S U C C E S S I V E A P P R O X I M A T I O N I N DESCRIPTIVE ECOLOGY
51
useful data for a natural classification of plant communities are descriptions which include as much information as possible about all
attributes of the community. Nevertheless, any description of a “uniform” community which includes the minimum acceptable number of
facts should be usable for this purpose; although the greater the amount
of information, the more valuable the description. There are difficulties
introduced by the comparison of lists obtained by a variety of techniques, but these are not insuperable. The task of describing and codifying plant communities is so vast that one should be prepared t o accept
all accurate data of whatever provenance.
Whether he is using the descriptive method as a research tool or
merely t o collect the raw material for classification, whether he is using
strictly objective methods or subjective, the ecologist must exercise his
judgement in the choice of features t o be observed, features t o be rejected and the scales of measurement t o be used. All his descriptions,
however detailed, are abstractions from the data; all his results and
hypotheses are approximations. This should not be forgotten. Objective
methods have the advantage that they can be repeated and checked;
they may be used t o examine assumptions about the nature of the community (Cain and de Oliviera Castro, 1959), and they reduce certain
kinds of bias. The objections t o them are less frequently realized.
There is a brief review of the problem in Poore (1955b) in which the
conclusion is drawn “that current statistical methods are inappropriate
for [the description of stands for classification] and that the plant
sociologist should have recourse t o the most accurate methods of
estimation available to him. It should of course be realized that the
results he obtains will be suitable only for qualitative comparison, and
that any more rigid treatment is illegitimate. The proper province of
plant sociological studies should be t o describe vegetation and t o discover and define problems for solution by more exact methods; in
addition, they will often indicate what lines of future research will
prove most fruitful”. I see no reason t o modify this conclusion substantially, but I would not exclude data collected by statistical methods
provided that, in other respects, these measure up to the standards of
information required.
There are a number of descriptions of the range of information which
is normally required in order that the “method of successive approximation” may be successfully applied. (Braun-Blanquet, 1951 ; Poore,
1955a,b ; Ellenberg, 1956; and, for a very comprehensive treatment,
Emberger, 1957.) Success will vary according to the extent t o which
these requirements are met.
Much space has been devoted in the literature to the inaccuracy of
subjective estimates of cover (e.g. Hope-Simpson, 1940 ; Smith, 1944)
51
useful data for a natural classification of plant communities are descriptions which include as much information as possible about all
attributes of the community. Nevertheless, any description of a “uniform” community which includes the minimum acceptable number of
facts should be usable for this purpose; although the greater the amount
of information, the more valuable the description. There are difficulties
introduced by the comparison of lists obtained by a variety of techniques, but these are not insuperable. The task of describing and codifying plant communities is so vast that one should be prepared t o accept
all accurate data of whatever provenance.
Whether he is using the descriptive method as a research tool or
merely t o collect the raw material for classification, whether he is using
strictly objective methods or subjective, the ecologist must exercise his
judgement in the choice of features t o be observed, features t o be rejected and the scales of measurement t o be used. All his descriptions,
however detailed, are abstractions from the data; all his results and
hypotheses are approximations. This should not be forgotten. Objective
methods have the advantage that they can be repeated and checked;
they may be used t o examine assumptions about the nature of the community (Cain and de Oliviera Castro, 1959), and they reduce certain
kinds of bias. The objections t o them are less frequently realized.
There is a brief review of the problem in Poore (1955b) in which the
conclusion is drawn “that current statistical methods are inappropriate
for [the description of stands for classification] and that the plant
sociologist should have recourse t o the most accurate methods of
estimation available to him. It should of course be realized that the
results he obtains will be suitable only for qualitative comparison, and
that any more rigid treatment is illegitimate. The proper province of
plant sociological studies should be t o describe vegetation and t o discover and define problems for solution by more exact methods; in
addition, they will often indicate what lines of future research will
prove most fruitful”. I see no reason t o modify this conclusion substantially, but I would not exclude data collected by statistical methods
provided that, in other respects, these measure up to the standards of
information required.
There are a number of descriptions of the range of information which
is normally required in order that the “method of successive approximation” may be successfully applied. (Braun-Blanquet, 1951 ; Poore,
1955a,b ; Ellenberg, 1956; and, for a very comprehensive treatment,
Emberger, 1957.) Success will vary according to the extent t o which
these requirements are met.
Much space has been devoted in the literature to the inaccuracy of
subjective estimates of cover (e.g. Hope-Simpson, 1940 ; Smith, 1944)