20 THE BRAUN-BLANQUET APPROACH ‘Vieror Westuorr axp Eppy van pir Maaret. Contents 20.1 Introduction. es 289 20.2 History... 200) 20.3 General Concepts . te 293 20.3.1 Concrete vs. Abstract Units |.) 1 ae 208) 20.3.2 Floristic-Sociological Classification Units... | 295 2033 Diagnostic Species... . . 2. 296 20.3.4 Sociological Progression... 2... 1... 298 203.5 Natural Classification ee e209 20.4 Analytical Reseatch Phase . . 6 301 20.4.1 Reconnaissance and Choice of Plots . 20.4.2 Boundaries. . Lee 204.3 Minimal Area and Plot Size |... 20.4.4 Description of Structure and Strata 20.45 Floristic Analysis. 20.4.5.1 Cover and Abundance. . 20.4.5.2Sociability ... . . . 20.4.5.3 Vitality and Fertility. 204.5.4Periodicity, 2. 20.4.6 Relevé Protocols... 20.5 Synthetical Research Phase Sees 3B 20.5.1 Steps in Synthesis... a 33 20.5.2 Primary Table... : 2 315 20.5.3 Presence, Constancy, Homotoncity Pile. 316 20.5.4 Species Weights. Fee 319 20.5.5 Table Rearrangements... 2... 320 20.5.6 Mechanical Means for Rearrangement 323 20.5.7 Determination of Fidelity. 2... + 324 20.5.8 Geographical Effects on Fidelity Fs 326 20.5.9 Phytocoenon Characterization tes 328 20.6 Syntaxonomical Research Phase... . . . «329 287206.1 20.62 2063 20.64 2065 2066 20.6.7 20.6.8 20.6.8.1 0.6.8.2 206.9 207 20.7.1 20.73 2073 30.74 20.7.5 20.76 20.7.7 20.77, 20.7.7 208 208.1 208.2 2083 2084 20.85 20.8.6 2087 2088 209 209.1 209.2 20.10 20.11 20.12 288 Syntaxon Table Characterization « Characterization of the Asociation << Geographic Problems with Associations Higher Units (Alliance, Order, Class). Units above the Class te Horizontal and Vertical Classification Lower Units (Subassociation, Variant, Facies) Nomenclature Construction of Names. : Validity, Changes, and Authors” Names Scheme of Syntaxa, 7 Extension of the Approach, Other Geographic Areas Biotic Communities ‘Microcomminities and Synusiae Community Complexes Applied Phytosociology Structural Considerations Indicator Groups at Indicators for Classification. | | Indicators for Gradient Relations Numerical Techniques Storage of Relevés. Species Correlation, Fidelity Tests Relevé Similarity ee Measurement of Homotoneity. Numerical Clasification Table Rearrangement ‘Numerical Syntaxonomy. Hearest Ordination. Informal Ordination : Formal Ordination © 222). | Conclusion ‘Summary - Additions to the Second Edition | ec 20. THE BRAUN-BLANQUET APPROACH 20.1 Introduction ‘We shall, in this last chapter, treat the florstie-sociological or Brauy-Bravquer approach to classification and interpretation ‘of communities. Before elaborating the details of the approach, we may state its esence in three ideas G@) Plant communities are conceived as types of vegetation, recognized by their floristic composition. ‘The full species compositions ‘Of communities better express their celationships to one another land environment than any other characteristic, (ii) Amongst the species that make up the Goristic composition ‘of a community, some are more sensitive expressions of a given relationship chan others. For practical classification (and indica tion of environment) the approach seeks to use those species whose ecological relationships make them most effective indicators; these fare diagnostic species (character-species, dlferential-species, and con stant companions) (iii) Diagnostic species are used to organize communities into a hierarchical elasifeation of which the association is the basic unit. ‘The vast information with which phytesociologists deal must, of necessity, be thus organized; and the hierarchy is not merely necessary but invaluable for the understanding and communica tion of community relationships that it makes possible ‘The reader will sce how far the elaboration of this three-part theme has led the members of the ‘school of Baavx-BLANQUET.” ‘The results are (like community relationships) complex. We fear lest the reader should lose sight of the heart of the approach — the floristic perspective — for the abundance of technical details that follow. The latter are only the skeleton of the approach. Further ore, if we do not press the physical comparison too far, its body is the stillegrowing corpus of basic and applied research and the resulting knowledge of and insight into communities. This review ‘must deal mainly with the skeleton of the approach; we cannot in the space we have discuss adequately the kinds of understanding, ‘of communities that have come from it, We ask the reader (0 recognize wherein this account must stop short, and to seek if he is interested a further fecling for what the approach offers by observing or reading studies applying it. On the other hand, we 289shall pay some attention (especial recent developments in phytosociology that are not part of che Bravy-BLanguer approach as such but may be of value since they Introduce concepts and methods that relate the floristie-sociological approach te ather approaches, ‘Textbooks and reviews an the floristic-sociological approach have been published in most European countries, Argentine, India, Japan and the US.A. (see the bibliography by Maske, Tosi & Westuorr, 1970, Exc). Recent accounts include Gousor (1969), Westuorr & Hit (1969), and Kware (1971). Useful accounts in English are Becerye (1957), Warrraxer (1962), Klute (1967) and Siomwunt (1972), 20.2 History The origin and history of the approach have bees reviewed elsewhere and can be stated only briefly here (sce Gaws 1918, BrauxeBrasguer 1921, and Du Ruerz 1921, a well as Brcxine 1997, WureraKee 1962, and SiosweLt. 1972). The systematic de- scription of plant communities and the idea of community types can ‘oe waned from great students of plant_geogeaphy, Hewounr (1805), Sewovw (1828), Hsex (1835) and Grisenaca (1838). From their work two main approaches developed, the physiognomic and the floristic. The physiognomic approach developed 28 1s principal unit the formation, a unit characterized by physiognomy or vegeta tion structure (Post 1862, Grsepact 1872, Waxatne 1895; see article 13). Through the work of students dealing with smallerscale tunis (esp. Txcog 1844, 1835, Truretann 1815, Lorenz. 1858, Kenwen 1063 and Davbe 1890, in southern and central Eueope, Pose 1882, Hct 1881 and CAJaxbER 1903 in northern Europe) there developed the essential idea of the florstic-seiological approach: plant communities as units of clasifieation based. primarily on Species componition. From Loxixz (1863), Moares (1877), and Dame (1908) a parallel approach to biote communities can be traced, (See BaLoou 1958 and Wisrrraxen 1962 for review and references.) Much of the further development leading to the Baste Bianguer approach was eentered in Zarich (Stauux & ScunOren 1003, Scunomen W804, Scunoren & KinexiveR 1902, Beocxwea °) This and many other bibliographies to be mentioned inthis gomsbtion| have been published in Eseaple Banca Sc, 8 Scloiee (R. Txex ¥a)- They wil ot belted i the ferences, but indicated with “Exe 290 Jenoieu 1907) and Montpellier (Fianaver 1893, 1898, 1901, Paviniarn 1901, 1912). A main outcome of this period was a full hierarchy for vegetation classification —vegetation-type, formation group, formation, subformation, stand-type down to locab variants [Nebentypen} and geographic variants (Fazies). ‘The stand-type \was ealled the ‘association’ and considered the basic unit. FLABAULT & Scunoren (1910) agreed on the following definition, presented to the Third International Botanical Congress in Brussels, ‘An as- sociation is a plant community of definite floristic composition, presenting a uniform physiognomy, and growing in uniform habitat conditions. The association i the fundamental uni of synecology.” ((ranslation following Paviztarp 1935). From this background BravN-(BLaxguer) carried out a monographic study of alpine vegetation (1913). In this, and the cesay by Bravy-Braxguer & Foaee® (1913), attention was focused ‘on ‘Charakterpflanzen’ or character species species that possess fidelity’ (felative restriction) ta given asociation, GRaMANy (1.909) hadl advocated the approach through floristie composition, and the use of character species (‘Leitpflanzen’), The key ideas in Bravy-Buaxguer's treatment were: (2) The study of communities should be based on a fundamental unit, comparable te the species, Gi This unit should be the association, and associations should be defined by their possession of character-species. (jit) Each aswocia- tion consists (like a species) of “individuals,” and the association (like the species) ean be described from samples of its individuals liv) Bach sample (‘Aufnahine,’ relevé) should be chosen so as to represent adequately such an individual, and i¢ should include analysis of the complete species assemblage. (v) Associations should be grouped into higher unils not by phiysiognomy, but by floristic composition. ‘Additions to the approach came in further publications. BRavn- Branguer (1915) added the ‘Assoziationsgruppe’ (later called the Verband, alliance) as a unit above the association, also de- fined by character-species. BRAUN-BLANQuEr (1918) added the sub- association, as a deviation from the typical association expressed in a constant floristic difference, and the facies as a subordinate unit possessing merely quantitative differences. BRaus-BLANQvET (1921) outlined essentially the full system, including the analytical seales (204.5) and the ‘sociological progression’ (20.3.4) for ay- ranging communities by their levels of organization. Beavy- Buangver (1925) pressed the claims of fidelity or ‘Gesellschafts- treue’ as the key to vegetation systematics, against the disagreement of northern schools, In the same essay he introduced the ‘charac teristic species combination’ (charakteristische Artenkombination 201after Saran, 1923) as the ultimate community diagnosis, comprise ing both character-species and constant companion species. Koc (1925) and Braus-Braxgus & Jexwy (1926) added the concept of differential-pecies for the characterization of subordinate units, fand the order a8 a unit sbove the alliance. BRAUN-Braxguer & Paviztarp (1922, 1925, 1928) in their Vocabulaire codified the analytic and synthetic procedures of the approach. Finally the Science of plant sociology and its application were spelled out more fully in the first edition of the textbook of BravN-BLANovxr (1928, 1932).!) From these beginnings the influence of the approach spread through western and central Europe, We cansot review that spread, but will mention suck leaders as Pavitiagp, AtLoRGe and Mouxte® in France, Tox and OneRvorre in Germany, Szarsn and Paw= zowsa in Poland, Funner in Switzerland, Soin Hungary, Kxsxain Cerchoslovakia, Horvat, HorvaTié and Weauee in Yugoslavia, Boxza in Rumania, Giacowint and Towassiits in Haly, BoLos in Spain, Leoron in Belgium, and oe Lewuw in the Netherlands, References for these authors are foundl in BRaus-Branguer (1964, 1968). Braus-Braxguer’s own work began at Ziirich and com- tinued from 1927 onwards at Montpellier, where he still Ieads the S.1.G.M.A., Station Internationalede Géobotanique Méditerranéen= neet Alpine. A second centre was established 1932 atStolzenau, West Germany under Reinhold Toxex, a pre-eminent leader of the ap- proach and former director of the Zentralstelle (later, Bundesn- stalt) fr Vegetationskartierung, who has done much to give the approach its value in application. From TOxen’s centre the ap- roach spread to various non-European countries as well, and notably to Japan, During its spread the approach has been known by several names: French-Swiss (or Swiss-French) school, Zirich-Montpellier school, Middle Buropean-Mediterranean school, and Sigmatism (Tanstzy 1922, Dv Ruerz 1936, Beavx-Braxguer 1953, 1968, Eoxus 1954 Beextse 1957, Wurrtaxee 1962, Toxew 1969). Most of these terms lack specificity or are unclear. Moreover we feel that the time has passed for the word ‘school with its implications of a fixed system, We prefer in this article the most direct name, vi Bravy-BuaxgueT approach, whilst we may characterize the es- sential ideas in the tetm, the ‘floristic-sociological approach.” °) One notes dough this development of the full range of concepts the continuing contribution of HauseBeange himself chat has geen the ap= proach is present character (Gaus 1972 tthe contest) 292 203 General Concepts 20.3.1 Coxensire vs. Ansreacr Ustts ‘Much dispute has centered on the nature of the plant commu nity, This will be briefly reviewed as far asthe Baavx-Braxours ap- proach is concerned, (See further Pavizr arb 1935, Westhorr 1951, 1970, WisrrraKer 1962, Braux-Branovet 196+) ‘Concepts of the plant community include: (i) the organismal concept (Crranewré 1936, Taxsuiy 1920): the coramunity as a “superorganism.” (i) the concept of social structure (Paczoskt 1930) and many early Russian authors such a5 Suxatscatew (1929). (ii) the individualistic concept (Gruason 1926): the community as a changeable mixture of ‘individualistically’ distributed plant species. liv) the concept of population structure (WarrzaxeR 1953, 1962, 1970): the community asa system of interacting species and vege tation as a complex popalation pattern. The BRAvs-BLaxguer approach takes a practical, inter mediate position that recognizes the heterogeneity of species di tributions but emphasizes nonetheless the interactions between plants in the community, which has a certain individuality because ‘of relative discontinuities between communities in the field. Def initions of plant community or phytocoenase range from the more superficial, ‘any collection of plants growing together which has fas a whole a certain unity’ (Tasstey 1935), to the more pro- found, ‘a plant community (-animal community — biotic com- ‘munity) isa working community: the species composition of which is in a sociologic-dynamical equilibrium in competition for space, ‘minerals, water and energy, in which each component affects all others and which is characterized by harmony between environ ment and production and phenomena expressing fe in form, colour and temporal course’ (Tuxtx 1957). ‘As shown by Txrx’s definition, the community concept may be broadened from the plant community to the biotic community of plants and animals: producers, consumers and decomposers, Yn a further broadening of perspective the biotic community (= bio- ‘cocnose) plus its environment or biotope is treated as a functional unit, the ecosystem. The concept of ecosystem has become most familiar from its expression in English-language ecology by Taxs- tev (1935), but the concept is largely identical with Fraperuc’s (1927, 1958) ‘holecaene’ and the ‘biogeacoenose” of SUKACHEV (cxg. 1954, 1960) and other Russian authors, Though by far the greatest development of the Braux-Braxguer approach has becu 293in application to plant communities, biotic communities are clearly amenable to parallel study (section 20.7.2). TOxes (1957, 19656} hhas emphasized this possibility and ‘stressed the influence of Frieoentens and Tanexeman (e.g. 1956) on his own views ‘The term plant community and its German equivalent ‘Pflan- ‘zengesellschaf have been used in both conctete and abstract senses, which has caused dispute and confusion (gee Taxstey 1920, 1935, De Rirez 1921, Aueciun 1926, Pavctano 1935, Westuorr 1951, 1965, WutrTAKER 1956, 1962). The floristic-seciological approach hias always stresved the distinction between the concrete and the abstract community, The classification units were, naturally, ab- Stract units, Concrete plant communities were often, especially in forestry, referred to as stand (Bestand), whilst concrete representa- tives of associations were even called association-individuals (PaViL- ano 1912, BRaun-BLanguer & Pavintaxo 1922), This term has been much disputed and criticized (most thoroughly by WarrTaKu 1962) and we consider it now as of only historic interest. ‘Westiorr (1951, 1965) has acknowledged the distinetion be- ‘ween concrete and abstract communities in separate definitions, which were linked to general definition of vegetation. For the concrete plant community the term phytocoenose (Gams 1918) was proposed, for the abstract community the wrm phytocoenon (MAAxer 1965, specifying the general term ‘coenon" propased by Bankwan et al, 1958, Wesrsoss et al. 1959). ‘The term {phyteleaenon may be proposed as a suitable inter= national term ‘which may replace the terms community-type (Warrraxex 1956, 1962) and nodum. The latter term has been sue gested by Poort (1956, 1962), but has been used in @ more spe cific meaning by Wintsaws & Lauper (1961) in their “nodal analysis’ (cf Ivimy-Coox & Proctor 1966). For biotic communities parallel terms: biocoenase and bio coenon may. be used (Maanet 1965). For partial communis Moxze Bruyns' (\950) term merocoenose can be used with ‘merocoenon as the abstract equivalent. Layer communities could be called, stratocoenose-stratocaenon. For specific ubcommunities con~ sisting of plantsof the same stratum, lfe-form and seasonal xelations the couple, society-synusia may be reserved. (See further article 16.) In conclusion we present Westuorr’s (1951, 1970) definitions in a slightly adapted form. Vegetation is defined as a system of largely spontaneously grosing plant populations, growing in co herence with their sites and forming an ecosystem with thesesites and all other forms of life occurring in these sites. Thus are excluded all, asemblages of mobile plants and collections of plants growing in arrangements set up by man such as flowerbeds and arboreta.) 294 A phytocoenote is defined as a part of a vegetation consisting of interacting populations growing in a uniform environment and showing a floristic composition and structure that is relatively, ‘uniform and distinc from the surrounding vegetaticm, A phytocoenon is defined asa type of phytocoenose derived from the characterization of a group of phytococnoses corresponding with each other in all characters that are considered typologically relevant. 20.3.2 Fronistie-socrorocicat. CLassiieation Ustts Plpiocoena include such various kinds of vegetation units as formations defined by physiognomy, dominance-types defined by major species, forest site-types delined hy undergrowth composition, and noda derived by quantitative comparisons or numerical pro cedures. The approach of Braun-Brangver has its own, formal hierarchy of waits, of which none of those just mentioned is & part, ‘The fundamental unit of the hierarchy is the association, a unit that corresponds in function 10 the species as the fundamental unit of ‘diotaxonomy, or the classification of individual organisms. The word ‘association’ has had a long, complex, and argument-aflicted history as different schools sought to determine its meaning to their preference (Warrraxen 1962, Sumwets. 1972, sce also articles 14, VF and 18)- In an eatlier definition by Buavx-Branguer (1921) ‘the association is a plant community characterized by definite floristic and sociologiest (organizational) features which shows, by the presence of character-species (exclusive, selective, and. pref erential) certain independence.’ Mryjer Taxes (1951) defined. the association as a plant community identified by its characteristic taxon combination, including one or more (local) character-taxa, or differentiating taxa,” A similar conception was agreed on during a Symposium on Plant Sociological Systematics at Stolzenau in 1964 (in Texen 1968b, especially OnemborFER 1968) and a later colloquium at Rinteln (see Diexscuxe 1971). We shall retura to characterization of the association (20.6.2) but emphasize that, in keeping with the floristiesociological perspective, the association fn defined by its characteristic species combination including character- and differentialspecies as well as companions (Begleter) with high presence values (over 60%). The Sixth Botanical Congress at Am- sterdam, 1935, accepted definition of the association by charac teristic or differential species in the sense of RRAUN-BLANQUET as fone of three resolutions given in article 18.3.6. 295As basic units associations, like species, are to be grouped into a hierarchy of higher units. Associations are classed into alliances, alliances into orders, orders into classes, and classes into divisions (20.6.4-3); associations are divided into lower units ofthe hierarchy. (20.6.7). All these units are coena as defined above; but the coena of the formal system of Beavr-Buaxquer may be termed syntaxa (Barxaax et al. 1958, Wesrmorr etal. 1959). Thus the parallelism in the classification of organisms and phytocoenoses is realized: to the species as a fundamental unit corresponds the association, t0 the taxon for units on any level corresponds the syntaxon, to (idio)taxonomy for the practice of classification. corresponds syntaxonomy, to (idio)systematies as the broader study of relation- ships among organisms corresponds synsystematics, 20.3.3. Diacwosnie Seeciss Syntaxa are defined by diagnostic species (character-species, differential species, and constant companions). Characterspeces are species that are relatively restricted to the stands (or samples) of a given phytocoenon, and therefore characterize it and inelicate its environment (Fig. 1). In ideal, a group of character-species is ased for the characterization of a syhtaxon of the BRauN-BLaNquer clase sification on any level from the association to the class. Ta serve as character-species for an association, a species should have a relatively narrow distribution even if it is not simply restricted to the as sociation. (Degrees of restriction to @ given syntaxon, which are termed degrees of fidelity or Treue, ate discussed in 20.5.6.) Note that the concept does not say that the species need be important in hytocoenoses; very minor species may have diagnostic val ‘The German term is *Charakterart; this has been variously translated into English as characteristic species (BRavn-BLaxovet 1932), which seems unsatisfactory, faithful species (Poors. 1955a, Banxnan 1958a, Brarrins 1962, Moors 1962, Wesvutorr 1959), and character-species (Becxive 1957, Warrraxzr 1962). The latter, ‘most direct translation, is our proferenee. HziMaxs (1939) intro- duced *kensoort’ in Dutch to avoid the germanism “karaktersoort’; and the Dutch kensoort returned to German as Kemart (TOXEN 1950: 99}. Most phytosociologists writing in German now use that term, which is accepted in the third edition of BRAwN-BLANQUEN’s (1964) text. Syntaxa are most often defined by the fundamental units of idiotaxonomy — species —, but this is not always the case. Sometinies plant subspecies, varieties, or ccotypes may contribute to the definition of lower-level syntaxa or geographically vieariant 296 [Associations “A—+f s——__|c |"zabassocations cee Dee: secre: wet moist | fresh | dry (| very dry Fig, |, Diagnartie species along a moisture gradient. Species 1 and 2 {i characerayece or asocaton Bad ave thei populations ecatered in {er lrgely confined ty tha anoclcio, Specie 3 and 4 are characterspecies fer association dy and species 5 2 characterapeciea for socation Species ‘and 6 ate diferentnbspects for subawoiation s of asacation B, and species 4 ad 8 are diferentia-speciee fr subasocation of asecationB. Tn each ease presence ofthe diferentaspeie ditnguishes the mont or dry, subasocation From the “typical” (fesh) subasocistionb. Species 9s amore widely dstebuted species that right help characterie awoeiation B at constant companion for ‘Rat awociation, Species 9 rnght aso be chacacte-apecis for & higher ye taxon, such as an aliance uniting svocations B and C with other associations. asocations. Sometimes a genus or subgens may be used to define a higher synlaxon, 26 the alliance Spartinion i defined by the Fidelity of any speics of the genus Spartina (Beerrasx 1965, 1968, see also Pioxar 1968272, Wesruoreapud Pionarrt 1968a: 74) 1t ie therefore appropriate to broaden charactorspecies to character taxon (Banasan eal. 1958) as we shall do here following Dutch tse of Kentaxon’ (Wasror® & Hato 1969). The German equlv- alent is ofcourse ‘Kenntaxan’ (Gharacter-species charactrize symtaxa by thelr normal oc- currence in phytocoenoses of that sytaxon, contasted with ther Absence less requent occurrence or smaller total eatimate in phyto- Coenoss fall ther syntana tis powablealo to dng sly related syntaxa by the presence and absence ofeersin species withost Concern for the Dronder distributions of those species Two subess™ ions of a association may be characterized by the fact that Samples of one normally include species #6, whereas samples of the ether normally lack these speces (Fg. 1). These species function as dif sees that distinguish the 00 subassciations. Di 297fereniabapecen, a the second class of diagnostic spect, define symtasn on the bis ofthe diaributonal boundaries of epeces (withoun regard to fdeliy Yor the oytana in queaon) an ne tbe primo to define lower symana Since bing introduced by Koen (1925) and BratneBranguer (1925) a» ‘Dillrenlare, the concept has ensred Unghie eae fered species Baava-Buawany 1982, Wirvanee. 16d) oe aitereniting spcin (Bncarse 1957), whilst in German te ter Tremart has tome igo general use (TUxen 1050; 195% Biaore Banque 1958, 1964) in parallel wth Rennar, The se of te ferenalspecis has been urther developed. by, Senwixensnn (1531, 1942, 1954a) and “Ux (1987, 190, 1052) Dilreatal speci and Trennart ae synonyns ard thee ae ngely ee tno luo with aifrentitog speci; ut we shall Here ee ee feremlating spies Tor the mote informal wee of dlagronte pcs of unassigned syntaxonomie postion in the saute pies ot rescarch 208), and speak of aiterontebacer ane ae Species group forthe formal characereston of oensce one Aurtherrecem development the concept ot Coeteae oo bination” (ater the Duteh Kencombinatlootendceal here wort & Hit (1969) following Buurtive (106s, aioe iy species combination). The exemial concept the exeieonea ee isyntaxon ofthe combination of spetes or tana) slit see 6 them need be a character-taxon. ae : " ‘Aird group of species wit diagnose value ate the constant companions (20.8) which eceur it most tenés off one bat ate not desgoated as character or ailfreniatapectoy Ce slant companions are added to the charactertoss 6 formas ‘charatertieapecles aon) combination Br soca higher taxa. Altiough Beaus-Brangue® (1985) Mneloeed se concept rater cary and repeatedly fe. 1959, wee Moos 1963) Senet elevance there has been debt amongst ete oe Poo. 1955150). OF couse, as Baaus- Bangor eae 1992. 08 remarked or auociationg, these yntana are aed beat winch pouest a Nigh proportion of bout characte ‘and consoet Sede. 20.3.4 Soc GICAL PROGRESSION Bravn-Bianguer (1921) sketched an ‘arrangement of the plant communities according to theit sociological progression.’ The idea of this arrangement was agzin a parallel with idfobiology, specifically with the arrangement of taxonomic groups by evolne 298 onary level. Levels of structural and organizational development were chosen as criteria for arranging phytocoenoses from moving aaero- and hydroplankton communities on the lowest level to the tropical rain forest at the highest level. Other communities are ar- ranged between these according to increasing stratification, com- plexity, presence of dependent communities, species richness diversity of growth-forms, and assumed intensity of species teractions, Thus Braus-BLaxguer (1921) summed up many of the structural-suecessional trends much later commented on by Onum (1969)! The scheme was later modifed in detail, and increasing stability was added as a criterion. BRawy-BLaNover (1964) ar- ranges higher syntaxa (classes and orders) in the progression, ‘Wacxer (1968) called che sociological progression a ‘purely artificial division principle’ (cf. Etrexere 1963). Of course, no phylogenetic meaning should be imputed to its sequence. Yet the sociological progresion is useful, as a principle at right angle to the levels of the hierarchy, in aeranging syntaxa in surveys. Its wee is illustrated ia vegetation monographs by TUxex (1937), Onex- ponren (1957), Entespere (1963), Onerporrer ct al. (1967), Westiorr & Hao (1969), and others. It may be of interest that ithas appeared ako as an axis of broad-scale ordination of syntaxa (Maanit 1972a, see 20.9.2). 20.3.5. Navunat. Crassrication ‘The problem of ‘natural’ classification has been discussed in another article (12.1.2). It has been accepted by Bkauy-Biaxguer and others that associations are abstract units. A degree of con- tinuity or gradualness in transitions between phytocoena is rec» ognized, but considered not to preclude classification. ‘We are convinced that the plant cover of the earth in all its dimensions can be divided into phytosociological groupings of higher and lower rank; their delineation may be either sharp or less sharp and gliding.” (Bxaus-Brasguer and TOxry in comment on Goooaut 1963, translated.) ‘American research in gradient analysis has led to interpreta tons that may’ seem in conflict with those of Brauy-BLanouer. The American work of Warrraxee (1951, 1954, 1956, 1962, 1967, articles 2 and 3) and Corrs (Cortis & Melvrosu’ 1951, ‘Corns 1959, MeLvrasut 1967, article 7) developed in parallel to Russian work of Raazvsey (1930, article 17) toward the concept expressed by Wutrraxer (1970) as ‘the population structure of vegetation.” In this concept the population structure of the individual phyto- 299cocnose may be analysed through dominancediversity and specer- importance relations, and niche diferentition among plant species expressed in stratification and fuseionaldiferences among then, Thephytocoenesisthen conceived asa system ol interacting nce: differentiated and partially competitive species” The vegeon of an area, on the other hand, sto be approached through gradient analysis, studying the manner in whith speces. peptlatens are distributed. along environmental gradient and combined into phytocoenons. Species ae lferently distribute from one amotio, $0 that undisturbed phytacoenosesincergrade continvonaly fy mont svc nel fo pre ehuronmental aden co munities form complex ad largely continuous population pateres? (Warrraxer 1970, atile 3.9.) ae . Ics farther recognized that, Because of environmental inter ruptons and some relative discontinuities inherent in vegetation itself, the pattern may ako be considered a complex mixture ot continuity" and relauve discontinuity” (Wurraxen 1036: 33) Vegetational continuity by no means preciades well clasification (Wurreaxsi 1956, 1982, 1970) Gradient anayss in of vegetation, regarded as a coherent. patie. ofintergrading phytocoenosey, diferent om that of phytxociologiss who tend tosee it in tens ofa typology of phytocoena (sce, however, 20.74 en complexe}. Yet we think tat the ference betyeen te concept of the gradient approach and BratseBussgcer's clearly ope of Alegree ~ of emphasis of continuity vs discontinuity where both ae present, of specie individuality vs, species groupings where both te realise, and of gradient analyis vs. clasiieation where both ae posible: We judge the diferece between less extreme students of gradient analysand of the DkavseDuasguer. approach, te bye one of emplhais and perspective, not one of Tact or under. Given the individuality ofspeces distributions and some degeee of continuity between phiytacoenoses, clasifeaton is not stly tural in the sense defined in ardcle 121-2. BaawseDLangune ((95ta: 561, 1959: 147) has considered argument onthe nateraness of elasifcaton poids. TOxus (1955), in response to a eici of Euiesmene (1054b), interpreted phytosocclogical clasieaton through the concept of ter as deal Concept, recognized in an Eanpiical way trom “ortelation concentrate Le- groups of cor, related characters, ‘That which is evident and characterte of 3 ‘ype is always it nucleus, ot ie periphery; types are not pigeon. Hotes but foe ina field ef variation. Geant (1968, se also Rac scitext 1969) clauorated the concept of vegetation type, dating ing as its three aspect (i) The vegetation type as Henly: tepet= 300 Lions of certain observations are approached via intuitive integra ‘ion, resulting in vegetation-type concepts based on recurring combinations of species. (i) The vegetation type as maximal cor relative concentration: The joint floristic-sociological and ecological approach leads (© types as correspondences of recurring species combinations with recurring combinations of environmental fae- tors. (il) The vegetation type as systematic category: A hierarchic system is feasible as a result af an integrative, induetive process ‘grouping the initial types into higher and higher ranks on the basis ‘of common species combinations. The similarity ofthis statement to the theory of elasification developed at greater length by Warr crake (1962) may be noted. Relative naturalness of classification in this perspective is to be judged by success in embodying the maximum number of significant relationships among phytocoena and species in the structure of the hierarchy. We assert that cam- pared with other classifications the Bravs-Branguer approach, ‘with its floristic emphasis and maximum use of species distributional relationships for classification, does not fll short. 204 Analytical Research Phase 204.1 Ruconnatssace AND Cxtorce oF Pots An analysis of the vegetation is as a rule preceded by a pre« liminary survey of the area, when this is lie known to the in- vestigator. This reconnaissance (see Cans & Casto 1959) includes a study of the general vegetation pattern, and the establishment of the apparent relations of the various vegetation types with geology, topography and soil conditions. The next step, ‘primary survey,* inchiding a superficial description of the major communities, i mostly passed over in the Braun-BLanourr approach, A vegetation analysis starts with the choice of stands (phyto= coenoses) on the basis of the reconnaissance. Within the stand one sample plot is laid down, often covering a large part of the stand, ‘The analysis of this plot is called the raeeé. The Baaun-Banourt approach has often been criticized for the subjectivity of its sampling procedure. However, subjectivity of stand choice must be accepted in the procedures of many empirical sciences. A selection of relevés is desired chat will effectively represent the variation in the vegetation under study, the samples being so chosen that they will not represent different phytocoena disproportionately and will not include mixed, incomplete, or unstable stands. For this purpose ‘of equitable representation of different kinds of communities with 301most useful relevés, a subjective, ‘stratified sample selection is fr Superior to sample choice by random points on map. Relevés may be ndevtaken without clssieaton as a purpose ‘They may then serve Tor the study of vegetation dynamics on a given area (perediiy, Muctution, or suceestion) or another Ecological purpose, e.the study of the ecology aa certain taxon ‘or ccotype by describing its pater in relation to the paler ofthe ‘egcttion. In ant cases, however, elves are intended to be sed fr some rt of station or ordination. In that ststion, the only preconception in this choice isthe demand for unfermf ofthe stands fee “Uniformity’ may ewer expres what is sought than Shomo- gencity? This chapter i not the place for a treatment of the mathematical aspeets of homogeneity in vegetations we may refer to Poows (1955, 1956), Eutenseno (1955), Datu (1987, 1960), Broxina (1957), Lanter Dave. (1064), Baavs-Beangury (1964), Manet (19665) and Gocnor (1969) for general conse eration, to Goopatt. (1952, se also his bibliography, 1062 Bee) Dan & Hapae (1949), Das. (1957), Gaase-Surn [1968 toe pects of homogeneity ta dhe distribution of plant individes, to Goopass (1961) and Graso-Surru (1968) fer patern seaa to Raonssase (eg. 1954), Guinocner (1955), Dale (1057), Gon & Gastuo (1955) forthe relation beeen homogenety snd Genaency distibution and to Cen (1959), Goonon’ (1960) sad Goons (1969) for infastand similariics and information measutcs ae spe roaches to homogeneity. “The fist condtion that no obvious structaral boundaries or variation in ratieation ate visible within the stand: The second sriterion is uniform foie compesition, It is stat Task for Joint patteros of dominant andor abuadant specie and then to Geimit a stand where qualitative ckanges in patterns occur, ie where one or more species drop ovt and others come in In any Case an experienced Bld worker is able to jedge this rapidly In many eases however the changes in pattern are quantative rather an qualtative. The species eompesin difew hitlefevs onesie to an adjacent one, bat the relative proportions of ab dance and cover do vty. Ie is usual vo make separate eleva in any ease where the abi habitat factors show a lear dacontnuty of at least 8 marked transition, “Many aces eae pattern heterogeneity only by thir growth form or thei aggregation or shoot clustering. In such eateone may ley at fin to study this bitiealy heterogeneous pattern sone single stand, ie with one relevé. However, the situation becames eileen assoon asthe erowdig of one major species leads tothe eaable 302 ent of one or more other species which appear to be leally hound Toit, A simple example is the establishment of dvear shrub patch na hericcous vegetation, eg. a clone of Sal repens in pen dune grassland, When one er more species are found particularly bound 1S that dvart shrub patch we prefer to consider Uhe latter a8 8 Separate phytocaenose which has thus to be analysed with & NParate rele. Hoveever,there-are exceptions to this rule. Ghusvonenx (i970, Canaries) and Wrxokn (1973, South Aifics) feport sable savanina mosaics consisting of @ dwarkshrub and frustund and a low tree and shrub phytocoenoye. Such mosais Se considered one vegetation type, since one docs not find leally tnore extensive patches of cther phytocaenese without adjacent patches of the other one In intermediate or doubtful cases, such as a swamp com- rnunity with tall socks of sedges (Care dion) alternating with srette hollows, i wll be advisable, at least when the typeof vege tion pattern i unknove, to follow both procedures, viz analysing the pattern as a whole as well a televés of tusocks and hollows Separately. Tull then tant out later which rclevé is more wsefl for the clasication purpose. 20.4.2 Bouxparurs As was said belore, the recognition of distinct stands may pre= suppose the occurrence of discontinuities in the field. Although 2c- cording 10 the opinion of Braux-BLaxguer workers such dis- continuities aze mostly to be abserved, it is obvious that boundaries Between stands are less sharp in some eases than in others. In such cases the boundary may be detected by means of a belt transect fnalyss or a ‘Tine taxation.” Such transects consist of a series of small quadrats laid down at right angles to the extension of the ‘boundary zone. By comparion of quantitative data on species oc- currences in the quadrats the boundary zone ean be discerned from the vniform phytocoenoses. Maarrt (eg, Maagen & Leem- ouwen 1967) and Farsco (1972) refined boundary analysis by constructing “aliferenial profile, in which can be shown degrees of change between adjacent quadrats along a gradient, For similar techniques see article 5.4.4. ‘In general we may distinguish between two types of boundary zones which are, according to the relation theory of LeEowex (1965-1970, Werenorr 1971a, b, Wastuorr & Lusuwex 1966, Stunts 1972), the limes convergens and the limes divergens. The limes conoergens one, oF convergent limit, i characterized by sharp 303| 304 Vegetation profile on limes convergens < diverges anes ( Leste 1965), Vegetation profile on limes divergens Vegetation profiles corresponding to the limes eonvergens and limes vegetational boundaries on either side of which uniform phytor oenoses occur, with few species represented, some by many in- dividuals, in coarse-grained patterns. The corresponding environ- iment is unstable, ie. with sharp fluctuations in vertical direetion (eng. water-table oF phreatic level) or in horizontal dieection ( deposition of flood marks). A typical synsystematical unit OF this type is the Agropyro-Rumicion crispi. The limes divergens zone, or divergent limit, is characterized by numerous small-scale boundaries in phytocoenoses merging con- tinyously intp each other, with many species generally represented by few individuals in fine-grained patterns. The corresponding environment is stable and determined by a gradient. A typical series of divergent synaystematical units iy Mesobromion-Tri folion medii-Carpinion bevuli. Fig, 2 shows a vegetation profile belonging to these types of oundary. According to Maan (e.g, 1966a), Leeuwen (1966) and Wesriorr (19713) the ecotone concept of Lavivostons. and Cuanets (e.g. WeavEn & Crests 2038, Corrs 1959) could be applied to the limes convergens zone, ‘whilst the ceocline concept as derived from Huxtxy’s cline concept (WesruorF 1947, Mumer Daas 1951) may be considered an equivalent of the Times divergens zone ‘The delimitation of stands within ecotones genetally gives no difficulties, except for pattern problems already discussed. Gradients ‘of vegetation-and-biotopes (ecoclines) can he approached by belt transects following the direction of variation, whiel éransects have to fe divided into sample plots (not necessarily quadrats) small enough to be uniform. Recently Jaxucs (1972) gave examples of this approach in ‘grassland-shrub-woodland” transitions within the range of the Quercetea pubescenti-petcacae in Hungary. 204.3 Masman Ansa ano Pror Size The size of the sample plot is largely dependent on the structure of the vegetation under study, but may be affected also by the size of the stand. In gany cases the stand is sufficiently small to be analysed largely in one relevé. In many other cases, however, this is not possible, and then the plot should be large enough that all species of regular occurrence in the stand should be present in the sample plos. Ta extzemely uniform phytococnoses which are very poor in species, eg. salt marshes, where differences in abundance and cover are of major concer, these dillerences should be Considered in establishing the size of the sample plot. 30‘These conditions refer to the minimal area concept (see also antcles 5.2.1, and 18.3.3). Without treating this concept in deta We may give some comments relevant to the phytocdenose de= scription, First the minimal area as an analytical concept {for the Phytocoenose) has to be distinguished from the symthetie minimal area (for the phytocoenon). The analytic minimal area should be defined for a stand under study as a representative area, eas an adequate sample of species of regular occurrence in the stand; the size decision may, but will not necesarily, be based on the total tuumber of specics in the stand or the wwual minimal area curve (eavs-Buaxguer 1964, Cain & Casrno 1959, Dv Risrz 1921, 1980, Goooat1 1952, 1961, Gueto-Sure 1964, Horxins 1957, Maanst 19666, TOxsy 1970e and 1970 Exc, Visran 1949). & synthetic minimal area, in contrast, is subject to many current definitions (eg. Bravn-Buangust 1954, Cain & Casto. 1989, and Westiore 1951 — who explicitly defined it as ‘the minimal surface which has as a rule to be occupied by a sample of plant community if the normal specific assemblage will be able to develop’). ‘The size decision in this ease involves not merely the stand itself but the ‘normal’ composition of stands representing. a Phytocoenon. Ie is clear that before judging this in the fed one should know from synthesis the characteristic species combination Nevertheless many authors, including Bratx-Brasaver (1964), describe the normal specis-atea determination, applied analytically to the stand, as the way to find the minimal area, ‘Tex (1970e) presented numerous examples of speciesarca curves ending in a horizontal part, implying thae at the sample size where the horizontal part begins fll species representation or saturation has been reached, Taxen consequently spoke of saturated ‘communities,’ which he found among homogeneous commurities of various onganization levels, He cited various authors who found ‘curves of the same type, including Du Rtevz (1921) Most authors, eg BRat-Branguer (1968), Cats & Castro (1959), Hoprixs (1955), and Maanat, (9663) have found curves hat never reach an asymptote. Te is (ue that_ many of thels curves refer to cither relatively small or relatively large areas. It remains unclear in which situation which type of species relation can be expected — the Anamentus(ype of # log-log relation (ef, Pustow 1962), the Roustictype of a logclinear relation. (ef Horkiss 1955, Dar 1957, Wititans 1961, Maater, 19660), oF the Kyumtype of saturation relation, Batoct (1958), and MaAREL (29566), quoting Frey (1928), who presented the three types, interpreted them as refering to very" uniform and species poor environments (limes convergens type!), highly variegated ‘and 306 asus 1 Minimal ares ves in square meters fr vais coments anc sommes Foetal mom communi Fapostylos poner commits (oct Nancjuees) See eeands (Rocke Crsnesbarete) Salman (Aner pth Parare (Lal Cynonrer) Nib nstaldone cannes (Ammophetes) Ir iendon (Astra) lay Nordea) Metts tneedv' and dare (Blyo-Sedeis) Gahran Shapurlsemperate sero subi (ie communis Seay Seb commune (Ramo Prete) cpp communes Fervece decd oes (QereoFagete) et dco re (North Ameria onal wendy rire pe pecs slumtons (limes dvergens (pe), and intermediate Spay ot (aoc oa fomogenou Han ope. ‘Tube I present some mini area value for comma ype taken fom various sources. (See TOten, 1970 Exe, for a bllogtaphy). They range from a few dn? for certain epiphytic com- muti to one heeate or more for climax tropical tain forests, Tn guncal the plo sizes aken somewhat larger chan the minal are Tel beelear that plo see sould notary too much within ne vegetation type The thape of the plo i nt standarcized and tiny depend on the suatons I possible, a quadrat of rectangular Shape ito be peterced 20.4.4 Descrrprion oF StRUCTURE AND SrhaTa ‘Vegetation layering is an imporlant structural character. Moaly only four prinepal layers are distinguished. as the Ue, shrub, herb, and moss layers, The latter is also named field layer fr thallophyte layer (which is less adequate, because many of the larger fungi do not belong to it), These principal layers may be further subdivided, In the relevé as many layers are distinguished asis considered appropriate or necessary. For each layer height and. coverage degree in per cent, mostly with 510% intervals, are estimated. For tree and shrub layers the age is estimated and ad- ditional data on stem diameter, number of dead or fallen trees, and ‘occurrence of epiphytic communities are gathered. 307Stratification can be more accurately described with the help of diagrams in which the total or combined cover of each layer is indicated (Hutr 1881, Braux-Braxquect 1928, ef. Care & Casto 1959, Knapp 1971). These diagrams are related to vertical profiles, fas have frequently been presented for tropical rain forest. (sce article 13.4.2) and by Zosnzvetn (1960) for the freshwater tidal delta of the Rhine. Both however have their own special uses, The profile can show exactly the structure of a given phytocoenase} in tie coverage-stratilication diagram a pattern typical for a phyto- coenon ean be generalized (Cams & Castito 1959). A more formal method is that of Daxseiezav (1957, Daxstitzav & Antos 1959) for the description and recording of vegetation on a structural basis (see article 13.3.3) Its possible to classify different strata into separate synusial units. This approach is treated in detail by Barxwan (article 16). ‘The Braun-Buaxguer approach however considers the main strata ofa given stand as a single phytocoenose which has to be analysed as a whole, at least as far as terrestrial communities are concerned, The main arguments are: (i) The different strata are rooting in a common substratum; (ii) The layers are ecologically closely inter- related; (ii) The plants of all other layers have originally been part of the field layer and they must pass through one or more layers as they develop from seedlings to their mature lifesform stature. The layered community is, then, a dynamic totality (see ‘Wenn et al. 1967) on the vertical integration of the rain forest), The situation in aquatic communities of higher plants is less clear however. Here the coherence of layers is much looser, of absent, Du Rrerz (1930a) has proposed astratificationseheme, which was modified by Hanros & Szcat, (1964) (sce also article 16-4.3), Layers are distinguished according to rooting or non-rooting af plants and the positions of leaves in relation to the water surface ‘The layers are named by growth-forms after representative genera, eg. the isoetid layer, An additional analysis can ve made of subterrancan layers. For a treatment of root stratification see Beaus-Brawguer (1064; 59-62). Bibliographies of root studies were given by WiLMANys (1959 Bxe., 1966 Exe.) and Toxen & Wrnatanns (1973 Exe), 20.4.5 Forisnic Anatysis 20.4.5.1Cover and Abundance The description of structure is followed by an inventory of taxa, at least of phanerogams, pteridophytes, bryophytes and lichens, 308 oe SnrBcangier,coverabundsace octnal Dowie : “ trandiorr Fone individual, reseed vigor one oF few indica 1 rare fccasional and lew than 5% 2 sparse Grist plot area 1 Sisndatt ad with very tow 3 3 over, or less abundant but swith higher covers any ease Irs than 5% cover of ttl ot ares 2 ery abundant and les than 3%, Ser or 8.25 Sp cover co total plot area on ery abundant 45 reguent 2a 55125, cover, a 4 5-10 ierespeetive of arbor ct individual : 2 125-25 %eover, 6 5 188% irrespective of number of individuals 3 25-80% cover aftotal plot 7 6 633% fares, erepectve of number of 7 38504 Individuale once 4 50-75 sj emer of total plot 8 8 SLT % area, irrepective of smb of individuale 5 75-100% cover of wtal plot 9 ° tea, irrespective of aumber 10 inv “The taxa ae listed according to the layer in which they grow. Hants which appear to be structurally traggrestive, i. occurring inmore than on layer, have to be recorded in each ofthese layers separately PerTaxa occurring only outside the sample plot (but within the stand) are noted in parentheses, Next, the hantaive occurrence ofeach taxon i estimated. Inthe BaacncBeangoet approach two criteria ae considered mot seu: abundance and coverage degre, ‘Abundance relates to the density of the individuals of gen species ina plot Cover degre is measured asthe vertical projetion ofl serial par of plans ofa given spectes asa peroentage of te 309total plot arca, The term ‘dominance,’ often used as a synonym for coverage, i less appropriate Abundance and cover degree are asally estimated together in 2 single ‘combined estimation’ oF ‘cover-abundanee’ scale The five-point sale in Table II, from Bratx-Buangunr (1920, see alee article 18.4.1) is in general wae, Several authors (eq. Tuoxosns 1042, Dave 1954, Evans & Dan, 1955, Bannon etal. 1964 used tore detailed scales of combined estimation; the Beatx-Brangont Symbol 2 specially eau rofina. The more claborae scale are useful for special purposes, eq. for an accurate fecord of chat of abundance and cover by sigeesion on permanent sample plots in the course of years (Doin 1954). However, they often suggest more accuracy than ean really be justified. Only values obtained by one investigator should then be compared {cf Cate & Casto 1959: 142-148, Maan, I966b). We have recently inoduced ss 4 refinement of scale interval 2 (taken from Danan etal. 1968) the seale subdivisions 2m, 28, and 2b given in Table Tl. This elaboration, which brings the total number of scale values (0 ® has proved to be uscfil and reliable. 20.4.5.2 Sucability Sovihiity or greariousness is aa expresion of horizontal pawcrn of species, Its a measure of the degre of clustering (contagion) of the plant units of speci, A plant unit (Waa 1964) may be an individual ora shoot ora sproutforming part of an individual. Measrement of sociability goes back. (age) to Hine (1899) In the florsesociological approach sociability is estimated with the following scale (Bestn-Bratgort 192, 193, 1951, 1908). Te growing solitary singly. 2: growing in small groups of afew individual, or in ama ose socks (Carspitse), eg Crna cases in shin sand 5. growing in small patches cUshiony or large tases, ee Corer Judson asa hummock bude in atrophic swamps; Sen cals and Sasirage apposite a alpine swan 4." growing in extosve patches, carpets of broken mats, e stands of Heda felis, Lamia golebiolon, dopa drat ee deciduous temperate forest 5. growing in great crow or exetsive mats completely cover the whole plot area; mostly pute populations, eg Eis woul bk Frceeath; Sphagnim nln oS poivum ire haga A variant of scale value 5 (5, loose or open 5) was propaseel 310 for populations with a cover degree of over 75% consisting, how- ever, of plants which are sufficiently separated as to leave space for ‘other species. Similar variants can be used for loose cover of 51-75% (4) and 26-50%, (3) (Mexrzer & Wesriorr 1942, Brawy-Braxguer 1964). Tn the relevé sociability values are written immediately behind the combined estimation values, e.g, Serpus martimus 1.1. During the last decade several investigators (see FuRAREK 196) expressed the opinion that the diagnostic value of sociability has been overestimated and that a certain sociability degree is a specific character of most taxa. Other authors (BRaUN-BLANQUET: 1964, Scamoxt & Passance. 1963, Westnorr 1965) do not agree with this view, Only a few species have a fixed degree of aggrega- tion based upon their innate manner of growth. The degree of igregariousness of most species is much influenced by habitat Conditions and competition, and therefore is of major phytosocio- logical importance. Sociability may also change considerably during the course of a succession; many examples are given by the authors quoted above. On the other hand, the variation in socia- Dility will be correlated with variation in cover degree to some ‘Sceiability & commonly considered as an expression. of vitality. However in various situations this is not so. Jaxues (1970, 1972) remarked that eharacter-taxa of the Trifolio-Geranictea (thermoxerophilous woodland fringe communities or ‘Saumgesell- schaften’) tend to grow with sociability 1 in their optimal habitat, ‘whilst they form polycorms with reduced vitality in suboptimal habitats. Facies of species are often connected with extreme or disturbed habitats (see 20.6.7). 20.4.5 3Vitality and Fertility Further variables in the performance of a species are its vitality and its fertility, representing vegetative and generative development respectively. Braux-BLaxguer (e.g. 1952, 1964) Aeveloped a scale of relative ‘thriving? (Gedeihen) with four categories indicated by symbols (BRAUN-BLaNguer 1932), '@, 1 Well developed, regularly completing the life eycle (an extra- ofdinary vitality is indicated with ‘lux’, Inxurious). ©, 2 With vegetative propagation but not completing the life eyele, 0, 3 Feeble with tow vegetative propagation, not completing the Iife eye. 3iL00, 4 Occasionally germinating but not vegetatively propagating. Braun-BLaNgusr (1964) slightly altered the symbols and gave them numbers from 1 to 4, as above, Baxkwaw et al. (1964), following Vaxescrn (1981) and Zorix (1954), stated that for many species vitality and fertility are i lependent or even negatively correlated parameters. ‘They proposed separate scales for these. 20.4.5.4 Periodicity In addition to vitality the seasonal phase in the life eyele of cach plant, its ‘phenological state’ is recorded. Various scales are in use, the first one being that of Gaus (1918), BRaun-BLANover (1964: 67 and 510) presented twa scales (which are very similar). As was stated by Gaus (1918), Etxenmeres (1959, 1954¢a) and others (sce also. the bibliography by Batsrova-Turackovs, 1970 Exc.) one incidental record during the relevé is really not sufficient, a complete phenological diagram should be desired for cach com munity. For a relevé at a given time, however, appropriate abbre- viations may be used — v. (vegetative), A. (lowering), fe. (feuiting), 204.6 Retevé Prorocors The various structural and liste data discussed so far are written in standardized form either in eld notebooks or on special rotocl forms. These notes are preceded by seme notes On he Flowing items: — 1. Date? running number; topographic locality (as detailed as posible); altiade: exposure and inclination; geology of substrate {posible the locaton should be indicated ona detailed map. 2, Size and shape both af the plot and ofthe entire uniform stand; character of adjacent vegetatin; sol profes phreatie level depth and differentiation of rot system, 3. Character and intensity of Iniman and animal influence, e asturing, burning, mowing, teading, mranuding, ieigation, Table II presents an example of relevé, taken rom Mun za & Wasruorr (1942). For each specie ae recorded the scale numbers for cover-abundance (belore the period) and sociability latter the petiod), the phenological abbreviation, and the vitality symbol. 312 Vanue HE ie fora Mrurzen & Wrsvuore 1942) Ne some 1 Tescheling, Basenplak § of beacon orar beach mark 6, Gdn. Gi61.43 in IVON-system (Tnsitute for Vegetation Reserel inthe Netherlands). Sand very uniform, Empetcur heath om slope of 6 rll parabolic dune, expastion NNE, ineination 30" Habitat: shadowed, moie soi, by day: not strongly heated. and racely desice ating. Slight shilling of sand. Title naman and enimal ithiene Profle: Air? em semidecayed material, ‘Ae 5 Gm dark hums containing sand. Co bright, white dune sand Sample plot 10044 m. ia layer cover 100 Polyps vulgare a3 Empey niga ry i. Hicracium umbeliacam i esuca en aba, arena a Hypochoors radiata at Cabarete epigeos Mee Fasooe Sonne at ‘Ammophia arenaria Sai sepe ‘Visi ctaa ve, daneniy ‘Mons layer cover 1002 - Hypnum copresiforme var ercsoram Hieron sheer! 33 Dieranum orp 23 23 22 cco Miu hormure Laphocolea bidencata urkynchigm wakes Plage dena Polyeichurn juniperinsm Patigera carina Parmetia phydet ladon slicers oo@ceceeoe 20.5 Synthetical Research Phase 20.5.1 Steps uy Sywries ‘The analysis of stands is only the fist step in the description of | vegetation units. After relevés have been collected, they must be compared. This is the start of the synthetical phase which leads to the distinction of coena and, if wished for, the final classification of syntaxa. To this purpase, a number of relevés are tabulated in a ‘matrix, which is usually named a rclevé table or community table. 313314 ee Tig: 8. Scheme af tes in the sythetie procedure. Such a table is also called a primary or raw table (Rohtabelle).. Each primary table is then rearranged into a structured table in which one or more uniform phytocoena are distinguished and characterized. After consultation of the relevant syntaxonomical literature (and possible further revision), this phytocoenon table or community table may then he presented as a syntaxon table, often for an association. ‘When enviroumental data such as soil analyses are available the vegetation types arrived at ean he characterized synecologically. Finally the coena, or syntaxa, are checked in the field, eg. whilst mapping the vegetation of the area under investigation or during. the reconnaissance of a neighbouring area. ‘This may lead to the collection of new relevés. ‘Thus veyetation classification in the BRAON-BLANQUET approach is essentially an iterative process, oF cone of successive approximation in the sense of Poors (1956). Fig. 3 shows the various steps in synthesis. Sunwext (197) also out Tines in English (cf. Karp 1958 and Extexssne 1956) the steps of the procedure. “The checking of units is an essential part of the method; this point seems not to be recognized by all critics of the BRawn-BraN- Quer approach (eg. Poors 1955-56). Moore. (1962) explained that the misapprehension that the fidelity concept rests on a circular argument (e.g. Poor 1955a) is based on the failure to recognize the checking step: ‘In present practice, associations are not distinguished in the field at all, but only when editing the tables fof relevés, The first step involves describing uniform uacts of vegetation, not representative stands of a presumed association. Only when sufficient relevés have been accumulated and analysed ‘one can discern the asiociations. Of course, in observing continental phytosociologists at work in their own homeland whose vegetation they know, one may be misled as to their methodology. They will now have reached the second stage of checking the reality of units already distinguished. They will not necessarily make this clear to 8 visiting inquirer 20.5.2. Pamary Taste In the primary table all taxa are listed at the let hand side of a sheet of squared paper; to each relevé a single vertical column is assigned, Each item in any column should contain either a dot or dash (in the ease that a taxon is absent in the corresponding relevé) or else a combined estimation value (or an actval coverage per cent) and preferably a sociability value too. It may be convenient 31sa this prliinary stage © group the species under separate headings according to stala and to separate phancrogane. aod éryplogans. ‘The ndiion of new species fro the Inet Flees produces a characterise tiling off the right ofthe table Bach ros of the table represent a species (in a given stratum, Come pation ofthe rows ofthe table enables us tojudge the dsteibuion bf any taxon over the relevess comparison ofthe eats may Tend to pecliminary canlusions on relative simarses ofthe elev "The primary table shold be rearranged several mea inorder to establish groupings of reeves by rearrangement of columns at well ay to group taxa with similar distibuions in the table by re frrangement of row. Before we discus this rearrangement proce- dire 20.5.5) some concepts and techniques shoud be discuss, 20.5.3 Presence, Constsncy, Honarowerry Presence is the oceurrence of a taxon in a vegetation table. Te is usually measured as a degree by the number of relevés in which the taxon occurs (regardies of abundance and cover) expressed as a percentage of the total number of relevés compared. Presence degree (Stetigkeit) can thus be caleulated from any aumber of rclevés, regardless of the difference in size of the plots. When plots of equal size are compared the corresponding per cent occurrence for a taxon is named ‘constancy’ (Konstanz). (Frequency, in contrast with these, is an analytical concept dealing with the distribution of a taxon within subsamples from a given stand.) Presence degree and constancy are determined from stands of different and often widely distributed localities. Constancy determinations on plots of equal size were especially favored by the Uppsala school (article 18), starting with Du Runt et al. (1920), although the first such measurements, by BRoCkMANN-JeRoscit (1907) had used plots of various sizes. In the BRavN-BLANQUET ap= proach presence degree has been termed “Stecigkeit’ and distin« {guished from ‘Konstan’; but confusion has resulted from the trans lation of both concepts as ‘constancy’ in some statements in English (cog: Braun-Braxgusr 1932, Moravec 1971). Constancy and presence degree can be given in exact per centages, or in percentage classes. Tt is usual 10 distinguish five classes, noted in Roman figures: 316 Class Percentage of plots in which taxon occurs: 1 1-20 u 21-40 uM 41-60 Vv 51-80. Vv 8-100 ‘The mumbers of xa falling into these classes are often presented in a ‘constancy diagram,’ Such diagrams are iraporvant charac- teristics of vegetation units and provide useful tools to check the uniformity or homogeneity of a table. This synthetic homogencity concept, however, should be distinguished from analytical homo~ sgencity. BaRKwAN (1958a: 316-317) designated these concepts as “intensive homogeneity” and ‘extensive homogencity.’ However, as Dau. (1970, apod Texes, 1970e: 101) pointed out, the latter term has been designated as homotonely (Noxouacen 1943, Das 1957, 1960). ‘This term was long overlooked in the BkaN-BLANQuEr ap- proath, but naw receives common use (e.g. TUxen 1970e, Moravec 1971). Homogeneity is an analytic concept, based on comparing, different plots of the same size taken from an individual stand, whereas homotoneity is a synthetic concept, based on comparing similar plots from different stands of the same community-type ‘or phytocoenon. Homotoncity (see also 20.8.4) has been judged with the help of constancy diagrams; especially in North-Buropean approaches (see further article 18). ‘The classical interpretation, based on Raoxzasr's (e.g, 1984) ‘law of frequency,’ ofa constancy diagrant js that the following relation between the constaney classes exists S,> Su> Su = Siw < Sy (the reversed J-chape). In very homotoneous tables class V may equal class I (U-shape). When classes LIL and IV include more species than class V, the table is considered hheterotoneous (ef. Guinocist 1955, Cam & Casteo 1959, Moravec 1971). According to Maanet. (1972, in Maat & Téxex 1972: 209), the reasoning of Wit1asts (1964) and ob- servations of BaRxwan (1958) showed that constancy class I is more tr less dependent on the total number of relevés. Between the other classes the following relation exists in omotoncous tables Sp4Sy)Sict Su is slightly over 1, whilst Sy} Si Sy/Sq is mostly about 2. ‘A second feature with which the homotoneity of a table can be easily checked is the variation in the number of taxa within the relevés of the table. In most cases well-developed stands of phytocoenon do not differ much in number of species. A high 317variance of that number is reason to suppose that the table is heterotoncous, which may mean that more than one phytocoenon is represented in it (Hornas & Passanoz 1968). After the relevés have been classed into groups according to their number of taxa, we may obtain the ea diibatons of thes eas, whieh will indicate this variation. A table comprising releves belonging to two diferent coena may produce eurve with two oF more pea (Keare 1971) ‘Apart from superficial inspection of homotoneity some simple measures have been proposed on the basis of constancy figures nd species numbers [sce MORAVEC 1971 for a survey). ‘The following symbols, all corresponding to one vegetation table, are used Af is the number of relevés: Sy isthe total number of species and the mean number of species pet relevé; S,, ete. is the number of species falling into constancy cass V, ete. C, refers to the constancy per cents for species of the constancy class(es) indicated. by the subseript to the summation sign, or for the prevalent species P those belonging to the group of species with the highest constancy values, the number of which (S,) approaches the average number of species (8). “Amongst the measures based on constancy figures are 1. ‘The ratio Sy/Siy (Dan 1957). ee 2 Basic homotoncitycooficient 5% (Moravec 1971), 1 3. ‘Homogeneity value’ C, (Raane eg. 1952). za 4 Index of homogeneity (Gunns 1859) 5 5. Mean eonstaney for all species 3 C)/S. 6. Corrected homotoncity coefficient (Moravec 1971). This is coefficient 2 with an ‘oscillation factor,’ determined by the dif ference in species number between the schest ad Uhe poorest Toxex (1970c) presented curves for various communities ng in a horizontal line (section 20.4.3); one may derive a imal relevé number’ from such curves. In fact TExen based his saturation quotient on tables containing about this minimal number of relevés. Other observations suggest that the number of species grows continuously with the number of relevés (e.g. BETTER 1949, Barkaiaw 1958a, Dau. 1957, 1960, Wiextaws 1964). Dait. (1961) based an ‘index of uniformity’ on’ this telation (sce 20.8.4) 318 and Barkstaw adapted Kuesevr's heterogeneity index by omitting species with a constancy ofless than 10 %, 7. ‘Saturation quotient’ 1005/5, (‘Tox=y 1970) As Tiixen mentioned this measure is the reciprocal of the homo- geneity (ic. heterogencity) value of Dani. & Hapa’ (1941) Wanrraxen (1972) suggested use of BD = (S/S) —1'as a measure of beta diversity (article 3.6), Prerrenn’s (1957) homogeneity Value was reduced to TUxen’s measure by Moravec. Krament (1942) used the same measure under the same name as Daut. & ‘Hanat did, as was mentioned by Barkaay (1958a), who concluded that this index is largely dependent on the total number of relevés in the table In conclusion we think it would be more realistic to base a saguration quotient explicitly on species with a minimum constancy value, e.g. species from constancy classes II-V, and to measure the rate of increase in total species number separately. For measure= ment of homotoncity we suggest the mean similarity coefficient for the relevés (ser 20.8.5). 20.5.4 Srecins Weiois Species may be weighted in synthetical procedures such as fidelity determination, assignment of syntaxa to higher units and caleulation of spectra, Weighting is usually with the combined cstimation value or some transformation of this. ScHWICKERATH (1931) used ‘Artmachtigkeit? — and when species groups were concerned ‘Gruppenmachtigkeit? (species and. group importance value, Maret. 19725) — based on BRavn-BLaNguer figures and arbitrary numerical valuesforsymbols 4+ andr. TUxEN & ELLENRERG (1987) and Bravy-Bravgver (1946) “used the Deckungswert {cover value) and Gruppenwert (group value) by taking average coverage percentages for BRAUN-BLANQUET values 2-5 and ar- bitrary values for J, + and r, The latter procedure has received ‘more application than that of ScnwickeRaTa, although TOXEN & ELLENDERG stressed its limited applicability, which was also criticized by Wesrnorr (1947), Met Daess (1949), Snstven (1950), Daoxenie (1960) suggested the use of an are-sine transforma- tion as is usual with percentage scores. His rounded figures appear to be identical with the original BeauN-Branguer values, Various other transformations have heen applied, including those of ErtER (1949), Bargwan (19582), Barkan et al, (1964), Maaren (19666), Scrap & Koy (1970), Moone. (Moone & O'SLLIVAN 1970) and Lowvo (1971). (See MaaRet, 1972b for a review.) 319