“ The Oecology of Plants." 223 
that beech and oak may grow equally well on the same soil, and in 
many parts of Europe form natural mixed or alternating woods, the 
reply must be that every case must be taken on its own merits. The 
conditions may lead to such a result in one country and not in 
another. There may be a deciduous forest-formation dominated by 
alternative trees under one set of conditions, and separate pure- 
formations each dominated by a single species under another set. 
To the objection that this is to lay undue emphasis on minor 
differences and to obscure the wide relations, it must be replied that 
it is perfectly possible to bring together similar formations, such as 
the deciduous forests of cool temperate regions, in a unit which may 
be called a type of formation. It is simply a question of a suitable 
nomenclature. But it is of the first importance to separate 
communities which are really “ the expression of certain defined 
conditions of life ” whether they be characterised by markedly 
different growth-forms or not. Taking the the taxonomic analogy, 
we prefer to consider the formation as the species. The species, 
after all, is the fundamental unit in taxonomy, as the formation 
should be in vegetation. 
Of previous discussions on the concept of formation we consider 
Drude’s (Deutschlands Pflanzengeographie, Bd I., pp. 285-289), 
Schimper’s (Pflanzengeographie, p. 175) and Clements’ (Research 
Methods in Ecology, p. 292) as the most illuminating. These authors 
all lay fundamental stress on the habitat. The differentiating factor 
between adjacent formations may be edaphic or climatic, though as 
Professor Warming rightly insists, any given formation, considered 
by itself , is determined by a sum-total of factors both climatic and 
edaphic. In general, the different geological formations giving rise 
to different types of soil most frequently bear different plant- 
formations (though in the case of a “ country ” covered with a sheet 
of glacial debris this differentiating factor may be lost), while in 
mountainous countries different formations are determined by the 
differences of climate corresponding with altitudinal zones. 
We may, however, well consider whether the concept should 
not be extended in accordance with Moss’s usage (Vegetation of 
Somerset, pp. 12, 68). On one and the same type of soil and under 
the same climatic conditions we may find widely different types of 
plant community, e.g., forest, scrub and heathland, or grassland. 
These differences are very often no doubt brought about by the 
action of man in disforesting and pasturing, but it is not certain 
that this is so in all cases. A definite relation always exists 
between a given type of forest and the corresponding scrub, heath- 
land or grassland, and the latter types of vegetation tend, in a 
general way, and subject to arrest by various causes, to go back to 
the higher type. The different members of each series exist in a 
habitat which is fundamentally the same, because it is conditioned 
by the same climate and underlying rock, though it is modified by 
the relations of the types of vegetation to the soil. It is a question 
whether, in accordance with the fundamental importance of the 
habitat, we ought not to include all the members of such a series in 
the same formation, since they are indissolubly connected, and pass 
imperceptibly one into another not only in space, but in time. The 
same idea would hold with regard to the natural development of 
vegetation on bare soils. All the phases of the natural succession 
