SVENSK BOTANISK TIDSKRIFT.

1930.

THE FUNDAMENTAL UNITS OF BIOLOGICAL

TAXONOMY.
BY:

G. EINAR DU RIETZ.
1.

Introduction.

Though the species is generally accepted by present-day biologists (some geneticists excepted) as the most fundamental unit of
biological taxonomy, certainly very few biologists of our days
would be able to give a clear definition of what a species really is.
The main causes to that vague species-concept certainly must
be sought for in historical facts. The evolution-theories of LAMARCK
(1802, 1809) and DARWIN (1859) had led to a general belief
in the gradual transformation of species. That Natura non
facit saltus was doubted by very few biologists of the later half
of the last century. The logical result was that species were
mostly regarded merely as arbitrary divisions of the continuous
and everchanging series of individuals found in nature. 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, and anyhow it did not much influence the general discussion. of the species-problem. Important
exceptions were formed by the works of WAGNER, GULICK and
ROMANES (comp. below pp. 393-394) as well as by DOLLO) 1893),
BATESON (1894) and some other authors. But first the rediscovery
of the Mendelian laws (1900) and DE VRIES' mutation-theory (1901)
brought the principle of discontinuity to its full right again in the
genetical discussion.
23 - 30531.

Svensk Bolanisk TidskTi/l. 1930.

334
The vagueness of the traditional species-concept has got a very
significant expression in the international rules of botanical nomenclature adopted by the international botanical congresses of Vienna
1905 and Brussels 1910. In those rules, no serious attempt at all
was made to define the species-concept. The extremely vague formulation of the articles dealing with the fundamental taxonomical
units is clearly shown by the following abstracts;
Art, 1O. Every individual plant belongs to a species (species),
every species to a genus (genus), every genus to a family (familia),
every family to an order (ordo), every order to a class (classis),
every class to a division (diuisio).
Art. 11. In many species we distinguish varieties (uarietas) and
forms ([orma), in the case of parasites special forms (forma specia lis) ; and in some cultivated species, modifications still more
numerous; in many genera sections (seclio), in many families tribes
(tribus).
Art. 1.2. Finally if circumstances require us to distinguish a
greater number of intermediate groups, it is easy, by putting the
syllable sub before the name of a group, to form subdivisions of
that group. In this way subfamily (subfamilia) designates a group
between a family and a tribe, subtribe (subtribus) a group between
a tribe and a genus, etc. The arrangement of subordinate groups
may thus be carried, for wild .plants only, to twenty-two degrees,
in the following order: Regnum vegetabile. Divisio. Subdivisio.
Classis. Subclassis, Ordo. Subordo. Familia. Subfamilia. Tribus.
Subtribus. Genus. Subgenus. Sectio. Subsectio. Species. Subspecies..
Varietas. Subvarietas. Forma. Forma specialis. Individuum.
If this list of groups is insufficient it can be augmented by
the intercalation of supplementary groups, so long as these do
not introduce confusion or error.
Example: Series and Subseries are groups which can be intercalated between subsection and species.
Art. 13. The definition of each of these names of groups varies,
up to a. certain point, according to individual opinion and the
state of the science, but their relative order, sanctioned by custom
may not be altered. 'No classification is admissible which contains
such alterations.
Examples of inadmissible alterations are, - a form divided into
varieties, a species containing genera, a genus containing families
or tribes.

335

Art. 14. The fertilization of one species by another gives rise

to a hybrid (hybrida); that of a modification or subdivision of a
species by another modification of the same species gives rise to
a hall-breed (mistus, mule of florists). (BRIQUET 1912 p. 33.)
This vague conception of the fundamental units is still found
in most taxonomical works (comp. DIELS 1921 and HITCHCOCK 1925).
The differences in i ndividual opinion mentioned in the rules
are constantly leading to divergences between different workers
regarding the right use of the various terms, the same unit being
treated as a species by one author, as a subspecies by another
one, as a variety by a third one, as a form by a fourth one, etc.
And certainly there is little hope to attain a better stability in
this- respect, as long as no better definitions than those of the
international rules have become generally accepted. Probably
most taxonomists think that just the i I' way of using the terms
concerned should be considered the right one -- but why,
probably very few of them would be able to tell. The situation
has grown still worse through the suggestion of various geneticists
to -replace the old terms of taxomomy with a lot of new ones
better defined. And certainly it is no exaggeration to say, that
the attaining of a stable and generally accepted system of well
defined terms for the designation of the various fundamental
units is one of the most urgent needs in present taxonomy.
The following attempt to a critical revision of the fundamental
units of taxonomy is founded upon a synthesis of about 20 years
active taxonomical field- (and museum-) work - mainly in lichens,
marine algae and vascular plants - with an extensive study of
taxonomical and genetical literature with a bearing upon this
problem. For many years I have occupied myself with the general
problems of taxonomical methodology (comp. Du RIETZ 1921 pp.
10-13, 1923 a pp. 235-241, 1924 a pp. 385-391, b p. 76, 1930
pp. 294-296). However, it is a great pleasure for me to acknowledge the most stimulating impulses I received during my recent
sojourn in New Zealand (1926-1927) from many discussions and
excursions with Dr. L. COCKAYNE and Dr. H. H. ALLAN, whose
recent work on those problems could not fail to attract my intense
interest. As I am going to take a good deal of the examples used
in the following discussion from the flora of New Zealand, I also
wish to use this opportunity to express my deep gratitude to
the New Zealand Government and to all my New Zealand botanical

336
friends - the names of which it is impossible to enumerate
. here - for the neverfailing interest they took in my efforts to
learn to know as much as possible of the flora and vegetation
of their wonderful country. - During most of the New Zealand
fieldstudies referred to below my wife, Mrs. GRETA Du RIETZ, was
an invaluable collaborator.

2. The taxonomical units as concrete populations.

For many years there has been much discussion among biologists regarding the concrete or abstract nature of the various units
of taxonomy. A very good summary was given in 1914 by PLATE
(pp. 115-118). Many biologists have completely denied the existence of any concrete units in taxonomy except the individuals,
while others have considered a II taxonomical units as concrete.
Between those two extremes there are numerous intermediate
gradations. A rather common compromise accepted also by PLATE
is the view that only s p e c i e s are concrete units, while all the
higher units are abstract. And many modern geneticists have
insisted upon the b i 0 t Ypes being the only concrete units existing
in nature.
The following words by PLATE probably express a rather common
view among those present day biologists, who have taken the
trouble of thinking about those problems (comp. also KLINGSTEDT 1928).
Die Worte Art, Gattung, Familie usf. haben einen doppelten
Sinn; sie bezeichnen einmal Beg I' iff e von G I' U P pen und sind
als solche wie alle Begriffe ahstrakt; sie bezeichnen abel' auch
die realen Objekte, welche diesen Begriffen zugrunde liegen, also
In d i v i due n k 0 m p l e x e, und solche Komplexe sind natiirlich
ebenso real, wie die Individuen, aus denen sie sich zusamrnensetzen. Diese Individuenkomplexe sind voneinander insofern verschieden, als derjenige Individuenkomplex einer Art vollig unabhiingig ist von del' menschlichen Analyse und daher objektiv in
del' Natur existiert, wahrend die iibergeordneten Individuengruppen
(Gattung, Farnilie usf.) nicht durch sich selbst, sondern nul' durch
den vergleichenden und nberlegenden Verstand des Menschen
gebildet werden. In diesem Sinne ist die Art etwas Reales, wahrend
die Gattung, die Familie, iiberhaupt die hoheren Gruppen abstrakter
Natur sind. (PLATE 1914 p. 118.)

337

In modern genetical literature species and their subordinate

units are often spoken of as populations (=PLATE'S Individuengruppen) - certainly a very sound and stimulating method
of treatment. In a previous paper (Du RIETZ 19:10 pp. 293-296)
I have, proposed to apply the population-concept to all sorts of
taxonomical and plantsociological units, a course that I am going'
to take also in the following discussion. Thus I am going to
treat all the various units of taxonomy discussed below simply
as concrete populations, leaving the discussion of the possibly
abstract nature of the con c e p t s aside to those biologists more
interested in formal logic than myself. Far more important than
those purely theoretical discussions is, according to my opinion,
the problem of the definite or arbitrary nature of the borderlines between the populations accepted as units (Du RIETZ 1923 a).
But of course even a most arbitrary delimitation cannot take
away the concrete nature of the population we are dealing with
(comp. HALL and CLEMENTS 1923 p. 9). Nor may we forget the
g en e a log i c a I con tin u i t Y of a species as is often done
by authors advocating a purely, abstract species-concept (comp.
KLINGSTEDT 1928).

3.

The individual.

The most elementary unit of taxonomy is the individual. The

limits of an individual are not always easy to define, but I think
most biologists of the present day agree that the soundest definition
is the physiological one, i. e. that the main criterion of an individual should be its physiological autonomy. Thus in cases of
vegetative propagation a new individual is formed with the break
of the connection with the mother-plant. Theoretical and practical
complications arise in several cases, especially in the case of
vegetative segregation and in colonies of closely connected individuals. I think, however, that we may leave those cases aside in
the following discussion. After the introduction of the clone-concept
(comp. below) certainly nobody will share the opinion of GALLESIO (1816)' and HUXLEY (1855), that every clone should be considered
one individual, a view rather widespread among earlier authors.
As to the various individual-concepts of classical morphology compare HAECKEL 1866 pp. 241-374.

338
4.

The clone.

A c Ion e (WEBBER 1903, SHULL 1912 a and b, JOHANNSEN

1913) is a population consisting of the vegetative
(asexual) descendants of one individual.
. There has been some doubt whether a clone must consist only
of genotypically identical individuals or if also individuals formed
by vegetative segregation (sbud-mutation- etc.) may be included
in it. SHULL (1912 a pp. 28-29) primarily adopted the first
view, but later on (1912 b p. 183) went oyer to the second one,
making his clone-concept a purely genealogical one. JOHANNSEN
(1913 pp. 200 and 209) joined the latter view, which appears to
be the prevailing one among modern geneticists (comp. WINGE
1928 p. 50). Practically, the matter is of minor importance, but
obviously this clone-concept is useless as a t a x 0 nom i c a I unit.
In order to avoid confusion and uncertainty, it may be recommended to use the term p U r eel 0 n e, when dealing with clones
as taxonomical units, i. e. when genotypical homogeneity is postulated.
A special type of clones of great importance for taxonomy is
the a p 0 m i c tic one, formed by apomictic propagation of one
individual.
The investigations of modern geneticists have clearly proved
that the various individuals of a clone are genotypically identical
(vegetative rnutations> excepted). Thus a clone apparently is a'
taxonomical unit of extreme homogeneity. At least many of the
species of Taraxacum and of Hieracium subgenus Archieracium
described by Scandinavian taxonomists probably are pure clones
(comp. below pp. 358 and 362).

5.

The pure line.

A pure line (JOHANNSEN 1903) is a population consisti n g 0 f the i n d i v i dual s form e d b Y s t r i c tl y aut 0gamous reproduction of one homozygotic individual.
There has been some doubt wether only populations formed by
autogamous reproduction of one hom 0 z y got i c individual should
be called pure lines, or if that concept may be extended also to
populations including heterozygotic individuals (LEHMANN 1914
pp. 286-287). However, it appears quite clear that the term pure

339

line must be maintained in the sense of JOHANNSEN - if nol, we

would have to invent a new term for the concept of JOHANNSEN,
as we certainly cannot do without it. On the other hand,. it
appears that we have little need of LEHMANN'S extended "pure
line -concept. Practically, the matter is of little importance, as
strictly autogamous reproduction within few generations leads to
(at .least practical) homozygosity (JENNINGS 1912 p. 487 etc., JoHANNSEN 1913 pp, 484-:-485, 496-499, cornp. also for instance
NILSSON-EHLE 1915 p. 35, LOTSY 1916 pp. 68-75, BAUR 1919 pp.
311-312, HAGEDOORN 1921 pp. 115-118)~
As JOHANNSEN (1903, 1913 etc.) and others have shown, a pure
line is a unit of the same homogeneity as a clone, i. e. all individuals of one pure line are genotypically identical, provided that
no rnutation takes place within it. However, the pure lineconcept of JOHANNSEN is a purely genealogical concept, i. e. if a.
mutation takes place, the descendants of the genotypically deviating individual arisen through this mutation are included in the
same pure line, which thus loses its genotypical homogeneity (comp.
JOHANNSEN 1913 pp. 209-210, SHULL 1912 a pp. 28-29). As a
t a x 0 nom i c a I unit, the pure line thus is as useless as the clone.
For taxonomical use, a concept corresponding to that of the pure
clone would be necessary, i. e. a concept including only pure
lines of genotypical homogeneity. In the following discussion I am
going to designate this concept by the term pure line sensu stricto
(or pure line s, str.).
Good examples of pure lines of apparent taxonomical value are
furnished by the petites especes of Erophila described by A.
JORDAN (1864, 18i3) and ROSEN (1889, 1910, 191I, 1925, comp.
WINGE 1926 a 1) . However, as a b sol u t e autogamy appears to
be at least extremely rare in nature (comp. LOTSY 1916 pp. 7778), the taxonomic value of pure lines appears to be rather restricted. At least in most pure lines a cross with another one
sooner or later will happen, giving rise to a lot of new pure lines
and thus increasing their number so much that their recognition
as taxonomical units will be a rather hopeless task - especially
if many of those pure lines are phenotypically overlapping and
practically indistinguishable from each other (SHULL 1929 b).
1 According to LOTSY (1926), some of those may be apomictic clones, in spite
of WINGE'S criticism of BANNIER'S (1923) attempt to prove the existence of apomixis
in Erophila.
.

340
6.

The biotype.

A biotype (JOHANNSEN 1909) is a pop u I a t ion con sis ting of individuals with ide n tic a I g en 0 t Ypic a I
co n s t i t u t i o n .!
The term biotype, now ill general use, was not immediately
generally accepted by geneticists, some of them preferring for a
long time to use JOHANNSEN'S term genotype also in this sense
(comp. for instance" LOTSY 1916, SCHMIDT 1917, HEI\(BERT-NILSSON
1918). In some previous works (Du RIETZ 1921,1923 a) I followed
those authors, while my present treatment of all taxonomical units
as concrete populations of course makes the accepting of JOHANNSEN'S
distinction between genotype and biotype quite necessary.
(Comp. also JENNINGS 1911 and SHULL 1912 a.)
As every pure clone and every pure line s. sir. consists of one
biotype, the" biotype in exclusively vegetatively propagating or
strictly autogamous populations is a unit of apparent importance
for taxonomy. In strictly allogarnous populations. without vegetative reproduction, however, there is little chance that any biotype
is represented by more than one individual (at least at the same
time), on account of the very great number of possible combinations of genes. This has been pointed out by many authors
(comp. for instance LEHMANN 1914 p. 291, LOTSY 1925 a p. 29); a
good example is furnished by the human population. In those
cases the biotypes are apparently of little taxonomical importance,
as biotype and individual will become practically synomymous.
Intermediate cases are found in populations with both vegetative
and allogamous or with both autogamous and allogamous propagation.
A biotype may be either homozygous or he terozygous.
According to our present knowledge, there is probably not much
chance to find completely homozygous biotypes in other populations than strictly autogamous ones (JOHANNSEN 1913 p. 627). This
means that the only really homozygous biotypes existing will be
those forming pure lines - and even the homozygosity of those
:is now seriously doubted by LOTSY, probably with full right
(LOTSY 1925 a p. 27).
In a series of papers published in the
1 In his earlier works (1905) JOHANNSEN used the Danish equivalent of blotype> (~livstype) as a synomym for both his later terms genotype> and biotype. The form biotype was first used in 1906 (comp. SHULL 1912 a p. 27).

341

years 1912-1916, however, LOTSY claimed that the homozygous

biotypes were the real fundamental units of. taxonomy, and thereIore the only units worthy of being called s p e c i e s . Theoretically, it was not LOTSY'S intention to make his new species-concept
synonymous with JOHANNSEN'S pur eli n e (s, str.i; as he at that
time apparently expected absolute homozygotes to occur also in
allogamous populations.'. He vigorously defended himself against
the accusal of LEHMANN (19.13, 1914 a, b) of having proposed a
restriction of the term species only to the pure lines (LOTSY
1914 b pp. 614-616) and clearly declared, that keineswegs aIle
Arten reine Linien sind. Nevertheless, it has become a habit
among geneticists to polemize against the absurd proposition of
LOTSY to call only pure lines species - without mentioning
anything about LOTSY'S first supposition that species in his sense
were common also in allogamous populations (HERIBERT-NILSSON
1918 pp. 136-137, HAGEDOORN 1921 p. 190, 'fURESSON 1929 a pp. 325
and 332). Certainly this is not quite fair, as a good deal of the
absurdity in LOTSY's the 0 ret i c a I species-concept is constructed
by those authors themselves. Practically, they probably are quite
right in their belief that species in LOTSY'S sense will be formed
only in pure lines. But in order to give a clear view of LOTSY'S
species-concept it is absolutely necessary to point out that its
founder expected it to be generally applicable both in autogamous
and allogamous populations. And even if this supposition may
look absurd in the light of our present knowledge, we must not
forget that it may have been a different matter in 1912-1916.
In his polemics against LOTSY, LEHMANN (1914 b), who wanted
to retain the term species in its old sense, proposed the term
isogene Einheit as a designation for LOTSY'S species. But the
ambiguity of that term appears to have been felt by LEHMANN
himself, for in one place (1914 b p. 291) he replaces it with das
Isogenhomoeygotische. Nevertheless LEHMANN'S term isogene Einheit: has been accepted by LOTSY (1920 p. 181) as a substitute
to his species, the latter term being completely avoided in
LOTSY'S more recent papers. Another synonym for the same concept isRAUNKIAER'S genospecies (RAUm<IAI'lR 1913, 1918). For our
present need, no other term for that -concept than homozygous
biotype appears to be necessary.
1 Later on, LOTSY gave up this belief and went over to the extreme contrary,
namely to the doubting of the existence of any absolutely hornozy go tic biotypes in nature (1925 a p. 27) even in pure lines.

342
As a complement to his new species-concept, LOTSY proposed
to call all he tel' 0 z y go u s biotypes hybrids> - which meant
the introduction of a quite new 'sense also for this venerable term.
Also this proposition must be seen in the light of LOTSY'S belief
in the common 'occurrence of homozygous biotypes (species sensu
LOTSY) also in allogamous populations. Certainly if was not LOTSY'S'
intention to propose to call eve l' y biotype in an allogarnous population a hybrid - as has turned out to be the practical consequence of his hybrid definition. Also RAUNKIAER (1918) applied
the same hybrid- (or bastards-) concept.
It is of great importance for the right understanding of much
of the recent literature on species and hybrids to keep in mind
the real origin of LOTSY'S new application of ' the terms species
and hybrid. That LOTSY'S species-concept was radically different from the traditional one, has been realized by all authors.
That just the same thing. must be said of his hybrid-concept, has
apparently been overlooked by many.
Theoretically, two or more biotypes may be morphologically
absolutely alike but still have different genotypical constitutions,
possible to reveal only by genetical analysis. Such biotypes
would form an is 0 I' e age n t in the sense of RAUNKIAER (1918).
According to RAUNKIAER, the isoreagent is die letzte Einheit del'
Systematik. The practical need of this concept, however, appears
rather doubtful. In strictly allogamous populations, as mentioned
above, there is little chance for any biotype to be represented by
more than one individual (or one clone). Probably the chance
that two individuals belong to biotypes with absolutely identical
morphological appearance is not much greater. Anyhow, what
we know about the human population does not give much hope
to find two such individuals.

7.

The form.

A form (forma) is a pop u I at ion 0 f 0 n e 0 I' S eve I' a I

biotypes occuring sporadically in a species-pop u I a t ion (n 0 t form i n g dis tin c t reg ion a I 0 I' I 0 c a I
fa c i e s 0 fit) and d iff e ri n g fr 0 m the 0 the r b i 0.t Y pes 0 f t his s pee i e sop 0 p u I a ti 0 n in 0 n e 0 I' S eve 1'a I distinct characters.

343
The term form s has been used by taxonomists in so many
different senses, that the possibility of uniting taxonomists in the
use of that term proposed above appears rather doubtful. If this
proposition should fail to become generally accepted, the best thing
to do will certainly be to abandon the term form altogether
(except perhaps in a very general sense, as proposed by COCKAYNE
1928 p. 7, 1929 p. 23). But it would not be easy to find a better
term for the concept concerned - nor to get this new term universally accepted. The term aberration, now frequently used
by zoologists about in the same sense (SEMENOV-TIAN-SHANSKY 1910, '
DOBRZANSKY 1924, RENSCH 1929), is used by WINGE (1928 p. 243)
and other modern geneticists (e. g. CLAUSEN 1927 p. 253) for' a
quite different concept <chromosome-mutation). ' Another newly
established synonyme is exotypus. (REMANE 1928, comp. RENSCH
1929 p. 10). Very often also the term variety has been used
in this sense.
In many taxonomical works (comp. for instance ENGLER 1916
p. 4) we find the term Iorm used for the designation of what is now
generally called rnodification (NAGELI 1865 pp. 277-278, 1884 p.
264, LOTSY 1916 p. 28, BAUR 1919 p. 9, DIELS 1921 p. 175) or epharmone> by COCKAYNE and ALLAN (1927 p. 263, compo also COCKAYNE
1928 p. 7, 1929 p. 22), i. e. units of purely phenotypical
nature. For this concept, however, the term mod i f i cat ion
appears to be so well established in the literature of the present
day, that we certainly do best in sticking to it., That I am
not giving the modification a place among the fundamental units
of taxonomy, is due to, my strong conviction of the absolute
futility of describing and naming modifications in taxonomical
works. If we describe and name one modification, we could
equally well describe and name thousands or millions of them.
Of course exact studies of the phenotypical variability of biotypes
may have their given place in botanical science - but such studies should certainly be made in a way quite different from that
of describing and naming modifications as taxonomical units.
According to the definition of the form given above, it would
be theoretically possible to describe every morphologically distinguishable biotype of a species as a form. If we keep inmind
the little chance of any biotype being represented by more than
one individual (or clone) in allogamous populations, the futility
of this task is easily seen. The describing and naming of single

344

biotypes within a species can certainly only be taken into consideration when one biotype is represented by a great number of
individuals, as is the case in clones and pure lines.
In allogamous populations with vegetative propagation leading
only to the formation of sma II clones it is obviously equally
futile to describe and name each of those clones as it would be
to describe each individual in an allogamous population without
vegetative propagation. The formation of such small clones is a
very common process in nature. An easily observed example is
. that furnished by the Anemone nemorosa-population around Upsala, in which a careful observation everywhere reveals clones one
or a few square meter large, and distinctly differing from each
other in colour (some are anthocyaninic and some are not), the
form and size of the sepals, etc.
If sexual reproduction remains unchecked, there will be little
hope for such clones to develop to larger monobiotypic populations
worthy of special descriptions and names. But if the clone-forming biotype has lost its power of sexual reproduction completely
or partly, the clone will easily develop into a monobiotypic population of sufficient -size, to be recognized by taxonomists. If the
power of sexual reproduction is only partially lost, as in the cases
of facultative apomixis described in Hieraciutn subgenus Pilosella
(OSTENFELD 1910, 1912, 1919, ROSENBERG 1917, comp. also TACKHOLM 1922 pp. 313, 352-353), the apomictic clones may become
large enough to be distinguishable as distinct f a I'm s, but between
those clones crossing takes place often enough to make it impossible to treat them as separate species. Quite different is the case
if the power of sexual reproduction is completely lost. In this
case many clones may develop into sharply delimited monobiotypic populations large, constant, and permanent enough to be
easily recognized by taxonomists as distinct s p e c i e s, each with
its characteristic distribution-area. The best known examples of
such populations are furnished by apomictic species in the section Vulgata of Hieracium subgenus Archieracium (comp. below
pp. 358 and 362).
However, it also happens that a clone completely lacking the
power of Sexual reproduction lacks the vitality needed for growing out into a population large, constant, and permanent enough
to be treated as a new species. Good examples of such clones
are furnished by lichens. Of many lichen-species normally pro-

345
ducing apothecia but no soredia there are found soredia-producing
forms without apothecia, occuring sporadically as very small isolated clones in the great population formed by the species concerned, and obviously of slight vitality and ephemeral appearance,
as larger clones are never found. In a previous paper (Du RIETZ
1924 a pp. 386-388) I have described several such clones (or
groups of near related clones) as va r i e tie s, while similar clones
(or groups of clones) of somewhat larger size and apparently of
somewhat greater vitality were described as sub s p e c i e s (I. c.
pp. 388-389). With the terminology proposed in this paper, both
types of clones must be called form s only. As will be shown further below, there are also found soredious clones with great vitality
and large continuous distribution-areas, having the typical appearance of s p e c i e s and being separated by taxonomists as such.
Analogous to the clones with retained power of sexual reproduction are probably most of the pure lines occuring in wild
plant-populations. As absolutely pure autogamy is hardly found
in nature, there is little chance for any pure line to grow out
into a real species, but very often it may grow out unspoilt by
crossing into a population large and distinct enough to be distinguished as a good for m. Good examples of such pure lines easily
distinguishable in nature are the petite especes of Erophila mentioned above (p, 339). Several of those pure lines are often found
together; crossing is apparently rare in nature but easily performed artificially, resulting in a large number of new biotypes
(ROSEN I. c.).
Such distinct m 0 nob i 0 t Ypic for m s as mentioned above
from autogamous or asexual populations are lacking completely
in typically allogamous populations. If we want to divide such
populations into forms, we will have to deal with pol Yb i ot Ypic form s, i. e. with groups of biotypes kept together by
certain characters lacking in other groups. Such polybiotypic
forms have no genealogic continuity; they simply consist of a
lot of s i mil a r biotypes arising independently of each other in
the everchanging stream of intercrossing biotypes, and disappearing again when crossed with biotypes belonging to other forms
of the same populations (if not disappearing because of lacking
vitality or sterility, as is often the case in the most aberrant forms
of a species-population). Thus the polybiotypic form can scarcely
be said to be a nat u r a 1 unit, being far more an art if i cia 1

346

means of studying the polymorphy of a given species-population.

We should not forget that when choosing a system for naming
and describing polybiotypic forms.
In many cases we only want to keep apart the most aberrant
forms of a species-population, leaving the rest of the population
undifferentiated under the name of the main form, f. tqpica;
or similar designations. Thus we often want to note the occurrence of sporadic white-flowering forms in a species with normally
coloured flowers, or of forms with laciniate leaves in species with
normally entire leaves. Such forms are now often described and
named in the same manner as species and other units of higher
rank, i. e. they are given an arbitrarily chosen Latin name and
a short Latin diagnosis. Certainly this is a very unpractical method. Much better is the (also common) method of naming all
white-flowering forms of normally coloured-flowering species forma
(loribus albidis (f. (lor. alb.), i. e. of choosing a stereotyped formula for all the analogous forms of different species. That this
method is practically applicable, is due to the remarkable parallelism in polymorphy found in the most different species of the same
class of the vegetable system, and often quoted in modern literature as VAVILOV'S law of homologous variation (VAVILOV 192~,
comp. also BAUR 1919 pp. 293-294, Du RIETz 1921 p. 13, PHILIPTSCHENKO 1926 pp. 63-81, SEYBOLD 1927, TROLL 1928 pp. 90-91,
HALL 1928 p. 8). Also for the naming of t era to log i cal forms
this method appears very suitable.
This method, which really implies a naming of the characters
of the forms rather than of the forms themselves, certainly is
much in advance of the ordinary practice of taxonomists, i. e.
that of naming each form by an arbitrarily chosen name, difficult
to invent and difficult to remember, and describing it with a
diagnosis (a good system of character-naming, of course, makes
any description superfluous). It is equally applicable when we want
more than distinguishing the most aberrant forms, i. e. when we
want to divide the whole species-population into forms suitable
for helping us to study the polymorphy of that population. In
the taxonomy of Sphagnum, this method was proposed already by
Russow (1888), and is now commonly used by sphagnologists.
Certainly it has done much for simplifying the taxonomy of this
polymorphic genus. However, a good part of the Sphagnum-forms
designated in this way may be mere modifications.

347

. In most cases, however, it will prove necessary for the intensive

study of the polymorphy of species-populations to invent a special
system for the division of the species-population concerned into
suitable forms and for the naming ofthose forms. Good examples
for such practical systems are furnished by H. E. PETERSEN'S studies
of the .polymorphy in the species Anthriscus siloestris (1914, 1922)
and Pimpinella saxifragu (1921), and also by CRAMPTON'S monographs of; the land-snail genus Partula in the Society and Mariana
Islands (1917, 1925 a, b). In Anthriscus siloestris, PETERSEN distinguished 16 rnain groups of forms marked by Latin names iBreoisecunda,
Laiior Breoidioisa, Latadioisa etc.), the distribution of which was
studied statistically; in each group he distinguished and named a
number of forrns> (Brevisecunda media, Brevisecunda Longibrevis
media, etc.), In a similar way, CRAMPTON divided his species of
Partula into a number of color-classes> 1 (sometimes those were
grouped to color-orders), which were used as units for the statistical studies of the regional variability of the polymorphy of the
species. In many cases it may prove practical not to give the
forms any names at all, but simply divide the population into
variation-classes marked by numbers (SIRKS and BIJHOWER 1919,
PETERSEN 1924, 1926), or by letters symbolizing the main characters, as was done already by COUTAGNE 1895 in his studies of
polymorphic mollusca (comp. BATESON 1913 p. 126) as well as by
RAUNKIAER (1925) in his study of the polymorphy of a Danish erataegus-population. It may also often be practical to study the
polymorphy of the different organs separately, using a dynamic
system in the sense of HAYATA (1921 a, b, 1928 b, c, comp.
below p. 409). The species Picea excelsa, for instance, is often divided
into one series of forms according to the form of the cone-scales
and other series of forms according to the ramification, the needles etc. (comp. SYLvEN 1916 a pp. 112-152).
In practical taxonomy, such intense analyses of the polymorphy
within a species-population as mentioned above is mostly out of
question, the average taxonomist having to be content with distinguishing the most conspicuous forms within each species. As a
matter of fact, even those most conspicuous forms of a species
are often completely neglected by taxonomists, as long as they
are regularly occurring together, thus being normal characteristics
1

Analogous to the color-patterns or elementary patterns distinguished by

and COOKE (1912-1914) in Hawaiian Achatinellidae.

PILSBRY

348

of every part of the species-population concerned. Very few

taxonomists, for instance, take the trouble of distinguishing
and naming the twocolour-torms of Orchis' sambucina (in the
floristic handbooks those are only rarely mentioned as f. typica
and f. purpurea) - it is simply considered characteristic for the
species that both those forms are regularly found together. More
or less analogous cases are offered by anthocyaninic and non-anthocyaninic forms of many other species: sometimes they are named
by taxonomists, sometimes not. Analogous forms differing in the
shape of the leaves, in the villosity etc. are, as a rule, estimated
somewhat higher, and are mostly mentioned in the floristic handbooks as f. lati{olia and f. angusti{olia, f. glabra and f. uillosa, etc.
But only if the forms concerned show marked differences in their
distribution, i. e. tend to the formation of more or less pure populations, they attract the interest of taxonomists in a more profound
way. In this case, however, they approach the next stage in the
series of taxonomical' units, namely the variety.
As the near relation between two morphologically distinct forins
of one species-population may be very difficult to recognize except
by a close field-study of the living population, various forms of
the same species (and even of the same variety) have often been
described as separate species in traditional museum-taxonomy.
Good examples are furnished by many of GULICK'S species of Hawaiian Achatinellidae, shown by PILSBRY and COOKE to be mere
forms of the same species. And if the two colour-forms of Orchis
sambucitia had been growing in an exotic country without local
hotanists and sent to a European museum in dry herbariumspecimens (or even in living specimens to a botanical garden), they
would probably have been described as two good species. The danger
of mistaking forms for taxonomic units of higher rank may be
avoided only by fie ld stu die s of the population concerned, or
to some extent by very large and properly labelled collections
demonstrating the polymorphy in each locality.

8.

The variety.

A variety (uarieias) is a population of one or several biotypes, forming a more or less distinct local facies of a species.

349

With the variety we have reached the lowest unit in the series of
taxonomical units forming more or less closed lntercrossing populations or s y n gam eon s (LOTSY 1918, 191 ~ a, b, 1921, DANSEH 1924).
While the various forms of an (allogamous) variety constantly
cross with each other, thus presenting a more or less ephemeral
appearance, the variety itself is subject to crossing with other
varieties only in the transition-zones to them.
The occurrence of morc or less distinct local facies of a speciespopulation ( local races) is a fact wellknown to biologists since
long ago. Many of the petites especes of A. JORDAN and his
followers were such local races, while others probably were mere
forms in the sense defined above. However, it was first by the
intense study of the polymorphy of animal 'species-populations
started towards the end of the last century, that the differentiation
of species into local races became more intimately known. Thus
GULICK (1872-1905), COUTAGNE (1895), and MAYER (1902) demonstrated the marvellous local differentiation in Mollusca, while HEINCKE
(1898) described Farnilien oder Stamme- of the herring, regarded
by him as forming die erste Gruppe des natiirlichen Systems und
del' notwendige Ausgangspunkt aller Untersuchungen abel' 'die
Entstehung del' Arten'. In 1910, SEMENOV-TIAN-SHANSKY proposed
the term natio for such besonders kleine lokal-geographische
Einheiten within a species. Somewhat later LLOYD (1912) described a most remarkable local differentiation of Indian rats, the
rat-population of each house forming a more. or less distinct local
facies of the species - a result later confirmed by HAGEDOOHN
(1921) for other countries. In a series of important papers, SUMNER
(1915, 1917, 1918, 1920, 1923, 1924) demonstrated the differentiation of the California deer-mouse (Peromyscus maniculalus) not
only in the main geographical races called subspecies but also in
a large number of genotypically fixed local races. J. SCHMIDT
(1917) probably was the first author to apply the concepts and
terms of modern genetics upon such local races, showing that
each local race of the fish-species Zoarces oioiparus consisted of a
great number of intercrossing biotypes, the average characters
of the races in first line being only the statistical expression for
a different mixture of various genotypes! (p. 334). SCHMIDT also
pointed out (p. 333), that if his interpretation of his Zoarces-ana1

= biotypes (comp, above p. 340).

24.- 30531. Suensk. Botaniek. TidBkri/t.

1930.

350
lyses was right, Ihere can hardly be any doubt that what has
been found for Zoarces has age n era I validity for a II populations of a II fish-species etc. Analogous results were obtained
by CRAMPTON (1917, 1925 a, b) in his studies of Pacific Partulaspecies, each colony (= variety in our sense) consisting of a
great number of color-classess in proportions differing from those
of all other colonies. - A similar differentiation of a p I a n t-species into local populations widely differing from each other in
the percentage of the various groups of forms composing the
species was shown in 1914 by H. E. PETERSEN (comp. also PETERSEN 1922) for Anthriscus siluestris. Some years later, analogous
results were obtained by TURESSON (1922) in his study of the local
races of Hieracium umbellatum differentiated in various parts of
the coast of Skane in South Sweden.
The works mentioned above are only some of the main ones
in this field of study. The differentiation of species into local
facies, each containing a great number of intercrossing biotypes,
has now been demonstrated by so many authors (especially zoologists, cornp. besides those mentioned above especially GOLDSCHMIDT,
SEILER and POPPELBAUM 1920), that there is little reason to doubt
that a closer investigation will reveal a similar local differentiation in most sexual species of both plants and animals.
As to the terminology of those local races, the term natlo
(SEMENOV-TIAN-SHANSKY 1910) has been used by several zoologists,
especially in Russia. It has the great advantage of never having
been used in any other sense, but appears to be too little known
to have much hope to become generally accepted. The term
jordanon, proposed by LOTSY in 1916, according to LOTSY'S
original definition (1916 p. 27) would be about synonymous with
RAUNKIAERS isoreagent (comp. above p. ;J42), but, on the other
hand, it was explicitly proposed by LOTSY to replace the term
species in the Jordanian sense, viz.: mikrospecies, elementary
species etc. It is not quite clear from LOTSY'S own works in
what sense the term jordanon is p rae tic a II y used by its
originator, but probably it may be considered as about synonymous
with our Iorm>. By COCKAYNE and ALLAN (1927, comp. also
COCKAYNE 1928, 1929) the term jordanon is used both synonymous
to our forrn and to our variety. PHILIPTSCHENKO (1927), on
the other hand, proposes to use jordanon for all units intermediate between the biotype and the Linneon> (= the species), and

351

suggests to distinguish between Jordanon I Ordnung (= our

subspecies) and Jordanon II Ordnung (= our variety). Thus
the term [ordanon , in spite of its youth, has already been
used in quite as many different senses as the old term variety
has, and cannot be given any preference from that point of view.
As mentioned above, the term variety has been used very
often simply as a synonym to our form. It has also been used
to designate a superior unit of the same character, i. e. a more
collective form or a group of similar forms. However, we also
find it used in taxonomical literature as a designation of a geographically limited part' of a species-population. In the system of
taxonomical units proposed in 1916 by the grand old man of
present-day taxonomy, ADOLF ENGLER, we find the term Iorrn
used to designate mere modifications, and Subvarietat- about as
a synonym to our Iorm, while three units of a more regional
character (geographically limited facies of the species-population)
are recognized: 1. Subspecies. 2. Varietat und 3. Rassenvarietat oder Proles (var. proles)>. None of the three terms last mentioned is used for units of purely local character, all of them far
more corresponding to our subspecies - or partly even to our
species, as the species-concept of ENGLER is an extremely wide one.
The term variety has thus been used in taxonomy for units
of the most different rank, ranging from our forms to our subspecies. I fully realize the boldness of my proposition to restrict
that term to a unit of more intermediate rank, namely to the
more or less 10 c a I facies of a species-population. However, if
we want to keep the old term variety for any of the units
of modern taxonomy, this will probably be the only one possible. If my proposition should fail to become generally accepted,
the best way will possibly be to abandon the term variety
altogether, taking our refuge to SEMENOV-TIAN-SHANSKY'S term natio
for the unit in question.
As has been pointed out clearly by HAGEDOORN (1921 pp. 136139), distinct varieties in our sense are formed only in species
consisting of comparatively stationary individuals. In species consisting of more or less rapidly migrating individuals no distinct
varieties are found. The differentiation of varieties must thus
be due to a certain amount of in b r e e din g, causing a reductlon
of the Total Potential Variability! in the sense of HAGEDOORN.
1 = the
number of genes in respect to which an individual or any group

352
The inadequacy of the common belief in selection as the main
acting force to the differentiation of varieties has been shown in
a most suggestive way by HAGEDOORN. To HAGEDOORN we also
awe the explanation of the aut 0 mat i c reduction of potential
variability or polymorphy that takes place quite independently of
any sort of selection in any allogamous population more or less
~ffectively cut off from crossing with populations of a different
genotypical constitution. This explanation may be summarized
in the following way:
1. In a group of freely crossing organisms, we can very easily
show mathematically that, supposing every 'mating to result in an
equally large number of offspring, and assuming an uninterrupted
steady increase in numbers, the potential variability remains undiminished. JENNINGS and PEARL have worked out these calculations
very thoroughly. However, how does it work out in reality? Is
it possible that there is an automatic reduction of variability in
freely crossing populations, of the same nature as the very great
reduction in self-fertilized organisms and in inbred series, slower,
but still appreciable?
In the first place it is obvious, that the multiplication of any
group of organisms cannot proceed indefinitely. A few heterozygotes in a mixture of organisms, of which the majority are homozygous will continue to exist as long as multiplication is unchecked.
But, wherever the group is continued from a fraction of the number
of individuals, or where a colony is started by a few individuals,
the chance of the heterozygotes to be included in the group, or
to have heterozygous children included, is proportionate to their
frequency. Heterozygotes will produce homozygotes, but not the
reverse.
The group of organisms chosen by fate to become the parents
of the next generation is usually, but always occasionally, considerably smaller than the number of individuals of their species.
Every case in which rare individuals, having genes, not present
in the majority, or in which rare individuals being impure for,
of individuals is not pure; homozygous> (HAGEDOORN 1921 p. 10!). For the same
concept DANSER (1929 a, b) uses the term >potential polymorphy, As the term
polymorphy is tending to replace vartabllity in the literature of the present
day as a designation of the gruppenwelse Varlahllltat (PHILlPT~CHENKO 1927) of
a population (co mp. COCKAYNE and ALLAN 1927 p. 239), I am following DANSER in
this respect.

353
or lacking in genes, common property of the majority, happen
to be excluded from the number of pro-creating individuals, the
total potential variability is lowered.
This, in our opinion, is the most important gain in knowledge
which we owe to MENDEL'S work, and to the biomechanical interpretation of his work. Reduction of potential variability, in other
words purity of species is automatic, and not dependant upon any
sort of selection> (HAGEDOORN 1921 pp. 119-120.)
. 2. If the population concerned is divided into two parts, and those
parts are isolated from each other in absolutely identical habitats, .
there are very great chances that the reduction of potential polymorphy taking place in those isolated parts of the population
leads to different results, as 1) the initial genotypical composition
of the two parts is hardly ever absolutely identical, some genes
probably being stronger represented in one part than in the other,
thus having greater chances to survive and to become homozygous
in the first part than in the other, and 2) even if a certain gene
is equally represented .in both parts, its fate in those paris is largely determined by pure chance, in extreme cases leading to its
elimination in one part and its homozygosity in the other. Of
course this automatic differentiation may be increased by selection,
if the habitats of the two parts differ from each other.
3. Even if no complete isolation occurs within a population,
the reduction of potential polymorphy easily leads to different
results in different parts of the population, causing a differentiation
of the population into subordinate units, if intercrossing between
more distant parts of the population is preven ted by the stationary
habit of the population. But while in cases of complete isolation the
differentiation results in units more or less sharply delimited from
each other (= species), it results in the latter case only in the
formation of more or less intergrading units (= subspecies aud
varieties).
At last it must not be forgotten that in plant-populations the
reduction of potential polymorphy, and thus also the differentiation
of varieties, is very often hastened by a certain amount of autogamy, or by the formation of clones.
As to the practical study of varieties, most taxonomists certainly
agree that Latin names and diagnoses are of -Iittle use for those
units. A variety is best named by its locality, and is best described by a statistical survey of the forms constituting it (comp,

354

above p. 347). Very good examples for the various modern methods
of studying varieties are furnished by PETERSEN'S monographs of
Anthriscus siluestris and other Danish plant-species (1914-1926,
comp. above p. 347), by CRAMPTON'S monographs of Pacific Partulaspecies (1917, 1925 a, b) and by E. ANDERSON'S (l"928) investigation
of two North American species of Iris.
9.

The subspecies.

A subspecies is a population of several biotypes forming a more or less distinct regional facies
of a species.
Like variety, the term subspecies has been used by taxonomists in most various senses. It has heen used for any unit intermediate between form and species, i. e. for morphologically distinct
groups of forms, for varieties in the sense defined above, for
sharply delimited populations considered to be too nearly related
to be raised to the rank of good species, for species of somewhat
dubious autonomy, or for distinct regional facies (geographical
races) of a species. The restriction of the term subspeciesi to
the sense last mentioned has been gradually inaugurated during
the last four decades, in the first line through the works of
WETTSTEIN (1896 a, b, 1898), K. JORDAN (1905), HARTERT (1910-1923),
SEMENOV-TIAN-SHANSKY (1910) etc., and now appears to be generally
accepted among zoologists, while botanists are still using the term
in a rather chaotic way.
Of other terms in use far the same concept, reference in first
line must be made to RENSCH'S geographische Hasse. (RENSCH
1929). As RENSCH restricts the species-concept to the species not
differentiated in regional facies, applying the term Rassenkreis.
to all geographically differentiated species (comp. below p. 359),
he cannot use the term subspecies. in our sense. Other synonyms
to our subspecies are KLEINSCHMIDT'S Form. or Hasse (comp.
below p. 359) and REICHENOW'S (1904) Konspecies> (comp. PLATE
1914 p. 134). By some authors the term variety has been used
either synonymous with our "subspecies (comp. for instance
CRAMPTON 1917) or as a designation for subspecies of lower rank
(ENGLER 1916, cornp. above p. 351).
Special reference must be made to the ecotypes described by
TURESSON (1922-1929). While in some cases they appear to be

355
identical with good species generally recognized by .taxonomists,
most of them may be regarded as sub s p e c i e s in our sense
(comp. RENSCH 1929 p. 78). In one of his later papers, TURESSON
(1925 b) also applies the term subspecies. to some of his earlier
ecotypes. Theoretically, however, the ecotype-concept of TURESSON
appears to be more comparable with the subspecies oecologica of
ALPATOV (1922, 1923, 1924). ALPATOV pointed out, that several
subspecies geographicae (= our subspecies) of a species often
contained analogous subspecies oecologicae, e. g. that quite analogous subspecies oecologicae truncicolae were found both in
the European and in 1ihe .Japanese subspecies of Formica rafa. In
the same way, an oecotypus campestris, oecotypus arenarius or
oecotypus salinus in the sense of TURESSON (1925 a pp. 224-225)
may be differentiated in more than one subspecies of a certain species.
If the term ecotype could be restricted to such cases, it might be
found useful '1:0 taxonomy, as it is obviously preferable to ALPATOV'S
term. As a synonym to our subspecies it appears to be of no
use whatever.
That one for m may be found in several subspecies of one
species, has been pointed out by several authors (comp. DOBRZANSKY 1924 p. 417). This means that it is not always possible to
tell from an isolated specimen in a collection to which subspecies
it belongs, if the locality of the specimen or the range of the subspecies is unknown (comp. SUMNER 1923 p. 239). Like varieties, subspecies are best described by a statistical survey of the forms constituting it (comp. CRAMPTON 1917, 1925 a, b). But unlike varieties, subspecies are conveniently named by Latin names, either after the'
pattern Pinus silvestris subsp, lapponicai (Pinus silvestris *lapponica)
or by a simple trinomial name like Pinus siloesiris lapponica, The
latter method is generally accepted by present zoologists and by some
American botanists (comp. for instance HALL and CLEMENTS 1923,
BABCOCK and HALL 1924, HALL 1928); it appears to be practical
enough to be accepted also by botanists in other countries.
Unlike the species, subspecies are neither sharply delimited from
each other nor connected by transitional populations of the hybrid--type (comp. below p. 391). The various subspecies of a
species are continuously intergrading into each other, their'delimitation thus being infinitely much more arbitrary than that of
the species. Still there may be more or less distinct zones with
increased change in the composition of a species-population, which

356
of course must be carefully taken into account at the delimitation
of the subspecies.
A multitude of good examples of the differentiation of animal
species into subspecies is given by RENSCH (1929). Wellknown
examples from the vegetable kingdom are the subspecies of Pinus
siloestris Uapponioa, septenlrionalis, borussica, scotica, engadinensis
etc., cornp. for instance SCHOTT 1907, SVLVEN 1916 a, b, CAJANDER
1921), and of Picea excelsa (the eastern subspecies of which is often treated as a separate species under the name of Picea obovata,
but according to CAJANDER I. C. continuously intergrades into the
European subspecies). Many alpine plant-species are found as
different subspecies in different parts of the European Alps (e. g.
Silene acaulis and several species of Saxifraga, comp. ENGLER 1916).
An analogous differentiation in subspecies is found in many alpine
and subalpine plant-species of the Southern Alps of New Zealand,
a good example being furnished by Celmisia pettolata Hook. f.
Of this species, which plays a prominent role in a good part of
the mountain-chain, I had the opportunity of examining at least
three different subspecies in the field (one in the upper Waimakariri region at about lat. 43, one in the Mount Cook region at about
lat. 43 40' and one in the Routeburn-Humboldt Mts region at the
western end of Lake Wakatipu at about lat. 44 40'). Each of those
subspecies was composed by a large number of forms, the general composition of the population appearing to change gradually from the
northernmost subspecies of the Waimakariri region to the southernmost one of the Northern Fiord District, which mayor may not be
connected with Celmisia verbascifolia of the Southern Fiord District
by a gradual transition.' - In many (but by no means in all)
1 In
the Waimakariri-popuIation (on Hills Peak N. of Arthur's Pass and Mt
Misery near Cass, both visited by me in the company of Dr. H. H. ALLAN), I saw only
forms with distinctly petiolate, rednerved and more or less serrulate leaves with
rounded base, in the Wakatipu-population (on Bold Peak, Lake Harris Saddle,
and the peaks around the latter, visited by me partly in the company of Mssrs
G. SIMPSON and J. s. THOMSON) only forms without distinct petiolae, with leaves
gradually attenuated at the base, not serrulate, or very ind-istinctly, and with the
nerve either green or red. The Mt Cook-population (on Sealey Range) distinctly
kept -In the middle between those two extremes, consisting partly of unpetiolate, partly of more or less petiolate forms, with the nerve red or green and more or
less indistinct serrulation. Some of the forms in the Wakatipu-population are very
much approaching C. oerbascifolia Hook. f., and have even been referred to this
species by CHEESEMAN and COCKAYNE (some of my specimens Dr. COCKAYNE named

357
New Zealand plant genera the differentiation has led. to a large
number of sharply delimited alpine and subalpine species with
comparatively small distribution areas and J.10 distinct differentiation of subspecies, while the species of the lower belts are more
widely distributed and .often (but not always) differentiated into
a number of continuously intergrading subspecies. The explanation certainly must be sought for in the lack of effective barriers
in the lower belts, while in the alpine and subalpine belts a certain amount of isolation is often brought about by broad valleys
etc. A good example is furnished by the genus Hebe, the most
common lowland-species Hebe salicifolia being distributed throughout New Zealand but differentiated in several subspecies (comp.
COCKAYNE and ALLAN 1926 a pp. '17-19, 1927 p. 274), while most
mountain-species show comparatively small distribution areas and
no distinct differentiation of subspecies.

10.

The species.

The sma 11est nat ural pop u I a t ion s perm a n e n t 1Y

separated from each other by a distinct discontinuity in the series of biotypes, are called species.
A species thus is a population consisting either
of one strictly asexual and vital biotype, or of
C. oerbascifolia val'. pubescens"). However, what I have seen of the latter species
(I studied it in the field on Wilmot Pass and Mt Barber at lat. 45 30' together
with Mr. W. R. B. OLIVER) is constanlly more glabrous than the most glabrous
individuals that I brought home from the Wakatipu-population. In the Dominion
Museum at Wellington, where D. PETRIE'S large collections are preserved, I saw forms
of C. oerbascifolia closely resembling my own from Wilmot Pass and Mt Barber,
collected as far north as in the Clinton Valley (lat. 44
while all the specimens
collected further north or further east in the Fiord and South Otago Districts far
more resembled those of my Wakatipu-population. Further field-studies must decide whether C. oerbascifolia and the southern subspecies of C. petiolata are connected only by a narrow hybrid-zone (to which then probably some of the forms in my
Wakatipu-population should be referred) or by a broad zone of more continuous
intergrading. In the latter case it may prove necessary to include C. petiolata
and C. oerbascifolia into one heterofacial species.. However, it must be pointed
out that my knowledge of the populations coneerned is far too incomplete to
guarantee that not further field-studies will reveal zones of discontinuity even
within the huge population here referred to C. petiolata, enabling future taxonomists to split it into several species - though what I have seen of it appears
to make this rather improbable.
0

050'),

358

a group of practically undistinguishable, strictly

asexual and vital biotypes, or of many sexually
propagating biotypes formingasyngameon1separated from all others by more or less complete sexual
isolation or by comparatively. small transitional
populations.
One of the greatest difficulties confronting every attempt to give
a generally applicable species-definition is formed by the fundamental difference between the constitution of sexual and asexual
populations. Owing to this difference, different alternatives for
sexual and asexual populations are necessary in the general definition of the species-concept. While there is a comparatively
good agreement (at least theoretically) among present-day botanists regarding the species-concept in sexual populations, very
different opinions are found regarding what should be called species
in asexual populations. Many geneticists find it absurd to call
single asexual biotypes or even small groups of practically undistinguishable asexual biotypes species, when most sexual species
contain an enormous number of biotypes (comp. for instance TUREsSON 1922 p. 232, 1923 p. 173, 1926 b pp. 187-190, 203, 1929 a,
CLAUSEN 1922, TEDIN 1925 pp. 355-356) - but as yet they have
failed to tell what unit in asexual populations should be better
corresponding to the sexual species. The definition recently given
by TURESSON for his new concept agamospecies> (1929 a pp. 332333: An apomict-population the constituents of which, for morphological, cytological or other reasons, are to be considered as having
common origin ) is obviously too vague to be of any practical
value at all, as it is equally applicable to any unit of higher rank
than the biotypes in the genera concerned (from a group of a few
biotypes up to the whole genus!). As K. JOHANSSON (1927) has
pointed out for the Hierocia Vulgata, it is often a rather hopeless
task to group the apomictic microspecies to higher units of any
taxonomical importance to which the name species could be
applied. Thus in those cases we simply have to apply the speciesconcept to the rnicrospecies, if we wish to use it at all in the
genera concerned. The claim of certain geneticists that the speciesconcept should be applied to units of about the same size and
the same polymorphy, will hardly be generally accepted by taxo1 = intercrosstng population. Compo
pp. 204-209.

LOTSY

1918,1919 a, b, 1921,

DANSER

1924

359
nomists, though the same view is sometimes found also among
taxonomists (comp. below). Among most present-day taxonomists
the view is firmly established, that a natural species-concept
can be founded only upon the principle of dis con tin u it y, i. e.
that we have to delimit our species simply after the lines or zones
of discontinuity found in nature, without any prejudice as to the
convenient size of the species. If Nature has made some species
larger and more polymorphic than others, we simply have to accept
this arrangement.
Far more difficult than the delimitation of species in strictly
asexual- populations (which in spite of the often very great number
of species probably is one of the easiest tasks confronting the
taxonomist) is the dealing with fa c u l tat i vel y asexual populations. As mentioned above (p. 344) the stability and permanency
of the clones in such cases is often hardly sufficient to give them
the rank of species, and the species-delimitation in facultatively
asexual populations may certainly be a most difficult task (comp.
for instance the genera Rubus and Rosa). Similar difficulties are
met with in facultatively autogamous populations (regarding the
delimitation of species in those comp. SHULL 1923, 1929 a, b).
As was mentioned above, the acceptance of dis con tin u i t Y
as the only base for the delimitation of species must lead to the
rejection of s i z e as a base for that delimitation. There is a
rather widespread tendency among present-day biologists (especially
among zoologists) to apply the term species> only to populations
of comparatively equal size (comp. for instance RYDBERG 1929 and
the authors mentioned below). A consequence of this tendency
is the method of certain modern zoologists to restrict the use of
the term species to comparatively homogenous populations not
differentiated into distinct geographical facies, substituting it with
the term Hassenkreis (RENSCH 1926, 1929, preceded by the typus
polymorphus of ENGLER 1872, 1916, the Hassenkette of F. and
P. SARASIN 1899, and the Forrnenkreis> of KLEINSCHMIDT 19011926) when dealing with more heterogenous populations (species
in our sense) differentiated in geographische Hassen (= our
subspecies, cornp. above p. 354). RENSCH is perfectly right in
pointing' out that his Arten, as regards size and polymorphy,
are far more corresponding to the geographische Hassen> than
to the Hassenkreise>. But if we accept discontinuity, not
size nor polymorphy, as the only base for the delimitation of

360
species, both RENSCH'S Arten and his Hassenkreise must be
included in the same species-concept and carry the name of
species.
The rejection of the principle of size and polymorphy as a
base for the delimitation of species also implies the rejection of
the widespread belief that species should be delimited after the
magnitude of the differences between the forms concerned.' According to the species-concept defined above, two forms may be
very different and still belong to the same species, if connected
by transitional forms, while two other forms may be much less
different and still belong to two different species.

In recent genetical literature we often find the term species>

substituted by Linnean species or Linneon (the latter term
proposed by LOTSY 1916). The necessity or even desirability of
those new terms is difficult to realize. LOTSY'S invention of the
term Llnneon> was made necessary simply by his radical proposition to apply the term species in a quite new sense, namely
as a designation for homozygous biotypes (comp. above p. 341).
The latter proposition was declined by practically all biologists,
but nevertheless the new term Linneon has been accepted by
many. If the term species is kept in its old sense, Linneon
in the sense of LOTSY of course is completely superfluous. COCKAYNE and ALLAN (1927, cornp. also COCKAYNE 1928 p. 7, 1929
p. 23) have proposed to use Linneon in a modified sense, i. e.
for a unit superior to the species, or for two or more closely
related species connected by hybrids and confused by earlier
taxonomists. Against this use of the term may be advanced not
only the confusion easily resulting from the use of the same
term in two different senses, but also that it appears to give a
somewhat wrong idea of what the species of LINNJEUS really
were. Surely many of the species described by LINNJEUS on the
base of bad and. insufficient herbarium-material from foreign
countries. were aggregates of several good species in our sense
(i. e. Linneons in the sense of COCKAYNE and ALLAN) - but in
the flora studied by LINNJEUS in the field most of them certainly
As I have myself been accused ot having advocated this principle by TURES(1925 a p. 232), it may be appropriate to point out that TURESSON'S accusal
is based simply upon a wrong citation of my criticized paper, namely by a
translation of schart genug abgegrenzt- and schart abgesondert into sufficiently different.
1

SON

361
were just identical with the species of present day taxonomists
(or in some cases even with units of still lower rank l). There
are also examples of Linnean species (in the true sense, namely
species described by LINNAWS) that were degraded to forms by
later taxonomists and reestablished as species first by the modern
monographers of the genera concerned (e. g. Poa anqustijolio).
Also the terms ecospecies> and )) coenospecies recently proposed by TUREssoN (1929 a) appear to be easily dispensable.
TUREssoN's ecospecies is practically only another synonym for
our species> , if the asexual species (and some very dubious sexual
species with prevailing autogamy) are excepted. His coenospecies> ,
which according to the definition should consist of species having
a common origin so far as morphological, cytological or experimental facts indicate such an origin, are examplified partly
by the dubious species Capsella bursa pas toris and Erophila verna
(both with prevailing, but not obligate, autogamy), partly by
groups of nearly related and frequently hybridizing allogamous
species (Viola. tricolor and V. aroensis, Geum rivale and G. urbanum,
Salix caprea and S. viminalis, comp. also the practical application
of the term in the latter sense in the genus Galeopsis by MUNZING
1930). According to TUREssoN's definition any taxonomical unit
superior to the species may be called a coenospecies, If any new
terms are needed for the lowest of those units, I prefer DANSEH'S
more definite terms commiscuum and comparium (comp.
below pp. 387 and 398).
The distinction between simple and compound species recently proposed by COCKAYNE and ALLAN (1927) certainly is very
useful, though it may well be discussed which species are really
worthy of being called simple. According to COCKAYNE and
ALLAN (1927 pp. 259-260, comp. also COCKAYNE 1928 p. 7, 1929 p.
22) a simple species is a species that contains only one ))jordanon
(comp.. above p. 350), while a compound species is made up of
two or more [ordanons, But. even in the simplest sexual species
certainly intensive analyses will reveal a very great number of
slightly differing forms, and in strictly allogamous species even
every individual probably represents its own biotype (comp. above
p. 340). Thus the only strictly simple species existing would be
those consisting of a single strictly asexual biotype, while all
other species would be more or less compound.

362
As the species-concept is taken in this paper, it is necessary to
distinguish at least four main types of species:
1. Asexual species (species consisting of one or several strictly
asexual biotypes).
a. S imp 1e as e x u a 1 s pee i e s (species consisting of one
strictly asexual biotype). To this type probably many apomictic species of Hieracium sect. Vulgata, Taraxacum, etc.,
in first line those with very small distribution-areas, must
be referred, and possibly also a part of the lichen-species
never developing apothecia but constantly propagated by
so redia, isidia, or undifferentiated thallus-fragments (comp.
above p. 345).
b. Com p 0 u n d as e x u a 1 s pee i e s (species consisting
of several, practically indistinguishable, strictly asexual biotypes). To this type belong probably many of the apomictic species of A lchem illa, many of the more widely.
distributed apomictic species of Taraxacum and Hieracium
sect. Vulgata, and many of the asexual lichen-species
mentioned above.
As has been pointed out by SAMUELSSON (1922) and K.
JOHANSSON (1927), many of these apomictic species in
Alchemilla, Hieracium, Taraxacum etc. were recognized
by taxonomists long before the discovery of apomixis in
the genera concerned. Quite independently of the interesting explanation furnished by the apomixis, they were
thus' recognized in the field as natural, sharply delimited
units analogous with other species. Certainly it would
involve a good bit of absurdity, if, following the advise
of certain modern geneticists, we would take the discovery of their apomixis as sufficient reason to reduce them
to units of lower rank than that previously given to them
on "the base of unprejudiced field-studies.
2. Sexual species (species consisting of a great number of freely
intercrossing - or partly autogamous - biotypes).
a. Homofacial sexual s p e c i e s O'Arten sensu RENSCH).
Regionally undifferentiated, i. e. consisting of only one
regional facies (subspecies)..
b. He tel' 0 f a cia 1 sex u a 1 s p e c i e s ( Hassenkreise sensu
RENSCH). Regionally differentiated, i. e. differentiated into
several regional facies (subspecies).
As a close survey will probably reveal local facies
(varieties) also in most species of the first of those groups
(at least in those with more or less stationary individuals),
the term homofacial s is not quite adequate - but as I
have not been able to find out any better term, it may
\

363
be used at least provisionally. Many homofacial sexual
species give the impression of a simple species sensu
COCKAYNE and ALLAN, as they consist of biotypes differing
only very slightly from each other. But many would be
referred to COCKAYNE'S and ALLAN'S compound species,
obviously containing biotypes differing enough from each
other to be easily recognized as different forms. This will
be the case even in many highly inbred species with
very small distribution-areas. In Gentiana antipoda, for
instance, endemic in Antipodes Island, one anthocyaninic
and one non-anthocyaninic form are easily recognized,
(KIRK 1895), and I found just the same in G. antarctica,
endemic in Campbell Island.
Examples of heterofacial sexual species are given above under
the subspecies (pp. 356-357). Of course they all belong to COCKAYNE'S
and ALLAN'S type compound species.
As was mentioned above already in the definition of the species,
the limits of sexual species may he of different kinds. While
some species are separated from each other by complete sexual
isolation, others are connected by (comparatively small) transitional populations of various kinds (hybrids). It appears most convenient to discuss those two kinds of species-limits separately.
I.

Species separated from each other by complete sexual isolation.

The limits between such species may be said to form distinct

lin e S 0 f dis con tin u i t Y in the series of biotypes. They may
be formed either by geographical or by merely physiological isolation. In the former case, the species concerned may be perfectly
fertile with each other when brought together in an artificial
way (comp, DANSER 1929 b p. 425), or the geographical isolation
may be guarded also by physiological isolation, i. e. by an
inherent unability of the individuals from one species to cross
with individuals from the other. In the latter case (i. e., when
there is no geographical isolation but only physiological isolation
of species growing together in the same area), isolation may be
due to any of the following causes. 1. Differences in the .breeding
(or flowering) season or breeding places (the latter only in animals,
cornp. PLATE 1913 pp. 527-531, ROBSON 1928 pp. 122-124). 2.
Morphological, physiological or in animals - psychological
unability of the individuals of one species to cross with those of
the other species (comp. ROMANES 1886, 1897, GULICK 1872-1905,

364
PLATE 1913 pp. 531-549, ROBSON 1928 pp. 122-130). 3. Anyone
of the two most severe types of diplontic sterility described by
MUNZING (1930 p. 318), i. e., the dying of the diplont at the zygote
or embryo stage, or the incapability of the seedlings of an independent existence.
In cases of purely geographical isolation it is often extremely
difficult to decide whether the isolated populations should be
treated as separate species or as mere parts of the same species.
Here four different cases must be discussed separately.
1. If all the individuals of one population differ both greatly
and constantly from those of the others, everybody will agree that
the two populations represent different species. Thus everybody
who has studied the Gentians of the subantarctic islands of New
Zealand will agree that Gentiana antipoda of Antipodes Island,
G. antarctica of Campbell Island, and G. cerina of the Auckland
Islands (of which G. concinna of the same islands probably is only
a form; comp. KIRK 1895) are good species, though clearlyrepresenting one natural group (series) of vicarious species (Artenkreis

sensu RENSCH).

2. If all individuals of one population differ constantly but

very slightly from those of the others, many taxonomists will
treat the two populations as parts (subspecies or varieties) of one
species, while others (and among them myself) will keep them
apart as separate species. According to the literature, such a case
appears to be represented by the middle-european Pinus cembra
and the Northeast-European and West-Siberian Pinus sibirica, by
many authors (and also by myself in a previous paper, comp.
Du RIETZ 1930 pp. 320-321) treated as mere subspecies or even
as varieties of one species, Other cases are commonly found in
species isolated in the different islands of one island-group (cornp.
RENSCH 1929 p. 100 etc.) or in the different mountain-massifs of
one mountain-chain. A good example is furnished by Pleurophyllum Hookeri (Compositae) in the subantarctic islands of New Zealand, by previous authors regarded as one species occurring in
identica], forms in the various islands, but according to my own
observations in the field (in the Auckland and Campbell Islands,
supplied with studies of herbarium-material from those islands
and Macquarie Island) represented by slightly but probably constantly differing populations (= species) in those three island-groups.
3. If the average composition of one population is distinctly

365
different from that of the others, but certain biotypes from one
population so closely resemble biotypes from the others that they
must be referred to the same forms, most taxonomists will probably agree that the populations do not represent different species,
but mere subspecies of the same one. Theoreticallv, the opposite
opinion would be quite possible to defend, but practically it appears
imperative to delimit species in such a way that it is possible to
determine any well-developed individual also without knowing its
provenance. At present I am not able to demonstrate any botanical
example of the case concerned - the average composition has
been properly investigated in very few plant-species - but a very
good zoological example is afforded by CRAMPTON'S (1925 b) thorough
analysis of the Partula gibba -population of the two Marianaislands Guam and Saipan.
4. If the two populations differ only slightly in average composition, nobody of course would think of treating them as
different species or even subspecies, in spite of their isolation from
each other.
II.

Species connected with each otber by transitional hybridpopulations.

In this case the limits between the species may be said to form
not lines but z 0 n e s 0 f dis con tin u it Y in the series of biotypes. Those zones may be of several different types.
1.

The transitional hybrid-population consists

o f mer ely s tel' i lei n d i v i d u a I s.

This type of transitional hybrid-populations is certainly very

common, and very many examples are described in botanical and
zoological literature. The sterility may be of many different degrees. In extreme cases, leading over to the two severest types of
diplontic sterility mentioned above (p. 364), no reproductive organs
at all are formed, though the vegetative development may be
quite normal. Examples of this subtype are mentioned by FOCKE
(1881 p. 477: some hybrids of Rhododendron, Epilobium, Cereus
and Hymenocallis), HERIBERT-NILSSON (1918: Salix oitninalis X [ragilis), and MUNZING (1930 pp. 318-319). To the same SUbtype
possibly belong the following New Zealand Celmisia-hybrids,' of
which I only succeeded to find non-fl6wering individuals (I do not
25 -

30531. Sven;k Bolanisk 7'idskrijt. 1930.

366
know whether they have been found flowering by my New Zealand colleagues, to which they are probablyall well-known, though
little has been published about them yet): Celmisia coriacea X Lyallii
(mentioned already by COCKAYNE 1923 and 1925, found by myself
in several individuals on Sealey Range in the Mt Cook region and
on Bold Peak in the Humboldt Mts '}, C. coriacea X Petriei (Bold
Peak, only one individual), C. Lyallii X petiolata (Sealey Range in
several individuals), C. petiolata X Petriei (Lake Harris Saddle,
only one individual), and C. petiolata X sessilijlora (on the ridge
S. of Lake Harris Saddle, only two leaf-rosettes). The same may be
the case with the remarkable hybrid Olearia Crosby-Smithiana X
ilicifolia, at the time of my New Zealand visit known only in a
single and - as far as I know - non-flowering shrub at Wilmot
Pass in the Fiord District (discovered by Mr. LESLIE MURRILL,
who brought Mr. W. R. B. OLIVER and me some branches during
our excursion to Doubtful Sound).
Another subtype is represented by the hybrids in which reproductive organs are developed but more or .less defective, either
morphologically or only physiologically (comp. MUNZING 1930 pp.
314-322). The sterility of many hybrids has been well known
since very long ago (comp. FOCKE 1881 p. 477 and ROBERTS 1929),
and from time to time even the theory has been advanced that
a II true specles hybrids should be at least more or less sterile.
Thus the investigation of the seed-production and of the quality
of the pollen has become a well-established method in taxonomy for the testing of supposed hybrids. Abundant examples
of sterile hybrids are easily found in botanical (and zoological)
literature (comp. for instance FOCKE 1881, KERNER 1898, DE VRIfs
1903, MUNZING 1930, and modernfloristical handbooks, especially
those of Scandinavian countries). A very interesting account of
the hyhrids in the section Lapathum of the genus Rumex is given
by DANSER (1924); some of those hybrids are completely sterile,
while others are nearly sterile, producing only very few germinable seeds (comp. below p. 3(8). As to the New Zealand flora, I
do not know any hybrids of this subtype with certainty (comp.
COCKAYNE 1929 p. 11), but there is little reason to doubt that it
is common there too, though not as well known yet as fertile and
highly polymorphic hybrids. As mentioned by MUNZING (I. c.),
1

G.

On

Bold Peak my Celmisia-studies were made in company with Mssrs

and J. S. THOMSON.

SIMPSON

367
there are ~any degrees even of complete sterility in hybrids, from
types with the flowers degenerating already in the bud-stage to
types with flowers of quite normal appearance, but with no seedproduction. In the latter case, according to MUNZING, either diplontic or haplontic sterility may be responsible.
Hybrids consisting of perfectly sterile individuals are always
fairly homogenous, though some polymorphy may. be brought about
by intraspecific differences between biotypes involved in the interspecific crosses. They do not, as a rule, play any considerable
role in nature, and do not considerably disturb the sharp limits
of the species concerned.
The t r a n s i ti 0 n a I h Yb r i d - pop u I at ion con sis t s
of more or less fertile individuals.
The existence of more or less fertile hybrids has been well
known, though often doubted, since the time of KOLREUTER (1761,
comp. KERNER 1871 pp. 34-35). FOCKE (1881 p. 477) enumerates
quite a lot of genera in which fertile hybrids are more common
than sterile ones. The considerable polymorphy of many of those
fertile hybrids was also well known (cornp. FOCKE 1. c.). Still it
is first in recent years that highly polymorphic hybrids have
come to their full right in the discussion of the species-problem.
Also in this type of transitional hybrid-populations two subtypes
may be distinguished. In the first subtype the transitional hybridpopulation is obviously less vital than the two species-populations concerned, while in the second subtype the hybrid-population exhibits
a vitality quite comparable with that of those species-populations.
In the first subtype of fertile hybrid-populations, comprising those with a vitality clearly
lower t han t hat 0 f the s p e c i e s co n c ern ed, the
hybrid-population may obscure the limit between the two species
only rather- slightly, as the hybrid-population will playa rather
modest role in nature compared with that of the two speciespopulations, and will not be able to enter into serious competition
with them in any habitat. The lack of vitality in the hybridpopulation may be due to several causes. Though not completely
sterile, the biotypes of the hybrid-population may in many cases
exhibit a rather low fertility. Both in those cases and in the
cases of full fertility the vi t a lit Y of the hybrid-individuals may
he too low to enable them to compete successfully with the species.
2.

368

Some boreal examples of hybrids of this subtype may first be

mentioned. According to HERIBERT-NILSSON (1918), the hybrid between Salix caprea and S. viminalis cannot in spite of its fertility
compete with the parent species, owing partly to certation phenomena, and partly to characters lowering the vitality in various
ways. Many of the Salix-hybrids found in the indigenous vegetation of Sweden obviously belong to the same type. - In the hybrid
Viola arvensis X tricolor according to CLAUSEN (1926) there are no
biotypes able to compete successfully with the. parent species,
neither in experimental plots nor in nature. - Similar cases are
found in some of the hybrids of Rumex sect. Lapathum, e. g. in
Rumex aquaticus X hydrolapathum (R. maximus, comp. DANSER 1924
Pr>. 186-188). Here fertility is always much lower than in the
parent species, and the hybrids have therefore no chance in the
competition with those species. - Just the same, according to
E. ANDERSON (1928) is the case in the North American hybrid
Iris versicolor X virginica.
Analogous cases are easily found in the flora of New Zealand.
Among the best examples are those provided by the hybrids
Ranunculus Buchananii X Lyallii (R. MaUhewsii Cheesem., comp.
COCKAYNE 1921 p. 359, 1923 p. 118, 1925 p. 63, 192~ p. 240, ALLAN
1926 c p. 534) and Ranunculus Buchananii X Simpsonii (comp.
COCKAYNE 1928 pp. 241, 301, 302) both of which I have had excellent opportunities to study in the field in the upper Routeburn
Valley in the Fiord District of South Island (in the company of
Mssrs G. SIMPSON and J. S. THOMSON). Ranunculus Lyallii is an
important dominant of the sub a I pin e meadows throughout the
wet mountains of the southern and middle parts of South Island,
R. Buchananii forms small meadow-fragments in wet rock-crevices,
rock-ledges, rock-slides etc.' in the a I pin e belt of the Fiord and
South Otago Districts 1 only, and R. Simpsoniii is one of the most
As to the Botanical Districts of New Zealand cornp. COCKAYNE 1928.
Possibly R. Simpsonii Ckn and Allan (1926 b p. 58) will turn out to be better
treated as a subspecies' .of R. sericophyllus Hook. f. (Western District), with which
it was united by CHEESEMAN (1925 p. 444) and which it replaces in the Fiord
District. On Sealey Range in the Mount Cook region, where R. sericophyllus is
abundant, I found a continuous series of biotypes connecting the typical villose
forms with nearly glabrous forms indistinguishable from the true R. Simpsonii of
the Fiord District. Further field-studies by New Zealand botanists must decide
whether there is a comparatively narrow hybrid-zone between the pure R. sericophyllus - population of the north and the pure R. Simpsonii - population of the
1

369

important dominants of the open chionophyte-meadows ' in the

alpine belt of the Fiord District. R. Buchanonii and R. Lyallii
frequently meet on steep slopes along the alpine-subalpine border
in the Fiord District, and in those places the highly polymorphic
hybrid is found, forming a very fine series of transitional forms
connecting the two species. In my rather rich collection of hybrid-specimens, the development of the pollen varies from individuals with n ear I y all pollen-grains normally developed to such
with more than 50 % shrivelled. In one specimen a great reduction
of the seed-production was observed, not 50 % - and in some
flowers much less - of the achenes being fertile; if this is the
rule or not I do not know. I never saw the hybrid dominating. Even.
in the transitional belt the hybrid obviously has no chance in compe.tition with the two species. - This is still more the case with
the hybrid R. Buchananii X Simpsonii. Though those species are
frequently growing quite near each other in the alpine belt (they
are even frequently met together in the same community), the
hybrid was found only in one place, namely on a large wet
shingle-slip suitable for both species. The hybrid was very polymorphic, connecting the two species by a series of transitional
forms; four extravagant individuals with more or less transformed
petals were found. The pollen was bad: in many specimens nearly
all pollen-grains were shrivelled, while in others nearly 50 % (but
not more) could be normally developed. Though a good number
of individuals of the hybrid were found (growing among the mixed
parent species), it was quite obvious that it had no chance at all
in the competition.
None of those two Ranunculus-hybrids can be said to disturb
the delimitation of the species in any considerable way. The
hybrid-individuals are quite easy to pick out, and compared with
the large populations formed. by the pure species they playa
very insignificant role in nature.
south, in which case the two populations must be kept as species, or a more
continuous transition, in which case they would be better treated as subspecies
of one species (comp. above pp. 356-357).
I Chionophgtes (from X!rov= snow) are plants dependant of a very long snowcovering ,(= Schneebodenptlanzen> in German). According to their relation to
the length of the snow-covering (certainly one of the most important ecological factors in the alpine belt of all mountains) the alpine plant-communIties may be
divided into four gro~ps: chiotiophilous, subchionophilous, subcbionophobous and
chionophobous,

370
In the New Zealand species of Nothofagus perhaps still better
examples of this subtype of species-limits are found. Nothofagus
Solandri and N. truncata have about the same main distributionarea in the Beilschmiedia tawa - belt of New Zealand (= the
lower parts of North Island, except in the extreme north, and.
northernmost South Island), and to a large extent form mixed
forests, but nevertheless the hybrid between those two species is
only found in single individuals or a few trees side by side.
Though the hybrid is highly polymorphic and evidently more or
less fertile, forming a series of transitional forms connecting the
two species, its vitality and power of competition are apparently
so inferior to those of the two species, that the species can stand
out as sharply delimited units even when growing together in the
same forest, e. g. in the beech-forests near Wellington, where
Dr. L. COCKAYNE and Mr. E. ATKINSON kindly introduced me to
this hybrid (comp. COCKAYNE and ATKINSON 1926, COCKAYNE 1926
and 1929, COCKAYNE and. PHILLIPS TURNER 1928.) The hybrid
Nothofogus cliffortioides X fusca is evidently more vital. It is common in the mixed N. cliffortioides-fusca-forest of South Island, is
obviously quite fertile and exhibits an extreme polymorphy (CockAYNE 1923, 1926, 1929, COCKAYNE and ATKINSON 1926, COCKAYNE
and. PHILLIPS TURNER 1928). Generally in Nothofagus cliffortioidesfusca forest, hybrids form a very small percentage of the trees,
but perhaps 5-}O per cent occur with fair frequency. In some
places, however, there may be at least 50 per cent of hybrids,
but such spots are small in extent. Nevertheless, it is clear that
the hybrids can well hold their own, a statement substantiated
by the presence of hybrid seedlings in great numbers beneath
hybrid trees. After a forest fire hybrids may occur in great profusion and the reinstated forest will certainly contain more hybrid
trees than did the former community (COCKAYNE 1929 p. 14).
Still even this hybrid is not vital enough to prevent the two species from standing out as quite distinct units. When studying it
in the Dart Valley and at Kinloch TW. end of Lake \\;akatipu)
I got the impression that the remarkable frequency of the hybrid
in you n g forest mentioned by COCKAYNE and easily observed at
Kinloch was only ephemeral, the vast majority of those young
hybrid-individuals getting lost in the long run in competition with
the parent species. In the mature Nothofagus-forest there is room
only for. quite few of the trees of the young forest, and in this
j

371

final tree-layer of giant Nothofagus-trees the hybrid-trees appear Lo

have small chances to get a place. In old, mature Notliofaqus
cliffortioides-fusca-forest I never found more than single individuals
of the hybrid. If the habitat is suitable for N. fusca, which is
much more pretentious than N. cliffortioides, even the latter species
appears to have small chances to survive in the very' old forest.
In the second subtype of fertile hybrid-populations, comprising those with the vitality of the
hybrid-population equalling that of the species
con eel' ned, we meet the most difficult cases of species-delimitation that may confront a taxonomist. In those cases, the
species are saved from being completely swamped only by occupying different distribution-areas or at least different habitats in
the same distribution-area. In the areas (or habitats) where they
meet, they fleet together into one vast syngameon, the greater
part of which may clearly belong to the hybrid. If this hybridarea is m u c h smaller than the areas occupied by the pure speciespopulations, the species may still stand out as quite distinct units.
But if the hybrid-population is of a size comparable with that of
the pure species-populations, the taxonomical treatment of the
populations is extremely difficult. And still worse is the case
when the hybrid-population tends to outnumber the pure species.
But as long as so me areas of considerable size are occupied by
the pure species, there is still a possibility left to keep the species
distinct, though in all those cases the question may arise whether
it would not be a better method to treat the two species and the
hybrid only as different subspecies of one heterofacial species. It
is in those cases that pure herbarium-taxonomy proves to be
utterly unable to solve the problems (and pure garden-genetics
too 0, and that the necessity of age 0 g I' a phi c a I fie I d - m etho d will stand out in the most imperative way.
It is often rather difficult to decide whether a certain hybrid
should be referred to this or to the preceding subtype. This is
especially the case when the species concerned occupy only
different habitats in the same distribution-area. In the hybrid
Geum riuale X urbanum, for instance, it appears to be somewhat
-doubtful whether the comparatively small role played by the
hybrid in nature is due partly to fertility- and vitality-factors
(ALMQUIST 1926 pp. 67-68, TUREssoN 1929 a p. 329) or only to
the differences between the habitats of the two species and the

372

supposed scarcity of intermediate habitats (WINGE 1926 b p. 594)

combined with complex heredity of certain groups of characters
(WINGE 1928 p. 231). In the following discussion, however, I will
try to choose such examples where little doubt is possible regarding the full fertility and' vitality of the hybrids.
Let us begin then with some examples of species forming
hybrids with full fertility and vitality, but still kept distinct by
the adaptation of the two species to different habitats in the same
main distribution-area.
The example of such species most frequently met with in
genetical literature is the species-pair Melandriutn album (Melandrium
praiense. Lychnis diurna) and M. rubrum (M. silvestre, Lychnis
vespertina). While geneticists of the purely experimental gardenschool may fail to recognize those two species as natural units,
owing to the full fertility and vitality of the hybrid (comp. SHULL
1923), field-working taxonomists (and geneticists combining the
experimental garden-work with comparative field-studies) find little
or no difficulty in keeping the two species distinct. In the areas
where both species occur, they occupy so different habitats that
the hybrid cannot play any considerable role compared with the
pure species (comp. HAGEDOORN 1921 pp. 132-134). In Sweden
only M. rubrum belongs to the indigenous vegetation, being common
in somewhat wet forests of the meadow-type in most parts of the
country, and extending even into the open meadows of the lowalpine belt (according to TURESSON 1925 a pp. 184-189 it is
differentiated in four more or less distinct ecotypes or subspecies).
M. album. here is an introduced weed hardly entering the indigenous
plant-communities. According to. ALMQUIST (1926 pp. 66-67), the
hybrid is especially common in cultivated fields in some parts of
North Sweden, but it may be found everywhere where the two
species grow near each other. It does not enter the indigenous
plant-communities, and the wild M. rubrum - population mostly
keeps quite pure.
.
A good example of species occupying so different habitats within the same area that the limit between them is kept distinct
in spite of the vitality of the hybrid, is afforded by Salix
herbacea and Salix lapponum in the mountains of Scandinavia. Salix lapponum is one of the dominants of the willow-scrub
covering much of the moist ground in the low-alpine belt, while
S. herbacea is a minute dwarf-shrub or even hemicyptophyte,

373
being one of the most important dominants of the middle-alpine
and high-alpine belts (comp. Du RIETZ 1930 p. 354), bnt in, the
low-alpine, belt playing a role only -in the extreme chionophytecommunities. The hybrid is quite fertile and exhibits a remarkable
polymorphy, connecting those so different species by a series of forms
(comp. FLODERUS 1909 pp. 25-27). It may be found wherever the
species meet, and is mostly quite common especially along streams in
the low-alpine belt. According to FLODEHUS (I. c.) it may occasio.nally even be so abundant that it forms large con ti n uous mats
lying so close to each other that they give a dominating character
to the vegetation-picture>. But this is only in especially favourable
districts where there are large tracts. of country belonging to the
low-middle-alpine transition-belt. The vast majority of the Salix
herbacea- and s. lapponum - populations are kept comparatively isolated from any chance of hybridizing with each other, and
compared with those immense pure populations, the hybrid plays
such a small role in nature that nobody would think of uniting
S. herbacea and S. lapponum to one species or even considering
the limit between them as bad.
Analogous cases are easily found in the flora of New Zealand.
One of the best examples is certainly that formed by Dracoplujllum
[ili[oltum and D. recurvum (OLIVER 1928, ALLAN 1928), which I had
splendid opportunities to study on Mount Conspicuous in the
Ruahine Range (under the guidance. of Dr. H. H. ALLAN) and
especially on Ruapehu (in the central part of North Island). Both
species have their main distribution-area in the subalpine belt of
the central mountains of North Island, though' the area of D.
recurvum extends to Hikurangi in the East Cape District and that
of D. [ili[olium to Mt Egmon t and to Cook Strait. D. [ililolium
is one of the dominants of the closed subalpine scrub, while D.
recurvum is one of the two leading dominanls of the open dwarfshrub heath covering most of the wind-exposed ridges. In its
typical state the latter species is a prostrate dwarf-shrub, while
the former is an erect shrub. They. are connected by a hybrid
consisting of a most polymorphic series of forms and, if I may
rely upon my own observations on Ruapehu, these exhibit a vitality
fully equal to that of the pure species. In the dwarf-shrub heath
.of the wind-exposed ridges D. recurvum keeps pure, as D. [ilifolium
does mostly in the closed scrub (though not to the same high
degree), but in the open tall tussock (Danthonia Raoulii) vegetation

374
dotted with low shrubs the two species meet, and there the limit
between them is completely smoothed out by the hybrid in its
innumerable forms, mostly quite outnumbering the pure species.
As much of this open tussock-country may be due to fire, the
role played by the hybrid has possibly increased in comparatively
recent years.
Another similar example is found in the same genus, namely
the species-pair Dracophyllum Olioeri! and D. prostratum, both
occurring on more or less boggy ground in the South Otago and
1 Under this name I propose to raise OLIVER'S Dracophyllum lonqifolium f. 3"
(OLIVER 1928 p. 702) to the rank of a species. Though OLIVER is pointing out
the compound character of his (and previous authors') D. longifolium, he provisionally keeps it as one species, dividing it into three forms only. As far as my
experience goes, those three forms are quite as good and definite units as many
Dracophyllum-species recognized by OLIVER and previous authors, and the lumping of them to one species seriously hampers the discussion of their most
interesting hybrids with other species. I therefore propose to raise them all to
the rank of species. The name D. lonqifolium then must be kept for OLIVER'S
I. 2, as FORSTER'S Epacris longifolia (and ROB. BROWN'S Dracophyllum longifolium)
clearly belongs to this species (type.J.ocality Dusky Sound). Certainly this is a
very compound species even in this restricted sense. - OLIVER'S r. 3 is a rather
different population, by COCKAYNE (1928 p. 205) apparently referred not to
D. longifolium but to D. Urvilleanum (sens. ampl.). It consists of much smaller
shrubs with much shorter leaves closely resembling those of D. rosmarinifolium
(= D. uniflorum Hook. f.) and with smaller racemes somewhat approaching those
of the D. Urvilleanum - group. When Mr. OLIVER and I stayed together at
Manapouri, he kindly supplied me with good material of it from the bog E. of
Manapouri mentioned by COCKAYNE. My own collections contain specimens of it
from a bog in the lower Routebourn Valley (comp. below p. 375); probably also
some somewhat deviating specimens with longer leaves from Maungatua should
be referred to this species. When Mr. OLIVER and I discussed it in Wellington before my departure from New Zealand, he was also inclined to raise it to the
rank of a species. I propose to attach his name to it as an appropriate acknow.ledgement of his excellent and much-needed revision of the genus. - As to
OLIVER'S f. 1, which includes the D. longifolium>-population of the Auckland
and Campbell Islands, HOOKER already pointed out the difference between it and
the D. longifolium - population of New Zealand proper, and in his Flora novaezelandiae> (1853) he even described them as two separate species. However, he
referred 0 n e specimen from New Zealand proper, collected by FORSTER in Dusky
Sound, to the subantarctic species, and therefore applied FORSTER'S species-name
longifolium to that species, while he gave the name D. Lyallii to the common
species of New Zealand proper. Later on (1867) he again united the two species.
Considering the geographical isolation of the two populations concerned, even
the smallest con s tan t difference between them should be sufficient to assure
them the rank of separate species (comp. above p. 364). And as far as I know

375
Fiord Districts of South Island, the former species probably also
extending further north. D. prostratum is a prostrate dwarf-shrub
creeping and rooting in the peat of the extreme bog-communities
(Donatia-Oreobolus-bog, Sphagnum-bog etc."), while D. Oliveri is a
low erect shrub hardly entering those 'communities. As far as I
know, the two species mostly keep quite pure (the hybrid is not
mentioned by OLIVER 1928), but in a Sphagnum-bog in the montane belt of the lower Routeburn Valley (W. of Lake Wakatipu,
near the junction of tbe Routeburn and the Dart Valley), I found
a most bewildering multitude of forms gradually connecting the
two species and evidently belonging to the hybrid, which here
showed a great vitality and quite outnumbered the pure species.
This bog was of a very peculiar type, apparently offering both
species suitable conditions in the same plant-communities.
I cannot refrain from mentioning a third example from the
same genus in this connection, namely the Dracophgllum-population
of Campbell Island, which I had some opportunities to study in
collaboration with Mr. W. R. B. OLIVER. It consists of the two
species D. scoparium (= D. subantarcticum Ckn) and D. Cockaqnianum (= D. longifolium f. 1 in OLIVER'S revision, comp. the note to
there are even g l' eat constant differences between them: to those mentioned
by OLIVER there must be added one very important difference, namely the constantly pubescent upper surface of the adult leaves (but not of the juvenile ones,
a remarkable diUerence noted already by HOOKER 1847 p. 46) in the subantarctic
population. In the true D. lonqifolium of New Zealand proper I have never seen
anything like this pubescence; only in specimens from Stewart Island (Port
Pegasus) I found a vel' y s I i g h t pubescence on the upper surface of the
leaves (of much shorter hairs than in the subantarctic population). I therefore
do not hesitate in proposing to raise also the DracoplujllumLonqifolium of the
Auckland and Campbell Islands to the rank of a separate species. Could any
name be more appropriate lor this beautiful tree than Dracophyllum Cockasjnianum, recalling the everlasting picture of subantarctic vegetation given by the
grand old man of New Zealand Botany?
1 In
the subalpine belt, D. prostratum may be found also on dryer ground.
It may then assume the life-form of a cushion-plant (on Bold Peak at the W.
end of Lake Wakatipu I saw a big cushion sending out prostrate stems of the
normal type). I have not been able to find any .constant difference between this
lorm and D. muscoides Hook. I. Specimens of the latter from the Old Man Range
in Central Otago, kindly supplied by Mssrs G. SIMPSON and J. S. THOMSON, have
just the same leaf-size as D. prostraium, not smaller as should be the case according to OLIVER 1928. If further Held-studies should confirm the suspicion that
D. prostratum and D. muscoides are only forms of one species, this species must
be called D. muscoides Hook. f.

376

p. 374), and the hybrid between them (COCKAYNE 1928 p. 335,

D. insulare Oliver 1. c.). According to my own notes D. Cockayniatuun was the dominant species (but mixed with D. scoparium
and the hybrid in many forms) in much of the comparatively tall
scrub traversed by myself on the southern side of Northeast
Harbour, while in the literature the other species is. always mentioned as the main dominant. Probably D. Cockaynianum may
domifolate only in the most protected tallest scrub, while the more
low-growing D. scoparium plays the chief role in most of the scrub
of the island; the latter species also grows dotted about in the
open tussock-country (comp. KIRK 1891 p. 223, COCKAYNE 1904
pp. 274, 277). The hybrid appears to be abundant, fertile, vital,
and very polymorphic, connecting the two species by a multitude
of transitional forms. Nevertheless the two species have given all
visitors the impression of two good and distinct units. This may he
due to the comparative isolation of a large D. scoparium - population in the more open parts of the island. However, it may be
well worth to look more closely into the D. Cockaynianum - population of the island, as I have some suspicions that all of it may
be at least slightly infected with D. scoparium (the specimens I
brought home from Campbell Island as pure D. lotiqifolium
have clearly shorter and more pubescent leaves than those from
the Auckland Islands, where D. scopariutu is lacking).
In this connection probably also the classical New Zealand
hybrid Coprosma propinqua X robusta (C. Cunninghamii Hook. f.)
should be mentioned (comp. COCKAYNE 1923, 1928, 1929, ALLAN
1924, 1926 a, b, 1929 a). Though both C. propinqua and C.
robusta are distributed over at least most of the lowlands of New
Zealand, they mostly appear to grow in different plant-communities (I got the impression that C. robusta is more at home in
dense forest, while C. propinqua is frequently found also in more
open scrubland). Especially in more or less swampy and not
too dense forest they often meet (e. g. in Riccarton Bush at
Christchurch, where this interesting hybrid-case was demonstrated.
to me by Dr. COCKAYNE and Dr. ALLAN), and then are 'connected
by a very fertile and vital hybrid-population forming a complete
series of transitional forms.
The examples given above all referred to species forming
hybrids with full fertility and vitality but kept distinct by the
adaptation of the two species to different habitats in the same main

377
distribution-area. The existence of large tracts of habitats suitable
to one of the species but not to the other (and vice versa) guarantees, of course, the existence of pure populations of each species
large enough to prevent the swamping of the two species to one
large undifferentiated hybrid-population (comp. HAGE:DOORN 1921).
Just the same thing will happen when two species, forming a
fertile and vital. hybrid, have geographically different distributionareas overlapping only to a comparatively small degree.
It is quite easy to find any number of good examples of such
species-pairs. One of the best European examples is certainly
the species-pair Veronica longifolia and V. spicaia. V. longifolia is
a comparatively hygrophilousspecies characteristic for the alluvial meadows of North-eastern .Europe, while V. spicaia is a
xerophilous stepp-species of more south-eastern distribution (STERNER 1922). In Sweden V.spicala has a typical southern distribution,
while V. longifolia has its main distribution-area in the Northeast (comp. STERNER I. c.). A comparatively distinct subspecies
of the latter species (subsp. maritima) is very common on the
small exposed islands of the Swedish east-coast N. of lat. 58.
In the southern part of this distribution-area, V.longifolia suhsp.
maritima meets V. spicata and is connected with it by a fertile,
vital and very polymorhic hybrid-population. In the province of
Upland, only pure V. longifolia subsp. maritima is. found in the
outer belt of the archipelago, and (some clearly synanthropous
colonies of V. longifolia excepted) only pure (or at least nearly
pure) V. spicaia in the inner part of the mainland. But between
those belts there is a transitional coastal belt where the hybrid
is common and mostly quite outnumbering the pure species (comp.
ALMQUIST 1929 pp. 589-590). In a very rich collection made
by me in the summer of 1908 in one of the classical. localities
for this hybrid (Fjaderhclmarna in the harbour of Stockholm)
there is a most bewildering multitude of hybrid-forms but not a
single specimen that may be referred to either of the pure species.

Probably both parent-species have entirely disappeared from this

locality, the hybrid here appearing as a Ganzwaise- Tn the sense
of GAMS (1923). Still this hybrid-zone is of such a small extension
compared with the vast areas in which only one of the pure species
are found, that probably no botanist would get the idea to call
V. longitolia and V. spicaia bad species.
Another good example is the species-pair Heliantbemum nummu-

3iS
larium and H. ova tum in South Sweden (comp. Du RIETZ 1923 b,
1925). H. nummularium has a more eastern, H. ovatum a more
south-western distribution in South Sweden. In the province of
Blekinge, only pure H. nummulariumis found in the east and
only pure H. ovatum in the west, but in the middle there is a
zone with both species growing together and connected by a fertile, vital and very polymorphic hybrid-population. In this hybridzone certainly no distinct limit can be traced between those species
so distinct in the greater part of their distribution-areas.
A multitude of analogous cases have been described during the
last decade from New Zealand in a series of important papers by
COCKAYNE and ALLAN (1923-1929).
In several of those cases one of the two related species has a
more northern and the other a more southern distribution overlapping only to a comparatively small extent. Those species may
be quite distinct and very different-looking units in the areas
where only one of them is present, and still join into one vast
and most polymorphic hybrid-syngameon in the area where they
are meeting.. One of the best known examples for this type is
the species-pair Mgrtus bullaia and M. obcordata (COCKAYNE 1918,
1923, 1929, COCKAYNE and ALLAN 192i, COCKAYNE and PHILLIPS
TURNER. Hl28, ALLAN 1929). From the south of South Island up
to about Iat. 41 only pure M. obcordata is found, and in the
northern part of North Island (north of lat. 36) only M. bullata
(except in two very small areas), but between lat. 41 and 36
both species frequently grow together connected by a most luxuriant transitional population of the very polymorphic hybrid
(previously known as a species under the name of M. Ralphii
Hook. f., demonstrated to me by Dr. ALLAN in Kitchener Park at
Feilding).
In other cases one of the two species is restricted to the coastal
scrub while the other has a wide inland distribution. Hebe ellipfica and H. salicijolia form the best example for this type, the former
one of the chief dominants of the coastal scrub of South Island
(extending its distribution-area to the subantarctic islands of New
Zealand and even to the Falkland Islands and subantarctic South
America), the latter common in the lowlands throughout New
Zealand and frequently meeting H. ellipfica in the coastal scrub
of South Island. In those places a highly .polymorphic hybrid
connects the two species (COCKAYNE Hl23, 1925, 1928 p. 105,

379
COCKAYNE and ALLAN 1926, 192'7, ALLAN, S~MPSON and THOMSON.
1926). Upon various forms of this. hybrid the old species Veronica
(Hebe) amabilis Cheesem, and V. blanda (Gheesem.) Pennell were
founded. The high polymorphy and vitality of the hybrid is
clearly demonstrated by the thorough analysis of AI,.LAN, SIMPSON
and THOMSON (1. c). I saw it only in two localities (Punakaiki in
the Northwestern District, where it was introduced to me by Dr.
COCKAYNE and Dr. ALLAN, and Deep Cove in Doubtful Sound)' but
I got the impression that it does not enter the most exposed scrub,
this being formed by pure Hebe elliptica.
In the New Zealand flora there are also several cases of one
lowland and one mountain species connected by a luxuriant and
highly polymorphic hybrid-population confined to a certain altitudinal belt. Arisioielia serrata is a large-leaved low tree common
throughout the forested part of New Zealand up to about the
lower limit for the prealpine forest-belt ', A. [rullcosa a smallleaved divaricating scrub having its main distribution-area in the
prealpine and lower subalpine belts and in the dry low tussock
region of the Eastern side of South Island, but also frequently
occurring in the montane forest-belts (comp. COCKAYNE 1928, 1929).
In these montane forest-belts the two species may not only be
connected but even outnumbered by their very polymorphic hybrid (comp. COCKAYNE 1923, 1925, 1928, 1929, COCKAYNE and PHILLIPS TURNER 1928, ALLAN 1927 a, previously known as a species
under the name A. Colensoi Hook. f.) which may ascend even a
good bit higher than the pure A. serrata 2 (according to observations
made in the Routebourn Valley together with Mssrs G. SIMPSON
1 I am using the term prealpine belt. for the uppermost forest-belt of the
New Zealand mountains (characterized by Nothofaqus cliffortioides, N. Menziesii,
Libocedrus Bidwillii, Podocarpus Hallii and Dacrijdiumbiforme as main dominants
in the tree-layer, by the constant absense of ll'einmannia racemosa, Metrosideros
lucida and other trees common in the montane forest-belts, and by a shrub-layer
mostly very dense and more or less resembling the subalpine scrub, with which
it has many species in common), restricting the term subalpine belt to the
scrub and tall tussock belt above timber-line. COCKAYNE'S (1928) subalpine belt
thus includes both my prealplne and -subalpine belts as well as parts of
my "upper montane' belt. My terminology in this matter is based upon an
attempt to work out a generally applicable terminology for the altitudinal vegetation-belts of the earth, that will be published shortly.
t Hybrids occurring in this way outside the range of 0 n e parent-species are
called Halbwaisen by MURR (1919) and GAMS (1923).

380
and J. S. THOMSON). On the Huahine and Tararua mountains,
where this hybrid was first demonstrated to me by Dr. ALLAN, it
appeared to be one of the most frequent shrubs of the montane
and lower prealpine belts, and certainly posessed a vitality at
least equalling that of the two pure species. - A similar case is
that of Coprosma grandi{olia and C. lenui{olia. The former is a
large-leaved forest-shrub common in the lower forest-belts of North
Island and northernmost South Island (= the Beilschmiedia lara ireand B. lama - belts). The latter is a more small-leaved shrub confined to the higher forest-belts of the middle part of North Island
(according to CHEESEMAN 1925 from Mt Te Aroha and Hikurangi
in the North to Mt Egmont and the Ruahine Mts in the South,
and not recorded below 1000 It). In this area also the hybrid is
found (ALLAN 1928, COCKAYNE and PHILLIPS TURNER 1928, COCKAYNE
1929). After having made its first .aquaintance on the Ruahine
Mts with Dr. ALLAN, I got good opportunities to study it more
closely during a stay in the Tongariro National Park (between
Raurimu and Waimarino, on Hauhungatahi and at Ohakune).
The hybrid is very polymorphic, forming a fine transition-series
between the two species. As to its vitality, I got the impression
that it was about as equally common and vital as both species.
According to my notes, both species were common together with
the hybrid up to the lower prealpine belt (= a little above 1100
m), where they disappeared simultaneously. However, what I
brought home from the montane belt as pure C. grandi{olia proved
to be considerably more small-leaved than the guaranteed pure
C. grandi{olia of the lowland. I therefore have, some doubts
whether absolutely pure C. grandi{olia really occurs on the Volcanic Plateau much ahove the, upper limit of the tawa-belt (about
600 m). On the Tararua Mts, where C. tenuifolia is absent, I
noted the highest isolated individuals of C. grandi{olia at about
730 m, and in South Island it apparently does not much exceed
the limits of the tawa-belt. So it may well look possible that
the existence of a montane Coprosma grandi{olia - population on the
Volcanic Plateau may be due to the crossing of the pure C. grandifolia of the lowland with the more cold-resistant C. lenui{olia. As
to the latter species,' I hardly ever saw it isolated from the hybridpopulation. It does not seem impossible that the whole C. tenuifolia - population has got the opportunity of being infected with
C. qrandifolia. So we may here have one of those troublesome

381
cases where hybridization has gone far enough to become a serious menace to the autonomy of at least one of the species
involved.
This leads us over to the extreme cases in which the species
are more or less lost in an overwhelming majority of highly
polymorphic hybrid-populations. Many such cases have been
described during the last decades by the Swedish school of salicologists (S. J. ENANDER, B. FLO DERUS etc., comp. especially
FLO DERUS 1909, 1912, 1923, 1926 a, b) and their views certainly cannot be lightly dismissed, as they are founded upon
intensive field-studies carried out during half a century and
extended over large parts of the Northern Hemisphere. In the
later works of FLO DERUS there is an increasing tendency to class
most of the Salix-population of the arctic and subarctic regions
as hybrids, and to restrict the species-concept to include only
populations of the most extreme purity. Thus there are quite
a lot of Swedish Salix-species that according to FLODERUS are
hardly ever found pure in Scandinavia. As long as pure populations of those species are admitted to occur in other parts of
their distribution-area, this method of treatment may be easily
defended. But, if I have not misunderstood FLODERUS' recent
papers, many of his species never form pure .populations of any
extension, those species thus being known 0 n I y as single individuals or very small populations accidentally found here and
there in the highly polymorphic syngameons classed by FLODERUS
as hybrids. In those cases it may well be asked whether we are
not on a dangerous road that may easily Icad to complete dissolution of any practically applicable species-concept in those
populations.
If, as often happens, two Salix-species frequently hybridize in a
certain region, and the hybrid is fertile, vital and easily crossed
back with the two species, the genetical constitution of the two
species-populations is easily changed to such an extent, that practically every individual of one of the species will show at least
some slight traces of the other species. 'rhus, according to FLODERUS, the whole S. nigricans-population of the Scandinavian mountain-districts shows distinct traces of S. glauca and S. plujlicifolia, .
i. e., nearly every individual of S. nigricans in those districts
has taken up at least some genes from the neighbouring S. glaucaand S. phylici{olia - populations. And according to the same author
26 -

301131. Soensk: Botanisk: Tidskritt.

1930.

382
the S. nigricans - population of South Sweden is constantly intected>
in the same way by S. cinerea (and often also with some other
species). As a consequence of those statements FLODERUS refers
practically the whole S. nigricans - population of North Sweden to
the hybrids S. nigricans X phylici{olia, S. glauca X nigricans and S.
glauca X nigricans X phylici{olia (or to hybrids of still more complicated constitution), while the S. nigricans - population of South
and Middle Sweden is referred to the hybrid S. cinerea X nigricans
or to other, more complex, hybrids. - In the same way FLODERUS
treats the very polymorphic Salix-population of Greenland considered by him to be formed by the hybridizing of S. arctica,
S. chloroclados and S. glauca. Most of the population of North
Greenland is classed as S. arctica X glauca, most of that of Middle
Greenland as S. arctica X chloroclados X glauca, and most of that
of South Greenland as S. chloroclados X glauca. The pure species
are considered to be rare, S. glauca being found over the whole
country, but S. arctica lacking in the south and S. chloroclados
in the extreme north.
This method of treatment, of course, involves the theory that
the species distinguished are the primary units and the main
population classed as hybridsis younger than those. This, however,
is not proved. It appears quite possible that the smaller and
more uniform populations classed as species are sec 0 n d a l' y
units differentiated from the highly polymorphic syngameon classed
as a complex of hybrids, or even only extreme forms accidentally
appearing, disappearing and reappearing within this syngameon.
In such a highly polymorphic syngameon any form of sufficient
vitality, may simulate a primary species if isolated, and to some
exten t even if not isolated. In a case like that of Veronica longi{olia and V. spicaia (comp. above p. 377), where it is quite obvious that the two species have immigrated from different directions
(comp. STERNER 1922), the secondary nature even of a most vital
hybrid-population may be taken as proved. The same may be
said even of other highly polymorphic hybrid-populations occupying a comparatively narrow zone between two species widely
distributed in their pure state. But as soon as the highly
polymorphic hybrid-population is of a size equalling or surpassing that of the pure species-populations, suspicion may arise
that the hybrid is older than the species, and that the
latter are simply secondarily isolated segregates of the highly poly-

383
morphic hybrid. And this suspicion will be still stronger
when the species do not even form pure populations but only
appear as single individuals here and there in the highly
polymorphic hyhrid. Good examples of cases in which very
little or nothing talks in favour of the primary nature of the
supposed species> , are furnished by GUNNARSSON'S monograph of
the genus Betula in Scandinavia (1925,comp. LOTSY 1925 band
ARWIDSSON 1930) - an unfortunate attempt to apply the Floderian
Salix-method to a rather hopeless, highly polymorphic syngameon
in another genus.
Polymorphic syngameons more or less analogous to those found
in Salix and Betula are found also in several New Zealand genera.
The genus Alseuosmia (Caprifoliaceae, forest-shrubs of the northern
part of North Island), for instance, appears to consist of one vast
and immensely polymorphic syngameon (COCKAYNE 1923, 1925,
1928, 1929, ALLAN 1926 C, COCKAYNE and ALLAN 1927), only conventionally divided in a few species of which COCKAYNE writes as
late as 1929 (p. 21): Whether there are really the above species
of Alseuostnia, I have no idea. What I saw of it in the field
made me seriously doubt that it will ever be possible to get a
natural division of it into species and hybrids. - Senecio Lyallii
and S. scorzoneroides are two species of subalpine meadow-herbs
distributed throughout the mountains of South Island and Stewart
Island, mostly occurring together and then always connected by a
highly polymorphic hybrid (COCKAYNE 1923, 1925, 1928). If I may
rely upon my own observations, this hybrid quite often outnumbers the pure species. In this case it appears equally probable
that the species are segregates of the hybrid as that the hybrid
was formed by the meeting of two species originally isolated. If
we follow the modern New Zealand taxonomists in considering
the two species as valid, we probably here get a case similar to
those of northern .Salices, namely a hybrid playing a greater role
than the pure species. In the genus Celmisia several analogous
examples are found. While in some sections of that genus the
species are rather distinct, and the hybrids either sterile (comp. above
pp. 365-366) or fertile but not vital enough, or at least not abundant
enough to spoil the species-limits (Celmisia coriacea X Trauersii,
and C. spectabilis X Trauersii, cornp. COCKAYNE 1915, 1923,1925,1928 1) ,
1 A series of forms 01 these hybrids, collected by myself on the southeastern
spurs of Mt Captain (Northeastern District), showed no reduction of the pollenproduction.

384

there are other sections giving the impression of rather hopeless, extremely polymorphic syngameons analogous to the worst
northern Salix- and Betula-populations.
Thus the series of
forms treated in CHEESEMAN'S Manual as C. brevifolia and C. novaezelandiae gave me the impression of forming one very polymorphic
syngameon (splendidly developed on Mt Wakefield in the Mount
Cook region), probably only in some peripheral parts differentiated
into pure populations of short-leaved (C. brevifolia) or long-leaved
(C. novae-zelandiae) forms. Very nearly related and possibly intergrading into this syngameon is the very large and exceedingly
polymorphic syngameon probably formed by C. incana Hook. f.,
C. discolor Hook. f., C. intermedia Petrie, C. Bonplandii (Buch.)
Ckn, C. Du Rietzii Ckn and Allan, and perhaps some other species
too. It is distributed throughout the New Zealand mountains.
Round its margins smaller parts of it appear to have become isolated either in outlying mountains or on coastal cliffs [e. g. the
northernmost subspecies of C. incana in North Island (comp.
COCKAYNE and ALLAN 1927 p. 269), C. holosericea in the Fiord District,
C. Lindsayii on the south-eastern coast of South Island], some
of them differentiated as quite distinct species, but in its more
central parts the syngameon presents, at least to the botanist making a comparatively short visit like myself, the picture of a most
intricate network of rather different forms hopelessly mixed with
each other and apparently all crossing. If this extremely polymorphic syngameon has been formed by hybridization of a number
of more uniform species, the latter appear to have got lost at least
to the same degree as many of the arctic species of Salix - but
if my very incomplete impression of this vast syngameon may
be of any value, I would rather believe that the small more uniform
local populations occasionally found are simply secondary segregates of the highly polymorphic syngameon. How this syngameon
first was formed is another matter - of course it may have been
formed by the crossing of species differing from all those local
populations presently found. But about this matter we had better
confess our complete ignorance.
The practical taxonomical treatment of those large and highly
polymorphic hybrids-eyngameons in which the species have got
more or less lost - or never have been differentiated - is one of
the most difficult matters confronting the taxonomist of the present
day. The Floderian method of picking out some more or less

385
rare forms as species, and classing the bulk of the syngameon as
more or less complex hybrids between them,.is possibly the most
logical consequence 'Of the methods applied to less complex syngameons. But the danger of this method is also quite obvious.
Even if we admit that the frequency of vital extravagant forms
in wild hybrids has been exaggerated by HERIBERT-NILSSON and
other geneticists, attempting to prove the impossibility of determining wild hybrids without experiments, there is always a chance
that a certain form intermediate between two others may have
arisen from a cross of two other forms than those expected. And
certainly the reconstruction of all those triple- and quadruplehybrids in salicological literature is a very dangerous business.
With full certainty we shall probably never know whether those
highly polymorphic syngameons are not the result of early crosses
of quite other species than those now differentiated in the outskirts
of the syngameons.
The opposite extreme would be to treat the highly polymorphic
syngameons simply as what they are at present, without any
hypothetical considerations as to the species from which they are
supposed to have originated by crossing. This would imply the
naming of the highly polymorphic syngameons simply as species
and the dividing of them into subspecies, varieties and forms, just
as other compound and heterofacial species are divided. The
extremely polymorphic Salix-population of Greenland mentioned
above would thus be treated simply as 0 n e species and be divided into one' northern subspecies (= FLODERUS' S. arctica X glauca
+ the rare individuals of pure S. arctica and pure S, glauca found
in this region), one middle subspecies (= FLODERUS' S. arctica X
chloroclados X glauca + the rare individuals of these three Floderian species in this region), and one southern subspecies ( = FLODERUS'S. chloroclados X glauca + rare individuals of S. chloroclados and S. glauca). But I strongly fear that the consequent
application of this method would lead to the dividing of the boreal Salix-population into very few and extremely compound species, very difficult to handle and each consisting of so extremely
different forms that they would be of little value as units for
ecology and all the other branches of botany that taxonomy
has to serve. And I must confess that even my taxonomical mind
strongly revolts against the idea of lumping two so different populations as pure Salix arctica and pure S. glauca into one species.

386
Still I think this method is the only one applicable in the case
of Betula, the Scandinavian Betula-population probably containing
only two recognizable species, Betula nana and Betula alba (connected by a highly polymorphic hybrid of too moderate vitality
to spoil the species-limit). Also in the cases of Alseuostnia, Senecio
Lyallii-scorzoneroides, and the Celtnisia discolor - group it may be
sound to try this method parallelly with the others.
The third method possible is a compromise, used at present by
many Scandinavian taxonomists and ecologists, whose limited
knowledge in the mysteries of modern salicology has simply
forced them to adopt it in their field-work. According to this
compromise-method, the species are simply delimited in a less
rigorous sense than the Floderian one, and only the forms clearly
intermediate between the pure species are included in the hybrids.
The establishing of triple- and quadruple-hybrids is generally
avoided except in quite undisputable cases. This method implies
the admission that a part of a species-population may be changed
s I i g h t I Y by hybridogenous infection>, of other species and
still remain within the limits of the species concerned. Considering
the great polymorphy admitted in many species when due to
other causes, this appears to be a rather sound admission. Of
course it will then be a purely practical question where the limit
between the hybrid sensu stricto and this species sensu amplo
should be drawn. By this method most of the Salix nigricanspopulation of South and Middle Sweden is simply classed as S.
nigricans, in spite of its apparent hybridogenous infection of S.
cinerea etc., and most of the North-Swedish S. nigricans too, in
spite of the infection of S. phylici{olia and S. glauca. How the
Salix-population of Greenland mentioned above should be treated
according to this method I do not know, but possibly a good deal
of it would come in under the species S. arctica sens. ampl., S.
glauca sens. ampl., and S. chloroclados sens. ampl. [ = S. groenlandica
(Anderss.) Lundstr.]

Syngameon, commiscuumand convivium.

In the above discussion 1 have consequently used LOTSY'S term
syngarneon. in a rather wide sense, i. e. as a short and handy
term for any intercrossing population not divided by distinct
lines or zones of discontinuity. A syngameon in this sense may
be either a com m i s c u u m or a part of a commiscuum in the

387
sense of DANSER (1929 a, b). According to DANSER, a commiscuum
includes a II individuals connected by a genetical possibility of
Vermischung, i. e. exchange of genes. Populations prevented
from crossing only by geographical isolation according to DANSER'
still belong to the same commiscuum, but each of them forms
a separate con v i v i urn (DANSER 1929 b). According to the
definition and the examples given by DANSER, his convivia are
identical with the species> of HAGEDOORN 1921, i. e., any population
more or less differentiated from its relatives and tending to reduce
its potential polymorphy may be called a convivium. As the
convivia of DANSER must only be mehr oder weniger scharf
unterscheidbar, they may in some cases correspond to our species,
in other cases to our subspecies or even varieties (this is clearly
shown not only by DANSER'S definition but also by his examples).
A commiscuum, on the other hand, is often identical with a species,
but it may also include several species, if they are actually
connected by fertile hybrids, or if they are prevented from being
so only by geographical isolation.
An important fact pointed out by DANSER is the possibility of
Vermischung (exchange of genes) of two biotypes, even if they
nre not fertile with each other, but both form fertile crosses with
a third biotype. Thus it is not necessary that all biotypes in the
same commiscuum (or in the same syngarneon) should be capable
of d ire c t crossing. The occasional occurrence of intraspecific sterility pointed out by several geneticists (comp. for instance WINGE
1926 a, TURESSON 1929 a, MUNZING 1929, 1930) does not therefore
affect the validity of our species-definition.
Practical methods of species-taxonomy.
From what has been said above regarding the delimitation of
species it :will be clearly seen that the pure abstraction v-method
of species-making by sorting specimens into groups merely after
their morphological resemblance, as still practised by many taxonomists [and still supposed by LOTSY (1921, 1925 b) and many other
geneticists to be the method of a II taxonomy], must be firmly
rejected, and that the"morphologic-geographical method so
brilliantly advocated and demonstrated in the 'nineties by WETTSTEIN
and his followers, must conic to its right again not only in zoology,
where it has been practised with eminent results during the last
decades, but also in botany, where it unfortunately appears to

388
have fallen into some discredit or oblivion after the Wettsteinian
period. It must not be forgotten that taxonomy not only deals
with abstract concepts, but in the first line with living populations.
And certainly those populations must be studied in the field
much more than is mostly the case in present-day taxonomy.
Though I am inclined to think that COCKAYNE and ALLAN are
somewhat underrating the possibilities of studying a population
by means of herbarium-material (if this material is well preserved
and rich enough to give not only examples of the populations
concerned but also some idea of the regional and local variability
of its polymorphy), their strong emphasizing of the need of Fie I dT a x 0 nom y certainly cannot be supported enough (comp. also
FRIES 1852 pp. 194--210, SAMUELSSON 1922 p. 229, HALL and CLEMENTS
1923 p. 15, Du RIETZ 1930 p. 295). And certainly there is something
fundamentally wrong with taxonomy as long as many of those who are
put to carry out most of the monographical work, namely the officials.
of the great botanical museums, are forced by antiquated regulations to spend practically all their time in indoor museum-work
- while a modern botanical museum-official, of course, ought not
only' to be allowed but even forced to spend at least several
.months of the year in field-study of the populations he has to
deal with in his museum during the rest of the year.
The importance of genetical methods in species-taxonomy cannot,
of course, be overrated - as long as those methods are not
supposed to r e p I ace the study of wild populations in their
natural habitat, or the necessary comparative museum-work with
those populations. If we take home a single specimen (or a few
specimens) of each species, of each subspecies, or even of each
variety of a genus, plant them in our gardens, cross them with each
other and play with them in various ways, they will still tell us nothing about the variability of polymorphy in the wild populations
concerned, i. e. nothing of what is most essential for grasping
the species and their subordinate units. The field-study of the
variability of polymorphy in wild populations can be replaced as
little by the study of a few living individuals brought home 10
a garden as by a few dry individuals in a museum, and it is
often of. more use to get anumbel' of dry individuals sufficient
for getting at least some idea of the variability of polymorphy in
the wild population than it is to get an ins u f f i c i en t number
of living individuals to a garden. The claim of TURESSON (1926 c)

389

and other geneticists to have replaced the old deecrlptive taxonomy with a new experimental taxonomy must therefore be
firmly disputed, In this statement there lies no underrating of
the importance of garden-experiments a s a co m pie men t to
the comparative field-study combined with comparative museumwork. But even the elimination of the phenotypical variability
of a population can certainly be done to a much greater extent
by means of a comparative field-study, than is admitted by most
geneticists believing this to be possible only in their experimental gardens.
The practical methods of des c rib i n g s p e c i e s cannot be
discussed here in any detail. Still it must be mentioned that the
general recognition of species as populations, mostly consisting
of a great number of forms often grouped to distinct local and
even regional facies (varieties and subspecies), makes a revision
of the traditional method .of describing species rather necessary.
At present a diagnosis of a species may be founded upon one
individual or upon several individuals belonging to one or several
forms, varieties or even subspecies - and about this mostly nothing at all is told in the diagnosis. When the diagnosis tells us
that the leaves have a certain variability, nobody knows whether
this refers to one individual o.r to many individuals, nor what
the relation is of those many individuals to each other. The
recent proposition of GALLOE (1928) that a diagnosis of a species
should be founded upon one single individual, must certainly be
taken into serious consideration - but of course this description
of one single individual cannot be sufficient to characterize the
species concerned. The best method would probably be to give 1)
one diagnosis of the type-individual, 2) one description of the
polymorphy of the type-variety (i. e. the local population to which
the type-individual belongs), and 3) one description of the variability of polymorphy in the whole species. If distinct subspecies
are differentiated, also the variability ofpolymorphy in those
should, of course, be separately described. If many branches are
taken from a shrub or a tree and placed upon different herbarium-sheets, all of these must, of course, be considered parts of
the type (not co-types, as often practised at present). If, on the
other hand, several individuals of small plants are placed upon
one herbarium-sheet, only 0 n e of them may be designated as
the type (not the whole sheet, as is the present practice). The

390

term co-type ought then to be restricted to individuals taken

in the same locality as the type and supposed to belong to the
same biotype or being at least practically indistinguishable from
the type (very often the type forms a part of a clone or a more
or less 'autogamous and inbred population of practically undistinguishable individuals).
The practical application of this method I hope to be able to
demonstrate shortly in another paper.

11.

The hybrid.

A hybrid is a population intermediate between

two 0 r Seve r a I d iff ere n t s p e c i e s, sup p 0 sed t 0 b e
originally formed by the crossing of biotypes
bel 0 n gin g to d iff ere n t s p e c i e s.
This hybrid-concept probably coincides with what most biologists have called and still call a hybrid. As mentioned above
(p. 342), however, also another very different hybrid-concept has
been advanced during the last decades. According to LOTSY (1916
etc.) and RAUNKIAER (1918) every heterozygous biotype
should be called a hybrid. As the strict consequence of LOTSY'S
and RAUNKIAER'S hybrid-definition turns out to be that every sexual
species in our sense would consist exclusively of hybrids sensu
LOTSY (even according to LOTSY'S own opinion, cornp. above p.
341), the practical applicability of this hybrid-concept appears at
least rather dubious. Practically, even LOTSY himself does not
use it strictly in his own hybrid-studies, using the term hybrid
only for biotypes intermediate between two Linneons (= our
species, comp. for instance LOTSY and GODDIJN 1928). Still this
is something rather different from our hybrids. For highly polymorphic hybrids in our sense LOTSY (1. c.) uses the terms hybridogenous populations (s populus hybridogenus), hybridogenous
crowd, hybrid swarm etc. LOTSY'S hybrid-terminology is adopted
also by COCKAYNE and ALLAN (1923-1929), who use the term
hybrid> for any biotype intermediate either between two species
or between two iordanons (comp. above p. 350), and the term
hybrid swarm for any highly polymorphic hybrid in our sense.
The various types of hybrids in our sense have been dealt with
in detail above in connection with the species; so no further dis-

391
cussion of them is necessary. Like. a sexual species, a hybrid
may be more or less polymorphic, consisting of a very great number of biotypes. In fertile and highly polymorphic hybrids these
biotypes form a syngameon just like the biotypes in a sexual
species. Between a highly polymorphic hybrid and a highly polymorphic species there is no fundamental difference, and therefore it is
often very difficult to decide if a highly polymorphic population
should be classed as a species or as a hybrid. A highly polymorphic hybrid may develop to one or several species by more
or less effective isolation from the parent species. Still most
highly polymorphic hybrids are considerably more polymorphic
than most species, and the high potential polymorphy of the hybrid is then protected from being reduced by the constant crossing back with the two parent species, and the constant appearing
of new crosses between them. Using HAGEDOORN'S (1921) speciesdefinition we may say that a species tends to reduce its potential polymorphy, while a hybrid does not tend to reduce it if not isolated from the parent species, in which case it either
disappears or assumes the character of a new species.
The discoveries of cytological anomalies in many hybrids during
the last decades have made some botanists believe that cytology
is always able to tell whether a certain biotype belongs to a
species or to a hybrid. Unfortunately this is not the case. There
is no constant fundamental cytological difference between species
and hybrids. So ill e hybrids (between species with different chromosome-number) are easily recognized by cytological anomalies
(but then mostly also by reduced fertility or at least vitality), but
other hybrids (between species with the same chromosome-number)
are even cytologically indistinguishable from their parent species.
During the last years there has been much discussion regarding
the methods of naming hybrids. The traditional method of using
a formula consisting of the names of the parent species (e. g.
Veronica longifolia X spicaia, Dracophyllum [ilifolium X recuroumi
is clear but not very handy. Many botanists therefore give a
binary name also to hybrids (e. g., X Veronica hybrida, X Dracophyllum arcualum) which is used alternating with that somewhat
complicated hybrid-formula." This may be convenient when it is
nncertain whether the population concerned should be classed as
a species or as a hybrid, but in other cases it appears to imply
the bnrdening of nomenclature with a rather unnecessary name.

392

In the International Rules, the giving of binary names to hybrids

is recommended whenever it seems useful or necessary (Art. 31).
The method of naming hybrids by abbreviated formulae as proposed by COCKAYNE and ALLAN (1926 c, 1927 etc.), is. certainly
very handy in verbal discussions, notebooks etc., but for botanists
less familiar with the flora concerned it may often be hard to
understand that Nothojaqus cliffusca means Notho(agus cliffortioides
X fusca, Hebe ellipsala Hebe elliptica X salici(olia, Aristotelia [ruserrata Aristoielia (ruticosa X serrata, Ranunculus Buchlyallii Ranunculus Buchananii X Lyallii, etc. In scientific literature I therefore
prefer writing out the whole formula. COCKAYNE and ALLAN severely criticize the method of giving binary names of the ordinary
type to hybrids, pointing out that most of those binary names
in use were originally given only to 0 n e form or a small group
of forms within a highly polymorphic hybrid-population, and
therefore cannot be considered representative for the whole hybrid.
This is certainly quite right, but just the same may be said of
most species-names also, though the matter perhaps becomes more
disturbing in polymorphic hybrids because of the great differences
between the form originally described and many other forms of
the hybrid. To me it appears to be merely a matter of taste and
convenience, whether we use those old species-names or the abbreviated formulae of COCKAYNE and ALLAN as more handy substitutes for the complete hybrid-formulae.

12.

The units superior to the species.

The taxonomical units superior to the species can hardly be

called Iundamental s to the same degree as the species and its
subordinate units, since the delimitation of the higher units is to
much greater an extent a matter of taste and convenience. Every
taxonomist knows how hopelessly opinions differ regarding the
delimitation of genera and other units of higher rank. It is not
my purpose to discuss those higher units in any detail in this
paper. Only some of the general principles for the delimiting of
those units will be briefly' discussed here.
As was shown above' (pp, 352-353), it is quite easy to understand the differentiation of a species into regional and local facies in
the light of what I would propose to call HAGEDOORN'S I a w 0 f

393
differentiation by means of automatic reduction
of pot e n t i a I pol y mol' ph Y - even without the assumption
of any role played by induced or non-induced mutations, nor by
selection. In this connection it was also mentioned that [ust the
same process may often lead to the differentiation of a syngameon
into different species, if the syngameon is split up into parts
more effectively isolated from each other. This effect of isolation
was clearly recognized already by LEOPOLD VON Bucn (1825, compo
WAGNER 1889 p. 344) and MORIZ WAGNER (1868-1889), and this
.idea was built out to an elaborate theory by the latter author
(the migratlon-theory, later changed to the separation theory).
But the mechanism of differentiation in the absence of any
mutation could, of course, not be recognized by those authors
(though the importance of inbreeding was clearly pointed out
by WAGNER, comp. for instance 1889 p. 401), as this recognition
depends upon genetical experiences not available at that period
(cornp. HAGEDOORN 1921 p. 120). Of course nobody knows what
would have happened, if MENDEL'S work (1865) had been known
to WAGNER, and HAGEDOORN (1921 p. 120) may be right in assuming
that WAGNER would have greatly appreciated it, and could have
been trusted to incorporate it into a really logical evolution-theory.
As it was now, however, WAGNER had to look for another explanation to the effect of isolation, and he believed to find that explanation in the production of new forms (mutations), to a great
extent directly induced by the environment (as assumed also by
VON Been) and saved by isolation from being swamped in the
original main population [the clearest summary of his views in
this respect is probably found in pp. 472-473 in his Gesammelte
Aufsatze (1889)]. He strongly opposed DARWIN'S selection-theory
(1859), considering the struggle for life to be nur eine unwesentliche begleitende Erscheinung del' Differenzierung, d. h. ein nul'
zeitweilig mitwirkender und nebensiichlicher Faktor dieses Prozesses (WAGNER 1889 p. 462'). Unfortunately WAGNER'S ideas found
1 Many years after WAGNER'S death, his opposition in this point was characterized by HAGEDOORN (1921 p. 120) in the following words:
-DARWIN lacked the necessary key at the time when he needed it most, and
when he came into touch with WAGNER'S work, it could not shake his faith in
selection as the cause of stability of species. All the recent work in Genetics,
MENDEL'S law, the things we have since learned about the nature of the genes,
the selection experiments with the most diverse material, have shown us that
WAGNER in opposing DARWIN in this fundamental point had the right wholly
on his slde.

394

little recognition both in the dominating Darwinian school of his

time (comp. for instance DARWIN 1869 pp. 119-122, HAECKEL 1868,
WEISMANN 1868, 1872, SEIDLITZ 1871, and WALLACE 1889) and among
anti-selectionists like NAGELI (1872, 1884). But the idea ofisolation
as the dominating factor in the formation of new species got new
strength by the fundamental works of GULICK (1872-1905) and
ROMANES (1886, 1897). While WAGNER had recognized the importance of spa t i a I isolation only, GULIcK-and ROMANES - independently of each other - were the first to realize the full importance of ph Ys i 0 log I c a I isolation, and to work out the general.
theory of isolation summarized by ROMANES (1897 p. 145) in the
following words: )) Without isolation or the prevention of free
intercrossing, organic evolution is in no case possible. In the
first years of the new century, similar lines of thought were
followed by MAYER (1902), D. S. JORDAN (1905), K. JORDAN (1905)
and several other authors, and in the last decades the idea of
isolation as the main factor in species-formation has steadily won
new ground among biologists '.
Parallelly with this development the traditional theory of a
divergent evolution has gradually lost ground, not only among
geneticists but also (earlier than among geneticists and certainly
unnoticed by most of them) among those ecologists carrying out
field-studies of plant and animal distribution. In 1906, Guppy
advanced his d iff ere n t i at 0 nth eo r y, founded upon extensive
studies of plant-distribution and worked out more elaborately in
a series of subsequent publications (comp. for instance Guppy 1907
and 1917), and in 1907 very similar views were expressed by
WILLIS (comp. also WILLIS 1922, 1923), independently of Guppy but
founded upon analogous researches in other parts of the world.
According to Guppy'S differentiation theory, the role of the
polymorphous species belongs alike to the plant and to the bird.
A species that covers the range of a genus varies at first in every
region and nltimately gives birth to a new species in some parts
of its range. Then the wide-ranging species disappears and the
original area is divided up into a number of smaller areas each
with its own group of species. Each smaller area breaks up again,
and forms, yet more specialized, are produced; and thus the
process of subdivision of range and of differentiation of form goes
1 Of more recent authors clearly opposing this idea, H.
must be specially mentioned.

DE VRIES

(1903 p. 507)

395
along until each island in an archipelago OWns its bird and each
hill and valley has its separate plants. (GuPPY 1906 p. 522.)
The contrast between this theory and the traditional theory of
divergent evolution is summarized by WILLIS (1922 p. 221) in the
following very suggestive way: Chiefly important among these
is the new view of evolution, first proposed by Guppy in 1906,
and by the writer in the following year, that evolution did not
proceed from individual to variety, from variety to species, from
species to genus, and from genus to family, but inversely, the
great families and genera appearing at a very early period, and
subsequently breaking up into other genera and species.
As to the mechanism of this differentiation-process, Guppy and
WILLIS accepted mutations arising independently of the environment
as the evident cause, and rejected both selection and direct
adaptation as factors of any importance. Similar views are found
in the works of CRAMPTON (1917, 1925 a, b). As to isolation,
none of these authors appears to have fully recognized its fundamental role in the differentiation-process. And probably there
is not much need of isolation for the explaining of this differentiation-process, if we count with .mutation to such an extent as
these authors did - especially if with I a r g e mutations like those
assumed by WILLIS - just as DARWIN'S strong belief in the great
power of selection made him regard isolation as a factor of minor
importance.'
1 As has
recently been pointed out by EKMAN (1928, 1929), many recent
geneticists have made a great mistake in accusing DARWIN of having believed
selection to be able to pro d u c e new biotypes. The production of new biotypes forming the material for selection DARWIN explained by non-induced mutation (vvarlatlon), and to some extent (especially in his later works) by mutation induced by the environment. But certainly he believed in the power of
selection to change a population by favouring even mutations differing in the
most minute and insignificant characters from the main stock of the population
- and in this belief he was followed by WALLACE, WEISMANN and other selectionists of the Darwinian period. Therefore he thought himself able to explain the
remarkable differentiation of species in the Galapagos Islands by the minute
differences in the com pet i t ion fa c tor s of the different islands, which
he thought sufficient to lead to the favouring of different mutations in the
different islands (even if the same mutations were produced in them all). However, he confessed that the lack of differences in the h a bit a t of the different
islands for a long time appeared to him as a great difficulty (DARWIN 1859 p.
400). What a small role he alotted to the pure isolation-factor is clearly seen
both in his publications (comp, for instance 1859 p. 402, 1869 pp. 119-122) and

396

Quite independently of Guppy and WILLIS, the dilferentiationtheory was advanced again in 1916 by CAJANDER (1916 pp. 561562, comp. also 1921) and LOTSY (1916, cornp, also 1912-1914).
CAJANDER explained the differentiation-process partly (and apparently
in first line) by selection, partly by mutation; he also pointed
out that it resulted not only in geographical races and geographically vicarious species, but also in e colo g i c a I races and ecologically vicarious species (s Standortsrassen> and Standortsarten).
LOTSY, on the other hand, did not accept any mutation-theory,
but explained the differentiation-process simply by isolation
(spatial or physiological) in connection with selection. Similar
views were expressed by HERIBERT-NILSSON (1918). Already in
H117 J. SCHMIDT had pointed out that besides selection pure
accidents must be an important factor in the differentiation of local
races in Zoarces viviparus (comp. above p. 349) But it was first
HAGEDOORN (1921) who showed the way to a real understanding
of the differentiation of a syngameon by isolation even in the absence of any selection (and of any mutation), t h u sma kin g
the differentiation-theory quite independent of
all the chan.ging theories regarding the existence
of effective selection as well as of induced or
n 0 n- i n d u c e d m uta t ion.
During the last decade, the differentiation-theory has been accepted by several authors. TUREssoN (1922-1929) followed the
same main channels of thought as LOTSY, J. SCHMIDT and CAJANDER,
studying the differentiation of species into ecotypes (partly
corresponding to CAJANDER'S Standortsrassen, partly to his geographical races, comp, also above p. 355) and explaining this differentiation mainly by selection. The present author (Du RIETZ 1923 a)
applied the same point of view upon the differentiation of species,
pointed out the insufficiency of selection for its explanation, and,
rejecting mutation as a probable factor of importance, emphazised
in his letters to WAGNER (comp. WAGNER 1889 p. 350, ROMANES 1897 pp. 103106). Under those circumstances it is important to note that in earlier years
DARWIN had expressed a much higher appreciation of the role of isolation; in
1844 he declared in a letter to HOOKER, that with respect to original creation
or production of new forms... isolation appears the chief element (ROMANES
1897 pp. 101-102). Evidently this was the primary impression left by the
field-studies of his youth, but under the influeuce of his later speculations and
work with domestic animals and plants it gradually faded away and was replaced by the theory of natural selection.

397
the importance of purely accidental differences in the original
composition of isolated parts of a syngameon. Among other recent contributors to the differentiation-theory, BERG (1926), DANSER
(1929 a, b) and BAUR (1930 pp. 399-400) must be specially mentioned.
If we take the existence of this process of differentiation of a
highly polymorphic syngameon into species as a fact - andT
really think we have the right to do this now - we may leave
the question of how those syngameons were origirially formed
provisionally open, and try to look into the possible bearing of
this differentiation-process upon the practical delimitation of the
taxonomical units superior to the species. In those cases where
it is obvious and undisputable that a number of species have
been differentiated out of one original syngameon, the desirability
of uniting all those species to a unit. of higher rank has been
felt by several authors, If the differentiation is due only to
geographical isolation, i. e. if a series of vicarious species (echte
"Vikaristen> sensu VIERHAPPER 1919) with only 0 n e species in
each of the isolated districts (islands, mountains etc.) is formed,
those species form an Artenkreis in the sense of RENSCH (1929)
and a Formenkreis in the sense of KLEINSCHMIDT (1901-1926),
who does not recognize the difference between, the two types of
units separated by RENSCH as Artenkreise and Rassenkreise>.
A good example of a natural group of species of this type is formed by the three vicarious species of Geniiana in the subantarctic
islands of New Zealand mentioned above (p. 364).
If the differentiation is due not only to geographical but also
to ecological or physiological isolation, there is often not 0 n e
species 'in each district corresponding to one species in each of
the other districts, but a group of several nearly related species
in each district corresponding to an analogous group in each of
the other districts. In such a case I have mostly found RENSCH'S
Artenkreis--concept unapplicable, as it is often impossible to say
that one of those nearly related species in a certain district is
more corresponding to a certain species in another district than
the .others. In the genus Dracophyllum in New Zealand, for instance (comp, OLIVER 1928), there are only a few examples of
clearly vicarious species forming distinct Artenkreise sensu RENSCH
(one of the best is that of the two tuft-tree species D. recurvatum
in North Island and D. Traoersii in South Island). But most of
27 - 30531. SvenBk BotaniBk Tid'krilt. 1930.

398
the New Zealand species of Dracophyllum form a network of species
combining certain characters in the most various ways, and
occurring together in more or less overlapping distribution-areas
in such a way that it is not possible to group them into distinct
Artenkreise, but only into dubious Artenkreise ins eve I' a I
a I tel' nat i v e way s. If, for instance, life-form and habitat are
taken as base for the grouping, the subalpine dwarf-shrub heath
species D. recuruum (in the central part of North Island, cornp.
above p. 373) and D. pronum (South Island) may be grouped to
one Artenkreis, while D. filifolium (one of the dominants of the
North Island subalpine scrub, comp, above p. 373) forms another
Artenkreis together with D. rosmarinifolium (= D. uniflorum of
the older literature, one of the dominants of the South Island
subalpine scrub). But if we take the leaves or the flowers as
base, D. recuruum and D. pronum must be placed in quite different
Artenkreise>, and the same will be the case with D. filifolium
and D. rosmarinifolium. According to the position of the flowers
(a character considered to be of great taxonomical importance in
Dracophyllum), D. rosmarinifolium could also be united with D.
subulatum (on the Volcanic Plateau of North Island) and some
other species into an Artenkreis not less natural than the one
mentioned above.
The method of grouping geographically vicarious species into Artenkreise appears thus to be applicable only in special cases.
Another method is the grouping of the species forming hybrids
fertile enough to allow Vermischung> (exchange of genes) in
nature or in cultivation, into one com m is c u u m in the sense of
DANSER (comp. above p. 387), and of the commiscua forming more
or less sterile hybrids into one com p a I' i u m or Bastardierungsgenossenschaft (DANSER 1929, a, b). The whole section Lapathum
of the genus Rumex, for instance, according to DANSER forms one
comparium. And of the New Zealand species of Nothofagus four
(N. fusca, N. cliffortioides, N. Solandri and N. truncata) probably
form one commiscuum, while the fifth species (N. Menziesii) forms
not only its own commiscuum but probably even its own comparium (comp. COCKAYNE 1926, COCKAYNE and ATKINSON 1926).
Probably it will be found possible to work out a synthesis of
. the method last mentioned with the geographical method mentioned
above, enabling us to reconstruct many of the old syngameons
from which the present species have been differentiated. As a

399

matter of fact, many of the sections, subgenera and genera

distinguished in present-day taxonomy, especially those with a rather
restricted distribution, give the impression of being old welldefined syngameons split up into a number of species in the
way described above.
It cannot be discussed here in any detail what role selection
and induced or non-induced mutation 1 may have played during
this differentiation-process. In the discussion above, I have only
talked of the process of differentiation by means of automatic
reduction ofpolymorphy, bee a use no 0 the r pro cess appears
to ben e e d edt 0 ex p 1a i n w hat we see inn at u r e, and
because this is the only process about which we know anything
with full certainty. But, as mentioned above (p, 353), this differentiation process may be supported by selection (or rather by heterogenous
elimination), if the isolation takes place in obviously differing
habitats. And it would be unscientific to deny the po s sib i lit Y
that even induced or non-induced mutation may play a role in
that process, though I must admit that I (like LOTSY, HERIBERTNILSSON, HAGEDOORN, DANSER and other recent authors) am unable
to find much talking in favour of this assumption. As to the
role of selection, my field-experience in various parts of the world
and in various groups of the vegetable kingdom has brought me to
exactly the same conviction as that of WAGNER (1868-1889), Guppy
(1906-1917), WILLIS (1907-1923), HAGEDOORN (1921), HAYATA (1921
-1928), TOWER (1922), LOTSY (1925 a), PARR (1926), and many
other authors, namely tbat the role of selection has been much
overrated,' and that it mainly consists of the elimination of biotypes
absolutely unviable in a certain habitat (Du RIETZ 1929 p. 422).
Nobody has expressed this better than LOTSY (1925 a p. 59): We
. have . no evidence that it is the fittest, the best adapted, which
1 The term
mutatton is used here to designate 1) any genotypical change,
small or large, continuous or discontinuous, that is not due merely to recombination of genes, and 2) any new biotype produced by such a change. Especially in America that term is still sometimes used also in the sense of BURBANKS,
i. e. as a designation for biotypes originating from a cross and exhibiting characters not present in any of the parents (comp. HAGEDOORN 1921 pp. 84-85,
154-155 etc.), or 'extravagant biotypes in the sense of HERlBERT-NILSSON (1918).
Such appearing of new characters simply as the result of a recombination of
genes (Amphimutation> in the sense of PLATE 1913 p. 497, 1914 p. 138) is not
included in the mutation-concept used here (which corresponds only to PLATE'S
Neo- oder Idiomutation),

400
survives in the struggle for life; all we can say is that it is the survivals which survive, and this is probably, just as in a railway accident, more determined by a happy position than by any particularity
of structure. A critical study of the correlation between life-form
and habitat in various parts of the world has made me believe
that this correlation is much overrated in botanical literature, and
that only life-forms absolutely unviable in a certain habitat are
eliminated in this habitat, while chance. and historical factors,
but not superior or inferior adaptation, mostly determine the main
composition of the' vegetation (Du RIETZ 1929 p. 422) -- just
as is the case in the human population! This certainly 'speaks
neither in favour of the theory of positive selection nor of that
of mutations partly induced by the habitat.. But to all those
problems I hope to come back in another publication.
As to the role of hybridization in the differentiation of syngameons
into separate species, we unfortunately -know very little. Evidently
it must often happen, that the differentiation-process is counteracted by the occasional' meeting and crossing of two species
spatially - geographically' or ecologically ...,- isolated, but still belonging to the same commiscuum (e. g. Veronica longi{olia and V.'
spicata, comp. above p. 377) - or Abies alba and A. cephalonica,
which according toMATTFELD (1930) were geographicallydifferentiated from a common stock during the Tertiary period but
brought together again by the Glacial period, producing the highly
polymorphic hybrid A. Borisii regis now connecting the A. cephalonica-population of Greece with the A. alba-population of the
northern Balkan Peninsula). Equally clear is that this must often
give rise to extravagant> forms, and even "that some of them
- as well as some of the more intermediate forms - may happen
to become isolated and then be vital enough to develope into new'
species. 1 It is quite clear, that the act u a I pol y m 0 r ph Y
1 The theory of species-production by means of hybridization can be traced
back as .Iar as to LINN.lEUS, who already in 1744 (Peloria) expressed the belief
that new species, and even genera, could be formed by hybridization. In 1760
(De. sexu plantarum) he advanced the theory that there had originally been only
one species in each genus, all the other species being formed by the crossing of
that species with species belonging to other genera. (A good summary of the
genetical views of UNN.lEUS, which were fundamentally different from those
ascribed to .him in most of the evolutionistic and genetical literature, is given
by ALMQUIST 1917). In 1820 W. HERBERT, on the base of extensive hybridization-experiments, expressed the view that the present species of .each main

401
of a corrimiscuum may be heightened in this way, but not its
potential polymorphy (comp. Danser 1929 b), and all the
new forms produced in this way may very well have been reagroup of plants had been produced by variation of soil, temperature, or humidity", and that such diversity might have been further multiplied by hybrid
intermixture, as the species were brought together by the natural progress of
their diffusion (HERBERT 1820 p. 17). Later on, the hybridization-theory has
been repeatedly advanced by many authors. Already in 1863 CH. NAUDlN, who
had 'expressed similar views himself as early as 1852 (comp, FOCKE 1881 pp. 505
and 509), wrote that plusteurs botanistes d'une grande autortte croient que
certains hybrides fertiles, sinons tous, peuvent se fixer et passer a I'etat de'
varietes constantes, c 'est-a-dire de veritables especes (NAUDIN 1863 p. 196);
among those botanists E. REGEL was specially mentioned. In 1871, A. KERNER
published his first essay on the formation of species by hybridization, and in
1881, he .and L. REICHENBACH were mentioned as the chief advocates of this
theory by FOCKE (1881 p. 506), who also believed, that ein grosser Theil del'
neuen Arten zwar nicht plotzllch, abel' doch mittelbar aus Racenkreuzuugen
hervorgeht (p. 509). In 1891 (and 1898), KERNER gave a more complete account
of his Vermischungstheorle stating that alle in del' Nachkommenschaft slch
erhaltenden Veranderungen del' Gestalt durch Kreuzung, beziehentlich durch
Vermischung zweier ihrer Konstitution nach verschiedener Protoplasten zu
stande kommen (KERNER 1898 p. 536). - In recent literature WEISMANN is often
mentioned as one of the originators of this theory - WIESNER (1902 p. 259) even
talks of dle WElsMANN-I{ERNER'sche Vermlschungstheorle. This appears to be
in bad accordance with the following passage in WEIsMANr;'s last work (1913 II
p. 304): Einzelne Botaniker, wie I{ERNER v. MARILAIJN, sehen in del' Vermischung
del' Arten ein Mittel zur Bildung neuer Arten mit besseren Anpassungen; sie
lassen frochtbare' Bastarde unter Umstanden die Stammarten verdrangen und
selbst zur Art werden. Man wird zugeben konnen, dass solche Falle vorkommen
. .. abel' man wird dennoch behaupten mussen, dass die Artbildung im grossen
und ganzen auf anderen Wegen erfolgt als durch Bastardierung. Solche Falle
sind wohl ohne Zweifel nichts anderes als seltene Ausnahmen. In some of his
earlier works, however, WEISMANN had attempted to make amphimixis (or Vermischung- ) responsible for all the material of individual differences through
which selection produces new species (WEISMANN 1886 p. 29, comp. also 1891 p.
135), a position also accepted by WALLACE (1889 p. 439). But in. this connection
hybridization between different species was never mentioned, only Vermlschungv
between slightly differing individuals of the same species. And with his logical
mind \VEISMANN realized later on, that by such Verrnischung alone it would
hardly be possible to explain the evolution in which he - contrary to KERNER
believed so strongly. On p. 175 in his last work (1913, part II) he clearly
declared, that he did not regard amphimixis as die eigentliche Wurzel del'
Variationen selhst, but considered evolution possible only by Umblldung,
Verkiimmerung. und Neublldung von Determlnanten, i. e. by m uta t ion. During the last two decades the hybridization-theory has got a very able and
energetic advocate in LOTSY (1912~1929), who has gone further than any of his
predecessors, attempting to explain e v 0 I uti 0 n exclusively by means of

402

lized in the original commiscuum before the differentiation of the

species concerned (comp. HERIBERT-NILSSON 1918 p. 141). - It is
also known that new commiscua may be accidentally formed by
the crossing of biotypes belonging to separate commiscua within
the same comparium (KARPETSCHENKO 1928, DANSER 1929 b pp.
424-425). In the light of recent geneticists' successful production
of polyploid biotypes by crossing (comp. for instance WINGE 1917,
1926 b, 1928, ROSENBERG 1928, RENNER 1929) it appears rather
probable that the polyploid species found in many genera may
have been produced by crossing. of different commiscua d uri n g
the .process of differentiation of the syngameon originally formed
by the genus concerned (though it also looks quite possible that
they have been present in the syngameon from the beginning of
its existence). In this way the potential polymorphy of certain
commiscua may be accidentally heightened during the differentiation-process. But as crossing between different com par i a is
absolutely excluded, the potential polymorphy of a com par i u m
can never be heightened, only reduced (DANSER 1929 b pp. 427428). As a comparium only rarely includes more than one goo d
genus, this implies that most genera (or at least genera not connected by hybrids with any other genus) can never heighten their
potential variability, but only reduce it.
. This view appears to be supported by the fossil evidence available in several plant-genera, e. g. in Fagus and Quercus. Only
the example offered by Fagus will be discussed here. The Faguspopulation of the Tertiary period apparently formed one widely
distributed and very polymorphic syngameon (Fagus Feroniae Ung.,
comp. ETTINGSHAUSEN 1894, STOJANOFF and STEFAN OFF 1929, etc.),
containing all the forms of the present Fagus-species but dominated by other forms only appearing accidentally in the present Fagus-populations (vatavistic forms in the sense of ETTINGSHAUSEN
and KRASAN 1888-1889, comp, also KRASAN 1888). By isolation and
reduction of polymorphy this highly polymorphic syngameon has
become differentiated into the less polymorphic present species
hybridization. This was never attempted by KERNER, who did not believe In
evolution, but only in einer Umgestaltung, einer Umpragung des Vorhandenen
(1898 p. 5~6, comp, below p. 407). To this view also HERIBERT-NILSSON (1918 p.
143) returns, accepting the rest of LOTSY'S theory, but declining the theory of
evolution as impossible to reconcile with the 'results of modern genetics. Compo
also the theory of HAYATA reviewed below (pp. 405-410).

403

of Europe, Western and Eastern Asia, and North America, but

still the characters of the dominating Tertiary forms as well as
those of the other present species may crop up accidentally in
some of those species. Nothing appears to be talking in favour
of the assumption of a Post-Tertiary evolution> in this genus,
neither by crossing nor by mutation. If all the Fagus-populations
now isolated in Europa, Asia, and America were brought together
again into one large syngameon, it appears rather probable that
we would get back much of the act u a I polymorphy of the Tertiary
Fagus-population - but the pot e n t i a I polymorphy of the genus
would remain just the same, and would probably prove much
lower than that of the Tertiary population.
The difficulties in reconstructing the old syngameons, from
which the present species may have been differentiated and into
which they may be naturally grouped, are rapidly increasing as
we get along to genera or units of still higher rank. According
to the very suggestive theory of LOTSY (1916, comp. also Du RIETZ
1921 p. 90), each main class of animals and plants originally formed
one vast syngameon (originating from an accidental cross of widely
differing biotypes of an older syngameon), but even if this be true,
the reconstruction of tbe whole process of differentiation of those
immense syngameons of decreasing size appears anyhow to be a
rather hopeless task. Our division of the main classes of animals
and plants into orders, families, genera etc. must therefore to a
large extent be carried ont in the rather artificial way of grouping
the species, genera etc. simply after their morphological resemblance.
The same method must often be used even for the grouping of the
species within a genus, at least until the commiscua and comparia
within that genus have become sufficiently known. The more this
purely morphological method can ~e combined with geographical,
paleontological, genetical and cytological methods, the greater is the
chance that it will ultimately be possible to reconstruct the old
syngameons within the group concerned.
According to current opinion among taxonomists and morphologists of the present day, this morphological method of delimiting taxonomical units of higher rank should be founded upon
the theory of d i vel' g e n t e v 0 I uti 0 n, i. e. the taxonomical
units should be delimited and arranged according to their supposed ph Ylog e n y. In this ph Ylog e net i c met hod 0 f t a x onom y a central role is played by the distinction between Organisationsmerkmale> and Anpassungsmerkmale (NAGELI 1884). To

404
the first group (also called indifferent> or constltutional> characters) are referred characters supposed to have arisen independently of the environment [by some sort of Vervollkornnungsprinzip (NAGELI 1884) or Orthogenesis. (HAACKE 1893, EIMER 1897)J,
to the latter group (= the epharmonic or adaptive characters of
many authors 1) characters supposed to have arisen as adaptations>
to a certain environment. Taxonomical units of higher rank are
allowed to be founded only upon the former group of characters.
If two species are found to have a character of this type in common
(e. g. a floral or especially an embryological. character) this is
taken as a proof of their phylogenetical relationship, and they are
placed in the same genus, family etc. But if two species are found
to have a character of the second type in common (e. g. life-form,
bud-type, or leaf-type), this is said to be a case of epharmonic
convergence and is considered to be of no taxonomical importance
whatever.
This distinction - and with it the whole natural plant-system
of the present day - of course stands or faUs with the theory
of an evolution by means of mutations partly induced, and partly
not induced, by the environment. To this main theory belong
all the various evolution-theories of DARWIN, NAGELI, and modern
neolamarckists of selectionistic or anti-selectionistic type. Even
in the theory of evolution only by means of selected non-induced
mutations (i. e. the pure selection-theory of WEISMANN, cornp. above
p. 401), it is hard to conceive how that distinction could be upheld
strictly. And in the theory of evolution by means of hybridization there is. certainly no room for it at all; according to LOTSY
(comp. also HERIBERT-NILSSON 1918 p. 135, Du RIETZ 1921 p. 90)
phylogeny is a mere illusion, no science but a product of phantastic
speculation (LOTSY 1916 p. 140) - and certainly nobody could be
more competent a [udge in this question than LOTSY, who hasdevoted
a considerable part of his life to the reconstruction of phylogeny
after orthodox lines. It is [ust the same in the anti-evolutionistic
succession-theory of HAYATA and his predecessors (KERNER etc.,
cornp, below p. 407). But as practically all the leading botanical
taxonomists of the last half century have strongly believed in the
theory first mentioned, all that has been of no moment whatever.
1 Both those terms are sometimes used also for ph en 0 t Y pic a I characters
(comp, DIELS 1921 p. 137, COCKAYNE and ALLAN 1927 pp. 263-267). These, of
course, do not concern us here.

405
Considering the attitude of intense distrust taken by most
. geneticists of the present day towards the theory of mutations
induced by the environment, the method of making this theory
a fundament of higher botanical taxonomy may he called at least
a rather dangerous one. And this will still more be the case,
considering that the doubts of the geneticists arc supported by
the results of ecological field-studies. As was mentioned<above
(p. 400), my own comparative field-studies of life-forms, budtypes, leaf-types etc. in widely differing parts of the world have
brought me to the conviction that such characters - belonging
to the typical epharmonic characters (or Anpassungsmerkrnale)
of orthodox taxonomy - are mostly far less correlated .with the
environment than generally believed. Certainly this does not speak
in favour of their formation by means of mutations induced hy
the environment. The material upon which those statements are
based cannot be presented here, but I hope soon to have it ready
for publication.
A most remarkable revolt against the traditional phylogenetic
method of taxonomy was started about a decade ago by B. HAYATA
(Professor at Tokyo Imperial University and Director of the University
Botanic Gardens), whose imposing taxonomical knowledge and
wide field-experience of tropical, subtropical, and temperate vegetation certainly ought to make European botanists read his contributions (HAYATA 1921 a, c, 1928 a, b, c) with greater attention than
appears to have heen the case until now. But certainly they are
revolutionary enough to make an orthodox phylogenetical taxonomist shudder. Some extracts of HAYATA'S works will probably be
the best way to demonstrate the main views of this remarkable
reformator.
As I stated in the preceding paper, I have been reflecting on
a system of natural classification to which my attention was
drawn during the score of years that I devoted to the study of
the flora of Formosa. Current opinion demands that such a
natural system be a static one like BENTHAM-HoOKER'S or ENGLER'S
and that there is possible only. one true ideal system, to which,
however, we are as yet far from attaining, as but one phylogenetic
tree is possible. Much against my will, I have come to entertain
strong doubts as to the effectiveness of the modern systematizers'
effort to attain the ideal system; and my twenty year's experience
in systematic botany has steadily led me into quite a different

406
channel of thought. This I now venture to make public, though
I am aware that it will meet with a great deal of opposition.
All systematizers regard the natural system as a static one
with a definite form and believe that all species, genera or
families have their fixed natural positions, so as to be arranged
between this and that, according to their natural relations. My
idea is quite different from this current opinion. I regard the
natural system as a dynamic one, changing with the view of the
systematizer and subject to alteration, according to the way in
which it is considered, and I believe that none of the species,
genera or families has a fixed natural position, but has changeable
positions, subject to alteration according to the criterion for
comparison. It is neither natural nor necessary that a species
should in all cases be arranged between this limit and that; but
should be placed between this and that according to one view, or
between another this and another that according to another view.
In the present paper, it is my desire to explain what I propose
to call the dynamic system showing the natural relations of plants>
(HAYATA 1921 b pp. 97-99.)
HAYATA strongly objects against the orthodox view that only two
main theories about the origin of species are possible, namely
the theory that all the species are created by God separately and
independently and the theory that all the species are formed
by evolution from a single origin, i. e. the evolution theory.
He points out that there is one more way possible for the formation of species, namely that an innumerable number of species
of organic beings have existed from the eternal past and will
exist to the eternal future; they unite with or separate Irom one
another, and produce many different organisms by different combinations of the genes; or they change by themselves, as the genes
change (HAYATA 1921 a p. 94). To this theory HAYATA gives the
name of the par tic i pat ion the 0 ry ', and explains the theory
more funy in the following way:
1 To the first part of this theory HAYATA later on (1928 a) gave a separate
name, namely the succession theory. While the participation-theory sensu
stricto appears to be possible to reconcile with the evolution-theory in the form
given to it by LOTSY, the succession-theory stands in clear
opposition to any form of the evolution-theory. According
to the evolution-theory, the organic population of the present day has been
produced by d i v erg e n t e v 0 I uti 0 n, starting with very primitive organisms
and successively proceeding to higher ones. According to the s u c c e s s i 0 rr-

407

)The theory is in fact but one theory, yet for convenience' sake
.I shall treat it as two, namely: - The theory of the mutual participation of the gene, and the theory of the mutual sharing of the
gene. Literally speaking, the word 'participation' seems to express
a united action of genes to produce a certain result. Different
the 0 r y, on the other hand, a divergent evolution from lower to higher orga-,
nisms has not been proved, but only a succession without progression, and
there is no reason to believe that the original organic population of the earth
was less manifolded and of a "lower type than the present one. - Like the
hybridization-theory (comp, above p. 400), this theory can be traced back as
far as to LINNiEUS, who believed in the transformation of species, and to some
extent even of genera, by means of hybridization (comp. above p. 400), but
certainly not in any evolution. Similar views were widely spr.ead among the
botanists of the first half of the last century - in spite of the traditional tale
of evolutionistic literature, that botanists before DARWIN believed that all the
present species were unchangeable, simultaneously created by God independently and separately. In 1820, W. HERBERT (comp. above p. 400) formulated
the theory fhat in the early periods of the world, there existed only the
distinct genera of plants, or heads of families, not, however, exactly according
to the present divisions of Botanists; who indeed, are perpetually at variance
with each other; as to the features which are sufficient to constitute a variety,
a species, or a distinct genus. The lapse of centuries and diversity of soil and
climate have probably wrought the most wide and permanent distinction
between vegetables, that have originated from a common stock, possibly even
between the arborescent Ferns of the Andes, and the herbaceous inhabitants of
our Iorests. (HERBERT 1820 p. 16; this theory is mentioned by DARWIN in the
historical introduction to the later editions of his "Origin of speciesv.) In 1852
ELIAS FRIES one of the leading botanical taxonomists of his time - wrote
that "we now know with full certainty, that many quite different vegetationperiods have succeeded each other on earth, not a single one of .our present
species thus being original, but all of them being, as was observed already by
the prophetic eye of LINNiEUS, the dauqhiers of the time ('filia temporls'j> (E.
FRIES 1852 pp. 189-190. But like LINNiEUS, FRIES was inclined to accept the
g e n era as original (pp. 191-193), and certainly never believed in any progressive evolution. Of Post-Darwinian authors clearly opposing the evolutiontheory, A. KERNER takes the foremost place. His" Vermlschungstheorie (comp.
above p. 401), first published in 1891, certainly comes nearer the views of HAYATA
than any other previous theory: Sie setzt voraus, dass von [eher zahlreiche
verschiedene Pflanzenformen nebeneinander bestanden haben, was durch die
fossilen Reste auch tatsachllch bestatigt wird. Was sich von Pflanzen aus friiheren Perioden erhalten hat, weist durchgehends darauf hin, dass zu allen
Zeiten eine grosse M.annigfaltigkeit von Pflanzenformen die Erde bevolkerte,
Es bedurfte daher keiner Entwickelung, sondern nur einer Umgestaltung, einer
Umpragung des Vorhandenen. (KERNER 1898 p. 536.) However, KERNER never
drew the taxonomical consequences of those views drawn by HAYATA, nor did
he oppose the theory of natural selection,

408
genes participate in the effort to produce the resulting plant or
plant organ. Different plants or plant organs on the other hand
are found to share in the work of certain genes, or combinations
of genes; or perhaps we may say that the word 'participation'
points to the future,while the word 'sharing' points to the work
accomplished in the past. Thus, different genes participate in
the work of producing a certain result, while different plants
share with one another the work of certain genes. It must be
admitted, however, that my theory does not necessarily agree
with that used in genetics, but is rather to be regarded as the
latter theory expanded to the utmost limit. (HAYATA 1921 b pp.
101-102.)
All individuals alike possess innumerable genes or factors. The
former present various phenomena according as, on the one hand,
the latter are potent or latent; and on the other, according to the
different combinations or segregations of potent genes.. Consequently, the relation of one individual to the others in phenomenal
appearance is the relation of mutual participation or sharing of
potent and latent genes in individuals.
Then, too, ali genes in individuals may be apparent at one time
and latent at another, and may change their characters, according
to conditions. As genes change, so change individuals. Yet, no
new gene is ever created or produced; no existent gene ever
vanishes. The genes now present are those that have existed
from the eternal past and will continue to exist throughout the
infinite future. (HAYATA 1921 b pp. 102-103.)
As for mankind, and especially for anyone human family, a
genealogy is the history of descent from an ancestor. The customary
manner of making up a genealogy for a family is to trace its
history through descent on the father's side, only one side being
taken into consideration, if not the father's, then the mother's.
Thus, a genealogy is usually considered to be something like a
tree with spreading branches, the ancestor being here compared
to the trunk, and the descendants, to the branches. It is clear
that this graphical method is altogether useless when descent
through both parents is taken into consideration.> (HAYATA 1928
b p. 1869.)
Frorn the Participation Theory, it follows now that it becomes,
in any case, impossible to admit the existence of a single or a
few forms of a particular nature ancestral to the innumerable

409
species which at present we have. Ancestors should (in their
phenomenal appearances) be as different from' one another as are
their descendants. These forefathers, varying in different ways,
presumably by. crossing, one being succeeded by another; must
have given way to the species of the present day. Even supposing
ancestors to have been much simpler than their descendants 'and
more resembling one another, they must have potentially been
sufficiently different to succeed to one generation after another.
(HAYATA 1928 C p. 1R83).
The practical application of the dynamic system is demonstrated
by HAYATA by the following general introduction:
Now, turning back our attention to what we have said before
of the natural classification, we now see that our first attempt in
the course of the latter is to find how species share their genes
with others. Then, the second attempt is to group species into
small or large groups' according to the extent with which they
have shared their genes with others [LoTSY, P ..J. - Evolution
by Means of Hybridization (1916) pp. 137-138], - into groups
which should be subject to alteration,as to their limits as well
as their members, according to whether we select this or that gene
as a criterion for classification, - simply speaking, into dynamic
groups. Finally, our third attempt is to arrange the dynamic
groups thus obtained, such as species, genera or families, according
to the relations of the mutual sharing of genes, - to -arrange
them, not in a fixed order, but in orders varying with views, or
simply speaking to arrange them in a dynamic manner. The arrangement in' dynamic forms is what I have spoken asa dynamic
system. I shall explain' it with examples further on.
Now, let us take four individuals representing four different
species. These four are composed of numerous different genes;
but amongst the latter genes only four, represented symbolically
by a, b, c and d, are so far known to us. We shall try to systematize these species. Let a gene in parentheses denote a latent
gene. Then, the four species may be symbolized as follows: -(ajbcd, a(b)cd, ab(c)d and abc(d). One way of classifying these
four is to group them into two groups, one containing the latent
(a) gene and the other, the potent a gene. There are more
three different ways of making up such' groups of the same category by substituting (b) (c) or (d), in the place of (a). Also, there
are certainly many other ways of grouping the four species so as

410
to classify the species by taking a combination of genes as a
criterion instead of a single gene, such as ab, eb, ac, be, cd, or bcd,
aed, abd, abc, and so forth, - each way being in agreement with
a natural relation according to each respective view. So the groups
themselves are changeable or dynamie,aeeording to whether we
take this gene or that, or this combination or that, as a criterion
for classification, One way of grouping can not be said to be
more natural than others. Anyone is natural so far as it is
understood as dynamic and changeable according to views. But
as soon as it is understood to be the only fixed, unchangeable
one, admitting no other way, it becomes unnatural. (HAYATA 1921
b pp. 105--106).
After this introduction HAYATA demonstrates his principles more
in detail by presenting a dynamic system of natural classification
applied to the Angiosperms, with ENGLER'S system as a Iramework.
(HAYATA 1921 b pp. 159-216). Everyone of ENGLER'S orders and
families is put in the centre of those families to which, according
to HAYATA, it shows a distinct relationship. As to this practical
application of the system, however, I must refer the reader to the
original treatise.
I have quoted HAYATA'S work to such a length because his views
appear to have been unduly overlooked by most present-day taxonomists [the only reference to them that I have seen is the reviews of KRAUSE (1922 a, b) and a few w~rds by CROW (1926)], and
also because those views are very similar to those in which my
own field-experience has resulted. When I learnt to know HAYATA'S works in 1929, I was really surprised to find how my own
field-experience, culminating in my voyage around the world in
1926-1927, had brought me to views in many respects identical
with those expressed by HAYATA several years earlier (Du RIETZ
1929 pp. 421-422). I am mentioning this because I fully realize what opposition those views will meet, and because I must.
take it as a good sign that they have crystallized independently
in the minds of two botanists with wide field-experience in widely
differing parts of the world's vegetation.
When my own field-studies had brought me to the conviction
that the epharmonic characters (s Anpassungsmerkmalea] did not
show the high degree of correlation with the environment postulated in orthodox taxonomy and ecology, I could not help drawing the conclusion that the method of founding the higher taxo-

411
nomical units only upon the traditional Organisationsmerkmale>
could not be maintained. The conflict with the traditional system
arising 'from this consideration brought me to try various alter- .
natives for the grouping of the species in certain genera. During
those attempts I discovered some remarkable regularities in the
g e 0 g r a phi c a I dis tr i but ion of various characters. Studying the New Zealand flora, for instance, I found that in the genus
Celmisia brown and yellow hairs showed a remarkable frequency
in the northern part of South Island, being found there in species
widely differing in other respects (C. Traversii, C. cordatifolia, C.
Macmahoni, C. Dallii, and C. bieracifoliai, but of much rarer occurrence in other parts of New Zealand. In the genus Dracophyllum similar regularities were found (comp. above p. 398): terminal panicles or racemes are found in many species of the
northern and central part of North Island (from the small dwarf- .
shrub species D. recurvum to the big tuft-tree species D. laiifolium, D. recurvatum etc.) but only in one species of South Island
(D. Traoersiii and in no species at all in the southern part of
South Island, where lateral panicles or racemes are characteristic
even in the large-leaved tuft-tree species D. fiordense (comp, OLIVER 1928).
In the genus Euphrasia I found that digitate leaves
(a character found in all the South American species of that genus but very rare elsewhere) were characteristic for several species (rather different in other respects) of the southern part of
South Island, but in the rest of the Australian-New Zealand region
only occured in one Tasmanian species (comp. WETTSTEIN 1896
b), i. e. that this South American character was found in widely
differing species only in the parts of Australasia geographically
most related to South America. In the same genus I found
certain life-forms concentrated to certain parts of New Zealand,
where they could impossibly be correlated with a certain climate,
but in those parts appearing in the most different groups of the
genus (just the same if those groups were founded upon the flowers or upon the leaves). - To all those regularities (and many
others that cannot be mentioned here) I could not find any other
possible explanation than that in the syngameon originally formed
by each of those genera, certain genes had been generally distributed only in certain districts, but in those districts had inlected
the whole population. In this respect there was. no difference
whatever in the behaviour of Organisationsmerkmale> and An-

412

passungsmerkrnales in the traditional sense. But if epharmonic

characters could behave in this way within a certain genus, it
did not then look impossible that they could do the same also
in larger populations, i. e. that many cases of epharrnonical convergence could be explained simply by the accidental distribution
of a certain gene in the whole population formed by a certain
family, order or even class in a certain region already at the
time when this population formed 0 n e vast syngameon.
Just the same point of view may be applied to the many
examples known of VAVILOV'S law of homologous variation (comp,
above p. 346). If homologous forms appear in species belonging
to quite different genera or families, why may not this simply
be the result of the presence of the same gene in all those forms?
For nearly related species this explanation already appears to be
generally accepted among geneticists (comp. PHILIPTSCHENKO 1927).
But as soon as we come to species belonging to different higher
units of the traditional system, PHILIPTSCHENKO finds its impossible
to apply this theory of genotypical parallelism, .postulating a
principially different morphological parallelism, auf deren Eigentiimlichkeiten del' Begriff del' Gene und del' genotyplschen Struktur iiberhaupt nicht anwendbar ist (PHILIPTSCHENKO 1927 pp.
74-75.) As an intermediate type of parallelism PHILIPTSCHENKO
accepts the oecotypical parallelism, which bald von gleichen,
bald auch von vcllig verschiedenen genotypischen Strukturen
abhangen kann> (1. c.). As a result of this negative attitude
towards the explanation of the morphological parallelism PHILIPTSCHENKO finds it possible to explain only the evolution of
biotypes, jordanons and linneons>, or the rnicro-evolution in the
light of modern genetics, while he considers the evolution of
higher taxonomical groups or rnacro-evolution> to lie completely
ausserhalb ihres Gesichtsfeldes> (pp. 93-94). May be he is right,
but certainly my own experience makes me more inclined to
follow HAYATA in looking for the explanation also of morphological parallelism in genotypical resemblance.
.
Of course we must not forget that it is not allowed to translate
character> into gene. We can 'only see the characters but not
the genes, and a certain character is supposed to be the result of
the coaction of an immense number of genes, a llof which are
necessary for. the formation of the character. The disappearing
of a certain character may thus be the result of the lack of quite

413
different genes in different cases; and vice versa (comp. for instance
JOHANNSEN 1913 pp. 666-667, 'and especially the very. clear discussion by HAGEDOORN 1921). The matter looks still more complicated
in the light of GOLD.sCH~IDT'S(1920; 1928) theory ofq u ant ita t i v e
gene-differences. 'Futurevresearch must -show to what extent those
difficulties .maybe overcome in practical taxonomy; Theoretically
we may attempt to follow the suggestion of HAYATA to found the
higher units of taxonomy upon the distribution of genes, put
practically it looks quite possible that we will still have to be content with founding them upon the distribution of characters.
Anyhow, the main reform that is needed in higher taxonomy
certainly is that we will have to take genotypical characters as
they are without any subjective valuation, neither in Organisationsmerkrnale and Anpassungsmerkrnale> nor in "progressive
and regressive> characters. And here HAYATA certainly is right
both in pointing out that every natural system must form a
ne two r k and not a phylogenetical tree,' and in rejecting the
idea of 0 n e single static system.
My main objection against the principles of HAYATA is that I
think he is underrating the taxonomical importance of the lines
or zones of discontinuity formed in the series of biotypes by
geographical, ecological or physiological isolation. As has been
the main purpose of this paper to show, I think that the fun d amen t a I un its of taxonomy should be founded just upon those
lines or zones of discontinuity, and the purely morphological
grouping taken as a refuge first when we proceed to populations
of higher rank in which that method is failing. I am also more
inclined than HAYATA to leave the question open whether no new
gene is ever created or produced and whether no existent gene
ever vanishes. How life originated on earth in the eternal
past we shalI probably never know, nor how it will develop or
if it will even exist in the eternal future.
But with those restrictions (and perhaps some others of minor
importance), I feel convinced that the participation theory and
the dynamic system of HAYATA will prove a most excelIent workingbase for the great reform of higher botanical taxonomy that is
so greatly needed. I cannot refrain from citing in this connection
the folIowing critical remarks ma.de by KRAUSE in his review of
1 DIELS (1921 pp. 187-189) has also pointed out, that the relationship of plants
is far better illustrated by a network than by the traditional phylogenetical tree.

28 -

30531. Svensk Botanisk: Tid8krift.

1930.

414

first work on those subjects, because they give the best

characteristic possible of the prejudices dominating present taxonomical botany:
.
Die angefiihrten Proben zeigen, dass Verf. einmal Analogien,
wie z. B. Anordnung der Bliiten in Kolben, erikoide Blattgestalt,
ahnliche Fruchtentwicklung, Kleinheit del' Samen usw. als Grundlagen fUr die Annahme einer natnrlichen Verwandtschaft ansieht.
Denn nicht selten hat man den Eindruck, als ob Verf. rein
ausserliche, vielleicht nur durch gleiche Lebensverhaltnisse bedingte
Ahnlichkeiten, wie Blattform, Sukkulenz oder dergleichen, auf
eine Stufe stellt mit anderen wesentlichen Merkmalen im Bliitenbau,
wie Beschaffenheit und Stellung del' Samenanlagen usw. (KRAUSE
1922 a p. 51).
I do not think that the fundamental difference between the
view of HAYATA arid myself and that of traditional taxonomy could
have been expressed better. For my own part, it is just in the
raising of those neglected epharmonic characters (or Anpassungsmerkrnale) to the same taxonomical rank 11S that so long
monopolized by the floral characters only, 'that I see the only
way to an unprejudiced study of the higher taxonomical units.
But HAYATA'S excellent idea of the dynamic system certainly ought
to make it possible for old and new systems to exist in friendship
side by side in future taxonomy.
HAYATA'S

Plantbiological Institution of Upsala University, June 25th, 1930.

LITERATllRE CITED.
H. H., On the Hybridity of Coprosma Cunninghamii Hook. f. New Zeal. Journ. Sci. & Technol.,7. Wellington 1924.
-')-, Illustrations of Wild Hybrids in the New Zealand Flora. - Genetica, 7. 's-Gravenhage 1925.
-')-, The F l Progeny Resulting from Crossing Coprosma propinqua ~
with C. robusta if. - Ibid., 8. 1926 (a).
.
.
-')-, Illustrations of Wild Hybrids in the New Zealand Flora IL - Ibid.,
8. 1926 (b).
-,>-c-, Illustrations of Wild Hybrids in the New Zealand Flora III. - Ibid.,
8. 1926 (c).
-')-, Illustrations of Wild Hybrids in the New Zealand Flora IV.
Ibid., 9. 1927 (a).
ALLAN,

415
ALLAN, H. H., Illustrations of Wild Hybrids in the New Zealand Flora V.
- Ibid., 9. 1927 (b).
----, New Zealand Trees and Shrubs and How to Identify them. Wellington 1928.
---, The F 2 Progeny Resulting from the Crossing of Coprosma proplnqua
~ with C. robusta if. Genetica, 11. 's-Gravenhage 1929 (a).
- - , Illustrations of Wild Hybrids in the New Zealand Flora VI. - Ibid.,
11. 1929 (b).---, SIMPSON, G., and THOMSON, J. S., A Wild Hybrid Hebe-Community
in New Zealand. - Ibid., 8. 1926.
ALMQUIST, ERIK, Upplands vegetation och flora. - Acta phytogeographica
suecica, 1. Upsala 1929.
ALMQUIST, ERNST, Linnes Vererbungsforschungen. - Engler's Bot. Jahrb.,
55. Leipzig 1917.
- - , Zur Artbildung in der freien Natur. - Acta Horti Bergiani, 9: 2.
Upsala 1926.
ALPATOV, W. W., Die untersten systematischen Kategorien und die
Variabilitat bei den Ameisen, - Berichte der II. Russischen Entomophytopatologischen Versammlung. Leningrad 1922.
---, Decapoda, aus den Sammlungen der Expedition des Instituts im
Jahre 1921. Berichte des Wissenschaftlichen Meeresinstituts, 7.
Moskwa 1923.
---, Die Definition der untersten systematischen Kategorien vom
Standpunkte des Studiums der Variabilitat der Ameisen und der Crustaceen. - Zoolog, Anzeiger, 60. Leipzig 1924.
ANDERSON, E., The Problem of Species in the Northern Blue Flags, Iris
versicolor and Iris virginica. - Ann. Miss. Bot. Garden, 15: 3. Lancaster, Pa, 1928.
ARWIDSSON, TH., Om nagra Betula-bestarnningar. - Upsala 1930.
BABCOCK, E. B., and HALL, H. M., Hemizonia congesta. A Genetic, Ecologic, and Taxonomic Study of the Hay-Field Tarweeds. - Univ. of
Calif. Publ. in Bot., 13: 2. Berkeley 1924.
BANNIER, J. P., Untersuchungen tiber apogame Fortpflanzung bei einigen
elementaren Anten von Erophila verna. - Rec. des travaux botaniques
neerlandais, 20. Amsterdam 1923.
BATESON, W., Materials for the Study of Variation Treated with Especial
Regard to Discontinuity in the Origin of Species. - London 1894.
---, Problems of- Genetics. - New Haven 1913.
BAUR, E., Einfuhrung in die experimentelle Vererbungslehre. 3:e u. 4:e
Auflage. - Berlin 1919. 7-11 Auflage. - Berlin 1930.
BERG, L. S., Nomogenesis or Evolution Determined by Law. - London
1926.BRIQUET, J., Regles Internationales de la nomenclature botanique adoptees
par le congres international de botanique de Vienne 1905. Deuxieme
edition mise au point d'apres les decisions du eongres international
de botanique de Bruxelles 1910. Publiee au nom de Iacommlsslon de
redaction du congres. - Jena 1912.

416
BUCH, L. v., Physikalische Beschreibung der kanarischen Inseln. - Berlin 1825.
CAJANDER, A. K., Metsanhoidon perusteet. I. Kasvibiologian ja kasvimaantieteen paapiirteet. - Porvoo 1916.
-')e--, Einige Bemerkungen iiber die Entstehung der Arten insbesondere
innerhalb der Gruppe der Holzgewachse, - Acta Forestalia Fennica,
21. Helsingfors 1921.
CHEESEMAN, T. F., Manual of the New Zealand Flora, ed. 2. - Wellington
1925.
CLAUSEN, J., Studies on the Collective Species Viola tricolor L. II. - Bot.
Tidsskr., 37. Kobenhavn 1922.
- l ) - , Genetical and Cytological Investigations on Viola tricolor L. and
V. arvensis Murr. - Hereditas, 8. Lund 1926.
-,~, Nogle nyere Synspunkter vedrerende Evolutionslteren. - Naturens
Verden, 11. Kobenhavn 1927.
COCKAYNE, L., A Botanical Excursion during Midwinter to the Southern
Islands of New Zealand. - Trans. New Zealand Inst., 36. Wellington
1904.
- l ) - , Some New Species of New Zealand Flowering-Plants. Ibid., 47.
1915.
- l ) - , Notes on New Zealand Floristic Botany, Including Descriptions of
New Species &c. (No.3). - Ibid., 50. 1918.
-')-, The Vegetation of New Zealand. - Die Vegetation der Erde, 14.
Leipzig 1921.
- l ) - , Hybridism in the New Zealand Flora. The New Phytologist,
22. London 1923.
-')-, List of Supposed Wild New Zealand Hybrids amongst the Vascular
Plants. - Appendix to J. P. LOTSY, Evolution Considered in the Light
of Hybridization. Christchurch 1925. -,)-----.:, Monograph on the New Zealand Beech Forests. Part I. The Ecology of the Forests and the Taxonomy of the Beeches. - New Zeal.
State Forest Service Bull. 4. Wellington 1926.
-~, Monograph on the New Zealand Beech Forests.
Part II. The
Forests from the Practical and Economical Standpoints. - Ibid. 1927.
- l ) - , The Vegetation of New Zealand. ~ Die Vegetation der Erde, 14.
2d ed. Leipzig 1928.
-,)-, Hybridism in the Forests of New Zealand. - Acta Forestalia Fennica, 34. Helsingfors 1929.
- l ) - , and ALLAN, H.H., The Present Taxonomical 'Status of the New
Zealand Species of Hebe.- Trans. New Zealand Inst., 57. Wellington
1926 (a).
- l ) - , -~, Notes on New Zealand Floristic Botany, Including Descriptions of New Species, &c. No.5. - Ibid., 57. 1926 (b).
- l ) - , - l ) - , The Naming of Wild Hybrid Swarms. Nature, 118. London 1926 (c).
--, -')-, The Bearing of Ecological Studies in New Zealand on Botanical Taxonomic Conceptions and Procedure. - Journ. of Ecology, 15.
Cambridge 1927.

417
COCKAYNE, L., and ATKINSON, E., On the New Zealand Wild Hybrids of
Nothofagus. - Genetica, 8. 's-Gravenhage 1926.
-)}--, and PHILLIPS TURNER, E., The Trees of New Zealand; - New Zealand State Forest Service. Wellington 1928.
COUTAGNE, G., Recherches sur Ie polymorphisme des mollusques de France.
- Ann. Soc. d' Agric. Sci. et Industr. Lyon 1895.
CRAMPTON, H. E., Studies on the Variation, Distribution, and Evolution
of the Genus Partula: The Species Inhabiting Tahiti. - Carnegie Institution of Washington, Publ. 228. Washington 1917.
-'>--, Contemporaneous Organic Differentiation in the Species of Partula
Living in Moorea, Society Islands. - The American Naturalist, 59.
New York 1925 (a).
-')-, Studies on the Variation, Distribution, and Evolution of the Genus
Partula. The Species of the Mariana Islands, Guam and Saipan. Carnegie Inst. of Wash., Publ. 228: A. Washington 1925 (b).
CROW, W. B., Phylogeny and the Natural System. - Journ. of Gen.,
17. Cambridge 1927.
DANSER, B. H., Uber einige Aussatversuche mit Rumexbastarden. - Genetica, 6. 's-Gravenhage 1924.
-'>--, Ueber die Niederlandisch-Indischen Stachytarpheta-Arten und
ihre Bastarde, nebst Betrachtungen iiber die Begrenzung der Arten
im allgemeinen. - Annales du Jardin Botanique de Buitenzorg, 40.
1929 (a).
-')-, Ueber die Begriffe Komparium, Kommiskuum und Konvivium
und ueber die Entstehungsweise der Konvivlen. - Genetica, 11. 's-Gravenhage 1929 (b).
DARWIN, CH., On the Origin of Species by Means of Natural Selection,
or the Preservation of Favoured Races in the Struggle for Life. London 1859.
5th edition. - London 1869.
DE VRIES, H., Die -Mutationstheorie. Versuche und Beobachtungen iiber
die Entstehung von Arten im Pflanzenreich. Leipzig 1901-1903.
DIELS, L., Die Methoden der Phytographie und der Systematik der Pflanzen. Abderhaldens Handbuch der biologischen Arbeitsmethoden,
Abt. XI, Teil. 1, Heft 2. Berlin und Wien 1921.
DOBRZANSKY, TH., Die geographische und indlvlduelle Variabilitat von
Harmonia axyridis Pall. in' ihren Wechselbeziehungen. - Biolog. Zentralblatt, 44. Leipzig 1924.
DOLLO, L., Les lois de I'evolutlon. - Bull. Soc. Belge de Geologie, 7. 1893.
Du RIETZ, G. E., Zur methodologischen Grundlage der modernen Pflanzensoziologie. - Diss. Upsala 1921.
-')-, Der Kern der Art- und Assoziationsprobleme. - Botaniska Notiser
1923. Lund 1923 (a).
-'>--, De svenska Helianthemum-arterna. - Ibid. 1923 (b).
-'>--, Die Soredien und Isidien der Flechten. - Svensk Botanisk Tid- ,
skrift, 18. Stockholm 1924 (a).
-'>--, Kritische Bemerkungen iiber die Parmelia perlata-Gruppe.
Nyt Magazin for Naturvidenskaberne, 62 . Oslo 1924 (b).

418

Du RIETZ, G. E., Helianthemum nummularium (L.) Dunal X H. ovatum

(Vlv.) Dunal in Stidschweden gefunden. -

Bot. Not. 1925. Lund 1925.

--, Den svenska botaniska Australasia-expeditionen 1926-1927.

Ymer, 48 (1928): 4. Stockholm 1929.
Vegetationsforschung auf soziationsanalytischer Grundlage.
Abderhaldens Handbuch der biologischen Arbeitsmethoden, Abt. XI,
Teil 5. Berlin und Wien 1930.
EIMER, TH., Orthogenesis der Schmetterlinge. - Leipzig 1897.
EKMAN, S., Die Selektionstheorie und die Selektionsversuche W. Johannsens in sog. reinen Linien. - Zeitschr. f. indukt. Abstamm.- u. Vererbungslehre, 48. Berlin 1928.
-')-, Utvecklingslaran och den nyare forskningen. - Stockholm 1929.
ENGLER, A., Monographie der Gattung Saxifraga. - Breslau 18n.
-')-, Beitrage zur Entwicklungsgeschichte der Hochgebirgsfloren erlautert an der Verbreitung der Saxifragen. Abhandl. d. Konigl.
Preuss. Akad. d. Wissensch., 1916, Phys.-Math. Klasse, Nr, 1. Berlin
1916.
ETTINGSHAUSEN, C., Freih. VON, Die Formelemente der europaischen Tertiarbuche (Fagus Feronire Ung.). Denkschriften d. math.-naturw.
Classe d. k. Akad. d. Wiss., 61. Wien 1894.
--, und KRASAN, F., Beitrage zur Erforschung der atavistischen Formen
an lebenden Pflanzen und ihrer Beziehungen zu den Arten ihrer Gattung. - Ibid., 55, 56, 57. Wien 1888-1889.
FLODERUS, B., Bidrag till kannedomen om Salix-floran i Tome Lappmark.
- Arkiv fOr Botanik, 8: 9. Upsala and Stockholm 1909.
-')--, Bidragtill kannedomen om Novaja Sernljas Salices. - Sv. Bot.
Tidskr., 6. Stockholm 1912.
-')-, Om Gronlands Salices. - Meddelelser om Gronland, 63. Kobenhavn 1923.
----, On the Salix Flora of Kamtchatka. - Arkiv f(ir Botanik, 20 A: 6.
Stockholm 1926 (a).
--, Salix. - Lindman, Svensk Fanerogamflora. Stockholm 1926 (b).
FOCKE, W.O., Die Pflanzen-Mischlinge. Ein Beitrag zur Biologie der
Gewachse. - Berlin 1881.
FRIES, E., Botaniska Utflygter, 2. - Stockholm 1852.
GALLESIO, G., Teoria della riproduzione vegetale. Pisa 1816.
GALLOE, 0., Individforskning i Planteriget. - Dansk Botanisk Arkiv,
5: 12. Ktibenhavn 1928.
GAMS, H., Noch einmal die Herkunft von Cardamine bulbifera (L.) Crantz
und Bemerkungen tiber sonstige Halb- und Ganzwaisen. - Ber. d.
Deutsch. Bot. Ges., 40. Berlin 1923.
GOLDSCHMIDT, R., Die quantitative Grundlage von Vererbung und Artbildung. - Roux' Vortrage u. Autsatze, 24. Berlin 1920.
--, Physiologische Theorie der Vererbung. - Berlin 1927.
-')-, SEILER, I., and POPPELBAUM, H., Untersuchungen zur Genetik
der geographischen Variation. 1 . - Arch. f. mikr. Anat, u. Entwicklungsgesch., 101. 1924.

~-,

419

GULICK, J. T., The Variation of Species as Related to their yeographical

Distribution, illustrated by the Achatlnelllnse, - Nature, 5. London
1872 (a).
->)-, Diversity of Evolution under One Set of External Condition. .Iourn. Linn. Soc., Zool., 11. London 1872 (b).
--, Divergent Evolution through Cumulative Segregation. Ibid.,
20. 1~87. (Also in Rep. Smithson. Inst. 1891.)
---, Intensive Segregation. - Ibid., 23. 1889, (Reprinted in GULICK
1905.)
-)-, Divergent Evolution and the Darwinian Theory. - Amer. Journ.
of Science. 1890.
-)-, Evolution, Racial and Habitudinal. - Carnegie Inst. of Washington,
Publ. 25. Washington 1905.
GUNNARSSON, J. G., Monograti over Skandinaviens Betulee. - Malmo
1925.
Guppy, H. B., Observations of a Naturalist in the Pacific between 1896
and 1899. II. Plant-Dispersal. London 1906.
-:"')--, Plant-Distribution from an Old Standpoint. - Trans. Viet. Inst.
London 1907.
-,)-, Plants, Seeds, and Currents in the West Indies and Azores. London 1917.
HAAGEDOORN, A. L., and HAGEDOORN-VORSTHEUVEL LA BRAND, A. C.,
The Relative Value of the Processes Causing Evolution. - The Hague
1921.
HAACKE, W., Gestaltung und Vererbung. - Leipzig 1893.
HAECKEL, E., Generelle Morphologie der Organismen. I. Allgemeine
Anatomie der Organismen. Berlin 1866.
---, Nattirliche Schopfungsgeschichte, - Berlin 1868.
HALL, H. M., The Genus Haplopappus. A Phylogenetic Study in the
Composites. - Carnegie Inst. of Washington, Publ. 389. Washington
1928.
-)-, and CLEMENTS, F. E., The Phylogenetic Method in Taxonomy.
The North American Species of Artemisia, Chrysothamnus and Atriplex.
- Ibid., Pub I. 326. 1923.
HARTERT, K, Die Vogel der palaarktischen Fauna. Bd. 1-3 und Nachtrag I. Berlin 1910-1923.
HAYATA, B., An Interpretation of Goethe's Blatt in his Metamorphose
der Pflanzen, as an Explanation of the Principle of Natural Classification. - leones Plantarum Formosanarum, 10. Taihoku 1921 (a).
- - , The Natural Classification of Plants According to the Dynamic
System. - Ibid., 10. 1921 (b).
-')-, Succession in the Vegetation of Mt. Fuji and the Formulation of a
New Theory, the Succession Theory, in Opposition to the Natural
Selection Theory. Proceed. Third Pan-Pacific Science Congress
Tokyo 1926, Vol. II, Tokyo 1928 (a).
-)-, The Succession and Participation Theories and their Bearings upon
the Objects of the Third Pan-Pacific Science Congress. - Ibid. Tokyo
1928 (b).

420
HA YAT A, B., The Relation between the Succession and Participation Theories
and their Bearings upon the Natural System. - Ibid. Tokyo 1928 (c).
HEINCKE, FR., Naturgeschichte des Herings. - Abhandl. d. Deutsch.
Seefischerei-Vereins, II. Berlin 1898.
HERBERT, W., On theProductton of Hybrid Vegetables; with the Result
of many Experiments made in the Investigation of the Subject. ~
Trans. Horticult. Society of London, 4. London 1820.
HERIBERT-NILSSON, N., Experimentelle Studien tiber Variabilitat, Spaltung, Artbildung und Evolution in der Gattung Salix. - Lunds Universitets Arsskrift, N. F., Avd. 2, Bd. 14, Nr. 28 (Kungl. Fysiografiska
Sallskapets Handlingar, N. F., Bd. 29, Nr. 28). Lund 1918.
HITCHCOCK, A. S., Methods of Descriptive Systematic Botany. - New
York 1925.
HOOKER, J. D., Flora Antarctica. - London 1847.
- l ) - - , Flora Novse-Zelandise. London 1853.
- 1 ) - - , Handbook of the New Zealand Flora. London 1867.
HUXLEY, TH., Upon Animal Individuality. - Proceed. of the Royal Institution. Nov. ser, Vol. I. London 1855.
.
JENNINGS, H. S., Genotype and Pure Lines. - Science, 34. New York
1911.
- 1 ) - - , Production of Pure Homozygotic Organisms from Heterozygotes
by Self-Fertilization. - American Naturalist" 46. New York 1912.
JOHANNSEN, W., Ueber Erblichkeit in Populationen und in reinen Linien.
- Jena 1903.
- 1 ) - , Arvelighedsleerens Elementer. Kobenhavn 1905.
---, Elemente der exakten Erblichkeitslehre. Ed. 1. Jena 1909.
Ed. 2. - Jena 1913.
JOHANSSON, K., Enumerantur Hieracia vulgata Sueciae. - Arkiv for Botanik, Bd 21 A, N:o 15. Stockholm 1927.
JORDAN, A., Diagnoses d'especes nouvelles ou meconnue, tom. I. - Paris
, 1864.
- l ) - - , Remarques sur Ie fait de l'existence ensociete, a I'etat sauvage des
especes vegetates affines. - Lyon 1873.
JORDAN, D. S., The Origin of Species through Isolation. - Science, 22.
New York 1905.
JORDAN, K., Der Gegensatz zwischen geographischer und nichtgeographischer Variation. - Zeitschr. f. wiss. Zoologie, 83. Leipzig 1905.
KARPECHENKO, G. D., Polyploid hybrids of Raphanus sativus L. x Brassica oleracea L. - Zeitschr. f. indo Abstamm.- u. Vererbungslehre, 48.
Berlin 1928.
KERNER, A., Konnen aus Bastarden Arten werden? - Osterreich. Bot.
Zeitschr., 21. Wien 1871.
- 1 ) - , Pflanzenleben.
Zweiter Band. Die Geschichte der Pflanzen. --'Leipzig und Wien 1891. Ed. 2. 1898.
KIRK, T., On the Botany of the Antarctic Islands. - Rep. 3d Meeting of
the Australasian Assoc. f. Advanc. of Science. Wellington 1891.
' - 1 ) - , A Revision of the New Zealand Gentians. Trans. New Zealand
Inst., 27. Wellington 1895.

421
KLEINSCHMIDT, 0'-, Der Formenkreis Falco- Hierofalco. ~ Aquila, 8.
Budapest 1901.
-,}--, Die Formenkreislehre und das Weltwerden .des Lebens, Halle
1926.
KLINGSTEDT, H., Der Begriff der Art an dem Zeltkorper-Begrlff erlautert;
- Memoranda Soc. pro Fauna et Flora Fennica, 4. 1927-1928. Helsingfors 1928.
KRASAN, F., Uber- continuirliche und sprungweise Variation. ~ Engler's
Bot. Jahrb., 9. Leipzig 1888.
KRAUSE, K., Review of Hayata 1921 a and b. - Engler's Bot. .Iahrb.,
57. Leipzig 1922 (a).
-)}--, Eine neue Form des nattirlichen Systems. - Die Naturwissenschaften. - Berlin 1922 (b).
KOLREUTER, J. G., Vorlaufige Nachricht von einigen das Geschlecht der
Pflanzen betreffenden Versuchen und Beobachtungen. - Leipzig 1761.
LAMARCK, J. DE, Recherches sur l'organisation de corps vivants et par"
t.iculierement sur son origine. - Paris 1802.
-)}--, Philosophie zoologique. - Paris 1809.
LEHMANN, E., Lotsy's Anschauungen tiber die Entwickelung des Deszendenzgedankens und den jetzigen Standpunkt der Frage. - Zeitschr.
f. indo Abstamm.- U. Vererb.-Lehre, 11. Berlin 1913.
-,}--, Bemerkungen zu der vorstehenden Entgegnung Lotsys. - Ibid., 12.
1914 (a).
.
-)}--, Art, reine Linie, isogene Einheit. - BioI. CentralbI., 34. Leipzig
1914 (b).
LINNlEUS, C., Dissertatio botanica de Peloria, ~ Upsalise 1744. [Reprinted in Amoenitates academicse 1 (1749).]
- l } - - , Disqvisitio de qusestlone ab academia imperiali scientiarvm petropol. in annvm 1759 pro prrernio proposita: Sexum plantarum argumentis
et experimentis novis ... -Petropoli 1760. [Reprinted in Amoenitates
academicse 10 (1790) under the title Dlsquisitio de sexu plantarum.
English translations by J. E. Smith (London 1786)and R. J. Thornton
(London 1807).]
LLOYD, R. E., The Growth of Groups in the Animal Kingdom. - London
1912.
LOTSY, J. P., Versuche tiber Artbastarde und Betrachtungen tiber die
Moglichkelt einer Evolution trotz Artbestandigkeit. - Zeitschr. f. indukt. Abstamm.- u. Vererb.-Lehre, 8. Berlin 1912.
-,}--, Fortschritte unserer Anschauungen tiber Descendenz seit Darwin
und der jetzige Standpunkt der Frage. - Progressus rei botanicse, 4.
Jena 1913.
- l } - - , Meine Anschauungen tiber die Entwicklung des Descendenzgedankens seit Darwin und den jetzigen Standpunkt der Frage, eine Entgegnung zu der daran von Prof. Dr. E. Lehmann geiibten Kritik. - Zeitschr.
f. indukt. Abstamm.- u. Vererb.-Lehre, 12. Berlin 1914 (a).'
- 1 ) - , Prof. E. Lehmann tiber Art, reirie Linie und isogene Einheit. .
BioI. CentralbI., 34. Leipzig 1914 (b).
--,}--, Evolution by Means of Hybridization. - The Hague 1916.

422
LOTSY, J; P., Quintessence de la theorie du croisement. - Archives neerlandaises des sciences exactes et naturelles. Ser, III B, Tome III. 1918.
-.)-, Proeven en beschouwingen over evolutie. - Genetica, 1. 's-Gravenhage 1919 (a).
-:-l)-, Cucurbita-strijdvragen. Ibid,1. 1919 (b).
- . - , Heribert Nilsson's onderzoekingen over soortsvorming bij Salix
met opmerkingen mijnerzijds om trent de daarin en in publlcaties van
anderen uitgeoefende kritiek aan mijn soorts-definitie. - Ibid., 2. 1920.
-.)-, Evolutie-factoren. - Ibid., 3. 1921.
-.)-, Evolution Considered in the Light of Hybridization. - Christchurch,
N. Z., 1925 (a). - [Deutsche Ubersetzung (Evolution im Lichte der
Bastardierung betrachtet) in Genetica, 7.J
-.)-, Species or Linneon. - Genetica, 7. 's-Gravenhage 1925 (b).
-')-, Has Winge Proved that Erophila is not Apogamous? - Ibid., 8.
1926.
-.)-, and GODDlJN, W. A., Voyages of Exploration to Judge ofthe Bearing
of Hybridization upon Evolution. I. South-Africa. - Ibid., 10. 1928.
MATTFELD, J., Uber hybridogene Sippen der Tannen nachgewiesen an den
Formen der Balkanhalbinsel. - Bibliotheca Botanica, 100. Stuttgart
1930.
MAYER, A. G., Some Species of Partula from Tahiti; A Study in Variation.
- Memoirs of the Museum of Comparative Zoology at Harvard College,
26: 2. Boston 1902.
MENDEL, G., Versuche iiber Pflanzen-Hybriden.- Verhandl. d. naturforsch.
Vereines in Briinn, 4. Brunn 1865.
MURR, J., Hinterlassene Halbwaisen in unserer Flora. - ' Feldkircher Anzeiger 1919.
MUNTZING, A., Cases of Partial Sterility in Crosses within a Linnean Species. - Hereditas, 12. Lund 1929.
- l ) - , Outlines to a Genetic Monograph of the Genus Galeopsis. -Ibid.,
13. 1930.
NILSSON-EHLE, E., Den modarna arftlighetslaran och dess betydelse for
vaxtodlingen. - 'Stockholm 1915.
NAUDIN, CH., Nouvelles recherches sur I'hybridite dans les vegetaux. Annales des sciences naturelles, 4me serie, Botanique, 19. Paris 1863.
NAGELI, C. v., Ueber den Einfluss ausserer Verhaltnisse auf die Varietatenbildung im Pflanzenreiche. - Sitzungsber. d. konigl. bayer. Akad.
d. Wissensch. zu Munchen, Jahrg. 1865, Bd. II. Munchen 1865.
- l ) - , Das gesellschaftliche Entstehen neues Spezies. Ibid. 1872.
-:-')-, Mechanisch-physiologische Theorie der Abstammungslehre, Munchen und Leipzig 1884.
OLIVER, W. R. B., A Revision of the Genus Dracophyllum. - Trans.
New Zealand Inst., 59. Wellington 1928.
OSTENFELD, C. H., Further Studies on the Apogamy and Hybridization
. of the 'Hieracia. - Zeitschrift f. indukt. Abstamm.- u. Vererb.-Lehre, .
3. Berlin 1910.
- l ) - , Experiments on
the Origin of Species in the Genus Hieracium
(Apogarny and Hybridism). - New Phytologist, 11. London 1912.

423
OSTENFELD, C. H., Kimdannelse uden Befruktning og Bastarddannelse
hos nogle kurvblomstrede samt disse Forholds Betydnlng for Formernes
Konstans. Kgl. Veterinter- og Landbohojskolens Aarsskrift, 1919.
Kobenhavn 1919.
PARR,A. E., Adaptiogenese und Phylogenese, Zur Analyse der Anpassungserscheinungen und ihrer Entstehung. - Abh. z. Theorie d. org. Entw.
Neue Folge. 1. Berlin 1926.
PETERSEN, H -. E., Indledende Studier over polymorphien hos Anthriscus
silvestris (L.) Hoffm. - Diss. Kobenhavn 1914. (Also in Dansk Bot.
Arkiv, 1: 6. Kobenhavn 1915.)
- t - - , Nogle Studier over Pimpinella saxifraga. -Bot. Tidsskr., 37.
Kobenhavn 1921.
,
--, Etudes ulterieurs sur la polymorphie de l' Anthriscus silvester (L.)
Hoffm. - Dansk, Bot. Arkiv, 4: 2. Kobenhavn 1922.
-.}--, Studier over Polymorphien hos Vaccinium uliginosum. - Bot.
Tidsskr., 38. Kobenhavn1924.
-.}--, Uber die Variation der Potentilla erecta (L.) Dalla Torre: - Ibid.,
39. 1926.
PHILIPTSCHENKO, J., Variabilitat und Variation. - Berlin 1927.
PILSBRY, H. A., assisted by COOKE, C. M., Manual of Conchology. Vol. 22.
Achatinellidse. Philadelphia 1912-1914.
PLATE, L., Selektionsprinzip und Probleme der Artbildung. Ein Handbuch des Darwinismus. 4:e Aufl. - Leipzig und Berlin 1913.
- - , Prinzipien der Systematik mit besonderer Beriicksichtigung des
Systems der Tiere. - Die Kultur der Gegenwart, herausgegeben von
Paul Hinneberg, Teil III, Abteilung IV, Bd. 4. Leipzig und Berlin
1914.
RAUNKllER, C., Art. -Salmonsens Konversationslexikon. Kobenhavn
1913.
-)}--, Uber den Begriff der Elementarart im Lichte der modernen Erblichkeitsforschung. - Zeitschr. f. indukt. Abstamm.- und Vererb.Lehre, 19. Berlin 1918.'
- t - - , Eremitageslettens Tjdrne. Isoreagentstudier I. Det Kgl. Danske
Videnskab. Selskab, Biolog. Meddel., V, 1. Kobenhavn 1925.
REMANE, A., Exotypus-Studien an Saugetieren. I. - Zeitschr. f. Saugetierk., Bd. 3. 1928.
REICHENOW, A., Uber die BegrifIe der Subspecies. - Journ. f. Ornithol.,
52. 1904.
RENNER, 0., Artbastarde bei Pflanzen. - Handbuch der Vererbungswissenschaft, II A. Berlin 1929.
RENSCH, B., Rassenkreisstudien bei Mollusken I. Der Rassenkreis der
Felsenschnecke Campylaea zonata Studer. - Zool. Anzeiger, 67. 1926.
- - , bas Prinzip geogsaphischer Hassenkreise und das Problem der
Artbildung. Berlin 1929.
ROBERTS, H. F., Plant Hybridization before Mendel. - Princeton 1929.
ROBSON, G. C., The Species Problem. An Introduction to the Study of
Evolutionary Divergence in Natural Populations. - Edinburgh and
London 1928.

424
ROMANES, G., Physiological selection. ~ Journ. Linn. Soc., Zool., 19.
London 1886.
- l ) - - , Darwin and after Darwin.
III. Post-Darwinian questions. Isolation and Physiological Selection.' - London 1897.
ROSEN, F., Systematische und biologische Beobachtungen tiber Erophila
verna. ~ Botanische Zeitung, 47. 1889.
------, Uber Bastarde zwischen elementaren Spezies der Erophila verna.
- Ber. d. deutsch. Bot. Gesellsch., 28. Berlin 1910.
-')--, Die Entstehung der elementaren Artenvon Erophila verna. Cohn's Beitrage zur Biologie der Pflanzen, 10. Breslau 1911.
~.)--, Das Problem der Erophila verna. Bibliographia Genetica, 1. 1925.
ROSENBERG, 0., Die Reduktionsteilung und ihre Degeneration in Hieracium. - Sv. Bot. Tidskr., 11. Stockholm 1917.
-')-, Speziesbildung mit Vervielfalfigung von Chromosomen. - Verh.'
d. V. intern. Kongr. f. Vererbungswiss. Berlin 1927. Leipzig 1928.
Russow, E., Ueber Studien an einheimischen Torfmoosen. - Sitz.ber. d.
Dorpat. Naturf. Gesellschaft, Bd. 8, H. 2. Dorpat 1888.
RYDBERG, P. A., Scylla or Charybdis. - Proceed. of the Internat. Congr.
of Plant Sciences, Ithaca, New York, 1926, Vol. II. Menasha, Wise.,
1929.
SAMUELSSON, G., Zur Kenntnis der Schweizer Flora. - Vierteljahrsschrift
d. N aturforsch. Gesellsch. in Zurich, 67. Zurich 1922.
SARASIN, F. u. P., Die Landmollusken von Celebes. - Materialien zur
Naturgeschichte der Insel Celebes, 2. Wiesbaderr 1899.
SCHMIDT, J., Race-Undersogelser. I: Zoarces viviparus L. og dens lokale
Racer. - Meddelelser fra Carlsberg Laboratoriet, 13: 3. Kobenhavn
1917.
SCHOTT, P., Rassen der gemeinen Kiefer. - Forstl. Centralblatt 1907.
SEMENOV-TIAN-SHANSKY, A., Die taxonomischen Grenzen der Art und
ihrer Unterabteilungen. - Berlin 1910.
SEYBOLD, A., Untersuchungen tiber die Formgestaltung der Blatter der
Angiospermen. I. Die homologen Konvergenzreihen der Blatter und
allgemeine, kritische Bemerkungen tiber das Gestaltproblem. - Bibliotheca Genetica, Bd. XII. Leipzig 1927.
SEIDLITZ, G. VON, Die Darwirr'sche Theorie. - Dorpat 1871.
SHULL, G. H., Genotypes, Biotypes, Pure Lines and Clones,
Science, 35. New York 1912 (a).
- l ) - , Phenotype and Clone. Ibid. 35. 1912 (b).
-')-, The species Concept from the Point of View of a Geneticist. American Journ. of Bot. 10. Lancaster, Pa., 1923.
-')-, Species Hybridizations among Old and New Species of Shepherd's
, .Purse. Proceed. intern. congr. plant sciences, Ithaca, New York
1926. I, pp. 837-888. Menasha, Wisconsin 1929 (a).
- l ) - , Significance of Taxonomic Units and their Natural Basis.
Point
of View of Genetics. - Ibid., Aug. 16-23, 1926, Vol. II. 1929 (b).
SIRKS, M. J., and BIJHOUWER, J., Onderzoekingen over de eenheid der
Linneaansche Soort Chrysanthemum Leucanthemum L. - Genetica, 1.
's-Gravenhage 1919.

~25

STERNER, R., The Continental Element in the Flora of South Sweden. . Geografiska Annaler, 1922. Stockholm 1922.
~TOJANOFF, N., and STEFANOFF, B., Beitrag zur Kenntnis der Pliozanflora der Ebene von Sofia. - Zeitschr. d. bulgar. geol. Ges., 2. Sofia
1929.
SUMNER, F. B., Genetic Studies of Several Geographic Races of California
Deer-Mice. - The American Naturalist, 49. New York 1915.
- J ) - , The Role of Isolation in the Formation of a Narrowly Localized
Race of Deer-Mice (Peromyscus). - Ibid., 51. 1917.
- 1 ) - , Continuous and Discontinuous Variations and their Inheritance in
Peromyscus. - Ibid., 52. 1918.
-.)--, Geographical Variation and Mendelian Inheritance. - The Journal
of Experimental Zoology, 30. Philadelphia 1920.
- 1 ) - - , Some Facts Relevant to a Discussion of the Origin and Inheritance
of Specific Characters. - The American Naturalist, 57. New York
1923.
-,)-, The Stability of Subspecific Characters under Changed Conditions
of Environment. Ibid., 58. 1924.
SYLVEN, N., De svenska skogstraden. I. Barrtraden. - Stockholm 1916 (a).
-----:-, Den nordsvenska tallen. - Skogsvardsforeningens Tidskrift 1916
(Meddel. fro Statens Skogsforsoksanstalt, 13-14). Stockholm 1916 (b).
TEDIN, 0., Vererbung, Variation und Systematik in der Gattung Camelina.
- Hereditas, 6. Lund 1925.
TOWER, W. L., Darwinism. An Analysis by Observation and Experiment.
A Digest and Preliminary Statement of Result. - Genetica, 4. 's-Gravenhage 1922.
TROLL, W., Organisation und Gestalt im Bereich der Blirte. - Berlin 1928.
TURESSON, G., The Genotypical Response of the Plant Species to the Habitat. ~ Hereditas, 3. Lund 1922.
--, The Scope and Import of Genecology. - Ibid., 4. 1923.
-.)--, The Plant Species in Relation to Habitat and Climate. - Ibid., 6.
1925 (a).
-----:-, Studies in the Genus Atriplex. I. - Lunds Universitets Arsskrift.
N. F. Avd. 2, Bd. 21, Nr. 4. Lund 1925 (b).
->)--, Die Betrachtung der Rassenokologie fur die Systematik und Geographie der Pflanzen. - Beihefte zu Fedde's Repertorium, 41. . BerlinDahlem 1926 (a).
->)--, Studien tiber Festuca ovina L.
I. Normalgeschlechtliche, halbund ganzvivipare Typen nordischer Herkunft. - Hereditas, 8. Lund
1926 (b).
--, Experimentell eller beskrivande vaxtsystematik, - Lund 1926 (c).
->)--, Contributions to the Genecology of Glacial Relics. Hereditas, 9.
Lund 1927.
--, ZurNatur und Begrenzung der Arteinheiten. - Ibid., 12. 1929 (a).
- - , Ecotypical Selection ill' Siberian Dactylis glomerata L. Ibld.,
12. 1929 (b).
TXCKHOLM, G., Zytologische Studien tiber die Gattung Rosa. - Acta
Horti Bergiani, Bd. 7, N:o 3. Upsala 1922.

426
VAVILOV, N. I., The Law of Homologous Series in Variation. - Journal
of Genetics, 12. Cambridge 1922.
VIERHAPPER, F., Ueber echten und falschen Vikarismus. - Osterr. Bot.
Zeitschr., 68. Wien 1919.
WAGNER, M., Die Darwin'sche Theorie und das Migrationsgesetz der
Organismen. - Leipzig 1868; (English Translation: The Darwinian
Theory and the Law of Migration. London 1873.)
-'>---, Die Entstehung der Arten durch raurnliche Sonderung. Gesammelte
Aufsatze, - Basel 1889.
WALLACE, A. R., Darwinism. 2d edition. - London 1889.
'VEBBER, H .. J., New Horticultural and Agricultural Terms. -r-r- Science, 18.
New York 1903.
WEISMANN, A., Ober die Berechtigung der Darwinschen Theorie. - Leipzig 1868.
-'>---" Uber den Einfluss der Isolierung auf die Artbildung. - Leipzig
1872.
-'}-, Die Bedeutung der sexuellen Fortpflanzung fUr die Selektions-:
Theorie. - Jena 1886.
-'>---, Amphimixis oder: Die Vermischung der Individuen. - Jena 189!.
-'}-, Vortrage tiber Deszendenztheorie.3:e Aufl: - Jena 1913.
WETTSTEIN, R. v., Die europaischen Arten der Gattung Gentiana aus der
Sekt, Endotricha und ihr entwicklungsgeschichtlicher Zusammenhang.
- Denkschriften der math.-naturw. Classe d. k. Akad. d. Wissensch.,
64. Wi en 1896 (a).
-l>---, Monographie der Gattung Euphrasia. - Arbeiten d. bot. Inst.
d. k. k. deutsch. Univ. in Prag, 9. Leipzig 1896 (b).
-'}-, Grundzlige der ,geographisch-morphologischen Methode der Pflanzensystematik. - Jena 1898.
WIESNER, J., Biologie der Pflanzen. 3:e Aufl. - Wien 1902.
WILLIS, J. C" Some Evidence against the Theory of - - - Natural
Selection of Infinitesimal Variations. - Ann. R. Bot. Gard. Peradenyia,
4. Peradenyia 1907.
C-'>---, Age and Area. A Study in Geographical Distribution and Origin
of Species. - Cambridge 1922.
-'>---, The Origin of Species by Large, rather than by Gradual, Change,
and by-Guppy's Method of Differentiation. - Ann. of Bot., 37. London
1923.
WINGE, 0., Studier over Planterigets Chromosomtal og Chromosomernes
Betydning. - Meddelelser fra Carlsberg Laboratoriet. 13: 2. Kobenhavn 1917.
->1---, Das Problem der Jordan-Rosen'schen Erophila-Kleinarten. Cohn's
Beitrage zur Biologie der Pflanzen, 14: 3.' Breslau 1926 (a).
-'}-, Artkrydsningsproblemer i Planteriget. - Beretning -om Nordiske
Jordbrugsforskeres Forenings Tredie Kongres Oslo 1926. Nordisk
.Iordbrugsforskning 1926, 4-7 hefte. Kobenhavn 1926 (b).
-'>---, Arvelighedslrere paa experimentelt og cytologisk grundlag. Kobenhavn 1928.

427

CONTENTS.
Pages

Introduction. . . . . . . . . . " . . .
The taxonomical units as concrete populations
The individual
'
The clone . .
The pure line
The biotype.
The form . .
The variety .
U. The subspecies
10. The species . .
1. Asexual species
2. Sexual species.
I. Species separated from each other by complete genetical
isolation
.
II. Species connected with each other by transitional hybridpopulations . . . . . . . . . . . . . . . . . . . . . . . . .
1. The transitional hybrid-population consists of merely sterile individuals . . . . . . . . . . . . . . . . . . . . .
2. The transitional hybrid-population consists of more or
less fertile individuals. . . . . . .
Syngameon, commiscuum, and convivium .
Practical methods of species-taxonomy
11, The hybrid . . . . . . . . . .'. .
12. The units superior to the species.
Literature cited . . . . . . . . . . . .
1.
2.
3.
4.
5.
6.
7.
8.

333
336
337
338

338
340

342
348
354
357
362
362
363

365
365
367

386
387
390
392

414

,Postscript.
During the printing of this paper there has appeared a new
paper of N. HERlBERT NILSSON (Sind die mutierenden reinen Linien
auch rein?, Hereditas, Bd. XIV, H. 1, Lund 1930), containing
such importants results and views that I cannot refrain from
adding a few words about them. By extensive cultures HERlBEHT
NILSSON has found that different individuals of a certain so-called
pure line of oats produce pure lines differing considerably in
their production of rnutations, some of them giving a much
higher percentage of mutations than others. From this result
HERlBERT NILSSON draws the conclusion that the original pure
line did not contain only homozygous individuals, thus being no
real pure line at all, and that the rnutations appearing in it

428
(and in analogous false pure lines) are no real mutations but
simply some sort of segregants caused by the heterozygosity.
This certainly does not speak in favour of the current theory of
the rapid production of pure (i .. e. homozygotic) lines by autogamous ' reproduction (cornp. above p. 339). On the contrary,' it
gives a strong support to' LOTSY"S 'recent doubting of the existence
of a ny' absolutelyhomosygotic biotype :i~ nature even in pure
lines (comp, .above p., 341), . This implies. that the existence .of
real, pure. lines in .nature may be a mere illusion, and that the
taxonomical 'importance' of . the 'pure' line-concept thus. would
be still 'smaller than: postulated above (p. 339):

/\