It should not be thought that the differentiation of a species-population into hereditary habitat types is a phenomenon peculiar to the species discussed above. The cultivations hitherto made of other species, including for instance Rumex acetosa L., Solidago virgaurea L., Artemisia campestris L., Campanula rotundifolia L., Ranunculus acer L., indicate precisely the same behaviour. The same will very likely be found to hold true for the majority of common plant species. It is in fact to be assumed that the rarity of certain species is in great measure due to a decreased power of genotypical response to habitat differences, climatic and edaphic, within their area of distribution. Thus, as a result of genotypical responses of the species-population to different habitats, isolated units are formed within the species much in the same way as contemplated by JORDAN (1905) and HAGEDOORN (1921). However, to speak of such units as species, as is done by these writers, is largely to strip the ordinary species, as found in nature, of one of its most characteristic qualities, viz, the ability to respond genotypically to a wide range of different habitats with such units or habitat types, representing various combinations of Mendelian factors. The further discussion of this subject leads to a consideration of the species and the variety as ecological units. The concept of the species has undergone various changes since the time of LINNAEUS, Darwinism, as also Lamarckism before it, by the nature of the theory led to the conception of the species as mere conventions, set up in order to facilitate a classification of ever-changing and therefore not clearly defined groups. The genetical analyses of Linnean species have again brought about a change; the genotype has now become the real unit, while the collective species is still to a large extent considered a conventional conception. This general notion is also reflected in the species-concepts recently propounded (LOTSY, 1916; HAGEDOORN, 1921). Whether this is a legitimate procedure, or whether it does not at the same time reflect the failure on the part of these investigators rightly to appreciate the ecological side of the species problem may perhaps be gathered from the following. It is well known that during the last two decades great progress has been made with regard to the experimental study of the species problem from the genetical point of view. Not only has Mendelism gone far to show that species follow the same laws as varieties with regard to segregation and combination; it has also been able to demonstrate and to a certain extent copy Nature's own course in the building up of new species. This has been admirably shown by LOTSY (1916) in the well known case of Antirrhinum rhinanthoides, produced from a cross between A. glutinosum and A. majus and so different from its parents that a trained botanist would refer it rather to the genus Rhinanthus than to Antirrhinum. It is constant in certain characters but varies in other in the same way as the Linnean species. The extravagant types produced by HERIBERT-NILSSON (1918) from various Salix crosses belong to the same category of facts. All of them (Salix amerinoides, S. pendulifolia, S. monandra) demonstrate in a striking way the process by which new and morphologically very remarkable organisms arise.

Thus, while the belief that the Linnean species of the present genetically represent complicated products of recombined Mendelian factors, or genotype compounds, has been strengthened, few would maintain that the problems connected with the formation of the Linnean species are exhausted by this demonstration. Most of these species are, as every earnest inquirer will find, in their natural areas of distribution rather circumscribed products, which do not live in any extensive connubium with congeners of other species. The bridgeless gaps found between species of the same genus, the final moulding of the Linnean species, remain then to be explained. The Darwinian idea of selective processes at present offers to most minds a plausible explanation of the differentiation of Linnean species. Although very little is known with regard to the actual play of these selective processes, certain facts likely to demonstrate the complex nature of selection have been brought to light. KOLREUTER (1761) showed that a species pollinated simultaneously with its own pollen and pollen from another species breeds true to type, in spite of the fact that it otherwise gives hybrids when crossed with that species. That the native pollen is favoured as compared with foreign has been shown by HERIBERT-NILSSON (1920) in the case of Oenothera Lamarckiana. He found that pollen tubes of O. gigas grew more slowly in the styles of O. Lamarckiana than the O. Lamarckiana's own pollen tubes. The terms elimination, certation, prohibition, and substitution discussed by HERIBERT-NILSSON refer to phenomena which give rise to aberrant types of segregation. The importance of such gametic and zygotic complications has been discussed more recently by NILSSON-EHLE (1921). They are all particularly well calculated to throw light upon selective processes of great weight. To whatever extent this pre-natal selection may limit the output of new organisms, hybrids between already existing species would no doubt be more numerous and more widely distributed in nature were it not for the controlling effect of living and non-living factors of the outer world. Various disturbances involving different organs are frequently seen in hybrids and in artificial species, and this fact does not support the idea that such organisms are able to hold their own with nature. We are thus forced to the conclusion that the present-day species represent the necessary outcome of the complex processes of selection in this epoch of the earth's history (cp. HERIBERT-NILSSON, 1918). As a natural consequence we are led to the inference that a change in the non-living world brings about a corresponding change in the living, inducing a recombination of Mendelian factors now distributed in organisms, and resulting in the formation of new genotype compounds or species (=evolution).

The species problem is thus seen to be in a large measure an ecological problem. As such it has hitherto remained almost unattacked from an experimental point of view. While at present the purely genetical side of the problem is fairly well understood, we have only to a slight extent been able to arrive at an understanding of the ecological questions involved. So much appears certain, however, that the Linnean species are units of the same importance ecologically as their constituent elements are genetically. In view of the necessity of keeping the distinction between ecological and purely genelical units in mind, the term ecospecies has been proposed (TURESSON, 1922) to cover the Linnean species or genotype compounds as they are realized in nature.

Fig. 78.

The ecological sub-units of the ecospecies distinguished in the first part of this work have now to be briefly considered. They have been spoken of as types or habitat types. These terms for several reasons, and especially because of their vagueness, do not seem appropriate. The term variety might have been employed.

Fig. 79.

It is obvious, however, that this term gives an inadequate conception of the result of the genotypical reaction of an ecospecies to a specific habitat. It is rather arbitrarely used to denominate different kinds of hereditary deviations from a valid or supposed type, and cannot therefore be advantageously employed in any ecological sense. The term ecotype seems to be an adequate expression, and is therefore proposed for the ecological unit, to cover the product arising as a result of the genotypical response of an ecospecies to a particular habitat (cp. TURESSON, 1922).

The importance from an ecological point of view of the differing genotypical reactions of the ecospecies, when distributed over a continuous area comprising different types of localities, seems to have been sufficiently demonstrated at this stage. We have been able to show, by the help of cultures made, and by the differentiation of dune-ecotypes, sea cliff-ecotypes etc. that one and the same ecospecies succeeds in populating widely different habitats. It should not be argued, however, that the differing phenotypical reactions may not also be of great moment from an ecological point of view. This is all the more the case when it is found that a population in an extreme habitat responds as a whole with a reaction-type which is suggestive of a specialized ecotype, although it may be found upon culturing to be wholly or partly due to the modificatory effect of the particular habitat factors. The cases of the shade forms of Lysimachia vulgaris and Solanum Dulcamara on Hallands Väderö, and the Centaurea and Succisa dwarfs, analyzed above, illustrate this mode of behaviour. The reaction-types of the ecotypes called forth by the modificatory influences of extreme habitat factors may be termed ecophenes.

It becomes necessary briefly to consider the genetical analogues of these units. The term genospecies has been proposed (cp. TURESSON 1922) to embody the facts of the genotypical construcion of the ecospecies. The term, however, is properly to be employed for the genetical analogue of the coenospecies. The genotypes (JOHANNSEN, 1909) are, further, the Mendelian sub-units of the genospecies, as the ecotypes are the ecological sub-units of the ecospecies. The various reaction-types of one and the same genotype becomes then the phenotype (JOHANNSEN, 1909). The different units distinguished thus group themselfes in the following series.

It should be said in conclusion that the study of the species along the lines developed in the present work is intended to furnish a necessary complement to the Mendelian study of the species problem. The impblem The importance of this line of research for the understanding of bio-geographical and bio-sociological questions is also evident.

1. Material of a number of Linnean species, brought under culture, has revealed the presence of a considerable hereditary variation within each species.
2. The hereditary variations of the species have been found in nature to be grouped into different types confined to definite habitats. Certain characteristics have been found to be common to all or to the majority of the individuals of each type, while other characteristics vary.
3. The following facts bring out particularly well the intimate relation of these habitat types to the habitat factors:
   1. The characteristics of the types are apparently such that they specialize the types for certain habitats.
   2. With regard to Hieracium umbellatum a close correspondence is found between the alternation of habitat and habitat type.
   3. Cases have been found where a local differentiation of the habitat type from the species-population has in all probability taken place.
   4. There is a morphological parallelism between the modification and the hereditary variation.
4. The evidence at hand does not support the idea that the habitat types have originated through sporadic variations preserved by chance isolation. On the contrary the facts go to show that the habitat type represents the genotypical response of the species population to a definite habitat.
5. The Linnean species represents an ecological unit of great importance. It is specified as a coenospecies (defined above), narrowed down to the ecological combination-limit. A genotype compound of this order is here termed an ecospecies.
6. The term ecotype is used as an ecological sub-unit to cover the product arising as a result of the genotypical response of an ecospecies to a particular habitat.
7. The term ecophene is proposed to cover each of the reaction-types of the ecotypes arising through the modificatory influences of the combinations of extreme habitat factors given in nature.
8. The genetical analogues of these units are briefly considered, and the concepts of the genospecies and the genophene, defined above, are introduced.