Notes on the vegetation of Suriname 1-Vegetation changes in a dammed up fresh water swamp in NW Suriname

The vegetationnl changes in a newly dam:;·.ed area on the Nanni Creek Basin in NW Suriname are here considered . The effects o f periodical floocüng on the different vegetation types are studied, and the types of succession ocurring at present are described. It is postulated that marsh forests that are now being drowned will be replaced by a Triplaris swamp forest. Peat levels will probably remain constant, as any accumulation of peat during high water seems to be counter-balanced by peat decomposition and destruction by fire during Jow water leveis.


INTRODUCTION
The Nanni Creek in NW Suriname, approximately 150 km long, empties into the Corantijn River about 12 km south of the settlement of Nieuw Nickerie (fig.1).
Because the Nanni drainage basin has, for the greater part, a swampy ( 1 ) <.haracteristic, and the creek ltself is covered over a long distance with floating vegetationl the entire basin is usua lly ca lled the Nanni Swamp.Tho Nanni Swamp is bordered in the north-west by the Nieuw Nickerie polder area ( 2 ), in the north-east by the lower Nickerie River, and it lies between the Corantijn River in the west, the Maratakka River in the east and continues south until the boundary of a higher landscape.The whole drainage basin covers an area of about 1700 km 2 • P. A. Teunissen ("*) Within the scope of a study on the Hydrological behaviour of shal/ow lakes and marshes in sma/1 watersheds in the humid tropics, with special reference to the Nanni Swamp in Suriname by Sevenhuysen (l), the present authcr studied the influence of the recently changed hydrological behaviour of the northern Nanni basin on the vegetation .Field work and aerial surveys in this area were carried out on May 25-26.July 29, 1972; Sept. 19-29 and Oct. 17-22.1973; and Dec. 12-14.24, 1974.

ENVIRONMENTAL DESCRIPTION OF THE STUDY AREA
In analyzing the factors influencing the vegetation in this part of Suriname, emphasis was placed on landscape ( 4).hydrology and fire.For general climatic data of Surinamc see Braak (1935) I Ostendorf (1953)(1954)(1955)(1956)(1957) I Lindeman and Moolenaar (1959) I Schulz (1960).and the "Climatological tables" and "Precipitation maps" for Suriname by the Metereological Service, Paramaribo ( 1960).Comprehensi • ve climato logica l and metereological data from the Nanni area in particular will be dealt with by Sevenhuysen (in preparation) .

. GENERAL DESCR1PTION OF LANDSCAPES
The northern part ot the Nanni Swamp is located in the recently formed (holocene ) cocstal region of Suriname.In holocene times marine, estuarine and riverain clays and loams   Notes on the vegetation ...

•-119
alternating with sand and shell r idges were deposited on a pleistocene stratum (Brinkman and Pons, 1968).
Two landscapes can be distinguished in the northern Nann i basin (Van der Eyk, 1954Eyk, , 1957)): the Nickerie-landscape or Young sea and river clay landscape, and the Kwata-landscape or Ridge landscape [fig.2).

A. Nickerie-landscape
The principal landscape element ( 5) is formed by vast swamps .The soils consist mainly of half ripened to nearly ripened r.lays, mostly without sa lt, to a depth of 1 m (Pons, 1964).The soils are covered with a peat ( 6) layer of about 10-30 em thickness (DBK = Soil Survey Department Suriname, 1964).
These soils are classified as Fresh w ater clay swamps soils (DBK, 1964).Their vegetation consists of swamp forest, swamp wood, swamp scrub, or they are covered with a herbaceous swamp vegetation.Another lands• cape element is formed by marshy (1) river levees along the Corantijn Nanni, Nickerie and Mar::~takka Rivers.The soils cons ist of neariy ripened and fully ripened fine sandy, si lty and heavy clays.These soils are classified as River levee soils (DBK, 1964).Their vegetation is a river levee forest.fhe last landscape element is formed by the period ically submerged mangrove be lts along the lower rivers.The soils are brackish.slightly ripened clays.These are classified as Brackish periodically inundated clay swamp soils (DBK , 1964).These soils are covered by mangrove forest and swamp scrub.

B. Kwatta-landscape
Especially the southern and north-western part of the investigated area is formed by the Kwatta-landscape.The most important elements are fanshaped bundles of ridges: low dry and marshy bodies originated as coastal bars.The soi ls of the higher parts consist of dry fine sands, class ified as Ridge soils (DBK, 1964) .The ma rshy parts h ave soi ls o f more or less heavy, f ine sandy loam and are classified as Ridge flank soils and Ridge feet soi ls (DBK. 1964).Between the ridges interridgeswamps occur.Their soils are fresh water clay swamp soils and Levee Soils (DBK, 1964) The higher parts of the ridges are as a rule.covered with ridge forest.the flanks and feet are covered with marsh forest, the interridge swamps have swamp forest, swarnp wood, swamp scrub, or an open herbaceous swamp vegetation.

HYDROLOGY
The whole Nanni area is hardly accessible and the creek itself is difficult and partly impossible to navigate because of hcavy floating vegetation.Since early times the area has been visited only by some hunters.fishermen.ba lata-bleeders and woodcutters.Since 1915 the area around Nieuw Nickerie has been cultivated for rice growing.In the early twenties, people became interested in the Nanni swamp area as a water supp ly for their rice fieicls.In 1920/ 1921, for the first time, a topogrélphic survey of the Nann i Creek, from its mouth ti I! 55 km upstream, was carried out by topographer Sniphout (Geijskes, 1941).In 1926, at about 6 km from its mouth, the so called Boonacker Dam was built in the Nanni Creek to lead the water via the Roonacker Canal mto the cu lture area (fig.3).As this area has been enlarged severa!times since, frequently a water deficit occurred, especially with regard to the second r ice harvest in relatively dry years.The water supply has been improveci gradually.
In 1936 the whole drainage basin was surveyed by plane by irrigation engineer Van Wouw.In 1939 another reconnaissance flight was carried out by the airplane "Snip" of the Roya l Dutch Airlines (KLM).During that trip ( 5) -A land$Cape element is a smaller unit w ithin a lan dscape.Landscape elements are strongly contrasting in topography, soils and vegetation (Van der Eyk, 1954Eyk, , 1957)).( 6) -Undecayed organic matter.( 7 ) -Mars~y soíls ~re inundated part of the year and in the other part they lie well above the ground-water table allowmg aerat10n ot the topsoll, but w ithout dessication (Líndeman and Moolenaar, 1959).aerial photographs were taken of the higher course of the creek.From ãll these data the lrrigation Service of Nickerie designed a preliminary map of the entire Nanni basin.With the new information they built a new aam called "Nanni Dam" about 6 km further upstream to lead the Nanni water via the "Van Wouw" canal to the culture area.Construction of dam and canal was carried out in 1940/ 1942.
Because the swamp area is very flat, from that time on part of the dammed up water drained away via the surrounding swamp inlo the lower course of the Nanni Creek whi le part of it disappeared east of the Nickerie polders into the Nickerie River.However.since 1942 there is evidence that the water levei has been raised and somewhat stabilized.
Because the poider area was enl:1rged again, the Nanni Dam was improved in 1965, after which a 11 km long "Lekbeteugelingsoam" (drainage diversion dam) was constructed) from the Nanni dam to the ridge complex of the Corantijn River, to prevent drainage into the lower Nanni River (photo 1).From that time on part of the dammed up water still drained away into the lower Nickerie and Mnré::t3kka Rivers.To prevent this, between 1971 and 1972 a final dam was built in an easterly direction.this time 33 km long.This new dam has improved the water supply conciderably.For the first time since hydroiogical dat3 are available (1956), the swamp levei at the Nanni Dam r os e above + 1 O. 30 m Nickerie Levei (N.P.) which means a water depth o f 1 . 1 O m above peat levei for that lo• cation.The first period of this extremely high swamp leve i lasted 25 weeks (March 19-September 9, 1972).At Nanni Dam maxima of + 10.43 (1972) 3) are given below (fig.5) .Fig. 4 and 5 show that .
-------Fig. 4 Average monthly swamp level in m + N.P. (Nanni Dam> . Period 1957-1974, after figures supplied by the Suriname Hydrological Service, Paramaribo. in a relatively short time ali dammed up water can be used for irrigation.At swamp levers lower than + 9. 80 m N. P. no optimal irrigation can take place for the entire culture area.At swamp leveis lower then + 9. 20 m N. P. only the Nanni Creek bed still contains free water, while the peat layer of the surrounding swamp starts drying out (Sevenhuysen, personal communication).

FIRE
Primitive agricultura!methods of the Amerindians, who during early times had been living on the higher ridges and levees around the swamp, could have caused burning of swamp vegetation in dry seasons, and of peat in extremely dry seasons.Since 1915 land along the northern edge of the Nanni Swamp is being reclaimed.A common method of removing vegetation and the peat layer from the land is by burning.These fires pushed by the ever present N. E. trade winds can easily move over a long distance.Since colo-Notes on the vegetation . . .nial times , hunters, fishermen , balata-bleeders and woodcutters have visited the areas and a few still do so at this time.To improve th e navigability of the Nanni Creek, it is worthwhile trylng to burn the floating vegetation to the water levei.This method can be especially successfull for grassy mats, which are mostl y mixed with a lot of dead and dry material .This method was, and still is, being practiced Sometimes, for no particular reason except willful destruction.vegetation is burned.In severely dry periods even the peat layer can burn causing extensive swamp woods and swamp forest damage; this is not because of thc burning of the trees but because of the loss of iheir root-subs ~~ate which will cause trees t.o fali over.Bubberman ( 1973)  The opinion is gaining ground that for most fresh w:~ter swamps in Suriname which are presently not covered with clímax swamp fcrest, the vegetation has been developed after one, or repeated burns, by man.

. EARLIER VEGETATION STUDIES
In 1933 Lanjouw (1936) visited the southern border of the Waterloo sugar plantation, and made the first botanical collection of the Nanni Swamp along its northern border.This part of the swamp, however, has now been impoldered.
On request of the lrrigation Service. Geijskes (1941, unpublished) tried to reach the source of the Nanni Creek in 1941, but because of the heavy floating vegetation from the 4th hairpin turn (60 km upstreams) on, he was forced to return without reaching his goal.This trip however can be considered as the first terrestrial biological survey in the northern part of the Nanni drainage basin.In his report Geijskes gives an account ot soil, vegetation , flora and fauna of the visited area .Some of his observations concerning flora ano vegetation were published !ater (Geijskes. 1945;Ge:jskes and Schols, 1948).A herbarium collection was made but partly unidentified material was d istributed to herbaria abroad during the Second World War.From his 1941 trip, and aerial surveys in 1942 (Geijskes, 1942. unpublished) and 1946(Geijskes and Schols, 1948).Geijskes sketched a vegetation map, scale 1 :200.000,.ofalmostthe entire.Nanni basir.(Geijskes 1946, unpublished).Geijskes placed his unpublished data at the author's disposal.They are used in this paper thanks to his kind permission.
Early in 1949 Lindeman (Lindeman, 1953: Lindeman and Moolenaar, 1955, 1959) visited the fresh water are.asouth of the Nickerie River between the Nieuw Nickerie culture area and Maratakka River.Lindeman recorded 29 vegetation plots in this area.Fourteen plots are now lost as a result of impoldering.A vegetation map was given.The vegetation types Lindeman (1953) distinguished.however, occur also in the now investigated, somewhat more southei n area.

VEGETATION DESCRIPTION ( 8 )
A short description of the vegetation of the northern part of the Nanni Swamp is given below.extracted from above mentioned sources.
Unpublished observations of Geijskes will be dealt with in more detail.The vegetation un.ts described below are arranged according to the legend of the present map of vegetation and vegetation changes (see back flap) .
Between the mangrove trees Machaeriwn lunatum (branti-maka) scrub exists along the Nanni Creek mouth (Geijskes, 1941) as well as along the Nickerie River and the Maratakb River where it does not quite reach the Amerindian village of Cupido (Lindeman, 1953) .
Hibiscus tiliaceus (maho) scrub is mentioned by Geijskes (1941) from the Nanni Creek mouth and it was found by Lindeman (1953) in the Cupido-line.
( 8 ) -For the convenience of those readers, who are not acquainted with scientific plant names, local names (if available) are added in brackets .A complete list of species mentioned in text and tables is given as Appendix 1.

B. Herbaceous swamp vegetation
Geijskes ( 1941) remarks that open swamps ~long the lower course of the Nanni Creek, tn::~t is from its mouth till the first hairpin turn, have a similar vegetation as the swamps do south of the polders of Nieuw Nickerie.This vegetation consists of herbaceous plants in víhich the giant sedge Cyperus giganteus (pajagrasi l and Cyperus articu/atus (adroen) àominate.According to Lindeman (1953), this type of vegetation which he describes as Cyperus giganteus-Typha-ScJeria swamps, also occurs south of the Nickerie River.According to the dominance of one or a few species, he distinguishes various communities.lmportant species are Typha angustifolia (langa-grasi), Cyperus articulatus (adroen), Cyperus giganteus (paja-grasi) and Montrichardia arborescens (mokomoko).In some places shrubs and small trees appear in these swamps as a firsi' sign of development into swamp wood arct swamp forest (Lindeman, 1953).

•
Fl0ating vegetation also exists, mainly in creeks.Lanjouw ( 1936) mentions some species of open water and of swamps near Nieuw Nickerie.For the lower course of the Nanni Creek, Geijskes ( 1941) describes f loating vegetation which grows out from the banks and sometimes covers the total water surface.H e mentions especially two "creek grass" species as dominant in these mats: "bamboe" grass with a loose panicle ( 9 ) and "soppoe" grass with a long closed panicle ( 10 ).As accompanying species he gives Ludwigia leptocarpa and L. nervosa and two species of "matrozenroos" one of which is Hibiscus sororius.Further, a dull lilac flowering Convolvulaceae (" perhaps lpomoea subrevoluta") and locally a dark green herb named "Louisa or segodrowiwiri ( 11) are noted.Geijskes also makes mention o f "wara pa" grass (1 2 ) from the mouth of Kamisa and Two Mouth Creeks.From the Nanni Creek further upstream he reports f loating Montrichardia arborescens mokomoko) islands with accompr:mying shrubs and lianas, to which Alf•amanda cathartica (Wilkensbita) and Cydista aequinoctialis belong.Sometimes these islands contain Pterocarpus officinalis (watra-bebe) treelets as well.Between the 1slands floating mats of creek grasses are found again with Hibiscus sororius, both Ludwigia species and a Xyris species ( 13 ) .The more open patches are covered with Hydro• cctyle umbel/ata (waternavel) , Nymphoides humbo/dtianum and in a few cases with Lentibu lariaceae spp.. Lindeman and Moolenaar ( 1959) also make mention of the float ing mats in the Nanrri Creek.They report that thesE: mats sometimes form a solid floating peat layer up to 2 m thick, which occurs over considerable stretches of the creek.

-
Bactris spec.(keskes'maka) and Heliconia bihai (paloeloe).The trees had flood marks on t he trun ks at one foot high.Lindeman ( 1953) makes mention of marsh fore&t from marshy strips between the drier parts of the Cupido ridge and the levees of Nanni Creek, Maratakka and Nickerie River.

F. Ridge forest
The rea l ridge f orest , a poor form of tropical rain forest.is reported by Lindeman (1953) as found on drier parts of the ridges west of the Maratakka River near Cupido.

V EGETATION MAP 1972 (see map in back flap)
After some additional field work and aerial surveys the above mentioned vegetation units (A-F) were mapped at the present scale 1:100.000f rom aerial photographs scale 1:30.000infrared (1972).Legend: This map is also used to show the burned forest areas (period 1948-1972), and drowned forosts (period 1941-1972, 1972-1974).
A more detailed vegetation map of tne whole Nanni drainage basin will be pub lished in the near future in the framework of an inventory of ecosystems in Suriname for the purpose o f nature conservation pol icy.
VEGET ATION CHANGES 1. VEGETATION CHANGES 1941-1972 Studying the already mentioned literatura, vegetation maps and aerial photographs of the area, consiáerable changes in vegetation were found to have taken place between 1941 and 1972, mostly caused by fire (swamp wood and swamp forest), but also by changes of the hydrological behaviour of the swamp (marsh forest).
Geijskes ( 1946, unpubl ished) sketched some levee forests along the Nanni Creek and close to the Kaaiman Creek mouth, ali between the Nanni Dam and the hairpin turns.For the greater part these levee forests were drowned within the period 1941-1972.These forests have been mapped (see map in back f lap) .lhe ir total area is estimated at about 5 km 2 .Photo 1 shows part of the drowned river levee close to the Nanni Dam where the forest has disappeared completely.

VEGETATION CHANGES, 1972
On May 25, 1972, after the eastern drainage diversion dam was completed, a boat trip was made through the excavation canal along this dam from the Maratakka River to the Nickerie polder area.During this trip the leved of the swamp at Nanni Dam was extremely high ( )t + 1 O. 30 m N .P. ) and had been for about 5112 weeks.At that time the swamp woods south of the dam were dying.Subsequent te, this trip the Nanni Creek was vis ited from the Nanni Dam up to the Kaaiman Creek mout h.At one place the river was fringed by a narrow strip of dying levee forest.Only some Cordia trees and Attalea maripa palms seemed to survive.The levee forest at Kaaiman Creek was also dying.h.July of the same year the author joined a reconnaissance flight over the Nanni basin.The visible effect of the constructon of the Notes on the vegetation ... eastern dam on the vegetation appeared to be limited to the northern part of the drainage basin, i. e. the area north of the Nanni Creek hairpin turns.In that area a lot of swamp scrub and wood and marsh forests (on levees and along the northern edge of the Cupido ridge bundle) were dying.However, along the western part of the dam the swamp woods looked rather healthy on both sides.Here the swamp woods, as was found out later, grow on a somewhat higher elevation.Herbaceous vegetation in between the drowned woody vegetations looked healthy.

VEGETATION CHANGES 1973-1974
In Apri l, 1973, WOTRO placed funds at the author's disposal to study the vegetation changec; in detail.In September, 1973, the c0mbined hyd rological-vegetational study of the Nanni basin could start.

. Selection of sample plots
Along both sides of the eastern drainage diversion dam plots were selected, the study of which could give an idea of the vegetation changes which took place in the past 1V2 years.Herbaceous swamp vegetations and swamp scrub were excluded from the study as they form a very irregular mozaic on both sides of the dam.lt could not be justified to assume that these vegetations on both sides of a certain poi nt at the dam were identical at the time the dam was closed.lt was considered too late to study probable changes in these types of vegetation.Selection of plots was done during a few trips in an air-propeller boat.From the highly elevated chairs it is easy to Jook over the dam.Three "types" of swamp wood and one type of marsh forest cou ld be then recogn!zed.O. Marshy ridge forest.
On pair of sample plots (one plot insidFJ and one outside the dammed up swamp, and opposite to each other) was chosen in each of these types as follows.With the aid of the aeri al photographs taken in 1972 a unit of each vE.getation type was chosen which looked homogeneous in structure, texture and grey tone , and which was intersected by the dam.Within each unit a sample plot was chosen.Thi5 was considered the most justified method of selecting pairs of sample plots which werc very probably comparable at the time the dam was constructed and closed.In this way th ~ following plots were selected: For location of these plots see vegetation m3p in back flap.

. Sample method
In September, and October, 1973 ali plots were sampled as well as possible.At thet ti me.access to the plots inside the dammed swélmp was rather difficult because, in addition to the discomfort caused by clouds of mosquitoes, the area was covered with ll mixture of dead wood, floating pegasse, and water reaching to breast or shoulder height (see photo 4).In each plot, far enough from the zone influenced by the vegetation of the dam e: nd the excavation canal, a spiral trajectory was walked outwardbound until no new species were observed.The area seen in this manner varied in extent from 100 m 2 12A -(Erythrina swamJ) wood) to SOO m 2 (marsh forest).From the plots the following dat<' l were collected: a. Number and date of record.b.Stratification.In each plot outside the dammed up swamp the following strata have been distinguished: one or two tree layer (s) .a shrub-layer, a herb-layer.a layer of rooting and or floating aquatics (if any) .In each plot inside the stor3ge basin: one o r two tree-layer (s) .a shrub-layer, a layer of plants rooting in floo• ting peat and/ or on dead wood and a layer of rooting and/ or floating aquatics.In ali layers.epiphytes, lianas and vines reaching 1hose layers were included.

Description of sample plots
The vegetation plots were described and compared for the first time during the period of September-October, 1973 (see tables A-DJ.In December.1974, another visit was paid to the plots.In the meantime changes have been described below.

Swamp wood
The vegetation changes which took place in A s. B s and C s are well illustrated by comparing tables and photos of A s and A n (Table A, photos 2-5), B s and B n (Table B, photos 6-9) and C s and C n (Table C, photos 10-13).
1!1 this description ali three plot pairs are considered together as the noticeable ch:mges are rather similar, especially in the shrub-layer and lower layers.

a. Tree layers
In plot A s most Pterocarpus (watra-bebe; and Tabebuia (zwamp-panta) trees survive<.lbut crowns became partly deciduous during high water leveis (photo 3).About 10% ot these trees fell over (photo 5) during the period of observation (September 1973) • December 1974) I possibly by weakening of the clay and peat layer in which they root.From the fallen trunks new shoots (photo 3) were observed in Decemberl 1974.In plot B salmos~ ali Pterocarpus and Tabebuia treelets died immediately.
In C s the dominating Erythrina (kofimama) trees survived although their crowns lost many leaves during high water leveis (photo 1 O and 11) .Some trees fell over but imrnediately started to make new robust suckers in an explosive wayl just like fallen Erythrina trees react after a peat fire when they alsa fali over.Also root suckers played a role!ir. the ultimately developedl dense undergrowth (photo 12).Trees which still stood upright have survived in another way.Around the trunks, at the height of the high water leve!, floating, corky breathing roots (pneumatophores) were observed in Octoberl 1973 (photo 13).These roots bent down during lower Notes on t he vegetation ... water leveis and developed as prop roots.which were seen for the first time in December, 1974 (photo 14) .Triplaris (miraoedoe) .Viro/a (baboen) and Cecropía (boesipapaja) trees survived extreme inundation very well .A Ficus tree fell over .Euterpe (pina) palms died after a struggling year with cnly a few leaves left.Eight species of cpiphytes, Phllodendron linnaei, Philodendro11 cf.acutatum, Philodendron pedatum, Philodendron jenmanii, Monstera oblíqua, Monstera cf .expilata, Monstera 14472 and Syngonium vel-/ozianum var.poeppigii (ali Araceae) were found on both sides of the dam therefore they survived very well .Not so with the lianas Entada polystachya and Dioclea virgata (pikin kaw-ai), as they were not seen in the dammed up swamp .

b. Shrub /ayer
Montrichardia arborescens (mokomoko) and Heliconia marginata (paloeloe) survivea high inundation; they were recorded on botn sides of the dam.Between September/ Oct ober , 1973 and Oecemberl 1974, Montrichardia ar• borescens stems in A s and B s seemt::d to have increased somewhat as well in numbe:-as in height.In C S 1 however, Montrichardia rece ived strong competition from new Erythrina glauca (kofimama) shoots (photo 12) .Four species of the scrub layers from northern plots: Costus arabicus (sangrafoe).Piper divaricatum and the ferns Acrostichum danaeifolium (tabaka-tiki) and Thelypteris serrata were not recorded from the southern plots .Epiphytes recorded on both sides of the dam are: Peperomia elongata var.piliramea lsecond record from Suriname) and Margraviõ coriacea One unknown vine recorded in the northern plots was not SP.en in the southern ones .

e. Layer of aquatic plants
A few specimens of three species of floating aquatic plants, the Nile lettuce Pistia stratiates, and two spec ies of duck weed Spirade/a biperfarata and Lemna minar, were found in northern plots.Pistia has especially developed explosively in southern ones.Together with the above mentioned species and Sa/vinia auriculata, Walfiella lingulata and a few specimens of Geratapteris thalictraides there, it forms a dense layer on the water surface not occupied with floating pegasses and dead wood.The water li ly, Nymphaea 130-blanda, has not been recorded from thr. southern piots.
The thickness of the peat layer in the northern, as well as in the southern, sample plots amounted to 30 em.The fluctuation o f lne water levei and water depth (above peat layer) for plot A n and A s are given in fig. 5.The fluctuations are similar for plot pairs B and C, with the restriction that during high water leveis the water in B s and C s is estimated to be 5-15 em deeper than in A s, duo to the former wate rshed between the Nanni Creek and the lv1aratakka River; this is also célused by the inclined water surface going down in the direction of the Nanni Dam.
To find out why some swamp woods within the Nanni storage basin still do not seem to be intluenced by the new hydrological behaviour, a healthy lookmg Pteracarpus Tabebuia swamp wood along the western part of the cana l was visited.The swamp wood appeared to grow on a somewhat higher elevation (20-40 em measured) .

Marshy ridge forest
The vegetation changes in D w are illustrated in Table D and photo 14-17.lnside the dammed up swamp (O w) most trees from both t ree-layers died, such as Manilkara bí• dentata (boletri), Spondias mambin (mope), older Garapa (krapa) trees and Euterpe (pina) palms.For other trees in D w it was too late to recognize individual species during the first visit in September, 1973.Trees such as Geiba pentandra (kankantri), the palm Atta/ea maripa (maripa) , young Garapa (krapa) individuais, Rinaraa pubiflara and Gustavia (konikonioedoe) treelets barely survived the first long period of inu ndation.Triplaris surinamensis (mira-oedoe) and Cecropia obtusa (boesi-pa• paja) resisted it very well.New Triplaris seedlings carne up, justas often happens when a swamp forest is destroyed by peat fire.A new shrub and herb-layer developed, tangled with twiners.Jpamaea phi/lomega (patat!ltite) developed explosively and by December, 1974, it covered almost the entire herb and Rhrub-layer and even the dead trees (photo 17) .
Triplaris surinamensis (mira-oedoe) seedlings, however, were able to compete with the new herbs and tanglers.After 13 months the young Triplaris trees already measured up to 3 met ers in height (photo 17) .The floating aquatics observed in October, 1973, were not seen again in December, 1974.
In May, 1972, the water depths ranged from 65-115 em depending on the distance to the ridge crest.In October, 1973, water depth varied from 40-90 em.In May, June and December, 1974, the soil fel I dry.

MAPPING OF ATTACKED AND DROWNED VEGE'l'ATION
From aeria l surveys the area containing the attacked and drown swamp scrub, swamp wood, swamp and marsh forest is sketched in the vegetation map (in back flap).Within this area of 190 km 2 about 1 O km 2 of marsh fores't drowned while another 60 km 2 of swamp scrub, swamp wood and swamp forest were severely attacked, or even died.

ANO PEAT GROWTH
There are indications that recently drowned marsh forests (now 10 km 2 ) will be succeeded by Triplaris (mira-oedoe) swamp forest.
Most attacked swamp forests (now 60 km 2 ) wi ll not survive because mature trees are starting to fali over while the natural tree regeneration is being drowned.A new herbaceous or shrubby swamp vegetation will possibly develop from this vegetation.Nothing can be said of the speed of vegetational succession towards the new clímax: possibly a two-story swamp forest on a thick peat layer.
Vegetational succession is accompanied by peat growth.Since early times, during severely dry periods, the vegetation was damaged locally, and peat was decomposed by fire.This brought the vegetation back to earlier stages of development, while it deepened the swamp and increased the water-storage capacity of the basin.
Notes on the vegetation .. .Recapitulating , it can be said that peat growth will be in balance with peat decomposition if decomposition by fire and :iving decomposers during low water leveis ( < + 9. 20 m N. P.) compensates for peat accumulation durlng high water leveis ( > + 9.20 m N. p.).
Only if the swamp levei will fluctuate élround + 9. 20 m N. P. for approximately equal time periods above or below this levei, and if leveis > + 10 .30 m N .P. can be avoided, peat accumulation and decomposition wi 11 be in balance.Beca use this will not be quite desirable for the water supply, in the long run a net peat growth has to be expected.This peat growth can be measured in permanent plots.lf necessary and in a justified way peat growth can be nullified only by controlled burning.RESUMO O pântano de Nanni, no Surinarne, é urna.área natural, medindo 1700 km.2.Segundo os dados disponíveis, a vegetação e a camada turfosa têm sido fortemente influenciadas pelo fogo.A partir do início deste século, ::> pântano vem sendo usado corno fonte de abastecimento de água para a área de rizicultura que circunda a vila de Nieuw Nickerie A medida que a área cultivada foi sendo gradual-mentP ampliada, proporcional deficit de água ocorreu.Em decorrência disso, o :;uprirnento de água foi melhorando com a construção de barragem e canais de irrigação, e o comportamento hidrológico do pântano alterou-se consideravelmente.

146-Teunissen
Photo 13: Cut-of corky, floating air roots as formed at the trunks of Erytirrina trees at high water level.October 1973 . ..
lt is an intermediate stage between open herbaceous swamps with scattered shrubs and treelets, and swamp wood, which is defined here as a low, open to closed, one story forest, 5-15 m high.In species composition swamp scrub has an intermediate position between open herbaceous swamps and (low) swamp wood.D. Swamp wood and swamp forest Lindeman (1953) mentions open swamp wcods south of the Nickerie River.These oper.woor!s have a dense undergrowth ciosely rela-Th e followíng suggestions are made based upon recent field observations by the present author: ( 9 J -Luzi o la spruceana .(10) -Saccíolepis striata and / or Hymenachne amplexicaulís.(11) -Dianthera obtusifolia.(1 2) -Laersia hexandra.( 13) -probably Xyris macrophala.ted to the veget ation of the open swamps in the vicinity, often with numerous vines ( Lindeman and Moolenaar, 1959).
Along t he western part of the dam: A. High Pterocarpus-Tabebuia swamp woods Along the central part of the canal: B. Low Pterocarpus-Tabebuia swamp woods (in between Pterocarpus-Tabebuia swamp scrub and herbaceous swamp vegetation).Along the eastern end o f the canal:C.Erythrina swamp woods (in between 11erbaceous swamp vegetation).
A .Pairs As ( = south) and A n (= north) in + 12 m high Pterocarpus-Tabebuia swamp wood.B. Pairs B s and B n in low (± 6-8 m) Pterocarpus-Tabebuia swamp wood.C. Paris C s and C n in a more or les closed Erythrina swamp wood ± 8 m high.O. Pairs O w ( = west) and O e ( = east) in closed marshy ridge forest.
c. Heigth in m and % of coverage d.List of species per layer c.Frequency of each species using the following scale : = liana (woody) gr = graminoid pln ~t aq = aquatic si = seedling t = dead jv = juvenile Water depth, peat thickness in em h .Oblique aerial photographs and terrestrial photographs were taken.lf necessary, plant specimens were col lected for identification and herbarium study .Two almost compl ete sets were collected, of which the first has been deposited in the Herbarium of the Surinam Forest Service in Paramaribo and the other in the Botanical Museum and Herbarium of the State University of Utrecht.

Future
peat growth depends on future water fluctuations which can be managed now to a certa in degree.a.After a few weeks with swamp leveis lower than + 8.50 m N.P. (at Nanni Dam) the upper peat layers dry out to such an extent that their rapid destruction by fire becomes possible.b.At swamp leveis lower than + 9. 20 m N .P. decomposition of accumui3ted pe~t can take place by aerobic organisms.c .At swamp leveis between + 9. 20 m N. P. and + 1 O. 30 m N. P. litter will accumulate and turn into peat.d.After a few weeks with swamp leveis higher than + 1 O. 30 m N. P. a quick accumulation of debris can take place as more woody vegetation dies which cause3 an extra quantity of organic matter to fali into the water.

F
. C. Bubberman and Dr. J. P. Schulz for their permission to cooperate in the hydrologicai project; Drs.M.C.M. Teunissen-Werkhoven and Ms. C.J. Vreden of the herbarium for their assistance in the field as well as for most taxoncmical identifications; Mr. R. O. de Graav and his staff for drawing the illustrations and Mr. R. N. Kromoidjojo for his photographic work.I would like to express my special thanks to Dr. O. C. Geijskes of the Rijksmuseum van Natuurlijke Historie at Leiden, Hollsnd, for his willingness to place his manuscripts, maps and photographs at my disposal; also to the Staff of the Botanisch Museum en Herbarium of the State University of Utrecht, Holland for additional plant identifications.The manuscript was kindly reviewed by Dr. J. C. Lindeman.I am greatly indebted to Ir. R.J. Sevenhuysen for the fruitful cooperation and to his wife for her Nickerian hospitality.Finally I thank Mr. H. A. Reichart M. Se. for the correction of the English text.
Photo 4: Pt.erocarpus-Tabebuia -;wamp wood.Plot As.September 1973.Water levei at breast height .entire water surface is covered with Pistia str atiotes and other floating aquatic plants.