64
J . M. LAMBERT AND M. B . DALE
phytosociological studies, the last of them is especially relevant. The
more complex a problem, the more difficult it may be to disentangle
assumptions, concepts, postulates and the like from the situation
actually requiring statistical examination. This is not to imply that a
mixture of concepts may not occasionally be used when speed of operation is the limiting factor; but exceptional care is then required in
interpreting the results and the pitfalls must be f d y realized (wide, for
example, the discussion on the use of frequency data in Greig-Smith,
1964, pp. 9-19). Similarly, the more elaborate the method of analysis
itself, the more it is necessary to understand the underlying assumptions
and logical consequences of each mathematical step. The use of statistics
in phytosociological work has thus an additional hidden function in
that it forces the investigator to clarify his mind as to the precise nature
of his current problem: the demands of exacting methods can prove a
powerful antidote to any tendency to loose or superficial thinking.
11. THE NATURE OF PHYTOSOCIOLOGICAL DATA
A. THE NATURE O F VEGETATION
Since phytosociology is concerned with vegetational phenomena,
the concept of vegetation itself must first be examined in a statistical
context. Briefly, vegetation may be defined as plant material growing
on the earth’s crust. There are thus three elements to be distinguished:
(1) the plants themselves; (2) the sites (defined by position in space) in
which they occur; and (3) the environmental features associated with
these sites. This threefold plant/site/habitat distinction seems only
vaguely recognized in most vegetation studies ; it nevertheless provides
a useful mechanism for differentiating between strictly phytosociological phenomena (i.e. plant/plant relationships in a number of different
sites) and wider ecological phenomena concerning the whole of the
vegetational system.
It may at first seem inappropriate t o include environmental features
in any general concept of vegetation. However, we rarely think of
vegetation as a set of plants completely isolated from their normal
habitat. For instance, we do not envisage saltmarsh vegetation
divorced from a saline soil and represented merely by a collection of
halophytes laid out on a laboratory bench, or even growing in a garden:
although the plants occupy spatial positions and a plant/site relationship thereby exists, a further site/habitat relationship is necessary
before the system fulfils our usual concept of such vegetation.
For statistical purposes, the sites may be regarded as a number of
individuals with certain attributes or properties potentially in common.
The particular plants attributes which concern us here are, first, the
J . M. LAMBERT AND M. B . DALE
phytosociological studies, the last of them is especially relevant. The
more complex a problem, the more difficult it may be to disentangle
assumptions, concepts, postulates and the like from the situation
actually requiring statistical examination. This is not to imply that a
mixture of concepts may not occasionally be used when speed of operation is the limiting factor; but exceptional care is then required in
interpreting the results and the pitfalls must be f d y realized (wide, for
example, the discussion on the use of frequency data in Greig-Smith,
1964, pp. 9-19). Similarly, the more elaborate the method of analysis
itself, the more it is necessary to understand the underlying assumptions
and logical consequences of each mathematical step. The use of statistics
in phytosociological work has thus an additional hidden function in
that it forces the investigator to clarify his mind as to the precise nature
of his current problem: the demands of exacting methods can prove a
powerful antidote to any tendency to loose or superficial thinking.
11. THE NATURE OF PHYTOSOCIOLOGICAL DATA
A. THE NATURE O F VEGETATION
Since phytosociology is concerned with vegetational phenomena,
the concept of vegetation itself must first be examined in a statistical
context. Briefly, vegetation may be defined as plant material growing
on the earth’s crust. There are thus three elements to be distinguished:
(1) the plants themselves; (2) the sites (defined by position in space) in
which they occur; and (3) the environmental features associated with
these sites. This threefold plant/site/habitat distinction seems only
vaguely recognized in most vegetation studies ; it nevertheless provides
a useful mechanism for differentiating between strictly phytosociological phenomena (i.e. plant/plant relationships in a number of different
sites) and wider ecological phenomena concerning the whole of the
vegetational system.
It may at first seem inappropriate t o include environmental features
in any general concept of vegetation. However, we rarely think of
vegetation as a set of plants completely isolated from their normal
habitat. For instance, we do not envisage saltmarsh vegetation
divorced from a saline soil and represented merely by a collection of
halophytes laid out on a laboratory bench, or even growing in a garden:
although the plants occupy spatial positions and a plant/site relationship thereby exists, a further site/habitat relationship is necessary
before the system fulfils our usual concept of such vegetation.
For statistical purposes, the sites may be regarded as a number of
individuals with certain attributes or properties potentially in common.
The particular plants attributes which concern us here are, first, the