PLANT COMMUNITIES. 895 



certain greo;arious species thafc, where a plant of the one kind occurs, the presence of 

 its companiou species of the other kind may be inferred with certainty, and obviously 

 due regard must be paid to such constant companions in any description of par- 

 ticular communities. A further circumstance which must not be overlooked is that 

 up to a certain point the suViordinate species may replace one another. Thus, for 

 instance, that variety of carpet in which the Mat-grass (Nardus stricta) forms 

 the dominant species, includes as subordinate species both in the Alps and the 

 Carpathians Homogyne alpina, Hieracium alpinwm, Campanula Schev^hzen, &c. 

 Certain other subordinate species, e.g. Potentilla aurea, Hypochairis Helvetica and 

 Campanula harhata, which are almost invariably present in the community in 

 question when it grows in the Alps, are, however, replaced in the Eastern Carpathians 

 by Potentilla chrysocraspeda, Scorzonera rosea, and Campanula abietina. 



Special attention must be drawn to the fact that species which are gregarious in 

 one plant-community occur as a mere sprinkling in another. When this observation 

 is first made in Nature it is apt to lead one to suppose that it is a vain task to 

 attempt to arrange the ditferent communities in groups, and to define and describe 

 each separate kind. But more careful study soon con\'inces one that the pheno- 

 menon in question, far from interfering with the scientific treatment of the subject, 

 actually assists it, and that the history of plant-communities is elucidated thereby. 

 It has been repeatedly remarked that mud newly deposited by water, exposed soil, 

 and naked rock do not forthwith deck themselves in their permanent mantle of 

 vegetation, but that first of all small Algee, Lichens, Mosses, and various annual 

 Phanerogams establish themselves and prepare the way in the course of time for 

 other plants. This preparation, which was described in vol. i. pp. 257-268, consists 

 not only in mechanical changes in the ground, but also in the admixture of humus 

 derived from the parts of the first settlers which die oflf. The only species which 

 take possession of ground so modified are such as differ entirely from the first 

 colonists, and, curious to relate, the latter are gradually ousted by the new arrivals 

 and driven from the field. But the second settlement has no permanence either. 

 The quantity of humus arising from the death of plants increases from year to year, 

 the soil becomes too rich for the plants in possession, and the process of eviction is 

 now undergone by them at the hands of fresh species, which thrive on the rich soil, 

 and gradually possess themselves of it. At least three successive series of settlers 

 may thus be traced on every spot, and not infrequently the number is four or five. 

 Xow, if each of these groups corresponds to a particiliar community, which is as a 

 matter of fact the case, the phenomenon described must produce the same impression 

 as though the communities became transformed into one another in course of time. It 

 is therefore necessary to recognize the existence of the incipient and decadent stages 

 as well as that of predominance. In the incipient stage relics of the community 

 which previously occupied the same spot are still to be foinid, and in the stage of 

 decadence the fii-st pioneers of the community that is to succeed make their appear- 

 ance. When, for example, a carpet is invaded here and there by individual plants 

 belonging to species indicative of communities which pertain to some other type, 



