44
M. E. D. P O O R E
order t o classify or to argue by successive approximation it is necessary
t o group together things that are equivalent. It is no use including buttons or zip-fasteners in a classification of garments, although both may
be an essential part of certain articles of clothing. I n the same way communities of bark epiphytes or the moss flora of an erratic boulder
should not be classified with forests; nor should the chemical elements
be included in a classification of igneous rocks.
Phytosociological or ecological descriptions are usually made of
“uniform” communities. This is again an approximation. Absolute
uniformity is foreign to vegetation; but several scales of relative uniformity can be distinguished, the data from which should not be mixed
up in classification or inference. Much valuable statistical work has
been carried out to investigate pattern in relatively simply communities, and there is a growing understanding of the va.rious
kinds of pattern which can occur. (For references see Greig-Smith,
1957.)
If we consider an imaginary idealized habitat which is uniform in all
the physical factors of the environment in so far as they affect vascular
plants, various scales of pattern, or of relative uniformity, can be
distinguished.
(i) At the scale of soil micro-organisms or the exploring root tip of
a vascular plant the smallest volume of soil is heterogeneous; and
the least irregularities may provide micro-environments for the
establishment of seeds. If we wish to examine the micro-distribution
of bacteria or mites and to classify the communities in which they live,
the sampling unit will clearly have t o be much smaller than that required
for examining the distribution of vascular plants; so will the unit of
classification.
(ii) Next there is a pattern determined by the size of the individual
vascular plants in the community. I n simple communities this may
affect the distribution of other smaller plants and animals and of parasitic organisms. But in more complex communities, such as forest, the
distribution of individuals of the larger and more important species may
determine the distribution of the societies of the forest floor and various
societies of epiphytes. According t o available statistical studies, individuals of a species rarely occur at random but are more often distributed or occur in randomly distributed contagious groups. The
synusiae which are partially or totally dependent on individuals of larger
species also follow this pattern of distribution.
(iii) This kind of pattern is seen not only in space but in time. Watt
(1947) has described a cyclical pattern of micro-habitats caused by the
life cycle of the dominant, each phase of which has its own characteristic
associated flora; and this kind of pattern is now recognized as occurring
M. E. D. P O O R E
order t o classify or to argue by successive approximation it is necessary
t o group together things that are equivalent. It is no use including buttons or zip-fasteners in a classification of garments, although both may
be an essential part of certain articles of clothing. I n the same way communities of bark epiphytes or the moss flora of an erratic boulder
should not be classified with forests; nor should the chemical elements
be included in a classification of igneous rocks.
Phytosociological or ecological descriptions are usually made of
“uniform” communities. This is again an approximation. Absolute
uniformity is foreign to vegetation; but several scales of relative uniformity can be distinguished, the data from which should not be mixed
up in classification or inference. Much valuable statistical work has
been carried out to investigate pattern in relatively simply communities, and there is a growing understanding of the va.rious
kinds of pattern which can occur. (For references see Greig-Smith,
1957.)
If we consider an imaginary idealized habitat which is uniform in all
the physical factors of the environment in so far as they affect vascular
plants, various scales of pattern, or of relative uniformity, can be
distinguished.
(i) At the scale of soil micro-organisms or the exploring root tip of
a vascular plant the smallest volume of soil is heterogeneous; and
the least irregularities may provide micro-environments for the
establishment of seeds. If we wish to examine the micro-distribution
of bacteria or mites and to classify the communities in which they live,
the sampling unit will clearly have t o be much smaller than that required
for examining the distribution of vascular plants; so will the unit of
classification.
(ii) Next there is a pattern determined by the size of the individual
vascular plants in the community. I n simple communities this may
affect the distribution of other smaller plants and animals and of parasitic organisms. But in more complex communities, such as forest, the
distribution of individuals of the larger and more important species may
determine the distribution of the societies of the forest floor and various
societies of epiphytes. According t o available statistical studies, individuals of a species rarely occur at random but are more often distributed or occur in randomly distributed contagious groups. The
synusiae which are partially or totally dependent on individuals of larger
species also follow this pattern of distribution.
(iii) This kind of pattern is seen not only in space but in time. Watt
(1947) has described a cyclical pattern of micro-habitats caused by the
life cycle of the dominant, each phase of which has its own characteristic
associated flora; and this kind of pattern is now recognized as occurring