Pores;f;ology

Management
ELSEVIER Forest Ecology and Management 72 ( 1995) 6 l-69

Edge-interior differences in vegetation structure and composition

in a Chaco semi-arid forest, Argentina
Javier Lopez de Casenavea,*, Juan Pablo Pelottob, Jorge Protomastro”
“Depto. de Biologlia, FCEyN, Universidad de Buenos Aires, Piso 4, Pab. 2, Ciudad Universitaria, 1428, Buenos Aires, Argentina
bCentro de Estudios Farmacokigicos y Bothicos, CONICET, Serrano 66.5 Piso 7, 1414, Buenos Aires, Argentina
“Universidad National de Salta, 4400, Salta, Argentina
Accepted3 May 1994

Abstract

The differencesin speciescomposition and horizontal and vertical vegetation structure betweena natural forest
edgeand the adjacent forest interior were analyzed in an old-growth Chaco forest of Argentina. Total basalareain
the shrubstratum wasslightly greater at the edge,whereastotal basalarea in the arboreal strata wasslightly greater
in the forest interior. The abundanceof trees and thick stemswere significantly greater in the interior, and total
stemdensity, speciesrichnessand abundanceof thin stemswere significantly greater at the forest edge.Compared
with the interior, the edgeshowedgreater cover at low heightsin the shrub layer and lower cover in the canopy.
Edge-interior differencesare similar to those reported for both tropical and temperate forests, and appearto be
related mainly to the environmental conditions at edgesand perhapsto vertebrate-related dispersionmechanisms
associatedwith edges.The similarity of this forest edgewith other modified areaselsewherein the Chaco forest
and the relationship with fragmentation processes are discussed.

Keywords: Forestfragmentation;
Edgeeffect;Conservation;
Neotropicalforest;Old-growthforest

1. Introduction an increment in light intensity and duration and,

therefore, both soil and air temperatures can be
Studies of forest edgeshave been traditionally both higher and more variable (Wales, 1972;
linked to analysis of ecotones (Ranney et al., Ranney et al., 1981; Lovejoy et al., 1986; Wil-
1981). Some authors have emphasized recently, liams-Linera, 1990a; Brothers and Spingarn,
however, the compositional, structural and func- 1992). High temperatures and wind action de-
tional differences between the edges and the in- crease the relative humidity of both soil and air
teriors of forests (Wales, 1972; Ranney et al., at edges,coupled with low rates of transpiration
1981; Whitney and Runkle, 1981; Williams-Li- (Lovejoy et al., 1986; Williams-Linera, 1990a;
nera, 1990a,b; Chen et al., 1992). Brothers and Spingarn, 1992 ).
Compared with the forest-interior environ- Vegetation associated with forest edges shows
ment, solar incidence on the edges brings about several features in response to edge conditions.
Light exposure stimulates germination and en-
*Corresponding
author. hances growth of pioneer or shade-intolerant
0378-l127/95/$09.500 1995ElsevierScience B.V. All rights reserved
SSDIO378-1127(94)03444-3
62 J. Lopez de Casenave et al. /Forest Ecology and Management 7-7(1995) 61-69

species (Wales, 1972; Levenson, 1981; Ranney temperature is near 22 ‘C, with high mean max-
et al., 1981; but seeWilliams-Linera, 1990a). In- imum temperature (more than 30°C). During
creased light availability results in higher pro- the summer we recorded an absolute maximum
ductivity, and hence in higher densities and basal temperature of 46°C in the forest, whereas ab-
areas (Wales, 1972; Ranney et al., 198 I; Lovejoy solute minimum temperatures are less than
et al., 1986; Williams-Linera, 1990a). Edge - 7‘C in winter. Eighty percent of the 635 mm
plants frequently have greater fruit production, of annual precipitation falls through October-
higher numbers of fruiting individuals, longer March, although year-to-year variation in rain-
fruiting periods and greater propagule and seed fall is large.
outputs (McDiarmid et al., 1977; Ranney et al., The soil of the forest is a well-drained silt loam
1981). Additionally, edges show higher species (Morello and Adamoli, 1974). In this region the
richness and increased tree mortality (Ranney et forest is frequently interrupted by belts (approx-
al., 1981; Lovejoy et al., 1986; Williams-Linera, imately 1 km wide) of natural grasslands of
1990a). Elionurus spp. arranged in a NW-SE direction.
Although detailed descriptions of plant com- These grasslands grow in low topographic posi-
munities and phytosociological studies of Chaco tions relative to the forest, occupying ancient
forests are relatively common (see review in river beds filled with alluvial sands (Morello and
Prado, 1993 ), the structure and importance of Adamoli, 1968, 1974). The change in edaphic
edges had never been assessedin these South conditions at the forest-grassland border is rela-
American forests. In this paper we analyze the tively abrupt.
differences in speciescomposition and horizon- Our study site was located on an old-growth
tal and vertical vegetation structure between the forest area adjacent to a natural grassland (ap-
edge and the interior of a Chaco semi-arid forest proximately 1300 m wide). Sampling was con-
of Argentina. ducted along the border (forest edge) and within
the forest. The forest interior was always sam-
pled more than 200 m from the border.
2. Study area

We conducted this study in the Copo Natural 3. Methods

Reserve (25”55’S, 62”05’W), located in the
Province of Santiago de1Estero, Argentina. The Seven plots ( 15 m x 15 m) were randomly lo-
reserve lies in the semi-arid Chaco forest, which cated at both edge and interior during Septem-
is included in the Occidental District of the Cha- ber-October 1991. Diameter at breast height
quenian Biogeographic Province (Cabrera and (DBH) of trees over 1.5 m tall, and diameter at
Willink, 1980). This large, xeromorphic old- 20 cm height (DBS) of every stem of shrubs and
growth forest (‘Quebrachal of two quebrachos’; saplings of tree species (less than or equal to 1.5
seePrado, 1993 ) is dominated by red quebracho m tall) were recorded in each sample plot. With
(Schinopsis quebracho-colorado) and white que- these data we calculated basal areas (m’ ha-’ ),
bracho ( Aspidosperma quebracho-blanco ) , which tree density (trees ha- ’ ) , and stem density in the
reach heights of 16-l 8 m. Mist01 (Ziziphus mis- shrub stratum (stems ha- ’ ). We were able to
to/) is the main subcanopy tree (up to 8- 10 m ), analyze separately both the shrub and the arbo-
and Capparis retusa, together with the thorny real (canopy and subcanopy, see below) strata,
shrubs Acacia praecox and Ceitis pallida charac- providing that those individuals of tree species
terize the shrub layer. with DBH> 45 mm reached heights of over 4 m,
The climate is continental semi-arid, remark- so entering into the arboreal layers.
ably seasonal,’with a distinct dry season in au- We calculated values of several variables
tumn-winter (April-September), and a warm measuring features of horizontal structure and
and wet season (October-March). Mean annual cover at each plot: two kinds of basal area (total
 J. Lopez de Casenave et al. /Forest Ecology and Management 72 (I 995) 61-69 63

basal area for arboreal individuals and total basal sectswere randomly located in both the edge and
area in the shrub stratum), total number of trees, the interior of the forest. We sampled ground and
number of trees with DBH > 200 mm, and num- shrub strata (O-4 m) at 60 random points on
ber of stemsin each of four stem sizeclasses(less each transect, by erecting on each point a rod
than 25,25-50,50-100, and 100-200 mm DBS). marked at 0.25 m intervals and recording the
Moreover, we considered the number of species height of contacts of the vegetation on the rod.
present in the plot, and stem diameter diversity, Canopy and subcanopy strata (over 4 m) were
calculated as sampled at 15 random points on each transect,
D= l/Cp’ but in this case the rod was marked at 2 m
(1) intervals.
where pi represents the proportion of stems with We calculated vegetation cover at each of the
diameters in the ith diameter class, based on 5 four principal strata present in the forest
mm sizeclasses. (ground, O-O.5 m; shrub, 0.5-4 m; subcanopy,
Vertical structure was measured along 16 tran- 4-8 m; canopy, over 8 m) as the percentage of
sects 15 m long during March 1992. Eight tran- points with contacts at each height. Two vegeta-

Table 1
Mean basal area ( m2 ha- ’ ) of woody species on plots from the edge and the interior of the forest at Copo Natural Reserve

Interior Edge

Canopy trees
Aspidosperma quebracho-blanco ( Apocynaceae) 6.13 * 2.75
Schinopsis quebracho-coforado (Anacardiaceae) 1.89 2.01
Subcanopy trees
Ziziphus mistol (Rhamnaceae) 6.35 5.65
Prosopis nigra (Leguminosae) 0.34 * 1.45
Shrubs
Schinusfasciculatus (Anacardiaceae) 0.63 0.61
Capparis retusa (Capparidaceae) 3.03 * 1.39
Capparis atamisquea (Capparidaceae) 0.26 ** 1.09
Capparis saficifofia (Capparidaceae ) 0.17 0.08
Capparis speciosa (Capparidaceae ) 0.05
Capparis tweediana (Capparidaceae) -a
Acacia praecox (Leguminosae ) 2.01 * 0.86
Acacia furcatispina (Leguminosae) 0.58
Mimosa detinens (Leguminosae) 0.16
Cercidium australe (Leguminosae ) 0.13
Ceftis paffida (Ulmaceae) 1.01 *** 2.19
Achatocarpus praecox (Achatocarpaceae) 0.16 * 1.22
Castela coccinea (Simaroubaceae ) 1.47 * 0.62
Jodina rhombifofia (Santalaceae) -a -a
Maytenus spinosa (Celastraceae) 0.13 0.10
Low shrubs
Solarium sp. (Solanaceae) 0.02 *** 0.30
Lycium spp. (Solanaceae) 0.02 *** 0.13
Afoysia chacoensis (Verbenaceae ) **** 0.22
Afqvsia casadensis (Verbenaceae ) 0.03

Total 23.61 21.64

* Basal area values less than or equal to 0.0 1 m2 ha- ’ .
Asterisks indicate significant differences between sites (Mann-Whitney U Test, n = 7): *Pi 0.05; **P-K 0.0 1; ***PC 0.005;
****p<o.o01.
64 J. Lopez de Casenave et al. /Forest Ecology and Management 72 (I 995) 61-69

Table 2
Mean basal area and mean tree density of woody species present in canopy and subcanopy, from plots located in the edge and
the interior of the forest at Copo Natural Reserve

Species Basal area Tree density

(m’ha-‘) (trees ha- ’ )
_~_-____---___^.
Interior Edge interior Edge

Aspidosperma quebracho-blanco 6.10 2.61 152.38 88.89

Schinopsis quebracho-colorado I .84 2.00 69.84 44.44
Ziziphus mist01 6.32 5.62 304.70 ! 33.33
Prosopis nigra 0.34 1.43 6.35 19.05

Total 14.59 Il.66 533.33 285.71

Table 3
Mean basal area and mean stem density of woody species present in shrub stratum, from plots located in the edge and the interior
of the forest at Copo Natural Reserve

Species Basal area Stem density

(m* ha-‘) (stems ha-‘)
__--
Interior Edge Interior Edge

Aspidosperma quebracho-blanco 0.03 0.14 558.73 1301.59

Schinopsis quebracho-colorado 0.06 -a 584. I3 139.68
Ziziphus mist01 0.03 0.04 285.71 247.62
Prosopis nigra 0.02 69.84
Schinusfasciculatus 0.63 0.61 1053.97 774.60
Capparis retusa 3.03 1.39 4520.64 3568.25
Capparis atamisquea 0.36 1.09 1257.14 3644.44
Capparis salicifolia 0.17 0.08 1377.78 558.73
Capparis speciosa 0.05 171.43
Capparis tweediana - -a 31.75
Acacia praecox 2.01 0.86 1619.05 3466.67
Acacia furcatispina .- 0.58 38.10
Mimosa detinens .- 0.16 755.56
Cercidium australe 0.13 63.49
Celtis pallida 1.01 2.19 10463.49 14165.08
Achatocarpus praecox 0.16 1.22 469.84 4730.16
Castela coccinea 1.47 0.62 1955.56 1149.21
Jodina rhombifolia -a -a 126.98 38.10
Maytenus spinosa 0.13 0.10 571.43 488.89
Solanum sp. 0.02 0.30 742.86 4914.29
Lycium spp. 0.02 0.13 1250.79 3244.44
Aloysia chacoensis 0.22 4165.08
Aloysia casadensis 0.03 863.49

Total 9.02 9.99 26838.10 48590.48

--~
a Basal area values less than or equal to 0.01 mz ha-‘.

tion profile diversities (foliage height diversity, height interval i with respect to total coverage of
FHD) were calculated with Eq. ( 1), using as pi the profile. We considered the cover of the four
values the proportions that represent coverage of principal strata in one case (ground, shrub, sub-
 J. Lopez de Casenave et al. /Forest Ecology and Management 72 (1995) 61-69 65

Table 4
Values of vegetation variables measuring structure and coverage on plots and transects from the edge and the interior of the
forest at Copo Natural Reserve. Table values are means of seven plots (first ten variables) and of eight transects (last six vari-
ables) in each site

Variables Interior Edge

Basal area by stratum ( m2 ha- ’ )

Arboreal layers 14.59 11.66
Shrub layer 9.02 9.99
Tree abundance (number per plot)
Total 12.00 * 6.43
Trees > 200 mm DBH 3.43 2.86
Stem abundance (number per plot)
< 25 mm DBS 517.14 *** 1001.00
25-50 mm DBS 66.00 78.14
50-100mmDBS 28.43 ** 16.14
100-200 mm DBS 6.14 3.71
Species richness (spp. per plot) 13.57 **** 16.71
Stem diameter diversity 5.11 4.20
Percent cover by stratum (O/o)
Ground layer (O-O.5 m) 38.75 40.52
Shrub layer (0.5-4 m) 15.06 18.35
Subcanopy layer (4-8 m ) 21.92 30.00
Canopy layer ( > 8 m) 26.67 ** 12.50
Foliage height diversity
All intervals 11.85 11.38
Four layers 3.18 * 2.64
Asterisks indicate significant differences between sites (Mann-Whitney UTest ): *PC 0.05; **I’< 0.01; **P<O.OOS; ***Pi 0.001.

canopy and canopy, four layers FHD), and all area between sites (Table 1). Aspidosperma que-
height intervals (0.25 m intervals up to 4 m, 4- bracho-blanco, Capparis retusa, Castela coccinea
6, 6-8 and over 8 m) in the other (all layers and Acacia praecox had greater total basal area
FHD), so avoiding assumptions about the num- in the interior, while Celtispallida, Achatocarpus
ber of vegetation layers actually present in the praecox, Capparis atamisquea, Solanum sp., Ly-
forest. cium spp., Aloysia chacoensis and Prosopis nigra
Differences in basal area, stem density, tree had greater basal area at the edge.
density, and structural features between interior Four tree speciesform the forest canopy and
and edgesampleswere evaluated with the Mann- subcanopy (Table 2 ) : Aspidosperma quebracho-
Whitney U test. blanco, Schinopsis quebracho-Colorado, Ziziphus
mist01 and Prosopis nigra. The total basal area
contributing to these layers was slightly greater,
4. Results and tree density was significantly higher
(PC 0.01) in the interior than at the edge. Aspi-
4. I. Coverage and composition dosperma quebracho-blanco and Ziziphus mist01
were better represented in the interior of the for-
Mean ( rt SD) total basal area of the interior est, while Prosopis nigra was commoner at the
plots was 23.612558 m2 ha-‘, whereas edge edgethan in the interior.
plots had a mean total basal area of 2 1.64 t- 4.88 Twenty-three woody specieswere found in the
m2 ha- ’ (Table 1) . The difference was not sig- shrub stratum, including saplings of the four tree
nificant (P= 0.26 15). Nevertheless, several plant species (Table 3 ) . All of these specieswere pres-
species showed significant differences in basal ent in edge plots, but only 15 were present in the
66 J. Lopez de Casenave et al. /Forest Ecology and Management 72 (I 995) 61-69

was due to the same reason, even though adult

trees were better represented in the forest inte-
rior (Table 2 ) .

4.2. Horizontal structure

Differences in coverage and horizontal struc-

ture between forest interior and edge (Table 4)
were associated with greater abundance of trees
and stems 50-100 mm DBS in the interior, and
with increased species richness and number of
thinnest stems in the edge. Allocation of overall
basal area between arboreal and shrub layers was
opposite in both sites: interior plots had high
cover in arboreal layers and low in shrub stra-
tum, while the edge plots presented high cover in
shrub stratum and low in the arboreal layers
(Table 4).

4.3. Vertical structure

i0 LO Vertical distribution of vegetation in the inte-

“I. COVER
rior of the forest (Fig. 1) showed high ground
Fig. 1. Average profiles of foliage cover in the edge and inte- cover, followed by a peak at 1.75 m (because
rior of the forest at Copo Natural Reserve, expressed as the nearly all shrub speciesconcentrate foliage at this
percentages of points with touches at each height (see
Methods). height ), and then a slow decreaseup to 4 m high.
The low portion of subcanopy showed more
amount of vegetation, mainly of Ziziphus mistol.
interior of the forest. Stem density was signifi- Canopy cover, finally, was due to both que-
cantly greater (PC 0.001)) and total basal area bracho species (Aspidosperma quebracho-blanco
in the shrub stratum was only slightly greater at and Schinopsis quebracho-colorado ) .
the edge. Comparatively, the edge of the forest showed
Capparis retusa, Capparis salicifolia, and Cas- greater cover at low heights in the shrub layer,
tela coccinea showed greater basal areas and stem where low shrub species absent in the interior
densities in the forest interior, while at the edge (Solanum sp., Lycium spp., Aloysia chacoensis,
the same was true for Aspidospermu quebracho- Aloysia casadensis) increased foliage cover up to
blanco, Celtis pallida, Achatocarpus praecox, 2 m high (Fig. 1). Over 2 m, cover was slightly
Capparis atamisquea, Solanum sp., Lycium spp., lower in the edge, owing almost exclusively to the
Aloysia chacoensis, Acacia Jicrcatispina, Cerci- absence of mature Acacia praecox individuals,
dium australe and Mimosa detinens (Table 3 ) . which typically reach 3.5-4 m high in interior
Acacia praecox had smaller basal area but greater plots (this species was represented almost en-
stem density at the edge because large numbers tirely by seedlings in the edge). Percent cover in
of seedlings and saplings in this site increased the the shrub layer (OS-4 m) did not differ signifi-
number of stems with minor contribution to to- cantly between forest interior and edge (Table
tal basal area (J.P. Pelotto and J. Lopez de Cas- 4)) but this resuh was due to the opposite trends
enave, personal observation). Greater coverage above and below 2 m highbetween both sites (see
of Aspidosperma quebracho-blanco in the shrub Fig. 1) . Ground and subcanopy cover was mark-
layer of the edge in relation to the interior plots edly similar between sites, but the absence of
 .I. Lopez de Casenave et al. /Forest Ecology and Management 72 (1995) 61-69 61

canopy trees at edges (only a minor contribution miento, 1963; Morello and Adamoli, 1974).
of Aspidosperma quebracho-blanco) determined Other shrub species associated with edges and
an open canopy with significantly less cover in well represented in second-growth forests and
this site (Table 4 and Fig. 1) . openings are Capparis atamisquea, Celtis pallida
Foliage height diversity values are a measure and Achatocarpus praecox. LOW shrubs species
of heterogeneity of vertical cover, and they were (Solanum sp., Lycium spp., Aloysia chacoensis
always lower in the edge (Table 4)) where foliage and Aloysia casadensis) are able to take advan-
cover was more concentrated at certain heights tage of edge environmental conditions for estab-
(O-2 m, 4-6 m, see Fig. 1). These values were lishment, and grow exclusively on this site. The
statistically different when they were considered association with edgecould be enhanced in some
for a priori defined strata (four layers, Table 4)) species as a consequence of differential disper-
but they did not differ when all the sampling lay- sion rates by vertebrates, principally birds. Cer-
ers were taken into account (all layers). The use tain dispersion-related mechanisms like frugi-
of predetermined strata could enhance the small vory are particularly important in forest edges
actual differences in vertical heterogeneity be- (McDiarmid et al., 1977; Thompson and Will-
tween both forest sites. son, 1978). The abundance of fruit-eating birds
is significantly greater in this edge than in adja-
cent forest interior (J. Lopez de Casenave et al.,
5. Discussion unpublished data, 1992). In other areas within
the Chaco, some speciesare associatedwith trail-
The edgeof the forest studied here is a mature forest borders due to the dispersion of their fruits
one (sensu Ranney et al., 1981), with an edaph- eaten by cattle (M. Aizen, personal communi-
ically controlled boundary with the adjacent cation, 1993). However, interior plots are char-
grassland (Morello and Adamoli, 1968). It is acterized principally by quebracho-blanco trees
characterized structurally by a very dense cover (Aspidosperma quebracho-blanco) and three
of thin stems up to 2 m high, with few trees and, shrub species: Castela coccinea, Capparis retusa
therefore, scarceoverstory cover. In contrast, the and Acacia praecox. The last has been recog-
forest interior shows less cover in the shrub stra- nized as a characteristic species of old-growth
tum and a well developed canopy. This trend ex- forests in the Chaco (Sarmiento, 1963). Surpris-
plains the lack of significant differences in total ingly, we observed high seedling density of Aspi-
basal area between both sites:vegetation concen- dosperma quebracho-blanco and Acacia praecox
trates in the lowest layers at edges, and in the at edges.This situation might be related with in-
highest ones in the interior. Moreover, density of creased seed outputs and/or increased germina-
thin stems is greater at the edges, while greater tion rates at edges, coupled with low survival
densities of thick stems and trees are character- there becauseof harsh conditions. Aspidosperma
istic of the forest interior environment. These quebracho-blanco has good regeneration and dis-
features, along with increased species richness, persal under situations like those found in forest
are similar to those reported in previous studies edges (Luti, 1982; Saravia Toledo, 1982), but
for both tropical and temperate forest edges little is known about Acacia praecox establish-
(Wales, 1972; Ranney et al., 1981; Williams-Li- ment, survival, and differential growth rates.
nera, 1990a; but see Chen et al., 1992). This forest edge is structurally and floristically
Floristic differences between sitesare also well similar to modified areas elsewhere in the Chaco
marked. Most edge speciesare colonizers or pi- forest. Second-growth forests (selectively logged
oneers. Prosopis n&a, Acacia furcatispina, Mi- in the 195Os), also present in the study area, re-
mosa detinens and Cercidium australe (all found semble the studied edge, mainly because of the
almost exclusively in edge plots) are typical of high density of thin stemsin the shrub layer and
degraded forests, and they quickly invade open of similarities in speciescomposition (J. Proto-
areas contributing to shrubland formation (Sar- mastro and J.P. Pelotto, unpublished data,
68 J. Lopez de Casenave et al. /Forest Ecology and Management 72 (1995) 61-69

1991). Western zones of the Chaco forest were for agricultural and cattle raising activities (Mo-
subjected to natural recurrent fires before hu- rello and Saravia Toledo, 1959; Luti, 1982, Sar-
man activities became important (Morello and avia Toledo, 1982; Aizen and Feinsinger, 1994).
Saravia Toledo, 1959; Morello and Adamoli, If these trends are maintained, there will be an
1968). Fires originated in lowland areas, reached increment in edgehabitats and in second-growth
the borders of higher ones occupied by tall que- forests which present structural and floristic
bracho forest, thus producing a narrow strip of changes that resemble edge-related conditions.
shrubs (‘brotal’) (Morello and Adamoli, 1968, Knowledge about edge structure and dynamics
1974). These strips were characterized by Cer- therefore becomes essential for conservation and
cidium australe, Prosopis nigra, Mimosa deti- management purposes in these Chaco forests.
nens, Acacia furcatispina and Celtis pallida
(among others), and showed good regeneration
of the tree species, although only with juvenile Acknowledgments
individuals up to 5 m tall (Morello and Ada-
moli, 1968 ). These old fire-controlled edges ap- We are grateful to J. Perez, B. Perez, J. Mal-
peared to be remarkably similar to the edaphi- donado and their families for the great help dur-
tally controlled edge that we studied. In this last ing field work. We also thank M. Aizen, J. D.
case, woody species are unable to colonize the Deans, H. Ginzo, L. Marone, A. Ribichich, and
adjacent grassland growing on alluvial sands, and one anonymous reviewer for providing valuable
there are no signs of recent or past fires that pre- comments on manuscript, and L. Berenguer for
vent colonization. Similarities among forest bor- drawing the illustration. The Consejo National
ders independently of the factors that keep them, de Investigaciones Cientificas y Tecnicas
suggestthat these features are broadly spread out (CONICET) of Argentina, the International
in the Chaco forest when edge conditions are Council for Bird Preservation (Pan American
present. Therefore, it seems likely that condi- Section), and the US Fish and Wildlife Service
tions observed at edges in the study area reflect partially financed this work.
characteristics of both artificial and natural bor-
ders in the whole austral portion of the Chaco
semi-arid forest. References
Edge-effects become a very important compo-
nent of landscape dynamics in fragmented for- Aizen, M.A. and Feinsinger, P.; 1994. Forest fragmentation,
ests (Levenson, 1981; Ranney et al., 1981; Lo- pollination, and plant reproduction in a Cbaco dry forest,
vejoy et al., 1986; Wilcove et al., 1986; Harris Argentina. Ecology, 75: 330-351.
and Silva-Lopez, 1992). Fragmentation reduces Brothers, TX and Spingarn, A., 1992. Forest fragmentation
and alien plant invasion of central Indiana old-growth
the sizeof natural habitat, increasing the propor- forests. Conserv. Biol., 6: 9 l- 100.
tion of edgeareas, so that fragmentation rates are Cabrera, A. and Willink, A., 1980. Biogeografia de America
correlated with the increment of edge-effects Latina, 2nd edn. Monograffas O.E.A. 13, Washington, DC,
(Levenson, 198 1; Lovejoy et al., 1986; Wilcove 117pp.
et al., 1986; Harris and Silva-Lopez, 1992). Semi- Chen, J., Franklin, J.F. and Spies, T.A., 1992. Vegetation re-
sponses to edge environments in old-growth Douglas-fir
arid Chaco forests have been increasingly re- forests. Ecol. Appl., 2: 387-396.
duced, modified and fragmented by human ex- Harris, L.D. and Silva-Lopez, G., 1992. Forest fragmenta-
ploitation and poorly developed management tion and the conservation of biological diversity. In: P.
strategies. Most of these ancient forests in Ar- Fiedler and S. Jain (Editors), Conservation Biology. The
gentina have been degraded or lost in the last 200 Theory and Practice of Nature Conservation+ Preserva-
tion, and Management. Chapman and Hall, New York,
years (Sarmiento, 1963), and remaining virgin pp. 197-237.
areas are subjected to diverse threats, including Levenson, J., 198 1. Woodlots as biogeographic islands in
logging to produce timber, posts, firewood and southeastern Wisconsin. In: R. Burgess and D. Sharpe
charcoal, and clearing of second-growth forests (Editors), Forest Islands Dynamics in Mandominated
 J. Lopez de Casenave et al. /Forest Ecology and Management 72 (I 99s) 61-69 69

Landscapes. Springer, New York, pp. 13-39. Ranney, J., Bruner, M. and Levenson, J., 198 1. The impor-
Lovejoy, T., Bierregaard, R., Rylands, J., Malcolm, J., Quin- tance of edge in the structure and dynamics of forest is-
tela, C., Harper, L., Brown, K., Powell, A., Powell, G., lands. In: R. Burgess and D. Sharpe (Editors), Forest Is-
Schubart, H. and Hays, M., 1986. Edge and other effects lands Dynamics in Man-dominated Landscapes. Springer,
of isolation on Amazon forest fragments. In: M. Soul6 New York, pp. 67-95.
(Editor), Conservation Biology. The Science of Scarcity Saravia Toledo, C., 1982. Chaco semihrido. In: Conserva-
and Diversity. Sinauer, Sunderland, MA, pp. 257-285. ci6n de la vegetation natural en la Repliblica Argentina.
Luti, R., 1982. Chaco semiarid0 y centro. Presentacibn. In: Sociedad Argentina de BotAnica, TucumBn, pp. 29-34.
Conservacibn de la vegetaci6n natural en la Reptiblica Sarmiento, G., 1963. Las comunidades vegetales de1 Chaco
Argentina. Sociedad Argentina de Bottinica, TucumBn, pp. semiarid0 santiagueiio. Doctoral thesis, Universidad de
25-28. Buenos Aires, Buenos Aires, Argentina.
McDiarmid, R.W., Ricklefs, R.E. and Foster, MS., 1977. Thompson, J. and Willson, M., 1978. Disturbance and the
Dispersal of Stemmadenia donnell-smithii (Apocyna- dispersal of fleshy fruits. Science, 200: 116 1-l 163.
ceae) by birds. Biotropica, 9: 9-25. Wales, B., 1972. Vegetation analysis of northern and south-
Morello, J. and Adamoli, J., 1968. Las grandes unidades de em edges in a mature oak-hickory forest. Ecol. Monogr.,
vegetaci6n y ambiente de1 Chaco argentino. Primera parte: 42: 451-471.
Objetivos y metodologia. Serie Fitogeogrifica, IO. INTA, Whitney, G. and Runkle, J., 198 1. Edge versus age effects in
Buenos Aires, 125 pp. the development of a beech-maple forest. Oikos, 37: 377-
Morello, J. and Adamoli, J., 1974. Las grandes unidades de 381.
vegetaci6n y ambiente del Chaco argentino. Segunda parte:
Wilcove, D., McLellan, C. and Dobson, A., 1986. Habitat
Vegetaci6n y ambiente de la Provincia del Chaco. Serie
fragmentation in the temperate zone. In: M. Soul& (Edi-
FitogeogrBfica, 13. INTA, Buenos Aires, 130 pp.
Morello, J. and Saravia Toledo, C., 1959. El bosque cha- tor), Conservation Biology. The Science of Scarcity and
queiio. I. Paisaje primitivo, paisaje natural y paisaje cul- Diversity. Sinauer, Sunderland, MA, pp. 231-256.
tural en el oriente de Salta. Rev. Agron. Noroeste Argent., Williams-Linera, G., 1990a. Vegetation structure and envi-
3: 5-81. ronmental conditions of forest edges in PanamB. J. Ecol.,
Prado, D.E., 1993. What is the Gran Chaco vegetation in 78: 356-313.
South America? I. A review. Contribution to the study of Williams-Linera, G., 1990b. Origin and early development
flora and vegetation of the Chaco. V. Candollea, 48: 145- of forest edge vegetation in Panam& Biotropica, 22: 235-
172. 241.