SUCCESSIVE A P P R O X I M A T I O N I N DESCRIPTIVE ECOLOGY
39
builds up parallel classifications of vegetation types and environmental
situations and becomes able t o relate the two. As his experience grows
he will be able t o distinguish more finely and to draw more detailed
inferences. This is the “method of successive approximation”. The
criterion which must be used to judge the validity of the inferences made
is that they should be eonsishnt with all the facts as far as they are
known. No analysis can expect t o be the ultimate truth, and all hypotheses are flexible and may be modified if new and contradictory evidence
comes t o light.
In this way a framework of points of reference is built up. Unexplained relationships, of which there are likely t o be many, due, for
example, t o vagaries of establishment, unknown historical factors and
other “accidents”, may be noted and reserved for future attention. The
differences that are significant will become apparent and can be subjected t o more detailed and experimental examination.
The method of successive approximation would appear t o be the
most economical way of obtaining a comprehensive understanding of
vegetational variation. It is par excellence a reconnaissance method but
is not limited t o reconnaissance. When the main framework is determined, more detailed investigations either of the community or of the
habitat will reveal finer relationships. This, not earlier, is the proper
stage for the entry of more strictly quantitative methods; for only at
this stage will they produce results which are commensurate t o the work
they involve.
If we briefly compare alternative approaches we can see what advantages the method of successive approximation offers.
The very detailed examination of one area or situation, though frequently very valuable, may easily prove t o be time partially wasted,
because the situation investigated is unrepresentative; or it may give
negative results because, through inexperience, the wrong features of
the environment have been observed.
Attempts have been made recently (Goodall, 1953ab, 19548; Hughes
and Lindley, 1955) t o put extensive primary survey on a quantitative
basis. These workers followed detailed quantitative investigation of the
vegetation and the habitat by factor analysis or ordination. This involves much more computational work and may, in the end, be less
revealing because only two or three factor complexes can be abstracted
and many of the incidental variations the causes of which are plain
in a detailed field investigation do not become apparent.
The detailed comparison between these different research methods
w i l l be taken up later. It is a cardinal rule, however, that the methods
used should be commensurate with the scale of the problem and with
the results that are t o be obtained.
39
builds up parallel classifications of vegetation types and environmental
situations and becomes able t o relate the two. As his experience grows
he will be able t o distinguish more finely and to draw more detailed
inferences. This is the “method of successive approximation”. The
criterion which must be used to judge the validity of the inferences made
is that they should be eonsishnt with all the facts as far as they are
known. No analysis can expect t o be the ultimate truth, and all hypotheses are flexible and may be modified if new and contradictory evidence
comes t o light.
In this way a framework of points of reference is built up. Unexplained relationships, of which there are likely t o be many, due, for
example, t o vagaries of establishment, unknown historical factors and
other “accidents”, may be noted and reserved for future attention. The
differences that are significant will become apparent and can be subjected t o more detailed and experimental examination.
The method of successive approximation would appear t o be the
most economical way of obtaining a comprehensive understanding of
vegetational variation. It is par excellence a reconnaissance method but
is not limited t o reconnaissance. When the main framework is determined, more detailed investigations either of the community or of the
habitat will reveal finer relationships. This, not earlier, is the proper
stage for the entry of more strictly quantitative methods; for only at
this stage will they produce results which are commensurate t o the work
they involve.
If we briefly compare alternative approaches we can see what advantages the method of successive approximation offers.
The very detailed examination of one area or situation, though frequently very valuable, may easily prove t o be time partially wasted,
because the situation investigated is unrepresentative; or it may give
negative results because, through inexperience, the wrong features of
the environment have been observed.
Attempts have been made recently (Goodall, 1953ab, 19548; Hughes
and Lindley, 1955) t o put extensive primary survey on a quantitative
basis. These workers followed detailed quantitative investigation of the
vegetation and the habitat by factor analysis or ordination. This involves much more computational work and may, in the end, be less
revealing because only two or three factor complexes can be abstracted
and many of the incidental variations the causes of which are plain
in a detailed field investigation do not become apparent.
The detailed comparison between these different research methods
w i l l be taken up later. It is a cardinal rule, however, that the methods
used should be commensurate with the scale of the problem and with
the results that are t o be obtained.