Range Type:  Defined and Described

Description and Designation of Vegetation vs. Biotic Communities: The Nature Conservancy and the United Nations Systems of Vegetation Classification vs. Brown et al. Biotic Community Classification

Associations and Consociations

Cover Types are Dominance Types

Origin of the term 'Type'

Historical Note:  "Founding Fathers" of Range Management

Role of Habitat

The “lost” legacy of George Nichols   

Conclusion

The Nature Conservancy hierarchy of vegetation classification system (NVCS) for terrestrial ecological communities is:

Formation Class

                  Formation Subclass

                                    Formation Group

                                                       Formation Subgroup 

                                                                                        Formation

                                                                                                          Alliance.

The first four units are physiognomic levels and the last two units are floristic levels.

Vegetation type and/or cover type units are obviously absent.

A “working draft (do not cite)” of “An Initiative for a Standardized Classification of Vegetation in the United States” by the Ecological Society of America (ESA) Vegetation Classification Panel (1997) stated unequivocally that there is a lack of agreement on terms used in describing and classifying vegetation:

“Vegetation science has had neither consensus nor national direction for its basic vocabulary” (p. 9).

The Panel working draft further commented:

“Agreement on terms often is not possible without agreement on the concept to be used, or at least respect for alternative perspectives on concepts” (p. 15)

With introduction of the U.S. National Vegetation Classification (USNVC) System The Nature Conservancy did compare or relate (briefly) their system to the Society of American Foresters (SAF) cover types:

“SAF cover types are ‘a descriptive classification of forest land based on present occupancy of an area by tree species’ (Eyre 1980). By contrast the USNVC uses all vascular plant species present in a community to help define vegetation units. Where dominant tree species are also diagnostic for a community type, similar, though not identical, units are identified by SAF cover type and USNVC alliance. Examples include the Jack Pine Forest Cover Type and the Pinus banksiana Forest Alliance, and the Post Oak-Blackjack Oak Forest Cover Type and the Quercus stellata-Quercus marilandica Forest Alliance. In other cases, the SAF cover types are more broad-ranging over both structural and environmental gradients than are the alliances recognized in the USNVC. For example, the Black Spruce Cover Type is separated by the USNVC into an upland Picea mariana Forest Alliance, a Picea mariana Saturated Forest Alliance, and a Picea mariana Saturated Woodland Alliance. The primary difference between these systems is that the SAF cover types are defined exclusively for forest vegetation, whereas the USNVC classifies all terrestrial vegetation.” (Grossman et al., 1998, p. 44).

Both The Nature Conservancy (Grossman et al., 1998, ps. 7-8) and the Society for Range Management (Shiflet, 1994, p. xii) cited classification systems of potential natural vegetation, noting especially the well-known maps of A.W. Kuchler, and clarified that their units of vegetation sometimes coincided with climax or potential natural vegetation. Shiflet (1994, p. xii) was explicit:

“Some of the cover types included here may bear some resemblance to those of Kuchler, but most do not since they are, for the most part, the result of human influence since the mid-1880’s. If plant succession were allowed to take place unimpeded, in time the vegetation of these should approximate those communities described by Kuchler”.

The Society of American Foresters offered a similar assessment for the forest cover types it described:

“Without disturbance, the present forest would tend to move toward the Kuchler potential and some SAF descriptions of climax types strongly resemble Kuchler’s phytocenoces” (Eyre, 1980, p. 3).

The UNESCO system in turn was a direct application of the Physiognomic- Ecological Classification of Plant Formations developed by Ellenberg and Mueller-Dombois in 1967 and shown in detailed revised form in Mueller-Dombois and Ellenberg (1974, ps. 466-488). The vegetation hierarchy of the Ellenberg and Mueller-Dombois scheme is:

Formation Class

                     Formation Subclass

                                        Formation Group

                                                           Formation

                                                                            Subformation

                                                                                                   Further Subdivisions.

The USNVC System also differs from the Mueller-Dombois and Ellenberg Physiognomic-Ecological System in that the former is based almost solely on the single factor of vegetation:

“A pivotal decision made by Conservancy ecologists was to develop a terrestrial classification system that was based primarily on vegetation” (Grossman et al. 1998, p. 14). 

The UNESCO Structural-Ecological Formation System is also based primarily on vegetation but “some environmental-geographic information” is incorporated. It’s hierarchial levels “are real vegetation units” but the system is “artificial” (Mueller-Dombois and Ellenberg, 1974, ps. 163-164). Mueller-Dombois and Ellenberg (1974, ps. 157-168) reviewed structural vegetation units and systems like UNESCO. In this context vegetation structure refers to “physiognomic criteria” such as plant life (=growth) form (which is the most important), plant size or height, deciduousness vs. evergreenness, and leaf features (eg. shape, size, and texture). These are essentially the same features of vegetation physiognomy given by Warming (1909, ps. 137-140 passim). They have traditionally formed the basis of distinguishing formations:

“Plant communities that are dominated by one particular life form, and which recur on similar habitats, are called formations [in the physiognomic-ecological sense]” (Mueller-Dombois and Ellenberg, 1974, p. 157). 

A longer historical examination suggests that the dichotomy between European and Anglo-American views of formation and physiognomy is not as clear as Muller-Dombois and Ellenberg (1974) suggested. Warming (1909, p. 139) credited Grisebach with the introduction of the term “formation” or “vegetative formation” in 1838 “in the form of ‘phytogeographical formation’”. The original coinage and usage by Grisebach with the adjective of “geographical” is clearly consistent with the interpretation of Clements suggesting that it was the European and not the American (Clementsian) school that took formation away from it’s seminal usage. This is further suggested by the definition by Warming (1909, p. 140):

“A formation may then be defined as a community of species, all belonging to definite growth-forms, which have become associated together by definite external (edaphic or climatic) characters of the habitat to which they are adapted”.

Currently (and probably for the forseeable future) the system of classification for North American vegetation that is both comprehensive and most consistent with the traditional biome-association-dominance type model used in Range Management and Forestry (eg. SRM and SAF cover types) is that of Brown et al. (1998). These workers specified that their classification was "a biotic-community approach" in contrast to "purely vegetative classification systems", but their seven-level system published as A Clsssification of North American Biotic Communities included examples only of the biotic community (fourth level) and the series (fifth level) for most communities and to the association (sixth level) for some communities (Tables 4 and 5, ps. 36-48). All of these were designated by dominant plant species. Association was defined consistent with it's usage traceable to the International Botanical Congress in 1910, but spatial scale of association was "more or less local" (Brown et al., 1998, p. 35) and thus not consistent with the Clementsian-Tansylian association that was by and large the scale, interpretation, and usage of association adopted by the professions of Range Management and Forestry. Brown et al. (1998, p. 35) specified that the sixth level refered to "a distinctive association" with such associations "generally equivalent" to the Daubenmire habitat type. The few examples of the association (sixth level) were "plant associations". Nonetheless, the series (fifth level) corresponded very closely with the SRM and SAF cover types. Brown et al. (1998, p. 14) stated that their "general series of actual or potential plant dominants" were referred to as cover types by the Society of American Foresters (and alliances by The Nature Conservancy).

As was explained above, cover types and alliances are not synonyms and are not really interchangeable or completely convertible/translatable from one to the other. Many (probably most) of the Brown et al. (1998) series were very similar and often essentially synonymous with SRM and SAF cover types. These two units of vegetation are far rmore similar to each other than is either to The Nature Conservancy alliance because Brown et al. (1998) series and cover (= dominance) types both are essentially Clementsian in origin and outlook while The Nature Conservancy classification system is fundamentally based on the Braun-Blanquet model (Zurich-Montpellier School of Phytosociology). Brown et al. (1998, ps. 9-12) briefly reviewed the approaches to vegetatation classification more commonly used in the United States (including that of The Nature Conservancy). They also explained how and why A Classification of North American Boiotic Communities had adopted the Clementsian biome and the terms/concepts of disclimax and subclimax.. They then explained "advantages of a biotic-community approach over purely vegetative classification systems" (Brown et al., 1998, ps. 12-14).

Brown et al. (1998, p. 12) stated that "Clements was a great ecologist" and "... many of his contributions remain valid in modified form". (This conclusion was enthusiastically shared by your current author as was made evident throughout this publication.)

The Brown et al. (1998) hierarchial classification system was highly recommended and extremely useful. The earlier publication, Biotic Communities- Southwestern United States and Northwestern Mexico (Brown, 1994), was even more utilitarian and applicable in classifying and describing flora-fauna communities and, especially, vegetation in that region of the continent.

Unfortunately, A Classification of North American Biotic Communities (Brown et al., 1998) did not include enough series to correspond with all SRM and SAF cover types. This limitation was especially noticable and problematic for North American grasslands. The Brown et al. (1998) classification however did include (without major emphasis) subclimax and disclimax plant associations as well as climax vegetation (Brown et al., 1998, p. 35-36). This was particularily useful for Range Management as the biotic community system included such important disturbance climaxes as the California annual grassland (Annual Disclimax Series) and former climax bunchgrass-shrub steppe depleted to cheatgrass range (Cheatgrass Disclimax Series). Here again, though, Brown et al. (1998) omitted the climax Stipa bunchgrass prairie of the Central Valley, Coast Range, and Sierra Nevada foothills. This omission was consistent with omissions in the published SRM cover types (Shiftlet, 1994) and U.S. Forest Service forest and range ecosystems (Garrison et al, 1977) and inconsistent with the potential natural vegetation mapped and described by Kuchler (1964). Interestingly, for the Kuchler manual describing the climax California bunchgrass prairie, California steppe, there was apparently not even a readily available photograph of this pre-Columbian grassland so that a U.S. Forest Service photograph of California foothill annual grassland was used (Kuchler, 1964, p. 48). This was the same photograph used to accurately illustrate annual grassland disclimax in the manual by Brown et al. (1998, Plate 62, p. 88)! This rather comic situation was symptomatic (and symbolic) of the fact that the disclimax grassland designation was a more correct one for practical management than the designation of the pristine-- and essentially "extinct"-- virgin range.

Brown et al. (1998) series and associations were also quite limited for conifer forests of western North America. This was especially obvious (and bothersome) for SAF cover types of the Northern Rocky Mountains. This situation may have been partly a result of the Brown et al. (1998) emphasis on climax vegetation whereas several SAF cover types were "based on existing tree cover", "... forests as they are today...", such that some of these types are transitory or temporary (ie. seral) to climax forest communities (Eyre, 1980, p. 1). That partial explanation was largely unsatisfactory however because: 1) Brown et al. (1998) classification did encompass units of disclimax and subclimax forest vegetation and 2) Brown et al (1998) omitted several distinct forest communities long recognized as climax by workers like the "great ecologist" Clements. For example, Brown et al. (1998, ps. 36- 38) omitted series and associations in the Rocky Mountains that applied to SAF cover types 212 (western larch), western white pine (215), western redcedar- western hemlock (227), and western redcedar (228) that are all parts of the larch-pine forest association (thus climax) recognized by Wesver and Clements (1938, ps. 503-504).

Overall, the Brown (1994) and Brown et al. (1998) classification appeared to be more relevant-- more applied and more similar-- to range and forest cover types as both a hierarchial system and designations/descriptions of plant communities (= vegetation) than those of The Nature Conservancy and United Nations. In some respects, the Brown et al. classification of biotic communities was also more comparable to the Kuchler (1964) map and manual of Potential Natural Vegetation (disclimax and subclimax of the former being exceptions). Kuchler units of vegetation served as the basis for Forest Service forest and range ecosystems (Garrison et al., 1977, ps. 1-3) and for comparisons with SAF (Eyre, 1980, p. 3) and SRM (Shiflet, 1994, p.xii) cover types.

On the European mainland sociation was seen as a basic unit of vegetation classification where a sociation was any plant community that essentially had a “homogenous species composition with at least certain dominant species in each layer”. Sociation was one of the lower units in the Clementsian hierarchy of climax vegetation largely based on or describing seasonal sequence or phenological aspects (hence “aspect”) (Clements, 1936, ps. 276-277, 281). 

Consociation was a major unit of vegetation in both European thought and in the Anglo-American (Tansley-Clements) school of dynamic vegetation (Clements 1916a, p. 120; Tansley, 1926, ps. 34-36; Clements, 1936, p. 274). In the Clementsian model  consociation was “the unit of the association” (Clements, 1916a, p. 120) though the association “is to be regarded as divided into definite groupings of dominants or faciations” where faciation ”is the concrete subdivision of the association” (Weaver and Clements, 1938, ps. 94-95).  The distinguishing or defining characteristic of consociation was presence of a single dominant, one dominant species, or at least a single major dominant if several domiants were present. In an abstract sense each major dominant of an association was a consociation (Weaver and Clements, 1938, p. 94). Tansley (1926, p. 34) explained that an association usually had “more than one dominant species” or co-dominants. The community (the unit or level of vegetation) within an association formed by a single dominant species was the consociation. Pinus ponderosa in the montane forest formation and Stipa spartea or Agropyron smithii within true prairie were used as examples of consociations in Weaver and Clements (1938, p. 94-95).         

Mueller-Dombois and Ellenberg (1974, ps. 172-173) concluded that as a unit consociation had “no universal applicability”. Consociation was difficult to use in communities having high species richness or, conversely, a single species “may become dominant under different habitat conditions” (ie. when a single dominant species is used to define or delineate a community type “very different habitats” may get grouped together). The first difficulty is self-evident and an obvious limitation. The second condition of lumping together different habitats is a ridiculous criticism, at least at larger spatial scale (within the same or similar climate). Yes, it is a true statement of the actual situation but: 1) vegetation and not habitats are being classified, described, mapped etc. and 2) at that large scale (the first subdivision of a formation) “jillions” of habitats are included and these habitats, in turn, will be subdivided out at progressively lower, more specific levels. In other words with regard to 2) no unit of vegetation is site-specific unless it also includes habitat criteria, but progressively smaller units of vegetation have progressively less heterogeneous habitats. Incidentially, recall from discussions in the Range Site section that this same sort of grouping together (inclusion) of different habitats in the same taxonomic unit of vegetation is typical of the habitat type approach.

The obvious difficulty with consociation was the confusion in the different ways in which  association was applied. As with— actually originating with— formation, the vegetation unit of association was applied differently in Continental schools of vegetation than in the Anglo-American school (the predominate Anglo-American perspective).  As a concept and vegetation unit, association has been second to none or only to formation (or climax and, later, biome in Clementsian Ecology). Association has been a— perhaps the— major unit and unifying concept in description, classification, and mapping of vegetation. Association is traceable to the early days of Plant Geography. Warming (1909, ps. 137, 144) credited the origin of plant association (plantes associees) back to Humboldt (1807). Warming’s Oecology of Plants (1909, English translation) has often been regarded as the world’s first textbook in Plant Ecology. In it Warming listed several other plant geographers and ecologists who employed association (or its concept by other names). He defined association as “a community of definite floristic composition within a formation; it is, so to speak, a floristic species of a formation which is an oecological genus” (Warming, 1909, p. 145).  Associations were “smaller, often more-localized subdivisions or kinds of the formation” (p. 144). As was species to genus so was association to formation. To Daubenmire (1968, p. 29) “the association serves the same purpose in synecology as the species does in taxonomy”. Daubenmire stated that there was a long-standing analogy of association to species and that while “[n]either is susceptible to rigorous definition…” both remain equally essential to their fields. Tansley (1926, p. 30) succinctly put the association is perspective by describing it as “the fundamental unit of vegetation”.

In 1910 in Brussels the Third International Botanical Congress adopted by official resolution the following definition of plant association:

“ … a plant community of definite floristic composition, presenting a uniform physiognomy and growing [or “when occurring”] in uniform habitat conditions” (Daubenmire, 1968, p. 27; Mueller-Dombois and Ellenberg 1974, ps. 173-174).

After a five page review of association Daubenmire (1968, p. 32) concluded that association was “a type of climax phytocoenosis”. He then remarked that “very little” climax vegetation remained (due largely to human-induced disturbance) yet “each climax can normally regenerate itself repeatedly following destruction” so it was the area belonging to each association that is of ecological importance. This area of one plant association or that had been in one association or that had potential to regenerate that association was Daubenmire’s habitat type (synonym, homoece).

***Note: if habitat type is equal to plant association (or previous or potential plant association) and association is a subdivision of a formation, the largest unit of vegetation in the original usage of the term, then there is absolutely no way whatsoever that habitat type could be the equivalent of or a synonym for range site, the smallest unit, the most distinctive kind, of native grazing land. Given the historic definitions of these vegetation units, the attempt to rationalize the equivalency of habitat type to range site (discussed in the Range Site section) seemed ludicrous to the present author. [It will be explained below that the Daubenmire association was generally closer to a subassociation of the Clementsian association, a unit more similar to a cover type.]    

Unfortunately, the 1910 Botanical Congress definition, which was nearly verbatim (with an addition or two) that of Warming, did not specify the basis for delineating an association and the term/concept came to mean two different things to the Continental and Anglo-American views of vegetation (Mueller-Dombois and Ellenberg, 1974, ps. 174-175). In his last major paper on the climax Clements (1936, p. 273) noted this difference between the two schools of thought in defining association:

“Under the climax concept this represents the primary division of the biome or formation, and hence differs entirely from the generalized unit of the plant sociologists, for which the term community is to be preferred. Each biome consist regularly of two or more associations …”.

The Clementsian school retained the meaning of association specified by Warming (1909, ps. 144-145) as a subdivision of the formation or as species of the formation genus.

Mueller-Dombois and Ellenberg (1974, p. 175) described the Clementsian association as “the general plant cover in a given macroclimatic region (ie. a vegetation mosaic)” or “more or less a climatic subregion”. The latter description is apt because the formation was based on, determined by, regional climate; the former description was not of an association but of a formation which includes several associations. Mueller-Dombois and Ellenberg (1974, p. 175) were correct, and specifically so, in a synopsis statement:

“Each climax was subdivided into a few ‘formations’ (regions) and each ‘formation’ was subdivided into two or more ‘associations’… Clements defined an association floristically by joining the names of two regionally dominant species and then implied that an association was a grouping of two or more consociations”.

As Clements (1916a, p. 128) first used association it was a floristic subunit of the physiognomically-developmentally derived formation. Fundamental development and life-forms were the same for all associations in a formation (these two features unified or served as basis for formation); differences in dominant species distinguished the various associations. In Plant Ecology Weaver and Clements (1938, p. 93) introduced the association with this description:

 “Every climax formation consist of two or more major subdivisions known as associations. These are climax communities associated regionally to constitute the formation. The number of associations in a particular formation is naturally determined by the number of subclimates within the general climate of the formation.  Each association is marked by one or more dominants peculiar to it”.

One of the most explicit descriptions of plant association was in Plant Indicators:

“Each climax formation falls readily into two or more major subdivisions known as associations. Toward their edges these blend into each other more or less, making a transition area or ecotone. The latter is broad in the case of relatively level regions, and narrow in that of the climax zones of mountain ranges. The associations have one or more dominants in common, or at least belonging to the same genus, and there is a certain degree of similarity as to subdominants also. Each association consist of several dominants as a rule, though there may sometimes be as many as eight or ten or more, as in scrub and chaparral. Each dominant constitutes a consociation. It may occur alone, though as a rule it mixes and alternates with the other dominants of the same association This is the direct outcome of the similar requirements of the dominants, and hence it is a guiding principle that two or more consociations are regularly associated in the larger unit (Clements, 1920, p. 107).

This view expressed by Clements came to be the interpretation of most British, American, and Canadian ecologists as well as those in nations with ties back to “mother countries” (eg. South African ecologists like J.F.V. Phillips were largely of the Anglo-American school). The various “schools” of ecological thought were defined largely by geographic location. These were reviewed exhaustively by Whittaker (1962). Shimwell (1971, ps. 44-62) also discussed the different Schools or Traditions and drew pedigrees or lineages of them. The English Tradition (Shimwell, 1971, ps. 47, 54) which dominated Range Ecology and Forest Ecology during their formative years and formed the foundation of modern Range Science and Forest Science can be traced back through Clements (and to a lesser degree Tansley, C.E. Moss, and H.C. Cowles) to Warming, Oscar Drude, and finally to the fountainhead of August Grisebach, an oft under-rated plant geographer who described vegetation types physiognomically and in relation to climate and who was the first to use the term formation (Clements, 1916a, p. 116).

The English-American-South African Tradition came to be known as the holistic or organismic school of vegetation. It was clearly the “dominant species” of ecological thought in its day and probably, though in less pure form, today. A smaller, less influential school was the individualistic or continuum school begun by the American ecologist Henry A. Gleason and more or less perpetuated— at least experiment-wise— by Robert H. Whittaker. Whittaker actually defied pigeon-hole categorization being somewhat of a hybrid. He did expand the Gleasonian individualistic concept through gradient analysis but as discussed above (Biome section) he successfully reconciled Clements’ monoclimax and Tansley’s polyclimax. This, plus his views on communities (Whittaker made much use of biomes), would place him squarely in the Anglo-American “pedigree” (Shimwell, 1971, p. 54).

The Gleasonian branch of the English Tradition (Shimwell, 1971, p. 54) became known for its “individualistic concept of the plant association”  (Gleason, 1917; 1926). Allen (1998) contrasted the Clementsian and Gleasonian associations by remarking that Clements used association as an abstraction for a unit of climax vegetation while ignoring seral stages leading to climax (as a subdivision of formation or climax, association was not a seral unit). By contrast Gleason used association as a concrete unit of vegetation based on species composition at an instant of time and at a site while ignoring climax. Allen (1998, p. 320) overlooked or ignored the fact that in his elaborate scheme of vegetation units Clements (1936, ps. 278, 281) included as the equivalent of associations the seral unit, the associes, which was “the major unit of every sere”. (Clements was  always one jump ahead of potential criticism and covered about every conceivable weakness in his pet theories.) Nonetheless, Allen’s observation on Clements’ association being an abstraction was partly correct because Clements (1916a, p. 126) stated that association was used “in both abstract and a concrete sense”. “The general use of association in the concrete has fixed it definitely in ecological terminology”. On the other hand, Allen’s interpretation of the Clementsian association as an abstraction seems to have been a view either anticipated by Clements or made to him, and he seems to have denied it:

“… the formation is necessarily an organic entity, covering a definite area marked by a climatic climax. It consist of associations, but these are actual parts of the area with distinct spatial relations. The climax formation is not an abstraction, bearing the same relation to its component associations that a genus does to its species” (Clements, 1916a, p. 127).

Shimwell (1971, p. 53) commented on these words and concluded that Clements “conceived the formation not as an abstraction”.

Baker (1950, p. 44) concluded that Clements (1928) used association “in a somewhat abstract sense to cover a section of a formation occupying a subclimatic zone and therefore having climax dominants dictated by that subclimate…”.  Nichols (1923, ps. 15-17, 172) discussed the abstract vs. concrete concept of plant association. He cited a  resolution routed by letter to 85 ecologists (“mostly botanists and foresters and mostly members of the Ecological Society of America”) in 1921 which read in part: “That the term Plant Association be recognized as applicable both to the abstract vegetation concept and to the concrete individual pieces of vegetation on which this concept is based”. Of the 76 respondents, 67 favored adoption of the recommendation (Nichols, 1923, p. 15).

For his part, Clements (1916a, p. 117) concluded that the formation of Grisebach and the association of Humboldt “meant practically the same thing by their respective terms”.  Clements (1916a, p. 118) quoted Moss (1910) as dividing formations into associations following “many previous authors” but these were traced back to Warming as being the first to publish this subdivision (Clements 1916a, p. 121).

Obviously association so defined and perceived would be of immense size, considerable  environmental heterogeneity (as noted earlier), and of a mosaic pattern among other associations of the same formation. Warming (1909, p. 145) specified: “Associations may occur irregularly as patches in the formation; or may exhibit a zonal arrangement”. Tansley (1926, p. 31) stated it thusly: “Associations are, on the whole, large units, with wide extension, though this may be, and practically always is, interrupted by the occurrence of other associations”. Daubenmire (1952, p. 302) described this same pattern when he identified associations in coniferous forests.

It seems that Warming, Tansley, and Clements shared a similar interpretation of association. For example, all used the example of “reed swamp” which was a community dominated by Phragmites communis, Scirpus lacustris or S.validus, and Typa spp. Though Warming (1909, p. 145) referred to “reed-formation” he used these individual species as examples of associations within that formation which would thus be consociations in Clementsian terms. Tansley (1926, p. 31) referred to the “reed swamp association” but did not specify single species units (= consociations) as a form of association. Clements (1936, p. 264, 278) noted that “reed swamp” was “the universal example” of a “serclimax” and Weaver and Clements (1938, ps. 62-63, 100) treated the “reed swamp stage” as a seral stage and not climax so that each unit of vegetation dominated by a single one of these “reed swamp” species was a consocies, the seral equivalent of the climax consociation.

The difference between association vs. associes (or consociation vs. consocies) between Tansley and Clements was the difference between Clements’ monoclimax (where meso-,  mesic, level would theoretically be the terminal regional climax) and Tansley’s polyclimax (where hydric-, mesic-, xeric-levels could all lead to local, say, edaphic climaxes). The important point is that all of these ecologists of the English Tradition saw associations (associes) at the same spatial scale. Since then a younger “descendent” of the Anglo-American school cited this same example of a consocies (Polunin, 1960, p. 334, 505). Likewise, Polunin (1960, p. 333-335) gave the classic Clementsian units of vegetation verbatim from Clements (1936) climax paper illustrating the vast profusion of the Clementsian view of vegetation and its dynamic development and the persistence of the Anglo-American Tradition of Vegetation Science.

Critical note: the important sweeping— and inevitable— conclusion is not that one ecologist or school (= tradition) was “right” or “more nearly correct” and some other school or model was “wrong” but rather which one(s) formed the basis and had most impact on current applied ecological thought. The fact is that the paradigm (some would say “philosophy”) of the Clements-Tansley group going clear back to Warming and Grisbach was the view that dominated American, British, Canadian, and South African ecological thought. And this domination— almost an exclusion of counter thought— was profoundly influential on the fledging disciplines and professions dealing with management of native vegetation and, for that matter, natural resources in general (eg. Soil Science, Wildlife Management).

To understand current systems treating of range and forest vegetation, students of these resources must understand the basis of this Vegetation Science upon which Range Management and Forestry are built. That basis is overwhelmingly Clementsian.

Range types (both rangeland and forest cover types) are one of the best examples of this lasting Clementsian influence. As with range site, there is no specific level or unit of vegetation in any classification or vegetation mapping system that corresponds exactly and consistently to range type (rangeland and forest cover type). It was shown previously in this review that range site is a unit of vegetation and certain environmental or habitat factors, especially soil (Range Site section). By definition, range site cannot be “retrofitted” into any system or classification scheme that is devoted solely to vegetation and that ignores habitat. By contrast, both plant formation (climax vegetation at regional or zonal scale) and plant association (the Anglo-American association at any and all scales) are both units strictly, solely, of vegetation with no environmental factors (except the biotic interactions of the vegetation) directly included. As explained earlier, this is the exact situation with the Daubenmire habitat type which is the equivalent of the  Daubenmire association. Vegetation cover types are also strictly, exclusively, units of vegetation and therefore could theoretically match exactly one or more units of vegetation in the numerous schemes or hierarchies of vegetation classification, description, or mapping. At present none do match. Range cover types do not correspond directly and equivalently to any unit of vegetation in existing vegetation classification systems. For example, The Nature Conservancy unit of alliance in the USNVC System corresponds but approximately to the SAF forest cover types.    

Where do rangeland and forest cover types fit into vegetation classification? To which—if any— units of vegetation do they correspond? From which— if any —did they arise?

Once again it was the genius of Whittaker (1975) to synthesize and clarify the various uses of concepts in context of vegetation types or cover types and to suggest the origin of forest and range types. In making more sense than most out of plant community classification Whittaker (1975, p. 128, 135) stated that the biome or formation is a major kind of community classified or defined, in principle, solely by physiognomy but, in practice, a combination of physiognomy and environment is required. “The physiognomic approach classifies communities by structure—generally by the dominant growth-form or the uppermost stratum or the stratum of highest coverage in the community”. Further:

 “Structural or physiognomic classification is the usual approach to description of the communities of a continent, or of the world, and it is widely used by geographers, climatologists, and soil scientists as well as ecologists” (Whittaker, 1975, p. 128).

Consistent with the preceding discussion, Whittaker (1975, p. 135) explained that “on a given continent” biome (or formation) is the broadest possible classification level or structural/organizational unit of plant and animal community (or of plant community, vegetation, only). He then went one unit of scale larger and one  organization level of community higher to define biome-type or formation-type as “[t]he still broader [thus the broadest of all] grouping of convergent biomes or formations of different continents”.

Obviously the biome-type or formation-type is the broadest unit or most general kind of  vegetation (eg. “grassland biome type” and “desert biome type” include all grasslands and deserts on Earth). This is not to be confused with vegetation or cover type which is the major subdivision of  biome or formation. But the use of the term “type” at different levels of organization and scale for the grouping of several units at these respective levels is consistent. Type is the all-inclusive word for the unit that encompasses all forms or kinds of plants and animals or of vegetation at a given taxonomic level (taxon) of organization.

Whittaker (1975, p. 128) followed the biome and formation unit with the unit he labeled as dominance-type:

“Classification by dominant species is a natural and widely used approach.  Community-types defined by their dominant species can be termed dominance-types, but often they are called simply ‘types’”. 

Whittaker noted that plant communities are usually dominated by more than one species “and subjective decisions on what combinations of major species ought to be recognized as dominance types are necessary to make the classification work”. The subjectivity (and considerable experience and familiarity with the vegetation involved) notwithstanding,  dominance types are “quite workable … and physiognomy and dominance can be used together, with dominance-types subordinate units within formations”.

In this last phrase Whittaker described the typical arrangement or organizational pattern of the major units of vegetation as commonly used by foresters and rangemen, at least in North America. Whittaker (1975, p. 128-129) pointed out that ecologists like Clements and Braun used “as units very broadly defined dominance-types, or groups of dominance-types”. Their term for these broad dominance-types was, of course, the association as defined by dominant species or genera such as the oak-hickory, beech-maple, maple-basswood, and oak pine associations or, as they are synonymously known, “types”. This usage was also noted by Shimwell (1971, p. 62). In other words, type is often at the association level as the largest or broadest unit or subdivision of a given formation (eg. the preceding associations within the deciduous forest formation of eastern North America). This is the exact usage or organization (community hierarchy) of Braun (1950, p. 10-11). As quoted previously, association was explained in considerable detail in Plant Ecology (Weaver and Clements, 1938, p. 478-482) after having first defined association on pages 93-94:

“Every climax formation consists of two or more major subdivisions known as associations. These are climax communities associated regionally to constitute the formation. The number of associations in a particular formation is naturally determined by the number of subclimates within the general climate of the formation. Each association is marked by one or more dominants peculiar to it…An association is similar throughout its extent in physiognomy or outward appearance, in its ecological structure, and in general floristic composition [the mixed-prairie association was one example used; likewise Stipa spartea and Sporobolus heterolepis were two dominants of the true prairie association]”.

This was requoted at extended length to show the key roles of physiognomy and floristic compostion which Clements retained from the association of Warming (1909, p. 145).

In turn:

“The visible unity of the climax is due primarily to the dominants or controlling species. All of these belong to the same life form. … Each formation is named after two of its most widely spread and important dominants” (Weaver and Clements, 1938, p. 91).

A concise summary by Shimwell (1971, p. 53) explained this arrangement and the entire Clementsian hierarchial organization (complete with seral units and climax units mentioned above):

“Formations were subdivided into associations which could be characterized by their codominant species. Within the association, if a single species was the physiographic dominant then the term consociation was used, while these were further subdivided into societies each characterized by subordinate species.  However, these terms were restricted to the considered climax vegetation type.  All other developmental units leading up to a climax, all seral stages, were given parallel terms such as associes, consocies and socies”.

From this passage it is the association that is the vegetation unit relevant to the current discussion. The Clementsian association (or consociation where there was a sole dominant) was the unit of the “climax vegetation type” which was determined by dominant or co-dominant species. This description showed that vegetation type corresponded to association which was defined or identified by its physiographic dominant(s). Ergo, the unit Whittaker (1975, p. 128) identified as dominance-type.  

Whittaker (1975, p.129) was careful to specify that “the American association of Clements and Braun” was distinct from the floristic association of Josias Braun-Blanquet or the Zurich-Montpellier School of Phytosocioloy (= Releve Method). Shimwell (1971, ps. 57, 59) drew the same distinction. Use of the terms association and alliance for the lowest and second lowest units of vegetation in The Nature Conservancy USNVC System followed in similar format the hierarchy of the Braun-Blanquet Zurich-Montpellier (Shimwell, 1971, ps. 56-62).

The association in the USNVC System is the Braun-Blanquet association and not the American association which Clements adopted from Warming. It is the Warming-Clements association (often subunits thereof) that is the basis of the cover or dominance types of range and forestry practitioners.

The Nature Conservancy United States National Vegetation Classification System is primarily an application of the Zurich-Montpellier Tradition (elaborated by Shimwell, 1971, ps. 45-47, 50-53, 56-62) to the vegetation of North America (specifically the United States of America).The USNVCS is largely an adaptation of the Ellenberg and Mueller-Dombois (UNESCO) classification from which the former is modified at upper levels (Grossman et al., 1998, p. 20). It follows that the SRM and SAF cover types, which are essentially dominance-types derived primarily from the Anglo-American Tradition of Clements and Tansley (Shimwell, 1971, ps. 45-62 passim), will inherently correspond poorly to the The Nature Conservancy System. This was bound to be the case because the SAF/SRM cover types and The Nature Conservancy USNVCS alliances are units derived from vegetation classification systems of two schools or traditions of plant community ecology that have been distinct and separate for 90 to 100 years (again, Shimwell [1971]). This was shown even before The Nature Conservancy USNVC System was developed when (as discussed above) the Ecological Society of America Vegetation Classification Panel (1997) in its initial report rated, as a level or unit of vegetation, the SRM cover type dead last.     

This rating was understandable given that vegetation units are from two distinct traditions. TNC and ECA rated cover type relative to their tradition and found it to be  unacceptable or least acceptable. SAF and SRM with their units derived from another tradition should— for purposes of clarity and consistency— reciprocate toward the USNVC System and rate the USNVCS association and alliance at the bottom of their lists.

In a review of its initial report (ie. its second report) on an initiative for a classification of United States vegetation the Ecological Society of America Vegetation Classification Panel (1999, p. 11-12) concluded that the SAF and SRM cover types and dominance types were close approximations of each other: dominance types “are very similar in concept to ‘cover types’ for simple, efficient inventory and mapping purposes”. Community-layer dominance types “are simply a convenient, descriptive subdivision of Cover Types or Domanance Types based on the dominant species in a lower, conspicuous layer of vegetation”. In this second report the ESA Vegetation Classification Panel did not relate or equate cover or dominance types to any units of vegetation in The Nature Conservancy USNVC System. The Vegetation Classification Panel did discuss briefly the unit of plant association, but only from the standpoint of “[t]he two major association classification approaches in North America that are based predominantly on composition of the entire plant community” which are the 1) Braun-Blanquet and 2) Daubenmire systems. Specification of “entire plant community” automatically eliminated any of the systems based on dominance (distinction based on dominant species) such as that of Weaver and Clements (1929, 1938). Yet these dominance-based systems are the systems or classifications which have dominated vegetation schemes and hierarchies used almost exclusively in Range Management and Forestry in North America. In the USNVC System both cover and dominance types were disregarded or ignored as was plant association as defined and designated by the Anglo-American Tradition (ie.the association as developed over the Grisebach-Warming-Moss-Nichols-Tansley-Clements-Weaver-Phillips lineage).

In other words, The Nature Conservancy USNVC System made no use of any of the units of forest and range vegetation traditionally used by American workers except that of the habitat type as defined by Daubenmire which includes all layers of vegetation and not just the layer of dominants. Furthermore, given that “the USNVC is of existing, not potential, vegetation” (Ecological Society of America Vegetation Classification Panel, 1999, ps. 8, 14) even the Daubenmire association and habitat type was irrelevant for the USNVC System because the Daubenmire association is a “type of climax phytocoenosis” and “based on samples restricted to ‘near-climax’ vegetation” [ie. potential natural vegetation] (Ecological Society of America Vegetation Classification Panel, 1999, p. 12). Thus, the USNVC System is a modification of and based on the Braun-Blanquet releve phytosociological system (the Zuric-Montpellier Tradition or School of Phytosociology) of continental Europe and in final effect totally ignored the entire heritage of American research on classification and description of range and forest vegetation. As asserted by the Ecological Society of America Vegetation Classification Panel (1999, p. 13), a partial exception to the preceding fact were studies of vegetation “sampled with releves from late successional stands …”. Yet, examination of the investigations cited by the panel revealed that this sampling too was related back to “potential climax vegetation”. The USNVC System also differs from such vegetation maps as those by Kuchler which are of potential natural vegetation and not existing vegetation.

Beyond doubt or debate, the United States National Vegetation Classification System is a system separate from such units as range cover types and the Anglo-American unit of plant association on which range and forest cover types are based. “Classification of existing vegetation and classification of potential natural vegetation are complimentary and of equal importance in synecology” (Ecological Society of America Vegetation Classification Panel, 1999, p. 14).  So said and homage given to classification systems for climax vegetation, the obvious fact is that they are two separate though complimentary systems.

As recorded above, both the SAF and SRM cover types are of “existing vegetation”, but also as noted previously most of the rangeland cover types described are, upon close reading, closer to climax or potential natural vegetation than to seral communities (any disclaimer-like statements made by Shiflet [1994, p.  xii] notwithstanding). By way of example, compare how many rangeland cover type names begin with and list as dominants “Bluebunch Wheatgrass” or “Idaho Fescue” versus how many begin with or list as the dominant species “Cheatgrass”. Point made. Likewise, certain SAF forest cover types are climax types and correspond closely to potential natural vegetation such as that mapped by Kuchler (Eyre, 1980, p. 3).  Undoubtedly some of the USNVC units will be of climax vegetation, but this new classification scheme offers little to Range Management professionals that is not in existing range literature.      

Conclusion: the unit(s) or level(s) of vegetation traditionally used in North America that most closely correspond to cover (= dominance= vegetation= grazing= range) type as applied in Forestry and Range Management is the association or, its seral counterpart, the associes. Cover type and association remain the basic descriptive units of vegetation intermediate between biome or formation and range site. These hierarchial levels are still in use and of great value because much of the most important and detailed reports of range and forest vegetation used these units and these studies remain cornerstones of Range and Forest Ecology.

The actual connection of the forest and range types to the Clementsian association (or the seral unit associes in cases where the forest or range vegetation is seral not climax) is obscure and seems not to have been documented other than indirectly. Neither Eyre (1980) nor Shiflet (1994) included even a brief review of the type concept.

Toumey and Korstian (1947, ps. 356-397) and Baker (1950, ps. 22-47) reviewed units and systems of forest vegetation at a time when such approaches were as utilitarian as they were academic and when the history of their development was fresh in the “collective mind” of Forestry. These silviculturalists discussed such forests or forest vegetation units as the various kinds of formations (eg. climatic vs. physiognomic), forest regions or forest vegetational areas, associations, and forest types. There were different kinds of all units and hierarchial levels depending on purpose and author or mapper even at this relatively early stage or mid-age of American Forestry literature. (The American Society of Range Management was just “germinating”.) Both Toumey and Korstian (1947) and Baker (1950) distinguished forest cover type from the other kinds of forest types: 1) management, 2) permanent, 3) temporary, 4) physical, and 5) indicator. It is interesting from a historical perspective that then, as now, forest cover type was of the present vegetation (could be seral or climax) and had the current meaning (Toumey and Korstian, 1947, ps. 382-383; Baker, 1950, ps. 34-35; Munns, 1950, p. 88). At this time the Society of American Foresters (Baker, 1945) and Eyre (1954) published its revised  descriptions of forest cover types in North America, a project begun in 1929 (Eyre, 1980, p. v).

According to Clements (1920, p. 337) and Toumey and Korstian (1947, p. 379 in a portion that is essentially the review from Clements [1920]) the term forest type was used first by Graves (1899) of the United States Department of Agriculture Division of Forestry to mean a dominance (= cover) type of natural forest vegetation. Baker (1950, ps. 31-33) determined that the Society of American Foresters decided upon this concept for forest cover types at a special symposium (Dana et al., 1913) in the beginning days of their profession.

Even though the current concept and usage of forest cover type was established at the onset of Forestry and Forest Science in North America and while this coincided with the dawn of Ecology and the emergence of Clementsian Ecology (compare 1913 to the publication of Clements’ Plant Succession and Plant Indicators in 1916 and 1920), cover type was not an ecological unit or an equivalent to any ecological unit. Toumey and Korstian (1947, ps. 383-384) wrote:

“Cover type, as at present recognized and based on floristics and economics, has no equivalent in ecology. It takes no account of site factors, consequently temporary and permanent forest communities with unlike site factors may fall into the same cover types so long as the dominant species are the same”.

Also recall that cover types are based on present vegetation and not necessarily on climax or potential natural vegetation. In contrast, permanent forest type is the equivalent of climax forest vegetation so its synonym is climax type (Toumey and Korstian, 1947, ps. 380-384; Baker, 1950, ps. 44-46). Forest cover type can include both permanent and temporary forest types. Examples of the latter include subclimax forest types or ecologically transitional forest types like those of Pinus strobus of the Great Lakes forest region and Pseudotsuga menziesii of the Pacific Northwest Coastal forests. Thus even where species like eastern white pine and Douglas fir are clearly seral to the climax forest vegetation they are regarded as forest cover types because they are so valuable economically even though they are, when classified as to ecological forest type, temporary forest types and not permanent (= climax) forest types. 

Apparently it was through the correspondence or equivalence of permanent (= climax) forest types to the Clementsian associations (and temporary [= subclimax] forest types to the Clementsian associes) that eventually led to the close approximation for forest cover types to a generic, nonspecific, “more-or-less” Clementsian association with both terms used in less restrictive, less precise meanings than originally intended (Toumey and Korstian, 1947, ps. 382-384; Baker, 1950, ps. 44-46). Both of these Silviculture textbooks noted that the term association had different meanings to different ecologists or foresters (Toumey and Korstian, 1947, p. 377; Baker, 1950, p. 44). “Forest types as conceived by one writer are not the same as those of another” (Toumey and Korstian, 1947, p. 379).

Toumey and Korstian (1947, p. 382) noted that there appeared “wide divergence in opinion as to what constitutes a forest type”, but they interpreted Clements (1920) in Plant Indicators as having concluded “that the classification of forest into associations by ecologists and into forest types by foresters differs more in emphasis than in fact” (Toumey and Korstian, 1947, p. 382). Technically Clements did not relate or equate associations to forest types. The section of Plant Indicators referred to and interpreted by Toumey and Korstian (1947) was entitled “Forest Types” (ps. 337-345) and when Clements (p. 342) wrote “… they differ more in emphasis than in fact” he referred to the first part of that same sentence which was “… of the opinions just summarized”. These were opinions of numerous foresters all discussing various versions of forest types including cover, management, permanent, temporary, physical types. There was no reference whatsoever to associations. However, a sizable portion of Plant Indicators (ps. 105-235) was devoted to description of actual (not abstract) vegetation and this was organized on the basis of formations and associations. Most importantly, both Toumey and Korstian (1947) and Baker (1950) concluded that the plant association as defined or interpreted in the Clementsian concept (including being the main subdivision of the formation) was the rough equivalent and theoretical basis of forest cover types. This seems to have been a widespread perception during the formative years of the Clementsian concept of association and the invention and adoption of vegetation cover type. This is indicated by the fact that forest cover type corresponded to plant association in the common textbooks of the day.

A better “indicator” that vegetation types as used in America corresponded to Clementsian units of vegetation is to be found in the “Grazing Types” subsection of the “Grazing Indicators” section of Plant Indicators (Clements, 1920, ps. 270-283). Clements grew up on the central continental grasslands when virgin sod was still widespread and he was educated at the great prairie land-grant University of Nebraska which became the nucleus of the “founding school of American plant ecology” (Tobey, 1981). As such Clements was much more a rangeman than a forester. Predictably he dealt with grasslands in more detail than forests. (This was also the case in Bio-Ecology co-authored with Shelford from the adjoining prairie state and land grant University of Illinois.) In Plant Indicators Clements did relate grazing types to formations and associations in the portion that preceded forest types. It is logical that foresters like Baker, Korstian, and Toumey drew conclusions from the preceding and more detailed section and applied it to forest types. Thus while Clements did not relate forest types to associations and formations, foresters could have made this rational adjustment as a logical and consistent extension of Clements’ more comprehensive treatment of grazing types.          

It is obvious from the “Grazing Types” portion of Plant Indicators that Clements’ grazing types were of different scales and that his grazing types corresponded to both formations and associations (Clements, 1920, ps. 273-275) as well as at the smaller scale (s) of  “local grazing types” or “practical grazing types”.  This was written so as to leave no doubt:

“Local grazing types.--While the main grazing types, such as the formation and association, indicate the comparative value of great regions, as well as the groupings possible in any one, it is the local groupings which determine the carrying capacity of a particular ranch and the proper system of management to be employed upon it. For this reason, they may well be termed practical grazing types. In areas relatively uniform, a single grazing type composed of the two or three major dominants of the association may cover a wide extent.”

Clements ended this portion discussing the faciation and the facies, the subunits of the association and the associes or consociation and consocies, respectively. It is clear that Clements used grazing types as he did vegetation types in a generic, multi-level/multi-unit sense. The careful reader can also detect in the local grazing types the forerunner or germ of the smallest unit or kind of range, the range site. Clements was always ahead of his time and his influence can be found in almost all major concepts dealing with units and analysis of range and forest vegetation.

In this period when Range Management was emerging as a profession built on the infant discipline of Range Science, Frederic Clements was most active as an author. Like Sampson, Clements was an ecologist and a researcher (an “escaped professor” as he dubbed himself). Clements and Sampson had nearly identical, as it were, backgrounds and academic “pedigrees”. Both men were born in Nebraska (Clements 10 years earlier) and both were products of the Botany program directed by Dr. Charles E. Bessey at the University of Nebraska. Clements received his Ph. D. from Nebraska in 1898; Sampson earned his MS from Nebraska in 1907 and his Ph. D. from Johns Hopkins and George Washington University in 1917 (Tobey, 1981, ps. 238-239; Casamajor, 1965, p.115).

As an author of ecological monographs and scientific papers for prestigious journals Clements’ titles did not readily emulate the land grant ideal of “practical education” (from the Morrill Act) or an industry orientation. Such appearance was misleading. For all of Clements’ elaborate, sweeping concepts and classification of vegetation, he always sought to apply theory to on-the-ground practice. “It is believed that succession and indicators constitute the most essential and useful form into which the results of research can be put for practical use” (Clements, 1920, p. iii). Clements (1920) wrote Plant Indicators- The Relation of Plant Communities to Process and Practice, his second monumental monograph as “a companion volume to ‘Plant Succession’”, aimed at application of the science of Plant Ecology to the arts of farming, lumbering, and range-based ranching. The “indicator concept” was application of the ecological viewpoint that species of plants (and formations, associations, consociations, primary seres, and secondary seres) indicated certain conditions of climate, soil, etc. as well as human modifications such that there were “agricultural indicators”, “forest indicators”, and “grazing indicators”. In Plant Indicators, for example, is to be found the genesis of the Dyksterhuis (1949) categories of decreasers, increasers, and invaders in the concept of seral types as grazing indicators:

“In the grassland climax, where the effects of overgrazing have been most studied, it is possible to recognize three or four stages. The first is marked by the decrease or disappearance of Stipa or Agropyron, or of both of them, and the corresponding increase of the short-grasses wherever these are associated; the second stage is characterized by the greater vigor and abundance of the normal societies, as well as by the increased importance of some; the third stage begins with the replacement of the grasses by annuals, while the fourth is marked by the spread of annuals and of introduced weeds generally over the area….These four stages indicate so many primary degrees of overgrazing, while minor degrees are denoted by the dropping out of particular dominants or subdominants…. Palatability is the chief factor in determining the sucessive disappearance of species, and hence the indicators of the corresponding degrees of overgrazing, though the sequence is often disturbed by the vigor of certain dominants. Since there are few species that are wholly unpalatable or inedible, it becomes possible to construct for a particular community a complete sequence of indicators, reflecting each appreciable degree in the process of overgrazing” (p. 298). [Note here even the importance of “vigor” as a factor in considering the ecological status of plant communities.]

Clements (1920, ps. 298-308) then followed this with discussions of changes in range plant communities with overgrazing and lists of species indicating degrees of overgrazing (departure from climax). He discussed subshrubs, cacti, shrubs, and annuals as indicators (eg. “Annuals are typically indicators of serious disturbance, and hence serve to mark the existence of serious overgrazing when abundant. They are the universal pioneers of secondary succession.” [Clements, 1920, p. 301]).

Section VI. (Grazing Indicators) of Plant Indicators (ps. 270-335) was a mini-textbook in Range Management. It included subsections on “proper stocking”, “range improvement”, “eradication of brush”, “water development” “rotation grazing”, “herd management”, and even a line diagram showing arrangement of corrals and sheds. This section is as much a manual or “bible” on Range Management as was Jardine and Anderson (1919) or Barnes (1913), but it was within a massive scientific volume of Plant Ecology or what could most precisely be termed Vegetation Science.

Like Sampson, Clements combined science with practice to form the fra