THE UNITS OF VEGETATION. 121 



even in Warming's summary statement of the formation. Instead of a single 

 fundamentum divisionis, Warming puts forward two tests of the formation, 

 namely, definite plant forms ('growth-forms') and definite characters of the 

 habitat. It is not clear, either from his definition or from his general treat- 

 ment of formations, what Warming precisely means by the term 'definite 

 growth-forms.' In any case, the definition is defective, as plant form is not 

 necessarily related to habitat: and therefore the two tests put forward in the 

 one definition will frequently yield contradictory results. Warming (p. 232) 

 insisted that a salt marsh characterized by suffruticose Salicornias 'must be 

 set apart from' salt marshes characterized by herbaceous Salicornias 'as a 

 separate formation' merely because the plant form in the two cases is different. 

 Such paradoxes occur throughout the whole of Warming's book; and indeed 

 this Janus-like 'formation' is inevitable if plant form is to be allowed to enter 

 into competition with habitat in the determination of formation. Warming's 

 view might find some justification if definite plant forms were invariably 

 related to definite habitats; but it is quite certain that this is not the case. 

 For example, on salt marshes in the south of England, it is no unusual thing to 

 find associations characterized (a) by herbaceous species of Salicornia, (6) by 

 suffruticose species, and (c) by a naixture of these. To place these associations 

 in separate 'formations,' however, simply because of the different nature of the 

 plant forms, is to reduce the study of formations to an absurdity." 



The foregoing criticism is as vaUd from the developmental viewpoint as 

 from that of the habitat. It brings out in clear rehef the fallacy of using a 

 single basis for the recognition of formations, as is the usual method in most 

 systems, notwithstanding statements to the contrary in defining the concept. 

 There is of course no real contradiction between habitat and physiognomy, in 

 spite of the fact that two or more life-forms may appear in the same habitat, 

 and that the same life-form may recur in widely different habitats. The error 

 lies in assuming that all species must make the same structural respones to 

 a habitat, and that the general character of the life-form necessarily indicates 

 its actual response. Nowhere in the field of ecology is there a more striking 

 confirmation of the fact that development is the sole clue to follow through 

 the maze of apparent and real, of superficial and basic relations between habi- 

 tat, floristic, and physiognomy. 



Moss (1910 : 39) commends Warming for adopting, in connection with the 

 division of the formation into associations, "a view which has forced itself 

 on the minds of nearly all close students of vegetation." This is the view of 

 Cowles (1899 : 111), in which he regarded the relation between the formation 

 and association as similar to that of the genus to the species. The wording 

 of Cowles's statement is as follows: "One might refer to particular sedge 

 swamp societies near Chicago, or to the sedge swamp formation as a whole; by 

 this application, formation becomes a term of generic value, plant society of 

 specific value." It is an open question whether the relation of particular 

 local associations to an actual floristic entity is not really intended. In any 

 event, it seems clear that there was no expressed intention of building up an 

 artificial concept like the genus by placing in it aU the swamp associations of 

 the entire globe. Yet this is a legitimate if not necessary assumption from 

 this comparison, and it is illogical to commend the acceptance of the principle 

 and to object to the application of it. In the meaning of formation as used by 

 those who regard it as a definite entity, the sole relation of the genus to the 



