38
M. E. D. POORE
they might be used for the solution of ecological problems by a process
of “successive approximation”, a series of inferences each closer t o the
truth than the preceding. This theme will be restated and elaborated in
this paper.
What is meant by the methods of successive approximation? Extended illustrations of the application of the method t o field problems
are given by Dahl (1956) and Poore (1955~); here we may look instead
at a generalized example.
Let us for a moment imagine that an ecologist is starting work in a
new and unknown area. What are his problems and how can he best
solve them? He is faced with a variable cover of vegetation, here forest,
there grassland, marshes, dwarf shrub communities and so on. On the
other hand the country has marked topographic relief, different soils
and a variable intensity of grazing. His first aim is broadly to relate the
variation of vegetation to that of the habitat, and then progressively t o
refine his knowledge so that he understands the causes underlying
smaller and smaller vegetational differences.
The variety is baffling t o one unfamiliar with the species and the
environment, and so the ecologist tries to orientate himself, to give
himself some point of reference. He chooses some community (A) and
describes both it and its environment in as much detail as he is able.
He chooses the community because he considers that it is typical of a
much wider sample; but, because he is inexperienced in this region, his
choice may be wrong. This does not matter for he will have plenty of
opportunity to modify or reject his choice as the investigation continues.
Having become familiar with community A, he moves t o community B
and describes it in the same way. This consists of different species and,
let us say, occur8 on a completely different soil. By comparing these
two (A and B) he makes the provisional inference that the difference in
vegetation between A and B is due t o a difference in soil. He realizes
that this may be en inaccurate or incomplete explanation, but i t is
consistent with his observations. After he has examined a number of
these stands and found in every case the same difference in habitat, his
inference becomes more sure and he can erect classes of communities and
classes of habitats and generalize that community type A is distinguished
from community type B by a certain consistent difference in species
content and that this is correlated with a difference in habitat. Nevertheless, as he Continues, he is likely to find inconsistencies. Theso should
be carefully examined, because knowledge grows by the explanation of
inconsistencies. He may find, for example, a community C in a habitat
which appears identical t o that of B; and by the presence of charcoal
he may be able to relate this to the recent occurrence of a fire. By the
continued examination of similarities and differences he gradually
M. E. D. POORE
they might be used for the solution of ecological problems by a process
of “successive approximation”, a series of inferences each closer t o the
truth than the preceding. This theme will be restated and elaborated in
this paper.
What is meant by the methods of successive approximation? Extended illustrations of the application of the method t o field problems
are given by Dahl (1956) and Poore (1955~); here we may look instead
at a generalized example.
Let us for a moment imagine that an ecologist is starting work in a
new and unknown area. What are his problems and how can he best
solve them? He is faced with a variable cover of vegetation, here forest,
there grassland, marshes, dwarf shrub communities and so on. On the
other hand the country has marked topographic relief, different soils
and a variable intensity of grazing. His first aim is broadly to relate the
variation of vegetation to that of the habitat, and then progressively t o
refine his knowledge so that he understands the causes underlying
smaller and smaller vegetational differences.
The variety is baffling t o one unfamiliar with the species and the
environment, and so the ecologist tries to orientate himself, to give
himself some point of reference. He chooses some community (A) and
describes both it and its environment in as much detail as he is able.
He chooses the community because he considers that it is typical of a
much wider sample; but, because he is inexperienced in this region, his
choice may be wrong. This does not matter for he will have plenty of
opportunity to modify or reject his choice as the investigation continues.
Having become familiar with community A, he moves t o community B
and describes it in the same way. This consists of different species and,
let us say, occur8 on a completely different soil. By comparing these
two (A and B) he makes the provisional inference that the difference in
vegetation between A and B is due t o a difference in soil. He realizes
that this may be en inaccurate or incomplete explanation, but i t is
consistent with his observations. After he has examined a number of
these stands and found in every case the same difference in habitat, his
inference becomes more sure and he can erect classes of communities and
classes of habitats and generalize that community type A is distinguished
from community type B by a certain consistent difference in species
content and that this is correlated with a difference in habitat. Nevertheless, as he Continues, he is likely to find inconsistencies. Theso should
be carefully examined, because knowledge grows by the explanation of
inconsistencies. He may find, for example, a community C in a habitat
which appears identical t o that of B; and by the presence of charcoal
he may be able to relate this to the recent occurrence of a fire. By the
continued examination of similarities and differences he gradually