a killer in variety 8, which kills without mating, and
also killers in variety 6. Thus killers of one kind or another are
now known in varieties 2, 4, 6, and 8, but not in any other va-
riety.4

As a rule, all nonkiller strains of all varieties are sensitive to
all killers. Only strains that can conjugate with mate-killers can
be tested for sensitivity, but this restriction does not apply to
other killers. Sensitive strains of variety 2 alone show heightened
resistance to killers; the degree of resistance varies from strain to
strain, but it is sometimes very great (Preer, 1950). The signifi-
cance of this variation in resistance is somewhat complicated by
Austin's (1951) discovery that variations in resistance occur
within a single sensitive strain of varieties 4 and 8 in correlation
with the serotype they happen to be expressing at the time (see
under "Serotypes" ) . Nevertheless, the commonly greater resistance
of variety 2 may well be independent of this.

For purposes of identifying strains, the preceding information
is of some value. If one has a killer strain, it probably belongs

PROTOZOA

to variety 2, 4, 6, or 8; if it is a mate-killer, probably it belongs
to variety 8. If a nonkiller strain is extraordinarily resistant to a
known killer, it almost surely belongs to variety 2.

Serotypes. As is well known, serologic techniques have been
used in the study of evolutionary relationships, and they are cur-
rently used in routine identification of certain microorganisms.
The possibility of using such techniques in identification of the
varieties of P. aurelia therefore needs to be considered, particu-
larly since serological traits have been much studied in several
of the varieties. These studies have exposed a serologic situation
which is far from simple and not yet completely known; but
enough is known to indicate that serologic identification of the
varieties of P. aurelia may become possible.

Present knowledge is confined almost exclusively to antigens
which are detected by immobilization of the paramecia in the
presence of appropriate antisera obtained from injected rabbits.
Any one individual seems to manifest, as an almost invariable
rule, one and only one immobilization antigen at a time. This
antigen defines its serotype in the sense in which that term is
here used.

The serotype of an individual commonly persists among its
progeny; but, as in phase variation of bacteria, changes of sero-
type may occur within a single line of descent. However, there
are more phases or serotypes expressible in a line of P. aurelia
than in a line of bacteria. As many as 14 different serotypes have
been brought to expression in different individuals of a single line
of descent (Margolin, 1956b), and nearly as many serotypes have
been found in other lines by others. Change of serotype occurs
without genetic recombination or gene mutation. The whole ar-
ray of serotypes characteristic of a line of descent appears in
individuals of identical — and even entirely homozygous — geno-

pe. The changes of serotype are reversible and can be experi-
mentally controlled and directed by appropriate manipulation of
environmental conditions. In order to discover the array of sero-
type!; producible by any line or strain, prolonged study with a
wide variety ol cultural conditions is required.

Different strains of the same variety show similarities in their

T. M. SONNEBORN 181

systems of serotypes. If a serotype, A, is found in one strain, the
homologous anti-A serum is found often to be capable of immo-
bilizing one of the serotypes in each of many other strains of
the same variety. These are therefore all designated as serotype
A. In like manner, one finds many strains have serotype B, C, and
so on. Genetic analysis reveals another kind of similarity be-
tween strains of the same variety. The serologic specificity of each
serotype is determined by a single gene. Different strains may
carry different alleles. The whole array of specificities controlled
by alleles at one locus is given the same serotype designation re-
gardless of whether cross reactions are detectable serologically.
There are thus genetic as well as serologic similarities between
serotypes of different strains.

There are also differences between the serotypes of different
strains of the same variety. These are of three main sorts. First,
corresponding serotypes of the different strains, although desig-
nated by the same letter, may differ more or less in serologic
specificity. For example, the immobilization titer of an anti-D
serum against the D serotype of the homologous strain may be
greater than against the D serotype of another strain. This is not
always true; but when it is, the differences in titer vary in ex-
tent from very small to very great. In the extreme case, there
may be no cross reaction at all, and the inference as to corre-
spondence then depends upon genetic considerations. Second,
the cultural conditions required to bring a particular serotype
to expression may be alike or different for different strains. For
example, serotype B may require for its expression cultivation
at high temperature in some strains, cultivation at low tempera-
ture in others. Third, some strains may be capable of expressing
a serotype which other strains cannot be made to express at all
under any of the many conditions of culture employed. This
means then that the discovered array of serotypes may differ
somewhat from strain to strain.

As might be expected, different varieties also show similarities
and differences in their systems of serotypes, and the degree of
resemblance and differences runs parallel to the closeness or re-
moteness of evolutionary relationship as inferred from other

• * * * - *

182 PROTOZOA

lines of evidence. These independent evidences of evolutionary
relationship will be discussed later. Figure 1 anticipates that dis-
cussion by diagramming the inferred relationship of the six va-
rieties that have been studied serologically: varieties 1, 2, 3, 4,
8, and 9. The serologic results are summarized in Beale ( 1954 ) ,
except for the subsequent publications by Finger (1956; 1957a,
b) on variety 2; by Melechen (1955) on variety 3; by Austin et

Fig. 1. Inferred evolutionary relationships of varieties 1, 2, 3, 4, 8, and
9 of Paramecium aurelia, the only varieties that have been studied serolog-
ically.

al. (1956), Margolin (1956a,b), and Skaar (1956) on variety 4;
and by Pringle (1956) on variety 9; and unpublished work of
Preer on variety 2 and Sonneborn and associates on varieties 4
and 8.

The most closely related varieties are 4 and 8 and their sero-
type systems show the greatest resemblance. For the most part,
each serotype found in one variety is matched by a serologically
recognizable corresponding serotype in the other variety; but a
few serotypes have thus far been found only in one or only in
the other variety. Further, the corresponding serotypes are sel-
dom if ever serologically identical although they commonly cross-
react. The cross reactions are usually weaker than those involv-
ing corresponding serotypes of different strains of the same
variety.

The differences between these two varieties and variety 2 are
much more marked. Several serotypes (A, C, E, F, G, and II)
have been recognized in variety 2 which correspond to serotypes
ol varieties 4 and 8, but they show marked differences in specific-
it}. There are, moreover, some general resemblances in the sys-
tem ol serotypes, which, as will appear, distinguish these three

T. M. SONNEBORN 183

varieties from the other three more remotely related ones (i.e.,
varieties 1, 3, and 9). First, in varieties 2, 4, and 8 each strain
usually possesses several alternative serotypes that can be stably
maintained under appropriate conditions. Second, so far as now
known, corresponding serotypes of different strains of the same
variety seldom if ever totally fail to cross-react to each other's
homologous antiserum. Third, in any one strain, usually more
than one stable serotype may be expressed in different sublines
under the same conditions; that is, there is a strong cytoplasmic
element controlling the persistence of a given serotype. Fourth,
corresponding serotypes of different strains may require very dif-
ferent conditions for their expression.

In all four of these respects the varieties 1, 3, and 9 of the
other group seem to be like each other but different from va-
rieties 2, 4, and 8. The phenomena have been most fully worked
out in variety 1 (Beale, 1954). Although each strain of variety
1 has a considerable array of potentially producible serotypes,
only three serotypes, D, G, and S, are very stable, and these are
regularly found in all natural strains. The specificities of any
one of these serotypes in different strains may be much alike,
but often the differences in specificity are so great that no cross
reactions occur and correspondence is detectable only by genetic
and physiological means. Further, the expression of these three
serotypes is in a uniform way in all strains closely dependent
upon temperature. Thus, in any strain serotype S requires for
its expression lower temperature, and serotype D higher tempera-
ture, than serotype G. This results regularly in the expression of
only one stable serotype in any one strain under given conditions
and evidence of a cytoplasmic element in the persistence of a
serotype is as a rule inconspicuous or lacking. In so far as evi-
dence is available, these features of the serotype system appear
also to be operating in varieties 3 and 9.

The differences among varieties 1, 3, and 9 concern mainly
the array of serotypes in each and the specificities of correspond-
ing serotypes when such are found. These differences are com-
parable to those found among varieties 2, 4, and 8. The two varie-
ties 1 and 3 (which are more closely related than either is to

PROTOZOA

variety 9, but which are not as closely related as are varieties 4
and 8) show greater resemblances than either does to variety 9.
Further, their serologic resemblances do not seem to be as close
as those which exist between varieties 4 and 8.

In spite of the profound differences between the two groups of
varieties, some serologic similarities are still detectable. Thus, se-
rotypes D and G are found in both groups. Moreover the D sero-
type of a strain of variety 1 may be more like the D of a strain
of variety 4 than it is like the D of another strain of variety 1.
These serotypes may be ancient and widely distributed. Preer
(unpublished) has found serotype G even in the distinctly differ-
ent species P. caudatum.

The body of knowledge now available on the serotypes of P.
aurelia, although limited to only 6 of the 16 varieties, is already
of some practical value in identification. If one has a battery
of antisera against the various major known D, G, and S sero-
types of variety 1 and grows an unidentified strain at low, inter-
mediate, and high temperatures, the strain could be identified
with reasonable confidence as belonging to variety 1 if the ani-
mals are immobilized by one of the D antisera at high tempera-
ture, one of the G antisera at intermediate temperature, and one
of the S sera at low temperature. Even a less complete series of
responses would be highly indicative. In like manner identifica-
tions of strains belonging to varieties 3 and 9 could probably be
made with appropriate antisera; but at present work on these va-
rieties is less extensive and doubtless a fuller battery of antisera
than now available would be required. The procedure with varie-
ties 2, 4, and 8 would have to be somewhat different because of
the absence of the regular temperature relations and, conse-
quently, the great strain differences in serotypes which are stable
at any particular temperature. One would have to use all avail-

le antisera against the whole array of serotypes known in the
variety. In variety 4, there are 17 such serotypes and a number of
variants of some of them. Moreover, new serotypes arc1 continu-
ally being found as more1 strains are studied. Because o\ this great
variability and the incompleteness of present knowledge, the
probability of success in attempts at serologic identification is

T. M. SONNEBORN 185

perhaps less than in variety 1. However, I have already repeat-
edly proved the usefulness of this serological method of distin-
guishing closely related varieties, such as 4 and 8, which are the
most difficult to distinguish in the usual way by their mating
specificities, and the serologic identification has never failed to
be confirmed by later mating tests. Beale (unpublished) has had
similar success in serologically distinguishing varieties 1 and 9.

There can be little doubt that the practical usefulness of the
serological method of identification of varieties will increase as
our knowledge of the serotypes is extended to more strains and
more varieties, thereby providing more complete sets of antisera
for use in this task. In principle, it should become possible even-
tually to identify the varieties of P. aurelia in this way with as
much assurance as the so-called species or serotypes of Salmo-
nella are now identified. But no one should underestimate the
magnitude of the task, even if all of the necessary antisera were
available. Their number would be very great, perhaps about 20
for each of the 16 varieties, or a total of over 300.

Mating Type Inheritance. Two groups of varieties, A and B,
are distinguished by their systems of mating type determination
and inheritance. Group A includes varieties 1, 3, 5, 9, and 11 and
also probably the varieties 7 and 15 which have been little stud-
ied. Group B includes varieties 2, 4, 6, and 8 and also probably
the little studied varieties 10, 12, and 14. Nothing is known about
mating type inheritance in variety 16. 'A Variety 13, as will appear
later, cannot yet be assigned to either group A or B.

The outstanding features of the group B system are ( 1 ) the
conspicuous role played by the cytoplasm in the determination
of mating type and (2) the consequent rarity of change of mat-
ing type after fertilization. A fertilized animal usually gives rise
to a clone of the same mating type as the parent. The group A
system shows no conspicuous effect of the cytoplasm on the
determination of mating type. Consequently there is no signifi-
cant correlation between the mating type of parent and offspring
arising after sexual processes. Instead of being a clonal trait, as
in group B, mating type is a caryonidal trait in group A; that is,
each new macronucleus that arises in a fertilized animal is inde-

186 PROTOZOA

pendently determined for mating type. Hence, the unit of mating
type inheritance is the caryonide, the descendants that carry
products of division of one original macronucleus. Since usually
two new macronuclei arise in a fertilized animal and one passes
to each of the two products of the first fission, a clone includes
two caryonides that arise at the first fission after fertilization.
Because of the independent determination of the macronuclei,
the two caryonides from a fertilized animal often are different
in mating type. In group B this also happens sometimes, but so
rarely as to be inconspicuous. For a detailed account of the
similarities and differences between the group A and group B
systems, as well as of a number of exceptional phenomena that
are omitted here, the reader is referred to a review by Nanney
(1954). Further details and a more general interpretation will
also be given below in other connections.

As a rule, the differences between the group A and group B
methods of determination and inheritance of mating type are so
striking that an unidentified strain could be assigned to one or
the other group on the basis of a study of half a dozen or so
pairs of conjugants and their vegetative descendants. It would be
required only to isolate the pairs, then to isolate the exconju-
gants, to cultivate separately the products of the first (better, the
first two) fissions, and finally to ascertain the mating types of the
various cultures. If the cultures from the two exconjugants of a
pair are usually of different mating types, the strain belongs to
a variety of group B. If the two caryonides from a single excon-
jugant often differ in mating type, or if the same mating type
frequently occurs among the descendants of both exconjugants
of a pair, the strain probably belongs to a variety of group A.
There is, however, one serious source of possible misinterpreta-
tion. In group B, if the two conjugants of a pair exchange much
cytoplasm, they may produce clones that arc alike in mating
type. This difficulty can be avoided by watching the conjugating
pairs as they conclude conjugation and eliminating those that
remain long united posteriorly after separating at the anterior
ends. Only those that remain long united posteriorly exchange
enough cytoplasm to affeel the inheritance of mating type.

T. M. SONNEBORN 187

Life Cycle Differences. Evidence for the existence of a life
cycle in Ciliates was first obtained in abundance by Maupas
( 1888, 1889 ) and was subsequently confirmed by many investi-
gators. But it was immediately challenged on theoretical grounds
by Weismann and eventually on experimental grounds by many
investigators. The chief objection grew out of the experimental
prolongation of the cycle as cultural conditions were improved
and the maintenance of some Ciliates in culture for very long
periods without indications of deterioration or ultimate extinction.
From my own observations on P. anrelia (Sonneborn, 1954a) and
my analysis of the reports of others on other Ciliates, I have con-
cluded that there is indeed a life cycle in these organisms but that
it may vary greatly in extent and in its manifestations in different
species and varieties. These conclusions are well illustrated by
the diversities in life cycle among the varieties of P. anrelia. The
same fundamental pattern may be seen in the life cycle of all of
them, but there is great variation in the duration of the various
stages and, in some respects, in the processes that mark them.

The life cycle, as Maupas discovered, begins with fertilization
and proceeds successively with immaturity, maturity, senility, and
death. Each of these stages will be taken up in order, and the
differences among the varieties will be pointed out. At the start,
it is necessary to emphasize that the course of the life cycle de-
pends upon the conditions of culture, and especially upon the
abundance or scarcity of available food. What happens at a par-
ticular stage may be decisively determined by this and by other
conditions which will be mentioned later. In order to follow the
entire potential cycle, it is necessary to maintain an excess of
food at all times, and this is best accomplished by the laborious
method of re-isolating single animals daily. On the other hand,
since depletion of food is so decisive and since this is doubtless
commonly recurrent in nature, to get a more complete picture of
the cycle it is also necessary to starve subcultures taken fre-
quently from the overfed daily re-isolation lines and to note their
responses to starvation. Complete studies of this sort have been
made on but few varieties, yet partial information is available
on all of the varieties. To the extent that is now possible, all this

188 PROTOZOA

information will be utilized in the following comparison of the
varieties of P. aurelia.

1. Fertilization. Varietal differences with respect to fertiliza-
tion are of several sorts: (a) the kinds of fertilization processes
that occur; (b) the tempo of the processes associated with fer-
tilization; (c) the environmental conditions required for the
occurrence of conjugation; (d) the number of interbreeding mat-
ing types; and (e) the number of micronuclei and macronuclear
anlagen formed from products of the fertilization nucleus. In
previous sections, (d) and (e) have been set forth and will not
be further mentioned.

(a) The kinds of fertilization processes. Basically there are
only two kinds of fertilization processes, conjugation and autog-
amy. Conjugation involves union of the animals in pairs. Each
mate undergoes meiosis and one of its reduced nuclei divides to
form a male and female gamete nucleus. The male nucleus of
each conjugant migrates into the mate with whose female nucleus
it unites to form a diploid fertilization nucleus. Autogamy in-
volves exactly the same nuclear processes as conjugation except
that cross fertilization is impossible because the animals are not
united in pairs. Instead, the fertilization nucleus is formed by
union of the male and female gamete nuclei of one and the same
individual. Both conjugation and autogamy can initiate a new
life cycle, but an appreciable period of immaturity never occurs
after autogamy. Conjugation occurs in all varieties. Autogamy
has been found in all varieties except varieties 13, 15, and 16. It
may yet be discovered in variety 13;6 but it probably will not be
found to occur as a regular normal event in varieties 15 and 16.

(b) The tempo of the processes associated with fertilization.
The time that conjugants remain united, i.e., approximately until
the time of the first division of the fertilization nucleus, varies

nversely with temperature. At any one temperature, the time
varies with the variety. Thus, at 22°, conjugants remain united
for about 7 to 8 hours in variety 4, about 12 hours in varieties 13
and 15. Correspondingly, the time from separation of the con-
jugants until the first fission is about 15 hours in variety 4, 24
to 25 hours in varieties 13 and 15. In both cases the interval is

T. M. SONNEBORN 189

much longer than an ordinary interfission interval. Also pro-
longed, but not so much so, is the time from the first to the sec-
ond fission. This is about 10 hours in variety 4, about 16 to 17
hours in varieties 13 and 15. These are probably the extreme dif-
ferences among the varieties. Similar but smaller differences may
well exist among certain other varieties. Information is at present
lacking on that point. In general the timing of autogamy seems
to be about the same as for the comparable nuclear stages of con-
jugation in varieties in which both processes occur.

(c) Environmental conditions for conjugation. The tempera-
ture optima and ranges for the occurrence of conjugation differ
for certain varieties (Sonneborn, 1938, 1939, 1941, 1956a, unpub-
lished; Sonneborn and Dippell, 1943, 1946). Varieties 2, 3, 11,
and possibly a few other less studied varieties, have relatively
low optimal temperatures for mating and conjugate with reduced
frequency or not at all at high temperatures within the tolerated
range. These varieties mate avidly at 19° or less, poorly or not at
all at 27° or more. Other varieties, such as 1, 4, 8, and 9 conju-
gate avidly at 27°, or even higher temperatures, as well as at
19°. Sonneborn and Dippell (1946) compared the sexual reactiv-
ity of representatives of several varieties over a wide range of
temperatures and found for each one a unique pattern of re-
sponses to temperature. Possibly extension of such a study to all
varieties would reveal unique features for each. Nothing is yet
known as to the possibility of strain differences in these respects
in a variety — like 1, 2, 15, or 16 — which occurs in nature over a
wide range of temperatures.

Visible light inhibits sexual reactivity in variety 3 (Sonneborn,
1938, 1939). Under the usual alternation of day and night, this
variety is sexually reactive only between about 1 a.m. and 1 p.m.
Diurnal periodicities also exist in some other varieties; the period
of reactivity is during the night in variety 2; from about 4 p.m. to
10 a.m. in variety 11; and from about 8 p.m. to 10 a.m. in va-
riety 15, peak reactivity after 11 p.m. (Sonneborn, 1956a; unpub-
lished ) .3 There are also indications of periodicities in varieties 12
and 13, but the details are not yet known. Some of these varieties
can become sexually reactive at any hour after they have been

PROTOZOA

^rown for some days in darkness, and they totally lose sexual
reactivity if grown several days in continuous illumination. Other
varieties show no diurnal periodicity.

Variety 16 conjugates better when grown in test tubes than
when grown in culture slides (Giese, 1939, 1957). Variety 15
shows similar relations.3 I have observed that animals from a
highly reactive tube culture lose their capacity to mate in an
hour or so after transfer to culture slides, while the animals left
in the tube remained highly reactive. These observations suggest
a decisive role of dissolved gases. Other varieties do not require
cultivation in tubes to become sexually reactive; they become
strongly reactive when grown in depression slides as well as
when grown in tubes. Perhaps, however, detailed study would
reveal quantitative differences among the varieties in this re-
spect.

None of the conditions which have just been discussed is known
to affect the occurrence of autogamy, nor are there known any
varietal differences in external conditions required to induce
autogamy. In the proper stage of life only starvation seems to be
needed in order to induce autogamy. More or less depletion of
the food supply is also necessary to bring about conjugation. The
same stimulus leads to these two different results at different
stages of the life cycle, as will appear.

2. Immaturity. Following the initiation of a new life cycle
by conjugation there may be a longer or shorter period during
which fissions occur, but during which the animals so produced
are unable to conjugate even when provided with mature poten-
tial mates under conditions most favorable for mating. This
defines the period of immaturity. It is doubtful whether the first
fission or two after conjugation should be included in this pe-
riod, for during them the nuclear reorganization consequent upon
fertilization is still in progress. If they be excluded, it can be
said that immaturity has never been found after autogamy, but
only after conjugation. Even without this exclusion, as has long
been known, animals of some strains can conjugate again while
still in the process of nuclear reorganization following the last
conjugation or autogamy.

T. M. SONNEBORN 191

An immature period appears to be totally lacking in varieties
4, 8, 10, and 14, and perhaps in some others. On the other hand,
some varieties are not uniform in this respect. Some strains may
totally lack an immature period and others may have a short
one, its duration varying from strain to strain. Thus, Siegel (un-
published) finds that the period of immaturity in variety 1 lasts
about four to five days in most strains, although in some strains
it may last eight or nine days. I was unable to find any immature
period in strain R of variety 1. Comparable variations among
strains seem also to occur in varieties 2 and 3. Translated into
fissions, these longest periods of immaturity represent about 35
or more successive fissions. Much longer immature periods occur
in varieties 15 and 16. Giese (1957) reports that some strains
of variety 16 did not begin to conjugate until they had been in
the laboratory for 6 months (fission rate during this period not
stated). In variety 15, I have thus far studied only a relatively
small number of crosses, but among them considerable variation
in the extent of the immature period was noted even among dif-
ferent caryonides of the same exconjugant. In some, it extended
for only 8 to 12 days, in others for a little more than two months.
However, these cultures were being grown slowly, and this repre-
sents a range of about 20 to 100 fissions or more. Curiously, there
were relatively few immature periods of intermediate duration.
This variability needs further careful study.5 There thus appear
to be at least three different conditions as to immaturity in
different varieties of P. aurelia: some have no period of immatu-
rity, some have variable but short periods of immaturity, and
some have much longer periods of immaturity, at least in most
lines.

3. Maturity. The onset of maturity is marked by the acquisi-
tion of the capacity to mate. When there is a preceding immature
period, the capacity to mate develops gradually, as evidenced by
mating reactions of increasing intensity, i.e., involving larger
proportions of the animals. When a period of immaturity is lack-
ing, maximal mating reactions occur as early as adequate tests
can be made, certainly by the 8th fission after fertilization. Of
course, in order to bring about mating, the food supply must be

192 PROTOZOA

more or less depleted and the other conditions necessary for
certain varieties, which were mentioned above, must be pro-
vided. The period of maturity terminates when the animals fail
to respond to these conditions by mating, but instead give one
of the responses characteristic of senility, which will be set forth
in (4) on senility. The transition from maturity to senility is not
sharp, but gradual, like the transition from immaturity to matu-
rity. It is marked by decreasing intensity of the mating reaction
under conditions most favorable for mating. Maturity will be
considered to have ended when conjugation with the comple-
mentary mating type of another culture becomes difficult or
impossible to obtain. The following account of varietal differ-
ences in the period of maturity takes up first the variations in
duration of this period and then variation in the system of mating
during maturity.

The period of maturity is exceedingly brief in some varieties.
In varieties 10 and 14 it lasts only until about the 15th fission and
in varieties 4 and 8 until about the 25th fission. Under the stand-
ard conditions of culture, the time from the end of postzygotic
nuclear reorganization to the end of maturity (no immature pe-
riod existing in these varieties) is about 3 days in varieties 10
and 14, about 6 days in varieties 4 and 8. Maturity lasts longer
in most varieties, much longer in some. It lasts until about the
10th to 13th day (30 to 40 fissions) in variety 2 and until about
the 17th to 25th day (50 to 75 fissions) in variety 1. According
to observations of Powelson (unpublished), it may last more than
30 days (more than 80 fissions) in variety 12. Current and still in-
complete observations on variety 15 indicate that it has a still
longer period of maturity. The first cultures to be followed are
still mature after three months since the end of immaturity.
Giese's (1957) observations on variety 16 indicate that its ma-
ture period may last more than a year. There are reasons to be-
lieve that the duration in time of the various periods of life
depends hugely upon the fission rate, the characteristic feature
being a certain average number of fissions. Eventually, therefore,
comparisons will have to be made in terms of number of fissions.

Varietal differences in the system of mating during maturity

T. M. SONNEBORN 193

concern the incidence of "selfing" caryonides. A selfing caryonide,
or selfer, is one in which conjugation can take place between
members of the same caryonide in the absence of mixture with
another caryonide. Selfers occur in all varieties, but with very
different frequencies. Kimball (1939) found that about 4% of
the caryonides of strain S of variety 1 are selfers, while Sonne-
born found a higher frequency in strain R of the same variety.
In the group B varieties, or at least many of them, selfers are
relatively rare in exautogamous caryonides that have not recently
been derived from conjugants, but are much more common in ex-
con jugant caryonides. Variety 5 (group A) presents what is thus
far a unique situation. Caryonides pure for either mating type
(IX or X) are relatively rare, the great majority being selfers. Of
the latter, most are predominantly type X. In these selfers, usu-
ally the great majority of the animals are of one type, so that
relatively few seLfing pairs are formed, and the predominant type
remains characteristic for any caryonide. The most extreme con-
dition is found in variety 13, every caryonide, so far as now
known, being a selfer. ,; Moreover, although I have demonstrated
that two mating types occur, the predominant type does not
remain the same from day to day in a caryonide. It seems as if
mating type is continually liable to change in both directions
during multiplication by fissions. Whether this inconstancy of
mating type appears at the very beginning of maturity or only
after a period of constancy as one type is not yet known. How-
ever, if there is a constant period, it must be brief. The incon-
stancy of mating type in variety 13 is the reason for present
inability to assign it to group A or B.

4. Senility. In varieties that undergo autogamy, senility be-
gins when the animals respond to starvation by undergoing
autogamy instead of becoming ripe for conjugation. Senility ex-
tends until death of the culture. At the beginning of senility,
fission rate is high and virtually all animals that undergo autog-
amy survive well. As senility progresses, the fission rate declines
and the frequency of survival after autogamy decreases. Even-
tually no animals can survive autogamy, and still later fissions
cease, the culture dying out. In variety 4, the period of senility

194 PROTOZOA

may extend through nearly 300 fissions or three to four months.
It is by far the longest stage of the life cycle. By the time it has
lasted for 170 to 200 fissions, no animals can survive autogamy.
During advanced senility the autogamy response to starvation
also declines and eventually disappears; concomitantly, the re-
sponse of conjugation reappears but in weak form. However,
survival after conjugation is low and approaches zero. The clone
thus becomes genetically dead about 100 fissions before it ceases
to exist. The length of senility is not greatly different in varieties
1 and 2 from what it is in variety 4, when measured in number
of fissions, but the lower fission rates of these varieties make it
longer in time.

The period of senility in varieties 15 and 16 is very different
from that in the other varieties. The response to depletion of food
is selling, not autogamy. The phenomena are most marked in the
usual kind of caryonide which is pure for one mating type during
the long period of maturity and then becomes a selfer during
senility. Caryonides that are selfers during maturity have not
yet been studied in this respect; but it is possible that there are
characteristically different features of selfing in the two periods
of life. During senility, it is impossible to isolate from a caryonide
individuals that yield cultures entirely or even predominantly of
different mating types. Moreover, as senility progresses, several
changes take place which do not occur during maturity. First,
the proportion of animals that self when the food supply declines
increases slowly but steadily. Second, the amount of depletion
in food necessary to bring about selfing decreases. Third, the
percentage of viable clones produced from selfer pairs decreases:
at first survival is high, eventually it decreases to zero. Giese
(1957) describes these changes which he had observed in va-
riety 16 as early as 1941. My recent observations on variety 15,
though far from complete, agree as I'ar as they go with his
observations on variety L6.3

Of particular interest are two strains of variety 15 which are
very old but which came into my hands only a few months ago.
One strain came from the University of Washington about 4 years
ago through Dr. Vance Tartar. The other came from Stanford

T. M. SONNEBORN 195

University where it was isolated about 1941. This 15-year-old
culture is the well-known strain long maintained in axenic cul-
ture under the name of P. multimicwnucleatum by Dr. Willis
Johnson (1952), who gave it to me. Both of these strains I
found at once to be predominantly mating type XXIX but selfers,
and in both all exconjugants died. Dr. Ruth Dippell stained
these for me and found, as I suspected, that both are now ami-
cronucleate. These strains are thus genetically dead. How long
they have been so, I do not know. But it is clear that the total
life cycle in variety 15 is at least 15 years and that life persists,
perhaps very long, after a culture has become genetically dead.
As mentioned above, the same sort of relation exists on a smaller
scale in varieties with shorter life cycles, e.g., clones of variety
4 may live for about 100 fissions after they are unable to survive
sexual processes. In both cases, deterioration and loss of the
micronuclei, the germ plasm, mark genetic death which precedes
somatic death by a considerable period. Whether variety 15 and
variety 16 clones ever reach the stage of somatic extinction is as
yet unknown. If they do, it must be after ages measured in dec-
ades and therefore rarely if ever discovered by an investigator.

Should the Varieties of P. aurelia Be Assigned Species Names?
The preceding detailed account of the known and, so far as can
now be foreseen, the knowable differences among the varieties
of P. aurelia had as its first objective marshalling all evidence
which might be pertinent to the question: Can the varieties of
P. aurelia be readily identified? A positive answer to this question
would justify changing the status of the varieties to species and
assigning to each a species name. A negative answer would seem
to require that we abstain from assigning species names even if
we regard the varieties as species.

So far as now known or now foreseeable, all the 16 varieties
could be uniquely identified either by mating type tests alone
or by serologic means alone, but not by any other single charac-
teristic. To carry out the mating type tests, it is necessary to
employ successfully standard cultures of each of the 34 mating
types. This involves having available enough strains to yield
all these types, being able to obtain and isolate the various mating

196 PROTOZOA

types, and having the knowledge and skill required to get all
the types in sexually reactive condition. This is a colossal task
and beset with many difficulties. Beyond question, it is too
much to expect of a zoologist who merely wishes to conform
to the mores by attaching the correct species name to the or-
ganism about which he wishes to publish his observations. The
only alternative is for him to send his culture to an expert for
identification. Although this could be done now — it is no easy
job even for the expert and could take weeks or months of work
in the case of varieties with long immature periods — the solu-
tion is one of limited dependability. There is no guarantee that
there will always be a laboratory maintaining a full battery of
live standard cultures. If identification depends upon such a
hazardous procedure, this basis for setting up species is much
too insecure. Type specimens would have to be kept perpetually
alive. I therefore reject mating type specificity as a basis for de-
fining named species.

At present, serologic specificity cannot serve the purpose either.
However, this may be largely due to insufficient knowledge and
to unavailability of enough diagnostic antisera. If these lacks
were ever supplied and the diagnostic antisera were routinely
maintained by a number of established institutions well dis-
tributed around the world, perhaps identification of the varieties
might become practicable and within the limits of routine taxo-
nomic procedure. The situation would then be comparable to
the one that now exists for certain microorganisms such as Salmo-
nella. But P. aurelia is not Salmonella. There is no reason to be-
lieve that the demand for these means of identification would
ever justify the great labor and expense of maintaining a supply
of the 300 or more necessary antisera. I therefore also reject
serologic characteristics as a basis for defining named species
in this material.

To aid in judging whether there arc any combinations of
characteristics which could routinely be used to achieve success-
ful identification, Table III presents the differential traits of the
varieties in summary fashion. Most of the differential characters
involve for their recognition prolonged study and highly elabo-

T. M. SONNEBORN

197

so

so

e

Oh

co

-^>

.2

>

CO

i— i

cu
+3

CO

o

CO

"s-i
cu

o

o3

O

o

CO

►J

«

o

-^>

_cu
>

V.

i
<

B

<

v

a

-.

iS

-a
si

a

-a!

SI
M

oj o3

as s

s- - -

S

Cj£
SB

"T3"o -c m
OU O V

c3 c3 ai

c

03 o O

. 1 t-» 1 — 1

00 g

.5 -<j

I <

00 — 1 y.

is

7.

CM I •<

v E;2 x

SZ5 X

.0

is ^

ocoo

iS is

Z rt-

c — n .

. < g 03 03 o o

< is x a ^ £ ?. ° x.

00

o-< o v i-J' — ;

G — i. -73 ^

:' =: :° V

O

&3iS

000

— O00 >_

<■ CM CO VS

+ +

o m

00 «e

-1 M CO

2 + +

-r r < x 8 5 x

"5 ^

O O CO — "*,

MNP!

+ +

OO .

ccoo ■<

NOfP OO whi;

CM s_ - <| <r, /. ,-1 -r <

;S tiifflffl^ 00 coS -»!

_. ^ ~* ^s ' *" ^f

is

in - e e ^. 0001

o

. 10 «t

sSan ^ 66 62 <j

n ^ ^ ^ A ' °°^

C
S3

ft

B

S M »

h:e

03 M

Q.S

GO

^^

to

j

-c

^0

03

^3

CO

0

—

(3

0

J5

CO

«

0)

5

c

03

0 CJ

+-•

—

>1

^

"*

"u

0

0>

V £}

fc£

»*1

OQ

^ Z

u

J

C *-
•- u

03'E

i- —

^ E

03 -

"5 «

P

X

-

—

-

s

*rH 03

l-l

§

*—
^

X

M

<.

0J

6

t^

06

r!

3

-

01

CO

-r

x r; ^ M C l.^

ft^ S . 0 o?"E'S

O = >. S 03 ■

n t C 3

-

« at" O'

«-£g2gfe3.2g
aj£«o3o3ai'-c2

G«Orf03+^--3o3c3
' ^"O »-? C « co

C , "" .. 4) ^ o =

J S - «^CQ g-M

CO- C fr, X^ r-C^g

:^ fl fi §

1 fto:

rf B S)

g 0J-H j. --W j,

J- ^, c -d ^ o3 ca

O g °
- Mi'
fc 2 * ~ " "ai S e

CJ O 0)y « 3

. • -tj 03 o3P--- ..j- £
X'S^-2 . dfL, t,

C030Of;i -/

o3-- a cjo a 0_» c

Oc

cu o "

6
u

0

n

CU

-

>.

X

^

:
0

=

CD
—

a

a

CU

"H

03

CU
CO

£
0

0

is

fi
C
CU

cu

CQ
CU
■-

CU

CO

CU

CU

CU

T3
CU
+3

CO"

"d a
0 t.

C

<

T3

CO

I

c

CU

03

B

—
/.

CU

-

B

B

Q
co

1

0

t-

CU
—

13

CO

~

^J

--

3

-.

CU

c

'_

~

0

C

—

N

a

73

4J

+^

<-^-

—

O

-

a

-1

-

—

a

S3

0

~z

CO

-f^

/

—

•S?t)J2

■2 o o e S ciK a;
c ai S. S ° •— 03

"coos-;-:

>.

r.

cu

-T)

go cd P5 S =- 32 -

Si<^ • r-— n

a— '

cu

03 —

cj Ox m ^

Sis.

• o a

,-- o
! 03 '-P

(- 03

3a

03 03 cic^^ ^ 0-tOJT4

198 PROTOZOA

rate and technical procedures: obtaining conjugants, following
the life cycle, genetic studies, serological procedures, and other
physiological procedures. These are totally unsuitable for use
in routine identification. Mean body length alone, among all the
traits listed, could be used routinely. And this does not identify
all the varieties; it could only limit an unknown to one of three
or four groups of varieties. I therefore also reject combinations
of the traits in Table III as a basis for assigning species names
to the varieties.

Nevertheless, the information summarized in Tables I, II, and
III make it possible to define the varieties and to prescribe a pro-
cedure for identifying them without recourse to living standard
cultures. Sonneborn (1950) explained this in detail for the 8
varieties known at that time. The same general approach, sup-
plemented by much new information, would make possible
identification of all 16 varieties now known. The first step would
be to obtain many strains from regions that have provided the
known varieties. From each strain mating types would have to
be isolated and, in the way originally done, the various strains
would have to be sorted into varieties. When 16 have been ob-
tained, they could be identified with the 16 now known in the
following ways. The only variety with 4 mating types would
be variety 16.3 Varieties 2, 3, 11, and 15 could be identified by
their distinctive diurnal periods of sexual reactivity and their
sizes. The remaining large animal varieties would be 12 and 13.
Variety 122 could be identified as the only large variety that regu-
larly formed 2 macronuclear anlagen and 2 micronuclei after
fertilization. Variety 13 could be identified by its regular selfing.6
Of the 4 smallest animal varieties (4, 8, 10, and 14), 8 could be
singled out because of the reaction of one of its mating types
with one of those of the already identified variety 3. Variety 4
would then be identifiable by its two-way sexual reactions with
variety 8. The remaining two varieties ( 10 and 14 ) have been but
recently discovered and I do not yet know enough about them to
prescribe how to differentiate them. The system of reactions in
Table I shows how variety 3 could be used to identify varieties

T. M. SONNEBORN 199

1, 5, and 7. This leaves only varieties 6 and 9. These could be
identified both by their nonoverlapping distributions and by the
fact that 6 shows group B inheritance of mating type, while 9
shows the group A pattern. Thus, without either living standard
cultures or diagnostic antisera, the complete identification of all
varieties (except 10 and 14) could be accomplished.

It is a great satisfaction to know that now or centuries from
now, anyone who is willing to give the time and effort needed
for the job would be able to identify strains in conformity with
our descriptions. This, however, would involve a major research
effort over a period of years. I think every reader will agree that
this is too much to require of anyone who merely wishes to
know to what variety the one stock on which he has been work-
ing belongs. To make matters worse, it seems all too likely that
we do not yet have in hand by any means all the varieties of
P. aurelia that exist in nature. Until we do, there is always the
chance that anyone who attempts to follow the procedure out-
lined above will come out with 16 (or more) varieties some of
which are different from those we have tried to define. Under
such unhappy conditions, I contend that assigning species names
to varieties of P. aurelia is indefensible because it is totally im-
practicable.

Should the Varieties of P. aurelia Be Referred to as Unnamed
Species? Even the most devoted adherents of the modern bio-
logical species concept would probably agree with the con-
clusion that the varieties of P. aurelia should not — or at least not
yet — be assigned species names; but they doubtless would also
tenaciously hold that the varieties are species nevertheless and
should be referred to as numbered species, not numbered vari-
eties. This is a much more fundamental and important question
than the one considered in the previous section. The answer to
be given to it must be based on reason and judgment. Let us
review the major considerations pro and con.

The chief argument for the modern biological species concept
is that it refers to a real level of biological organization, a group
of populations among which gene flow is potentially possible.

PROTOZOA

This contrasts with the typological or morphological species con-
cept which is to a degree arbitrary, depending upon what traits
the taxonomist can recognize and what ones he believes to be of
taxonomic importance. This arbitrariness, the existence of inter-
grades, and other difficulties have often led to the conclusion
that species so defined are but figments of the imagination and
have no objective validity. The definition of species in terms of
a potentially common gene pool provides an objective criterion
which has the great merit of delimiting a natural entity with
evolutionary significance.

On the other hand, there are strong objections to this other-
wise attractive point of view. First of all, the task of discovering
what constitutes a potentially common gene pool and the rela-
tion of such groups to morphological species is very great. For
example, after 20 years of research on P. aurelia, we know it
consists of at least 16 distinct potentially common gene pools
and suspect that very many more remain to be discovered. Ob-
viously only a very small proportion of the enormous number of
morphological species can in the foreseeable future be sorted out
into common gene pool species. This means inevitably that there
will always be a double standard of species, relatively few ever
being defined on the modern biological species concept. Second,
as pointed out at the beginning, a very considerable proportion
of organisms, those which are obligatory self-fertilizers and those
which totally lack sexual reproduction, are totally outside the
possible domain of the modern biological species concept. Po-
tentially common gene pools simply have no meaning in connec-
tion with them. This leads logically to the view that such organ-
isms cannot be sorted into species. Proponents of the modern
biological species concept, when faced with this situation, retort
that difficulties and exceptions do not destroy the value of a
concept, but this is not the point at issue. I for one certainly do
not challenge the value of the biological species concept. It has
great value. Hut its value is very narrowly limited, in the first
place to outbreeding organisms in principle and in the second
place to the very small proportion of them that will in the fore-
seeable future be studied sufficiently. What this boils down to is

T. M. SONNEBORN 201

a division of opinion on the most desirable usage of the techni-
cal term species. Until recently it has been used as a universally
applicable designation for pigeon-holing all organisms. Propo-
nents of the gene pool species concept have used the term in a
second sense which is inapplicable in principle to a large pro-
portion of all organisms and which can strictly be extended to
but an infinitesimal fraction of the eligible organisms because of
the great labor involved in applying it. The last difficulty is
minimized on the assumption that taxonomists have probably
already defined most species of outbreeders in such a way that, if
breeding studies were made, the taxonomists' species and the
biological species would be found to be identical. However, in
the absence of breeding studies, this remains only a more or less
reasonable conclusion. In any case, the gene pool species is more
restricted in applicability than the species of routine taxonomy.
Many taxonomists are unwilling to abandon the universal use of
the term for a limited use, and I agree with them.

The alternatives seem clear. (1) Accept the double standard
of applying the biological species concept whenever information
justifies it and using routine taxonomic procedure in all other
cases, i.e., in the great majority. (2) Restrict the term species to
the limited group and find a new term for the majority. (3) Con-
tinue to use the term species in its universal pigeon-hole sense
and find a new term for the limited number of cases in which
common gene pools are known or will be discovered. I reject the
first alternative on the ground that a technical term should have
a single usage, and the second alternative on the grounds of pri-
ority and generality. Accepting, therefore, the third alternative,
I propose the term "syngen" for the potentially common gene
pool, for organisms capable of "generating together." I recom-
mend that it or a better term, if one can be found, should be
adopted. I am therefore prepared to abandon the use of the term
variety in P. aurelia in favor of syngen or whatever other new
term meets with the approval of the experts. But I am not will-
ing to abandon the term variety in favor of species, for the
reasons just given.

202 PROTOZOA

Paramecium aurelia: Breeding Systems and Evolution

Disagreements concerning terminology must not be permitted
to obscure the biological problems that underlie them. The ques-
tion discussed at the end of the preceding section — whether the
varieties of P. aurelia should be called species or should be given
some other designation such as syngens — is of relatively minor
importance. More important are the evolutionary questions raised
by the recognition of these biological units of organization and
by the nature of the diversities among and within them. These
are the questions to be dealt with in the present section. In doing
so the biological units will have to be referred to by some term.
I shall continue to call them varieties so that the reader will not
have to change horses in the middle of the stream. However, I
urge that a term such as syngen eventually be agreed upon both
for the varieties of Ciliates and for the so-called biological species
of other organisms.

Basic Biological Frame of Reference in P. aurelia. Probably
the most fundamental evolutionary requirements of any organ-
ism are to take full advantage of opportunities for reproduction
and to possess and maintain mechanisms for providing enough
genetic variability to meet the demands imposed by changes in
the environment. The common mechanism for the latter is of
course mutation. But organisms differ in the way mutations are
handled. Among lower organisms, two major ways are observed.
Some combine haploidy with large and rapidly produced popu-
lations. This permits both an adequate supply of mutations and
an effective way of bringing them to immediate or rapid expres-
sion. Other lower organisms, like most higher organisms, combine
diploidy with genetic recombination by sexual means. This per-
mits storing an accumulation of mutations and selecting among
their various combinations. The latter is the alternative generally
adopted by Ciliates, including Paramecium. It is further rein-
forced by the occurrence of asexual reproduction between sexual
processes and by the existence of nuclear dimorphism. The new
mutations are stored in the micronuclens, which has little or no
effect on the phenotype, and they are multiplied during asexual

T. M. SONNEBORN 203

reproduction so that they become widely distributed prior to
recombination.

In Paramecium, and in Ciliates generally, the occurrence of
sexual processes and genetic recombination is linked to decline
in the food supply. This must be an inevitable and frequently re-
curring event in the lives of such organisms. The logarithmic in-
crease in numbers resulting from repeated binary fission quickly
yields populations that no conceivable food supply could support.
Not even an amount of food equal to the bulk of the earth could
support reproduction from a single individual for as much as
100 successive fissions. With due allowance for predation and
other causes of accidental death, it is difficult to see how asexual
reproduction could long continue without the onset of famine.
The prevailing pattern of life must be to increase, multiply, and
spread while the harvest lasts, forgetting about sex while enjoying
food to the full, and to use hard times when reproduction comes
to a virtual standstill as occasions for grasping opportunities to
carry out sexual processes.

Little is yet known about the alternation of asexual reproduc-
tion and sexual processes in nature. We are at present largely
limited to inferences based on laboratory study. What is worse,
we systematically carry out the laboratory studies under con-
ditions that differ in important respects from those which prob-
ably exist in nature. For example, we routinely and at once trans-
fer fertilized animals from the starvation conditions that induced
fertilization to conditions of excess in available food. In nature,
the conditions that induce fertilization would hardly be expected
to change so radically in the course of the hours occupied by the
processes of meiosis and fertilization. Yet we know very little
about the effects of this difference between natural and labora-
tory conditions. That it does have some effects on the nuclear
processes following fertilization and even on the subsequent
breeding relations is already indicated. The whole matter is in
great need of investigation.

Nevertheless the laboratory studies seem clearly to show that
there are two distinct and independent functions of sexual proc-
esses. First, they have the usual function of genetic recombina-

204 PROTOZOA

tion. Second, they serve to initiate new life cycles. In the absence
of sexual processes, paramecia grow old and either die or become
genetically extinct by losing their germ plasm ( micronuclei ) .
Survival depends upon the occurrence of fertilization before
senility has gone too far. The fertilized animals are reinvigorated
and start new, young clones.

Different sexual processes may occur in different stages of the
life cycle when food is depleted. This sequence of responses to
famine is a clue to the relative values and functions of the differ-
ent sexual processes. Invariably, the response of ripeness for con-
jugation precedes the response of autogamy when both occur in
the life cycle. In varieties that undergo selfing instead of autog-
amy, ripeness for cross conjugation precedes ripeness for selfing.
Cross conjugation thus has priority over selfing, and conjugation
has priority over autogamy. These invariable priorities indicate
the judgment of natural selection as to which sexual processes
best fulfill their functions. Since all are equally operative in initi-
ating new life cycles, the advantage of conjugation must lie in its
greater effectiveness in bringing about genetic recombination.

A priori it might have seemed possible to question this verdict
of natural selection. Autogamy results in conditions approaching
those associated with haploidy: there are no hidden recessives
masked by dominant alleles; all loci are homozygous in the one
case and hemizygous in the other. Selfing also approaches the
same condition, but less efficiently. Why then would not autog-
amy or even selfing work as well as haploidy? The answer lies
in the decisive differences from the conditions in haploids. In a
diploid with autogamy, the clone is the unit of expression of a
genotype, not the individual animal. This results in a considerable
lapse between the occurrence and expression of a mutation. More
important, the number of units or clones in a population must
be exceedingly small in comparison with the number of units or
individuals in the usual successful haploid organism. Hence,
neither selfing nor autogamy in Paramecium can be as efficient
as haploidy in large and rapidly produced populations. To achieve
the necessary genetic variability, natural selection has in P.
aurelia given first priority to cross conjugation, which however

T. M. SONNEBORN 205

cannot be guaranteed to occur; and second choice to genetically
less satisfactory, but physiologically equally satisfactory sexual
processes, which can be completely depended upon to occur.
The second choice is essential to short-term survival; the first
choice may be essential to long-term survival. The universality of
the occurrence of a first and a second response in time indicates
that this is not a mere laboratory phenomenon, but a double
assurance mechanism which operates in nature and has been de-
veloped and preserved by natural selection. Incidentally, if this
is so, the second choice response must occur often enough in
nature for selection to be able to preserve it.

However, the existence of these two sorts of responses in all
varieties of P. aurelia does not mean that they are equally well
developed in all. As will appear, most of the varietal differences
we have set forth in an earlier section bear importantly on just
this point. They have the effect of giving greater or lesser em-
phasis to one or the other of two successive sexual responses to
famine. The life processes of some varieties favor and prolong
the possibility of cross conjugation, making them outbreeders.
The life processes of other varieties contract these possibilities
almost to the vanishing point, making these varieties inbreeders.
And other varieties show intermediate conditions. Since out-
breeding may be considered advantageous for long-time survival
of diploid free-living organisms, it may be assumed that this
was the breeding system of the ancestors of the P. aurelia now
living. Further, those living varieties that now maintain the out-
breeding habit are most likely to provide the solid core of per-
sistence on the evolutionary time scale. Relatively few varieties
of P. aurelia conform to this exacting requirement; but they are
the ones with the widest distributions in nature. On the other
hand, inbreeding suffices well enough for short-time survival
while conditions remain relatively constant. This mode of life
is less exacting. The life cycles succeed one another rapidly so
that the system approaches the one in haploids. Evolutionary
divergence can and does take place at a rapid rate; but at the
sacrifice of some of the genetic plasticity needed for rapid adap-
tation. These are the views I shall now attempt to develop in

206 PROTOZOA

detail by considering the bearing of varietal differences on breed-
ing systems and evolution.

Varietal Differences in Breeding Systems. As just mentioned,
most of the varietal differences are related to their breeding sys-
tems. Aside from this relation, they would appear, as they have
until now, as mere chaotic brute facts of nature. Recognition of
their relation to breeding systems makes them fit together, like
the pieces of a jigsaw puzzle, into a meaningful picture. The
nature of the relations will be best illustrated by contrasting first
the most extreme inbreeders with the most extreme outbreeders,
even though less is known about them than about some of the
less extreme varieties. I shall then review the situation in the in-
termediate varieties progressing from those that appear to be
least, to those that appear to be most, committed to inbreeding.

1. The extreme inbreeders, varieties 10 and 14. These two va-
rieties have been found only once each in nature, in the south-
ernmost part of the United States. They are both apparently rare
and highly localized. This alone would limit narrowly their pos-
sibilities for cross breeding. Everything else that is known about
them points in the same direction. They have no detectable im-
mature period and the shortest known mature period which ex-
tends to about the 15th fission after conjugation or autogamy.
This means two things: they are able to cross-conjugate immedi-
ately after they complete the reorganization following fertiliza-
tion; but only for a short time thereafter. After that the response
to depletion of food is autogamy, the closest possible form of
inbreeding.

In estimating the chances for cross conjugation during the
short period in which it is possible, several considerations are
pertinent. The processes of meiosis and fertilization and the sub-
sequent process of nuclear reorganization are relatively rapid in
these varieties, and they afford little time for dispersal while they
arc in progress. The organisms occur in warm regions, and this
also hastens the processes. The fission rates are among the highest,
and this would lead to the most rapid exhaustion of available
food and cut down on time available for migration before star-
vation .main leads to fertilization. These varieties belong to group

T. M. SONNEBORN 207

B. Usually, therefore, each pair of exconjugants produces two
clones of complementary mating types. Since the factors pre-
viously mentioned operate to make the animals quickly ripe for
mating, this will happen when complementary mating types of
common descent have had little time to migrate. They will there-
fore be apt to mate with each other, if they find mates at all.
Failing this, they would have little opportunity to find mates
other than those available from the same population. Conjuga-
tion with individuals that are not closely related is thus almost
impossible except as a rare event; conjugation with close rela-
tives or inbreeding by autogamy would be expected as the almost
universal forms of fertilization. Possibly in a sparse local popula-
tion provided with much food, maturity could be passed before
starvation occurred and induced autogamy. The same might oc-
cur in a sparse population with little food, as a result of the
difficulty of finding mates. Only in a dense population would
conjugation be likely on a large scale. The various characteristics
of these varieties would then conspire to bring about mating be-
tween close relatives or within the population.

2. The extreme outbreeders, varieties 15 and 16.3,5 These vari-
eties are widely dispersed from cold to warm regions of North
America; they have not yet been searched for elsewhere. There
are thus many local populations, the first prerequisite for out-
breeding. Most important is the fact that these are the varieties
with the longest known periods of immaturity, lasting for months,
perhaps for a year, at the low reproductive rates that may usually
prevail in nature. This gives time for the animals to scatter and
migrate. The period during which they are most likely to be
near close relatives is the period in which they cannot mate. As
a corollary, the method of mating type inheritance is less im-
portant when there is so long a period of immaturity. Even if
complementary mating types arise among the descendants of a
single pair of conjugants, they are not so likely to meet each other
when they are mature if there is an intervening long period of
wandering about. Actually the method of mating type inheritance
is unknown in variety 16; but the group A method is indicated in
variety 15.

PROTOZOA

The period of maturity is probably even longer than the im-
mature period in these varieties. An outbreeder, especially one
with no known self-controlled method of getting from one body
of water to another, doubtless needs much time to find a suitable
mate. The chance of finding a suitable mate is increased in va-
riety 16 because there are three other mating types. In all other
varieties, there would be only one. Obviously if the chance of
finding any stranger belonging to the same variety is low, there
is a selective advantage in a system of multiple mating types
which greatly increases the probability that the first stranger en-
countered will, if mature, be a suitable mate. No such advantage
attaches to multiple mating types in inbreeders. Lacking imma-
turity and having either the group A or B method of mating type
inheritance, they would still have prospective mates at hand
even with a two-type system.

Two considerations apply to both the periods of immaturity
and maturity. First, varieties 15 and 16 have lower fission rates
than varieties 10 and 14. This makes both immaturity and
maturity last longer in time than is indicated by merely com-
paring the numbers of fissions in these periods with those in the
inbreeders, for at low fission rates it obviously takes longer to
use up the number of fissions allotted to each stage of life. The
times available for migration and for finding a suitable mate are
thus to this extent increased. Second, various collectors (e.g.,
Pringle, 1956) have reported that some varieties of P. aurelia
can be found in nature only during summer and fall, in certain
regions, and that the numbers found in samples of equal volume
have an annual rise and decline. In varieties 15 and 16, with im-
maturity and maturity lasting through one, two, or more such
cycles, this might drastically reduce the number of caryonides
inhabiting a particular body of water and so would perhaps re-
duce the probability that caryonides of complementary mating
type and common descent would persist in the same body of
water until they reached maturity and attained a sufficiently high
population density to be likely to encounter one another.

Finally, varieties 15 and 16 arc the only ones in which autog-
,ini\ appears to be replaced by old-age selfing. Thus, even this

T. M. SONNEBORN 209

second choice method of fertilization, which is required to fore-
stall clonal death, takes a form in which inbreeding is less in-
tense. There is no published evidence to decide with assurance
whether immaturity follows old-age selfing, but Giese's (1957)
comment about obtaining a series of successive inbreedings in
variety 16 following selfing suggests that immaturity after this
sort of conjugation may be lacking as it is after autogamy. If so,
this would permit rapid restoration of the heterozygosity lost by
selfing. The episode of selfing would thus serve to rejuvenate
and give more time for crossing without sacrificing heterozy-
gosity.

The contrast between these two inbreeding and two outbreed-
ing varieties illustrates how differences in some of the major
life processes bear on the breeding system. The same features,
with mainly quantitative differences, and some additional ones,
are encountered in the other varieties, as will now appear.

3. Variety 12. Very little is yet known about the only avail-
able strain of this variety. It has an immature period of unde-
termined length and a long mature period that probably extends
for more than 80 fissions. Yet its fertilization process during
senility is autogamy. Thus it seems to be a somewhat less extreme
outbreeder than varieties 15 and 16.

4. Varieties 1, 2, 3, 7, 9, and 11. These varieties are all less
extreme outbreeders than varieties 12, 15, and 16, and less ex-
treme inbreeders than varieties 10, 14, and those to be discussed
later. However, not enough is yet known about them to rank
them in a definite order. Little is known about variety 11, of
which only two strains have been found. It belongs to group A,
has a moderately short period of maturity, and undergoes autog-
amy. Immaturity has not been studied in it; but the relatively
short total interval until senility begins precludes any but a
very short immaturity. Variety 11 could not be a strong out-
breeder. Variety 7 has long been known only as a single strain
pure for mating type XIII. Recently a few more strains have
been found. These are not pure for mating type, but show the
group A method of mating type determination. All come from
Florida. Preliminary observations indicate no immaturity and

PROTOZOA

moderately short maturity. The little information available may
appear to be somewhat conflicting. Purity of a strain for one
mating type limits its conjugation to outbreeding. Yet it can
live on indefinitely without conjugation. I have had this strain
for 18 years during which it has undergone successive autogamies
at intervals of some 30 to 40 fissions. It is still as vigorous as ever.
Thus, purity for one mating type in a strain that undergoes autog-
amy has double significance: it conjugates only by outbreeding,
but without conjugation it undergoes the most intense form of
inbreeding. Further, the purity for type XIII depends upon a
recessive gene. Therefore, once an outbreeding has taken place,
this purity disappears except in segregants in later generations,
and these at once have available for inbreeding plenty of re-
lated segregants of complementary mating type. On the whole,
therefore, purity for one mating type in this context cannot be
considered as an important outbreeding feature. All other avail-
able facts (group A, no immaturity, short maturity, narrow
range ) indicate that variety 7 is prevailingly an inbreeder.

Variety 9 has been found only in western Europe. It belongs
to group A, has a moderately short mature period, with senility
beginning by about the 50th fission, and it undergoes autogamy.
With so short a period until the onset of senility, immaturity, if
it occurs, could only be brief. These features point to an in-
breeder. In this case, the inference can be supported by analysis
of Pringle's (1956) field studies. He made collections from a
small part of a large pond at approximately 20-day intervals over
a period of 15 months. At each collection, 24 to 55 samples of
30 ml. each were brought into the laboratory and searched for
variety 9. Not more than one animal of this variety per sample
was studied. It and its descendants were tested for the alleles
present at two loci. At each locus three alleles were found, but
one was very rare. The results on the two loci were similar and
may be combined. For the two principal alleles at each of these
loci, 151 homozygotes and 7 heterozygotes were found. At one
of the two loci, no heterozygotes were found. In a randomly
mating population, altogether 116 homozygotes and 42 heterozy-
gotes would be expected. Pringle cautiously remarks that his de-

T. M. SONNEBORN 211

tection of heterozygotes might have been imperfect, so that his
figure is to be considered minimal. Knowing the methods em-
ployed, I believe the error cannot be very great. Even if it is
assumed, as is by no means justified, that only one-fourth of the
heterozygotes were identified, their deviation from the expecta-
tion on random mating is still significant at the 1% level of
confidence. Accepting the data at their face value, it can be cal-
culated that at least 78.5% of the homozygotes arose by inbreed-
ing. Altogether I consider these field results to constitute strong
evidence for the occurrence of considerable inbreeding in nature
in variety 9.

The preponderance of homozygotes could be due to conjuga-
tion between relatives or to autogamy or both. Some calculations
which seem to bear on this can be made from data given by
Pringle ( 1956 ) in connection with the observations mentioned
above. Applying the Poisson formula to the proportion of 30-ml.
samples which contained no animals of variety 9, the peak den-
sities, found in late summer, amounted to only 25 animals per
liter or 1 animal per 40 ml. This calculation is surely subject to
some error for it assumes unjustifiably that the animals were
randomly distributed in the area sampled; but it is a good first
approximation, in spite of the occasional finding of local aggre-
gations by other collectors. During a considerable part of the
year, winter and spring, few or no variety 9 animals could be
found. With such low population densities, even at their peak,
the chances of animals of complementary mating type meeting
and mating would seem to be low except in so far as local aggre-
gations occur. Yet the occurrence of some heterozygotes proves,
as Pringle points out, that conjugation must occur sometimes be-
tween animals of different clones. The low population densities
and the excess of homozygotes on the other hand suggest that
autogamy probably also takes place. As mentioned earlier, the
invariable existence of autogamy or selfing during senility in
the laboratory indicates that there has been strong natural selec-
tion to maintain them, and this could be true only if the processes
do occur in nature.

The preceding analysis applies only to what goes on in a

PROTOZOA

single pond. Pringle's data also throw some light on what goes on
1 the range occupied by variety 9. The pertinent data appear in
his table that gives information on segregation of serotype alleles
in the F2 by autogamy after crosses between homozygotes for
different alleles. Here the proportion of nonviable F2 clones is
the significant information. A cross between a Scottish and
French strain gave 16.9% F2 mortality. Crosses between Scottish
strains from different sources gave 7 to 20% F2 mortality, mean
13.4%. Crosses between different isolates from the one intensively
studied pond gave 2 to 10% F2 mortality, mean 6.3%. As far as
they go, these data indicate that F2 mortality is less following
conjugation within a local population than following conjugation
between different local populations. This implies to a certain de-
gree that inbreeding has a selective advantage over outbreeding.
But the impediment to outbreeding imposed by F2 mortality
is much less in this material than in some varieties to be dis-
cussed later. Taking all the facts into consideration, variety 9 ap-
pears to inbreed to a marked degree, even within a single local
population; but it is capable of and apparently undergoes some
small measure of outbreeding, at least within the confines of a
local population.

Varieties 1, 2, and 3 show several features in common. All of
them show variations among strains with respect to the occur-
rence and duration of the immature period, from none at all up
to about a week or more. There is also some variation in the
duration of the moderate mature period, up to about 20 days or
about 75 fissions. These periods appear to be somewhat shorter
in variety 2. All undergo autogamy when starved during senility.
These common features indicate that inbreeding and outbreeding
occur to different extents in different strains of each of these
varieties. The main variable involved is the immature period. In
its absence, inbreeding is favored; in its presence, outbreeding is
favored in proportion to its duration. Each of these varieties thus
appears to be in process of evolutionary divergence with respect
to its breeding system and it may become necessary to subdivide
each variety into groups of strains differing with respect to the

T. M. SONNEBORN 213

immature period in order to obtain a satisfactory picture of the
breeding relations.

The diversity of conditions within variety 1 may here be noted
also for certain other features. In one stream in Woodstock,
Maryland, I found many years ago a strain pure for mating type
I. No such strain has been found among the many collections
of variety 1 from the rest of the world, including Asia, Aus-
tralia, South America, and Europe. Its significance for the breed-
ing system is the same as set forth above for the pure strain of
type XIII in variety 7. Of special interest, however, are two
further pieces of information. First, I collected from this same
source several times over a period of years and always found
there this same strain and no other variety 1 (although variety
2 was also always found with it). This shows that the strain can
and does persist for years in nature by repeated autogamous
inbreeding, a conclusion supported by 18 years of maintenance
in the laboratory. The same facts also indicate how constant a
local population can remain for a long period. Second, Beale has
examined the serotype loci of representatives of this population
and finds them to be entirely homozygous at all three major sero-
type loci. This is in marked contrast to his observations of hetero-
zygosity at these loci in representatives of other populations that
contain both mating types. It is therefore clear that within variety
1 there are populations that inbreed closely and almost exclu-
sively by autogamy and other populations that at least occa-
sionally undergo conjugation. The constancy of the Wood-
stock population further suggests that the conjugations in other
local populations are probably as a rule within the population.
Migrations and interbreeding between populations seem to be
relatively rare.

There are differences, as well as similarities, in the life features
of varieties 1, 2, and 3. Variety 2 belongs to group B, whereas
varieties 1 and 3 belong to group A. Variety 2, like variety 9,
seems to show relatively little F2 mortality in the strain crosses
that have been examined. On the other hand, variety 1 shows
exceedingly high F2 autogamous mortality after crossing certain

214 PROTOZOA

strains. For example, Beale (1952) found more than 60% F2
mortality after crossing strains 60 and 90. Field studies by
Beale ( 1954, and unpublished ) indicate that heterozygotes occur
in nature in variety 1 less commonly than expected from random
mating. Both close inbreeding and some measure of outbreeding
doubtless occur in these three varieties. Evaluation of the net
effect of the complex of their features on the breeding system
must await fuller study; but it is already evident that these
three varieties have breeding systems intermediate between
those of the most extreme inbreeders and those of the most ex-
treme outbreeders.

5. Varieties 5, 6, and 13. Nothing is yet known with pre-
cision about some of the most important features of the life cycle
in these three varieties, e.g., their periods of immaturity and ma-
turity. Variety 5 belongs to group A, variety 6 to group B, and
variety 13 cannot yet be assigned to either group. Every caryo-
nide of variety 13 becomes a selfer;6 most caryonides of variety 5
also do; and so do many caryonides of variety 6. However, the
frequency of selfing pairs in the selfing caryonides of varieties
5 and 6 is usually very low, while it is very high in variety 13.
Further, the predominant mating type in selfing caryonides of
varieties 5 and 6 remains constant for a caryonide, but it does
not in variety 13. Finally, autogamy occurs in varieties 5 and 6,
but it has not yet been found in variety 13. 6 On the basis of the
widespread occurrence of selfing in these varieties, they may all
be considered as strongly predisposed to inbreeding; but, since
selfing occurs universally in all caryonides of variety 13 and with
high frequency in each of them, variety 13 is by far the most ex-
treme inbreeder of this group. This picture may have to be
changed somewhat when variety 13 becomes more fully studied.1'
The possibility has not yet been excluded that a short immature
period exists and that the frequency of selfing increases with age,
though rather rapidly. If these things do occur, they would op-
erate to some extent toward favoring early outbreeding.

6. Varieties 4 and 8. These last two varieties to be taken up
are clearly very strong inbreeders, almost as extreme in this re-

T. M. SONNEBORN 215

spect as varieties 10 and 14, which were used at the start as the
examples of the most extreme inbreeders. They both belong to
group B and lack a period of immaturity. Hence they are capable
of mating when they are most likely to be near close relatives of
complementary mating type. Their period of maturity, however,
is short, about 6 days or 25 fissions, but this is twice as long as in
varieties 10 and 14. Such a short mature period might be consid-
ered to provide relatively little opportunity for conjugation to
occur. Yet, if all the individuals lived and food remained abun-
dant, the two clones of complementary mating type from a pair
of conjugants would yield many millions, possibly even billions, of
progeny before famine would induce 100% autogamy rather than
ripeness for mating. Applying the maximum density of population
calculated from Pringle's data on variety 9, it would take a lake
with a surface area of a square mile and an average depth of 25
feet to house so large a population. It would seem, therefore, that
even with so short a period of maturity, the progeny of a single
pair of conjugants would be likely to deplete the food and be ripe
to conjugate before maturity had passed. This would certainly
happen in the second generation for it would start, not with a
single pair of conjugants, but with a large population. This and
all such calculations fail to take into account predation and other
factors that eliminate part of the population and permit the resi-
due to multiply further before exhausting the food supply. On
the other hand, it also ignores competitors for food. Such errors
will have to be corrected before conclusions as to the relative
frequency of conjugation and autogamy in nature can be placed
on a firm basis. Meanwhile, considerations based on the limited
available knowledge have value in exposing needed types of fur-
ther investigation. For example, it is known that although mating
type changes relatively rarely at autogamy and conjugation,
changes are more frequent in one direction at autogamy and in
the opposite direction at conjugation in variety 4 at 27°. There
is need to try to discover the various factors involved in determin-
ing the relative frequencies of the two mating types in a popula-
tion, their quantitative effects, and the equilibria attained under

216 PROTOZOA

various conditions. Such information should be of great impor-
tance in estimating the relative frequency of occurrence of con-
jugation and autogamy.

Unlike the extreme inbreeders ( varieties 10 and 14 ) which have
been found once only, variety 8 has been found in numerous
localities, all but one in the extreme southern part of the United
States; and variety 4 occurs abundantly in cool to warm parts of
the United States, South America, Japan, and Australia. The mate-
rials for outbreeding are thus much greater in variety 8 than in
varieties 10 and 14, and they are enormously greater in variety 4.

Nevertheless, in both varieties crosses between strains from
different sources invariably lead to considerable mortality in the
F2 obtained by autogamy. In variety 4 this varies from about 15%
to about 98%, depending upon which strains have been crossed.
This is in striking contrast to the results of repeated autogamies
within a strain. Sonneborn and Sclmeller (1955) passed a strain
of variety 4 through 50 successive autogamies without an inter-
vening conjugation and without increasing mortality which re-
mained throughout virtually negligible (0.7%). Variety 4 thus
survives the closest inbreeding very well, but invariably it suffers
from outbreeding. This imposes a strong selective disadvantage
on crosses between different local populations.

The probable basis of F2 mortality following crosses of repre-
sentatives of different populations of variety 4 was revealed by
Dippell's (1954) karyological studies. Every strain of variety 4
that she examined had a different karyotype. The karyotypes dif-
fered in number of chromosomes in an aneuploid series from
about 33 to 51 chromosomes in the haploid set. Further, there
were regularly differences in the shapes and sizes of the chromo-
somes of different strains. In the laboratory, a strain remains as a
rule constant for all features of its karyotype. Whether variations
occur in detectable frequency within a natural population remains
unknown. Finally, Dippell roughly correlated the degree of F2
mortality with the degree of karyotypic diversity.

If, as now appears, F2 mortality has this basis, then the high
F2 mortality found in other varieties, such as varieties 1 and 8,
may be taken as prima facie evidence of the existence in those

T. M. SONNEBORN 217

varieties of karyotypic diversities among local populations. Clearly
this situation could develop only if crosses between local popula-
tions were rare events. Otherwise, interbreeding would swamp out
existing karyotypic diversities and prevent new ones from arising,
as we shall later find to be true in certain outbreeding Ciliates.
Thus Dippell's studies lead to the view that F2 mortality follow-
ing strain crosses may be one of the most potent evidences of the
existence of regular close inbreeding in a variety.

The general picture of variety 4 that now emerges is of a vari-
ety which is a loose assemblage of innumerable highly isolated
and highly inbred local populations. Grouping these together as
a variety is to some degree artificial. Although it is true that they
are potentially able to interbreed and do so in the laboratory,
everything we know about them points strongly to the conclusion
that such interbreeding occurs very rarely in nature, almost negli-
gibly except perhaps on a very long-range view. In effect, each
local population of variety 4 just falls short of having the same
status as a variety among outbreeders. The same conclusions ap-
ply even more extremely to varieties 8, 10, and 14, perhaps less
so to variety 1 and some others.

The importance of these conclusions in relation to problems of
species, breeding systems, and evolution in P. aurelia is self-
evident. Some varieties are in effect virtually obligatory inbreed-
ers, and each isolate in nature is an entity of some biological
stature. For such varieties, the local population is the natural
biological unit of organization, although still a part of a larger
unit (the variety) the members of which are in actuality all but
completely cut off from one another.

This situation raises many problems about which we can only
speculate at present; most of the information needed for more
solid analysis is lacking, though at least much of it is obtainable.
Soma speculation may stimulate efforts to supply the needed in-
formation. The problem of how paramecia get from one body of
water to another, as their distribution proves they must, is an old
but still unanswered one. Cysts capable of withstanding desicca-
tion have never been satisfactorily proved to occur in Parame-
cium. Human transportation of water must be a minor and recent

218 PROTOZOA

factor. Transportation by animal carriers, such as insects, am-
phibians, and terrestrial vertebrates, needs to be looked into far
more than in the past. Perhaps even occasional hurricanes and
floods and subterranean water connections between surface wa-
ters play a part. The pressing problem with respect to migration
is not just how or how far, but how often.

The next problem is what happens when a migrant succeeds
in getting away from home base into another body of water. This
is of course a complex of problems. What happens will depend
upon whether or not the new locale contains a population of the
same variety. If it does and the migrant is an outbreeder, the
events that follow will differ from those that would take place
with an inbreeder. Special problems arise if the intruder finds a
local population, not of the same variety, but of a variety with
which it nevertheless can mate.

It is clear that such interbreeding varieties can coexist for some
time. For example, I have found varieties 4 and 8 in a small sam-
ple of water from a single source. Since in the laboratory the hy-
brids are readily produced but either die or are very weak, one
would expect equilibrium to be reached only when one variety,
the less numerous one, dies out completely as a result of such
crosses. Persistent coexistence would seem possible only under
one of two conditions. (1) If corresponding mating types of the
two varieties were in excess, they could not interbreed, and the
minority types would be eliminated or kept down by crossing.
This would also enforce inbreeding by autogamy in each variety
because only one mating type would be present. (2) If there is
nonrandom mating between the two varieties and each preferen-
tially mates with its own variety, this would reduce the elimina-
tion which results from crossing.

No great problems are posed by the migration of an outbreeder
into waters occupied by another strain of the same variety. To the
extent that crossing occurs, a new equilibrium of the now com-
mon gene pools would occur. The low or nonexistent mortality
From strain crosses of outbrecders would impose little or no bar
to amalgamation. The situation lor an inbreeding intruder is of
course very different. Karyotypic diversity and F2 mortality

T. M. SONNEBORN 219

would tend to wipe out the intruder by reason of its small propor-
tion in the population, if extensive mating with the indigenous
strain took place. The degree to which such interbreeding would
occur depends upon the mating type frequencies in the two
strains. If some does occur, the progeny will include some viable
combinations, and these will give rise in further fertilizations to a
supply of genotypes and karyotypes for selection to work upon.
Eventually this could lead to a new relatively homogeneous popu-
lation of a single selected best adapted karyotype or to a balanced
mixture of diverse karyotypes and genotypes. Until field studies
of local populations of strongly inbreeding varieties and labora-
tory studies of mixtures of diverse strains of such varieties are
available, it is not worth pursuing the possibilities further.

If a migrant enters a new body of water not already occupied
by its variety or one with which it can mate, there is no difficulty
in seeing how both an inbreeder and an outbreeder could soon
give rise to a population containing both mating types. Both the
group A and group B methods of mating type inheritance provide
ready-made mechanisms for this except in the exceedingly rare
cases in group A (varieties 1 and 7) of strains pure for one mat-
ing type. The only real problem raised by migrations of this sort
is how to account for the development of a distinctive karyotype
by the inbreeders. This may happen very slowly by selection of
rare spontaneously arising chromosomal aberrations. In the latter
case, two factors present themselves as possibilities. First, one
thinks of the invariable production of chromosomal aberrations
with advancing senility (Sonneborn, 1955; Dippell, 1955). Ordi-
narily these would not be expected to occur in nature. In estab-
lished populations, famine would be expected to set in long before
the age at which the aberrations occur. This would lead to ferti-
lization and thereby to setting the age clock to zero (Sonneborn,
1954a). However, if the migrant is already no longer young and
the new waters provide food for much reproduction by a single
intruder (or very few), the requisite age might be reached. A sec-
ond possibility arises from some indications that fertilization at
extreme temperatures (10° and 35°) also results in chromosomal
aberrations in certain variety 4 stocks. If paramecia migrate via

PROTOZOA

animal carriers such temperatures might prevail, e.g., on the moist
surfaces of insects during flight or in the bodies of warm-blooded
animals. The small amount of available nutrient under such con-
ditions might at the same time induce autogamy. Both of these
possibilities may be wide of the mark, but they and others need
to be looked into. The existence of regular karyotypic diversities
among local populations of inbreeders is a phenomenon that needs
to be accounted for.

7. Conclusions and general comments. This survey of the
bearing of varietal characteristics on the breeding system is no
doubt incomplete. Increase of knowledge and further reflection
will probably bring to light the relevance of other characteristics.
Among those we have discussed, the period of immaturity stands
out as of major significance. The longer it lasts, the more oppor-
tunity close relatives have for dispersal before they can mate;
hence, the less likely that they will mate with each other. A very
long period of immaturity is thus an indicator of an outbreeder.
The absence of a period of immaturity marks an inbreeder. Being
ready at once to mate again, they are apt to mate with those
nearest at hand, i.e., their close relatives.

The duration of the mating period is also of much importance.
This limits the period for finding a suitable mate. Probably no
great period of time is required to find a mate in the same local
population. In accordance with this, a short mature period marks
an inbreeder. If it fails to find a local mate in time, it undergoes
autogamy. Outbreeders need more time to find a suitable stranger
to mate with and they have long periods of maturity. Their chance
of success is enhanced, as in variety 16, by multiple mating types.

Certain considerations apply to these two periods taken to-
gether. First, the available evidence indicates that migration from
one body of water to another is a relatively rare event; hence,
short combined periods of immaturity and maturity will as a rule
prevent breeding between different local populations and conjuga-
tion would occur between inhabitants of the same body of water.
The longer the duration of the two periods, the more time for
migration into or out of the local population and consequently
the more opportunity for outbreeding. Under the most favorable

T. M. SONNEBORN 221

conditions of the longest immature and mature periods, the
chance of outbreeding may still be small. But the occurrence of
selfing during early senility in varieties with such long periods
starts a new cycle and extends the time available for finding
suitable mates.

Second, if the total life prior to senility occupies a year or more,
the annual rise and fall in population density may reduce the
chance of inbreeding by rendering the population less hetero-
geneous in mating type.

Third, the immature and mature periods must be considered
together sometimes to avoid making an erroneous judgment based
on one of them alone. According to old observations of mine on
stock S of variety 1, it has an immature period that lasts almost
until the beginning of senility and that leaves but a very brief ma-
ture period. The immature period alone suggests an outbreeder,
but the mature period is so brief as to militate strongly against
mating with members of another population and even against
mating at all. Unless famine comes at just the right moment, its
consequence would be autogamy or no sexual process at all. How-
ever, since no immature period follows autogamy and the mature
period is then correspondingly longer, there is more opportunity
for the occurrence of conjugation after an autogamy in such a
strain. This example shows how the interplay of various aspects
of life must be taken into account before arriving at a judgment
of the breeding system.

In addition to the duration of the periods of immaturity and
maturity, there are two other important signs of the breeding sys-
tem. One is a direct examination of the frequencies of various
genotypes in a natural population. An excess of homozygotes and
a deficiency of heterozygotes, in comparison with expectations
based on random mating, is of course a sign of inbreeding even
within a local population. The other sign is the karyological pic-
ture or inferences about it based upon the nature of the F2 after
crosses between representatives of different populations. Kar-
yological differences between local populations and increased F2
mortality following crosses between them indicate that the popu-
lations are isolates of inbreeders. The indication is confirmed if

PROTOZOA

each population alone survives inbreeding very well. Outbreeders
would be expected to show inbreeding mortality as a result of
lethal and detrimental genie recombinations, but this should hap-
pen both when the inbreeding follows crosses and when it occurs
within a single population.

Certain other indices of the breeding system are perhaps rela-
tively minor in comparison with those already discussed, but they
are still significant. First, the range of distribution of a variety is
both a rough measure of the number of populations available for
interbreeding and an index of the genetic heterogeneity of the
variety. Outbreeders generally have a wide range of distribution
both in latitude and longitude. Inbreeders are generally restricted
either in latitude or in both directions. Second, the rates at which
the processes of meiosis, fertilization, postzygotic nuclear reorgan-
ization, and particularly asexual reproduction occur are relatively
slower in the extreme outbreeders than in the extreme inbreeders.
The slower the tempo, the more time for dispersal before matur-
ity and before senility, and therefore the more favorable this is for
outbreeding. Third, strong outbreeders become selfers during
senility; strong inbreeders undergo autogamy during senility.
Nevertheless, some degree of outbreeding is compatible with the
occurrence of autogamy. This is true because autogamy is never
followed by a period of immaturity. Hence, the homozygosity
resulting from autogamy can at once be compensated by restor-
ing heterozygosity through conjugation within the population.
The episode of autogamy serves to prevent the progression of
senility and to restore vigor, and it can do this without undue
sacrifice of heterozygosity, thanks to the clever device of sup-
pressing immaturity after autogamy.

This summary of the life features that affect the breeding sys-
tem has thus far failed to refer to what might, a priori, have been
supposed to be the most important single feature, namely, the
mating types and their inheritance. In spite of the large amount
ol information available on this complex subject, I find it the most
difficult one to relate to breeding systems. Before attempting to
do so, the units with which wc shall have to deal will be re-
viewed. The basic unit is the caryonide, the group of individuals

T. M. SONNEBORN 223

that possesses macro-nuclei descended from the same ancestral
macronucleus. The clone, embracing the vegetative descendants
of a single autogamous individual or a single conjugant, usually
includes two carvonides in most varieties but four in varieties 13,
15, and 16. A new term, synclone, is a convenient designation for
the vegetative descendants of two animals that have conjugated
with each other. Each of these three terms, caryonide, clone, and
synclone, designates a group of individuals sharing a common
genotype. (The two conjugants of a pair normally acquire iden-
tical genotypes.) Nevertheless, in spite of genie uniformity only
the caryonide is (aside from selfers) always phenotypically uni-
form with respect to mating type. The two caryonides of the
same clone are often unlike in mating type in group A varieties
but usually alike in mating type in group B varieties. The two
clones of a synclone are usually unlike in mating type in group B
varieties. The only larger unit we need to consider is the local
population which consists of unknown numbers of caryonides,
clones, and synclones.

The difficulty in judging the bearing of diverse systems of mat-
ing type inheritance on the breeding system stems from the fact
that in both the group A and the group B varieties, as well as in
the unclassified variety 13, any one synclone as a rule has a rea-
sonably high probability of including both mating types. Whether
inbreeding will occur or not thus seems to be independent of the
method of mating type inheritance at conjugation. The same dif-
ficulty is encountered if we start, not with synclones, but with a
single caryonide. When it goes into autogamy (and also presuma-
bly if it selfs in senility ) , both mating types will arise in varying
relative frequencies. The only known exceptions are the very
rare strains pure either for mating type I or mating type XIII; as
already pointed out, they are thereby forced, not to outbreed, but
to inbreed by autogamy. Thus, so far as the methods of mating
type inheritance alone are concerned, it would seem that all varie-
ties of P. aurelia should be close inbreeders. Outbreeding be-
tween local populations would be favored if each whole popula-
tion were pure for one mating type, but only if there were also a
very long period of maturity and no autogamy. Outbreeding

PROTOZOA

oithin a local population, that is, mating between different syn-
clones, would require a mechanism for preventing mating until
there has been time for dispersal, i.e., by a sufficient immature
period, or for a mechanism to make the synclone uniform in mat-
ing type.

In fact, both groups A and B have mechanisms for bringing
about synclonal uniformity in mating type. In group B, the mech-
anism is massive cytoplasmic exchange between mates. It has
been studied to some extent in variety 2 (with intermediate
breeding system), but most fully in variety 4, a strong inbreeder.
In this variety, the mechanism usually gives a synclone of mating
type VIII at high temperatures, while at low temperatures both
types of synclones are produced in high frequency (Nanney,
19.54 ) . Remarkably, there are strain differences in the capacity to
experience cytoplasmic exchange during mating. Wood ( 1953 )
found a single major gene difference between the two kinds of
strains. He also found that the gene for cytoplasmic exchange
has much higher penetrance if conjugation takes place at low
temperatures and if the conjugating clones are very young. These
facts could well be of importance in relation to the breeding sys-
tem. Outbreeders would be expected to have the genotype for
cytoplasmic exchange and to conjugate in nature under condi-
tions favorable for the penetrance of the relevant gene. Inbreed-
ers would be expected to lack this gene or to conjugate in nature
under conditions in which it rarely comes to expression. It is not
known whether these alternatives actually occur in association
with inbreeding and outbreeding.

The mechanism in group A is quite different. Here cytoplasmic
exchange plays no part whatever, for there is in group A no cy-
toplasmic difference between the two mating types. Each car-
yonide is independently determined for mating type and the
probability for being determined as the even-numbered mating
type increases linearly with the temperature prevailing during
pre- and postzygotic reorganization (Sonneborn, 1939). At high
temperatures, the even type has so high a probability that a con-
siderable proportion of synclones are uniform for it. However, it
seems likely that at the temperatures prevailing in nature, the

T. M. SONNEBORN 225

probability of each type may not be far from 0.5; and with this
probability only one-eighth of the synclones would be uniform
for mating type. However, this is the minimal proportion of syn-
clones uniform for mating type; all other probabilities give higher
proportions. Since the minimum is virtually zero in group B varie-
ties under some conditions, the group A method of mating type
determination is, under these extreme conditions, more efficient
in promoting outbreeding than the group B method. In agree-
ment, the most extreme outbreeder in which the method of
mating type inheritance is known ( variety 15 ) occurs in group A
and the most extreme inbreeders (varieties 10 and 14) in group
B.

The potentiality of the group A method for promoting out-
breeding is further indicated by some other facts. Different strains
of the same variety, e.g., variety 1, have characteristically differ-
ent proportions of the two mating types under the same tempera-
ture conditions (Sonneborn, unpublished; some data quoted by
Butzel, 1955). Butzel (1955) presented evidence that these dif-
ferences are due to genes other than the mating type genes. This
probably explains why hybrids between certain strains show in
the Fl a much higher frequency of type II than either parent
strain ( Sonneborn, 1942a ) , while hybrids between other strains do
not ( Butzel, 1955 ) . If the group A method of mating type inheri-
tance is ever a strong factor for promoting outbreeding, one might
expect outbreeding strains or varieties to yield high frequencies
of one mating type under the conditions in which they conjugate
in nature. As this discussion has attempted to bring out, there are
aspects of mating type inheritance which bear on the breeding
system, but little is known as to whether they do in fact operate
this way in nature.

With due allowance for the points on which we are still regret-
tably ignorant, the mass of information now in hand on the varie-
ties of P. aurelia has acquired for the first time coherence and
significance by the perception of its bearing on systems of breed-
ing. Instead of being merely so many interesting but isolated facts
of natural history or experimental research, they are now the
meaningful elements of the grand design by which this organism

226 PROTOZOA

achieves its primary function of self -perpetuation. In view of its
surpassing importance for every species, no surprise or suspicion
should be aroused concerning the apparent convergence of so
many aspects of the biology of P. aurelia upon it.

Evolution of Groups A and B and Their Varieties. ( 1 ) Origi-
nally groups A and B were recognized only on the basis of their
different systems of mating type inheritance ( Sonneborn and Dip-
pell, 1946). In the discussion of the serotypes, it was pointed out
that all the group A varieties examined (varieties 1, 3, and 9) show
one system of serotypes and that all the group B varieties ex-
amined (varieties 2, 4, and 8) show a different system. Thus,
group A varieties may equally well be distinguished from group
B varieties by either the mating type or the serotype system.
There is also a third difference between the two groups (Sonne-
born, 1956c). Killers are common in all group B varieties in
which many strains have been observed, and have been found in
every group B variety of which more than one strain is available,
i.e., varieties 2, 4, 6, and 8. Not a single killer strain has been
found among the large number examined in the varieties of group
A.4 Curiously, these three independent sets of traits show in their
inheritance a conspicuous cytoplasmic involvement in group B,
not in group A; but the mode of cytoplasmic involvement seems
different for each of the three sets of traits. One can hardly re-
frain from suspecting some at present unrecognized common ele-
ment in the situation which leads to this nonrandom concentra-
tion of cytoplasmically inherited traits in group B.

The differences between group A and group B varieties in three
sets of traits suggest that the cleavage into these two groups of
varieties represents an ancient and important evolutionary diver-
gence. If so, then one might expect the differences between the
group A and group B alternatives for each of the three sets of
traits to be rather simply derivable from one another. The al-
ternatives with respect to the killer trait are genetically simple.
The presence of this trait depends upon the presence of one major
gene (Sonneborn, 1943) and a few minor genes (Sonneborn,
unpublished; Balbinder, unpublished). Without these genes, the
cytoplasmic particles, kappa, ran, or pi, which determine the

T. M. SONNEBORN 227

killer trait and its various mutant forms, cannot be maintained in
the animals. The differences between groups A and B with re-
spect to the serotypes and mating types have been commented
upon by several authors (Beale, 1954; Nanney, 1954, 1956b; Son-
neborn, 1947, 1954b ) . Here I shall discuss only the mating types,
for they are of special importance in relation to breeding systems
and evolution.

The group A and group B systems of mating type determina-
tion and inheritance have been characterized in detail from vari-
ous points of view at several places in the present paper. On the
basis of these facts an attempt will be made here only to formu-
late these systems in an abstract and general way in order to see
how many and what kinds of changes would be required to evolve
one from the other. In making this formulation, I shall pass rap-
idly over the points on which evidence is good and interpretation
is fairly straightforward, but shall say more about certain other
points.

On a few points, common to both systems, there can be little if
any doubt. These are: ( 1) the mating types are controlled by the
macronucleus; (2) except in certain mutants, all macronuclei
carry a genotype for both mating types; (3) usually a macronu-
cleus becomes differentiated early in its development from a prod-
uct of the syncaryon so that only one of its mating type poten-
tialities will thereafter come to expression; (4) this macronuclear
differentiation is permanently or long irreversible and, as a result,
a caryonide is commonly pure for one mating type. Nanney
(1956b) interprets the macronuclear differentiation as an intra-
nuclear steady state between products of the alternative genes
for different mating types.

The differences between the group A and B systems involve
the agents that bring about macronuclear differentiation and the
genetic control of these agents. In group B, the cytoplasm clearly
is the agent of macronuclear determination, and the cytoplasmic
agent is clearly controlled by that macronuclear mating type geno-
type which is being expressed. This is what results in the usual
correlation in mating type between the two caryonides of a clone
and between both of them and the mating type of the sexual

228 PROTOZOA

parent. Superficially, the group A system seems very different:
there is no correlation between mating types of the two caryo-
nides of a clone or between these and the sexual parent. This lack
of correlation gives the impression that macronuclear determina-
tion is not brought about by a cytoplasmic agent, an impression
which is strengthened by the marked effect of environmental con-
ditions, such as temperature, on macronuclear determination.

However, the facts are not contradictory to the interpretation
of macronuclear differentiation by the cytoplasm or to the control
of the cytoplasmic agent by the macronuclear genotype. Further,
some facts point to just this sort of mechanism. Such a hypothesis
will therefore be developed in order to have a single unified inter-
pretation for both the group A and group B systems. There will
obviously have to be some differences between the two systems,
but the hypothesis will attempt to reduce these to the minimum
and preserve maximal similarity. If it be assumed that there is a
different cytoplasmic agent corresponding to each possible nu-
clear differentiation, i.e., two in a variety with two mating types,
four in a variety with four mating types, there need be only two
differences between the group A and B systems. ( 1 ) In group B,
the cytoplasmic agents or differentiators of the macronuclei are
subject to a steady state relationship; but in group A they are
not. This means that in group B one or the other of the cytoplas-
mic differentiators is always in relatively high effective concentra-
tion as compared with the other or others; but that in group A
a continuous series of relative concentrations is possible. (2) In
group B, the macronuclear genes for the expressed mating type
also as a rule determine that the corresponding cytoplasmic
differentiator will have a relatively high effective concentration;
but in group A, although the macronuclear genotype also controls
the relative concentrations of the cytoplasmic differentiators, the
pertinent part of the genotype is not that, or not only that, part
which controls the expressed mating type. The genes involved in
group A control of the cytoplasmic differentiators an1 assumed to
be the same in both mating types.

With tin's interpretation, group A would show the absence of
parent-offspring correlation that is observed. Further, the ob-

T. M. SONNEBORN 229

served linear relation between temperature and the frequency of
differentiation of macronuclei determining the even-numbered
mating type would be explicable as due to the rise in relative
concentration of the corresponding cytoplasmic differentiator
with increasing temperature. Exactly the same rise has been
noted by Nanney ( 1954 ) in group B when conjugants exchange
cytoplasm and thus set up an approximately equal concentration
of the two cytoplasmic differentiators. However, because of the
steady state mechanism in group B, a rise in the concentration of
the even-numbered differentiator leads to a steady state in which
it greatly predominates. Finally, Sonneborn (unpublished) and
Butzel (1953) showed that different genotypes in group A result
in determination of different proportions of the mating types at
the same temperature. Hence, the genes seem to be determining
something on which temperature also acts. On the present hy-
pothesis, this something is the relative concentrations of the
cytoplasmic differentiators.

Viewed in this way, the evolution of the differences between
the two mating type systems needs involve but few changes: the
establishment or loss of the steady state condition for the cyto-
plasmic differentiators, and the predominant control of these
differentiators by the expressed mating type genes versus control
by genes which are identical in both mating types. If Nanney's
( 1956b ) conclusion as to the existence in both groups of varieties
of an intranuclear steady state among the products of the mating
type genes is correct, the extension of this steady state mecha-
nism into the cytoplasm in group B and not in group A may rep-
resent a relatively simple difference.

The group B system seems to be the more ancient one, occur-
ring in all investigated varieties of other species of Paramecium
and in one variety of Tetrahymena pyriformis. It is associated in
some of these organisms with outbreeding and, as already men-
tioned, outbreeders were probably the ancestors of current in-
breeders. However, one variety of T. pyriformis shows the group
A system. Thus, either we are on the wrong track in supposing
groups A and B of P. aurelia diverged early, or such divergence
has occurred repeatedly and independently in different genera.

230 PROTOZOA

Since the group A system can arise merely by loss of two features
of the group B system, the repeated independent evolution of the
A from the B system seems not unreasonable.

2. Evolutionary relations within groups A and B. If the three
different features which characterize all group A varieties and the
alternative three which characterize all group B varieties indicate
a fundamental and ancient evolutionary divergence of the two
groups, we should expect to find evidences of this even in those
features which distinguish varieties of the same group. For them,
we should expect closer relations among varieties of a group than,
on the average, between varieties belonging to different groups.
This is indeed found with respect to the two sorts of chemical
specificity expressed as antigens and as mating types.

In the account given earlier of the serotypes, it was pointed out
that the specificities of the immobilization antigens were different
for each variety, but that the differences were least for varieties
4 and 8 of group B. Most of the antigens that occur in one variety
are matched by somewhat different but related antigens in the
other. Resemblances are detectable between several of the anti-
gens in these varieties and several of those in variety 2, the other
investigated group B variety, but the resemblances are fewer and
weaker. Of the three principal antigens in variety 1, two bear
some resemblance to the group B antigens, but these resem-
blances are recognizable serologically in relatively few strains.
Melechen ( 1955 ) reports that correspondences to all three prin-
cipal antigens of variety 1 are indicated by antigens in variety 3.
One of the two antigens found in variety 9 corresponds to one
of the three main antigens in variety 1. The serologic relations are
certainly closer for some varieties of the same group than for
some varieties of different groups. Whether this relation will
prove to be general remains to be discovered.

The relations of mating type specificity are indicated by the
capacity of mating types of different varieties to interact sexually.
Table I shows these interactions. Within group A, they occur
among varieties 1, 3, 5, and 7, most strongly between 1 and 5.
Within group B, they occur among varieties 4, 8, 10, and 14, most
strongly between 4 and 8. The close relationship of varieties 4 and

T. M. SONNEBORN 231

8 is thus attested by both their serologic and their mating type
specificities. The only sexual reactions between groups A and B
involve variety 8 of group B. It reacts and mates with variety 3
and gives barely perceptible reactions but no mating with varie-
ties 1 and 7. Sexual reactions between complementary mating
types of different varieties, without regard to their intensity, occur
in 6 of 42 possible combinations within group A, 6 of 42 within
group B, and 4 of 84 between groups A and B. They are thus three
times as frequent within each group as between groups. If only
those that lead to mating are counted, there are 6 among the 84
possibilities within groups, and only 1 of the 84 possibilities be-
tween groups. The mating type specificities are thus on the whole
less different within a group than between groups.

What light, if any, does the geographical distribution of the
varieties throw on their evolution? On this subject, knowledge
is of course meager, but it is suggestive. In group A, only variety
1 is known to be cosmopolitan; variety 9 has been found only in
Europe, varieties 3, 5, 7, and 11 only in North America. Variety
15 will be disregarded for little is known about it. These facts of
distribution suggest that varieties 3, 5, 7, 9, and 11 may have
arisen from variety 1. The antiquity of variety 1 is indicated by
its occurrence in Australia and Hawaii as well as in both Ameri-
cas, Asia, and Europe. The relative recency of origin of the others
is suggested by the fact that each is, so far as now known, con-
fined to a single continent. And the close relationship of varieties
3, 5, and 7 to variety 1 is indicated by sexual reactions that occur
between them. Serological evidence relates varieties 3 and 9 to
variety 1. Significantly, varieties 3 and 5 are localized near the
northern, and variety 7 near the southern limit of the range of
variety 1 in North America. Altogether it seems reasonable to
conclude that the other group A varieties (3, 5, 7, and 9) arose
from variety 1 relatively recently.

Several varieties of group B have a wide distribution, but only
variety 2 is cosmopolitan. Variety 6 occurs in India, Puerto Rico,
and southernmost North America and it may be assumed to be
subtropical around the world. Variety 4 has been found in Aus-
tralia, Japan, and both Americas, but not in Europe. These three

PROTOZOA

widespread varieties of group B give no sexual reactions with
each other. Of the varieties of group B with more limited distri-
bution, variety 12 has been found only once (in Madagascar)
and it seems to be very different from the other group B varieties.
The three remaining varieties, 8, 10, and 14, have been found
only in North America, almost exclusively in its southernmost
parts. As has been pointed out in detail, these three varieties are
much alike in a number of ways and are also much like variety 4.
These resemblances and distributions suggest that varieties 8, 10,
and 14 evolved relatively recently from variety 4 directly or
indirectly.

3. Gaps in knowledge. Many of the gaps in present knowl-
edge have already been mentioned. A most important one is the
paucity of information on the large animal varieties, both those
that form two micronuclei and two macronuclei after fertilization
and those that form four of each. Whether the latter constitute
a distinct and readily recognizable species (P. multimicronucle-
atum) or whether they intergrade into the others or cannot be
readily identified is one problem that needs solving.2 Another is
the possibility of mating and of serologic relations with the small
animal varieties, a subject barely touched as yet. A third problem
concerns how many varieties of large animals occur and what
their distributions are. Attention is just beginning to be directed
to this. When these matters are cleared up, a good deal more
about evolution in these organisms will doubtless become appar-
ent and some of the inferences in the preceding section may well
have to be changed. It may, for example, turn out that certain
large animal varieties are the closest to the ancestors of both
group A and group B. I have long suspected, on a priori grounds,
l hat the ancestors of these groups were more extreme outbreeders
than any small animal variety yet found; and some of the large
animal varieties are extreme outbreeders.

Isolating Median is ins. Geographic isolation appears to be the
most potent isolating mechanism in P. aurclia. Although it is
obvious that the organisms have become spread over the face of
the earth, the means by which they accomplish this remains un-

T. M. SONNEBORN 233

known. Moreover, observations snch as those I made over a
period of years on the population of the Woodstock stream, to-
gether with Beale's observations on the occurrence of only one
allele at each serotype locus in the samples taken in various years,
show how rare must be the effective migration of a foreign strain
into a preexisting population. In this case, moreover, the observa-
tions were made on a population of variety 1 in the midst of a
region well within the range of the variety and where many other
populations of the same variety were known to occur. If migra-
tions from one body of water to a neighboring one are so rare,
then migrations between continents must be exceedingly rare or
nonexistent except on a geologic time scale. The fact that strains
of the same variety (such as 1, 2, 4, and 6) from different con-
tinents can still interbreed freely when brought together in the
laboratory implies that the widely distributed varieties must be
evolving very slowly in spite of long isolation. Nevertheless, with
each local population virtually locked up in the tight compart-
ment of a river system or body of still water from which leakage
is slow and erratic, the stage would seem to be set for abundant
evolutionary divergence.

Differences in mating type specificity and differing conditions
for mating reactivity (temperature, light), once they arise, con-
stitute such effective barriers to interbreeding that different vari-
eties can and do intimately coexist in the same body of water
without losing their integrity. Even when the mating type speci-
ficities are only slightly different and interbreeding is possible,
as in the case of varieties 4 and 8, the varieties are found to co-
exist in the same body of water in nature. From these observa-
tions, it would seem that even relatively slight changes in mating
type specificity could lead to isolation of a new variety. Probably
the inbreeding propensities discussed earlier assist such isolation.
Outbreeders might not be so easily isolated in this way. Another
related possible isolating factor is assortative mating with respect
to body size (Jennings, 1911). There is some evidence that body
size is in a general way correlated inversely with the rate at
which the processes of meiosis and fertilization occur and marked

PROTOZOA

differences in these rates in two mates leads to abnormality
(Sonneborn, 1954b). Such abnormality would put the hybrids
between the two types at a selective disadvantage.

The role of the killer traits in bringing about isolation is per-
haps not as great as might at first be supposed. Killers that liber-
ate paramecin might be expected to eliminate or reduce other
strains in the population, especially when the population density
is high. To the degree that mating is at random, the excess of
killers would foster inbreeding among the killers, but outbreed-
ing among surviving sensitives of the same variety. The mate-
killers, known only in variety 8, present a special situation, which
has been discussed by Siegel ( 1954 ) .4 Gene flow occurs readily
from sensitive to killer, but not directly in the reverse direction.
However, since mate-killers readily outgrow the cytoplasmic
basis, [J-, of the trait sensitives of the same strain are likely to be
present and through them some gene flow from the killer strain
to other sensitive strains would be possible. Even so, gene flow in
the two directions would probably be very unequal unless the
proportion of mate-killers in the strain were low.

Whatever degree of isolation is due to killers is cytoplasmic in
basis; and cytoplasmic involvement in isolation has also been
found in another case. Levine (1953) showed that the death of
hybrids between varieties 4 and 8 is due to the lethal action of
the cytoplasm of each variety on the nuclei of the other. This is
also to be noted as an additional evidence of the importance of
the cytoplasm in the varieties of group B. I doubt whether the
same sort of mechanism plays any role in isolation of varieties of
group A. Butzel (1953) was unable to find any trace of such
cytoplasmic action in hybrids between varieties 1 and 7. Cyto-
plasmic differentiation may regularly bo a stage in genetic isola-
tion in organisms with well-developed cytoplasmic genetic sys-
tems like group B, not in other organisms like group A.

Probably the most important isolating factor in P. aurelia, aside
horn geographical isolation, is the widespread and strong in-
breeding habit. This is so important in relation to the species
problem that the two are considered together in the next section.

Breeding Systems and flic Species Problem. Among the L6

T. M. SONNEBORN 235

varieties of P. aurelia only three varieties, 12, 15, and 16, seem to
be adapted primarily for outbreeding. Unfortunately, these are
among the varieties which have been least studied from the point
of view of speciation. It might be expected that these outbreeders
and others like them, if such are yet to be discovered, constitute
the main line of descent in P. aurelia from the past and the main
line that will long continue into the future. They above all have
the mechanisms for maintaining the supply of genetic variability
needed for long-time survival. We may also expect such varieties
to show the minimal differentiation of local populations. Among
them, we may also eventually expect to find the best evidence for
the existence of races embracing the local populations in a con-
siderable land mass.

At present, evidence for such races is meager and is limited to
Beale's (1954) survey of serotype alleles in strains of variety 1.
This variety is intermediate in breeding habit. Beale's Table 19
shows that one serotype allele, designated G90, is confined thus
far to North American strains of variety 1: 7 of the 16 North
American strains examined possessed it, but it was found in none
of 17 strains from Japan, Europe, or South America. This is the
only shred of evidence I can find for the existence of racial differ-
ences in P. aurelia; but there is as yet little available information
pertinent to the subject.

The evidence that most varieties of P. aurelia are prevailingly
inbreeders has already been fully set forth. This, together with
microgeographic isolation, serves to go far toward making each
local population a species in the making. The setup is precisely
one which Wright (1940) considers so favorable for speciation.
Evolution is indeed in progress on a lavish scale in these organ-
isms. There must be many thousands or millions of such differen-
tiated local populations of P. aurelia in the world. Many or all of
them may be karyotypically diverse, so that laboratory crosses
among them would yield the characteristic F2 mortality. This is
one of the major things that makes the species problem so com-
plex and difficult in P. aurelia. And it is in this sense that the
species problem is tied so closely to the breeding system.

The species problem becomes more and more acute as one

236 PROTOZOA

passes from one breeding system to the next, in this sequence:
obligatory outbreeder, facultative out- and inbreeder, obligatory
inbreeder, exclusively asexual reproduction. Nearly the whole
sequence is found among the varieties of P. aurelia; they embrace
the range from nearly obligatory outbreeder to nearly obligatory
inbreeder. Exclusively asexual reproduction is not found, for sur-
vival depends upon fertilization or something associated with it.
The near-obligatory outbreeders have a wide geographic range
both in longitude and latitude and relatively few closely related
varieties of this sort are found in the same territory. The more
one approaches obligatory inbreeding, the larger the number of
varieties and the more restricted their range, especially in lati-
tude. Toward the end of the series, each local population becomes
almost as distinctive, differentiated, and isolated as a variety of
outbreeders. Correspondingly, the further one proceeds down the
sequence of breeding systems, the more difficult the species prob-
lem becomes until, in terms of the biological species concept, it
bursts into nothingness when the closest inbreeding becomes
obligatory or when sexuality is altogether abandoned.

Toward the end of this paper, we shall take up these extreme
situations. But first we turn to the other species of Ciliates in
which mating types are known. The points that have been made
in the detailed analysis of P. aurelia will serve as a touchstone for
orientation and interpretation in reviewing the status of knowl-
edge in the other Ciliates.

Paramecium caudatum

The Varieties of P. caudal urn and Their Distribution. In gen-
eral, the situation in P. caudatum bears many resemblances to
the one just set forth in detail for P. aurelia; but there is, as will
appear, at least one feature of special interest. Like P. aurelia,
P. caudatum consists of a number of varieties or syngens. Each of
these varieties has two mating types. At least 16 varieties have
been found: varieties 1 and 3, independently by Oilman (1939)
and by lliwatashi (1949); varieties 12 and 13 by Iliwatashi
( L949); and at least twelve more by Gihnan (1939,' 1941, 1949,
L9503 L954, L956a,b). Gihnan (1950) reported the relation of the

T. M. SONNEBORN 237

four varieties found by Hiwatashi to those lie had independently
found. Giese and Arkoosh (1939) found in western United States
a pair of interbreeding mating types which doubtless represented
a variety. Y. T. Chen (1944) found four varieties with two mat-
ing types each in China. Vivier (1955) found a variety with two
mating types in France. The relation of the varieties of Giese and
Arkoosh, of Chen, and of Vivier to those of Gilman has not been
reported.

Of the 16 numbered varieties, 1 and 3 have been found in
Japan and the United States, 12 and 13 in Japan, 2 in North and
South America, 8 in North America and Scotland, 15 and 16 in
Europe, and the rest only in the United States. Most varieties
seem to be scarce or absent in warm regions, but varieties 1 and
3 and probably 12 and 13 appear to be concentrated in warm to
moderate regions. Two varieties may have a very restricted range
( Gilman, 1956a and unpublished ) : variety 7 has been found only
on the island of Martha's Vineyard and neighboring Cape Cod;
variety 11 was found only once, in Connecticut.

Varietal Differences. Varietal differences are being investi-
gated by Gilman (1941, 1956b), but relatively little has been
published on this subject. There are some differences in the tem-
peratures that give optimal mating reactions. Varieties 4 and 5
are resistant, while varieties 1, 2, and 3 are sensitive, to the killer
strain H of variety 2 of P. anrelia. The animals of variety 3 appear
to be smaller than those of certain other varieties. A number of
characteristic varietal differences in the dimensions of the animals
and of the micronuclei appear to exist, but most of the details are
yet to be published. Thus far, it would be quite impossible to
identify a variety by any features yet made known, except their
mating types. Hence, the situation is in general the same as in
P. aurelia with respect to the question of whether the varieties
should be given specific names. This would be quite useless.

The Question of Gene Flow within and between Varieties.
Nine of the 16 varieties, which may be designated group 1, are
completely isolated from each other and from the other varieties
by inability to interbreed. Varieties 3 and 6, in one combination
of their mating types (VI with XI), give a faint sexual reaction

238 PROTOZOA

which does not lead to conjugation. These varieties, together with
varieties 1, 4, 5, 7, 11, 12, and 13, show no other sexual reactions
except between the two mating types of the same variety. Strains
of the same variety do interbreed, although little or no informa-
tion is in the record concerning later generations. Presumably,
however, each of these nine varieties constitutes a group of popu-
lations among which gene flow is possible, but between which
there is no gene flow.

The remaining seven varieties, 2, 8, 9, 10, 14, 15, and 16, con-
stitute what may be called group 2. (Gilman has spoken of it as
the "variety 2 complex.") Certain combinations of mating types
of different varieties in group 2 give sexual reactions and all the
varieties of this group are interlinked in this way directly or
indirectly. The intensities of these intervarietal reactions vary
from the weak sort, mentioned above as occurring between types
VI and XI, up to intensities equal to the maximal reaction nor-
mally obtained between complementary types of the same va-
riety. Nevertheless, according to Gilman, each mating type of
each variety has a uniquely defined pattern of mating reactions.

Whether the varieties of group 2 are sharply marked off from
one another genetically, as the varieties of group 1 are sexually,
seems to be as yet unsettled. Gilman has been studying this, but
has not yet published his results. Meanwhile, Pringle (1955) and
Johnson (1955) have been obtaining remarkable results on a
considerable number of European strains. All these strains belong
to group 2, but many of them have not yet been assigned to
definite varieties. Indeed Pringle and Johnson find it difficult or
impossible to sort their strains into varieties. Whether Gilman can
do this, as lie is now attempting to do, remains to be seen. The
two former workers have the impression that matings between
i titers of the same local population usually give a fair propor-

ii '»! viable V\ clones, whereas crosses between representa-

diflerent local populations give highly variable results

isually a less viable Fl. Pringle suggests that each local

popul ttion may have a gene pool which is to a considerable

degree isolated from the gene pools of other local populations.

Certainl) the results at present are difficult to fit into a system

T. M. SONNEBORN 239

in which there is free flow of genes among populations of a va-
riety, but none between varieties. The extent to which the same
difficulties exist in the American "varieties" of group 2 is un-
known.

The preceding relations constitute the feature of special inter-
est in P. caudatum to which reference was made at the start of
this section. It is the possibility to which Pringle referred, namely,
that group 2 or a part of it may be a group of differentiated local
populations more or less isolated by Fl mortality. This, as he
emphasizes, is different from the F2 mortality which serves to aid
in isolating local populations of some varieties of P. aurelia. It is
obviously most important for our understanding of evolution and
speciation in P. caudatum to have this situation clarified.

Breeding Systems. The observations on the European popula-
tions belonging to group 2, as already indicated, point to inbreed-
ing as the prevailing system in that material. There are other
indications that inbreeding is widespread among the varieties of
P. caudatum. With few possible exceptions, no immature periods
have been found thus far, though the methods Gilman has em-
ployed would not detect periods of immaturity of much less than
one month. Selfing is common in some varieties ( Gilman, 1941 )
and it may occur more often in cultures of one prevailing mating
type than in cultures of the other. The method of mating type
inheritance is apparently the kind characteristic of the group B
varieties of P. aurelia, at least in six of the seven varieties exam-
ined (Gilman, personal communication). This would assure the
presence of animals of both mating types in a synclone. All these
features speak for inbreeding. Of the varieties thus far examined,
all fall into the same pattern with the possible exception of vari-
ety 3. The latter may show a number of outbreeding features,
but the picture is not yet clear.

With such widespread occurrence of inbreeding features, au-
togamy might be expected. It, or what can now be interpreted as
autogamy, was reported by Erdmann and Woodruff (1916), by
Jollos (1916), and by Giese and Arkoosh (1939). On the other
hand, Gilman (1941 and unpublished), Pringle ( 1955), and John-
son ( 1955) have searched for autogamy and have been unable to

240 PROTOZOA

find it with certainty. These conflicting reports suggest that
autogamy may occur in some varieties and not in others; but fur-
ther investigation is called for. The relation of selfing to autogamy
seems also to differ in different materials. Giese and Arkoosh
(1939) found that autogamy was followed by the occurrence of
conjugation in a culture previously pure for one mating type;
autogamy had resulted in the differentiation of stable comple-
mentary mating types. Gilman (1941) studied cultures descended
from single individuals and could find no autogamy, although
selfing occurred within single cultures. It would seem that, as
with different varieties of P. aurelia, some varieties of P. cauda-
tum. differentiate mating types without prior autogamy, while
others do not. However, a special peculiarity in P. caudatum
seems indicated. Gilman ( 1939, 1941 ) observed that the cultures
which selfed, presumably without prior autogamy, in some cases
contained animals which, when isolated, gave cultures pure for
different mating types. In other cases, the selfing was apparently
due to transient change of mating type, for animals isolated from
the culture, even those that were coming together to mate, all
gave rise to the same mating type or to predominantly the same
type from each isolate. These varied observations indicate that
there may be several mechanisms in P. caudatum which bring
about mating between progeny of the same ancestral individual.
A different type of selfing was observed by Hiwatashi ( 1951 )
in variety 3, the variety that may be an outbreeder. He marked
cultures of the two mating types with different vital dyes and
observed that about 5% of the conjugant pairs obtained by mixing
the two cultures consisted of two animals of the same color. This
is selfing induced by mixture of pure types. It is not mediated by
detectable hormones in the medium. I doubt whether this sort ol
induced selfing plays much of a role in nature, for it depends
upon the formation of an agglutinated group consisting of at least
two animals of one mating type and one of the other. It is un-
likely that the formation of agglutinated groups occurs in the
diluted populations of natural waters to anything like the extent
observed with concentrated cultures in the laboratory. Yet this
sort o! selfing loo remains as a natural possibility.

T. M. SONNEBORN 241

In brief, nine of the 16 varieties, composing group 1, are sex-
ually isolated from each other and all others, and apparently they
conform to "biological species" or syngens. The other 7 varieties,
composing group 2, are not so sharply isolated from each other.
The status of these varieties is still unsettled. It is possible that
each local population is to a marked degree isolated from all
others. The inbreeding with which this extreme evolutionary
diversification is correlated is supported by a number of other
adaptations for inbreeding in both groups of varieties, though
some varieties may be outbreeders. Because of the widespread
close inbreeding, the species problem in P. caadatum is even
more difficult than in P. aurelia.

Paramecium bursaria

Mating Types and Varieties. Six varieties of this organism
have been reported: varieties 1, 2, and 3 by Jennings (1939),
varieties 4 and 5 by Jennings and Opitz (1944), and variety 6 by
Chen (1946b). The putative variety 5 is known from only one
collection, and its existence is suspected only from the fact that
the clones attributed to it do not mate with any other mating
type standards. However, they also did not mate with each other.
It is therefore dubious whether they were immature or impotent
members of a known variety or whether they were of a distinct
variety but all of one mating type. Because of the uncertain status
of variety 5, it will be largely ignored in the present account.

Of the five remaining varieties, only one, variety 4, has a sys-
tem of two mating types. The rest have systems of multiple mat-
ing types. Each of the varieties 1, 3, and 6 has four mating types,
and variety 2 has eight. The breeding relations are the same in
all: each mating type can conjugate with all the others in the
same variety, but cannot mate with animals of its own mating
type.

One and only one combination of different varieties can react
sexually and conjugate; this is the combination of varieties 2 and
4. One of the two mating types of variety 4, type R, reacts and
conjugates with four of the eight mating types of variety 2, the
types E, K, L, and M. Chen (1946a) and Sonneborn (1947)

242 PROTOZOA

pointed out that this indicates similarity between R of variety 4
and F, G, H, and J of variety 2, because all these mate with E,
K, L, and M; and between S of variety 4 and E, K, L, and M of
variety 2, because all mate with R. Obviously one cannot express
such similarities in terms of a basic binary system of plus and
minus types.

Metz (1954) suggests that mating reactions nevertheless de-
pend upon reactions between pairs of complementary mating
substances, even in varieties with multiple mating types. Accord-
ing to this hypothesis, there are three pairs of complementary
substances in a system of eight interbreeding mating types: alpha
substances A and a, beta substances B and b, gamma substances
G and g. Each mating type would have one alpha, one beta, and
one gamma substance. The formulas for the eight mating types
would therefore be: ABG, ABg, AbG, aBG, Abg, aBg, abG, and
abg. A system of two mating types would have only one pair of
substances, e.g., A in one type and a in the other. The mating
type with the A substance would thus be complementary to the
four mating types with the a substance in an eight-type set. This
hypothesis is consistent with the observation that one mating type
in the two-type system of variety 4 mates with four mating types
of the eight-type system of variety 2.

Isolation of the Varieties. The isolation of the varieties of P.
bursaria is partly sexual. All combinations of varieties, except 2
and 4, are completely unable to mate with each other. (Jennings,
1939, p. 223, remarks that very rarely a single pair of conjugants
is found when an immature clone is mixed with a mature clone
of a different variety. He suggests that such pairs are not crosses,
but sclfing in the mature clone. )

In the exceptional combination of variety 4 with variety 2
Jennings and Opitz (1944) found that all four of the possible
reactive combinations of mating types of these two varieties (R
with E, K, L, and M) led to the same result: most of the con-
jugants died without separating; those that eventually separated
died within four days without undergoing a single fission. Chen
(1946a) examined in detail the cytology of the invariably fatal
(losses. Although the nuclear processes proceed normally for a

T. M. SONNEBORN 243

considerable period, abnormality suddenly develops. The chro-
mosomes clump, meiosis does not proceed beyond the first divi-
sion, and the nuclei acquire abnormal shapes and behavior. These
abnormalities are not consequences of an intervarietal hybrid
genotype, for the nuclear processes do not reach the stage of
fertilization. As Chen pointed out, death is due to an interaction
between the conjugants, which is dependent upon contact and
possibly upon migration of substances from one mate to the
other. The situation has features in common with the mutual
nucleocytoplasmic incompatibility reported by Melvin (1949)
and Levine ( 1953 ) for the cross between varieties 4 and 8 of P.
aurelia. Melvin reported that when cytoplasmic exchange was
induced (by exposure to antiserum), the pairs died before sepa-
rating. The isolation of varieties 2 and 4 of P. bursaria is also
probably to be understood as due to mutual nucleocytoplasmic
incompatibility.

A viable Fl is thus never achieved in crosses between the
known varieties of P. bursaria. On the other hand, as in P. aurelia,
different strains of the same variety can interbreed freely and
yield (under proper conditions, see below) viable, normal prog-
eny. Hence, the varieties of P. bursaria, like those of P. aurelia,
conform to the modern biological concept of species. In our
terminology, each variety is a syngen.

Varietal Differences. As with P. aurelia and P. caudatum, the
question now arises as to whether the varieties of P. bursaria
could be identified without recourse to standard living cultures
of the various mating types and if so, whether such identification
is sufficiently practicable to warrant assigning species names to
the varieties. In order to answer this question, the varietal differ-
ences will be summarized.

1. Geographical distribution. Variety 1 occurs in China and
the United States ( Chen, 1956 ) ; varieties 2 and 3 in the United
States; varieties 4, 5, and 6 in Europe. Barring new extensions of
range and the discovery of additional varieties in the regions
explored, the problem of identification would be much simpler
in P. bursaria than in the other organisms thus far considered, for
not more than three varieties are known from any continent. If

244 PROTOZOA

the natural source of a strain is known, its identification is seem-
ingly limited to but few possibilities.

2. Karyotypes. Chen (1946b) reported that the chromosomes
of varieties 2 and 4 are thin and short while those of variety 3 are
larger and much longer. However, similar differences distinguish
strains of variety 6 collected from different parts of Europe, Eng-
land and Czechoslovakia. The usefulness of this criterion is ham-
pered not only by the possibility of differences within a variety,
but also by the fact that one must obtain conjugants to examine
the chromosomes satisfactorily. Late prophase or prometaphase
of the first meiotic division is the stage employed.

3. Number of mating types. As stated above, varieties differ
as to whether they have 2, 4, or 8 mating types. With regard to
the usefulness of this criterion, the pertinent question is: Can
the number of mating types in a variety be readily ascertained?
Success in answering this question depends upon finding at least
two different mating types in the variety, for Jennings ( 1941 )
reports that all types of a variety can be obtained among the
sexually produced descendants of any two. The most reliable way
of finding two types is to make a number of collections from
nature in the same region. If only one type is found in nature, it
is still possible to obtain all types, but this depends upon the oc-
currence of a rare event: self-differentiation of an additional
mating type in a culture previously pure for one mating type.
According to Jennings (1941), this happens on the average once
in about 2000 culture-days. Once a culture yields a second type,
conjugation between the two types can yield all the others in the
variety. Lee (1949) reported that exposure to high doses of x-rays
induced selling in a previously pure culture. He suggests that
the irradiation induced self-differentiation of the sort that occurs
rarely without irradiation. If this proves repeatable and general,
a method of finding all types in a variety would be in hand.
However, the process of doing this would still take much time
and labor because immaturity is apparently a regular occurrence
alter conjugation, and it may last a long time. Incidentally the
difficulty of getting all types from one (without Lee's technique )
is attested by the fact that in years of observation the one collec-

T. M. SONNEBORN 245

tion called variety 5 never yielded more than one mating type.

4. Other differences. At 26.5°, conjugation is completed in
20 to 22 hours in variety 1, but it lasts for 32 hours or more in
varieties 2, 4, and 6. Variety 3 shows no daily periodicity in sexual
reactivity; but varieties 1 and 2 are most reactive around noon
and are not reactive at night (Jennings, 1939; Wichterman, 1948;
Ehret, 1953). Reliable varietal differences in size, shape, or fine
anatomy have not been reported. Strains of the same variety ap-
pear to differ more commonly in size, shape, and certain physio-
logical traits than do strains of the same variety in either P. aure-
lia or P. caudatum,

5. Status of the problem of varietal identification. All the
differences among varieties occurring in the same region involve
mating types and conjugation. This imposes great labor and long-
continued laboratory investigation on the part of anyone who
sets out to identify a strain and therefore is, in my opinion, not
suitable as a basis for species naming. However, it is true that
with labor and time the varieties could be identified without
recourse to living standard cultures. The requirements are per-
haps not so great as they would be for a comparable accomplish-
ment in P. aurelia or P. caudatum, but they are still much too
great for routine taxonomic purposes.

Life Features and the Breeding System. The life features of
P. bursar ia form a striking contrast to those which characterize
most varieties of P. aurelia and P. caudatum. The differences are
by no means fortuitous. They are all clearly understandable as
adaptations to or consequences of a different breeding system. In
brief, P. bursaria is an outbreeder, while most varieties of the
other two species are inbreeders to a greater or lesser degree. In
this frame of reference, the life features of P. bursaria and their
contrast with those of most varieties of the other two species
make excellent sense.

The life history of P. bursaria is documented in great detail in
a series of papers by Jennings ( 1944a,b,c,d; 1945). Most of this
work was done on variety 1. My comments are therefore to be
understood as applying chiefly to that variety. There is no indica-
tion in these papers of any differences in the major life features

246 PROTOZOA

among the varieties, but that possibility needs to be kept in mind.
Particularly, one would like to know more about variety 3, which
is rarer and less widely distributed; and about variety 4, which
has only two mating types. A priori, I would expect that these,
if any, varieties might be less committed to outbreeding. The
main features of life and the breeding system in variety 1 follow.

1. Immaturity. Conjugation is invariably followed by an im-
mature period. It never lasts less than twelve days and usually
lasts three to five months when the animals are well nourished.
If the organisms are placed for a time under poor nutritive condi-
tions during immaturity, this may disproportionately delay the
onset of maturity. For example, 18 days of such conditions, with
good nutritive conditions during the rest of the time, resulted
in a total immature period of 10 to 14 months, and some clones
failed to mature during a period of years. Translating these labo-
ratory findings into what might be expected in nature, immatu-
rity probably lasts at least a year, for under natural conditions
poor nutritive conditions are hardly avoidable for long and this,
combined with the lower fission rate usually to be expected,
would extend immaturity at least to the 10-to- 14-month range.
This feature of its life is alone sufficient to mark P. bursaria as an
outbreeder for the features that could balance it are lacking, as
will be shown.

2. Maturity and mating types. The capacity to mate develops
gradually and reaches a peak at which it remains for very long
periods. In the next paragraph I shall define what brings maturity
to an end. Under laboratory conditions, maturity last about 3
years. In nature, it probably lasts much longer because of the
lower reproductive rate. As was mentioned in connection with
the outbreeding varieties .15 and 16 of P. aurelia, a very long
period of maturity is an adaptation to increase the probability
of finding a suitable mate. Further, like variety 16 of P. aurelia,
the varieties of P. bursaria (except variety 4) have multiple
mating types which serve to increase the probability of mating
when a mature stranger is encountered. Finally, unlike any va-
riety of P. aurelia, in about 97.4% of the cases all four caryonides
of a synclone are alike in mating type (Jennings, 1942). More-

T. M. SONNEBORN 247

over, caryonides that mature as selfers are totally unknown in
P. bursaria. Both the absence of selfing caryonides and the usual
synclonal uniformity in mating type serve to reduce the probabil-
ity of mating between close relatives.

3. Senility, autogamy, and selfing. Senility is marked by de-
creasing fertility, as in other species. However, unlike most vari-
eties of P. aurelia, variety 1 of P. bursaria is able to mate readily
throughout senility, almost until the clone dies. The probability
of death after conjugation is low, usually less than 10%, during
maturity; but with the onset of senility, as a rule at an age of 30
to 40 months, the probability of death after conjugation begins
to rise and continues to rise, often eventually approaching 100%.
There is much variation in the details from clone to clone, but
representative results are about 80% mortality after conjugation
at 4 years of age and 100%; or nearly so at 5 to 5/2 years of age
and thereafter. In the laboratory the period of senility lasts 4 to 5
years, making the total life cycle 8 to 9 years. (Of course many
clones are weak and die immediately or soon after conjugation;
the present account refers only to clones of full vigor. )

Comparable decreasing fertility with age is found in other Cili-
ates. Most varieties of P. aurelia, those that undergo autogamy,
exhibit an age-correlated mounting death rate after autogamy;
and varieties 15 and 16 show the same after selfing during senil-
ity. The varieties that undergo autogamy during senility can also
conjugate during that stage, but conjugation is difficult to obtain.
In this case, the parallel to the situation in P. bursaria is exact,
for the probability of death after these conjugations also increases
with the advance of senility.

The long period during which conjugation, but only cross
conjugation, can occur in the life cycle of P. bursaria, 7 to 8 years,
suggests how rare opportunities for crossing may be in nature.
Long ago Weismann suggested that the length of the reproduc-
tive period in organisms is adjusted by natural selection to permit
the production of just about a sufficient number of offspring to
perpetuate the species. Even allowing for sexual progeny that fail
to survive either because of inherent defects or because of extinc-
tion due to external causes, Weismanns suggestion, if valid,

2 18 PROTOZOA

would lead to the conclusion that P. bursaria "needs" a long re-
productive period because sexual reproduction rarely occurs.

Senility in P. bursaria may be marked by another feature,
which also has its analogy in the other species that have been
discussed. This is the phenomenon, mentioned earlier, which
Jennings (1941) discovered and called "sell-differentiation": a
clone previously pure for one mating type comes to contain two
mating types which can be isolated and shown to reproduce true
to type. Self-differentiation is a very rare event, occurring once in
about 2000 culture-days. It was not assigned to a definite stage in
the life cycle by Jennings; the available data are actually not
sufficient to make such an assignment with assurance. The uncer-
tainties are due to two facts. First, many of the observed cases
occurred in clones begun with a wild individual of unknown age.
Second, the clones of known age, in which self-differentiation
occurred, are heterogeneous with respect to the kind of matings
that produced them and there may be life cycle differences asso-
ciated with this. For example, clones produced by conjugation
between two mating types that arose by self-differentiation
within a single clone may themselves reach earlier than usual the
stage at which self-differentiation can occur. There is some indi-
cation of this in the data of Jennings, but the matter has not been
systematically explored. Tentatively, and largely by analogy with
other species, self-differentiation in P. bursaria may be considered
as a feature of senility which assures rejuvenescence for clones
that have failed to find suitable mates, by permitting conjugation
to occur within the clone.

Jennings (1941) assumed that self-differentiation was probably
a consequence of prior autogamy. There were two reasons for
this assumption. First, the differentiation of stable mating types
within a single line of descent is due to autogamy in /'. aurclia.
Second, Erdmann ( 1927) briefly noted the occurrence of nuclear
reorganization within a clone of P. bursaria, and her observations
have been later interpreted as representing autogamy. However,
neither ol these reasons now seems cogent to me. Self-differentia-
tion oi mating types within a caryonide in the absence ol autog-
amy has been demonstrated in varieties 15 and 10' of P. aurclia

T. M. SONNEBORN 249

and in P. caudatum. In the latter, stable types may arise in this
way. In the former, the differentiation can occur in variety 15 in
amicronncleate caryonides which are therefore obviously incapa-
ble of undergoing autogamy. Further, Erdmann's observations
were fragmentary and not well documented, and all could have
been stages of reorganizing exconjugants, the conjugation of
which had been overlooked. Finally, Chen's long and careful
observations on P. bursaria have never included a single instance
of autogamy in unpaired, i.e., non-mating, animals. It might be
possible now, if Lee's (1949) method of x-irradiation actually
induces self -differentiation, as he suggests, to put to a direct test
the question of whether autogamy precedes self-differentiation.
Previously, this was impossible because of its great rarity and
unpredictability. Meanwhile, until better evidence is forthcom-
ing, autogamy in single animals must be considered as not only
undemonstrated in P. bursaria, but as probably nonexistent.

Thus, as matters now stand, P. bursaria conforms to the pattern
of outbreeders by lacking autogamy and by using in its place self-
differentiation and conjugation between the mating types thus
differentiated, to restore vigor to clones that have reached senility
without finding opportunities for outcrossing.

4. Survival after inbreeding and outbreeding. The comments
made above about the relation of parental age to the frequency
of death after conjugation are correct, but the situation is greatly
complicated by the fact that other factors also influence the death
rate after conjugation. A most important factor is the relationship
of the parent clones themselves. In general, with one exception
to be noted later, outcrosses yield better survival than crosses be-
tween relatives. Jennings noted 5 to 14 or more times as much
death after conjugation between sibs (i.e., crosses between dif-
ferent synclones derived from the cross of the same two parental
clones) as after conjugation between clones derived from differ-
ent natural populations. Such comparisons were of course made
with clones of similar age. These relations obviously reinforce the
conclusion that P. bursaria is an outbreeder. They suggest that
the organisms are normally heterozygous for lethal and detrimen-
tal genes which come to expression through inbreeding and are

250 PROTOZOA

masked to a considerable degree by outbreeding. Jennings' efforts
to carry on series of successive inbred generations by sib matings
were frustrated by rapidly increasing death rates which ap-
proached 100%.

The exception referred to above concerns the closest possible
inbreeding, matings between the two types that self-differentiate
within a clone. This will be referred to as selfing. No greater
mortality appeared after such selfing than when one or the other
type was outbred. However, there is some indication that the
clones produced by selfing age more rapidly than clones produced
by outbreeding: the former may give at ages of 12 to 18 months
as much death after conjugation as the latter do at ages of 3 to 4
years. This, like the earlier occurrence of self-differentiation, in-
dicates shorter stages in the life cycle after selfing than after out-
crossing. The higher death rate from sib matings than from
selfing or outcrosses led Jennings to remark, "There is in clonal
self-fertilization some biological relation that prevents it from
resulting in high mortality" (Jennings, 1944c, p. 195).

What this biological relation may be is of great interest, and
this will be considered later. Here it suffices to note that it oper-
ates adaptively. The organism, as an outbreeder, survives in-
breeding but poorly; yet, to survive at all, it requires some form
of inbreeding as a last resort to forestall death of those members
of a clone that have failed to find a suitable mate. It has managed
to accomplish this by facilitating selfing through self-differentia-
tion and it has somehow endowed this saving sort of mating witli
protection against the usually disastrous effects of other sorts of
inbreeding. Whether something like this also operates in the
sellings that occur during senility in variety 15 of /\ aurelia needs
to be investigated. It is even possible that a similar basis under-
lies the perplexing observations of Pringle (1955) and Johnson
(1955) on inbreeding and outbreeding deaths in P. caudatum.

The Mechanism of Mating Type Determination and Its Bear-
ing on the Breeding System. Jennings (1942) maintained that
the mating types are determined by the chromosomal constitu-
tion. He based this conclusion on two facts: (1) conjugation
brings about identical genotypes in the two clones of a synclone;

T. M. SONNEBORN 251

(2) in nearly all cases, these two clones are alike in mating type.
Yet, when he tried to formulate genotypes which would agree
with the observed results of crosses, he was totally unable to do
so. The chief difficulty was that synclones of all possible mating
types were producible from the cross of any two mating types,
indeed from any single type which underwent self-differentiation.
For, like other matings, selfings also could produce all possible
types of synclones.

Sonneborn (1947, p. 340) pointed out that the inheritance of
mating type in P. bursaria conforms in principle to the group B
pattern in P. aurelia when there is massive exchange of cytoplasm
between mates. The only difference is that there are four or
eight alternative types in most varieties instead of just two. Son-
neborn (p. 335) further cited three evidences that cytoplasmic
exchange between mates does in fact occur regularly in P. bur-
saria. ( 1 ) Harrison and Fowler ( 1946 ) observed transfer of anti-
gens from one mate to the other during conjugation. (2) As men-
tioned earlier, the abnormalities that occur during conjugation
between varieties 2 and 4 appear to be due to interactions be-
tween the nucleus of each variety and the cytoplasm of the other,
and they occur before nuclei are exchanged ( Chen, 1946a ) . ( 3 )
Jennings ( 1944b ) observed that conjugation between an old and
a young clone usually kills both mates, abnormalities appearing
during conjugation itself. The evidence subsequently obtained
by Sonneborn and Schneller (1955) that nuclear abnormalities
are induced by old cytoplasm reinforces my interpretation of the
dependence of death of the young mate in Jennings' crosses on
the action of cytoplasm received from the old mate. Since cyto-
plasmic exchange is relatively rare in P. aurelia, usually the old
mate dies and the young one lives. To these evidences may now
be added the fact that Chen's numerous figures of conjugation
stages indicate the regular occurrence of a wide region of free
cytoplasmic continuity between the mates. There can thus be
little doubt that massive exchange of cytoplasm between mates
is the rule in P. bursaria.

This then provides the explanation for the uniformity of mating
type in a synclone. Nanney (1956b) has come to the same con-

PROTOZOA

elusion and shows in detail how P. bursaria fits the group B sys-
tem. Exposure of the four macronuclear anlagen to a common
cytoplasm differentiates all of them to control the same mating
type. If my formulation of the operation of the cytoplasmic action
in group B is correct, then there should be four, instead of two
cytoplasmic differentiators in flux equilibrium in variety 1 of P.
bursaria. Of the four, one is in high concentration in one mate,
another in the other mate; the mixture of these two cytoplasms
thus gives high concentrations of two of the four. This accounts
readily for the fact that usually over 80% of the synclones pro-
duced from a cross are like one or the other parent in mating
type, and that, of these, roughly half are usually like the one
parent and half like the other parent. Whether temperature can
affect these proportions is unknown. The hypothesis also explains
some other remarkable observations, as will now appear.

The phenomenon of self-differentiation has this peculiar fea-
ture: it always results in the same two types in any one clone,
but results in different types in different clones of the same
original mating type. That is, a clone of type A may differentiate
into A and B; another clone of type A may differentiate into A
and C. Yet if either the new A or the new B in the first clone
self-differentiates again, the A will always yield B and the B will
always yield A. Likewise, in the other clone, the new A and C
will always differentiate only into C and A, respectively. All
combinations of two types arc possible in different clones, but
any one clone has only one such combination of possibilities.

The physical basis of these remarkable relations has never been
elucidated. Kimball (1943) suggested that each possible com-
bination of alternative mating typos producible by self-differen-
tiation was determined by a different genotype. He agreed with
Jennings' conclusion that the usual uniformity of mating type in
a synclone bespoke genie determination and suggested two rea-
sons for Jennings' failure' to find workable genotypic formulas.
First, each formula should be not for a particular mating type,
but for a particular pair of alternative types. Second, the formu-
las should take into account Chen's (1940) finding of polyploidy
in P. bursaria. Kimball's proposals have not thus far led to the

T. M. SONNEBORN 253

further analysis he called for. In my opinion, his suggestion is
unsatisfactory, although the genotype probably does play some
part in the system. The chief difficulties with Kimball's proposal
are that it provides no insight into the basis for the choice be-
tween the two alternative types in each synclone or into the
mechanism by which either of the alternatives can be inherited
through great numbers of fissions.

These difficulties are precisely the ones which are accounted for
by the analysis of the group B method of mating type inheritance
in P. aurelia. In extending this to variety 1 of P. bnrsaria, four
instead of two cytoplasmic differentiators are assumed in order
to meet the requirements of four instead of two mating types. In
a steady state system, these four would stand in a hierarchy of
relative concentrations in each mating type, and the one initially
in highest concentration would be the one corresponding to the
expressed mating type. However, the relative concentrations of
the other three might well vary in different synclones of the same
mating type. With such a system of cytoplasmic constitutions,
the cytoplasmic differentiator present in second highest concen-
tration could be the one which determines the alternative mating
type producible by self-differentiation. On this view, there are in
each clone very rare occasions on which the two most concen-
trated cytoplasmic differentiators exchange relative positions in
the hierarchy. Then, when age renders the macronuclei suscep-
tible to redifferentiation, those animals which have the changed
hierarchy of differentiators change mating type. As in P. aurelia,
the determination of the specific hierarchy in any synclone or
subclone is a resultant of the action of the genotype and of ex-
ternal conditions.

This is perhaps not the appropriate place to develop in detail
how the previously unexplained peculiar features of mating type
inheritance in P. bursaria become intelligible in the light of this
approach to the problems; but, to indicate the fruitfulness of the
hypothesis, one further example will be cited. Jennings ( 1942,
Table 2) lists five crosses between mating types B and C, with
different clones involved in each cross. One of the five crosses
stands out as giving unusual results. Instead of the expected re-

254 PROTOZOA

suit, namely a 1:1 ratio of synclones of type B to synclones of
type C, Jennings reported 39 type B and none type C. The clue
to this exceptional result, in my opinion, is found in Jennings'
paper of 1941 which states that clone Lo 12, which was the type
C parent in the cross just mentioned, self-differentiated into types
C and B. On my hypothesis, the production of type B to the
virtual exclusion of other alternatives is a direct consequence of
mixture of cytoplasm between two mates, one of which had the
B cytoplasmic differentiator in highest concentration and the
other of which had it in second highest concentration, as indi-
cated by the type to which it could differentiate.

In like manner, the occasional production of synclones differing
in mating type from either parent can be explained. These new
types often correspond to the second most concentrated cyto-
plasmic differentiator in one or the other or both parent clones,
as inferred from the type to which the parent self-differentiates.
Although this information is available for only a minority of the
parents reported upon by Jennings, in these cases the agreement
with expectation is on the whole impressive. However, there are
also a few instances of apparent disagreement. The outstanding
exception appears in one of the five crosses of mating types B and
D. This yielded a 1:1 ratio of A to D instead of B to D. Since the
B parent self-differentiated into C and the D parent arose by self-
differentiation in a B clone, my hypothesis leads to the expecta-
tion of an excess of B progeny, whereas in fact B progeny were
totally lacking. However, it is to be noted in this case that the D
parent was derived from an inbred ancestor which had for a few
generations repeatedly differentiated only into A and D, the B
arising as a late exception (Jennings, 1941). Here then is a pos-
sible case of a less probable and stable cytoplasmic state revert-
ing to a more probable and stable equilibrium. In view of the
obvious complexities of the general system, involving a fourfold
steady state system acted upon by genie and environmental fac-
tors, I am of the opinion that an apparent exception of the sort
just mentioned is not serious enough to invalidate the hypothesis.
What is now needed is a number of experimental tests directed
specifically toward disentangling the various components of the

T. M. SONNEBORN 255

system as formulated in the hypothesis. Many such tests come
at once to mind. Meanwhile, this hypothesis is formally capable
both of accounting for most, if not all, of the known kinds of
facts and of leading to testable predictions.

One aspect of the breeding results still remains as puzzling in
the extreme. This is the contrast between the high viability after
selfing and outcrossing and the low viability after crosses between
sib clones. I cannot suggest a satisfactory explanation, but atten-
tion might be called to one point: selfing occurs between two
types which, on my hypothesis, are necessarily as much alike in
cytoplasmic constitution as two different mating types can be:
the most concentrated cytoplasmic differentiator in each is the
same as the second most concentrated differentiator in the other.
How this could influence the viability of the resulting synclones
is far from clear, but it would be of interest to compare mortality
in sib crosses and outcrosses between clones that have this sort
of cytoplasmic relation and clones that are more diverse in cyto-
plasmic constitution.

Evolution, Speciation and the Breeding System. The evolu-
tionary relation between the systems of mating type determina-
tion in P. bursaria and P. aurelia are close, and the differences
are simple. In principle, the system in P. bursaria is the most
complex, the one in the group A varieties of P. aurelia is the
simplest, and the one in the group B varieties is intermediate.
The single unique feature in P. bursaria is the regularity of mas-
sive cytoplasmic exchange between mates. In the group B varie-
ties of P. aurelia this feature occurs only in a small minority of the
conjugating pairs. In group A, the cytoplasmic role in mating
type determination is inapparent because of the lack of evidence
of flux equilibrium and the overriding influence of temperature
and of a genotypic action which is the same in both mating types.
All other features are found developed to varying degrees in
both species: caryonidal differences, nuclear differentiation, cyto-
plasmic differentiators and the genes affecting their concentra-
tion, dual and multiple mating type systems, and mating type
differentiation sometimes associated with and sometimes inde-
pendent of a preceding fertilization.

256 PROTOZOA

The regular exchange of cytoplasmic differentiators during
conjugation in P. bursaria is one of its key adaptations to out-
breeding. It almost completely prevents the occurrence of mating
within a synclone during maturity: only 2.6% of the synclones
are exceptions to this rule. This feature, the long period of
immaturity, the long life cycle, multiple mating types and the
high mortality from sib crosses, assure a highly developed system
of outbreeding. Variations from population to population occur,
but the populations are prevented from progressive divergence
by the forced interbreeding. A common gene pool is thus main-
tained over a wide range.

As a consequence, speciation occurs on a much smaller scale
than in the inbreeders, P. aurelia and P. caudatum. Only 3 va-
rieties of P. bursaria have been found in the United States; they
were discovered by 1939, within two years after the discovery
of mating types, and no additional ones have been found in the
large number of collections subsequently examined. At the same
time, the first three varieties of P. aurelia were found; but in the
following years ten additional varieties were discovered in the
United States. Probably few if any additional varieties of P.
bursaria will be found in this region; but doubtless more, possibly
many more, varieties of P. aurelia will be discovered, for several
new ones turned up within the past year. The difference is ac-
centuated by the existence of karyotypic differences between lo-
cal populations of P. aurelia and their regular production of F2
mortality after crosses between representatives of different popu-
lations. Speciation is indeed rich in the inbreeder, poor in the
outbreeder.

Tetrah ymena pyriformis

Mating Types and Varieties and Their Distribution. Like the
other Ciliates which have been studied in this connection, T.
pyriformis consists of a number of sexually isolated varieties.
Elliott and Gruchy (1952) discovered variety 1; Gruchy (1955)
reported varieties 2 through 8; and variety 9, which seems not to
occur in North America, was found by Elliott and Hayes (1955)
in Colombia and in the vicinity of the Canal Zone. In the latter

T. M. SONNEBORN 257

places and in Mexico, they also found variety 2, which is widely
distributed from Canada to Central America. The other seven
varieties are known only from North America. Gruchy (1955)
observes that varieties 1, 4, 5, and 8 have been found only in the
northern states; that variety 3 seems to occur mainly in running
water; and that varieties 1, 5, 6, and 8 have been found only in
standing water. In striking contrast to the varieties of P. aurelia,
of which two or more are often found together in the same body
of water, the varieties of T. pijriformis usually occur alone, two
varieties in the same collection having been found only once
(Elliott and Hayes, 1955). Altogether, cultures from about
eighty collections in the Americas have been identified as to
variety; no collections from other continents have yet been re-
ported upon.

The varieties of T. pijriformis are diagnosed primarily on the
basis of sexual reactions. If a clone from a new collection mates
well with a standard mating type culture, it is assumed to belong
to the same variety as the standard. Variations in intensity of the
mating reaction appear not to occur under optimal cultural con-
ditions (Gruchy, 1955). In the case of P. aurelia, it will be re-
called, varieties 4 and 8 interact sexually with almost normal
maximal intensity. It is therefore possible, but perhaps not likely,
that in T. pijriformis some different varieties have been, on the
basis of strong intermating, misclassified as belonging to the same
variety. Tests of genetic isolation, which have proved important
in P. aurelia and P. bursaria, have not yet been carried out ex-
tensively in T. pijriformis. (There are practical difficulties here,
as will appear. ) For this reason, the status of the nine reported
varieties, particularly in relation to the problem of gene flow
with which we are concerned here, is still dubious. However, it
seems clear that there are at least nine varieties among the ma-
terials examined, for these are all reported to be isolated by
failure to intermate. The study of gene flow within these varie-
ties could serve only to increase their number, if it changes the
situation at all.

Another systematic exclusion of possible additional varieties
comes from the practice of ignoring all isolates from nature which

258 PROTOZOA

prove to be selfers. Gruchy (1955) reports that about 5% of the
wild strains collected were selfers. If some varieties regularly
self, like variety 13 of P. aurelia, these would be overlooked."'' At
present there seems to be no satisfactory way to get around this,
for it has proved difficult or impossible to discover the variety
to which such strains belong. One wonders whether the tech-
nique of using the animals left over after a culture selfs and
before it is fed again would work, as it does with variety 13 of
P. aurelia. Nanney (1953) tried another and ingenious approach
to the problem in wild selfers of another species of Tetrahymena,
but without success. These selfers gave 100% death of exconju-
gants. On the assumption that exconjugant survival might be im-
proved by outcrosses, he mixed selfers derived from different
sources and examined the viability of the conjugants isolated
from the mixtures. No difference was observed from the results
obtained with selfmg conjugants from a single source. This should
not discourage further use of the method since it is the result
expected if the two strains of selfers belonged to different va-
rieties. Perhaps other combinations of selfers would give a differ-
ent result, indicating that they belong to the same variety. The
method could even be profitably extended to mixtures between
a selfer and a pure type.

Varietal Differences. The nine varieties appear to be alike in
many respects, even in some that distinguish varieties of other
species of Ciliates. No diurnal mating periodicities have been
found, and differences in temperature requirements for mating
seem not to exist (Elliott and Hayes, 1953; Gruchy, 1955). The
haploid chromosome set consists of one large, one small, and
three intermediate chromosomes, all with median or submedian
centromeres; and this seems to be the same in all varieties that
have been examined, i.e., in all but varieties 3, 7, and 8 (Ray,
1954, 1956; Ray and Elliott, 1954; Gruchy, 1955). There arc,
however, some differences among the varieties. Differences in
distribution and in preference for running or still water were
mentioned earlier in this section; others follow.

1. Size, morphology, and cytology. According to Gruchy
(1955), animals of varieties 3 and 7 are unusually large, and

T. M. SONNEBORN 259

those of variety 5 are unusually small. Gruchy further remarks
that Corliss found other peculiarities in varieties 3 and 7, the
cilia being arranged in fewer rows and part of the meridians
appearing heavier and more definite; but earlier, Corliss (1954)
mentioned that he could find no differences in structural details
among the varieties. Presumably, these differences are statistical
and not suitable for identification purposes.

2. Mating types, the mating reaction, and mating type inherit-
ance. The most striking differences among the varieties are in
the number of mating types they contain. There are two each
in varieties 5, 7, and 8; three each in varieties 4 and 6; five in
variety 9; seven in variety 1; eight in variety 3; and eleven in
variety 2 (Elliott and Gruchy, 1952; Elliott and Hayes, 1953;
Nanney and Caughey, 1953, 1955; Elliott and Hayes, 1955;
Gruchy, 1955). The system of possible matings within a variety
follows the patterns of P. aurelia and P. bursaria: no type mates
with other clones of the same type; each type mates with all the
others of the same variety. Of course, if any of these varieties
proves by genetic tests to be more than one variety, that would
reduce the number of types in the varieties involved.

On the other hand, there may well be more mating types in
some of these varieties than are now known. In other Ciliates,
the progeny test has been used to discover whether additional
mating types can occur. In T. pyriformis, primary reliance has
thus far been placed on finding the types in nature, although
some effort has been made to use the progeny test. Its importance
and also its limitations have been shown for variety 1 by Nanney
and Caughey (1953) and Nanney, Caughey, and Tefankjian
(1955). From crosses between the two original strains of type
I and type II, they obtained among the progeny types III through
VII. Only types I, II, and III have been found in nature so far
(Elliott and Gruchy, 1952; Elliott and Hayes, 1953; Gruchy,
1955). But Nanney and his co-workers have also shown that
clones of different genotypes are restricted as to the mating
types they can produce among their sexually derived progeny.
One genotype cannot produce type I; another cannot produce
types IV or VII. Thus, even the progeny test is limited, and the

260 PROTOZOA

full number of types in some varieties may not be known until
similar studies have been made on crosses among a number of
strains of the variety from different sources. The matter is fur-
ther complicated by the low frequency with which certain types
appear among the progeny and the somewhat different fre-
quencies of types under different conditions (Nanney, Caughey,
and Tefankjian, 1955). All these difficulties will of course be re-
solved in time. Meanwhile, Gruchy (1955) has tried to look for
new types from crosses and found a new type in variety 6 in
this way; but the difficulties of immaturity and viability have
made this impossible in some varieties.

Unlike Paramecium, the animals in all varieties of T. pyriformis
fail to agglutinate or react immediately when complementary
types are mixed. There is a refractory period of at least an hour
or two and, in variety 9, of 24 to 36 hours; then pairing is direct
without agglutination (Elliott and Gruchy, 1952; Elliott and
Hayes, 1953; Nanney and Caughey, 1953; Ray, 1955). This has
raised in several investigators' minds the question of whether
there may be sex hormones which induce selling. Elliott and
Hayes ( 1953 ) presented evidence that selling is not induced in
this way and proof was provided by Nanney and Caughey
(1953). Nevertheless, the latter do not totally exclude the pos-
sibility of some sort of sex hormone action; but the former and
Gruchy (1955) hold that there is no action of fluid in which one
type has lived on animals of the other type. It would be interest-
ing to know whether exposure to such fluids could reduce the
refractory period. In view of Rays (1955) report of a 24-to-36-
hour refractory period in variety 9, perhaps this would be the
variety of choice for such a test.

The method of mating type inheritance is known only for va-
rieties 1 and 2. The work on variety 2 has just begun, but Hurst
(unpublished) has made the important discovery that this variety
exhibits the same pattern of mating type inheritance as P. bur-
saria; that is, a synclone is usually uniform with respect to mating
type, some synclones from a cross having the mating type of
one parent and some having the mating type of the other parent.
This indicates that the basic mechanism is the same as in group

T. M. SONNEBORN 261

B of P. aurelia, but with regular exchange of cytoplasm between
mates. The existence of 11 mating types in variety 2 would, if
our hypothesis is correct, indicate a flux equilibrium among 11
cytoplasmic differentiators. Although this might seem extraordi-
nary and unduly complicated, it should be remembered that equal
complexity in flux equilibrium is present in the serotype system
of the group B varieties of P. aurelia. Nanney (1956b) has
pointed out the detailed resemblances and differences between
the mating type and serotype systems. Variety 2 of T. pyriformis
may well prove to be the material of choice for testing the inter-
pretations proposed above for P. bursar ia.

Mating type inheritance in variety 1 has been analyzed by
Nanney and Caughey (1953, 1955), Nanney, Caughey, and
Tefankjian (1955), and Nanney (1956b). In contrast with va-
riety 2, variety 1 shows the group A pattern of mating type
determination and inheritance. Leaving selfers aside for the mo-
ment, each caryonide seems to be independently determined for
mating type. Hence, among the four caryonides from a pair of
conjugants, there are usually present hereditary differences in
mating type; in fact, as expected, even the two caryonides from
a single exconjugant are usually different in mating type. With
seven possible mating types, a large number of combinations is
found among the two caryonides from an exconjugant or among
the four from a pair of conjugants.

There are, however, genie restrictions as to which mating type
potentialities can be realized. One gene permits the development
of any type except I; another gene permits the development of
any but IV and VII. These genes behave as alleles. The authors
point out some difficulties raised by these discoveries with respect
to Metz's scheme of pairs of complementary mating type sub-
stances (see earlier discussion under Paramecium bursaria), but
realize that these difficulties need not be fatal to Metz's hy-
pothesis. The mating type potentialities may be modified by
genes at different loci which act in different ways.

3. The life cycle. There appear to be some varietal differences
in the length of the immature period, but the literature contains
conflicting statements of the details. Part of this is clearly due

262 PROTOZOA

to the use of different methods of culture by different investiga-
tors or by the same investigator at different times. Sometimes
cultures are transferred (fed) once a month, sometimes once a
day. This leads to great differences in the observed immature
period. With rapid growth, the immature period is very brief,
if it exists at all, in varieties 2, 4, and 6 (Ray, Elliott, and Clark,
1955); exconjugant clones of these varieties could mate again
within 3 to 5/2 days. In contrast, Gruchy ( 1955 ) stated that a
few exconjugants of variety 6 were not mature after four weeks
of growth, although most matured within this period. Variety 1
has been most fully studied in this respect. Elliott and Hayes
( 1953) give two to four weeks as the immature period; but Nan-
ney and Caughey ( 1955 ) , growing the animals at a rate of about
10 fissions per day, claimed that 16% of the exconjugants mature
after 50 fissions, 43% after 60 fissions, 83% after 70 fissions, and
100% after 80 fissions. If the animals were starved before each
transfer (every 12-13 fissions), 100% were mature by 60 fissions.
Because of the time involved in starvation (transfers every third
day), the time till maturity was prolonged although the number
of fissions was reduced. Variety 8, according to Gruchy (1955),
matures in nine days of rapid growth, which is about the same
as variety 1; but with monthly transfers, they were still imma-
ture at six months. Gruchy further reports that exconjugants of
varieties 3 and 7 were still immature after eighteen and nine
months, respectively. These varieties seem to be difficult to cul-
tivate and grow slowly or erratically; this may be involved in
the seemingly very long period of immaturity.

Although there is no question about the existence of an imma-
ture period or about its being followed by a period of maturity,
the possibility that maturity is followed by a period of senility is
just beginning to engage attention. Before inquiring into the
existence of varietal differences in the period of maturity, the
question of whether maturity ever comes to an end must be dealt
with. This is by no means an uncalled-for question. Cultures of
T. pyriformis have been maintained in axenic medium for
decades, and this has created widespread conviction that this
species is potentially immortal and hence lacking a life cycle.

T. M. SONNEBORN 263

Recent discoveries have, however, made it necessary to reevalu-
ate the situation.

The first shock came with the report of Elliott and Nanney
(1952) that one of the long-cultivated strains, E, has been ami-
cronucleate for at least 20 years. But the full impact of this
discovery becomes clear only in connection with other recent
discoveries about T. pyriformis and P. aurelia. After mating types
were discovered in the former species, efforts were made to ob-
tain collections far and wide in the Americas in order to find the
mating types and varieties in this region. This search turned up
a surprising and, in my opinion, highly significant fact: one-third
to more than one-half of the collections of T. pyriformis from
nature proved to be amicronucleate. Thus 50 of the 154 collec-
tions examined by Gruchy ( 1955 ) and more than half of the
collections of Elliott and Hayes (1955) in Central and South
America were amicronucleate. To me these facts signify that
the amicronucleate condition is, at least in some varieties, a
regular and long-lasting stage of life.

The absence of micronuclei is clearly to be characterized as a
senile condition, for two reasons. First, an animal without micro-
nuclei is genetically dead. Its genotype, confined to the macro-
nucleus, could not be transmitted to sexual progeny, since gamete
nuclei cannot arise directly or indirectly from a macronucleus.
Second, as more and more amicronucleate cultures have been
studied, it has become clear that absence of a micronucleus is
invariably associated with inability to mate, i.e., even to unite
in pairs (Elliott and Nanney, 1952; Elliott and Hayes, 1955;
Gruchy, 1955). Hence, amicronucleates have no mating type
and, once they get into this condition, the variety to which they
belong can no longer be ascertained. However, one can guess
that the amicronucleates found in Central and South America
belong to variety 2 or 9 or both, because these are the only
varieties thus far found in that region. Since maturity by defini-
tion is the stage of life during which fruitful mating can occur,
amicronucleates are not mature and cannot later become mature.
In short, they are senile.

The situation is in part parallel to the one existing in variety

264 PROTOZOA

15 of P. aurelia: these organisms also seem to lose their micronu-
clei in old age and to be genetically dead even though the line
of descent can continue to reproduce a very long time in this
condition, if not indefinitely. It will be recalled that we find
Johnson's long-lived axenic culture of variety 15 (known in the
literature as P. multimicronucieatum) to be amicronucleate, and
other strains of this variety show the same condition. However,
amicronucleates of P. aurelia, unlike those in T. pyriformis, are
able to mate even though they cannot transmit nuclei to progeny
produced by their mating.

If the view developed here is sound, long-continued laboratory
observation should reveal the sequence of immaturity, maturity,
and senility, i.e., loss of micronuclei, and should further reveal
the time at which senility, so defined, begins. Such observations
are just beginning to be made. The results already indicate a
complication apparently resulting from two different causes of
micronuclear loss. On the one hand in some clones considerable
variation in the number of micronuclei, accompanied by the
production of amicronucleate animals, appears early. Although
Elliott and Hayes ( 1953) found no relation between micronuclear
number and age up to 40 days in variety 1, Nanney, Caughey,
and Tefankjian (1955) found that amicronucleate animals char-
acteristically arise with high frequency in some clones of the
same material, beginning about 40 to 50 fissions after conjugation.
Nanney ( 1956a ) has intensively studied these clones, which he
calls semi-amicronucleates. Amicronucleate animals seldom arise
before the fortieth fission and may not appear until the one
hundredth fission or later, i.e., at about ten days. All the amicro-
nucleate animals in these clones die within a short time. This
condition is clearly determined by the genotype; it is not by
any means an invariable feature of all clones.

On the other hand, according to Nanney (unpublished), nor-
mal clones, alter reaching 500 to 1500 fissions, produce amicronu-
cleate animals and show other signs of aging. These amicronu-
cleates, like those produced earlier in certain clones, may also
be non\ table, for the entire clone is deteriorating. However, some

T. M. SONNEBORN 265

amicronucleates are clearly capable of living and reproducing for
a long time, as mentioned above.

Whether senility can be characterized by criteria other than
loss of micronuclei is not yet clear. There are indications that
selfing may also be a feature of senility in some varieties, as it is
in varieties 15 and 16 of P. aurelia. My reason for suspecting this
is that Gruchy (1955) reports differences as to the viability of
exconjugants obtained by selfing in different collections of selfers
from nature. It will be recalled that Giese ( 1957 ) found the
selfed progeny of senile clones in variety 16 of P. aurelia to show
increasing nonviability with increasing parental age. The differ-
ences reported by Gruchy may be of the same sort. Unfortu-
nately, as already mentioned, the varieties to which the selfers
belong has not been, and perhaps cannot be, ascertained. Until
the phenomenon is discovered to arise with age in clones earlier
identified in the laboratory, the occurrence of selfing as a feature
of senility will remain uncertain and, if it does occur, the varie-
ties in which it occurs will remain unknown.

At present, the life cycle in T. pyriformis appears to be some-
thing like this: there is an initial period of immaturity which is
followed by a long period of maturity; senility may begin with
age-induced selfing, which at first yields viable progeny and
progressively yields lesser proportions of survivors until none
survive; or senility may begin with loss of micronuclei and last
at least for decades. Because of the paucity of available informa-
tion, little or nothing is known about varietal differences in senil-
ity. Some may well become selfers, others may lose micronuclei,
some may do first the one and then the other. All this remains to
be discovered.

There is also practically no available comparative information
on the period of maturity. Detailed studies are virtually confined
to variety 1. Many caryonides are characterized by purity for a
single mating type throughout maturity, but a considerable frac-
tion begin maturity as selfers according to Nanney and Caughey
(1953, 1955). These workers discovered that pure types can
often be isolated from such selfing caryonides and that starvation

266 PROTOZOA

during immaturity is highly effective in stabilizing potential sett-
ers as a pure type. They pointed out that early selfing is a devel-
opmental stage characterized by progressive limitation of
expressible mating potentialities. Early selfing has also been
reported in variety 6 by Gruchy ( 1955 ) and may well occur in
other varieties.

This early selfing is to be distinguished from age-induced
selfing, just as it was in P. aurelia. The selfers collected from
nature would probably in most cases be past the incompletely
determined early stage of development. Moreover, they would
probably have been much exposed to the stabilizing action of
starvation. In fact, since conjugation is induced by starvation,
the exconjugants would at once be exposed to this stabilizing
influence and early selfing would be much rarer in nature than
in the laboratory where exconjugants are routinely provided
with excess food. Finally, the available reports indicate that
selfers collected from nature are not stabilized as one type by
starvation in the laboratory. A sharp distinction between the two
kinds of selfers should therefore be made.

To sum up the situation on varietal differences in periods of
the life cycle, very little is yet known about this, except for the
existence of differences in length of the immature period. How-
ever, from what is known about the marked difference between
varieties 1 and 2 in the method of mating type inheritance and
in geographical distribution, one may expect them to differ in
breeding system and in many features of their lives adapted to
this difference. Since the first two varieties to be studied show
such marked differences, it is to be expected that others also will.
This is already suggested by the observations on selfers and
amicronucleates found in nature.

4. Are the varieties of T. pyriformis readily identifiable? Pres-
ent knowledge provides only one way to identify the varieties
and that is by their mating reactions. Other differences are scarce
and of little or no use in identification. There is, if anything,
even less ground for assigning them specific names than there
was in P. bursaria or P. aurelia. The question of whether they
should be considered biological species is in principle again the

T. M. SONNEBORN 267

same. However, the degree to which gene flow occurs or is possi-
ble between different populations is not yet well worked out in
T. piriformis. Some further facts bearing on this will be brought
out in the following section.

Breeding Systems. Some information is available with respect
to viability and variability of exconjugants after crosses between
representatives of a single local population of variety 1 and, with
respect to viability alone, after crosses between representatives
of different local populations of varieties 5, 6, and 9. Convincing
evidence that natural strains of variety 1 are highly heterozygous
has been obtained by Nanney, Caughey, and Tef ankjian ( 1955 )
and by Nanney ( 1956a and unpublished ) . Among the sexually-
produced descendants of two individuals of complementary mat-
ing types that had been isolated from the same natural source,
two alleles for mating type potentialities segregated. The anal-
ysis likewise yielded evidence of polygenic differences controlling
the production of the semi-amicronucleate condition, other de-
generative conditions, and low viability in general. Two of the
segregant families were further inbred separately. In spite of
selection, viability of exconjugants in each inbred family oc-
curred in lower and lower frequencies in successive inbred
generations; but then selection resulted in improvement, and
by the sixth inbred generation it was possible to select a strain
which gave 100% viability at conjugation. Usually, crosses be-
tween the different inbred lines, at the stage when they were
giving poor survival and much abnormality, yielded more normal
sets of exconjugants.

These results of Nanney show that variety 1 of T. pyriformis
differs from variety 1 of P. bursaria in the effects of inbreeding.
In the latter it seems to lead quickly to extinction, but in T. pyri-
formis completely normal and viable inbreeding lines can even-
tually be selected.

Crosses between representatives of different local populations
were made in varieties 5 and 6 by Gruchy ( 1955 ) and in variety
9 by Ray (1955). In the single parental combination tested for
each variety, all the exconjugants of varieties 5 and 9 died, but
viable exconjugants were obtained from the cross in variety 6.

268 PROTOZOA

Viable exconjugants from a cross in variety 3 were also obtained
by Gruchy, but I cannot find information as to whether the
clones crossed were from the same or different populations. Some
comment is in order on the meaning of their statements about
obtaining viable conjugants. Apparently this is not intended to
mean that all or even a large proportion of the exconjugant clones
lived, only that some did. This is inferred because Elliott and
Hayes (1953) reported that variety 1 exconjugants were viable,
while Nanney and Caughey (1953) and Nanney, Caughey, and
Tefankjian ( 1955 ) emphasize that crosses of the same strains
as those used by Elliott and Hayes yielded a high frequency of
abnormalities and few viable clones. In the absence of detailed
mortality data on the other varieties, it is difficult or impossible
to know the extent to which different natural populations are
isolated by genetic barriers to gene flow operating through mor-
tality in the Fl and later generations.

Ray's ( 1955, 1956 ) observations on the cytology of conjugation
in varieties 1 and 9, however, are very suggestive. In both va-
rieties, he found chromosomal abnormalities at the first meiotic
division. They were abundant in variety 9, which gave no viable
progeny, and less common in variety 1, which gave some viable
progeny. Among the abnormalities noted were clumping of chro-
mosomes, lagging on the spindle, multivalent associations, and
chromosome bridges at anaphase. The last two abnormalities of
course indicate that the strains were structural heterozygotes.
These abnormalities were observed when strains from nature
conjugated. The natural strains thus might have been hybrids
themselves, but another interpretation to be given in the next
paragraph is not excluded. Whether the strains being crossed,
which were from different sources, also had different chromo-
somal arrangements remains unknown.

The interpretation of the preceding observations is rendered
difficult if we keep in mind the complications that have been
exposed in the fuller studies on other Ciliates. In both P. aurelia
and P. bursaria age is a most important factor influencing sur-
vival or death after conjugation. Moreover, chromosomal aberra-
tions are induced by age in P. aurelia, and further study may

T. M. SONNEBORN 269

show that they are similarly induced in P. hursaria. The impor-
tant question therefore arises as to the extent to which age may
be a factor in the observed exconjugant mortality and in the
observed chromosomal aberrations. Until this is cleared up, any
conclusions as to the breeding system based on material in this
section, except the inbreeding studies on variety 1, would be
premature. The alleles for mating type potentiality discovered
within a single population are good evidence that some degree
of heterozygosity occurs within a population. The observation of
inbreeding degeneration followed by selection of fully normal
lines also probably indicates genie heterozygosity, though part of
these results might be consequences of age-induced chromosomal
aberrations in the starting material.

When we now look back over the various features that have
been discovered in T. pyriformis and ask ourselves whether this
organism is an inbreeder or an outbreeder, the situation seems
confused. The organism clearly has some outbreeding features,
and it shows evidence that they work. Among them are the
usual ones: a period of immaturity, multiple mating types (in
most varieties), a very long life cycle, and low viability after
inbreeding. The finding of heterozygosity for genes within a
single local population is good evidence that outbreeding does
in fact take place. On the other hand, inbreeding features also
seem to be present. Among them, the most outstanding are (1)
the occurrence of selfing clones in considerable proportions and
(2) the method of mating type inheritance in variety 1, which
assures the presence in most clones of animals capable of mating
with each other.

However, these various aspects of the biology of the organism,
as discovered in the laboratory, must be considered in the light
of conditions in nature. Three factors tend to reduce the impor-
tance of the inbreeding features. First, the likely recurrence of
starvation in nature would tend to reduce, if not prevent, the
development of selfing clones during early maturity. Second,
although mating type inheritance provides potential mates among
close relatives in variety 1, the immature period permits time to
spread and so reduces the chance of mating with closely related

270 PROTOZOA

individuals. Third, inbreeding nonviability tends to eliminate
those which do inbreed; and, if there are present simultaneously
outbred competitors, these would have an advantage in subse-
quent population increase.

Looking at the picture as a whole, therefore, I am inclined
to interpret T. pyriformis as an outbreeder, but as one that has
an already highly developed potential for evolving inbreeding
descendants. The fact that Nanney has been able in the labora-
tory to select inbred lines of variety 1 that give virtually no
exconjugant death shows that the organism could do the same
in nature, if conditions were propitious for it. The fact that some
selfing clones which give viable sexual progeny are found in na-
ture could mean that the same sort of selection in nature pre-
ceded their origin; but I am rather of the opinion that this
represents a last resort for old clones which have failed to out-
cross. Finally, it is of much interest that Corliss (1952) has in
fact discovered a species of Tetrahymena (T. rostrata) which
regularly responds to starvation by going 100% into autogamy —
the closest type of inbreeding. T. pyriformis is in fact of great
interest in our present discussion because it shows how an out-
breeder is preadapted, as it were, to become an inbreeder on
demand. This may be considered as strengthening our view con-
cerning the evolutionary derivation of inbreeders from out-
breeders. The mechanisms needed for an inbreeder are present,
but counterbalanced so that they operate relatively poorly; remove
the restraints and an inbreeder results. The restraints actually
seem to be developed to different degrees in different varieties,
as indicated, for example, by variations in the length of the
immature period and in the mode of mating type inheritance.

As the varieties of T. pyriformis are compared in more detail,
it may well be discovered that the different mechanisms for in-
breeding and outbreeding are quite differently developed and
balanced among them, as among the varieties of P. aurelia. In
this connection, comparison between variety 1 and variety 2 is
of special interest. The latter is clearly more of an outbreeder
than the former. It shows synclonal uniformity in mating type,
while variety 1 does not. For this reason, the report that its imma-

T. M. SONNEBORN 271

ture period may be a few days shorter is of less significance. Even
without an immature period, the method of mating type inherit-
ance inhibits the closest form of conjugation inbreeding. Variety
2 is also distinguished by having more known mating types than
any other variety and, as we have seen, multiplicity of mating
types is of adaptive value to an outbreeder. Finally, variety 2 has
the widest known latitudinal distribution in nature, and we have
previously seen how this too is correlated with outbreeding. The
fact that the first two varieties to be studied in detail show such
marked differences in features associated with different breeding
systems leads to the supposition that the species T. pyriformis,
like P. aurelia, consists of varieties with a large array of breeding
patterns, perhaps from rather extreme outbreeders to extreme in-
breeders.

Euplotes

Species, Varieties and Mating Types. Until Kimball ( 1939,
1942) discovered mating types in Euplotes patella and worked
out the breeding relations among many collections of organisms
that closely resembled or were identical with E. patella, there
was great confusion in the literature concerning those species of
Euplotes that look much like E. patella. Sorting his collections
into those that would or would not interbreed, Kimball came
out with at least five groups. One group did not conjugate. Each
of the other four included a number of collections that could
conjugate in mixtures with each other, but no strain of one
group would conjugate with any strain of another group. Pierson
(1943) made a careful morphological study of Kimball's five
groups of strains and of some additional strains belonging to these
groups. On the whole (exception noted below), the groups were
found to be morphologically distinct. The one which was not
observed to conjugate was identified as E. woodruffi. One proved
to be a new species, E. aedicukitus. One was E. eurystomus. Two,
this is the exception, could not be distinguished morphologically;
both were E. patella. There may have been more than one group
in some of the other species already mentioned, but this is not
clear from the published account. There is also some vagueness

PROTOZOA

as to whether there was a third group of E. patella. This vague-
ness is apparently due to the fact that the only group extensively
studied as to breeding relations was one of the groups of E.
patella.

The groups of strains of E. patella, although they are sexually
isolated, were not assigned specific names by Pierson because
she agreed with the views of Sonneborn (1938) and Kimball
( 1943 ) that this should not be done unless the species could be
readily identified. They correspond exactly to the varieties of
Paramecium.. Kimball (1943) and Pierson (1943) agree that
sexual isolation probably preceded morphological divergence in
the speciation of Euplotes (and other Ciliates as well). This is
based on the plausible assumption that the varieties of E. patella,
which are morphologically identical so far as known, are more
closely related than the previously mentioned species, which
nevertheless are much alike in morphology and well marked off
from other species of Euplotes. The very similar species are there-
fore considered' to be intermediate in evolutionary divergence
and relationship between morphologically very different species
and morphologically indistinguishable varieties.

The preceding work established the existence of mating sys-
tems based on mating types in E. patella, E. eunjstomus, and E.
aediculatus. Katashima ( 1952 ) reported mating types in E. harpa,
which is apparently not a member of the group of species that is
most closely related to E. patella. Thus far, no survey has been
made which would indicate the number of varieties in any species
of Euplotes. I suspect that it may be small. The full number of
mating types in a variety is also at present unknown. Katashima
(1952) reports upon only two cultures collected in Hiroshima
City; these were of different mating types. Kimball (1939, 1942)
also confined his studies to the progeny of two individuals of
complementary mating type obtained from neighboring ponds,
but he included clones derived sexually from the two original
cultures. In this way, he discovered four more mating types. It
might be supposed that this procedure would reveal all the
mating types of the species, as it appears to do in P. bursaria.
Kimball ( L942 points out, however, that his discoveries of the

T. M. SONNEBORN 273

genetic basis of mating type specificity (see below) do not at
all preclude the existence of more types. One might actually
expect a very large number of types in nature, as will appear.

Mating Relations and Inheritance of Mating Type. The mat-
ing relations reported by Kimball for the six mating types he
found seem superficially to be like those in P. bursaria, but they
probably differ in several respects. Conjugation occurs when any
two of the six types are mixed together. However, by using
visibly marked animals, Kimball proved that the conjugant pairs
are not always crosses between the two types that were mixed
together. In some combinations of cultures, one finds both crosses
and selfing. In fact, if one adds animals of one mating type to
fluid in which animals of certain other types have lived, the one
and only type introduced into the fluid selfs. The animals clearly
secrete "hormones" that induce conjugation in animals of certain
other types. Sonneborn (1947) concluded from a detailed exam-
ination of the data of Kimball (1942) and Powers (1943) that
crosses occur only when each type secretes into the fluid a hor-
mone which induces the other type to conjugate. As Kimball
showed, in such mixtures three kinds of conjugant pairs are
formed: crosses between the two types and selfing of each type.

This system has a simple genetic basis and remarkable conse-
quences. Altogether, among the six mating types, only three
hormones are produced. Each hormone is determined by a single
gene, and the three genes are codominant alleles. There are thus
three homozygotes, each producing one and a different hormone;
and three heterozygotes, each producing two of these same three
hormones in a different combination. These are the six mating
types. Because mating type is directly controlled by the geno-
type and each mating type has a different genotype, each syn-
clone is necessarily uniform for mating type because conjugation
has rendered it uniform in genotype. Without going into further
details, the system works out so that crosses between different
synclones take place (along with double selfing) only in three
of the five possible combinations of any one mating type with
the others. In the other two combinations one member selfs, the
other does not, and no crosses occur.

274 PROTOZOA

The Breeding System. Superficially the mating relations seem
like a very fumbling and inefficient arrangement to encourage
outbreeding, inefficient because not all combinations of diverse
types can mate, and fumbling because it can induce selfing as
well. However, there appear to be good reasons to suppose the
system is not really as bad as it seems. In fact, when one takes
into account all the information discovered in the laboratory, but
makes allowance for the differences between the laboratory con-
ditions used in analysis and those that probably prevail in na-
ture, the system appears to be well suited for outbreeding and
neither fumbling nor inefficient. Actually, it seems if anything to
be a more nearly perfectly developed outbreeding system than
that of any multiple type system in Paramecium or Tetrahy-
mena. I shall now try to make this clear.

First, the six known types were derived from but two neighbor-
ing individuals. Both were heterozygotes for the hormone genes
— in itself a little indication that we are dealing with an out-
breeder that does not prevailingly self in nature. Three of the
four genes at the hormone locus differed in these two individuals.
These are the three alleles the combinations of which yield the six
known mating types. Is it reasonable to suppose that by chance
these two individuals possessed all the hormone alleles that exist
in nature in this species? Decidedly not, in my opinion. It seems
far more likely that a wide sampling of various natural popula-
tions would have revealed a large number of such alleles. The
situation might even be comparable to the one known for incom-
patibility alleles in Basidiomycetes and higher plants, in some of
which a hundred or more alleles have been estimated to occur.
Certainly it would be a miracle if all the alleles that exist in this
variety of E. patella were the three found in the only two in-
dividuals studied. With many multiple alleles, there would be
many mating types; and multiple mating types are, as has been
repeatedly emphasized in this paper, an adaptation for outbreed-
ing-

Second, Kimball reports the regular existence of an immature
period of a month or more following conjugation and Katashima
(1952) found a comparable immature period (three weeks or

T. M. SONNEBORN 275

more) in E. harpa. The existence of a considerable immature pe-
riod is hardly to be considered fortuitous. As set forth above, it
exists also in T. pyriformis, in P. bursaria, and in outbreeding
varieties of P. aurelia. Its meaning appears to be to discourage
inbreeding by making mating impossible when the organism is
most likely to be near its relatives and by giving time to spread
and meet a stranger that would make a suitable mate. Of other
stages in the life cycle, nothing is known except that there is a
long mature period and a long life cycle, if there is a terminating
cycle at all, as I believe there is. All that is known is characteristic
of outbreeders.

Third, Kimball mentioned that there was high mortality in his
crosses, all of which involved inbreeding. As we have seen in
the accounts of Paramecium and Tetrahymena, inbreeding mor-
tality is a sign of an outbreeder.

All these strong indications that E. patella is an outbreeder
make me conclude that the superficially fumbling, inefficient,
and contradictory features may be misleading us. The major
difficulty seems to be the fact that mating types which can cross
also induce each other to self. This is no way for an outbreeder
to behave. Perhaps it is largely a laboratory phenomenon and
rare or nonexistent in nature. In the laboratory these observa-
tions were made in depression slides, i.e., in about 1 ml. of fluid
containing large numbers of euplotes. Contrast this with a pond
or river. The contrast is important, for we deal with extracellular
hormones acting through the surrounding fluid medium. Surely,
in a pond the dilution of the hormones would be enormous and
rapid. There would be a sharp concentration gradient away from
the position of the secreting animal. Further, the concentration of
animals producing hormones would be expected to be very much
less than in the laboratory. Consequently, the hormones could
probably work only for short distances, possibly only on two
euplotes that were close together, if not in actual contact. I can
see little likelihood of the hormonal mechanism leading in nature
to anything like the amount of selfing observed in the laboratory.

However, there may well be some selfing in nature as well.
This is indicated by two facts. First, selfing was observed re-

976 PROTOZOA

peatedly within single clones. Second, viable inbred progeny
were obtained in the laboratory. Although as mentioned, mortal-
ity was high among the inbred progeny of the two original iso-
lates, Kimball and Powers were able to pursue genetic analysis
with them. Whether it was possible to select in this way com-
binations relatively free from further inbreeding mortality is not
ascertainable from their papers, so far as I can discover. It is also
not clear whether continued inbreeding led to progressive de-
terioration, but Kimball eventually abandoned the organism be-
cause of difficulties in maintaining cultures. Either inbreeding
degeneration or aging or both may have played a part in this. To
the extent that inbreeding yields viable progeny and to the extent
that selfing occurs in nature, E. patella may be considered as a
facultative or potential inbreeder, although provided with life
features that make outbreeding the usual practice. Here, as in
Tetrahymena, we may see traces of how an organism which is
customarily an outbreeder contains the seeds of an inbreeding
evolutionary descendant.

Other Ciliates

Mating types and something about the breeding system are
known for relatively few Ciliates other than those already dealt
with. All can be quickly dealt with except Oxytricha bifaria about
which much of interest is known. A system of three mating types
which show the P. bursaria relations (each type mating with
both of the others) was found by Sonneborn (1938, 1939) in
P. trichium. Diller (1948) and Wichterman (1937) have found
strains of P. trichium which self very readily. 1 have also en-
countered such strains. On the other hand, those in which I re-
ported the three mating types did not self, and Wichterman
I 1953) has had one strain under cultivation for twelve years
without finding selfing. Whether the selfing and nonselfing strains
belong to different varieties is unknown. If they do, the former
would probably be found to be an inbreeder, the latter an out-
breeder. Sonneborn (1938) reported a pair of mating types in
P. calkinsi. Wichterman (1950, 1951) found a second variety
which also contained two mating types. I have found multiple

T. M. SONNEBORN 277

interbreeding mating types in an organism kindly identified for
me as Colpidium truncatum by Corliss. This organism, which is
closely related to Tetrahymena, likewise has its breeding system
much complicated by selfing. Downs (1952) reported a system
of five interbreeding mating types in Stylonychia putrina col-
lected from a single source and he ( 1956 ) now finds two varie-
ties with eleven types in one variety and fifteen in another. In
S. putrina, selfing is rare; only one clone of one mating type
selfed among 41 clones of the first five types studied.

In Oxytricha bifaria, Siegel ( 1956 ) reported a system of nine
interbreeding mating types. He investigated the possibility that
O. bifaria might manifest relations like those in Euplotes because
these organisms are both hypotrichous Ciliates and so are more
closely related to each other than they are to other (holotri-
chous) Ciliates in which mating types have been found. Siegel
looked for action of sex hormones, but could find no evidence of
their existence. He also devised means of discovering whether
the conjugating pairs in mixtures of two mating types were true
crosses or selfers and showed that they were regularly true
crosses. Indeed, selfing appears to be exceptional in his material.
Only one mating type (III) selfed, and this frequently showed
about 1% of the animals conjugating even when not mixed with
another type. The two members of selfing pairs were separated
before becoming firmly united. Both gave rise to cultures like
the parent, that is, they were predominantly or entirely type III,
but selfed slightly.

Crosses of certain mating types gave types among the progeny
which sometimes differed from both parents; but the genetic
details were not reported. Whether, as in Euplotes, mating types
are determined by a series of multiple alleles is still unknown.
The large number of mating types in O. bifaria and in S. putrina
suggests that such a basis may well exist. On the other hand,
Siegel found the same mating type in several natural sources,
and this indicates that the number of existing types may not be
great.

The life cycle of O. bifaria begins with a period of immaturity
lasting for variable periods, from less than one month to over

278 PROTOZOA

two years. Although Kay (1946) reported endomixis (autog-
amy?) in certain strains of this species, Siegel did not find it and
was unable to induce it. Perhaps Kay's strains belonged to a
different variety. In the absence of conjugation, Siegel's variety
of O. bifaria slowly weakens and eventually dies, the whole cycle
usually lasting more than two years.

Some information is available on the survival of exconjugant
clones. Those derived from the selfing of type III almost always
proved nonviable. Some crosses (types I by II, III by V, and V
by VIII ) gave high percentages of viable progeny; other crosses
(types I by III and II by V) yielded a majority of nonviable
progeny. It is to be noted that the type III culture, which yielded
few survivors when it selfed, gave high survival when crossed
to type V. This type III culture was one collected from nature,
not one produced from crossing in the laboratory. Its age was
therefore unknown. The possibility that selfing of this culture
and the death which is its consequence were due to age seems
excluded, for Siegel found no stage of selfing in other cultures
which were followed through senility to death. Had this informa-
tion not been available, one would have been tempted to suggest
that old-age selfing is a regular feature of the life cycle. This
should serve as a warning of the insecure basis of some of the
conjectures that I have made about life cycles in Tetralujmena
for which full laboratory studies of the whole cycle are not
available.

In discussing the significance of the facts about O. bifaria,
Siegel stresses those features of life that facilitate the occurrence
of mating, without differentiating between inbreeding and out-
breeding. He is particularly impressed by the occurrence of
autogamy and the existence of only two mating types in some
species of Ciliates and by the absence of autogamy and the
existence of multiple mating types in other species. Although
recognizing certain exceptions, he attributes this correlation to
the necessity for fertilization in preventing senescence and death
by starting new clones. Animals in species with multiple types
have more opportunity to meet an appropriate mate; since the

T. M. SONNEBORN 279

probability of doing so is less in two-type varieties, autogamy
may serve as a safety device to cover the risk of failure to meet
a mate.

Sonneborn (1955) agreed with some of these views and to
some extent also does so in the present paper. Yet, taking into
account other facts and more comprehensive correlations, I have
here developed in detail the thesis that all of them have their
primary significance as mechanisms adapted to different systems
of breeding, either inbreeding or outbreeding. The association
of a considerable period of immaturity with multiple mating
types is not interpretable on Siegel's hypothesis. These two fea-
tures seem to act at cross purposes: one favors mating, the other
prevents it. Their association is, however, adapted to outbreeding.
In varieties that have these features, it is important to find a
stranger, not just to undergo fertilization indiscriminately. The
absence of autogamy and the long life cycle in such varieties
have the same meaning, as do the other features discussed
throughout this paper. As implied here and developed in detail
above, the length of the life cycle is highly variable from variety
to variety; there is no need for quick fertilization in an outbreeder
and a short life cycle is adequate for inbreeders with their various
devices: little or no immaturity, much selfing, early autogamy,
and various genetic means of assuring presence of both mating
types. Actually, the situation seems to be just the reverse of the
one postulated by Siegel: in two-type systems, the whole biology
of the organism is designed to assure early meeting of related
mates; whereas, in multiple type systems, usually the meeting of
mates is prevented during early life so as to increase the chance
of mating with a stranger.

So far as O. bifaria is concerned, the variety examined by
Siegel seems clearly to be an outbreeder. It rarely selfs; it has
no autogamy; it has multiple mating types; it has a considerable
period of immaturity, long maturity, and a long life cycle. Only
the questions of inbreeding degeneration and outbreeding via-
bility remain unsettled. The high death rate after selfing fits the
picture, but the variable results of outcrosses need clarification.

280 PROTOZOA

That other varieties may exist with different breeding systems
is perhaps suggested by Kay's report of what might be autog-
amy.

Obligatory Inbreeding and Asexual Protozoa

The Transition from Sexual to Asexual Reproduction in the
Flagellates of Wood Roaches and Termites. The Ciliates illus-
trated gradations from extreme outbreeders to nearly the extreme
of inbreeding. The highly evolved and morphologically complex
Flagellates that live in the gut of wood roaches and termites
illustrate the transition from obligatory close inbreeding to
asexual reproduction. Cleveland (1934-54) has reported ex-
tensively on the Flagellates of wood roaches. They undergo
sexual reproduction at or near the time the host moults. Only at
this time do the roaches pass fecal pellets laden with the Flagel-
lates; and at the same time the eggs, which are laid earlier,
hatch. Sexual processes in the Flagellates are thus timed to
coincide with transmission to new individual hosts.

That the two processes should be timed in this way is hardly
without meaning. Obviously this would make possible selection
of those genetic variations in the symbionts which happen to be
adapted to variations of the individual infected host. Yet, as I
shall at once show, there is little opportunity for either host or
symbiont variability. It would therefore seem more likely that
the present correlation in time of symbiont sexual processes and
transmission to new hosts is a relic of earlier conditions when
both symbiont and host were more variable than they are to-
day.

The wood roach lives in small isolated colonies and migration
seems to be very limited. Under such conditions, the amount of
genetic variability expected among the individuals of a single
roach colony would be small. Correspondingly there would be
little need for genetic variability among the Flagellate symbionts
living in so constant a milieu.

In agreement with this, Cleveland's accounts of the sexual
processes in the Flagellates indicate little opportunity for genetic
recombination. Although some of the Flagellates, such as Euco-

T. M. SONNEBORN 281

monympha, show complete two-division meiosis, Cleveland de-
scribes remarkable abbreviations of, and variations on, the
standard sexual processes. According to him, Oxymonas, Saccino-
baculus, Notlia, Leptospiwnympha, and Urinympha undergo a
remarkable one-division meiosis. Neither centromeres nor chro-
mosomes duplicate, chiasmata are not observed, and homologous
chromosomes pass intact to opposite poles. This would clearly
reduce the possibilities for genetic recombination. Cleveland re-
ports for other Flagellates processes which would place even
greater restrictions on genetic variability. In Rhynconympha,
the first meiotic division is accompanied by cell division, but the
second is not, and the sister nuclei thus produced reunite in au-
togamy. In Urinympha he maintains that the diploid nucleus
undergoes meiosis in one division and that the two reduced nuclei
reunite in autogamy. In the haploid Barbulanympha, fusion of
sister nuclei is said also to occur sometimes. Some of these
Flagellates thus appear to have no possibility of outbreeding.
Organisms reach the ultimate limit of inbreeding when fertiliza-
tion always takes the form of autogamy. These are the considera-
tions concerning host and symbiont variability which led me to
conclude that present conditions are not adequate to explain
the coincidence in time between infection and fertilization, and
that earlier conditions of greater host and symbiont variability
probably provided the selective pressure necessary for its evolu-
tion.

A corollary of this conclusion is that the aberrant and abbre-
viated sexual processes now observed in the symbionts represent
stages in the atrophy and loss of meiosis and fertilization. Cleve-
land (1951b) on the contrary rejected this possibility and con-
cluded that they were stages in the origin and evolution of
sexuality. To him this seemed the more reasonable interpretation
of what he recognized as the comparative genetic uselessness
of the cytological processes he described. The major support for
his view comes from the juxtaposition of two facts : ( 1 ) these
symbionts are among the morphologically most complex and
highly evolved Flagellates; and (2) sexual processes are not
satisfactorily established as occurring in the simpler animal

282 PROTOZOA

Flagellates (see review by Wenrich, 1954). Whether the second
point is to be ascribed to incompleteness of investigation or to
actual absence of sexuality in the lower animal Flagellates re-
mains unknown. However, many examples of perfectly standard
meiotic and fertilization processes are well known among the
plant Flagellates. One could hardly assume therefore that
sexuality in Flagellates first arose in the complex symbionts of
wood roaches. Rather one needs to inquire why so many of the
simpler Flagellates lost it, if indeed they do lack it. That the
Flagellates of wood roaches show stages in the loss, not the
origin, of meiosis and fertilization is further indicated by a com-
parison of the situation in wood roaches and termites.

The wood roaches are primitive members of the group that
gave rise to termites. Hence, the later evolutionary stages of the
symbionts might be expected in the termites. Closely related
Flagellates are indeed also found as symbionts in the termites.
But without exception these show no sexual reproduction what-
ever. They reproduce asexually only. It is true that Cleveland's
evidence correlates the sexual cycles in the Flagellates of wood
roaches with the moulting hormones of the roach, and exactly the
same hormonal conditions would not be expected in termites.
But any organism which is under selective pressure to maintain
the hard won evolution of sexuality should not find such a change
of hormone pattern beyond its capacity to cope with. More
likely, the change in the mode of life of the evolving host pro-
vided no more pressure for genetic variability on the symbionts,
and possibly less.

Each colony of termites is isolated underground. The extent to
which the brief nuptial flight preceding the founding of new
colonies permits outbreeding of the host is not, so far as I am
aware, known. It seems that there is a considerable degree of
synchronization in swarming, however, so that interbreeding
between neighboring colonics might well occur. On the other
hand, the enormous number of eggs laid by a single female is a
powerful factor for assuring a high degree of uniformity among
the members of a colony and its descendant colonies. The rela-
tive!) small degree of host variability and presumably very slow

T. M. SONNEBORN 283

host changes during the course of evolution have obviously been
coped with adequately by the symbionts without recourse to
sexual reproduction (Kirby, 1949). Thus, under conditions of
relaxed selection for sexual reproduction, it seems that it de-
generated step by step from normal meiosis and crossing, to
one-step meiosis, to autogamy, and finally to asexual reproduc-
tion.

The Problems. In the series of transitional stages from non-
obligatory but preferred inbreeding to exclusively asexual repro-
duction, changes take place in the species problem and in the
problem of adjusting the features of life to assure maintenance of
an unbroken line of descent through evolutionary time. I am not
sufficiently familiar with the intimate details of the lives of the
Flagellates to attempt to work out for them, as I did for the
Ciliates, a full picture of how the major life features are adapted
to the mode of reproduction. A few of the more important and
obvious features will be mentioned below.

Those who have considered the species problem in relation to
types of reproduction often make a sharp contrast between spe-
cies in outbreeding organisms and in obligatorily inbreeding and
asexual organisms. For example, Dobzhansky (1937, 1941) draws
a sharp line between the two, maintaining that there is little in
common save the word "species," which has very different mean-
ings in the two cases. He further states, "in asexual groups either
every biotype is to be called a species or else species do not exist
there at all. ' These comments are of course based primarily on
the concept of a species as a common gene pool, which Dob-
zhansky has strongly urged for 20 years. The nature of species
in this special sense does not suddenly change, but undergoes
continuous, progressive change in a single direction as one passes
through the various degrees of outbreeding and inbreeding to
asexual reproduction. One might therefore expect to find equiva-
lent levels of biological organization in different systems of
breeding and reproduction. I shall explore that possibility after
taking up first the problem of asexual species in the routine
taxonomic sense.

Hoare ( 1952 ) maintains that only morphological differences

284 PROTOZOA

should be used as species differentials. Although his discussion is
primarily about parasites, he clearly intends to include a broader
domain. In essence, this amounts to denying any difference in
principle between criteria of taxonomic species in sexual and
asexual organisms. In fact, Hoare includes both sexual and
asexual parasites in his discussion. Physiologic differences he re-
jects in general as alone insufficient to mark off species. He par-
ticularly lists among such inadmissible traits differences in drug
resistance, virulence, disease produced, serotypes, and host
specificity- I shall add specific objections to most or all of these
traits, but I shall also bring out objections to certain morpholog-
ical or visible traits. My chief point in this connection will be
that the genetic difference between species in asexual organisms
should be as nearly as possible of the same kind and magnitude
as in sexual organisms. To achieve this equivalence some genetic
information gained from the study of sexual organisms will have
to be used in guiding the choice of differential traits. I shall illus-
trate my point of view by presenting and then discussing certain
observations on Trypanosomas and Rhizopods.

Trypanosomes. The Trypanosomes are of special interest in
relation to the soundness of serological traits as species differen-
tials. Since the time of Ehrlich, the complexity and variability
of serologic traits in Trypanosomes has been known, and this has
prevented their abuse as species differentials. Recently, Inoki
et al. (1952a,b) have shown particularly well that the multiple
alternative serotypes which can arise in a single clone are not
selected gene mutations, but alternative expressions of an array
of potentialities which were from the start potentially expressible.
In so far as the analysis has been carried, it has yielded results
which are comparable to those' obtained in P. aurelia. As set
forth above, the genetic analysis in P. aurelia shows that each
clone possesses at a series of loci genes lor the various serotypic
potentialities and that each individual and subclone expresses
serotypically genes at only one locus of the series. The parallel
phenomena in the Trypanosomes point to the same sort of genetic
basis though breeding analysis is not possible.

In view of this parallel, it seems likely that similar situations

T. M. SONNEBORN 285

occur in other asexual organisms, including some in which
serologic traits are used to define species. Failure to appreciate
that organisms with the same genotype can express serologically
unrelated antigens could easily lead to assigning organisms witli
the same genotype to different species. Only by knowing the full
array of potentially expressible antigens in a clone and by taking
into account the grades of difference in cross reactions between
corresponding serotypes in the arrays producible in different
clones could a sensible beginning be made in establishing sero-
typic criteria for distinguishing species. Lederberg (1955) points
out other weaknesses of the use of serologic species criteria in
bacterial taxonomy based upon his work on gene transduction.

Physiological differences such as virulence and host specificity
are used as species differentials in Trypanosomes and other para-
sitic asexual organisms. Both of these criteria may be objected to
on the ground that comparable traits in other organisms (e.g.,
bacteria and viruses) are known to arise as single mutational
steps and to be inherited as single gene traits. So simple a genetic
difference would surely not be acceptable as a species differen-
tial in sexual organisms, and there is no good apparent reason
why this should be done in asexual organisms. It is sometimes
argued that confinement to different hosts is equivalent to
geographic isolation. But this is only true if the adaptation to
different hosts is outside the range of simple genetic differences.
Otherwise the mutational array of host specificities within each
single line of descent will yield a common array of potential
hosts.

Restriction of species differentials to visible differences does
not necessarily avoid this difficulty. Some visible differences also
are of course simple in genetic basis. For example, certain species
of Trypanosomes are known to differ only with respect to the
presence or absence of a kinetoplast. This is a self-reproducing
cytoplasmic granule. Under the action of certain chemicals, the
kinetoplast can be irreversibly lost. Sublines of the same clone
can differ in this way. Yet they must be assigned to different
species if this is accepted as a valid species differential. The sit-
uation is closely parallel to the existence of killer and sensitive

286 PROTOZOA

paramecia in the same clone by reason of loss of kappa from some
individuals. And it is no more sensible to use the one or the
other as a species differential. These cases are of additional in-
terest iu showing that visible hereditary differences need not
necessarily be preceded by invisible physiological genetic differ-
ences, as many have maintained.

Rhizopods and the Results of Selection. Jennings (1916) and
his students were able to select strikingly different morphological
types within a clone in various Rhizopods. The different types
thus obtained tended to perpetuate themselves in the absence
of further selection. For example, Jennings obtained the most
varied shell (test) types in Difflugia: spineless and many spined,
with few teeth around the mouth and with many, and so on. Con-
trariwise, it was possible to start with two diverse forms and
select the same type from both. Here again there is no indication
that physiologic genetic difference precedes morphologic differ-
ence in the divergence of two forms. Even persistence of the dif-
ference among unselected progeny is no assurance that the two
types belong to different clones, and similarity of type is no
guarantee of belonging to the same clone.

Criteria of Species in Asexual Organisms. From the forego-
ing examples and discussion a few useful general principles seem
to emerge. Species differences in sexual organisms are based
upon complex, not simple, genetic differences. To reduce species
differences to a single trait dependent upon a single gene differ-
ence is to equate species differences in asexual organisms to the
level of individual differences in sexual organisms. From this
point of view, a single, genetically simple visible or morpholog-
ical trait is in no better case than a single, genetically simple
physiological trait. If asexual species are to represent a level of
evolutionary divergence comparable to the one used in taxonomy
of sexual species, the distinctions would have to be not only
morphological but genetically complex. To employ differentials
which <an be wiped out by one or even a few mutational steps
is iu in\ opinion indefensible.

Variation and Speciation in Asexual Organisms. The condi-
tions of life for an asexual organism are likely to be less varied

T. M. SONNEBORN 287

than those encountered by most sexual organisms. A large body
of water or a particular host, for example, acts as a buffer against
radical variation in the milieu. But some variations in conditions
are inevitable and the organism has to be prepared to adapt to it
or run the risk of extinction. Lacking the rapid and wide-ranging
variability obtainable by genetic recombination, other sufficient
mechanisms for yielding adaptive variations have been selected.
One of these is the combination of haploidy and large, rapidly
produced populations with mutation. This has apparently been
adequate for many asexual organisms.

Another and faster-operating mechanism is one which is less
widely appreciated. It is the one illustrated by the serotype
systems of Trypanosomes and Paramecium. This mechanism in-
volves ( 1 ) including in the common genotype of the species sets
of loci for alternative and mutually exclusive phenotypes and (2)
incorporating mechanisms for shifting phenotypic expression
from one to another on demand. This is clearly a very satisfactory
adaptation to the need for rapid variation during asexual repro-
duction. As such, it is found not only in organisms that lack sexual
processes, but also in those that have sexual processes along with
a well-developed asexual phase, such as the Ciliates. Further,
there is some evidence that it plays an important role in develop-
mental differentiation, the asexual phase of higher organisms.
For example, genes for both sexes are present in hermaphrodites,
but come to expression in different cells; and other aspects of
differentiation probably have a comparable genetic basis. Thus,
at every level of asexual reproduction, this mechanism of var-
iability seems to be in operation.

Speciation in both asexual and sexual organisms is probably
fundamentally the same. In both cases, the basic event is genetic
adaptation to different conditions. Perhaps asexual organisms
selectively live under relatively constant or very slowly changing
conditions. This selection of constant environments would tend to
retard conspicuous speciation, but beneath the visible uniformity
there might well be much accumulated genie and chromosomal
change, such as exists among local populations of strongly in-
breeding varieties of P. aurelia. On the other hand, selected vari-

288 PROTOZOA

ants can be rapidly multiplied and spread by asexual reproduc-
tion. The individual wastage of recombination and the swamping
of variations by crosses are avoided. As in any system which has
proved itself in nature, what is lost by one feature of the system
is compensated by others. But a short-term advantage is of course
no guarantee of long-term persistence. Very often temporarily
successful asexual organisms may have been wiped out on occa-
sions of stress. Those which have adapted to constant or but
slowly changing environments may, however, succeed even on
the geological time scale, as studies on the Flagellates of roaches
and termites have shown ( Kirby, 1949 ) .

The nonexistence of species in asexual organisms is asserted
only by those who define species as syngens. For reasons given
earlier, I have rejected this definition and have accepted the
necessity for readily recognizable distinctive features in a species.
In the present section, emphasis has been placed upon the com-
plex genetic basis of species differences in sexual organisms, and
genetic differences of the same kind and magnitude were de-
manded for species differentials in asexual organisms. On the
whole, the taxonomy of the Protozoa does in fact utilize such
species differentials. The chief infractions of the rule are found
among the parasitic and medically important Protozoa which
have come to be known in great detail. This intimate knowledge
has led the "splitters" to use such genetically simple criteria for
species differentials that their species become the equivalent
of individuals or families in sexual organisms. "Lumping" is now
needed until the discontinuities between groups become geneti-
cally complex as well as readily discernible.

Before proceeding to a consideration of the more important
question of the possible existence in asexual organisms ol the
equivalent of the syngen in sexual organisms, some further com-
ments on the usage of the term species in genera] and its paral-
lelism in sexual and asexual organisms are in order. Difficulties in
the application of the term species arise from the attempt to
make it do double duty in serving both as designating an evolu-
tionary unit, the one which shows minimal irreversible discon-
tinuity, and as designating a readily recognizable group. For

T. M. SONNEBORN 289

most of the last hundred years, it has been tacitly assumed that
these two aspects of species coincide and that, since the evolu-
tionary unit often represents a real level of biological organiza-
tion, species exist as real entities, not as human constructs made
for the convenience of biologists. Modern researches, on the con-
trary, have shown that the evolutionary unit is not always readily
recognizable. In such cases, the proponents of the modern biolog-
ical species concept admit the dual implications of the term
species by distinguishing named species, those which are readily
recognizable, from sibling species (Mayr, 1948), those which are
often not named because they are not readily recognizable. This
leads to the totally illogical procedure of referring to a group of
species as a species, i.e., a group of unnamed sibling species has
a single species name. Moreover, by reserving species names for
readily recognized groups, this procedure tacitly admits that the
convenience of biologists must first of all be served by the term
species. Agreeing with this, I propose, in the interest of logic and
clarity, to use a different term, syngen, for the evolutionary unit.
The same group of organisms would be both a species and a syn-
gen if both the criteria of ready recognition and minimal irre-
versible evolutionary divergence were met. Otherwise, not. The
reluctance of proponents of the modern biological species concept
to take this logical step seems to be based on their love of the
word species. Its early connection with evolution theory, its wide
usage, and general familiarity with it make it a great prize to cap-
ture and invest with absolute existence as a definite level of bio-
logical organization of the greatest evolutionary significance. This
cannot be done without sacrifice of logic and introduction of con-
fusion.

Once the brave but hopeless effort of the proponents of the
modern biological species concept is abandoned, it will again be
generally recognized that species in sexual and asexual organisms
have been and can be defined on essentially the same principle.
The principle is simply minimal irreversible evolutionary di-
vergence that yields readily recognizable difference. With this
principle guiding him, no greater task is placed upon the skill of
the taxonomist in classifying asexual organisms into species than

290 PROTOZOA

in doing the same with sexual organisms. He has in fact already
done it and, on the whole, has done it just as satisfactorily in the
asexual Protozoa as in sexual Protozoa or higher organisms, with
the readily corrected exceptions pointed out above.

The question of whether asexual equivalents of syngens exist
and can be recognized is more difficult, but not hopeless. The key
to progress in this direction is to recognize in the syngens of sex-
ual organisms a distinction between the means of ascertainment
and that which is ascertained. Breeding methods are the means
of ascertainment. The syngen which is thereby ascertained proves
to be a discrete group, the discontinuity of which with other
discrete groups is based upon complex genetic differences. The
discontinuity and the complexity of its genetic basis are the es-
sential features of the difference between closely related syngens.
This is what prevents the flow of genes between them and that
is why the test of gene flow is a means of ascertainment. The defi-
nition of a syngen on the basis of gene flow is an admirable opera-
tional definition. But conceptually the syngen could also be de-
fined on the basis of discontinuity and its complex genetic basis.
The same basis can be used for a concept of the asexual equiva-
lent of the syngen. Indeed, the same term, syngen, might be used
in both cases for it implies not only a common pool of genes but
also generation by a common group. Although another term might
be preferred for asexual organisms, the same one will tentatively
be employed here.

The obvious difficult) with the concept of the asexual syngen
just set forth is the lack of a simple, neat operational definition
comparable to the one that serves for the sexual syngen. This
statement of the difficulty of course assumes that there is no gene
flow among different lines of descent in organisms not known to
have sexual processes. In the light of modern researches on trans-
formation and transduction in bacteria, this assumption might be
questioned. If such mechanisms or others come to be discovered
in other asexual organisms, so much the better; but this cannot
be counted upon and may not occur. For the present at least,
oilier approaches must be sought. The problem then is how to
recognize irreversible evolutionary divergence in the absence of

T. M. SONNEBORN 291

a breeding test. In sexual organisms such divergence is associated
with complex and discontinuous genetic differences. In asexual
organisms also, such differences may be used as the sign of ir-
reversible divergence.

To discover whether such complex discontinuities exist within
a taxonomic species of asexual organisms, one would have to
make detailed comparative studies of many strains of the species
obtained from throughout its range. The comparisons would have
to include studies of the life cycle, cytology, morphology, physiol-
ogy, and ecology. Many features would have to be examined
experimentally under the same conditions to distinguish pheno-
typic from genotypic variation. As with "sibling species," syn-
gens discovered within an asexual species would not differ in
readily identifiable traits. If there were closely related sexual
species, the genetics of traits that differ in the asexual strains
should be studied in the sexual relatives. In short, every possible
means of arriving at a sound judgment about the complexity of
the genetic differences and their discontinuities would have to
be utilized. This would be a great labor; but great labor is also
involved in the analysis of a sexual species into its syngens. The
results of such studies on the species of Paramecium, summarized
in this paper, represent the work of 20 years by an increasing
group of investigators and the work is far from complete.

The concept of an asexual syngen as having passed the thresh-
old of irreversible evolutionary divergence, but not having
reached readily recognized differentiation, is thus the exact
equivalent of the sexual syngen and refers to the same level of
biological organization. In both cases, this level is the one which
embraces all the individuals that can potentially contribute to the
further evolution of the group. Thoday (1953) has presented a
comparable point of view. The evolutionary unit is as distinct and
isolated in asexual organisms as it is in sexual organisms, no more
and no less so. The syngen is sharp and definite in many out-
breeders; it becomes fuzzier and fuzzier the more inbreeding is
adopted. Similar differences in its distinctness may be anticipated
among asexual organisms. The difference between syngens in the
two cases is not in the concept or in its correspondence with

PROTOZOA

reality, but in the method of ascertainment which is the inevita-
ble result of the different methods of reproduction.

General Comments and Conclusions

Before passing to a general statement of the major conclusions,
two outstanding limitations of the present paper should be
pointed out. First, the restriction of the treatment of Ciliates to
those in which mating types have been found may well introduce
important biases. Such a discovery is likely to be made only in
organisms that are not too specialized in their conditions of life,
for the investigator has to be able to cultivate them in the labora-
tory. They also must conjugate under the conditions provided.
This excludes entirely asexual Ciliates, if such exist, and also
those with undiscovered fastidious requirements for mating. This
might also exclude organisms with the longest immature periods
or shortest mature periods. Further, even if conjugation occurs,
the demonstration of mating types is difficult or impossible if
there is excessive selfing. This would exclude some extreme in-
breeders. In fact, most work on conjugation and life cycles in
Ciliates prior to studies on mating types were performed on
selfers. In this category belong the classic studies on Uroleptus
mobilis by Calkins (1919, 1920) and on Spathidium spathula by
Woodruff and Spencer ( 1924 ) . These and other studies are rich
in information on life cycles. There is much need to bring them
into relation to the modern work and particularly to make in-
vestigations on breeding systems and speciation in these organ-
isms.

The second general limitation of the present paper is due to
the twin vices of ignorance and haste. As a newcomer to the fields
of taxonomy, speciation and evolution, I have only the most
superficial acquaintance witli the large and impressive literature
in these fields. As a result, I have doubtless failed in my obliga-
tions to make appropriate references to this literature. My onl\
defense is that strong pressure from editor and publisher allowed
me no time to correct my ignorance.

The major conclusions to be drawn from the present study con-
cern three general topics: (1) the problem of taxonomic species

T. M. SONNEBORN 293

and the modern biological species concept; (2) the relation of
breeding systems and methods of reproduction to the species
problem; and (3) the significance of differences in the major fea-
tures of life. These topics have been touched upon repeatedly
throughout this paper in connection with the treatment of the
various organisms discussed. The following resume brings to-
gether the relevant points on each topic.

The Problem of Taxonomic Species and the Modern Biological
Species Concept. The modern biological species concept of a
potentially common gene pool is in principle limited to a fraction
of existing organisms and can in practice be applied to but a
small fraction of these. It excludes all purely asexual organisms
and all obligatory self-fertilizers. Into this category fall perhaps
half or more of the Protozoa and a large number of plants in all
the major groups, as well as many invertebrates. Among the or-
ganisms to which the biological species concept applies in prin-
ciple, relatively few can ever be sorted with assurance into such
species because the task of accomplishing this in any one organ-
ism is so great.

The fact that closely related biological species sometimes can-
not be readily identified leads to the logical inconsistency of
grouping a number of species into a species. Thus the species
Paramecium aiirelia consists of at least 16 biological species (re-
ferred to as varieties in this paper). These can be identified only
under special conditions which are not generally available or only
with great labor and prolonged research. Biological species of
this sort are referred to as sibling species (Mayr, 1948) and may
remain unnamed. A single species name may be used for the
whole group of them. Regardless of the mental reservations about
what the "real" species are in such cases, this procedure is tanta-
mount to admitting that identifiability cannot be ignored in a
workable species concept. Beginning with logical inconsistency
we end in terminological inconsistency, using the word species to
denote three different things: taxonomic species not analyzed as
to gene flow, biological species, and groups of closely related and
practically unidentifiable sibling species.

The enthusiasm for the biological species concept in spite of

294 PROTOZOA

such limitations and confusions is perhaps to be explained in part
by the aura which surrounds the word species and in part by the
justifiable satisfaction of having delimited a biological level of
organization of great evolutionary importance. Darwin forged a
close connection between the term species and evolution theory.
Everyone knows the term. To confiscate it for an objectively
defined evolutionary unit is tempting. In the process, certain
broader considerations were put aside.

Classification into species preceded evolution theory and had
as its prime functions ordering and identification. It still has, and
for all kinds of organisms, not for outbreeders only. In spite of
the reproductive implications of the biblical account of creation,
with each kind of organism reproducing true to type, and in spite
of the early scientific usage of species as referring to the type-
maintaining original creations, it can hardly be denied that the
primary function of the word species was to designate a unit of
identification. This long usage has become deeply ingrained.
The universal need for it with respect to all organisms is recog-
nized by practically all biologists. However distasteful it may be
to some geneticists, there seems to be no possibility that biol-
ogists in general will ever agree to restrict the term species to
the carriers of a common pool of genes.

Even the proponents of that concept of species would probably
agree with most of my statements in the immediately preceding
paragraphs. Dobzhansky ( 1941 ) himself makes most of these
points. He says that the systematists task must be primarily to
pigeon-hole, that the term species is used in more than one
sense by being applied both to taxonomic species and common
gene pools, that species are discrete by reason of complex ge-
netic differences, and so on. But he is willing to accept this
situation along with its logical inconsistency, terminological in-
consistency, and confusion.

Only deep attachment to the word species seems to prevent
recognition that the needs of biologists would best be served by
having a separate term, such as syngen, for the unit of evolution
when it is known not to be the same as the unit of identification.
When the) are known to be the same, then they are at once both

T. M. SONNEBOHN 295

a species and a syngen. The reverence now accorded, often un-
justly, to the term species would in time be transferred to the
term syngen, or whatever term may be adopted, because of its
greater evolutionary significance. The taxonomist and the ge-
neticist would both be free to perform their different tasks with
a logical and consistent terminology, neither being frustrated by
conflict with the other. The major remaining problem is whether
the two terms can be generalized.

Generalization of the term species for the unit of identification
offers no great difficulty in principle. It simply requires a verbal
statement which embodies current sound taxonomic practice.
The first requirement is that the species be readily and visibly
identifiable as a group of organisms manifesting discontinuity in
morphology with all other groups. Secondly, the level of visible
discontinuity ideally is the simplest one judged to be untrans-
gressible by genetic means, either recombination or mutation.
Groups of organisms that are judged to show minimal, irrever-
sible, visible divergence thus are assigned to different species.

The generalization of syngen is more difficult. Failure to recog-
nize the equivalent of the syngen in organisms that cannot out-
breed is due to the identification of the method of ascertainment
with the thing ascertained. The discreteness and reality of the
latter is based upon the genetic complexity of the features that
mark off a syngen from closely related syngens. This is what pre-
vents gene flow between them, and it is the reason that the test
of gene flow serves for ascertainment. Conceptually therefore a
syngen can be defined by the complexity of its genetic distinct-
ness from other syngens. A syngen, like a species, has thus passed
the threshold of irreversible evolutionary divergence; but, unlike
a species, it need not show readily recognized visible differentia-
tion. This level of biological organization is difficult to discover
in any kind of organism, but it can be done with any kind. With
outbreeders, the method is to test for gene flow. With asexual
organisms, the method is to compare different strains of a species
from every possible point of view and, with the fullest possible
array of facts, to arrive at judgments on discontinuities and the
probable complexity of their genetic basis. In spite of the differ-

PROTOZOA

ence in method imposed by the difference in reproduction, cor-
responding levels of biological organization are thereby defined.
In both cases, the level of organization includes the group of
individuals that can potentially contribute to the further evolu-
tion of the group. Differences in the sharpness of delimitation
of the asexual syngen are to be expected. Similar differences are
found in sexual organisms, as will now appear.

The Relation of Breeding Systems and Methods of Reproduc-
tion to the Species Problem. The preceding attempt to general-
ize the biological species or syngen runs counter to the view of
proponents of the biological species concept (e.g., Dobzhansky,
1937, 1941 ) that biological species do not exist among obligatory
inbreeders or asexual organisms. This denial, as indicated above,
is based upon an operational definition of biological species. Since
the operation, testing gene flow, is impossible in asexual or-
ganisms, they deny the existence in them of the thing this opera-
tion discovers in sexual organisms, i.e., the biological species or
syngen. Their statement of the situation thus implies an abrupt
change in the organization of nature and in the units of evolu-
tionary divergence correlated with an abrupt change from out-
breeding to obligatory inbreeding and asexual reproduction. By
subordinating the method of ascertainment to the thing ascer-
tained and by seeking methods of ascertainment in asexual repro-
duction, the concept of biological species or syngens was
generalized. This implies the absence of an abrupt change in the
organization of nature and in the units of evolutionary divergence
with changes in breeding system or method of reproduction.

No such abrupt change is in fact found in the present review of
conditions in the Protozoa. On the contrary, this review shows a
progressive series of changes in the organization of nature and in
the units of evolution running parallel to the sequence of breed-
ing systems from extreme outbreeding through various degrees
of out- and inbreeding to extreme inbreeding. The passage from
inbreeding to asexual reproduction is accompanied by the final
gradual change in the series. Nowhere does an abrupt, marked
change occur. The continuity of the whole series stresses the
unity of the problems of species and syngens in both sorts of

T. M. SONNEBORN 297

reproduction and justifies the attempt to generalize these con-
cepts.

The progessive series of changes accompanying the sequence
of breeding systems in the Ciliates includes chiefly the following:
the number of closely related syngens per species increases; the
range of distribution of a syngen, especially in latitude, becomes
more and more restricted; the number of local populations per
syngen becomes smaller and smaller; the genetic divergence of
each local population becomes greater and greater. Thus, the
outbreeding species Paramecium bursaria includes only three
syngens in North America, while the inbreeding species P. cauda-
tum includes 12, and others include more, perhaps many more.
Outbreeding syngens, like variety 1 of P. bursaria, variety 15 of
P. aurelia, and variety 2 of Tetrahymena pyriformis, cover a
range from cold to very warm latitudes; while inbreeding syn-
gens, such as variety 10 or 14 of P. aurelia, may be restricted to
a small number of populations near the Gulf of Mexico. An out-
breeding syngen, such as variety 1 of P. bursaria, is genetically
so much alike throughout its range that crosses between widely
separated populations yield far more viable progenies than in-
breeding, and no differences in chromosome morphology or size
are found throughout its range. Different populations of an in-
breeding syngen, such as variety 4 of P. aurelia, are able to inter-
breed in the laboratory, but the F2 shows high nonviability,
sometimes approaching 100%, in contrast to the virtual absence
of nonviability after inbreeding within a population. Further,
each of its local populations that has been examined has a unique
chromosome number in an aneuploid series and shows unique
chromosome morphology and size. Syngens with intermediate
breeding systems, such as many varieties of P. aurelia, are inter-
mediate in range and in the genetic divergence of their popula-
tions. The extreme narrowness of range and divergence of
populations may be shown by the P. caudatum of Europe; each
local population gives more or less Fl nonviability when crossed
with other local populations. But that situation is not yet entirely
clear.

Actual, as contrasted with potential, gene flow thus becomes

298 PROTOZOA

progressively more limited with the passage through various
grades of breeding systems from extreme outbreeding to extreme
inbreeding. As the extreme of inbreeding is approached, each of
the many local populations of a Ciliate actually approaches in its
isolation and genetic differentiation the status of a syngen. Al-
though evolutionary divergence has not yet passed the point of
no potential for return, the actual level of biological organization
is as different as possible from that observed in outbreeding
syngens. Yet all gradations are found from one extreme to the
other. Although little information is available on asexual Protozoa,
the degree of isolation and genetic differentiation of local popu-
lations supposed to exist in them could be at most but little more
than actually occurs near the inbreeding extreme of the series
in sexual organisms. Both the potential and the actual group of
interbreeding populations contracts progressively with progressive
changes in the breeding system and the final change in method
of reproduction. The significance of the syngen as an objective
and real level of biological organization correspondingly de-
creases as the local population progressively increases in such
significance.

The Significance of Differences in the Major Features of Life.
Among the Ciliates, differences in the major features of life ap-
pear chiefly as quantitative variations in the duration of the
various stages of a fundamentally similar life cycle, as variations
in the details of a fundamentally similar mechanism of mating
type determination, and as variations in the form of fertilization
characteristic of the senile stage of life. Previously these varia-
tions were known as mere brute facts. In this paper they have
been shown to be of the greatest significance as adaptations to
the breeding system by which the syngen perpetuates itself.

Variations in the duration of stages in the life cycle are found
with respect to periods of immaturity, maturity, senility, and
total life. These are all long in outbreeders, short in inbreeders,
and intermediate in those with intermediate breeding systems.
The same sort of general relation has been pointed out in certain
plants by Stebbins (1950): self-fertilizers are chiefly annuals and
outbreeders are chiefly perennials. In Ciliates, long immaturity

T. M. SONNEBORN 299

serves to give outbreeders an opportunity to migrate away from
close relatives before they can mate; short or no immaturity per-
mits inbreeders to mate when they are still nearest to their close
relatives. Long maturity gives outbreeders time to find suitable
mates; short maturity in inbreeders restricts the opportunity to
mate with others. Long senility, during which mating is still
possible but less likely to yield viable progeny, occurs in out-
breeders and extends the opportunity to find suitable mates; the
relatively short senility of inbreeders is characterized by inability
to mate at all, or by greatly reduced ability to mate. During
senility in inbreeders, the form taken by fertilization is autogamy,
the closest form of inbreeding; and this stage of life is by far
the longest, immaturity being contracted virtually to nothingness
and maturity lasting only a matter of some days. Outbreeders do
not undergo autogamy. If they fail to find a suitable unrelated
mate, eventually complementary mating types arise within a
single line of descent and conjugation occurs within the line.
This serves to rejuvenate, initiating a new life cycle, and at the
same time serves to maintain heterozygosity, which is not accom-
plished by autogamy.

Outbreeders commonly have systems of multiple interbreeding
mating types, while inbreeders commonly have but two. Multiple
types serve to increase the possibility of mating when an out-
breeder meets a stranger. The method of mating type determina-
tion in outbreeders results in uniformity of mating type of the
synclone (the two clones from a pair of conjugants). This pre-
vents the closest forms of inbreeding. Inbreeders, however, regu-
larly yield more than one mating type in a synclone; this serves
to make the closest relatives capable of mating with each other.
The several systems of mating type inheritance (except the one
in Euplotes ) are based on a single fundamental system of mating
type determination, with relatively minor differences in detail.

In asexual Protozoa, the most important differences in life
features are with respect to the rapidity of reproduction, the size
of populations, the ploidy level, and the constancy or variability
of the conditions of life. Haploids that reproduce rapidly into
large populations can probably depend upon mutations to provide

300 PROTOZOA

sufficient genetic variability for adaptation to widely varied con-
ditions. They are thus comparable to the sexual outbreeders. On
the other hand, asexual diploids are comparable to sexual in-
breeders. They cannot achieve so readily genetic adaptation to
varied conditions. They are thus confined largely to relatively
constant or very slowly changing conditions, such as life in a
host or in a considerable body of water which acts as a buf-
fer against rapid change. A different sort of genetic basis for
variability is found in them: the possession of multiple gene loci
for alternative and mutually exclusive traits which are readily
transformed one into another in response to different environ-
mental conditions. Various intermediate types of Protozoa are
found to manifest various combinations of the mechanisms and
features just mentioned, e.g., haploid asexuals with small popula-
tions and slow reproduction share some of the diploid asexual
features, as does the well-developed asexual phase of some sexual
Ciliates.

All systems of breeding and both methods of reproduction have
proved themselves by existing and persisting in nature. Each
must be well adapted to the conditions under which it is found.
What one system or method lacks must be compensated by some
other feature it possesses. This does not mean that all systems
are equally likely to succeed on the evolutionary time scale.
Radical changes in conditions may periodically purge inbreeders
and asexuals which flourish under relatively constant conditions.
Their specialization to such conditions may give them a short-
term advantage over outbreeders. But those inbreeders and
asexuals that select the constant environments of hosts have
also persisted long on the evolutionary time scale, as is clearly
shown by the Flagellate symbionts of roaches and termites.

The survey of the Protozoa indicates that there an1 relatively
lew fundamental alternatives open to such organisms, perhaps
to an\ organisms. On the one hand, they can evolve systems of
genetic variability that permit adaptation to a wide range ol
conditions of life, as is done by sexual outbreeders and asexual
haploids that rapidly produce large populations. On the other
hand, the} can evolve highly specialized adaptations to relatively

T. M. SONNEBORN 301

constant and restricted conditions of life, as is done by sexual
inbreeders and asexual diploids. With the choice of one or the
other fundamental alternative go a host of further adaptations
to the system of breeding or reproduction, as set forth above in
detail for the Ciliates. The superficially chaotic differences in the
features of life of different closely related syngens acquire deep
significance as systematically selected means to the great end of
successful survival. The correlated problems of speciation that
tax the taxonomist and generate concepts from the geneticist are
trivial man-made difficulties that divert attention from the funda-
mental beauty and simplicity of nature's ways of meeting nature's
problems.

Summary

Statement of Problems and General Background. (1) Prob-
lems. Can biological species, i.e., the common gene pools, of
Ciliates be readily identified? Should they be assigned species
names? Should they be considered as species? Is it possible to
arrive at concepts of species and asexual equivalents of "biolog-
ical species" which would be generally applicable to inbreeders
and asexual organisms as well as to outbreeders? What is the
bearing of breeding systems and methods of reproduction on
species and evolutionary problems? What is the relation of the
varied life features of Ciliates to their breeding systems?

2. Background. Inbreeders and outbreeders among the Cili-
ates were early recognized. Since the discovery of mating types
(1937), each taxonomic species has been found to include a
number of "biological species" or varieties. Each variety includes
a number of populations which are potentially able to interbreed;
but the varieties are sexually or genetically isolated. Information
is fullest on several taxonomic species of Paramecium. The chief
differential characters of these species are given.

The Specificity of Mating Types and Isolation of Varieties in
P. aurelia. Of the 16 known varieties, all have two mating types
except variety 16 which has four.3 Although flow of genes can
occur in the laboratory between different populations of the
same variety, the F2 generation shows in some varieties con-

PROTOZOA

siderable mortality, approaching 100% in certain crosses. Parallel
and corresponding mating types occur in different varieties, but
each mating type is uniquely specified by its pattern of sexual
reactions. In most cases there is no attempt of different varieties
to interbreed or the attempt is unsuccessful; sexual isolation is
complete. In the few combinations of varieties that can inter-
breed, the Fl or F2 is almost completely nonviable, and survivors
are always weak. The latter are incapable of yielding normal
vigorous progeny either by inbreeding or by backcrosses. Ge-
netic isolation is virtually complete.

Only two varieties are known to have worldwide longitudinal
and latitudinal distributions; two have wide, but not so wide,
distributions. The rest are restricted to a single continent, some-
times to but a small part of it. The varieties are distributed in
temperature clines, many overlapping; not uncommonly more
than one variety is found in a small sample of a single body of
water.

Differences among the Varieties of P. aurelia and the Problem
of Identification. Differences of the following sorts are found
among the varieties: mating type specificity; geographical dis-
tribution; maximal and minimal tolerated temperature (these not
always being completely correlated with geographic distribu-
tion); fission rates at the same temperature; body length; number
of micronuelei in early life; the occurrence of killer and resistant
strains; specificity of serotypes; system of determination and
inheritance of serotypes; system of mating type inheritance and
determination; the existence of and nature of diurnal periodicities
in sexual reactivity; the temperature optima for mating; other
conditions for mating; the occurrence of and duration of a period
of immaturity alter conjugation (immaturity never occurs after
autogamy ); the duration of the period of sexual maturity; the
frequency of occurrence of lines that self-conjugate during ma-
turity: the occurrence of autogamy or selfing during senility;
the period of survival after loss of micronuelei during senility.

Each ol these many differences between varieties was de-
scribed in lull. No one of them, except mating type and serotype'
specificities, can be used to define or identify all the varieties,

T. M. SONNEBORN 303

nor can any combination of the others serve this purpose. Most
of the traits just distinguish a few "roups of varieties.

Only mating type specificity is sufficient to identify all varie-
ties. Serologic specificity may, with fuller knowledge, eventually
serve the same purpose. Great labor, technical knowledge, skill,
and elaborate materials are required to use either mating types
or serotypes for identification. Either living standard cultures of
all mating types or hundreds of specific antisera are required. The
difficulties and hazards involved in identification clearly preclude
assigning species names to the varieties.

Objections are also raised against considering the varieties to
be species. The term "syngen" is proposed to replace "biological
species."

Breeding Systems and Evolution in P. aurelia. Fertilization
serves two functions: genetic recombination and initiating new
life cycles. Death (genetic, somatic, or both) eventually occurs
in the absence of fertilization. Fertilization occurs either as con-
jugation or as autogamy. Varieties that undergo autogamy do
so at a later stage of life (senility) than the one (maturity) in
which they are ripe for conjugation. Varieties that lack autogamy
are adapted only to cross conjugation during maturity, but to
cross conjugation and self-conjugation during senility. Natural
selection has thus given prior choice to conjugation and cross
conjugation. The relative adaptation to cross conjugation and
to selfing or autogamy differs greatly among the varieties. Few
are strong outbreeders, many are intermediate, and many are
strong inbreeders. The different breeding systems of different
varieties are supported by differences in their life features. Out-
breeders have long periods of immaturity, maturity, senility, and
total life. This gives opportunity to migrate away from relatives
before mating and to have much time to find a suitable mate. In-
breeders have short lives, with no immature period and a very
brief mature period. They must mate quickly with the close rela-
tives near at hand or undergo autogamy. Other adaptations to the
breeding system, fully explained in the text, concern the number
of mating types (two in inbreeders, more in outbreeders), the
fission rate (low in outbreeders, high in inbreeders), the form of

304 PROTOZOA

fertilization during senility (setting in outbreeders, autogamy in
inbreeders), and the method of mating type inheritance (syn-
clonal uniformity in outbreeders, its absence in inbreeders ) . The
breeding system actually determined depends on the particular
combination of these features; when an inbreeding and an out-
breeding feature occur together, one may negate the other. Thus
the combination of a period of immaturity with a very brief ma-
ture period followed by autogamy leads to inbreeding in spite
of the usual association of immaturity with outbreeding. Other
examples are also cited.

Each local population examined in one inbreeding variety has
a uniquely different chromosome set. This seems to be the basis
of F2 mortality following crosses between populations. Strongly
inbreeding varieties are thus a loose assemblage of highly isolated
local populations, each of which has a status that just falls short
of being comparable to that of a whole outbreeding variety. An
outbreeding variety represents a clear-cut level of biological or-
ganization; inbreeding varieties do not. In the latter, the local
population, not the variety, is the significant unit of biological
organization.

Data are given to indicate the low population densities even
at the season of maximal populations; and to show that, within
a population of certain varieties, mating is not random; inbreed-
ing prevails. Other observations indicate that migration into and
out of a single body of water is a very rare event: the composi-
tion of a population remains constant over the years. Specula-
tions on means of migration and its consequences under various
conditions are presented.

The varieties are divisible into two groups, A and B, on the
basis of differences in mating type inheritance1. This is paralleled
by differences in the serotype system and the presence or absence
of killer strains.1 It seems to represent an early evolutionary cleav-
age in P. aurelia. Similarities in serotypic and mating type speci-
ficities are greater within group A and within group B than
between groups. Geographical distribution of the varieties in-
dicates that within each group a widely distributed variety has
given rise near the temperature limits of its range1 to narrowly

T. M. SONNEBORN 305

confined varieties. The group A and group B mating type systems
are found in other species of Paramecium and in Tetrahymena
pyriformis, suggesting repeated independent evolutionary diver-
gence along these lines. A general formulation of the two systems
is given from which it appears that one or a few changes in de-
tails could shift from one system to the other. The major isolating
mechanisms in P. aurelia are listed.

Paramecium caudatum. This species consists of 16 known va-
rieties, with two mating types each. Only four varieties occur on
two continents, the rest on one only. The varietal differences
include: mating type specificity, geographical distribution, ani-
mal size and shape, micronuclear size and shape, optimal tem-
peratures for mating, and resistance to a killer of P. aurelia. Only
mating type specificity serves to distinguish all varieties. None of
the other differences and no combination of them can, so far as
present knowledge goes, serve to identify the varieties. Assigning
species names would be futile.

With respect to gene flow, the varieties fall into two groups, 1
and 2. The 9 varieties of group 1 are absolutely isolated sexually
from all other varieties. The 7 varieties of group 2 show a limited
system of interbreeding, but the pattern of mating reactions of
each mating type is unique. The possibility of gene flow among
them has not yet been excluded. European workers are unable
to sort their collections into varieties. They find fair viability
after conjugation within a local population; but variable, though
usually less, viability after conjugation between different popula-
tions. This refers to the Fl only. Little is known about F2 results
in P. caudatum after crosses between populations of the same
or different varieties. The Fl results on the European material
suggest that each local population may be more or less isolated
from all others, but the matter is in an unsettled condition.

Most varieties, if not all, of P. caudatum appear to be inbreed-
ers. Immature periods are unknown and, if they exist, must be
relatively short. Selfing is common. The group B method of mat-
ing type inheritance assures the usual existence of both mating
types in a synclone. Some selfing is induced in mixtures of mating
types. Autogamy seems to occur in some varieties, but not in

306 PROTOZOA

most. Selfing appears to take its place. All these features, plus
outbreeding Fl nonliability, point toward inbreeding. There
may be one or more outbreeding varieties, but the evidence on
this is still not clear.

Paramecium bursaria. Five varieties of P. bursaria are well
known. One of these has two mating types, three have four
mating types, and one has eight mating types. The sexual isola-
tion of these varieties is complete except that one of the two
mating types of one variety conjugates with four of the eight
types of another variety. This exception is fruitless: all the con-
jugants die. Fruitful conjugation can occur between local popu-
lations of the same variety. Each variety thus constitutes a
potentially common gene pool which is completely cut off from
all others. Two of the varieties are confined to North America,
two to Europe; one is common to Asia and North America.

Varieties differ in the following traits: number of mating types,
specificity of mating types, diurnal periodicities in sexual reactiv-
ity; speed of conjugation; length and thickness of chromosomes;
and geographical distribution. Identification of varieties is simpler
than in the other species, partly because not more than three va-
rieties are known from any land mass. These three could be
readily distinguished if all the mating types were available, even
without living mating type standards. Since the task of obtaining
all mating types of a variety could be very laborious and time-con-
suming, this is not a satisfactory basis for a species description.

In variety 1, the only one very extensively studied, the life
features mark it clearly as a strong outbreeder. The immature1
period is 3 to 5 months with continuously favorable conditions,
but even a short unfavorable period triples the duration of imma-
turity. In the laboratory, maturity lasts about three years, senility
about four to five years. In nature, all these stages of life arc
probably much longer than in the laboratory. Unlike the prc\ ions
species, P. bursaria can mate readily throughout senility; but
during this period the fertility, i.e., survival after conjugation,
progressively decreases. This variety thus has remarkably long
periods for dispersal before mating can occur and much longer
periods (at least 7 years) during which mating can occur, as

T. M. SONNEBORN 307

befits an outbreeder. The selection of such long mating periods
suggests that the achievement of fruitful cross-mating is relatively
rare. Other adaptations to, and correlates of, outbreeding are the
systems of multiple mating types, the almost invariable synclonal
uniformity in mating type, the total absence of selfers of the
sort found in other species, the very much higher death rate
after inbreeding ( sib mating ) as compared with outbreeding, the
rapid rise in death rate through successive inbreedings. A puz-
zling exception is associated with the very rare production of a
complementary mating type in a clone. When this mates with the
original type in the clone, survival is as good as in outbreeding.
Whatever the basis of this exceptional inbreeding, the result is
highly adaptive, since, like autogamy and selfing in senility in
P. aurelia, it both rejuvenates and permits the retention of
heterozygosity.

The synclonal uniformity in mating type is based upon the
P. aurelia group B system of mating type determination plus
the regular massive exchange of cytoplasm between mates. The
rare exceptions (2.6% of the synclones) which yield more than
one mating type per synclone reveal the basic B system. A hy-
pothesis is presented to account for all the remarkable kinds of
observations on the inheritance of mating type which have
hitherto remained unexplained. The system of mating type de-
termination in P. bursaria is the most complex of those discussed.
Loss of the mechanism assuring regular cytoplasmic exchange
between mates transforms it to the P. aurelia form of the group B
system. Loss of one or two other features (explained in the text)
converts the B into the A system. The P. bursaria system is a
strong adaptation to outbreeding by almost completely blocking
mating within a synclone.

The highly developed complex of adaptations to outbreeding
in P. bursaria inhibits progressive divergence of local populations.
As a result of the wide distribution of potentially common gene
pools, speciation is poor in comparison with that of inbreeders.
In the United States there are probably only three varieties of
P. bursaria as compared with at least 14 (probably more) va-
rieties of P. aurelia.

308 PROTOZOA

Tetrahymena pyriformis. At least nine varieties of this species
have been found in a survey confined to North, Central, and
northern South America. There may well be more varieties in this
region, even among the materials already collected, because a
large proportion of the collections could not be classified for tech-
nical reasons ( selfers and amicronucleates ) and the basis of clas-
sification is solely the capacity to mate, genetic isolation not yet
having been systematically studied. All but two of the varieties
are confined to North America, four of them being concentrated
in the northern latitudes of the United States. Of the other two,
one (variety 2) ranges from Canada to northern South Amer-
ica, and the other (variety 9) is limited to the southernmost part
of this range. All varieties have five chromosomes in the haploid
set, and their distinctive morphology is the same in all varieties.
The number of known mating types is two in three varieties,
three in two varieties, and five, seven, eight, and eleven in one
variety each. These are minimal figures based chiefly on collec-
tions from nature; breeding analysis in the laboratory has re-
vealed some not yet found in nature and may reveal more.

Varietal differences include the following: number of mating
types, system of mating type determination, geographical dis-
tribution, preference for running or still water, size, possibly
ciliary patterns, and length of immature period. Little is yet
known about possible differences in the characteristics and dura-
tion of the periods of maturity and senility. In some varieties,
senility is marked by loss of micronuclei and capacity to mate;
but the organisms may live at least 20 years in this condition.
Variety 1 may begin to reach this stage at ages 500 to 1500 fis-
sions. Senility may be marked by the onset of selfing with de-
creasing survival accompanying increasing parental age in some
varieties; but this is not yet certain. Among the varietal differ-
ences, at present only mating type specificity can be used to
identify all the varieties. This requires live standard cultures of
at least two mating types of each variety. Since identification by
routine taxonomic procedures is impossible, assigning specific
names to the varieties is not practicable.

Knowledge of the breeding system is greatest for varieties I

T. M. SONNEBORN 309

and 2. Variety 1 with seven mating types shows the P. aurelia
group A system of mating type determination. A pair of alleles
limiting mating type potentialities has been found. This variety
has an immature period and is heterozygous in nature. It shows
hybrid vigor and inbreeding degeneration, but normal fertile
inbred lines can be selected eventually. During meiosis in crosses
of wild strains evidences of chromosomal aberrations are appar-
ent. The possible relation of these and the nonviability after
conjugation to age-induced chromosomal aberrations has not
been studied. Much selfing occurs during maturity in variety 1
when cultivated with excess food, but this is greatly reduced by
periodic starvation. In nature it therefore is probably not very
common. Taking all these features into account, variety 1 appears
to have an intermediate breeding system. It has many outbreed-
ing stigmata: immaturity, multiple mating types, natural hetero-
zygosity, and inbreeding degeneration. But it also has some
inbreeding features: distribution is concentrated in a narrow
range of latitude; synclones almost always include more than
one mating type; normal fertile inbred lines are obtainable; self-
ing occurs; and structural chromosomal differences between
natural strains lead to abnormality and death when they are
crossed. Some of these inbreeding features are to a degree com-
pensated by outbreeding features: synclonal nonuniformity of
mating type by immaturity and selfing by the stabilizing effect
of starvation. Others, such as chromosomal aberrations may be
peculiarities of aging. On the whole, therefore, variety 1 seems
to be primarily an outbreeder, but less extremely so than P.
bursaria. It has latent mechanisms for inbreeding, which it would
appear could easily give rise to inbreeding descendants.

Variety 2 is better adapted to outbreeding and shows more
stigmata of that system of breeding. It has an immature period,
the largest number of mating types (11), the widest latitudinal
range, and synclonal uniformity of mating type due to the P.
bursaria form of the P. aurelia group B mechanism of mating
type determination. Little is known about the breeding systems
in the other varieties. Crosses between different populations
within variety 5 and within variety 9 yielded no survivors; the

310 PROTOZOA

crosses of variety 9 showed much chromosomal abnormality at
meiosis. Some survivors were obtained from crosses of popula-
tions of variety 6. Again the role of aging in these results is
unknown. Further work may well reveal other marked differences
among the varieties in their breeding systems.

Enplotes. There is a group of morphologically distinct species
which are, however, much like E. patella. In three of these species
(E. aediculatus, E. eurystomus, E. patella) and in the less closely
related E. harpa, mating types forming one or more varieties
are known, but there has been no effort to discover the full
number of varieties in any region in any of these species. Knowl-
edge is chiefly confined to one variety of E. patella, indeed to
the progeny of two wild individuals from neighboring ponds.
These were of different mating type and yielded four other
mating types among their sexual progeny. Doubtless more, pos-
sibly many more, mating types of this variety exist in nature.

The method of mating type determination is entirely different
from the ones thus far described. The six types are determined
by the three homozygous and three heterozygous combinations
of three codominant alleles. Because of the genotypic identity of
the two exconjugants of a pair, each synclone is uniform in mat-
ing type. Mating is normally controlled by the action of three
hormones secreted into the medium, each hormone being under
the control of one allele. An animal is activated, or put into
mating condition, by exposure to a hormone which it cannot
itself produce. Animals thus activated can mate with each other
regardless of whether their mating types are alike or different.
Thus, "mating type" has a somewhat different meaning in this
material.

E. patella is an outbreeder in spite of sonic superficial indica-
tions to the contrary. Since any two activated animals can mate
with each other, mixtures of two mating types lead to much
selfing. But in nature, the lesser population densities and the
large volume of diluent probably restrict hormone action to
contiguous pairs, and these could conjugate only il each acti-
vated the other, a result which is possible only if they arc of
certain different mating types. Tin's is then actually an adapta-

T. M. SONNEBORN 311

tion to outbreeding, not inbreeding. Other indications of out-
breeding are: the system of multiple (probably many) mating
types, a regular immature period, and high mortality after in-
breeding. On the other hand, some setting may occur, as this was
observed even within a clone. Its possible relation to age is un-
known. Eventually work on E. patella was abandoned because
the animals became difficult to culture. Whether this was due
partly to aging and inbreeding is unknown.

Other Ciliates. In Paramecium trichium three interbreeding
pure mating types are known. Selfers occur, but their relation
to the former is unknown. P. calkinsi has at least two varieties
with two mating types in each. Colpidium truncation has multi-
ple mating types and much selfmg in the one variety observed.
Stylonychia putrina includes at least two varieties with at least
11 and 15 mating types, respectively, and selfmg is very rare.

One variety with nine mating types has been found in Oxy-
tricha bifaria. Some were collected in nature, others arose among
their sexual progeny. Selfmg is very rare, no hormones can be
demonstrated, and conjugation is regularly between unlike mat-
ing types. This variety is an outbreeder. In addition to multiple
mating types and the rarity of selfmg, other indices of out-
breeding are a considerable immature period and a long mature
period. The one culture which selfed a little usually failed to
survive selfmg, but usually survived when crossed to one of the
two other types reported upon. The system of survival and
death in relation to inbreeding and outbreeding is still unknown.

Obligatory Inbreeding and Asexual Protozoa. The Flagellates
which are symbiotic in wood roaches and termites seem to show
an evolutionary series from sexual to asexual reproduction. The
sexual species are found in wood roaches; the related asexual
species occur in termites. The coincidence in time between sexual
processes in the Flagellates and transmission to new roach hosts
seems like a mechanism for providing genetic variability in the
symbiont at the time it might be needed for adaptation to varied
hosts. But the biology of the roaches indicates little variation
among the individuals of a colony, and the sexual processes in
the svmbionts are such as to vield little or no variability in them.

312 PROTOZOA

Hence, these sexual processes are viewed as vestiges of more
nearly perfect recombination processes of the ancestors of the
symbionts when the latter were more variable. This is consistent
with the absence of sexuality in the closely related symbionts of
termites, which are more highly evolved than the roaches, and
with the fuller development of sexuality in some of the simpler
plant Flagellates.

There are two major species problems in asexual Protozoa. One
grows out of the gene pool concept of species, for there are no
common gene poles in such organisms and therefore no species
in that sense. The other grows out of the routine taxonomic
usage of the term species, for little attention has been given to
establishing similar usage in sexual and asexual organisms. In
the latter, physiological differences alone have been used in
some cases, though morphological differences are sometimes held
to be essential. In my opinion, similarity of usage requires recog-
nition of genetic differences of the same kind and magnitude in
both sexual and asexual organisms.

The situation in Trypanosomes illustrates the inadequacy of
both physiological and some morphological bases for distinguish-
ing species. Serologic differences in Trypanosomes can be inde-
pendent of genie differences, as in Paramecium aurelia, and
similar conditions may be expected in other asexual organisms.
Ignorance of the genetic identity of lines showing very different
serotypes could lead to assigning members of the same clone
to different species. Other physiologic differences, such as viru-
lence and host specificity, are used to distinguish species of
Trypanosomes; but in sonic organisms these are known to be
single gene differences. Confinement to different hosts is not
equivalent to geographic isolation when the adaptation to differ-
ent hosts depends on single mutational steps. Simple morpholog-
ical differences, such as the presence or absence of a kinetoplast,
are also used to distinguish species. These types also arise at a
single step by a mechanism comparable to loss of kappa in Para-
mecium. Tin's too is indefensible, for it could assign two products
of one fission to different specie's. The extent to which consider-
able morphological variations can quickly arise within a clone is

T. M. SONNEBORN 313

well shown by the rapid selection of very diverse types within
a clone in Rhizopods and the selection of similar types from
initially different clones. Thus genetically simple morphological
differences as well as genetically simple physiological differences
are equally inadmissible as species differentials if species are to
represent equivalent evolutionary divergences in sexual and
asexual organisms. Many criteria now used in asexual organisms
to distinguish species are comparable to individual differences in
sexual organisms.

Two major mechanisms of variation are recognizable in asexual
organisms. One is based upon mutation alone. It operates success-
fully in haploids that rapidly produce large populations. The
other works equally well in diploids. This involves having in the
genotype multiple loci for alternative, mutually exclusive, and
rapidly transformable traits. It is illustrated by the serotype sys-
tem in Paramecium and Trypanosomes. This mechanism operates
in purely asexual organisms, in sexual organisms with a well-
developed asexual phase, and in the asexual phase, that is, the
phase of development, in higher organisms.

Speciation appears to be fundamentally the same in sexual
and asexual organisms. In both it is based upon genetic adapta-
tion to changing conditions. It may proceed more slowly and
less abundantly in some asexual organisms because of their
choice of relatively constant conditions of life, such as in a host
organism or in a considerable body of water. The asexual method
of reproduction has proved its adequacy if not its advantage
under certain conditions, by its widespread occurrence in nature.
It has means of rapidly spreading and multiplying advantageous
variants, and it avoids the wastage of recombination and the
swamping of variations which results from crosses. However,
short-term advantage is no guarantee of long-term persistence.
While some asexual organisms may be wiped out in times of
stress, others have succeeded on the evolutionary time scale by
adapting themselves to a constant or very slowly changing en-
vironment, such as life in a host.

It is possible to adopt principles of classification into species
and syngens which are equally applicable to sexual and asexual

314 PROTOZOA

organisms. This requires reducing species to one of its current
meanings. By restricting it to designate a readily identifiable
group which shows the minimal irreversible discontinuity with
other groups, species can be set up just as well in asexual as in
sexual organisms. It is primarily designed for the convenience of
biologists and is not intended to represent a biological reality.
The latter is reserved for the term syngen. This designates the
group which is potentially able to contribute to the further evo-
lution of their descendants, i.e., a group within which irreversible
evolutionary divergence has not yet occurred. When this cannot
be defined by breeding methods as in asexual organisms, it is in
principle possible to do so by intensive comparison of every as-
pect of many strains and judging by every means available the
complexity of their genetic differences and the limits of untrans-
gressible discontinuities. This is precisely what is discovered by
breeding methods in sexual organisms. In both cases the syngen
is independent of readily recognized differences and depends only
upon complex genetic deviation associated with untransgressible
discontinuity.

Conclusions. The conclusions on the three major subjects
dealt with in this paper are given in full in the section and will
not be repeated here.

Addenda

1. In the six-month interim between writing and proofreading.
new discoveries about P. aurelia make necessary the following
changes.

2. The P. aurelia-multirnicronucleatum complex is not divisible
into well-marked species on the basis of currently used criteria:
body size and micronnclear number; but at least three distinct
types are distinguishable by size of micronuclei and structure of
their central chromatic area. The mieronuclear diameters are 2.6 /-,
3 /j. and more than 4 /a; their chromatic areas are solid, doughnut
shaped, and spongy, respectively. The first combination is typical
multimicronucleatum (varieties 13, 15, and new 16, see addendum
3) and it is usually associated with more than two micronuclei;

T. M. SONNEBORN 315

the second is typical aurelia; the third is characteristic of variety
12. (See pp. 161, 162, 169, 178, 179, 198, 232.)

3. My variety 15 and Giese's variety 16 are almost certainly
identical; both will hereafter be called variety 15. It has two con-
stant mating types (XXIX and XXX) with diurnal periodicity
stated in the text. It also has two other types (XXXI and XXXII).

XXXI behaves as if it was XXIX at night, XXX during the day;

XXXII behaves in the reverse way. There are thus four mating
types based upon two mating specificities. Not all strains require
tube culture in order to conjugate. Another variety (from Mexico)
is now called variety 16; it has the aurelia body size, but multi-
micronucleatum micronuclear size, number, and structure (see
addendum 2). (See pp. 163, 166, 169, 171, 178, 179, 185, 189,
190, 194,198,207,301.)

4. Beale ( in press ) has found a mate killer in variety 1 of group
A. (See pp. 179, 226, 234, 304.)

5. The short immature periods after these variety 15 crosses ap-
pear to be due to macronuclear regeneration. (See pp. 191, 207.)

6. Newly found strains of variety 13 contain caryonides some
of which stay pure for mating type XXV during the period be-
tween fertilizations. Autogamy has now been found in variety 13.
(See pp. 188, 193, 198, 214, 258, and Table III.)

REFERENCES

Austin, M. L. 1951. Sensitivity to paramecin in Paramecium aurelia in
relation to stock, serotype, and mating type. Physiol. Zool, 24, 196.

Austin, M. L., D. Widmayer, and L. M. Walker. 1956. Antigenic
transformation as adaptive response of Paramecium aurelia to
patulin; relation to cell division. Physiol. Zool., 29, 261.

Beale, G. H. 1952. Antigen variation in Paramecium aurelia, variety 1.
Genetics, 37, 62.

Beale, G. H. 1954. The Genetics of Paramecium aurelia. Cambridge
University Press, Cambridge, England.

Beale, G. H., and M. Schneller. 1954. A ninth variety of Paramecium
aurelia. J. Gen. Microbiol., 11, 57.

316 PROTOZOA

Butzel, H. M., Jr. 1953. Genetic Studies on Paramecium aurclia. Dis-
sertation, Indiana University, Bloomington, Indiana.

Butzel, H. M., Jr. 1955. Mating type mutations in variety 1 of Para-
mecium aurclia, and their bearing upon the problem of mating type
determination. Genetics, 40, 321.

Calkins, G. N. 1919. Uroleptus mobilis Engelm. II. Renewal of vital-
ity through conjugation. /. Exptl. ZooJ., 29, 121.

Calkins, G. N. 1920. III. A study in vitality. /. Exptl. Zool, 31, 287.

Chen, T. T. 1940. Polyploidy and its origin in Paramecium. /. Heredity,
31, 175.

Chen, T. T. 1946a. Conjugation in Paramecium bursaria. II. Nuclear
phenomena in lethal conjugation between varieties. /. Morpkol.,
79, 125.

Chen, T. T. 1946b. Varieties and mating types in Paramecium bur-
saria. I. New variety and types, from England, Ireland, and Czech-
oslovakia. Proc. Natl. Acad. Sci. U.S., 32, 173.

Chen, T. T. 1956. Varieties and mating types in Paramecium bursaria.
II. Variety and mating types found in China. /. Exptl. Zool, 132,
255.

Chen, Y. T. 1944. Mating types in Paramecium caudatum. Am. Nat-
uralist, 78, 334.

Cleveland, L. R. 1950a. Hormone-induced sexual cycles of flagellates.

II. Gametogenesis, fertilization, and one-division meiosis in Oxy-
monas. J. Morphol, 86, 185.

Cleveland, L. R. 19501). Hormone-induced sexual cycles of flagellates.

III. Gametogenesis, fertilization, and one-division meiosis in Sac-
cinobaculus. J. Morphol., 86, 215.

Cleveland, L. l\. 1950c. Hormone-induced sexual cycles of flagellates.

IV. Meiosis after syngamy and before nuclear fusion in Notila.
J. Morphol, 87, 317.'

Cleveland, L. R. 1950d. Hormone-induced sexual cycles of flagellates.

V. Fertilization in Eucomonympha. ./. Morphol., 87, 349.
Cleveland, L. R. 1951a. Hormone-induced sexual cycles of flagellates.

VI. Gametogenesis, fertilization, meiosis, oocysts, and gametocysts in
Leptospironympha. ./. Morphol, 88, L99.

Cleveland. L. R. 1951b. Hormone-induced sexual cycles of flagellates.
Nil. One-division meiosis and autogamy without cell division in
( 'rinympha. J. Morphol, 88. 385.

Cleveland, L. R. 1952. Hormone-induced sexual cycles of flagellates.
\ III. Meiosis in Rhynchonympha in one1 cytoplasmic and two nu-
clear divisions followed by autogamy. /. Morphol, 91, 269.

T. M. SONNEBORN 317

Cleveland, L. R. 1953. Hormone-induced sexual cycles of flagellates.

IX. Haploid gametogenesis and fertilization in Barbulanyrnpha. }.
Morphol, 93, 371.

Cleveland, L. R. 1954. Hormone -induced sexual cycles of flagellates.

X. Autogamy and endomitosis in Barbulanyrnpha resulting from
interruption of haploid gametogenesis. /. Morphol., 95, 189.

Cleveland, L. R., S. R. Hall, E. P. Sanders, and J. Collier. 1934. The

wood-feeding roach Cryptocercus, its Protozoa, and the symbiosis

between Protozoa and Roach. Mem. Am. Acad. Arts Sci., 17, 185.
Corliss, J. O. 1952. Le cycle autogamique de Tetrahymena rostrata.

Compt. rend., 235, 399.
Corliss, J. O. 1954. The literature on Tetrahymena: Its history, growth,

and recent trends. /. Protozool., 1, 156.
De Garis, C. F. 1935a. Lethal effects of conjugation between Parame-
cium aurelia and double monsters of Paramecium caudatum. Am.

Naturalist, 69, 87.
De Garis, C. F. 1935b. Heritable effects of conjugation between free

individuals and double monsters in diverse races of Paramecium

caudatum. J. Exptl. Zoo/., 71, 209.
Diller, W. F. 1948. Nuclear behavior of Paramecium trichium during

conjugation. J. Morphol., 82, 1.
Dippell, R. V. 1954. A preliminary report on the chromosomal constitu-
tion of certain variety 4 races of Paramecium aurelia. Caryologia,

suppl. 1954, 1109.
Dippell, R. V. 1955. Some cytological aspects of aging in variety 4 of

Paramecium aurelia. J. Protozool, 2 (suppl.), 7.
Dobzhansky, T. 1937. What is a species? Scientia, 61, 280.
Dobzhansky, T. 1941. Genetics and the Origin of Species. Columbia

University Press, New York, N.Y.
Downs, L. E. 1952. Mating types in Stylonychia putrina. Proc. Soc.

Exptl Biol Med., 81, 605.
Downs, L. E. 1956. Mating types and varieties in Stylonychia putrina.

Proc. Soc. Exptl Biol. Med., 93, 586.
Ehret, C. F. 1953. An analysis of the role of electromagnetic radiations

in the mating reaction of Paramecium bursaria. Physiol. Zool, 26,

274.
Elliott, A. M., and D. F. Gruchy. 1952. The occurrence of mating types

in Tetrahymena. Biol. Bull, 103, 301.
Elliott, A. M., and R. E. Hayes. 1953. Mating types in Tetrahymena.

Biol. Bull, 105, 269.
Elliott, A. M., and R. E. Hayes. 1955. Tetrahymena from Mexico, Pan-

318 PROTOZOA

ama, and Colombia, with special reference to sexuality. /. Protozool,
2,75.

Elliott, A. M., and D. L. Nanney. 1952. Conjugation in Tetrahymena.
Science, 116, 33.

Erdmann, R. 1927. Endomixis bei Paramecium bursaria. Sitzber. Ges.
naturforsch. Freunde Berlin (1925), 24.

Erdmann, R., and L. L. Woodruff. 1916. The periodic reorganization
process in Paramecium caudatum. ]. Exptl. Zoo/., 20, 59.

Finger, I. 1956. Immobilizing and precipitating antigens of Parame-
cium. Biol. Bull, 111, 358.

Finger, I. 1957a. Immunological studies on immobilizing antigens oi
Paramecium aurclia. J. Gen. Microbiol, (in press).

Finger, I. 1957b. Inheritance of immobilization antigens of Parame-
cium aurclia variety 2. /. Gen. Microbiol, 16, 350.

Giese, A. C. 1941. Mating types in Paramecium multimicronucleatum.
Anat. Record, 81 (suppl.), 131.

Giese, A. C. 1957. Mating types in Paramecium multimicronucleatum.
J. Protozool, 4, 120.

Giese, A. C, and M. A. Arkoosh. 1939. Tests for sexual differentiation
in Paramecium multimicronucleatum and Paramecium caudatum.
Physiol. Zool, 12, 70.

Gilman, L. C. 1939. Mating types in Paramecium caudatum. Am. Nat-
uralist, 73, 445.

Gilman, L. C. 1941. Mating types in diverse races of Paramecium
caudatum. Biol Bull, 80, 384.

Gilman, L. C. 1949. Intcrvarietal mating reactions in Paramecium
caudatum. Biol. Bull, 97, 239.

Gilman, L. C. 1950. The position of Japanese varieties of Paramecium
caudatum with respect to American varieties. Biol. Bidl, 99, 348.

Gilman, L. C. 1954. Occurrence and distribution of mating type vari-
eties in Paramecium caudatum. J. Protozool, 1 (suppl.), 6.

Gilman, L. C. 1956a. Distribution of the varieties of Paramecium cau-
datum. J. Protozool, 3 (suppl.), 4.

Gilman, L. C. 19561). Size differences among twelve varieties ol
Paramecium caudatum. ]. Protozool, 3 (suppl.), 4.

Gruchy, D. F. 1955. The breeding system and distribution of Tetra-
hymena pyriformis. ]. Protozool, 2, 178.

Harrison, J. A., and E. H. Fowler. 1946. A serologic study of conjuga-
tion in Paramecium bursaria. J. Exptl. Zool, 101, 425.

Uiwatashi, K. 1949. Studies on the conjugation of Paramecium can-

T. M. SONNE BORN 319

datum. I. Mating types and groups in the races obtained in Japan.
Science Repts. Tdhoku Imp. Univ., Biol., 18, 137.

Hiwatashi, K. 1951. Studies on the conjugation of Paramecium can-
datum. IV. Conjugating behavior of individuals of two mating types
marked by a vital staining method. Science Repts. Tdhoku Imp.
Univ., Biol, 19, 95.

Hoare, C. A. 1952. The taxonomic status of biological races in parasitic
Protozoa. Proc. Linnean Soc. London, 163, 44.

Inoki, S., T. Kitaura, Y. Kurogochi, H. Osaki, and T. Nakabayasi.
1952a. Genetical studies on the antigenic variation in Trypanosoma
gambiense. Japan. J. Genetics, 27, 85.

Inoki, S., T. Kitaura, T. Nakabayasi, and H. Kurogochi. 1952b. Studies
on the immunological variations in Trypanosoma gambiense. 1. A
new variation system and a new experimental method. Med. J.
Osaka Univ., 3, 357.

Jennings, H. S. 1910. What conditions induce conjugation in Parame-
cium? J. Exptl. ZooL, 9, 279.

Jennings, H. S. 1911. Assortative mating, variability and inheritance of
size in the conjugation of Paramecium, J. Exptl. ZooL, 11, 1.

Jennings, H. S. 1916. Heredity, variation and the results of selection
in the uniparental reproduction of Diffliigia corona. Genetics, 1, 407.

Jennings, H. S. 1938. Sex reaction types and their interrelations in
Paramecium bursaria. I and II. Proc. Natl. Acad. Sci. U.S., 24, 112.

Jennings, H. S. 1939. Genetics of Paramecium bursaria. I. Mating types
and groups, their interrelations and distribution: mating behavior
and self sterility. Genetics, 24, 202.

Jennings, H. S. 1941. Genetics of Paramecium bursaria. II. Self-differ-
entiation and self-fertilization of clones. Proc. Am. Phil. Soc, 85, 25.

Jennings, H. S. 1942. Genetics of Paramecium bursaria. III. Inherit-
ance of mating type, in crosses and clonal self fertilizations. Ge-
netics, 27, 193.

Jennings, H. S. 1944a. Paramecium bursaria: Life history. I. Imma-
turity, maturity and age. Biol. Bull, 86, 131.

Jennings, H. S. 1944b. Paramecium bursaria: Life history. II. Age and
death of clones in relation to the results of conjugation. /. Exptl
Zool, 96, 17.

Jennings, H. S. 1944c. Paramecium bursaria: Life history. III. Re-
peated conjugations in the same stock at different ages, with and
without inbreeding, in relation to mortality at conjugation. /. Exptl.
Zool, 96, 243.

320 PROTOZOA

Jennings, H. S. 1944d. Paramecium bursaria: Life history. IV. Rela-
tion of inbreeding to mortality of exconjugant clones. /. Exptl Zool,
97, 165.

Jennings, H. S. 1945. Paramecium bursaria: Life history. V. Some
relations of external conditions, past or present, to ageing and to
mortality of exconjngants, with summary of conclusions on age and
death. /. Exptl Zool, 99, 15.

Jennings, H. S., and P. Opitz. 1944. Genetics of Paramecium bursaria.
IV. A fourth variety from Russia. Lethal crosses with an American
variety. Genetics, 29, 576.

Johnson, M. 1955. Viability of Inbred and Outbred Paramecium cau-
datum. Thesis, University of Durham, England.

Johnson, W. H. 1952. Further studies on the sterile culture of Parame-
cium. Physiol Zool, 25, 10.

Jollos, V. 1916. Die Fortpflanzung der Infusorien und die potentielle
Unsterblichkeit der Einzelligen. Biol Zentr., 36, 497.

Katashima, R. 1952. Studies on Euplotes. I. Conjugation and cytogamy
induced by split pair method in Euplotes harpa. J. Sci. Hiroshima
Univ. Ser. B, Div. 1, 13, 111.

Kay, M. W. 1946. Studies on Oxytricha bifaria. III. Conjugation. Trans.
Am. Microscop. Soc, 65, 131.

Kimball, R. F. 1939. Mating types in Euplotes. Am. Naturalist, 73, 451.

Kimball, R. F. 1942. The nature and inheritance of mating types in
Euplotes patella. Genetics, 27, 269.

Kimball, R. F. 1943. Mating types in the ciliate Protozoa. Quart. Rev.
Biol, 18, 30.

Kirby, H. 1949. Systematic differentiation and evolution of Flagellates
in Termites. Rev. soc. mex. hist, nat., 10, 57.

Lederberg, J. 1955. Genetics and microbiology. In Perspectives and
Horizons in Microbiology, S. A. Waksman, editor. Rutgers Univer-
sity Press, New Brunswick, New Jersey, p. 24.

Lee, II. 1949. Change of mating type in Paramecium bursaria follow-
ing exposure to X-rays. /. Exptl. Zool, 112, 125.

Levine, M. 1953. The interaction of nucleus and cytoplasm in the iso-
lation and evolution of species of Paramecium. Evolution, 7, 366.

Margolin, P. 1956a. An exception to mutual exclusion of the ciliary
antigens in Paramecium aurclia. Genetics, 41, 685.

Margolin, P. 1956b. The ciliary antigens of stock 172, Paramecium
aurclia, variety 4. /. Exptl. Zool, 133, 345.

T. M. SONNEBORN 321

Maupas, E. 1888. Recherches experimentales sur la multiplication des
infusoires cilies. Arch. zool. exptl. gen., 6, 165.

Maupas, E. 1889. La rajeunissement karyogamique chez les cilies.
Arch. zool. exptl. gen. 7, 149.

Mayr, E. 1948. The bearing of the new systematica on genetical prob-
lems. The nature of species. Advances in Genetics, 2, 205.

Melechen, N. 1955. Antigenic Stability and Variability in Paramecium
aurelia, Variety 3. Thesis, University of Pennsylvania, Philadelphia,
Pa.

Melvin, J. B. 1949. Hybrid Mortality in Interparietal Crosses in Para-
mecium aurelia. Thesis, Wellesley College, Wellesley, Massachu-
setts.

Metz, C. B. 1954. Mating substances and the physiology of fertilization
in Ciliates. In Sex in Microorganisms, American Association for the
Advancement of Science, Washington, D.C., p. 284.

Muller, W. 1932. Cytologische und vergleichend-physiologische Unter-
suchungen iiber Paramecium multimicronucleatum und Paramecium
caudatum, zugleich ein Versuch zur Kreuzung beider Arten. Arch.
Protistenk., 78, 361.

Nanney, D. L. 1953. Nucleo-cytoplasmic interaction during conjuga-
tion in Tetrahymcna. Biol. Bull, 105, 133.

Nanney, D. L. 1954. Mating type determination in Paramecium aure-
lia. A study in cellular heredity. In Sex in Microorganisms, Ameri-
can Association for the Advancement of Science, Washington, D.C.,
p. 266.

Nanney, D. L. 1956a. Inbreeding deterioration in Tetrahymcna piri-
formis. Genetics, 41, 655.

Nanney, D. L. 1956b. Caryonidal inheritance and nuclear differentia-
tion. Am. Naturalist, 90, 291.

Nanney, D. L., and P. A. Caughey. 1953. Mating type determination
in Tetrahymcna pyriformis. Proc. Natl. Acad. Sci. U.S., 39, 1057.

Nanney, D. L., and P. A. Caughey. 1955. An unstable nuclear condition
in Tetrahymcna pyriformis. Genetics, 40, 388.

Nanney, D. L., P. A. Caughey, and A. Tefankjian. 1955. The ge-
netic control of mating type potentialities in Tctrahymena pyrifor-
mis. Genetics, 40, 688.

Pierson, B. F. 1943. A comparative morphological study of several
species of Euplotes closely related to Euplotes patella. J. Morphol.,
72, 125.

322 PROTOZOA

Powelson, E. E. 1956. Differences in the silver-line system and various

measurements in individuals, stocks and varieties of Paramecium

aurelia. J. Protozoal., 3 (suppl.), 9.
Powers, E. L., Jr. 1943. The mating types of double animals in Eu-

plotes patella. Am. Midland Naturalist, 30, 175.
Preer, J. R., Jr. 1950. The role of genes, cytoplasm and environment

in the determination of resistance and sensitivity. Year Book Am.

Phil. Soc., p. 161.
Pringle, C. R. 1955. Thesis, Univ. Edinburgh, Edinburgh, Scotland.
Pringle, C. R. 1956. Antigenic variation in Paramecium aurelia, variety

9. Z. Abst. Vererbsl, 87, 421.
Ray, C, Jr. 1954. Chromosome behavior during conjugation of mating

types I and II of variety 1 of Tetrahymena. Biol. Bull, 107, 318.
Ray, C, Jr. 1955. Irregularities during meiosis in variety 9 of Tetra-
hymena pyriformis. Biol. Bull., 109, 367.
Ray, C, Jr. 1956. Meiosis and nuclear behavior in Tetrahymena pyri-
formis. J. Protozool., 3, 88.
Ray, C, Jr., and A. M. Elliott. 1954. Chromosome number of four

varieties of Tetrahymena. Anat. Record, 120, 812.
Ray, C, Jr., A. M. Elliott, and G. M. Clark. 1955. The immaturity

period in three varieties of Tetrahymena pyriformis. Biol. Bull., 109,

367.
Siegel, R. W. 1954. Mate-killing in Paramecium aurelia, variety 8.

Physiol. Zoo/., 27, 89.
Siegel, R. W. 1956. Mating types in Oxytricha and the significance of

mating systems in Ciliates. Biol. Bull., 110, 352.
Skaar, P. D. 1956. Past history and pattern of serotype transformation

in Paramecium aurelia. Exptl. Cell Research, 10, 646.
Sonneborn, T. M. 1937. Sex, sex inheritance and sex determination in

Paramecium aurelia. Proc. Natl. Acad. Sci. U.S., 23, 378.
Sonneborn, T. M. 1938. Mating types in Paramecium aurelia: Diverse

conditions for mating in different stocks; occurrence, number and

interrelations of the types. Proc. Am. Phil. Soc, 79, 411.
Sonneborn, T. M. 1939. Paramecium aurelia: Mating types and groups;

lethal interactions; determination and inheritance. Am. Naturalist,

73, 390.
Sonneborn, T. M. 1941. The effect of temperature on mating reactiv-
ity in Paramecium aurelia, variety 1. Anat. Record, 81 (suppl.), 131.
Sonneborn, T. M. 1942a. Inheritance in Ciliate Protozoa. Am. Nat-
uralist, 76, 46.

T. M. SONNEBORN 323

Sonneborn, T. M. 1942b. More mating types and varieties in Parame-
cium aurelia. Anat. Record, 84, 92.

Sonneborn, T. M. 1943. Gene and cytoplasm. Proc. Natl. Acad. Sci.
U.S., 29, 329.

Sonneborn, T. M. 1947. Recent advances in the genetics of Parame-
cium and Euplotes. Advances in Genetics, 1, 263.

Sonneborn, T. M. 1950. Methods in the general biology and genetics
of Paramecium aurelia. J. Exptl Zool, 113, 87.

Sonneborn, T. M. 1954a. The relation of autogamy to senescence and
rejuvenescence in Paramecium aurelia. }. Protozool., 1, 38.

Sonneborn, T. M. 1954b. Patterns of nucleocytoplasmic integration in
Paramecium. Caryologia (suppl. 1), 307.

Sonneborn, T. M. 1955. Heredity, development and evolution in Para-
mecium. Nature, 175, 1100.

Sonneborn, T. M. 1956a. New varieties and mating types of Parame-
cium aurelia. Anat. Record, 125, 566.

Sonneborn, T. M. 1956b. Distribution of the varieties of Paramecium
aurelia. Anat. Record, 125, 567.

Sonneborn, T. M. 1956c. The distribution of killers among the varieties
of Paramecium aurelia. Anat. Record, 125, 567.

Sonneborn, T. M., and R. V. Dippell. 1943. Sexual isolation, mating
types, and sexual responses to diverse conditions in variety 4, Para-
mecium aurelia. Biol. Bull., 85, 36.

Sonneborn, T. M., and R. V. Dippell. 1946. The significance of race
31 as a link between group A and B varieties of P. aurelia. Anat.
Record, 96, 19.

Sonneborn, T. M., and R. V. Dippell. 1956. Giant Paramecium aurelia
(?)./. Protozool, 3 (suppl.), 9.

Sonneborn, T. M., and M. V. Schneller. 1955. Genetic consequences
of aging in variety 4 of Paramecium aurelia. Genetics, 40, 596.

Stebbins, G. L., Jr. 1950. Variation and Evolution in Plants. Columbia
University Press, New York, N.Y.

Thoday, J. M. 1953. Components of fitness. Symposia Soc. Exptl. Biol,
7,96.

Vivier, E. 1955. Contribution a letude de la conjugaison chez Para-
mecium caudatum. Bull soc. zool France, 80, 163.

Wenrich, D. H. 1954. Sex in Protozoa: a comparative review. In Sex
in Microorganisms, American Association for the Advancement of
Science, Washington, D.C., p. 134.

Wichterman, R. 1937. Conjugation in Paramecium trichium Stokes

324 PROTOZOA

(Protozoa, Ciliata) with special reference to nuclear phenomena.

Biol Bull, 73, 397.
Wichterman, R. 1948. The time schedule of mating and nuclear events

in the conjugation of Paramecium bursaria. Titrtox News, 26.
Wichterman, R. 1950. The occurrence of a new variety containing

two opposite mating types of Paramecium calkinsi as found in sea

water of high salinity content. Biol Bull, 99, 366.
Wichterman, R. 1951. The ecology, cultivation, structural character-
istics and mating types of Paramecium calkinsi. Proc. Penn. Acad.

Sci,, 25, 51.
Wichterman, R. 1953. The Biology of Paramecium. McGraw-Hill-

Blakiston Co., New York, N.Y.
Wood, H. K. 1953. Some Factors Affecting Delayed Separation of

Conjugants in Paramecium aurelia. Thesis, Indiana University,

Bloomington, Indiana.
Woodruff, L. L., and H. Spencer. 1924. Studies on Spathidium spa-

thula. II. The significance of conjugation. /. Exptl Zool, 39, 133.
Wright, S. 1940. Breeding structure of populations in relation to speci-

ation. Am. Naturalist, 74, 232.

AN EMBRYOLOGIST'S VIEW OF
THE SPECIES CONCEPT

JOHN A. MOORE: departments of zoology, barnard

COLLEGE AND COLUMBIA UNIVERSITY, NEW YORK

Nature never made species mutually sterile
by selection, nor will men. — Darwin to T. H.
Huxley, Jan. 7, 1867, in More Letters, vol. 1,
p. 277, Letter 197.

If an embryologist is asked to contribute his views to a sympo-
sium on "The Species Concept" it could be for any of several
reasons, such as his interest in the origin of things, in this case
species, or the possibility that information on gametes and early
embryos can aid our understanding of this taxonomic category.

The data that I will present have a bearing on the origin of
species and on the complex intergroup and intragroup relations
in natural populations. Most of these data have appeared pre-
viously (Moore, 1946, 1947, 1949a,b, 1950, 1954, 1955) but they
have not been used for the specific problem of this symposium.

Current Views on the Origin of Species

The work of systematists in the last part of the nineteenth cen-
tury and in the twentieth century has made a strong case that one
pattern of speciation involves divergence in geographically iso-
lated populations. The assumed events may be outlined in a
simple fashion as follows: A single species becomes distributed
over a large area, which is divided by barriers into a number of
smaller areas. The barriers are such that gene exchange between
the populations of the geographically isolated areas is greatly
reduced or almost absent. With the passage of time each popula-
tion evolves in an essentially different phyletic line. The course
of these separate evolutions is assumed to involve the selection

325

326 AN EMBRYOLOGIST'S VIEW

of genomes that best fit the isolated populations for survival in
their own environments. Eventually some of the isolated popula-
tions may diverge from others to such an extent that one would
call them different species.

Mayr (1948) has expressed these notions as follows:

Wherever extrinsic factors cause a retardation or interruption of
gene flow between portions of a species (geographical or spatial iso-
lation), these subdivisions of the species tend to drift apart geneti-
cally. The rate of this change is different in different species.

If the isolation is sufficiently complete and lasts sufficiently long, it
will permit the evolution of isolating mechanisms, which will inhibit
the interbreeding of the two daughter species after the elimination
of the extrinsic isolating factors.

This is, of course, the now familiar concept of allopatric spe-
ciation. Many have regarded it as an adequate explanation of
the common pattern of evolution in terrestrial vertebrates and
probably in many other organisms as well. It is this position that
I will attempt to defend. On the other hand, Dobzhansky and,
to a lesser extent, a few others have thought it necessary to in-
voke an additional phase in the formation of species from geo-
graphic races. They do not believe that the species level of
divergence can be reached while the population remains sep-
arated and that a culminating sympatric stage is necessary for
the development of full species differences. The problem for
them is whether it is possible to obtain fully effective interpopula-
tion isolating mechanisms as a consequence of evolution in
allopatric groups.

Thus, Dobzhansky (1940) develops his argument after first
asking the question ". . . whether isolating mechanisms develop
as a necessary consequence of the accumulation of genetic differ-
ences in general, or whether they represent a separate category
of genetic changes which appear and become established only
under certain special conditions." He believes that the second
possibility is the correct one, namely, ". . . that the origin of
isolation is a process separate from that of the origin of other
species differences." His reasons for so believing are as follows:

J. A. MOORE 327

This theory starts with the premise that each species, genus and
probably each geographical race is an adaptive complex which fits
into an ecological niche somewhat distinct from those occupied by
other species, genera and races. The adaptive value of such a com-
plex is determined not by a single or a tew genes, but is a property
of the genotype as a whole. Furthermore, the adaptive complex is
attuned to its environment only so long as its historically evolved pat-
tern remains, within limits, intact. It is true that interbreeding of dif-
ferent adaptive complexes may sometimes result in emergence of new
genotypes which fit into unoccupied or sparsely settled ecological
niches — hence the evolutionary role of hybridization. Nevertheless,
hybridization usually leads to the formation of disharmonious recom-
binations.

Considerations such as these have prompted some writers (Dob-
zhansky, 1937a,b; Sturtevant, 1938; ef. Fisher, 1930) to assume that
occurrence of hybridization between races and species constitutes a
challenge to which they may respond by developing or strengthening
isolating mechanisms that would make hybridization difficult or im-
possible. Where hybridization jeopardizes the integrity of two or more
adaptive complexes, genetic factors which would decrease the fre-
quency or prevent the interbreeding would thereby acquire a positive
selective value, even though these factors by themselves might be
neutral. Race formation is essentially the development of genetic pat-
terns which are adapted to a definite environment. Speciation is a
process resulting in fixation of these patterns through the development
of physiological isolating mechanisms. Clearly, raciation and specia-
tion should not be conceived of as entirely independent processes,
but the development of physiological isolating mechanisms must nev-
ertheless be supposed to intervene only after the divergence of the
adaptive complexes had been initiated. If races are to become species,
isolating mechanisms must arise when the distinct adaptive complexes
are exposed to the risk of disintegration due to interbreeding.

Dobzhansky has modified his view to some extent in more
recent publications. The quotation can be taken, however, as an
expression of a theory that is fairly widely subscribed to in

biology.

It is my feeling that in the majority of situations it is possible
for fully diverged species, adequately equipped with isolating

328 AN EMBRYOLOGIST'S VIEW

mechanisms, to evolve in allopatric populations. Furthermore,
I feel that it is unnecessary to assume a culminating sympatric
phase involving selection against hybridization as has been done
by Dobzhansky. In developing my reasons questioning these
views, I shall first cite a sequence of cases that can serve as a
model for the origin of fully effective isolating mechanisms in
allopatric populations.

Stages in the Evolution of Frogs

The processes involved in the evolution of one species into one
or more different species requires a very long time. It has been
suggested that one million years might be taken as a rough aver-
age. A period of this magnitude is equivalent to at least 20,000
lifetimes of work for a student of evolution. Clearly one individual
could not expect to reach any significant conclusions during
his own lifetime by measuring the changes in a single species.
Instead he must study a variety of organisms in different stages
of speciation and then reconstruct the probable course of
events. It must always be remembered that this procedure is
indirect, and that the answer depends to a considerable extent on
the choice of examples.

We will now list some data obtained from several frogs in a
manner suggesting a pattern of evolution.

There are five common species of the genus Rana in the New
York area. All are "good" species and no taxonomist questions
their distinctness. Each one has its own characteristic habitat,
morphological features, behavior pattern, and geographic dis-
tribution. Some years ago, I presented data showing the close
relation between the breeding behavior, geographic distribu-
tion, and temperature adaptations of the embryos of these species
(Moore, 1949b). There is no question of the importance of these
adaptations. Rana sylvatica, which is an early spring breeder,
lias embryos with a cold-adapted physiology that could not allow
survival during the1 summer when R. catesheiana breeds. Neither
could the heat-adapted R. catesheiana embryos survive in the
cold water of ponds in early spring.

In all but one species the temperature adaptations correlated

J. A. MOORE 329

so well with geographic distribution that it seemed clear there
must be a causal connection. The exceptional species was R.
pipiens, which had a geographic distribution much greater than
could be anticipated on the basis of the temperature adaptations
of individuals from the New England area.

Studies of R. pipiens from south of their predicted geographic
boundary showed that these southern forms differed from their
northern counterparts in their temperature adaptations. This
situation is by no means surprising. One would anticipate that
a widely distributed species would be broken into local popula-
tions, each adapted to its immediate environment. R. pipiens
embryos adapted for early spring breeding in the cold-temperate
environment of Quebec and northern New England would not be
expected to survive under the temperature conditions encoun-
tered in the semitropical zones of southern Florida or eastern
Mexico.

For our present purposes it is of considerable interest to note
that developmental (and presumably genetic) incompatibility
is associated with adaptation to different temperatures. If in-
dividuals from similar temperature-adapted populations are
crossed, the embryos are normal in their development. On the
other hand, if individuals from different temperature-adapted
populations are crossed, the embryos are very abnormal and
may die early in development.

The extreme results are obtained in crosses of Vermont x
Florida, Wisconsin x Texas, and Vermont x eastern Mexico in-
dividuals. In the last cross most of the embryos die as abnormal
gastrulae or neurulae. Thus, these southern pipiens and northern
pipiens are behaving as different species so far as their potential
ability to exchange genes is concerned. These are phenomena
associated with the crossing of individuals from widely separated
latitudes. Adjacent populations, however, can be crossed, and
they give entirely normal offspring.

We could consider the populations of R. pipiens to be one
species, two species, or to be one species in a crucial stage in the
process of splitting into different species. I believe the third
alternative is the most meaningful way to consider the data. The

330 AN EMBRYOLOGIST'S VIEW

point to be emphasized, however, is that the northern and south-
ern forms have developed isolating mechanisms, in reference
to each other, to such a degree that they could coexist and re-
main distinct. Since these differences that serve as isolating
mechanisms have developed in widely separated places, it is
not necessary to invoke Dobzhansky's hypothesis of a culminating
sympatric phase of speciation.

What is the basis of the differences between the southern and
northern frogs? Are they the incidental and fortuitous concom-
itants of divergence? I prefer to think not for the following
reasons. The northern and southern embryos differ greatly in
their embryonic temperature adaptations. They differ in their
range of temperature tolerance, temperature coefficient for
development, and rate of development. Clearly they are quite
dissimilar in their physiologies, which presumably are genetically
determined. When we cross individuals with these dissimilar
physiologies, the offspring are highly abnormal. It is my belief
that the developmental abnormalities in the "hybrids" are due to
the inharmonious association of two genetic systems that differ
in their- modes of action. If this is the correct interpretation, the
genetic differences that are the basis of the embryonic adapta-
tions are likewise the basis of an isolating mechanism, namely,
hybrid inviability. Two sets of data indirectly support this possi-
bility:

1. The magnitude of the differences in embryonic temperature
adaptations between any two populations parallels the degree of
abnormality shown by their "hybrids." This suggests that the
same genes might be the basis for both phenomena.

2. The genes that are responsible for the embryonic tempera-
ture adaptations and the genes that interact to produce the
hybrid malformations have their effects at precisely the same
time in the life cycle. This also suggests that the genetic basis
for the embryonic temperature adaptations and hybrid inviabil-
ity is the same.

If this interpretation is correct, the very process of adaptation
to diverse climatic conditions has resulted in genetic differences
that would also be completely effective isolating mechanisms.

J. A. MOORE 331

This brief synopsis of the situation in R. pipiens will serve as
the first stage in our model of geographic speciation. It is a wide-
spread species occupying a diversity of habitats. So far as the data
go, it appears that it is a continuous group of interbreeding popu-
lations. Some of the widely separated populations, however, have
diverged to such an extent that they behave as different genetic
species. This divergence could not have been promoted by selec-
tion against hybrids, according to Dobzhansky's hypothesis, since
there is no possibility of hybrids being formed.

Our question now is this: Where does the species go from
here? Some have regarded divergence in a continuous popula-
tion, such as R. pipiens, as a blind alley in evolution and, unless
some extrinsic factors can divide the species, it can never be
more than a group of interbreeding allopatric populations. I favor
this view, but the evidence for it is so incomplete that there is
little point in one being counted in adherence or in opposition.

There is no difficulty in imagining the formation of two species
in R. pipiens if we invoke some extrinsic factors which would
obliterate the populations in the middle portions of the species'
range or if there arose several well-placed barriers to gene flow.
A variety of physical and biological factors could serve this pur-
pose.

The next stage in our model of speciation will be a species
that at one time undoubtedly had a continuous distribution but
it is now found in two widely separated regions.

Crinia signifera occurs in eastern and in western Australia but
not in a wide band of unsuitable country that extends from the
Great Australian Bight to the northwest coast. There are no
detectable morphological differences between the eastern and
western individuals and no one thought they were other than
the same species until cross fertilization experiments were per-
formed. When one crosses eastern and western individuals, how-
ever, profound hybrid defects and generally so little survival are
observed that it is necessary to consider them separate species.

The fact that the eastern and western forms arc morphologi-
cally identical is best interpreted in terms of their descent from
a common ancestor. During the Pleistocene the climate was

832 AN EMBRYOLOGIST'S VIEW

such that there could have been a continuous population of
these frogs across Australia. There is good evidence for periods
during which most of the continent was moist. These humid
times were followed by dry intervals, some of which were drier
than the present-day climate.

We can imagine, therefore, that at one time in the past the
frogs ancestral to C. signifera occupied a continuous zone across
Australia. Somewhat later the continuous distribution was broken
by the development of arid central regions. The east and west
populations are now different species. There is no way of telling
whether they became different after being isolated or whether,
as in R. pipiens, the differences developed while the range was
continuous.

With either interpretation, or some combination of them, it is
clear that there is no need to invoke the concept of a culminating
sympatric stage in speciation. The differences that now serve
as isolating mechanisms arose in allopatric populations.

Other Arguments against the Hypothesis

We should not conclude from these two examples in frogs that
the hypothesis of the development of isolating mechanisms by
selection, in zones of overlap, against hybridization is incorrect.
For most cases, however, I do think it is unnecessary. The follow-
ing are some of the reasons.

Undoubtedly there have been many instances where a species
population lias become divided and has remained so for long
periods of time. When the land connection between North
America and South America was reestablished in recent geo-
logical times, the continuous ranges of many marine organisms
was broken. At the present we find many pairs of related species,
the two members of each pair being separated by the Central
American isthmus. The most probable interpretation of this situa-
tion is that divergence was a consequence of evolution in isola-
tion. Since the two diverging groups never came together there
would have been no possibility of their developing or reinforcing
isolating mechanisms through the selection against hybridization.

There must be numerous similar instances both for terrestrial

J. A. MOORE 333

and aquatic organisms. The chance dispersal of organisms from
a continent to a remote island, the breaking of connections be-
tween continents or bodies of water, the extermination by bio-
logical or physical causes of the individuals in portions of a
species' range are some situations that can divide an originally
continuous population. In many instances the isolated groups
will evolve with no further relation to each other.

The situations just described exclude the possibility of hybrid-
ization, and we will now consider those instances where once
separated populations become continuous.

When two once continuous and then separate populations re-
establish contact, a variety of consequences can ensue. If the
period of separation has been short, or if genetic divergence
has been slight, the two groups would be expected to become
parts of the same breeding population in the zone of overlap.
If the period of separation has been long and if genetic diver-
gence has been sufficient for effective isolating mechanisms to
develop, the two overlapping populations could maintain their
discreteness.

The two situations just considered are extremes and there is
little doubt about what will occur. It is with the intermediate
situation that we shall now give our attention, for it is here
that there is doubt as to the course of events. The situation will
be this: Two portions of a population become geographically
isolated and during their period of isolation an appreciable
amount of genetic divergence occurs. The amount of divergence
will be defined as sufficient to result in a hybrid less well adapted
than either parent population. It is in this situation that Dob-
zhansky has postulated that natural selection will promote the
development of isolating mechanisms that will decrease the wast-
age of gametes due to their combining in poorly adapted hybrids.

When the two populations reestablish contact, under the con-
ditions stipulated, they come together as differently adapted
groups. In their period of separation each evolved along paths
leading to better adaptation to their separate environments. The
very fact that the genomes are no longer able to collaborate
normally may be taken as indicative of considerable genetic di-

334 AN EMBRYOLOGIST'S VIEW

vergence and probably of considerable adaptational divergence
as well. Tims, in any zone of overlap the two groups may be
assumed to have somewhat different ecological niches. This may
take the form of differences in breeding time, habitat, food
preferences, and so on. The net effect will be to limit the pos-
sibility of forming hybrids.

The previously evolved adaptative differences will tend to
restrict competition between the two populations in the zone of
overlap. Where competition does occur, however, there are some
conditions under which further divergence of the populations
will be promoted. Let us designate the two populations A and B
and assume two habitats alpha and beta. Let us further assume
that A was previously isolated In a zone where alpha was the
commonest habitat and that the course of A's evolution was for
adaptation in alpha although it also occupied beta. We will
assume that B was found in an area where beta was the prevail-
ing habitat and B became better adapted to beta in spite of the
fact that it occupied the available alpha to a certain extent. If
the two populations became sympatric in a region with alpha
and beta, we would anticipate that each species would become
restricted to a single habitat. A being better adapted to alpha
would displace B from this habitat. Similarly, we would antici-
pate that B would displace A from beta. In the zone of overlap,
any genetic difference would have a selective advantage if it
tended to restrict each species to the habitat to which it was
better adapted. Thus, an isolating mechanism (habitat prefer-
ence) would be enhanced, but this would be a consequence of
natural selection reducing competition and not of natural selec-
tion reducing hybridization.

In the situation described, there seems no reason to believe
that the gene differences that promote habitat isolation in the
/one of overlap will spread throughout the population. The gene
differences will tend to be restricted to the /one where they
have an adaptive signifieance.

Much the same line of reasoning would lead us to question
the ability of selection against hybridization to produce isolating
mechanisms that characterize the entire population. Let us again

J. A. MOORE 335

postulate two populations, C and D, which were isolated for some
time during which they diverged genetically to such an extent
that any hybrids they might form would show reduced viability.
Let us further assume that C and D reestablish contact and in
the zone of overlap they hybridize. Since we have postulated
that the hybrids have reduced viability, let us assume with Dob-
zhansky, that natural selection would promote the development
of isolating mechanisms. These isolating mechanisms would have
selective advantage in the zone of overlap hut not elsewhere.
Consequently they could not become characteristic of the species
as a whole, for the following reason: If we assume with Dob-
zhansky that "Isolating mechanisms encountered in nature appear
to be ad hoc contrivances which prevent the exchange of genes
between nascent species," and that they are formed by selection
in zones of overlap, we can assume that they have adaptive value
only in the zone of overlap. If they were of adaptive value to
the two species in areas where they were not sympatric, they
would be selected but not for their function as isolating mech-
anisms. Genetic differences that are ad hoc isolating mechanisms
between two species can be selected for only in areas where the
two species occur sympatrically. These isolating mechanisms can
develop as a consequence of selection for genes that prevent
the formation of hybrids or by selection for genes that prevent
competition ( irrespective of whether or not hybrids are formed ) .

Outside of the zone of overlap any gene differences that serve
as ad hoc isolating mechanisms would probably be either neutral
or somewhat deleterious. (If they were advantageous they would
be selected for any way, but with no reference to any pleiotropic
functions as isolating mechanisms.) It is difficult to understand,
therefore, how they could become established as isolating mechan-
isms throughout the ranges of both species unless the ranges were
coextensive. Except on small islands, this last possibility would
be highly unlikely.

Several additional reasons can be cited for believing that the
origin of isolating mechanisms by selecting against hybrid forma-
tion cannot be of universal occurrence.

First, there are some species pairs that are able to form appar-

.336 AN EMBRYOLOGIST'S VIEW

ently normal hybrids, and yet they remain distinct even when
sympatric. In this situation there would be no adaptive advantage
to restricting hybridization, and hence the isolating mechanisms
that do prevent gene interchange could not have been perfected
by the natural selection of genes that prevent hybridization.

Second, in areas of overlap where inferior hybrids are formed,
we must remember that hybridization must be frequent if natural
selection is to increase the frequencies of gene differences that
reduce crossing. Consequently, as isolating mechanisms become
more effective, the ability of selection to augment them will
steadilv decrease.

Third, we must not assume that the wastage of gametes is
always a disadvantage to the species. There are some special
situations where the wastage of gametes might be a real advan-
tage to the population as a whole. As a hypothetical case let us
consider a species whose numbers are held in check by predators
and available food. If the source of predation is then removed,
there will be greater competition for the available food. In some
species severe competition may result in 100 per cent mortality.
In this, an unusual situation, some wastage of gametes would
be an advantage.

There are some observations that have been cited in support
of Dobzhansky's thesis, such as those of Koopman ( 1950 ) . He
kept Drosophila pseudoobscura and pcrsimilis together in a
population cage and found that it was possible to augment the
existing isolating mechanisms by removing all the hybrids that
formed. In other words he was able to select for mating prefer-
ence. This is not an unexpected result. It is highly probable that
all genetic differences that serve as isolating mechanisms are con-
sequences of natural selection (though in most instances I feel
they are not selected for as isolating mechanisms per se). The
I act that one can select against hybrid formation does not prove
that an analogous situation will occur in nature. It is possible to
select in artificial situations for all sorts of things that are not
selected for under natural conditions.

There are now a number of reports of species pairs which
overlap in their ranges and show greater genetic isolation from

J. A. MOORE 337

one another where they are sympatric than elsewhere. In no ease
that I know of is there good evidence to indicate that selection
against the formation of ill-adapted hybrids is involved. The
greater differences between the species in the area of overlap
might be due to greater genetic divergence resulting from com-
petition (as has been discussed before) or to still other factors.
There are also some reports of overlapping and closely related
species showing less genetic isolation where they are sympatric
than elsewhere. Volpe (1955) has described a case involving
Bufo americanus and B. fowled, and he refers to similar cases in
other organisms.

Summary

The pattern of evolution that involves the formation of new
species from geographical races is discussed. The conclusion is
reached that the most important aspect of this process is genetic
divergence associated with adaptation to the local environment.
It is felt that this divergence will involve differences that will
also serve as isolating mechanisms. There is no critical evidence
that isolating mechanisms develop as ad hoc contrivances that
prevent hybridization between incipient species. Furthermore,
there are some theoretical considerations that make it improb-
able that the genetic differences that serve as isolating mecha-
nisms commonly arise in this manner.

REFERENCES

Dobzhansky, T. 1940. Speciation as a stage in evolutional) divergence.
Am. Naturalist, 74, 312-21.

Koopman, K. F. 1950. Natural selection for reproductive isolation be-
tween Drosophila pseudoobscura and Drosophila persimilis. Evolu-
tion, 4, 135-48.

Mayr, E. 1948. The bearing of the new systematies on genetical prob-
lems. The nature of species. Advances in Genetics, 2, 205-37.

Moore, J. A. 1946. Incipient intraspecific isolating mechanisms in Rana
pipiens. Genetics, 31, 304-26.

338 AN EMBRYOLOGIST'S VIEW

Moore, J. A. 1947. Hybridization between Rana pipiens from Vermont
and eastern Mexico. Proc. Natl. Acad. Sci. U.S., 33, 72-75.

Moore, J. A. 1949a. Geographic variation of adaptive characters in
Rana pipiens Schreber. Evolution, 3, 1-24.

Moore, J. A. 1949b. Patterns of evolution in the genus Rana. In Ge-
netics, Paleontology, and Evolution. G. L. Jespen, G. G. Simpson,
and E. Mayr, editors. Princeton University Press, Princeton, New
Jersey.

Moore, J. A. 1950. Further studies on Rana pipiens racial hybrids.
Am. Naturalist, 84, 247-54.

Moore, J. A. 1954. Geographic and genetic isolation in Australian am-
phibia. Am. Naturalist, 88, 65-74.

Moore, J. A. 1955. Abnormal combinations of nuclear and cytoplasmic
systems in frogs and toads. Advances in Genetics, 7, 139-82.

Volpe, G. P. 1955. Intensity of reproductive isolation between sym-
patic and allopatric populations of Rufo americanus and Rufo
foivleri. Am. Naturalist, 89, 303-18.

THE SPECIES PROBLEM FROM THE
VIEWPOINT OF A PHYSIOLOGIST

C. LADD PROSSER: physiology department, university

OF ILLINOIS, URBANA, ILLINOIS

The principal task of physiologists is to describe life functions
as they occur in well-known organisms; physiologists are less
concerned with how organisms came to be than with how they
now are. Comparative physiologists attempt to ascertain the
ways in which different organisms solve their life problems. By
using kind of organism as one experimental variable when study-
ing a function, the comparative physiologist can arrive at different
kinds of generalizations from those obtained by the more in-
tensive study of single organisms. As part of the effort to describe
life processes, some physiologists investigate adaptations to en-
vironmental stresses; these adaptations constitute the functional
bases for animal distribution, and consideration of them leads
inevitably to the species problem. Success of a species at its en-
vironmental limits depends on adaptive capacity in structure and
function, but the morphology which is related to such essential
function is often at quite a different level from that used by the
systematist; it may be molecular morphology. At the same time,
the morphological characters of the systematist's key, whether
adaptive or not, are generally assumed to be associated geneti-
cally with physiological characters; the genetic basis for adaptive
physiological characters is usually multifactor and complex.

Most of the physiologists and biochemists who have considered
phylogeny have compared higher taxonomic units — phyla and
classes — and they have generally supported the conclusions
reached previously by comparative embryologists and classical
phylogenists. Relatively few physiological comparisons have
been made of closely related species or of populations living at

339

THE VIEWPOINT OF A PHYSIOLOGIST

the opposite range limits of one species. Most physiologists seek
homogeneous material, whereas in population studies, controlled
heterogeneity is desired. When differences are observed in similar
animals, the differences may be among individuals within a ge-
netically similar population, they may be in mutants carried at
some level in the population, they may be racial between eco-
types of the same species, or they may be between the total ge-
netic systems of true species. Our operative assumption has been
that speciation progresses from individual variation through
strains and races to species. Since the complete sequence cannot
be observed in any single species but is inferred, it is necessary
to piece together as much indirect evidence as possible to test
the basic assumption.

Physiological Criteria of Variation

Before examining hypotheses and experimental data it is im-
portant to consider what physiological criteria can best be used
in laboratory comparisons of animals drawn from different popu-
lations. We may list seven general categories of criteria of physio-
logical variation.

Survival. Tolerance of sudden environmental extremes is the
easiest and most widely used criterion of physiological variation.
Percentage survival at different times after application of a stress
is plotted for different intensities of stress, and LD.-,<> (median
lethal dose) values for the stress are obtained. More important
is the establishment of the maximum stress tolerated indefinitely.
The causes of death within short times after exposure to extreme
stresses are usually different from those after longer times, and
it is important that prior history ol the animals be known because
acclimation may strongly influence survival. For example, Fry,
Brett, and Clawson (1942) and Brett (1944) constructed useful
polygons that gave the upper and lower limits of temperature
tolerated by fish which had been differently acclimated, and they
found in goldfish a 1 C change in lethal temperatures for each
3°C. acclimation difference. When a stress is applied gradually,
the acclimation process increases the degree of stress which
can be tolerated, i.e., the rate of application of the' stress affects

C. L. PROSSER 341

the tolerated limits. Most survival tests are made with adults
rather than with young, a procedure which is easier but perhaps
less meaningful for speciation, since embryos are frequently more
sensitive to stresses. Despite the fact that survival data must be
qualified according to acclimation history, time of observation,
and rate of application of the stress, the extremes to which ani-
mals are subjected in nature are more important than are average
environmental values.

Reproduction. Ability to reproduce measures survival over

Fig. 1. Diagrams representing relation between an internal state I and
an environmental stress O. ax and a2 represent environmetal limits of two
acclimation levels corresponding to survival limits sx and s2 of internal
variation. Broken lines represent regions where only brief survival is pos-
sible. (A) Diagram for a conforming (adjusting) animal. (B) Diagram for
a regulating animal. From Prosser (1955).

a complete life cycle and approaches the natural situation better
than survival of adults only. Unfortunately, rearing animals in
the laboratory is possible for relatively few species, and the best
comparisons are of reproduction in populations occurring in
different ecological situations and acclimated by transplanting to
different field conditions. More significant than ability to repro-
duce is rate of growth of populations, and this can be measured
readily in the laboratory only for a few species of small animals,
particularly protozoans and insects.

Measurement of an Internal State as a Function of External
Stress. There are two basic patterns of physiological adaptation

342 THE VIEWPOINT OF A PHYSIOLOGIST

to a given environmental change (Prosser, 1955): (a) an animal
may alter itself so that it conforms internally with the environ-
ment (adjusts) (Fig. 1A), or (b) it may regulate its internal
state, and thus maintain relative constancy in an altered environ-
ment (Fig. IB). Familiar examples of these two patterns are
found in poikilotherms, in which the internal temperature
changes in conformity with the environment, and homeotherms,
which maintain relatively constant body temperature. The toler-
ated range of internal variation of a given character is greater
in a conforming than in a regulating animal (Fig. 1A, B). Accli-
mation can shift the range of internal variation in a conformer,
and it can shift the limits of regulation in a regulator; both pat-
terns of response are adaptive in that they permit tolerance of
an altered environment. The genotype fixes the point beyond
which acclimation is no longer effective. Conformity to the
environment and regulation can be analyzed with respect to
temperature, osmotic concentration, ionic balance, oxygen partial
pressure, and other aspects of the internal state. Some animals
may conform on one side of the tolerated range and regulate on
the other side; for example, some shore crabs are weak osmoregu-
lators in a dilute medium and osmoconformers in more concen-
trated media. Different kinds of animals and different populations
of the same species may profitably be compared with respect to
the pattern of response of internal state.

Mechanisms and Limits of Regulation. In a regulator any
environmental change starts feedback mechanisms which tend
to compensate for the change. At some limits, the feedback mech-
anisms fail and survival is no longer possible. The mechanisms
at the two limits of stress are not necessarily related and may
call upon different organ systems. More physiological studies
arc concerned with mechanisms of regulation than with any other
aspect of the interaction between organism and environment,
probably because man is such a good regulator. Yet there are
few systematic comparisons of regulating mechanisms in different
races and species.

Recovery from a Deviated Stale. Every animal, whether it is
a conformer (adjuster) or regulator, can withstand some brief

C. L. PROSSER 343

deviation from its "normal" internal state, and under natural
conditions every animal operates close to some "optimum." For
example, a homeotherm can be given either a fever or hypo-
thermia and then, when the stress is removed, it returns to its
"normal" temperature. An animal can have its normal water
load (water content) altered by dehydration or by excessive
hydration after which in its usual environment it will gain or
lose water until it arrives at its "normal" water load. One of the
most obvious and yet least understood generalizations in physiol-
ogy is that life processes go best within limited ranges of osmo-
concentration, at certain oxygen pressures, within narrow limits
of temperature and of concentration of specific ions; the condi-
tions which are most favorable differ for different animals, but
in a given animal there is always a tendency to return to the
"optimum" state after deviation. Numerous patterns of return
exist, some with an overshoot (Adolph, 1943), and comparison
of recovery patterns is a useful way to differentiate different kinds
of animals. Most of the comparisons thus far made have been
between distantlv related forms.

Rate Functions. Probably the most subtle and quantitative
method of differentiating physiological types is to compare reac-
tion rates within narrow and natural environmental ranges. Rates
of oxygen consumption, rates of heartbeat and breathing, rates of
growth, are affected by many factors, some more critical than
others. The most familiar use of rate functions is in respect to
temperature effects, and numerous natural populations have been
compared with respect to temperature characteristics (Qw, the
rate increase calculated per 10 °C.) for various functions. There
is need to extend rate measurements from organ systems to en-
zyme systems in order to learn to what extent physiological
variation depends on integration in the intact animal and to what
extent it may occur at the cellular or subcellular level. Rate func-
tions have meaning for speciation only if related to a critical en-
vironmental stress.

Behavior. Specific differences in reproductive behavior, par-
ticularly as concerns mating and care of young, have been useful
for differentiating certain insects and birds. In addition, taxic

344 THE VIEWPOINT OF A PHYSIOLOGIST

and orienting behavior, particularly in selection of an optimum
region in a gradient, for example of humidity or temperature,
may be of use in distinguishing races. Food selection, based on
taste "preferences," is partly established habit by "imprinting"
and partly based on genetic difference.

In general, the best examples of physiological differences from
population to population come from genera which have a wide
ecological distribution and which live in transitional environ-
ments. Care must be taken in selecting appropriate clines or
circles and in separating ecotypes from sibling species. Also there
may be interaction among several environmental factors, and it
must be remembered that animals react to a total environment
although certain factors may be more important than others for
a species.

Allopatric Speciation

The commonly assumed sequence of allopatric speciation of
sexually reproducing animals may be summarized as follows:

1. Phenotypic variation, both morphological and physiological,
occurs within populations of genetically similar individuals. Part
of this variation follows the normal distribution curve of a popu-
lation in an environment; part of the phenotypic variation is the
result of acclimation.

2. Primary adaptive variation becomes genetically fixed by
natural selection of randomly occurring genetic changes. Muta-
tions or chromosomal rearrangements are carried in an inter-
breeding population and genetic strains become established.

3. In a restricted environment or clinal situation ecotypes or
races become established. There may be gene flow between
races, but in one ecotype a given gene frequency may exceed
that in another ecotype.

4. Sufficient isolation, restrictive with respect to gene flow,
provides for the fixation of a selected genotype; such isolation
may occur between ends ol a eline or in insular populations, i.e.,
isolated by water or by unfavorable terrain surrounding breeding
sites. During isolation other differences become established, dif-
ferences whieh result in reproductive separation of the popula-

C. L. PROSSER 345

tions. Reproductive isolation may result from chromosomal
arrangement or mutation, from gross morphological, physiolog-
ical, or psychological incompatibility. Speciation is now effec-
tively complete.

5. Further changes resulting in habitat or ecological niche
selection may occur so that if the two species come to overlap
in range they may be separated ecologically as well as repro-
ductively. The initial or primary adaptive difference may have
lost its effectiveness.

Physiological variation is important at steps 1, 4, and 5.

Fig. 2. Schematic representation of distribution of a character in two
genotypes Gj and Go and in two populations of each genotype indicated
by phenotypes Pi and P2 for Gt and P', and P'2 for G2i

Figure 2 indicates two similar genotypes, each limiting the
range within which phenotypic variation can occur with respect
to some specified character. Two populations of the same geno-
type ( Gi ) living in two different environments, for example at
two latitudes, may show phenotypic patterns indicated by Pi
and P2. By acclimation, individuals of one of these populations
can shift their phenotypic expressions to correspond to those of
the other population. For example, the minimum lethal tempera-
tures can be altered for cold and warm water populations of
fish. In the other genotype or race (G2), one population P'i is
similar to Pi- in respect to the measured character, but by accli-
mation it is found that P'i cannot be converted to Pi but that it
can go beyond either Pi or Pj toward a new type P'^. According

346 THE VIEWPOINT OF A PHYSIOLOGIST

to the Baldwin effect, if a genetic change occurs in the direction
favored by selection, e.g., from Gi toward G2 in a population
already living near its limit with respect to that character, as in
P2, it is more likely to become fixed (step 2 above) than if such
a change should occur in a population far from this limit, as
Pi.* Conversely the genotype may become more limited. For
example, P2 may become genetically incompetent of acclimation
to Pi, and thus if these two populations are isolated by some
means they may become true species. There are many physiolog-
ical examples of Pi's and P2S but few physiological examples of
Gi's and G2's within single species. Distinction as to whether a
given difference between two populations is genetic or not can be
made only by acclimation or breeding experiments.

Inadequacies of the Simplified Scheme of Speciation

When one examines with care some of the physiological
analyses of populations and when one attempts field studies of
so-called races and related species and examines data from popu-
lation genetics, he begins to question the simple picture of specia-
tion which has been outlined above. None of the following
considerations is in itself decisive but in the aggregate they raise
serious doubts.

1. Every serious systematist is impressed more by the diversity
than by the uniformity of his species. Mutants of many sorts are
carried at certain levels in most breeding populations. We read-
ily recognize individual diversity in man and domestic animals
without knowing much about similar diversity in natural popula-
tions. There is pronounced heterozygosity in virile strains, and
balanced polymorphism permits the maintenance of several phe-
notypes in one population. Tt is unlikely that genetic factors
related to susceptibility to disease and to longevity are found
in man only.

' This docs not imply any causal relation between phenotype and genetic
change. Rather, individuals of P, would not be likely to move to the geo-
graphic range limit where the genetic change would be favored; pheno-
typically adapted individuals of Ps would tend to survive, hence reproduce
better in respeel to the critical stress, and if selection pressure is strong, G,
would permit migration to new range limits.

C. L. PROSSER 347

2. The notion that new characters are retained only if they
have adaptive advantage for survival is unrealistic, and there are
numerous examples of establishment of neutral morphological
and etiological patterns. Neutral characters may be retained
if they are linked or pleiotropic with advantageous characters;
according to the theory of genetic drift, neutral genes may be-
come established in small populations. As more neutral characters
are closely scrutinized, they appear to be associated genetically
with selective characters and hence they are present in ratios
different from those expected by random assortment (Sheppard,
1953). For example, human blood types are apparently corre-
lated with susceptibility to certain chronic diseases; type A is
more frequent and type O less frequent in people with cancer of
the stomach than in the general population (Aird et ah, 1953,
1954). The real problem of the physiologist is to recognize the
critical functional characters. In any case, the characters used
by taxonomists to distinguish species are often not adaptive ones.

3. While the range of a species may be very great, actual
breeding populations are often very small. Individuals in migrant
species tend to return to delimited areas. This restricts the prob-
ability of outbreeding. Territoriality is now recognized not only
among vertebrates — mammals, birds, and fish — but in some
crustaceans and insects, and it probably exists very widely. The
concept of gene flow throughout the range of a species is appar-
ently an oversimplification of the actual situation. In general,
breeding populations are more susceptible of isolation in fresh
water and terrestrial habitats than in marine environments, par-
ticularly if in the latter they have pelagic larvae.

4. Many of the so-called physiological races turn out to be
true species. When two populations differ enough to be physio-
logically distinct, they usually have developed sufficient differ-
ences to become reproductively isolated. The presence of a few
hybrids in zones of overlap does not invalidate the basic species
distinction, and in the laboratory there may occur interspecific
crosses which would be rare or nonexistent in nature. In clines
the terminal populations are often clearly different species, but
intermediate populations present difficulties for the taxonomist,

THE VIEWPOINT OF A PHYSIOLOGIST

and the cline as a whole may show progressive functional adapt-
edness. Usually sibling or morphologically similar species are
first considered as physiological races, but as they are studied
more intensively, they turn out to be natural species. Careful
observation usually reveals minor neutral morphological differ-
ences by which sibling species may be distinguished.

5. Physiological characters tend to be more sensitive to the
environment in their expression than do structural ones, and
the potential lability of individual animals is very great. The
changes which can be brought about by acclimation are extensive,
and as more populations are subjected to acclimation tests, more
and more do their differences turn out to be nongenetic. Many
so-called races are considered such merely because of their
ecology and they have not been subjected to adequate acclima-
tion tests.

6. Most physiological characters which have been analyzed in
animals are polygenic. In fact, a well-adapted animal is suited
to its entire environment, not merely to a single physical factor,
and it is adapted not through the action of one gene but by its
whole integrated genotype. A gene will often be expressed differ-
ently if the accompanying genes are changed. On this basis it is
perhaps futile to expect adaptive physiological characters to be
as susceptible of analysis as distinctive morphological ones.

7. It may be argued that the physiological criteria which
have been applied arc too gross, that measurement of oxygen
consumption by an animal is no more precise than stating the
number of segments on an appendage, that what is needed arc
more sensitive tests, measures of single enzymes. However, the
one gene-one enzyme hypothesis as developed so beautifully for
Neurospora may be ol limited applicability since it concerns
only defects in ability to synthesize relatively small molecules.
Very rarely is any animal with complex dietary requirements
dependent on any single enzyme pathway for either a specific
synthesis or degradation. There are numerous alternate pathways
l>\ which energy can be obtained from basic substrates. It is true
that interference with cytochrome oxidase1 would be disastrous
lor most animals, yet this enzyme must be under multiple control,

C. L. PROSSER 349

and many animals do function without it at restricted energy
levels. Also there are several alternate oxidative paths which can
bypass the cytochrome system. In other words, a genie change
interfering with single enzymes would not be readily detected
in complex animals, it would not likely become fixed because it
would be negative rather than positive or even neutral. However,
when enzyme changes have occurred in the direction of changes
in optima for a species, these have been selected; parasitic worms
can rely on anaerobic pathways, herbivorous insects tend to
have more active carbohydrases, and carnivorous ones more pro-
teases. However, there is no doubt that more sensitive functional
tests are needed. Also, it is not always easy to find the most crit-
ical or important functional characters.

In summary, physiological adaptedness to the environment may
exist in the absence of clear morphological speciation (a) in
balanced polymorphic populations (these include advantageous
heterozygotes and genie expression differing with genetic en-
vironment), (b) in clines, (c) in sibling species, and (d) in the
extensive phenotypic lability in response of individuals to en-
vironmental stress. Physiological characters may not conform to
"good" specific characters because (a) taxonomic characters are
often nonadaptive, (b) physiological characters are complex
genetically ( multif actor ) , (c) adaptive characters usually have
large safety factors and parallel or alternate paths, and functional
characters are either (d) quantitatively so sensitive to the en-
vironment that phenotypic variation exceeds genotypic, or (e)
qualitatively so complex that reproductive isolation has already
been established. If evolution is considered as the development
of adapted organisms and if adaptation is basically functional, it
is of prime importance to analyze populations in terms of physio-
logical characters. However, the origin of species per se may not
coincide with the development of adapted organisms.

Examples of Variation

We may next consider examples of populations which have
been examined for physiological variation with respect to various
limiting environmental factors. We shall consider first for each

350 THE VIEWPOINT OF A PHYSIOLOGIST

stress examples of phenotypic differences, then reported races,
and finally species adaptations.

Water and Ions. The transitions between sea and fresh
water and between regions of low and high humidity on land
permit much physiological variation in water economy. Closely
related to this is ion balance, although specific elements have
scarcely been considered as limiting factors in the distribution
of animals. Numerous examples of what appear to be phenotypic
osmotic differences between populations are known. North Sea
starfish ( Asterias rubens ) are much more tolerant of dilution than
populations from the Kiel Canal (Schlieper, 1929). Gills from
Mytilus from the North Sea (salinity 30°/oo) consumed 80 ml.
O2 per gram per hour, whereas after four weeks at a salinity of
15°/oo the corresponding value was 144 ml. O2 per gram per hour;
initial O2 consumption by gills from Mytilus living in the Baltic
at 15°/oo was 141 and after four weeks at 30°/oo it was 84 ml.
O2 per gram per hour ( Schlieper, 1953, 1955 ) . Blue crabs ( Calli-
nectes sapidus) collected in the upper parts of an estuary not
only tolerate lower salinities but regulate their internal osmo-
concentrations down to lower environmental limits than do crabs
from the mouth of the river, but after a week in normal sea water,
the crabs from the upper river approached in tolerance and regu-
lation those from the river mouth (Anderson and Prosser, 1953).

Several examples of possible osmotic races may be cited. Ne-
reis diversicolor in Denmark penetrate into nearly fresh water
and tolerate sudden transfer to fresh water better than do N.
diversicolor from England. These differences have been inter-
preted as indicating two races. However, tests of the regulation
of internal chloride as a function of environmental chloride show
similar limits of regulation for populations from Denmark, Scot-
land, and Finland (Smith, 1955a,b). The limiting salinity may
not be the summer value, which has been mostly studied, but
rather the spring dilution when the embryos are developing.
Perhaps sonic measure other than chloride regulation would re-
veal racial differences, but present evidence suggests only non-
genetic variation. Two brackish water gammarids, formerly
considered one species, are now recognized as two species, Gam-

C. L. PROSSER 351

marus zaddachi, which tolerates near-fresh water, and G. salinus,
which lives in the more saline regions of estuaries. The blood
concentration curves of the two forms have not been obtained,
but they are intersterile, and hence are effectively separate spe-
cies (Spooner, 1947; Kinne, 1954).

Genetically distinct clones of Euplotes vassus differ in their
ability to withstand transfer to high salinity corresponding to
saline lakes. Populations of Euplotes in salt lakes appear to have
become selected for salt-resistance (Gause, 1947). By analogy
with varieties of Paramecium (see below) these strains of Eu-
plotes may well be natural species.

The European stickleback, Gasterosteus aculeatus, occurs in
two subspecies (Heuts, 1947, 1949). One subspecies (gymnurus)
is predominantly a freshwater fish whereas the other (trachurus)
predominates in a brackish to marine habitat. The reproduction
(hatching) of gymnurus is better in fresh water, and that of
trachurus in salt water. Linked with salt-water tolerance is low
number of vertebrae, larger body size, and larger number of
lateral plates. Also the interaction between the effects of tem-
perature and salinity in affecting the number of dorsal fin rays
differs in the two forms. The physiological mechanisms of differ-
ences in salinity have not been examined, but this fish appears
to offer an example of the complex interaction of two environ-
mental stresses (salinity and temperature) and the linkage with
gross morphological characters. The two subspecies differ in many
ways, and despite the fact that they produce fertile hybrids,
natural selection favors the two extremes; they appear to be well
on the way to becoming natural species.

A similar example, now recognized as sibling species, is the
freshwater Anopheles gambiae and the morphologically similar
A. melas which normally breeds in brackish water and can
develop in 150% sea water. Hybrids are sterile, and the eggs
are morphologically distinguishable (Ribbands, 1944). The trop-
ical forest mosquitoes, Anopheles bellator and A. homunculus, oc-
cupy canopy levels which overlap on one side of each range;
bellator rises in the forest heights in the evening and descends
less far in the day because of its greater tolerance of low hu-

THE VIEWPOINT OF A PHYSIOLOGIST

midity ( Pittendrigh, 1950 ) . This difference is secondarily derived;
that is, the species were originally separated by some other fac-
tor.

Temperature. More different physiological criteria have been
used with temperature than with other environmental stresses.
Compensation for cold climates by poikilotherms has been re-
viewed and population differences have been described recently
(Bullock, 1955). The low lethal temperature of some temperate
zone marine animals may be above the high lethal temperature
of closely related arctic animals. However, lethal temperatures
can be altered markedly by acclimation, and familiar seasonal
variations in lethal temperatures reflect acclimation. Statements
about lethal temperatures of local populations need to include
information regarding previous temperature experience.

Many rate functions have been compared for animals from
different latitudes measured over a range of temperatures (Bul-
lock, 1955). The rate of oxygen consumption by cold-acclimated
poikilothermic animals is usually higher than for warm-accli-
mated ones, whether the acclimation is seasonal (Fundulus,
Wells, 1935; crustaceans, Edwards and Irving, 1943, as discussed
by Rao and Bullock, 1954), or latitudinal (Pandalus, Fox and
Wingfield, 1937), or altitudinal (planaria, Biasing, 1953). Ab-
sence of metabolic compensation for temperature change lias
been reported for other animals (hot springs fish, Sumner and
Lanham, 1942; certain terrestrial insects, Scholander et at., 1953;
an Alaskan pond gammarid, Krug 1954; and Rana pipiens, Fromm
and Johnson, 1955). Other rate functions such as heart rate,
breathing rate, and cruising speed indicate some compensation
by animals from cold environments. For example, My til us caJi-
fornianus from latitude 48°21' pumped water at the same rate
at 6.5°C. as those from 38°31' at 10°C. and those from 34°0' at
12°C. (Fig. 3) (Rao, 1953). Also the animals from colder waters
are less affected by temperature, i.e., their Qio is lower than for
animals from warmer regions. Adaptation by alteration in the
Qio, i.e., shift in slope as well as position of the rate-temperature
curve, occurs in some animals but not in others (Rao and Bullock,
1954; Scholander et ai, 1953; Bullock, 1955). In southern Cali-

C. L. PROSSER

353

fornia, limpets (Acmaea limatula) from the high intertidal zone
have consistently lower heart rates at a given temperature than
those from the low intertidal zone (Segal et oi, 1953). Mytilus
show a similar vertical variation in pumping rate; 75 cm. ver-
tically are equivalent in pumping to 330 miles latitudinally when
tested at 16°C. (Rao, 1953). By transplantation of limpets and
Mytilus in the vertical gradient, complete acclimation was

100
80
60

^?30

c
a. E

E

20

10

48° N.

-O 39° N.

34° N.

JL

_L

1

J

6 8 10 12 14
Temperature (°C.)

16 18 20

Fig. 3. Rate of pumping of water by 50-g. specimens of Mijtilus cali-
fornianus as a function of temperature for populations collected at three
latitudes of western North America. From Rullock (1955); redrawn from
Rao (1953).

demonstrated within a few weeks. Latitudinal transplantation
experiments have not been reported, but it is likely that many
of the latitudinal effects with respect to temperature are pheno-
typic.

Racial differences have been invoked to explain latitudinal
differences in reproductive rates; poikilothermic animals from
cold waters generally develop more rapidly than those from
warm waters when tested at the same temperature, especially in
the lower range. A low-temperature race of Daphnia atkinsoni
is found in pools in England where the breeding season is re-

354

THE VIEWPOINT OF A PHYSIOLOGIST

stricted by summer evaporation (Johnson, 1952). Oysters (Cras-
sostrea virginica) from Virginia, where normal spawning occurs
at 25°C, failed to spawn during two years in Long Island

10 0

O

X

<x>

7.0

5.0

40

c2 30

o i

o

>

cy

cr

D
O
0)

c
o

c

1.0

0.7

05

04

03

O
O

o

THAIS EMARGINATA

ALASKA

• •

so CALir.

-i i i u

j_

_i_

_i_

_L-

10 12 14 16 18

Te m p_ C

20

22

24

Fig. 4. Growth rates of shell length of veliger larvae of Thais emarginata
as a function of temperature for populations horn Mount Edgecombe,
Alaska and Big Rock, California. From Dehnel (1955), by permission of
University of Chicago Press.

Sound, where the temperature failed to reach 25°C; yet only
adults were transplanted, not spat (Loosanoff and Nomcjko,
1951). When spat of another species (Ostrea lurida) were trans-
planted from Puget Sound to Long Island Sound, survival and

C. L. PROSSER 355

breeding were possible only for those oysters brought into the
laboratory during the winter (Davis, 1955). Similarly, the oyster
drills, Urosalpinx cinerea of Virginia and Delaware, have differ-
ent temperature thresholds for spawning (Stauber, 1950). The
effect of latitude on rate of growth of embryos of various marine
invertebrates has been well summarized by Delmel (1955) who
added observations on certain species of snails from southern
California and Alaska. When growth rate was measured at
different temperatures in the laboratory, the Alaskan snails grew
significantly faster at corresponding stages ( Fig. 4 ) . In the middle
of the physiological temperature range the northern Qi0 values
were higher in those snails which showed no overlap in rates,
lower in the others. Similar findings have been reported earlier
for European marine invertebrates (Fox, 1938; Thorson, 1951).
These differential growth rates can partly account for the rela-
tively greater productivity of colder waters. It appears that the
Alaskan and southern Californian populations of the same species
of snails are genetically different, yet no transplantation experi-
ments were performed and the animals at intermediate latitudes
were not compared. It is likely that clines with respect to the
effect of temperature on development of marine invertebrates
may occur and that between the ends of such clines major
genetic differences exist. However, there is need for transplanta-
tion of larval stages and acclimation over a life cycle.

Among insects, several clines have been described where
temperature seems to be critical for development. In the moth
Lymantria dispar, Goldschmidt (1940) showed in his classic
studies that the rate of development in south European races is
slower than in north European populations. In pine-needle wasps
(Diprion) alpine races show faster development in the egg and
lower Qios, especially for conversion from pupa to adult, than
do populations from warmer regions (Elens, 1953).

Rana pipiens from northern United States breed at lower tem-
peratures, develop faster, have lower temperature coefficients of
development than frogs from extreme southern states and low-
lands of Mexico (Moore, 1949). Slight morphological differences
and a zone of hybridization confirm the conclusion that the

356 THE VIEWPOINT OF A PHYSIOLOGIST

northern and southern Rana pipiens are genetic races or sub-
species; in fact, hybrid abnormalities are so great that species
status is approached (Volpe, 1954).

In several species of DrosopJiila, e.g., pseudoobscura and per-
similis, natural populations contain individuals with several dis-
tinct arrangements of chromosome bandings. The proportion of
each pattern may vary with season and altitude, and clines are
observed with respect to certain inversions. Controlled breeding
in cages at constant temperature and humidity shows that the
percentage of certain types is in part temperature-dependent.
For example, when several types of D. persimilis were reared to-
gether, the high-altitude type became predominant at 16° C. but
not at 25° C. ( Spiess, 1950). When two gene arrangements of
D. pseudoobscura from the same locality were bred through
many generations, the heterozygote became more common than
the homozygotes. However, if the same two forms from different
localities were bred there was sometimes stabilization with one
homozygote predominating. Hence geographic differences in
gene contents must exist separate from the differences in chromo-
some pattern (Dobzhansky, 1954). The selective effects of tem-
perature on the chromosome polymorphism have been demon-
strated to occur in both pupae and adults (Levine, 1952), and
physiological studies are needed to learn the bases for the tem-
perature adaptedness.

Both genetic and nongenetic differences are shown in the fol-
lowing experiments on fish in which the zones of temperature
tolerance were tested after acclimation to different temperatures.
Fourteen species of freshwater fish from different latitudes
(Ontario, Tennessee, and Florida) were compared (Hart, 1952).
The tolerance polygons showed genetic differences for three of
these, the common shiner (Notropis cornutus), the large-
mouthed bass (Micwptcrus salmoides), and Gambusia affinis.
Morphological differences between northern and southern forms
of the first two are such that they are considered subspecies. Sev-
eral of the others such as golden shiner and blacknose dace are
morphological subspecies but they fail to show differences in
their temperature tolerance polygons.

C. L. PROSSER 357

In Paramecium aurelia there are eight varieties which are
really natural species (Sonneborn, 1950); two of these are dis-
tinguished by the temperatures optimal for their growth and re-
production ( Sonneborn and Dippell, 1943 ) .

Oxygen. A considerable amount of nongenetic variation with
respect to oxygen requirement is possible. For example, intestinal
parasites and such free-living animals as Tubijex and Daphnia
are able to alter their metabolism or their oxygen transport sys-
tem according to the environmental oxygen. It is entirely possible
that some of the effects on freshwater poikilotherms attributed to
temperature are in fact effects of oxygen, since in winter the
oxygen in stagnant ponds where there is no photosynthesis may
be very low. Altitudinal effects attributable to oxygen lack have
been reported only for birds and mammals where the phenotypic
response of increase in blood hemoglobin is well known. Of two
species of a Russian mouse, Apodemus, one, sylvaticus, lives in
both mountains and plains, the mountain populations having
higher hemoglobin concentrations, while the other, agrarius, is
restricted to the plains. Apparently at high altitudes members of
the latter species fail to acclimate by increasing their blood
hemoglobin (Kalabuchov, 1937).

Oxygen gradients do not in general provide good geographic
series, but microecologically oxygen supply may be important in
the isolation of populations. For example, in a series of chirono-
mid larvae, the stream species survived anoxia only a short time
whereas species from low-oxygen ditches survived anoxia for
about 100 hours (Walshe, 1948). Gammarids and planaria from
fast well-aerated water have higher rates of oxygen consumption
than do related species from sluggish water poor in oxygen. Some
snails and numerous parasitic helminths get along well on the
energy of glycolysis and excrete the acids which are formed. In
fact, several patterns of repayment of oxygen debt are recognized
(von Brand, 1952). The known genotypic variations recognized
with respect to oxygen are largely at the species level.

Foods and Toxins. There are marked interspecific differences
in digestive enzymes according to natural diets and a few in-
stances of slight differences in amino acid requirements, but in

358 THE VIEWPOINT OF A PHYSIOLOGIST

general, nutritional requirements are so similar and so general
that subspecific differences are slight. Selection of specific food
plants by herbivorous insects appears to be based on taste prefer-
ences for certain specific organic components of the plants which
are repellent or neutral to other insects. Thorpe (1939) pointed
out that selection of food plants can be developed by habituation,
as by repeatedly exposing the larvae to a particular food, and he
has experimentally habituated Drosophila to a medium contain-
ing peppermint and the ichneumid Nemeritis from Ephestia to
Meliphora. Dixippus morosus was "forced" to feed on ivy, initially
an unfamiliar food, instead of privet; initial mortality was high,
but after several generations the progeny accepted the ivy readily
(Sladden, 1934). Eventually natural selection can result in ge-
netic fixation of food preferences which are initially phenotypic
variants established by habituation. A population of sawfly, Pon-
tania salicis, which had been feeding on Salix andersonia, was
forced to eat only S. rubra, until after some four years the saw-
flies failed to make galls on S. andersonia when this species was
planted near-by (Harrison, 1927). This interesting experiment
deserves confirmation under carefully controlled conditions.

Most spectacular are tolerances of natural alkaloids by some
insects, such as those feeding on tobacco, and the tolerance of
modern insecticides by various strains. In some insects, resistance
to DDT seems to be carried by a single gene but in most it is
multifactorial. In many the genetic pattern necessary for resist-
ance was present in the population before the specific stress of
insecticides appeared (DDT resistance in Musca, Lichtwardt et
ah, 1955 ) . In a few the resistance appeared during many genera-
tions of selection ( Crow, 1954 ) . The complexities of resistance to
insecticides have been recently reviewed (Metcalf, 1955). Re-
sistant strains do provide a basis for natural selection of a particu-
lar race, which may ultimately lead to speciation, since areas
where the toxin exists are spatially separable from those without
the toxin and are well maintained.

Examples of species with very minor morphological differences
but isolated by food preferences are: the homopterans Psylla
mail on apple and peregrina on hawthorn (Lai, 1934); and two

C. L. PROSSER 359

species of the butterfly Colias, eurytheme on alfalfa and philodice
on red clover ( Hovanitz, 1949 ) .

Other environmental factors which may be important as iso-
lating agents are such sensory stimuli as light and mechanical
stimuli. For example, blind races of the Mexican characin, Ano-
ptichthys jordani, inhabit dark caves where lack of vision is not a
disadvantage ( Breder, 1943 ) . Wind and water currents, also sub-
strate composition, limit the distribution of a variety of animals.
Biotic Selection. Related to food plant preference as an iso-
lating mechanism is the selection of host organism or of symbiont.
Termites which are very similar morphologically are clearly dis-
tinguished by the species of staphylinid beetle living in their nests
(Emerson, 1935). Ascaris lumbricoides is morphologically similar
in man, pig, and monkey, yet attempts at reciprocal infection
have been unsuccessful ( Faust, 1949 ) ; these ascarids may be true
sibling species. Probably the sibling species most difficult to un-
derstand are those of the Anopheles maculipennis complex. Eight
different strains are recognizable by their selection of man or
other mammals as blood sources, also by their ability to develop
in saline water. These strains are morphologically similar except
for minor differences in egg structure. However, laboratory
breeding experiments demonstrate that in reality the "strains" are
sibling species (Hackett and Missiroli, 1935; Bates, 1949). An
American mosquito, Culex pipiens, has two strains, pipiens, which
for egg production requires a blood meal, and molestus, which
can produce eggs without it (Kitzmiller, 1953).

Serological specificity distinguishes closely related groups and
reveals incompatibility between more distant groups. "Blood
types" have been analyzed in detail in man, in pigeons, and in
cattle. In man, certain serological types, also certain hemoglobin
types, are correlated with other genetic factors, and each type is
carried at a level characteristic of the ethnic group ( Boyd, 1950 ) .
Present suggestions of linkage of blood type with incidence of
certain chronic diseases indicate the need for greatly expanded
medical statistics.

Behavior. Fully as important as physical and chemical factors
in maintaining (possibly not in initiating) isolation between pop-

360 THE VIEWPOINT OF A PHYSIOLOGIST

ulations are behavioral differences, particularly the behavior asso-
ciated with reproduction. For example, six similar species of the
willistoni group of Drosophila show mating habits so characteris-
tic that no interspecific mating occurs (Spieth, 1947, 1949). Two
sibling species of solitary wasps Ammophila campestris group, are
distinguished mainly by their mode of opening the burrow and
providing prey for their larvae (Adriaanse, 1947). Eastern and
western American meadowlarks can be distinguished by their
calls,* as can European chiffchaffs, several species of crickets, and
also two southern frogs in regions of overlap ( Blair, 1955 ) . These
behavioral traits could act as primary isolating characters after
becoming established by genetic drift as in "island" populations,
or they may secondarily separate species in which the initial iso-
lating agents are no longer active.

Mechanisms of Nongenetic Variation of the Phenotype

A stress physiologist is continually impressed by the phenotypic
lability of individual organisms. In regulating animals, acclima-
tion shifts the environmental limits for a stable internal state. In
conforming animals, acclimation raises or lowers the tolerable
internal variation or introduces some compensating mechanism,
for example, metabolic adaptation in poikilotherms. Since pheno-
typic variation may be important in initial stages of speciation
and since numerous population differences turn out to be non-
genetic, it is important in a discussion of the species problem to
point out possible mechanisms of phenotypic adaptation.

In describing any change induced in an individual by environ-
mental stress, the time during which the change occurs must be
precisely stated. For example, mammals exposed to cold show
initial constriction of surface blood vessels and piloerection. Next
they may shiver and then increase the activity of thyroid and
adrenal cortex, and they may maintain elevated metabolism lor
weeks. Long-range adaptation, however, consists of increased in-

0 A recent paper (Lanyon, W. E., 1956. Ecological aspects of the sym-
patric distribution of meadowlarks in the north-central states. Ecology, 37,
98-108) suggests that the primary mechanism isolating the eastern and
western species of meadowlark is indirectly related to moisture, particu-
larly spring precipitation.

C. L. PROSSER 36]

sulation in hair or subcutaneous fat. Thus, short-term regulation
and long-term acclimation are accomplished by very different
mechanisms. Long-term changes are often morphological — tad-
poles have larger gills when reared in low oxygen, certain mos-
quito larvae have larger anal papillae in freshwater than in salt
water. There may also be behavioral changes, such as aggregation
by hive bees in winter.

In addition, many physiological and biochemical adaptations
are known. Lungfish in estivation shift their products of nitrogen
excretion to urea from the ammonia they excrete when actively
swimming. Daphnia and Artemia reared in low oxygen develop
high levels of hemoglobin whereas in high oxygen their hemo-
lymph is colorless. Also, in low oxygen there is an increase in
tissue cytochrome (Fox, 1955). The amount of "free" water di-
minishes in cold-hardened insects and molluscs.

We have referred above to the increased metabolic rate in
aquatic poikilotherms acclimated to cold either seasonally or
latitudinally. That this change is at least partly cellular is shown
by the fact that tissue slices from certain organs of cold-acclimated
fish and crustaceans consume more oxygen than do those from
warm-water individuals (Piess and Field, 1950; Freeman, 1950;
Roberts, 1953).

It appears likely, therefore, that environmental stresses can re-
sult in biochemical adaptations at the cellular level. The enzymes
of thermophilic bacteria show higher temperatures of inactivation
than corresponding enzymes of mesophilic bacteria, and the in-
activation temperature differs for different enzymes from the
same organism; these differences are clearly genetic. However,
Precht and his associates ( Christophersen and Precht, 1950, 1951)
report that the highest temperature of activity in yeast measured
by oxygen consumption and by methylene blue reduction is
higher (65° C.) for yeast reared at 41° C. than it is (60° C.) for
yeast reared at 20° C.; also the Qo2 of 20° yeast is higher than
that of 41° yeast when measured at the same temperature. The
experiments with yeast agree with data from poikilothermic
animals that acclimation temperature may change the upper
limiting temperature, the absolute level of oxygen consumption

i

'■< UA .

"

♦ !•

362 THE VIEWPOINT OF A PHYSIOLOGIST

and, in some instances, the temperature characteristic of the total
metabolic system. The explanation of such acclimation effects is
by no means clear. It is unlikely that qualitative changes in en-
zyme proteins occur, since protein structure appears to be genetic-
ally fixed. There may, however, be quantitative changes in en-
zymes, a sort of enzyme induction.

Among microorganisms enzyme synthesis can be induced by
forcing the use of unfamiliar substrates, for example sugars which
the organism had not been able to metabolize. Enzyme induction
has been demonstrated in animals — the synthesis of tryptophane
peroxidase in rat liver induced by injection of L-tryptophane
(Knox, 1951; Lee and Williams, 1953). There is no evidence,
however, that physical factors such as salinity, temperature, and
oxygen partial pressure can similarly induce enzyme synthesis.
The following hypothesis might explain the effects of physical
agents in terms of enzyme induction: It is recognized but not
sufficiently emphasized that there are numerous alternate paths
in digestion and in intermediary metabolism. Some of these paths
are followed by certain tissues and animals and at certain stages
in a life cycle more than are other paths. Parasitic helminths in
hypoxia switch to glycolytic systems and excrete the acids
formed. Cecropia pupae use a cyanide-insensitive oxidative path-
way (presumably a flavoprotein ) until cytochrome oxidase is
activated by hormones of metamorphosis. If one path is affected
more than another by temperature, by oxygen pressure, or by
salinity, some of its substrates might pile up, and these might
serve as intermediate substrates of an alternate path. These sub-
stances might then induce synthesis of some limiting enzyme of
the alternate path. The limiting temperatures and the tempera-
ture characteristics of the overall metabolism might be changed
according to relative importance of given pathways. In fact, a
change in slope of the rate-temperature curve (change in tem-
perature characteristic) could effectively increase or decrease the
absolute rate of oxygen consumption at a given temperature.

What is needed is a quantitative assay of a variety of known
alternate paths under different acclimation states, a very difficult
task. An enzymatic analysis of adaptive phenotypic variation is

C. L. PROSSER 363

as important as the study of the relations between genes and
enzymes.

Conclusions

Examination of physiological characters, particularly by stress
tests, can provide much useful information about variation in
natural populations. When populations of animals which differ
markedly in their responses to a given stress are properly accli-
mated to each other's environment, the differences often are seen
to be nongenetic. In fact, the functional lability of animals is very
great. Part of this individual adaptability is behavioral, part oc-
curs at the organ system level, part of it is cellular, and part may
result from enzyme induction.

Some examples of genetic fixation of adaptive variations in
mutant strains are known, for example, insecticide-resistant in-
sects. Variants may become established as ecotypes or races, par-
ticularly in clinal situations; as an example, animals with different
temperature dependence of rate functions. More commonly,
when adaptive physiological differences between two popula-
tions are genetically established, the differentiation is sufficient
for reproductive isolation; that is, speciation occurs. Many sibling
species or morphologically similar species, which differ in physio-
logical adaptability, are recognized, and stress tests applied to
populations would probably reveal numerous natural species. In
summary, most functional variation among animal populations
appears to be either nongenetic or specific; relatively little is
racial.

Neutral morphological variations are sometimes genetically
carried with physiological characters by linkage or pleiotropism.
Also, neutral characters may provide the basis for distinguishing
typological species. Most adaptive functional characters of ani-
mals appear to be multifactorial; adaptation resides not in single
genes but in the balanced genotype; the compensatory reactions
to deviations from a "normal" physiological state make recogni-
tion of small genetic changes difficult.

There is urgent need for two kinds of research in the area of
physiological variation. First, there is need for more sensitive

364 THE VIEWPOINT OF A PHYSIOLOGIST

functional tests of variation and for cellular and enzymatic anal-
yses of the ways in which the environment induces phenotypic
change. Second, to better understand natural variation and to
distinguish nongenetic from genetic differences there is need for
careful acclimation experiments on populations collected over the
full range of a species, the examination of physiological clines,
in experiments extending over complete life cycles. The use of
physiological responses to various environmental stresses as cri-
teria of natural variation permits a close approach to the key
problem of evolution, the development of adapted types of plants
and animals, if not to the origin of species.

REFERENCES

Adriaanse, M. S. 1947. Ammophila campestris Latr. unci Ammophila

adriaanus Wilcke. Ein Beitrag zur vergleichenden Verhaltungsfor-

schung. Behaviour, 1, 1-34.
Adolph, E. F. 1943. Physiological Regulations. Jaques Cattell Press,

Lancaster, Pa.
Aird, I., and H. H. Bentall. 1953. A relationship between cancer of

stomach and the ABO blood groups. Brit. Med. J., 1, 799-801.
Aird, I., H. H. Bentall, and J. A. Mehigan. 1954. Brit. Med. J., 2, 315-

21.
Anderson, J. D., and C. L. Prosser. 1953. Osmoregulating capacity in

populations occurring in different salinities. Biol. Bull., 105, 369.
Bates, M. 1949. The Natural History of Mosquitoes. The Macmillan

Company, New York, N.Y.
Blair, W. F. 1955. Mating call and stage of speciation in the Micro-

lu/la olivacea-M. carolinensis complex. Evolution, 9, 469-80.
Biasing, I. 1953. Experimentelle Untersuchungen iiber den Umfang

der okologischen und physiologischen Toleranz von Planaria alpina

Dana und Planaria gonocephala Dugcs. Zool. Jahrh. Aht. (dig. Zool.

u. Phys., 64, 111-52.
Boyd, W. C. 1950. Genetics and the Races of Man. Little, Brown and

Co., Boston, Massachusetts.
Brand, T. v. 1952. Chemical Physiology of Endoparasitic Animals.
Academic Press, New York, N.Y., Chapters 10, L3, 14.

C. L. PROSSEK 365

Breder, C. M. 1943. Problems in the behavior and evolution of a spe-
cies of blind cave fish. Trans. N.Y. Acad. Sci., 5, 168-76.

Brett, J. R. 1944. Some lethal temperature relations of Algonquin Park
fishes. Publ. Ontario Fish. Research Lab., 63, 1-49.

Bullock, T. H. 1955. Compensation for temperature in the metabolism
and activity of poikilotherms. Biol. Revs. Cambridge Phil. Soc, 30,
311-42.

Christophersen, J., and H. Precht. 1950. Fermentative Temperatur-
adaptation. Biol. Zentr., 69, 240-56.

Christophersen, J., and H. Precht. 1951. Biol. Zentr., 70, 261-74.

Crow, J. F. 1954. Analysis of a DDT resistant strain of Drosophila.
J. Econ. Entomol, 47, 393-98.

Davis, H. C. 1955. Mortality of Olympia oysters at low temperatures.
Biol. Bull, 109, 404-6.

Dehnel, P. A. 1955. Rates of growth of gastropods as a function of
latitude. Physiol. Zool., 28, 115-44.

Delacour, J. 1954. The Waterfowl of the World. Country Life, Lon-
don.

Dobzhansky, T. 1954. Evolution as a creative process. Caryologia 6,
suppl, 1, 435-49.

Edwards, G. A., and L. Irving. 1943. The influence of temperature
and season upon the oxygen consumption of the sand crab, Emerita
talpoida Say. /. Cellular Comp. Physiol., 21, 169-81.

Elens, A. A. 1953. Etude ecologique des lophyres en campine (Bel-
gique) II. Agricultura (Louvain) n. s. 1, 19-32; from Biol. Ahst..
1954.

Emerson, A. E. 1935. Termitophile distribution and quantitative char-
acters as indicators of physiological speciation in British Guiana
termites (Isoptera). Ann. Entomol. Soc. Amer., 28, 369-95.

Faust, E. C. 1949. Human Helminthology, Lea and Febiger, Philadel-
phia, Pennsylvania.

Fox, H. M. 1938. The activity and metabolism of poikilothermal ani-
mals in different latitudes. III. Proc. Zool. Soc. London, 108, 501-5.

Fox, H. M. 1955. The effect of oxygen on the concentration of haem
in invertebrates. Proc. Roy. Soc. (London), B143, 203-14.

Fox, H. M., and C. A. Wingfield. 1937. The activity and metabolism
of poikilothermal animals in different latitudes. II. Proc. Zool. Soc.
London, 107, 275-82.

Freeman, J. A. 1950. Oxygen consumption, brain metabolism and re-

366 THE VIEWPOINT OF A PHYSIOLOGIST

spiratory movement of goldfish during temperature acclimatization,
with special reference to lowered temperatures. Biol. Bull., 99, 416-
24.

Fromm, P., and R. E. Johnson. 1955. The respiratory metabolism of
frogs as related to season. /. Cellular Camp. Physiol., 45, 343-59.

Fry, F. E., J. R. Brett, and G. H. Clawson. 1942. Lethal limits of tem-
perature for young goldfish, Rev. can. biol., 1, 50-56.

Gause, G. F. 1947. Problems of evolution. Trans. Connecticut Acad.
Arts and Sci., 37, 17-68.

Goldschmidt, R. 1940. The Material Basis of Evolution. Yale Univer-
sity Press, New Haven, Connecticut.

Hackett, L. W., and A. Missiroli. 1935. The varieties of Anopheles
maculipcnnis and their relation to the distribution of malaria in
Europe. Riv. malariol., 14, 45-109.

Harrison, J. W. H. 1927. Experiments on the egg-laying instincts of the
sawfly Pontania salicis and their bearing on the inheritance of ac-
quired characters. Proc. Roy. Soc. (London), B101, 115-26.

Hart, J. S. 1952. Geographic variations of some physiological and
morphological characters in certain freshwater fish. Puhl. Ontario
Fish. Research Lab., 72, 1-79.

Heuts, M. J. 1947. Experimental studies on adaptive evolution in
Gasterosteus aculeatus. Evolution, 1, 89-102.

Heuts, M. J. 1949. /. Genetics, 49, 183-91.

Hovanitz, W. 1949. Increased variability in populations following nat-
ural hybridization. In Jepson, Mayr, and Simpson, Genetics, Pale-
ontology and Evolution. Princeton University Press, Princeton, New
Jersey, Chapter 18.

Johnson, D. S. 1952. A thermal race of Daphnia atkinsoni Baird and
its distributional significance. /. Animal Ecol., 21, 118-19.

Kalabuchov, N. J. 1937. Some physiological adaptations of the moun-
tain and plain forms of the woodmouse (Apodemus sylvaticus) and
of other species of mouse-like rodents. /. Animal Ecol, 6, 254-72.

Rhine, (). 1954. Die Gammarus Arten der Kieler Bucht. 7,ool. Jahrb.
Abt. Syst. Okol., 82, 405-24.

kit/miller, J. B. 1953. Mosquito genetics and cytogenetics. Rev. brasil.
malariol. c. doengas trop., 5, 285-359.

Knox, W. E. 1951. Two mechanisms which increase in vivo the liver
tryptophan peroxidase activity: specific adaptation and stimulation
<»i the pituitary-adrenal system. Brit. ./. Exptl Path., 32, 462-69.

Krug, J. L954. The influence of seasonal environmental changes upon

C. L. PROSSER 367

the metabolism, lethal temperature and rate of heart beat of Gam-
marus limnaeus (Smith) taken from an Alaskan lake. Biol. Bull,
107, 397-410.

Lai, K. B. 1934. Psyllia peregrina Forst, the hawthorn race of the apple
sucker P. mali Schmidb. Ann. Appl. Biol, 21, 641-48.

Lee, N. D., and R. H. Williams. 1953. Protein turnover and enzymatic
adaptation. /. Biol Cliem., 204, 477-86.

Levine, R. P. 1952. Adaptive responses of some third chromosome
types of Drosophila pseudoobscura. Evolution, 6, 216-43.

Lichtwardt, E. T., W. M. Luce, G. C. Decker, and W. N. Bruce. 1956. A
genetic test of DDT resistance in field house-flies. Ann. Entomol.
Soc. Amer., 48, 205-10.

Loosanoff, V. L., and C. A. Nomejko. 1951. Existence of physiologi-
cally different races of oysters, Crassostrea virginica. Biol Bull,
101, 151-56.

Metcalf, R. L. 1955. Physiological basis for insect resistance to insecti-
cides. Physiol Revs., 35, 197-232.

Moore, J. A. 1949. Geographic variation of adaptive characters in
Rana pipiens. Evolution, 3, 1-21.

Piess, C. N., and J. Field. 1950. The respiratory metabolism of excised
slices of warm and cold-adapted fishes. Biol. Bull, 99, 213-24.

Pittendrigh, C. S. 1950. The ecoclimatic divergence of Anopheles
bellator and A. homunculus. Evolution, 4, 43-63, 64-78.

Prosser, C. L. 1955. Physiological variation in animals. Biol. Revs.,
Cambridge Phil. Soc, 30, 229-62.

Rao, K. P. 1953. Rate of water propulsion in Mytilus californianus as
a function of latitude. Biol Bull, 104, 171-81.

Rao, K. P., and T. H. Bullock. 1954. Q10 as a function of size and habi-
tat temperature in poikilotherms. Am. Naturalist, 88, 33-44.

Ribbands, C. R. 1944. Differences between Anopheles melas and
Anopheles gambiae. Ann. Prop. Med. Parasitol, 38, 87-99.

Roberts, J. L. 1953. Studies on thermal acclimatization in the lined
shore crab, Pachygrapsus crassipes Randall. XIX Intern. Physiol.

Congr., 706-7.
Schlieper, C. 1929. Ueber die Einwirkung niederer Salzkonzentratio-

nen auf marine Organismen. Z. vergleich. Physiol, 9, 478-514.
Schlieper, C. 1953. Zur Frage der Beziehungen zwischen osmotischer

Resistenz und Grundumsatz bei euryhalinen Meeresvertebraten.

Natunvissenschaften, 20, 538-39.

368 THE VIEWPOINT OF A PHYSIOLOGIST

Schlieper, C. 1955. Ueber die Physiol. Wirkungen des Brackwassers.
Kiel Meeresforsch., 11, 22-33.

Scholander, P. F., W. Flagg, V. Walter, and L. Irving. 1953. Climatic
adaptation in arctic and tropical poikilotherms. Phys. Zool., 26, 67-
92.

Segal, E., K. P. Rao, and T. W. James. 1953. Rate of activity as a func-
tion of intertidal height within populations of some littoral molluscs.
Nature, 172, 1108-9.

Sheppard, P. M. 1953. Polymorphism, linkage and the blood groups.
Am. Naturalist, 87, 283-94.

Sladden, D. E. 1934. Transference of food habit from parent to off-
spring. Proc. Roy. Soc. (London), B114, 441-49.

Sladden, D. E. 1935. Proc. Roy. Soc. (London), 119, 31-46.

Smith, R. I. 1955a. On the distribution of Nereis diversicolor in rela-
tion to salinity in the vicinity of Tvarminne, Finland and the Ise-
fjord, Denmark. Riol. Bull., 108, 326-45.

Smith, R. I. 1955b. Comparison of the level of chloride regulation by
Nereis diversicolor in different parts of its geographical range. Biol.
Bull, 109, 453-74.

Sonneborn, T. M. 1950. Paramecium in modern biology. Bios., 21, 31-
43.

Sonneborn, T. M., and R. V. Dippell. 1943. Sexual isolation, mating
types and sexual response to diverse conditions in variety 4, Para-
mecium aurelia. Biol. Bull., 85, 36-43.

Spiess, E. B. 1950. Experimental populations of Drosophila pcrsimilis
from an altitudinal transect of the Sierra Nevada. Evolution, 4, 14-
33.

Spieth, H. T. 1947. Sexual behavior and isolation in Drosophila. Evo-
lution, 1, 17-31.

Spieth, H. T. 1949. Evolution, 3, 67-81.

Spooner, G. M. 1947. The distribution of Gammarus species in estu-
aries. J. Marine Biol. Assoc. United Kingdom, 27, 1-52.

Spooner, G. M. 1951. /. Marine Biol. Assoc. United Kingdom, 30, 129-
47.

Stauber, L. A. 1950. The problem of physiological species with special
reference to oysters and oyster drills. Ecology, 31, 107-18.

Sumner, F. B., and U. N. Lanham. 1942. Studies of the respiratory
metabolism of warm and cool spring fishes. Biol Bull., 82, 313-27.

Thorpe, T. H. 1939. Further studies in pre-imaginal olfactory condi-
tioning in insects. Proc. Roy. Soc. (London), B127, 424-33.

C. L. PROSSER 369

Thorpe, T. H. 1937. Proc. Roy. Soc. (London), 124, 56-81; and 126,

370-97.
Thorson, G. 1951. Zur jetzigen Lage der marinen Bodentier Okologie.

VerhandL. detit. Ges. Willwlmshaven, 276-327.
Volpe, S. P. 1954. Hybrid inviability between Rana pipiens from Wis-
consin and Mexico. Tulane Studies Zool., 1, 110-23.
Walshe, B. M. 1948. The oxygen requirements and thermal resistance

of chironomid larvae from flowing and from still waters. /. Exptl.

Biol, 25, 35-44.
Wells, N. A. 1935. Variations in respiratory metabolism of the Pacific

killifish, Fandidus parvipinnis, due to size, season and continued

constant temperature. Physiol. Zool, 8, 318-336.
Wells, N. A. 1935. Biol Bull, 69, 361-71.

DIFFICULTIES AND IMPORTANCE OF THE
BIOLOGICAL SPECIES CONCEPT

ERNST MAYR: museum of comparative zoology.

HARVARD COLLEGE, CAMBRIDGE, MASSACHUSETTS

To bring into focus the picture that emerges from this sym-
posium one must consider very diverse material on which the
seven contributors have based their discussions — living organ-
isms and fossils, animals and plants, freshwater and terrestrial
organisms, vertebrates and invertebrates.* It would not have been
surprising if the seven speakers had represented seven entirely
different viewpoints. This did not happen. Indeed, the general
agreement among the speakers was quite far-reaching; for in-
stance, all speakers with one exception have emphatically
endorsed the biological species concept. Yet it is evident that
we have not yet reached a true synthesis. The approach of
every worker in this field is still largely colored by his intimate
knowledge of the material with which he himself works. Let me
illustrate this with a few examples. When an ornithologist speaks
of "species" he has a phenomenon in mind in which hybridiza-
tion or lack of sexuality are of no consequence. He has great
difficulty in determining whether or not species in birds and
species in plants are the same kind of phenomenon. Or let us take
another case. The population geneticist who has emerged in the
last thirty years from the typological thinking of mutationism
is likely to consider everything as new that he himself is learning
about species. He is unaware of the fact that thinking in terms
of populations, and indeed the whole biological species concept,
came from systematics into genetics rather than the reverse
(Mayr, 1955). Population genetics has, often quite independ-

° The revised version of T. M. Sonneborn s contribution was not avail-
able when this discussion was prepared.

371

372 DIFFICULTIES AND IMPORTANCE OF THE CONCEPT

ently, rediscovered much that has been considered axiomatic in
population systematics for seventy to eighty years.

Our thinking on these questions is one-sided not only because
it is so strongly affected by our working material, but also be-
cause it is based on a limited number of selected examples. An
author who works with a hybrid complex will be impressed by
the importance of hybridization, another one who works with
insular forms by the importance of geographic isolation. We now
have reached the point where we are badly in need of compara-
tive systematics and of a strictly quantitative approach. Analyses
such as were done by Verne Grant (this symposium) should be
clone for as many groups of animals and plants as possible. Ob-
viously, this method can be applied only where the group has
reached a considerable degree of taxonomic maturity, but there
are now many such groups in the vertebrates, insects, and plants.
The need for such comparative systematics emphasizes the im-
portance of sound orthodox taxonomic monographs.

Difficulties in the Application of Species Concepts

In the introduction I attempted to describe the three basic
philosophical species concepts which play a role in systematics.
Much of the discussion of this symposium dealt with the problem
of the utilization of such concepts in the taxonomic practice. It
may be useful at this point to say a word or two on the basic
method of the application of concepts. In particular it must be
emphasized that a concept is not necessarily invalidated if it
cannot be applied in an individual case. The concept tree is un-
questionably valid, yet one may have doubts whether or not to
include in this concept such plants as a spreading juniper, a
dwarf willow, a giant cactus, and a strangler fig. Some of our
most universally accepted concepts encounter the same diffi-
culties as the species concept, namely, borderline cases or insuf-
ficient information. Child and adult are two concepts which are
not invalidated by the fact that the adolescent is a borderline
stage. Father is a completely valid concept, but its application
sometimes encounters difficulties as is evident in paternity suits.
The student who attempts to apply the species concept to con-

E. MAYR 373

crete situations in nature faces, as all the speakers have empha-
sized, numerous difficulties. At first sight there appears to be a
bewildering diversity of perplexities, but these can be classified
into some major groups as was shown by Grant (this sym-
posium), and a study of the various classes of difficulties helps
considerably in an understanding of the species problem. Basi-
cally all obstacles in the application of a biological species con-
cept are due either to a lack of pertinent information on some
essential property of the investigated material or to its evolu-
tionary intermediacy.

Lack of Information

Different kinds of information are needed to permit the cor-
rect assignment of individuals to species. Most commonly the
question arises whether certain morphologically rather distinct
individuals belong to the same species or not. The long list of
synonyms, characteristic for some groups of animals and plants,
are a concrete expression of this difficulty (Mayr, Linsley, and
Usinger, 1953). This difficulty is particularly acute in three
branches of animal taxonomy. In many families of insects, par-
ticularly hymenoptera, males and females are so different that
separate classifications for males and females have to be adopted
until the proper associations have been made. Even more dif-
ferent are the larval stages in many groups of insects and aquatic
organisms. The same is true for parasites for which it is likewise
sometimes necessary to have two sets of names, one for larval
and one for adult stages, until association can be demonstrated
through elucidation of the life cycle. The difficulties caused by
sexual dimorphism, age differences, or nongenetic habitat differ-
ences which the neontologist faces in his work must be empha-
sized because some taxonomists seem to believe that paleontolo-
gists are the only ones who have to cope with the difficulty of
having to draw inferences from morphological types. That this
difficulty is particularly acute in paleontology no one will deny.
The worker who finds two or more essentially similar, yet some-
what different, morphological types in a single sample of fossils
is forced to make a somewhat arbitrary decision whether to con-

374 DIFFICULTIES AND IMPORTANCE OF THE CONCEPT

sider them as variants within a single interbreeding population
or rather as several similar species. There is no automatic solu-
tion. The splitting of every bimodal curve into two morphological
species would lead to the separation of males and females or of
age classes into separate species and to other equally unbiological
conclusions. Sylvester-Bradley (1956) and Imbrie (this sympo-
sium) have pointed to the fallacy of this approach. "The paleon-
tologist who classes members of a single interbreeding commu-
nity in more than one species is liable to confuse even himself
when he refers to hybridism. The distinction, in fact, between
morphological species and bio-species can only be overlooked at
the peril of utter confusion." Paleontologists cannot afford to for-
get that fossils are nothing but the remains of formerly living
organisms and that these organisms when still alive occurred in
the form of genetically defined populations exactly as the species
still living today. It is these populations which the paleontologist
attempts to classify with the help of fossil remains, and morpho-
logical criteria are used merely as a means to an end. Paleontolo-
gists classify their material on the basis of inferences. The better
they understand the nature of biological populations, the rela-
tions of genotype and phenotype, and the results of developmen-
tal physiology, the more skillfully they can interpret the available
morphological evidence. There is no justification for abandoning
the biological approach merely because it is sometimes difficult
to decide whether or not several morphological types in a popu-
lation are eonspecifie.

A second type of difficulty is introduced when there is insuffi-
cient information on the reproductive isolation of populations
that are not in contact with each other. Some paleontologists
have insisted that the classification of samples of discontinuous
vertical sequences introduces a new element in the evaluation of
populations, not appreciated by the neontologist. This is not alto-
gether true. The taxonomic, genetic, and ecological problems of
l lie multidimensional species arc quite the same whether one
deals with a series of populations in a chronological series, as
docs the paleontologist, or with geographically isolated popula-
tions, as does the neontologist. This has been recognized cor-

E. MAYR 375

rectly by the paleontologist Gwynn Thomas ( 1956 ) : "The essen-
tial similarity of temporal and spatial variation in fossils also
makes the paleontological species concept for practical purposes
almost identical with the neontological in fragmentary lineage
segments." In either case we are dealing with isolated samples,
and in either case we have to base a somewhat arbitrary decision
as to species status on a good deal of indirect evidence. Again the
difficulty is not with the yardstick but with its application.

There is a third group of cases where we lack information on
the reproductive isolation of individuals, namely, all forms of
asexuality. However, in this case more is involved than mere
lack of information because here we are not dealing with popu-
lations. This then is a much more formidable and more funda-
mental obstacle to the application of the species concept, and the
discussion of this difficulty shall therefore be postponed to a later
section. (See below.)

Evolutionary Intermediacy

A species definition postulates at least a discontinuity and,
under the most favorable conditions, a triad of characteristics:
reproductive isolation, ecological difference, and morphological
distinguishability. However, evolution is a gradual process and
so is the multiplication of species. As a consequence, there are
many instances where a population is on the borderline and has
acquired some but not yet all the attributes of a distinct species.
The gradual nature of the speciation process raises the following
difficulties for the student of multidimensional species.

Evolutionary Continuity. Species that are widespread in space
or time may have terminal populations which satisfy the criteria
of distinct species, yet are connected by an unbroken chain of
populations. Dr. Moore has described such a situation for con-
temporary populations of Rana pipiens; they are indeed not rare
in polytypic species. Theoretically this should be the standard
condition in paleontology, yet one has difficulty finding such
cases in the literature because breaks in the fossil record are suffi-
ciently frequent to prevent the piecing together of unbroken
lineages. One of the best substantiated ones is that of Micraster

376 DIFFICULTIES AND IMPORTANCE OF THE CONCEPT

discussed by Imbrie ( this symposium ) . On the other hand, there
are conspicuous gaps in most of the celebrated cases of so-called
unbroken lineages such as for instance Brinkmann's Kosmoceras.
In some cases of real continuity between morphological species
the differences are so slight that most contemporary neontologists
would not hesitate to consider these forms merely subspecies of
a single polytypic species. Even though the number of cases that
cause real difficulties is very small, the fact remains that an ob-
jective delimitation of species in a multidimensional system is an
impossibility.

Acquisition of Reproductive Isolation without Equivalent Mor-
phological Change. The reconstruction of the genotype which
is responsible for the reproductive isolation between two species
takes place sometimes without visible effect on the phenotype.
The resulting "sibling species" qualify as biological species, but
not as morphological species. Sonneborn has shown that the
varieties of Paramecium belong in this class. In plants most sib-
ling species appear to be allopolyploids. Where such sibling
species clearly differ in their ecology, nothing is gained by ignor-
ing them merely because the morphological difference is slight.
On the other hand, where morphological and ecological differ-
ences are not discernible, as seems to be the case in some in-
stances of allopolyploidy in plants or where the differences can
be established only by breeding in the laboratory such as the
varieties of Paramecium, it would seem impractical to separate
these forms as species in routine taxonomic work.

Strong Morphological Differences without Reproductive Isola-
tion. A number of genera of animals and plants are known
where even strikingly different populations interbreed freely
wherever they come in contact. Grant (this symposium) has
discussed this situation in detail and has given numerous illus-
trations. The attitude of calling every morphologically distinct
population a species, which was widespread among classical
taxonomists, is definitely losing ground. Yet to combine all mor-
phological species that freely hybridize in zones of contact also
leads to absurdity. Full agreement as to where to compromise
between these two extremes has not yet been reached. Such sit-

E. MAYR 377

nations occur also in the animal kingdom as, for instance, in the
snail genus Cerion (Mayr and Rosen, 1956). When the morpho-
logical differences between such populations far exceed those
normally found between good species in related genera, the
taxonomist is reluctant to unite them into a single species. These
cases are in a way the exact opposite to sibling species. The ac-
quisition of reproductive isolating mechanisms is lagging far
behind the general genetic divergence of these populations as
indicated by their morphological divergence.

Deficiencies in the Isolating Mechanisms. One, if not the most
important, attribute of a species is its possession of isolating
mechanisms. These isolating mechanisms are usually composite
and mutually reinforcing, but often they maintain only partial
isolation between populations. This becomes evident when a tem-
porary isolation between geographical isolates breaks down, re-
sulting in allopatric, or if the isolation was primarily ecological,
in sympatric hybridization. Numerous cases are known where
natural populations acted toward each other like good species in
areas of contact as long as their habitats were undisturbed. Yet
when the characteristics of these habitats were suddenly changed
in a drastic manner, usually by the interference of man, the re-
productive isolation broke down. Before this breakdown every-
one would have agreed that these populations were species, but
after the breakdown and after the loss of reproductive isolation,
they agreed better with the specifications of conspecific popula-
tions.

The frequency of hybridization, that is, the susceptibility to a
secondary breakdown of partial reproductive isolation, is very
different in different groups of animals and plants. In most
higher animals hybridization is sufficiently rare (or else the hy-
brids sterile) not to cause any serious difficulties in species
delimitation. In cases like the snail genus Cerion, discussed
above, it is a major source of difficulty. Grant has discussed in
detail ( in this symposium ) the effect of this phenomenon on the
species problem in plants and has supplied quantitative data
indicating its relative importance in different genera. It is obvious
that the delimitation of species becomes a serious problem in

378 DIFFICULTIES AND IMPORTANCE OF THE CONCEPT

genera where species hybridize freely with each other and
where introgression is a major factor.

The role of isolating mechanisms for the species problem is
brought into sharp focus by a consideration of hybridization.
Moore has paid special attention to this side of the species prob-
lem in his discussion of the origin of isolating mechanisms (this
symposium). He concludes that they originate as a consequence
of the genetic differences which accumulate among isolated pop-
ulations during adaptation to local conditions. This thesis seems
well substantiated and corresponds indeed to my own analysis
of the situation. Yet part of his evidence may have to be inter-
preted in a different manner, namely, the significance of the
climatic races of Rana pipiens. Moore has shown that the various
geographic races of this species are adapted in their embryonic
development to prevailing local water temperatures. When indi-
viduals of a cold-adapted race are crossed with ones of a warm-
adapted race, a more or less inviable hybrid will result. Moore
concludes from this "that the northern and southern forms have
developed isolating mechanisms, in reference to each other, to
such a degree that they could coexist and remain distinct." I am
not convinced that this conclusion is warranted. It is quite pos-
sible, if not probable, that much of this local adaptation is purely
ecotypic and essentially reversible. If a warm climate race would
reinvade a cool climate, its developmental rates and temperature
tolerance would have to be modified by selection to permit sur-
vival in the cooler waters. If it should subsequently come into
contact with a local cool water race, it might hybridize with it
and produce harmonious viable zygotes. That tin's is not pure
speculation is indicated by the Rana pipiens population in the
mountains of Costa Rica, which in the cool waters has acquired
the developmental properties of Vermont frogs and when crossed
with them produces normal viable zygotes. Differences that are
purely ecotypic local adaptations arc not necessarily good isolat-
ing mechanisms because they tend to disappear as soon as the
environmental differences disappear. This does not preclude the
possibility of other incipient isolating mechanisms in previously
isolated populations.

E. MAYR 379

Different Levels of Speciation in Different Local Populations
of the Same Polytypic Species. The amount of genetic diverg-
ence may differ in various isolated populations of a polytypic
species. This inequality in the level of speciation takes different
forms. Particularly spectacular are the cases of circular overlap,
an increasing number of which are cited in the literature. Other
cases are sympatric species, which are completely distinct at cer-
tain localities but hybridize freely at others. Lorkovicz ( 1953) has
suggested broadening the term semispecies (Mayr, 1940) to
include all isolated populations which on the basis of some cri-
teria have reached species level, but not on the basis of others.
Grant (this symposium) has made a similar suggestion to
broaden the application of the term semispecies. All these diffi-
culties demonstrate the fact that the clear-cut alternatives of the
nondimensional situation are absent in a multidimensional sys-
tem.

The species indicates a discontinuity above the level of the
individual, but a new difficulty is introduced by the fact that
there may be several such discontinuities, not all of them being
species. If we designate as isolate any more or less isolated popu-
lation or population segment, we can distinguish in sexually
reproducing organisms between geographical, ecological, and
reproductive isolates of which only the latter are species. Among
asexually reproducing organisms every clone and, in fact, every
individual is an isolate. Here, obviously, the species and the iso-
late can even less be synonymized with each other than in
sexually reproducing organisms.

Asexuality and the Species Problem

The essence of the biological species concept is discontinuity
due to reproductive isolation. Without sexuality this concept
cannot be applied. Asexuality then is the most formidable and
most fundamental obstacle of a biological species concept. In
truly asexual organisms there are no "populations" in the sense
in which this term exists in sexual species nor can "reproductive
isolation" be tested. Students of the species problem have neg-
lected asexual situations for a number of reasons. The geneticist

380 DIFFICULTIES AND IMPORTANCE OF THE CONCEPT

is interested only in sexual species because it is the recombining
of genetic factors through the sexual process which permits
formal genetics. The only event of genetic interest that happens
in asexually reproducing organisms is an occasional mutation
the effects of which will be invisible unless it is dominant or the
organism is haploid. Absence of sexuality in existing organisms
is almost certainly a secondary, derived phenomenon (Dough-
erty, 1955) and consequently does not require the setting up of
a primarily different species category. Finally, widespread though
it is in certain groups of organisms, asexuality is an exception
rather than the rule, and species can be defined and delimited
in most groups of organisms without any reference to loss of
sexuality.

What can one do if one wants to apply the species concept
to asexual organisms in view of the breakdown of the usual
criteria? A number of different solutions have been proposed.
The first is to find a species definition which would be equally
suitable for sexual and asexual organisms. Du Rietz (1930)
thought that the following definition was satisfactory: "The
smallest natural populations permanently separated from each
other by a distinct discontinuity in the series of biotypes are
called species/' This forgets that asexual organisms do not form
"natural populations and that in asexual organisms every in-
dividual and every clone is such a distinct biotype. The clear
realization of this difficulty has led some authors to go one
step farther and abandon the biological concept altogether, be-
cause of its inapplicability to asexual situations. Frankly, it ap-
pears to me that there is nothing to recommend this solution. In
exchange for the biological species with all its advantages, it
reintroduces the morphological species with all its weaknesses
pointed out by most of the speakers at this symposium.

A second solution is to restrict the term species to sexually
reproducing organisms and use it only in the sense ol biological
species. This proposal comes closer to a satisfactory solution, but
it still leaves open the question how to classify morphologically
differing individuals in asexual organisms. It has been suggested
to use for them a neutral term, such as the term binom, men-

E. MAYR 381

tioned by Grant (this symposium). The suggestion overlooks the
fact that the word species has not only the biological meaning of
a reproductively isolated population but also the purely formal
meaning "kind of," simply a classifying unit. The term "agamo-
species" has been used to designate "totally asexual populations"
but, as stated above, an "asexual population" is a biological
impossibility.

The most satisfactory solution in taxonomic practice has been
a frankly dualistic one. It consists in defining the term species
biologically in sexual organisms and morphologically in asexual
ones. There is more justification in this procedure than a mere
pragmatic one. The growing elucidation of the relations between
genotype and phenotype also justify this approach. Reproductive
isolation is effected by physiological properties which have a
genetic basis. Morphological characters are the product of the
same gene complex. Once this is clearly understood, a new role
can be assigned to morphological differences associated with re-
productive isolation, namely that of indicators of specific dis-
tinctness. This permits the assumption that the amount of genetic
difference which, in a given taxonomic group, results in reproduc-
tive isolation will be correlated with a certain amount of mor-
phological difference. If this is true, it is permissible to conclude
from the degree of morphological difference on the probable
degree of reproductive isolation. To base this inference on genetic
reasoning is new; the method itself, however, of determining
empirically with the help of morphological criteria whether or
not a population has reached species status goes back to classical
taxonomy. This inference method is by no means a return to a
morphological species concept since reproductive isolation always
remains the primary criterion and degree of morphological differ-
ence only a secondary indicator, which will be set aside whenever
it comes in conflict with the biological evidence.

It is possible to use the same kind of inference to classify
asexual organisms into species. Those asexual individuals are in-
cluded in a single species that display no more morpholog-
ical difference from each other than do conspecific individuals
in related sexual species. Criteria must be adjusted to individual

382 DIFFICULTIES AND IMPORTANCE OF THE CONCEPT

situations since there is great diversity in the forms of asexuality.
Where sexuality is abandoned only temporarily, as in the cases
of seasonal parthenogenesis in Daphnia and aphids, there is no
problem. It is customary and biologically sound to consider such
temporary clones as portions of the total gene pool of a species.
Nor is there any major difficulty where sexuality is lost in a
single species of a genus or in a number of lines which can be
clearly traced back to a common ancestor and where the mor-
phological differences are still slight enough to justify combining
these "microspecies" into a single collective species (Mayr, 1951).
By far more difficult is the situation in many microorganisms
where sexual reproduction or its genetic equivalents are totally
absent in large groups or at best highly sporadic. To include all
descendants of a common ancestor in a single collective species
is also impossible in groups like the bdelloid rotifers which, ap-
parently without sexual reproduction, have grown to an order
with four families, some twenty genera, and several hundred
"species" all reproducing partheuogenetically and possibly all
descendants from a single ancestor. If such a group were a
complete morphological continuum, any attempt to break it up
into species would be doomed to failure. Curiously enough there
seem to be a number of discontinuities which make taxonomic
subdivision possible. The most reasonable explanation of this
phenomenon is that the existing types are the survivors among
a great number of produced forms, that the surviving types are
clustered around a limited number of adaptive peaks, and that
ecological factors have given the former continuum a taxonomic
structure. Each adaptive peak is occupied by a different "kind"
of organism, and it is legitimate to call each of these clusters of
biotypes a species.

The large list of difficulties which the application of the species
concept faces may seem to confirm the opinion of those who
consider the species as something purely subjective and arbitrary.
To counterbalance this impression it must be emphasized (a)
that none of these difficulties of application invalidates the three
basic concepts of the species; (b) that these difficulties are in-

E. MAYR 383

frequent in most groups of animals and higher plants and that
the frequency and significance of their occurrence can be deter-
mined rather accurately; (c) that such difficulties are usually of
minor importance in nondimensional situations which are those
most frequently encountered by the taxonomist; and (d) that in
spite of these difficulties it is usually possible to classify doubtful
entities into taxonomic species which satisfy at least one or the
other species concept. To use the words of G. G. Simpson ( 1943) :
"A taxonomic species is an inference as to the most probable
characters and limits of the morphological species from which a
given series of specimens is drawn."

The Practical Importance of Species

Those who maintain that species are something purely sub-
jective, vague, and arbitrary sometimes ask: "What is gained by
recognizing species?" The answer is that the attempt to determine
species status has led in many cases not only to a more precise
formulation of a biological problem but very often also to its
solution. As an illustration I will call attention to only three areas
of biological research where this is true.

The Clarification and Simplification of Classification. The
need for classifying the morphological diversity of nature into
biological species forces an unequivocal decision how to handle
morphological variants. Accepting the biological species concept
no longer permits either describing all variants as morphological
species or listing the more pronounced ones as species and the
less distinct ones as varieties. Now only those variants are given
species rank which satisfy the criterion of the biological species
concept, namely, reproductive isolation. The result is a simplifica-
tion of classification which is not only of practical help to the
working taxonomist but also actually aids the understanding of
distribution, ecology, and phylogeny. In ornithology it has per-
mitted a reduction in the number of species from nearly 20,000
to 8,600, and a similar simplification is apparent throughout zool-
ogy. There are still some authors who resent having to make

384 DIFFICULTIES AND IMPORTANCE OF THE CONCEPT

such decisions, some of which by necessity will be incorrect, and
who prefer to classify all organisms into meaningless but com-
parable morphological species. It is quite evident that these
workers fight a losing battle because biologically trained tax-
onomists are unhappy to be degraded into pebble sorters. They
would much rather make an occasional mistake than be burdened
forever with a multitude of purely morphologically defined pi-
geon holes.

Fossil Species. The need to evaluate fossil specimens in terms
of biological species has led and is leading to a new outlook in
paleontological classification (Sylvester-Bradley, 1956). It forces
the paleontologist to make clear decisions: Different specimens
found in the same exposure (the same sample) must be either
different species or intrapopulation variants, "for two geograph-
ical subspecies cannot come from the same locality, and two
chronological subspecies cannot come from the same horizon"
( Sylvester-Bradley, op. cit. ) . The recognition of subspecies and
polytypic species in paleontology leads to the same simplification
and greater precision as it has in neontology. In all this work the
evidence is largely morphological, but the interpretation is based
on biological concepts. An occasional error of interpretation in
the synthesis of polytypic species in paleontology is vastly to be
preferred to the chaotic accumulation of morphologically defined
entities without biological meaning.

Biological Races. The shift of emphasis from morphological
difference to reproductive isolation has necessitated a reanalysis
of the whole complex of phenomena loosely referred to as "bio-
logical races." These are of great practical importance since most
of these so-called biological races were found during the study
of disease vectors or of injurious animals. Here again the need
for a clear decision, species or not, has led to clarification and sim-
plification. The study of sibling species which has been the major
outcome of this analysis has been, in many cases, ol great prac-
tical importance in applied biology.

The species is, however, of more than purely practical impor-
tance. It has a very distinct biological significance which has

E. MAYR 385

been described above (Introduction). Finally, a new field of
research is growing around the species.

The Science of the Species

It has become apparent within recent years that a new branch
of biology is developing, the science of species. It is devoted to
the many-sided aspects of the species and to the understanding
of the level of integration which is denoted by the term species.
This branch of biology is as legitimate as is cytology, devoted to
the level of the cell, or histology, dealing with the level of the
tissue. The species is an important unit in evolution, in ecology,
in the behavioral sciences, and in applied biology.

The study of the biological properties of species has within
recent decades led to the development of several new fields in
biology, each being a borderline field between the science of
species on the one hand and some other branch of biology, e.g.,
genetics, ecology, animal psychology on the other. It would lead
too far to paint a detailed picture of this recent development, and
I will content myself to outline it with a few bold strokes. These
new fields are:

The Study of Speciation. Darwin, as we saw, had fallen down
in his attempt to explain the multiplication of species because he
was not fully aware of the multiple aspects of species. Nor did
he realize that multiplication of species, an origin of discontinui-
ties, is not the same as simple evolutionary change. The study
of the mechanisms and modes of speciation has become an inter-
esting borderline field between systematics, genetics, and evolu-
tion.

The Study of Isolating Mechanisms. The appreciation of the
fact that in addition to sterility there are many other factors
which safeguard the genetic integrity of species has led to the
development of a new field of study. As far as animals are con-
cerned, contact was established with the field of animal psychol-
ogy because the ethologieal isolating mechanisms are among
the most conspicuous manifestations of animal behavior. The
study of courtship behavior has importantly contributed to the

386 DIFFICULTIES AND IMPORTANCE OF THE CONCEPT

concepts of Lorenz and Tinbergen and has led to a reinterpreta-
tion of many facts which Darwin had grouped together under
the heading "Sexual Selection."

The Study of Intraspeciftc and Interspecific Competition. The
realization that there is a subtle difference between the compe-
tition that goes on among conspecific individuals and that among
individuals belonging to different species has had a very stimulat-
ing effect in the field of ecology. That interspecific competition
is an important centrifugal factor in evolution has been dem-
onstrated by various recent authors. Attempts to determine and
to measure such competition more precisely have led to a much
more detailed study of the ecology and the population structure
of species than was attempted by the ecologists of preceding
decades.

Study of the Genetic Structure of Species. The realization
that local populations belong to a broader system, the species,
and that the total genetic content of these intercommunicating
populations are the gene pool of the species has had a profound
effect on population genetics. It has led to new questions and
has facilitated the understanding of certain intrapopulation
phenomena. Studies of the balance of the gene complex, of the
balance between local adaptation and compatibility with gene
dispersal and of hybridization between species have led to a
broadening of genetics which transcends considerably the ge-
netics of the early Mendelian period.

The Study of Physiological Species Differences. In physiology
a development is taking place which parallels that in genetics. It
involves the realization that the ecotypic physiological differ-
ences of local populations are of a different order of magnitude
from that existing between good species. Each species is a sep-
arate physiological system, and these physiological differences are
of different kinds as pointed out by Prosser (in this symposium).
With physiological characteristics of populations, special care
must be exercised to distinguish between genetically controlled
characters and others. Prosser lias shown that nongenetic ac-
climatizations are particularly frequent in marine animals with

E. MAYR 387

pelagic larval stages. Such organisms must be capable of coping
with the water conditions in the area where they settle. It would
be a mistake, however, to conclude from this special case that
all infraspecific population differences of physiological characters
are nongenetic. There is not only the excellent work of the
Stanford group (Clausen, Hiesey, and Keck) and their associates,
which clearly establishes the genetic nature of physiological
differences between local races of plants, and similar results have
been obtained elsewhere, but also a large body of fact demon-
strating geographic variation of physiological characters in ter-
restrial animals. The ultimate differences between species are in
part built up from such population differences. However, a dis-
tinction must be made between purely ecotypic local adaptation
to purely local conditions and a more general physiological
divergence which is almost always found in isolated populations
(Mayr, 1956). Physiological differentiation sometimes proceeds
more rapidly than morphological change, and this is the explana-
tion for the occurrence of sibling species in many groups.

The study of physiological adaptations and limitations is still
very much at the beginning. This seems to be one of the most
promising branches of comparative physiology. A study of the
physiological differences between species, particularly aquatic
species, with respect to rates of development, temperature adap-
tation, and temperature tolerance has already yielded most in-
teresting findings.

These are only a few indications of newly developing branches
of biology specifically dealing with the species level. They refer
to the development I had in mind when I said that we were wit-
nessing at the present time the growth of a "science of species."

Conclusion

The discussions of this symposium show clearly that species
are still a stimulating subject and that the species problem still
offers a challenge. There are many difficulties in applying the
species concept to the vast variety of discontinuities found in
organic nature. Yet many biological phenomena would make no

' 3 DIFFICULTIES AND IMPORTANCE OF THE CONCEPT

sense if individuals were not organized into populations, and
these populations into species. Species are an important bio-
logical phenomenon, important to every biologist because every
biologist works with species.

Note: For references, see those in the Introduction.

INDEX

Acclimation, 340 ff., 353, 361
Adaptation, 352

ecotypic, 378, 387

physiological, 340 ff., 349, 361
Adolph, E. F., 343, 364
Agameon, 61
Agamic complex, 70
Agamospecies, 61, 381
Agoseris, 47
Allophyllum, 47
Anadara, 135
Anderson, E., 71, 77
Anopheles, 351
Apomicts, 36
Archetype, 12
Arkell, W. J., 6, 19
Artemisia tridentata, 54
Asclepias, 53
Asexuality, 379 ff.
Asexual organisms, 283, 286, 288
Associations of species, 104, 115
Asterias, 350
Austin, M. L., 179, 315
Autogamy, 188

von Baer, K. E., 7, 19

Banta, A. M., 86, 121

Barrel, J., 143, 152

Bateson, W., 5

Beale, G. H., 182, 214, 235, 315

Behavior

relation to isolation, 359

reproductive, 343
Berg, L. S., Ill, 121, 122
Bessey, C. E., 4, 19
Beyrichia, 140-141
Binom, 46, 61, 380
Biological races, 384
Biological species, 160

characteristics of, 50

criteria of, 200

difficulties of concept, 57

in paleontology, 129

Biological units, 39
Birge, E. A., 83, 122
Brauer, F., 8, 19
Breeding system, 296

of Paramecium aurelia, 202, 206,
234 ff.

of Paramecium bursaria, 245

of Paramecium caudatum, 239

in Protozoa, 155-315

of Tetrahymena pyriformis, 267
Brett, J. R., 340, 365
Bufo, 337

Bullock, B. H., 125, 152
Burma, B. H., 6, 19, 44, 78, 125, 152
Butzel, H. M., Jr., 168, 315, 316

Cain, A. J., 10, 19

Calkins, G. N., 292, 316

Callinectes, 350

Camp, W. H., 10, 19, 44, 61, 78

Carbonicola, 128

Carson, H. L., 35, 37

Caryonide, 223

definition, 186
Categories

collective, 16
Ceanothus, 47, 53
Cerion, 377
Characters

neutral, 347

physiological, 348, 363
Chen,T. T., 241, 242, 316
Ciliates, 157 ff.

life cycle, 187
Cladocera, 104
Classification, 383
Cleveland, L. R., 280, 316, 317
Clines, 355
Clones, 159, 223
Coenospecics, 65
Colbert, E. H., 133, 134, 152
Colpidium truncatum, 277
Commiscuum, 65

389

390

INDEX

Comparium, 65
Competition

interspecific, 386
Concepts, 10 ff.

application of, 132, 372

biological, 42, 45, 57, 74, 126,
132, 155, 296, 371

conflicting, 6

critique of, 42

difficulties of, 371-383

evolution of, 41

genetic, 16

and hybridization, 62

importance of, 383-388

morphological, 126

multidimensional, 16

of naturalist, 41

nominalistic, 44

nondimensional, 14-16

philosophical, 11

typological, 11, 42, 43, 45, 126
Conjugation

definition, 188
Conodonts, 125
Constancy, 2, 7
Continuity

evolutionary, 375
Copepods, 104
Coregonus, 110-120

hybridization in, 118

spawning behavior, 118
Coregonus albida, 111
Coregonus lavaretus, 111
Crassostrea, 354
C rep is, 70
Crinia signifera, 331
Criteria

biological, 9

serologic, 285
Crurithyris, 141
Cubitostrea, 145
Cuvier, C, 7
Cytoplasm, 185

Danser, B. H., 65, 78
Daphnia, 82-110

characters of species, 105
distribution of, 82

helmet, 97

hybridization, 106 ff.

life cycle, 85

phenotypic variation, 90 ff .

reproduction modes, 86
Daphnia dubia, 106
Daphnia galeata mendotae, 106
Daphnia longircmis, 99 ff.
Daphnia longispina, 83
Daphnia middendorffiana, 107, 108
Daphnia parvula, 98
Daphnia pulex, 83, 84, 107, 108
Daphnia retrocurva, 93-99
Daphnia sch0dleri, 90, 91, 107
Darlington, C. D., 30, 37
Darwin, C, 3, 4, 7, 19, 325
Davidson, J. F., 44, 65, 78
Definitions, 16

biological, 17, 46

secondary speciation, 60
De Gads, C. F., 158, 317
Delimitation, 51
Delineation

borderlines, 54

difficulties of, 53
Deme, 28
De Vries, H., 5
Differences

interspecific, 357

morphological, 12, 376

nongenetic, 363

physiological, 386
Difflugia, 286
Dimorphism

sexual, 373
Dippell, H., 167, 175, 177, 179, 189,

216,219,317,323
Discontinuities

kinds, 379

morphological, 47
Dispersal

consequences of, 218
Dobzhansky, T., 17, 18, 20, 26, 59,
72, 78, 283, 294, 296, 317, 326,
337
Dougherty, E. C, 19, 380
Drosophila, 35, 356

behavior, 360

INDEX

391

Vrosophila persimilis, 57, 336
Drosophila pseudoobscura, 57, 336
Vrosophila robust a, 35
Du Rietz, G. E., 5, 17, 20, 51, 78,

380

Eagar, R. M. C., 125, 128, 129, 152

Eidos, 11

Eimer, G. H. T., 15, 20

Elliott, A.M., 256, 317, 318

Embryologist's viewpoint, 325

Engler, A., 42

Environment

influence on freshwater animals,
97
Enzvmes, 361
Euplotes, 271-276, 351

breeding system, 274

mating hormones, 273

mating types, 273

species and varieties, 271
Euplotes patella, Til

Fish, 356

hybridization in, 110
Flagellates, 280-283

sexual processes, 280 ff .
Fossil animals, 125-151, 384

application of biological concept,
132

incompleteness of record, 142

size-frequency distribution of pop-
ulation, 141

species, 384
Frequency

of good species, 52
Freshwater animals, 81-123
Frog, see Rana

Gammarus, 350

Gasterosteus, 351

Gene pool, 14, 26

Gene recombination, 23-37; see also

Recombination
Genes

drift, 360

flow, 32

Genotype, 24

Ci.sc, A. C, 190-192,209,318

Cilia, 48-50, 62-64

Gilia capitata, 62

Cilia leptantha, 47

Gilia millefoliata, 62

Gilia tenuiflora-latiflora, 63

Gillis, C. L., 10, 19

Gill rakers, 112, 117

Gilman, L. C., 236, 239, 240, 318

Ginkgo biloba, 65

Ginsburg, I., 12, 20

Gloger, C. L., 9, 20

Godron, D. A., 7, 20

Grant, V., 372, 376, 381

Greene, E. L., 3, 20

Gregg, J. R., 6, 20

Groupings, 42, 67

Gruchy, D. F., 256, 318

Habituation, 358
Haploidy, 202
Hayes, B. E., 257, 317
Hermaphroditism, 70
Heuts, M. J., 351, 366
History of problem, 1
Hiwatashi, K., 236, 240, 3*18, 319
Hoare, C. A., 283, 319 ^
Homogamic complex, 67
Hooijer, D. A., 133, 134, 152
Hubbs, C. L., 110, 122
Human species, 34
Hybridization, 32, 55, 59, 62, 70-72,
106, 118, 327, 377

in fishes, 110

introgressive, 110, 112

reasons for, 74
Hybrids, 330

methods of identification, 58

sterility of, 44

Identification

methods of, for hybrids, 58
Illiger, J. C. W., 8, 20
Inbreeders, 206 ff.
Inbreeding, 297 ff .

indications of, 220 ff.

obligatory, 280

392

INDEX

Intermediaey

evolutionary, 375
Introgression, 55, 109, 119
Isolation

agents, 359

mechanisms, 72, 167, 232 ft., 326,
330, 332 ft., 359, 377, 378, 385

microgeographic, 235

relation to behavior, 359

reproductive, 15, 47, 374, 376,
377

Jennings, H. S., 158, 159, 233, 241,
245, 248, 250, 253, 286, 319,
320

Jordan, K., 5, 15, 17, 20

Karyotype, 216
Kermack, K. A., 141, 148-152
Kimball, R. F., 252, 271 ft., 320
Kinne, O., 351, 366
Kiser, R. W., 84, 122
Koopman, K. F., 336, 337
Kosmoceras, 376
Kotaka, T., 135, 152
Kurten, B., 141, 152

Lederberg, J., 285, 320
Leitch, D., 128, 152
Levine, M., 167, 320
Linnaean species, 42
Linnaeus, 2, 3, 15
Lorkovicz, Z., 20, 379
Lotsy, J. P., 42, 67, 79

Margolin, P., 180,320

Mason, H. L., 44, 60, 79

Mating

Paramecium aurelia, 164
reactions, intervarietal, 165
types, 158, 164

Maupas, E., 157, 187,821

Mavr, E., 10, 17, 21, 50, 08, 79, 80,
289, 821, 826, 337, 382, 387

McKerrow, W. S., 138, 153

Meadow lark, 860

Meglitsch, P. A., 10, 14,21

Melvin, [. B., 167,321

Mendelian population, 26
Merging

of species, 60
Metz, C. B., 242, 321
Micraster, 141, 147-151
Micronuclei, 178
Microspecies, 70, 382
Moore, J. A., 378
Morphological concept, 126
Morphological differences, 12, 376
Morphological types, 373
Munz, P. A., 53,' 77
Mustela

comparison of osteological charac-
teristics, 133
Mytilus, 350, 352, 353

Nannev, D. L., 186, 227, 229, 251,

267, 321
Nereis, 350
Neurospora, 348
Newell, N. B., 125,

153
Niggli,P., 11,21

129, 143, 144,

Objectivity, 2
Oeder, G.L., 7
Oenothera, 56
Oken, L., 9
Opuntia, 56, 70
Ornithella, 139
Outbreeders, 206 ft.
Outbreeding, 297 ft., 347

indications of, 220 ft.
Overlaps, 48

circular, of races, 50
Oxytrichia bijaria, 277 ft.

breeding system, 278
Oysters (Crassest rca) , 354

Paleontology, 127

biological species in, 129

typological, 127
Paramecium, 158 (1.
autogamy, 188
characters of species, 161
conjugation, 188

INDEX

393

Paramecium (continued)

isolating mechanism, 167

killer strains, 179

life cycle, 187

macronuclear differentiation, 227

sexual processes, 203
Paramecium aurelia, 162-236

breeding system, 202, 206, 234 ff.

characteristics, 197

conjugation, 188

cytoplasmic exchange, 224

distribution of varieties, 169

evolution of groups A and B, 226

fission rate, 176

geographical distribution, 231

geographical variation, 216

identification of varieties, 173

immaturity, 190

inheritance of mating types, 185

karyotypes, 216

killers, 234

life cycle differences, 187

mating types, 163 ff., 227

maturity, 191

morphological differences, 177

mortality of intervarietal crosses,
172

races, 235

senility, 193

serotypes, 180, 230

taxonomic status of varieties, 195

temperature range, 171

temperature tolerance, 175

varieties, 15, 16, 163 ff., 173, 315

varieties as species, 199
Paramecium bursaria

breeding system, 245

characteristics of varieties, 243

distribution of varieties, 243

karyotypes, 244

mating types, 241, 244, 250

reproductive isolation, 242

self-differentiation, 248

speciation, 256
Paramecium calkinsi, 162, 276
Paramecium caudatum, 236-241

breeding system, 239

varieties, 236, 237

Paramecium multimil innuclcatum,

161, 163,314
Paramecium tricJuum, 162, 276
Parasites, 373
Parkinson, D., 142, 153
Parthenogenesis, 86
Pennak, R. W., 84, 122
Phenotype, 24, 360

effect of temperature, 97

plasticity of, 114
Phylogenetic tree, 130
Physiologist's viewpoint, 339-364
Physiology

comparative, 287
Pinus muricata, 60
Pinus remorata, 60
Pittendrigh, C. H., 352, 367
Plant species, 39-77
Plate, L., 8, 21
Plato, 11
Polyploidy, 54
Populations

competition between, 334

local, 27, 31

Mendelian, 26

recombination in, 29

size-frequencv distribution of fos-
sils, 141

in space, 32

in time, 32
Poulton, E. B., 2, 21
Preer, J. R., Jr., 179, 182, 322
Pringle, C. R., 182, 208, 210, 211,

238, 322
Prosser, C. L., 386
Properties, 14
Protozoa, 155-315

asexual, 280, 299

reproductive methods, 155-315

variety in, 160 ff.

Races

biological, 384

osmotic, 350

physiological, 347, 353
Bamsbottom, J., 3, 21
Rana, 328 ff.
Rana pipiens, 329, 355, 375, 378

INDEX

Rate functions, 343
Rawson, D. S., 103, 122
Ray, C, Jr., 268, 322
Ray, J., 7, 15, 21, 41
Recognition of species, 105
Recombination, 23-37, 202

amount of, 33

in man, 34

systems, 33, 36
Redi, F., 3

Regan, C. T., 127, 153
Regulation, 342
Rensch, R., 17, 21
Reproduction

asexual, 36, 56, 61, 70, 202

sexual, 203

uniparental, 88
Rhizomys, 134
Rhizopods, 286
Robinet, C, 3
Rotifers, 104

bdelloid, 382
Rowe, A. W., 147, 153

Sagebrush, 55
Salmonidae, 110
Schopenhauer, A., 3, 21
Self-pollination, 56
Semispecies, 68, 379
Senility

of Paramecium aurelia, 193
Serotypes, 180, 359
Sexual dimorphism, 373
Sexuality

loss of, 281
Sibley, C. C, 68, 80
Sibling species, 13, 54, 57, 69, 291,

293, 348, 359, 376, 384
Sicgel, R. W., 179, 191, 234, 277 (1.,

322
Simpson, G. C, 14, 15, 17, 21, 125,

127, 134, 153, 383
Sirks, M. T-, 3, 22
Sonneborn, T. M., 251, 322, 323
Spallanzani, L., 3
Spathidium spathula, 292
Speciation, 69, 256, 346, 37"). 385

allopatric, 49, 326, 344

in asexual organisms, 286

ecotypic, 387
Species

boundaries of, 150

genetic structure of, 386

multidimensional, 375

nondimensional, 131

secondary, 60

successional, 131, 147 ff.

typological, 127
Spencer, H., 292, 324
Spjeldnaes, N., 140-141, 153
Spontaneous generation, 2
Spooner, G. M., 351, 368
Stebbins, G. L., Jr., 36, 38, 72, 80
Stenzel, H. R., 145, 153
Sterility, 72
Strains

amicronucleate, 263

killer, 179, 234
Stresemann, E., 17, 22
Stress, 340
Stropliodonta, 141
Stylonycliia putrina, 211
Superspecies, 50, 67
Survival, 340

Svardson, G., 111-120, 122, 123
Sylvester-Rradley, P. C., 125, 129,

153, 374, 384
Sympatry, 49
Synclone, 223

Syngameon, 64, 66, 67, 71, 76
Syngen, 201, 289, 295

asexual, 290
Systematics

comparative, 372

Temperature, 97, 352

adaptations to, 328

tolerance, 175, 356
Tetrahymena pyriformis, 256-271

breeding systems, 267

characteristics, 258

life cycle, 261

mating types, 259

varieties, 257
Tetrahymena rostrata, 270
Thais emarginata, 354

INDEX

395

Thoday, J. M., 33, 38, 291, 323
Thomas, G., 375
Transitions between species, 59
Trueman, A. E., 129
Trypanosomes, 284
Turbulence, 97
Turesson, G., 65, 80

Uroleptus mobilis, 292
Urosalpinx, 355

Valentine, D. H., 10, 22
Variation

geographical, 98
nongenetic, 360
overlapping, 48
phenotypic, 90 ft., 112, 138
physiological, 340-343, 345, 349,
350, 352, 357, 359

polytypic, 43, 384

seasonal, 94
Variety, 4
Voigt, F. S., 9, 22
Volpe, G. P., 337, 338

Wagler, E., 83, 123

Ward, G. H., 55, 80

Weir, J., 127, 153

Westoll, T. S., 138, 153

White, M. J. D., 30, 38

Whitefish, see Coregonus

Wichterman, R., 162, 276, 323, 324

Wild type, 25

Woltereck, R., 99, 102, 123

Woodruff, L. L., 292, 324

Zirkle, C, 3