48
M. E. D. F O O R E
irrespective of the relative position of the samples, the vegetation may
be considered homogeneous. Independence of variance and spacing
would need t o be demonstrated for each species present in order that
the whole vegetation might be regarded as homogeneous. The third
method proposed is based on the joint distribution of species. If the
quantities of two species in a sample area are correlated, then the
vegetation is heterogenous. This, according t o Goodall, implies that
the environmental factors affecting at least one of them are not uniform
over the area - which may depend on both species having the same
(or complementary) habitat preferences, or on one modifying the
environment for the other. But this depends on the size of sample and
the scale of “uniformity” required of the habitat.
These three methods all aim at establishing a degree of homogeneity
within a stand. But an alternative procedure would be t o assess the
homogeneity of so-called uniform areas of vegetation relative t o the
transition areas between them.
After examining the existing data he concludes that “where homogeneity has been tested on various scales it has not been found” and
suggests that this should cause heart searching among those who readily
assume on the evidence of a general visual impression that the vegetation at which they are looking is homogeneous.
R u t before relegating all classifications of vegetation based on t,lie
subjective choice and description of stands t o the dustbin, we should
examine carefully exactly what these results do show us and how far
they affect the actual methods by which a plant sociologist works.
The principal data on which these conclusions are based are those of
Goodall (195423) and Greig-Smith (1952). In a later review Greig-Smith
( 1957) discusses Goodall’s results and concludes that “the samples
appear to have been taken within an area which would be regarded on
most subjective criteria as a single community, and the investigations
refer rather t o pattern. . . than to differences between communities”.
The investigation of Greig-Smith refers t o a type of vegetation, namely
tropical forest, where even the phytosociologists are not agreed about
the possibility of applying the subjective criteria which they find so
useful in the temperate, arctic and Mediterranean regions (Mangenot,
1958; discussion). Clearly a more useful check on the validity of these
methods would be t o apply an extensive analysis for homogeneity in a
region where distinct claims have been made for phytosociological
methods. For it is possible, indeed likely, that there are kinds of vegetation where the methods advocated by Goodall are more appropriate,
but this does not necessarily invalidate results based on the subjective
delimitation of stands elsewhere.
It is true that “plant sociologists anxious t o proceed wit,h the work
M. E. D. F O O R E
irrespective of the relative position of the samples, the vegetation may
be considered homogeneous. Independence of variance and spacing
would need t o be demonstrated for each species present in order that
the whole vegetation might be regarded as homogeneous. The third
method proposed is based on the joint distribution of species. If the
quantities of two species in a sample area are correlated, then the
vegetation is heterogenous. This, according t o Goodall, implies that
the environmental factors affecting at least one of them are not uniform
over the area - which may depend on both species having the same
(or complementary) habitat preferences, or on one modifying the
environment for the other. But this depends on the size of sample and
the scale of “uniformity” required of the habitat.
These three methods all aim at establishing a degree of homogeneity
within a stand. But an alternative procedure would be t o assess the
homogeneity of so-called uniform areas of vegetation relative t o the
transition areas between them.
After examining the existing data he concludes that “where homogeneity has been tested on various scales it has not been found” and
suggests that this should cause heart searching among those who readily
assume on the evidence of a general visual impression that the vegetation at which they are looking is homogeneous.
R u t before relegating all classifications of vegetation based on t,lie
subjective choice and description of stands t o the dustbin, we should
examine carefully exactly what these results do show us and how far
they affect the actual methods by which a plant sociologist works.
The principal data on which these conclusions are based are those of
Goodall (195423) and Greig-Smith (1952). In a later review Greig-Smith
( 1957) discusses Goodall’s results and concludes that “the samples
appear to have been taken within an area which would be regarded on
most subjective criteria as a single community, and the investigations
refer rather t o pattern. . . than to differences between communities”.
The investigation of Greig-Smith refers t o a type of vegetation, namely
tropical forest, where even the phytosociologists are not agreed about
the possibility of applying the subjective criteria which they find so
useful in the temperate, arctic and Mediterranean regions (Mangenot,
1958; discussion). Clearly a more useful check on the validity of these
methods would be t o apply an extensive analysis for homogeneity in a
region where distinct claims have been made for phytosociological
methods. For it is possible, indeed likely, that there are kinds of vegetation where the methods advocated by Goodall are more appropriate,
but this does not necessarily invalidate results based on the subjective
delimitation of stands elsewhere.
It is true that “plant sociologists anxious t o proceed wit,h the work