Studies of the vegetation of a white-sand black-water igapo (Rio Negro, Brazil)

The vegetation analysis of a Central Amazonian igapo, a forest under severe environmental stress — poor soil and seasonal flooding, reveals the existence of the dominant species Myrciaria dubia with a M.I.V. 75. The Shannon diversity index of this forest is 4.358 bits per individual, of which 75.73% is attributable to the evenness value. Species distributions change along the moisture gradient. Though many species exhibit the tendency of zonal distribution, the boundaries of zones are not abrupt. The authors suggest that species distribution is the result of physiological difference to flood tolerance. Further research should be directed to the comparison of flood tolerance of the plants which occur in the following three areas: permanent water-logged area, beach area with open canopy and upper area with closed canopy. The various adaptive strategies to seasonal inundation and drought, and the reproductive potential of the forest also merit further studies.

Since Takeuchi's preliminary investigation (1962), there h2ve not been any detailed studies dealing with the vegotation structure ot central amazonian igapós. The forests studied by Bluck et a/. (1950) near Rio Guamá and by Pires and Prance ( 1977) in the Catú Reserve (both in the vicinity of Belém). thouJh termed igapós, are in fact white-water várzea forests. Similarly, in his paper on igapó forests along the Rio Negro and its tnbutaries, Takeuchi treated the mixed water várzeas of Lago Janauacá as igapós. To a great extent. this confusion is the re.suit of inconsistent use of terminology for amazoni::~n vegetation types subject to inundation. The corrcct application of 1erms to am2zonian forcst types subjected to inundation was tully disct.:ssed by Prance (1979). He suggested that the term .. igapó n be restricted ta forests immdated by black or clear waters. In the present study, Prance 's definition of igapó will be used.
The seasonal igapós in the Amazon basin are flooded annually by the rise of watc;r levei of black water rivers. In the case of the Rio Negro, flooding is during the rainy season from December to May (Table 1 & Fig. 1); the water levei does not recede until well into August, whit:h is the midâle of the dry season (Fig. 2) . Although temperatura is uniformly high throughout the Amazon basin, soil shows great variety and thereby plays a more important role than climate in the different1ation o f vegetation types ( Ducke & Black, 1953) .
Igapó vegetation growing on different soil types exhibits different physiognomies. Some seasonal igapós occur on white-sand podzol and are f!ooded annually by black water, Shirley H. Kuo Keel ('') Ghillean T . Prance ("') which ranges in pH from 3.7 to 5.4 (Sioli, 1968). The investigation of amazonian podzol hr~s revealed that it is extremely deficient in nutrients. Rich in white quartz, this soil is highly porous and therefore rapidly leached by the usual heavy tropical rain. lts acidity results in low ion-retention capacity (Kiinge, 1965;Stark, 1971;Stark & Jordan, 1978). The poverty of this soil and seasonal flooding act as important edaphic factors which determine the vegetaticn of igapós. The purpose of this study was to investigate the effect of this environmental stress on the igapó vegetation, and especial ly to determine the relationship between stress and dominance.   "YR" (Yeor) represents the number of yeors of record over which the monthly ond onnuol precipitotlon volues hove been computed. "ELEV" (Eievotion) represents the elevotion in feet obove meon seo levei of the reporting stotion.

MATERIALS AND METHODS
The field work was carried out in 1977 from October to December. During t he dry season, the limit of the igapó can be easily recognized by the attachment of sponges (Photo 1) and flood marks on stems of trees and shrubs. A preliminary study was made to determine an adequate sample area. Figure 3 indicates that numbers of species increase sharply with the initial increase of area. However, the increace leveis off after a sample of 1.800 m 2 , with 68 species inc!uding vines and herbs (54 species exc ludi ng vines and herbs).
The igapó vegetation was sampled by means of 12 randomly chosen plots, each 10 x 15 meters. These plots were la id with their longer sides at a right angle to the shore. Within each plot, trees and shrubs higher than 1 m were recorded and their diameters at ground levei were measured. The importance  To invest1gate the effect of moisture stress on vegetation change, five linear transects, each 50 m long and 50 m apart. were placed at a right angle to the shore. One transect was only 40 m in length because of an abrupt 1 m rise above which terra firme vegetation was present. Trees and shrubs higher than 1 m and within 50 em on both sides of transect !ines were recorded, and their diameters at ground levei were measured . The data from each 10 m segment of 5 transects were pooled separately, and the importance values were determined. The flood leveis at each 10 m interval of transects were obtained by the highest points of sponge attachment on stems.
\louchers of sterile plant specimens from transects and plots were deposited in INPA, while those of fertile plants were deposited in INPA (1st set) and NY (3rd set) .

RESULTS
The study site is a low, open to closed one-story forest consisting of shrubs 2 -5 m high as the main life-form. The leading dominant in the area is Myrciaria dubia with an I.V. 75.78 (Table 2)  which contaial 54 species of trees and shrubs higher than 1 m with 1028 individuais. This value represents an estimate; the total number of species in the community studied is actually unknown. For this reason , 4 . 358 bits per individual is a slightly underestimated value (Pielou , 1975 Although spccies zonation is obvious , along the moisture gradient boundaries of the zones are not abrupt (Fig . 4) . Myrciaria dubia occurs only in the segments close to the water (between O -30 m), and its abundance decreases along each subsequent 1 O m segment . Eugema cachoeirensis and f. chrysobalanoides grow between the area delimited by the second and fourth segment, while Schístostemon macrophyllum can be found only up to the third segment . Although f. cf.
patrísii grows from the third to the fifth segment, Pera distichophylla is limited to segments four and five. Pithecellobium adíantifolium grows in every segment of the transects except the first, where only four species (M. dubia, Remijia tenuif/ora, Tococa subciliata and Turnera acuta) occur. The area covered by the first and second segments, with flood leveis that are 4 to 6 m high in the rainy season, appears to be more open with the individual plants widely spaced (Photo 2). i=urther up the slope, the space between individual plants decreases and species richness increases (Fig . 5) .

650-
thf:ir relative importance sequentially. The rapid change of dominant species within a short distance in a homogeneous substrate (white-sand) po ints to the influence of other important environmental factors, e. g . flood levei. The physiological basis of flood tolerance h as been discussed by Crawford ( 1976) .
McMannon & Crawford ( 1971 ) demonstrated that in anaerobic conditions flood-intolerant species accu mulate ethanol, w hereas floodtolerant species undergo a metabolic change, accum ulating malate rather then ethanol. To overcome the lack of Ch, flood to lerant species use more effectively nitrate as an alternate electron accept or (García-Noro et a/., 1973) Since the area has a gentle 10% slope, vegetation near t he shore is water-logged Jonger and deeper than t hat in t he upper area. Thus the change of spec ies along the moísture gré:ldient probably reflects the difference of physiological ability to flood tolerance.
The transects cover only sandy areas exposed in the dry season. There are perma· nent water-logged areas with sparse vegetation stret ch ing out from the shore into the water (Photo 3). Due to limited t i me ava'lable for field work, they were not subjected to the transect ancõlysis . Eugenia inundata, Sphinctanthus stmflorus and Securidaca longifolia are most frequent in the water. Takeuchi   Studies of . . of igapó forests along river shores and perma· nenlly inundated islands. However, according to the authors' observation, this species is mainly restricted to the permanent waterloggecl area rcgardless of whether the water is of black or white type. Along sa~1dy beaches which become dry annually for a period of time, different dominant species, such as Myrciaria dubia, occur.
The various adaptive strateJies of igapó vegetation to seasonal inundation and drought merit further investigat1on . The study a a woody swamp in Suriname has shown that some trees survived dunng h1gh water periods by becoming partly deciduous in the crowns or by corky breathing roots (pneumatophores) which bend down sometimes during low water period and develop as prop roots. Others, (e lled by the flood, regenerate vegetatively with root suckers (Teunissen , 1976) . In thc area studied, pneumatophores are seen on the trun k o f an unknown tree . The vegetation change during the rainy season remains to be investigated. Prance ( 1979) indicated that some igapó vegetation displays xerornorphic adaptat ion to seasonal dryness with sclerophyllous leaves. Since no plant with sclerophyllous leaves or other morphological adapté:ltion was noted, the drought effect on vegetation in this igapó may not be very severe. However, the s~udies of sap tension of flooded trees and shrubs along the Rio Negro near Manaus showed that many plants such as Ruprechtia sp., Parkia auriculata, Eugenia sp., etc. reached zero turgor without wilting at the peak of hot days (Scholander & Perez, 1968) . lt may well be possible that this physiological property contributes to the tolerance to drought during the dry season.
The species list (Table 4) provides preliminary information about the plants of the igapó studied. Many plants have been 1dentifieà on the basis of sterile material. In the dry season many seed lings and saplings are found in open canopy areas. They have been excluded from this study because of the difficulties in identifying them. However, if their identification should become possible, their abundance can serve as an indication of the reproductive potential of the ar ea.   Fig . 5 -A simplified profile of a transect of 1 x 50m2. Thin fine indicates basal d1ameter of stem < 10 em; triek llne indieates basal diameter of stem ~ 10 em but < 30 em; Ar: Anaeampta rupieola; Eea: Eugenia eachoeirensis; Ech : E. chrysobalanoides: Eo : E. cf. omissa; Lp: Leopoldi nia pulchra; Md ; Myrciaria dubla; Ml : Myrtaceae I; Os : Ormosia excelsa; Pa : Pithecelloblum adiantifolium; Pc: P. clavifo rum; Pd: Pera distichophylla; Sn: Sweetia nitens; Ta : Turnera acuta .

SUMMARY
The vegetation analysis of a Central Amazonian igapó, a forest under severe environmental stresspoor soil and seasonal flooding, revea ls the existence of the dominant species Myrciaria dubia with a M.I.V. 75 The Shannon diversity mdex of this forest is 4.358 bits per individual. of whlch 75.73% is attributable to the evenness valuc . Species distributions change along the moisture gradient. Though ;nany species exhibit the tendency of zonal distribution, the boundaries of zones are not 3brupt . The ~uthors suggest that species distribution is the re~ult of physiological difference to flood tolerance . Further research should be directed to the comparison of flood tolerance of the plants which occur in ~he foilowing three areas: permanent waterlogged area, beach area with open canopy and upper area with closed canopy. The various adaptive strategies to seasonal lnundation and drought, and the reproductive potential of the forest also merit further stuclies.