This was quite in keeping with the usual pattern and trend of scientific progress as most branches of science in their early stages of development are oriented strongly toward description and classification, and the earlier naturalists, botanists and phytosociologists quite naturally sought a framework within which to describe and classify the components of vegetation.
A belief in the discrete community-unit hypothesis enabled them to identify, describe and classify the plant communities. The earlier ideas of classification of vegetation were clearly set forth in Weaver and Clements, Plant Ecology (1929) and came to be styled as “Clementsian.”
The Clementsian views greatly influenced the development of British and American ecology, especially during the period 1915-1935. Clements was a great believer in the dynamism of vegetation and even more so of the strong influence of the environment on the vegetation.
Many contemporary ecologists notably on the Continent, India and other countries, still believe in this view and continue to study and describe the vegetation as an assemblage of distinct communities.
During the last few decades, however, certain ecologists began to dissent from the then widely-believed idea of vegetation discontinuity. During the 1920s, Gleason advanced a novel idea, called the Individualistic Hypothesis (Gleason, 1917, 1926) and stated that “presumptive associations or areas of uniform structure vary in composition from place to place and that the degree of variation increases with distance, making logical classification impossible.”
He further held the view that association of species results from coincidence of historical events and environmental tolerances of the species. In his later work, Gleason (1939) described the vegetation unit association as a temporary, fluctuating phenomenon, dependent in its origin, structure and disappearance, on the selective action of the environment and on the surrounding vegetation.
Thus the association was thought not to have any similarity to an organism or species. Gleason believed that a plant association was merely a fortuitous juxtaposition of plant individuals.
The idea of continuity (as opposed to discontinuity) of vegetation was given a serious consideration by many American ecologists, especially John Curtis and his coworkers at the University of Wisconsin. Gradually, many ecologists became inclined to believe in the continuum concept of vegetation, first put forward by Curtis and Mcintosh, (1951), viz., that vegetation changes continuously and is not really differentiated into distinct communities (Mcintosh, 1967).
The phrase “continuum concept of vegetation” means that an uninterrupted series of elements form a sort of continuum, there being no sharp transitions or ecotones between communities, and that species composition changes gradually from place to place or time to time (Mcintosh, 1967). Ecologists began to interpret vegetation not as communities, but as a complex but mostly continuous population pattern.
This new concept made it possible to study vegetation through a wider and more comprehensive approach than merely as assemblage of discrete communities. The new concept also created a basis for establishing a framework for a better and more efficient study of plant autecology.
During the last two decades the continuum concept has received a widespread acceptance and recognition and it can certainly be regarded as one of the more significant advances in the development of ecology in recent years, even though some phytosociologists and classical ecologists continue to ignore it.
Many careful quantitative studies of vegetation, by the application of variety of analytic and synthetic techniques have furnished strong support for this continuum concept. As opposed to the use of classification in the study of discrete communities, a new technique, called ordination, has been evolved for relating continuously variable communities to one another.
“An ordination is an arrangement of communities, species, or environments in sequence which, it is hoped, will reveal maximum information about the relationships among them and which will also reveal such classes as may exist” (Mcintosh, (1967). The concept of ordination has been further elaborated and extended by Whittaker (1967, 1973) in his “gradient analysis” to describe ordination methods applicable to the study of gradients of species, communities or environments.
Stated simply, ordination is a process by which communities or their samples are arranged in order (ordinated) along gradients of environmental change, or alternatively, the samples are ordinated along compositional gradients, with communities most similar to each other being placed continuously on a scale which defines the range of similarity values among all such samples.
In classification, similar samples are combined in the same category but in ordination, the objective is to consider sample differences rather than similarities, so as to dispose the samples in a linear or multidimensional network that will reveal the relationships between the samples and their environments.
Whittaker (1969, 1972) concluded on the basis of gradient analysis data that, while species are clumped in their habitat tolerances, no two species are quite the same in this respect. According to him, diversity in plant communities is made possible by a partitioning of the habitat, and his emphasis is almost wholly on diversification in the habitat niche.
More recently, he (1975) has suggested that groups of species with strongly similar habitat niches are not a general feature of vegetation, but this suggestion seems not to be borne out by the actual observations of many ecologists who have worked on grasslands, weed communities, deciduous forests and certain other situations (see Grubb, 1977).
Although gradient analysis and its associated quantitative multivariate techniques for the study of ecological relationships between plants and their environment have been widely employed during the last 15 years, direct gradient analysis requires collateral observations of environmental variables, e.g., quantification of soil moisture gradients which are quite difficult to measure.
Some workers have adopted an indirect gradient analysis using what is called the principal components analysis (PCA); explicit environmental measurement is not an essential pre-requisite for this kind of analysis. But, as pointed out by Whittaker, such gradients are difficult to interpret environmentally. Further, it is not proper to apply this multivariate normal technique of PCA to data which are not normally distributed.
Strahler (1978) has proposed a new method of analyzing plant species- environment relationships. This method is called ‘Binary Discriminant Analysis (BDA)’. It requires only the simplest forms of vegetation and environmental data. It can be used to group species having similar environmental preferences, and permits the construction of vegetation gradients.
BDA can be used on vegetation data in which species are recorded as present or absent within quadrats or plots, etc. It is used for identifying binary variables and their common trends that are most important for discriminating between groups.
It is applicable to lists of plant species to reveal similar patterns of preference or avoidance among species responding significantly to multistate environmental characteristic such as soil or rock type. BDA is really a two- step process, involving the construction of a set of contingency tables followed by PCA of standardized residuals derived from these tables (Strahler, 1978).
The technique has been demonstrated to work well on ecological data to produce either a continuous ordination of species in multidimensional factor space (Q-mode BDA), or grouping of species responding similarly to a given environmental variable being studied (R-mode BDA).
In the Q-mode technique, orthogonal multidimensional factors representing uncorrelated floristic trends best separating the groups are involved. Plotting of many species on the multidimensional hyperspace often reveals how each species responds to the floristic trends studied.
In contrast, in R-mode BDA, groups of species showing similar responses to the environmental parameter can be identified. These groups can be interpreted as community components consisting of ecologically-similar species (Strahler, 1978).
From the above account it should not be inferred that the continuum concept and the discrete community hypothesis are necessarily mutually exclusive. In contrast to an erroneous belief, even continuously variable communities do not exclude the classification approach.
The main objection to the continuation of the classical theory lies in the fact that all too often the classification of (discrete) communities becomes an end in itself (Monk, 1968). According to Gimingham (1968), both ordination and classification will continue to contribute materially to the elucidation of the complexities of vegetation.
Those who believe in the continuum concept do not imply that in nature species are necessarily always distributed independently of each other and unrelated to environmental factors. Mcintosh (1967) emphasizes the fact that in a discontinuous environment the vegetation need not be continuous.
He also duly recognizes the existence of physiognomically different communities but stresses that the major problem of continuity lies within rather than between physiognomic vegetation types (Mcintosh, 1968). The continuum concept also does not exclude the interactions among species within plant communities though it does question the exclusive joint occurrence of species which are not connected through food-webs (Vasilevich, 1968).
Langford and Buell (1969), on the basis of their studies of forest ecosystems, criticized the validity of the continuum concept. They believe that vegetation mostly exists as a pluridimensional continuum though rarely in some situations the graded expressions of a single factor may lead to the formation of a unidimensional or linear continuum.
According to Langford and Buell, the continuum concept though certainly useful, has failed so far “to resolve the problems as to which of (i) the individualist concept of the plant association, or (ii) the concept of sharply delimited communities, gives the most reasonable interpretation of the nature of climax vegetation.” They further express the opinion that the problem of continuity cannot be resolved by comparing communities that have not reached the climax stage.
If vegetation is continuous, why should it be classified at all? The answer to this poser may be that communities are continuous in an abstract sense but very rarely in a concrete sense. Even otherwise continuous vegetation can become discontinuous by changes in steep environmental gradients.
However, such discontinuities are really illusionary. Further, the interaction between two species often results in distinct boundaries or ecotones which may become pronounced if one species is conspicuous relative to the other.
Vannote (1980) has proposed the “river continuum concept” in an attempt to integrate predictable and observable biological features of biotic water systems with the physical geomorphic environments. He postulates that biotic communities of natural streams adopt processing strategies that seek to minimize energy loss with downstream communities capitalizing or thriving on upstream processing inefficiencies.
According to adherents of the continuum concept, the distribution of each plant species along environmental or geographical gradients occurs independently of other species and no distinct assemblages of species (communities) with common boundaries occur in nature. This view was strongly supported by the work of Whittaker and Niering (1965) who observed that the distributions of various species of oaks along an elevation gradient did not tend to associate with distinct communities.