|UFDC Home||myUFDC Home | Help|
ECOLOGY AND MANAGEMENT OF WETLAND FORESTS DOMINATED BY
Prioria copaifera INT DARIEN, PANAMA
WILLIAM THOMAS GRAUEL
A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY
UNIVERSITY OF FLORIDA
William Thomas Grauel
Dedicated to the memory of Jack Westoby, who knew that forestry is more about people
than about trees.
I express sincere gratitude to my committee chair, Jack Putz, for sharing many
insights on ecology, forestry, and writing. I thank my committee members, Eric Jokela
and Daniel Zarin of the School of Forest Resources and Conservation, Benj amin Bolker
of the Department of Zoology, and Thomas Kursar from the Department of Biology at
the University of Utah, for their time and guidance, and I am very grateful to Dr. Stephen
Humphrey of the School of Natural Resources and Environment for providing funding at
the University of Florida.
I particularly thank Claudia Romero for providing me with much of the Colombian
literature on these forests. Along with Claudia I am very thankful for friendship and
moral support from Geoff Blate, Clea Paz, and Kevin Gould, who, as fellow graduate
In Panama I received support and guidance from scientists and administrators at the
Smithsonian Tropical Research Institute, including Ira Rubinoff, Egbert Leigh, Leopoldo
Le6n, Stanley Heckadon, Rick Condit, Jim Dalling, Elena Lombardo, and Raineldo
Urriola. Manuel Rodes and Jose Solis administered the project and allowed me and my
co-workers to concentrate on the enormous amount of fieldwork. Among the many
people who assisted me in the swamps of Darien, I deeply thank Ricardo Pineda, Ivan
Cabrera of La Chunga, and especially Delfin Jaramillo of Tucuti for their staunch and
reliable fieldwork. Thanks to the personnel of the Panamanian Environmental Authority
in La Palma and Panama City for all their work in keeping the project moving. John
Leigh of the International Tropical Timber Organization agreed to allow me to use
proj ect data for this dissertation and provided timely direction during the four-year
Special thanks to Sarah Dalle for helping me keep faith in the Macondo world that
is Darien, and a thank you to Julie Velasquez-Runk for sharing the burden and for finding
Data for this study were collected as part of a proj ect funded by the International
Tropical Timber Organization (ITTO). The Panamanian Environmental Authority
(Autoridad Nacional del Ambiente, ANAM) and the Smithsonian Tropical Research
Institute (STRI) provided logistical support.
TABLE OF CONTENTS
ACKNOWLEDGMENT S .............. .................... iv
LI ST OF T ABLE S ............_...... .............. ix...
LI ST OF FIGURE S .............. .................... xi
LIST OF OBJECT S ..............._ ..............xiii........ ......
AB STRAC T ................ .............. xiv
1 LITERATURE REVIEW OF Prioria copalfera ....._._._ ........___ ........_.......1
Introducti on ........._ ....... ...............1.....
Distribution ........._ ....... ...............1....
Species Description .............. ...............3.....
Phenology .............. ...............4.....
Ecology ........._..... ....... ...............4....
Wood Uses and Properties............... ...............
Diseases and Insects .............. ...............10....
Y ield .......... _..... .......... ...............11....
Growth and Mortality ........._. ........_. ...............14...
Natural Regeneration ........._.. ........_. ...............16...
Artificial Regeneration .............. ...............17....
Conclusion ........._ ........_. ...............17....
2 EFFECTS OF LIANAS ON GROWTH AND REGENERATION OF Prioria
copaifera INT DARIEN, PANAMA .............. ...............19....
Introducti on ................. ...............19.................
Study Site ................. ...............22.................
M ethods .............. ...............23....
Re sults ................ ...............26.................
Discussion ................. ...............28.................
3 GROWTH AND SURVIVAL OF Prioria copaifera SEEDLINGS PLANTED
ALONG A HABITAT GRADIENT INT A PANAMANIAN SWAMP .....................40
Introducti on ............. ...... ._ ...............40...
Study Site ............ _. .... ...............42....
M ethods .............. ...............43....
Re sults............. ...... ._ ...............45...
Discussion ............. ...... ._ ...............47...
4 STRUCTURE, COMPOSITION, AND DYNAMICS OF Prioria copalfera-
DOMINATED SWAMP FORESTS IN DARIEN, PANAMA............... ................54
Introducti on ............. ...... ._ ...............54...
Study Sites .............. ...............57....
Principal Sites ............. ...... ._ ...............58...
Secondary Sites .............. ...............60....
M ethod s .............. ...............61....
Plot Descriptions .............. ...............61....
Sampling and Analyses .............. ...............61....
Re sults.............. .... ..._ .. ... .. .. ... ...........6
Tree Species Diversity and Stand Structure .............. ...............65....
Cativo Growth, Mortality, and Recruitment................ .............6
Growth of other Tree Species............... ... ...............6
Growth-dependent Mortality of Cativo Trees .............. ..... ............... 6
Cativo Regeneration .............. ...............70....
Discussion ................. ...............70.................
5 GROWTH AND YIELD PROJECTIONS OF Prioria copaifera FROM FOUR
SWAMP FORESTS IN DARIEN, PANAMA ................. .............................91
Introducti on ................. ...............91.................
Study Sites .............. ...............94....
M ethods .............. ...............95....
Re sults ................ ...............100................
Discussion ................. ...............102................
6 GEOGRAPHICAL, ECOLOGICAL, SOCIAL, AND SILVICULTURAL
CONTEXTS FOR CATIVO (Prioria copaifera) SWAMP CONSERVATION
IN THE DARIEN OF PANAMA ................. ...............115........... ...
Introducti on ................. ........... ...............115......
Timber Harvesting in Darien ................. ...............119...............
Forest Conservation Perspectives ................. ...............123................
A MODELING METHODOLOGY USED IN CHAPTER 5 ............. ...................133
B SOURCE DATA FOR CHAPTERS 4 AND 5............... ...............139...
LIST OF REFERENCES ............. ...... ._ ...............141..
BIOGRAPHICAL SKETCH ............. ...... ...............163...
LIST OF TABLES
1-1. Cativo wood properties ................. ...............10........... ...
2-1. Mean (+1 SE) annual diameter growth (mm) of Prioria copalfera one, two, and
two to four years after liana cutting. ............. ...............34.....
3-1. Mean canopy openness at 1.3 m above the ground as estimated with a spherical
densiometer, arranged by seedling age and habitat. ................ ..................5
3-2. Initial mean seedling height and diameter (at 20 cm above the ground; standard
errors noted parenthetically)............... ............5
3-3. Mean annual growth rates. ........._.._.. ...._... ...............51..
4-1. Total plot area measured for different minimum tree diameters and number
of tree species found. ........._.._.. ...._... ...............78...
4-2. Species diversity indices and relative dominance of cativo (Prioria copaifera).....78
4-3. Stem density and basal area of all species (above) and cativo only (below). ..........78
4-4. Incidence (%) of prostrate, inclined, broken stems and sprouts from prostrate
4-5. Forest-wide annual treefall and tree incline rates (i.e., partial uprooting) for small
(above) and large (below) trees for four sites............... ...............79.
4-6. Mean annual diameter growth (mm/year) of cativo trees of three stem types,
based on 1997-2001, 1998-2001, or 1997-2000 census periods. ..........................80
4-7. Mean annual diameter growth (mm/year) of cativo trees. ................. ........._......81
4-8. Ingrowth by stem type. Percentage of recruited individuals from broken stems,
undamaged stems, or sprouts from prostrate and inclined trees. ............. ................82
4-9. Cativo annual recruitment and mortality rates (%) by stem diameter class for four
census periods. ............. ...............83.....
4-10. Annual mortality rates (%) of cativo trees by stem type and stature
for four census periods. ............. ...............84.....
4-11i. Mean annual growth (mm/year) of cativo trees that were alive at the end of the
study and those that died during the study for which there was one or more
years of growth data ................ .............. .................. ...............85
4-12. Abundance of cativo trees < 1 cm dbh by height class............_._. .........._._. ...86
4-13. Annual mortality rates (%) of cativo trees by height class for the period
November 1998 November 1999. ............. ...............86.....
4-14. Mean annual height (cm/year) and diameter (mm/yr) growth rates for
cativo trees < 1 cm dbh, (sample sizes noted parenthetically). ............... ...............86
5-1. Characteristics of 4 cativo-dominated forests in Darien, Panama. ........................ 109
5-2. Cativo volume (m3 ha- ) and volume increment (m3 ha-l yr- ) of four cativo-
dominated forests in Darien, Panama ................. ...............110........... ...
5-3. Total volume yield after 65 years of growth and harvest simulations at three
different cutting cycles for three riverine swamp forests in Darien, Panama. .......110
5- 4. Percentage total volume reduction due to logging-induced damage for three
riverine forests at three cutting cycles ................. ...............110..............
A-1. Parameter estimates for the four sites .............. ...............134....
LIST OF FIGURES
1-1. Distribution of Prioria copalfera........___ .... .._.. ...............2..
2-1. Diameter distributions of ascending lianas in six 25 x 25 m plots in heavily
infested riverine Prioria copaifera forest degraded by repeated entry logging ........35
2-2. Mean (f 1 SE) density of Prioria copalfera regeneration (< 1 cm dbh) in areas
of high (N = 10) and low (N = 6) liana densities. ................ ......... ...............36
2-3. Mean (f 1 SE) Prioria copaifera seedling recruitment censused two years after
liana cutting in three control and three treatment plots. .............. .....................3
2-4. Mean (f 1 SE) annual Prioria copaifera diameter growth based on five annual
censuses of all trees > 4 cm dbh in three control plots and three plots in which
all lianas were cut at the beginning of the study. ............. ...... ............... 3
2-5. Mean (f 1 SE) annual Prioria copaifera diameter growth of cativo based on
five annual censuses according to liana infestation level in control and treatment
pl ots. ............. ...............3 9....
3-1. Mean diameters and heights (+ 1 SE) of planted Prioria copaifera (cativo)
seedlings. .............. ...............52....
3-2. Percent seedling survival, beginning with the first census (November 1997) after
planting (September 1997). ................ ...............53................
4-1. Principal study sites. a) Casarete, b) Sambu, c) Juanacati, and d) Naranzati. ..........59
4-2. Stem types: a) prostrate and inclined and b) vertical sprouts from a fallen stem. ....63
4-3. Darien Province, Panama showing principal sites (S= Sambu, N=Naranzati,
C= Casarete, J= Juanacati) and seven secondary sites. .............. ....................8
4-4. Annual diameter growth (mm/year) ................. ........... ......... ................88
5-1. Growth trajectories of cativo at four sites in Darien, Panama starting
at 10 cm dbh. ........._._ .... ._ ...............111.
5-2. Cativo volume projoooooooooections for three previously logged riverine forests in Darien,
Panam a. ................ ...............112......... ......
5-3. Year 2000 dbh for trees as they attain 60 cm dbh during harvest simulation
of a 20-year cutting cycle. ........._.__...... .__ ...............113..
5-4. Cativo volume projections for an unlogged inland swamp forest
in Darien, Panama. ........... ...............114.....
A-1. Quadratic regression curves fit to relative growth data for the four sites...............134
LIST OF OBJECTS
1 Comma delimited variable Cativo data (as a text file) ........_.._.. ... ......_.._.. .....139
2 Cativo data (as an Excel spreadsheet) ........._._.. ....__.. ...._.._._..........3
Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy
ECOLOGY AND MANAGEMENT OF WETLAND FORESTS DOMINATED BY
Prioria copaifera INT DARIEN, PANAMA
William Thomas Grauel
Chair: Francis E. Putz
Major Department: Natural Resources and Environment
The state of knowledge of Neotropical swamp forests dominated by Prioria
copaifera (cativo) is reviewed based on the available literature. Results of two
silvicultural experiments (liana-cutting and reforestation) are reported, forest-stand
dynamics are described, and growth and yield proj sections are presented.
Permanent sample plots at four sites in three watersheds in Darien, Panama were
installed in 1997 and 1998. Woody plants > 1 cm diameter at breast height (dbh, 1.3 m
above the ground) were tagged, mapped and measured, and censused annually to gather
demographic information on growth, recruitment, and mortality. Cativo regeneration
(trees < 1 cm dbh) was also censused at short intervals to monitor seedling dynamics.
Cativo growth was studied at an additional six sites along an inundation gradient that
varied in water salinity and hydroperiod.
Cativo growth was slow to moderate at sites flooded by tidal waters, moderate at
inland swamps flooded for long periods, and moderate to fast at riverine sites flooded by
fresh water; mean annual diameter growth ranged from 0.05 > 0.8 cm/year. Mortality
and recruitment rates of cativo varied widely among sites and years as well, with
mortality exceeding recruitment at two sites for two of the five years of monitoring. At
one slow growing site, trees that died during the study grew significantly slower prior to
death than trees that lived, demonstrating the occurrence of species-level growth-
dependent mortality for the first time in tropical forest.
Unlogged inland swamps contained very large standing volumes in trees > 60 cm
dbh (up to 180 m3/hectare). Riverine forests that were repeatedly logged during 1950-
2000 contained few large trees, but showed potential for future timber harvests in the
form of abundant regeneration and large standing volumes in trees 40-60 cm dbh. Sixty
year volume proj sections suggest that higher dbh cutting limits and longer cutting cycles
reduce residual damage, and can produce high timber yields for inland swamps and
riverine forests, respectively.
Volatile world timber markets and log shortages may be reducing incentives for
cativo logging (and thus swamp forest conservation) in the face of large development
projects and increased colonization in Darien, Panama. Cativo swamp forests have
important hydrological and carbon-sequestering values that should be incorporated in
LITERATURE REVIEW OF Prioria copalfera
Unlike the vast maj ority of tropical tree species, much is known about the ecology
of the swamp species Prioria copaifera Griseb. (hereafter cativo). Its potential
commercial value was recognized in the 1920s, but thousands of hectares of cativo-
dominated forests had already been cleared to make way for the banana boom of the early
twentieth century. By the end of the century much had been converted to agriculture and
of the remainder, most had been cutover and degraded.
Cativo is found from Nicaragua to Colombia, and is also present in Jamaica (Figure
1-1, Holdridge 1970). Although cativo was included in lists of the tree species of the
coastal region of Ecuador by Rimbach (1932) and Acosta Solis (1947) its presence in that
country is not confirmed. Barbour (1952) reported that the commercial range includes
the Atlantic coast of Central America from Nicaragua to Panama, the watershed of the
Bayano River, the rivers flowing into the Gulf of Darien (Golfo de San Miguel), and the
area around the mouth of the Atrato River in northwest Colombia. Cooper (1928)
referred to enormous stands of cativo in the Valle Estrella of Costa Rica and the Laguna
de Chiriqui in Panama. In 1987, it was estimated that cativo forests covered 49,000 ha in
Panama, with the maj or concentration in the easternmost Province of Darien but
including eastern Panama Province ( INRENARE, Instituto Nacional de Recursos
Naturales Renovables 1987). The same report noted 4000 ha of cativo on Coiba Island
and 17,000 ha in the Chucunaque River watershed; 9000 ha in the combined watersheds
of the Tuira, Balsas, and Marea Rivers; and an additional 4000 ha in adj acent swampy
areas. By 1999, the National Ministry of the Environment in Panama estimated an area
of only 15,000 ha of cativo in the country (ANAM 1999a).
Figure ~ ~ ~ ~ ~ A~s~ 1-.Dsrbto fPira cpafea.
In ColmbiaLinare reo(98 esrbdctv' anea nldnh
waesed fthe Aratoic an en iesinteUub eio nte otwsto h
conr.Acodn oEsoa n Vsuz(98) aio sas oudi h
Deplartet fAtouaadCoo oi(96 ttdta aiowsntfudo
Esoa ndClmba Vasquez (18)rieport98)dsed cativo intelo-y ng aras aroludnd the vlaeo
Santa Marta along the Nechi River and referred to a sample of cativo in the Gabriel
Gutierrez "Medel" Herbarium of the Agronomy Department at the National University in
Medellin collected in the Department of Magdalena in 1949, all of which are on the
Pacific. Colombia contained 363,000 ha of cativo forest (Linares Prieto 1987b, c,
Gonzalez Perez. et al. 1991). Subsequently, the area was estimated at 173,000 ha in 1978
(Linares Prieto 1987b); and in the late 1980s estimates ranged from 60,000 ha (Linares
Prieto 1987c), to 90,000 ha (Linares Prieto 1987b), tol63,000 ha (Linares Prieto 1988).
In Costa Rica, the areas of cativo-dominated forests on the Pacific (Allen 1956) as well as
on the Atlantic side of the country (Bethel 1976) have been severely reduced (Veiman
1982) and today the species is listed as threatened in the country (Jimenez Madrigal
Prioria copaifera is the only species in the genus, and is in the subfamily
Caesalipinioideae of the Fabaceae. Average adult height is 25-30 m, with a dbh
(diameter at breast height, 1.3 m) from 45 tol00 cm (INRENARE 1987a), although the
trees reach 180 dbh (Grauel, unpublished data) and 40 m in height (Del Valle 1972). The
branches of mature trees are somewhat arched and the foliage is distributed uniformly in
a round, thick crown (Escobar and Vasquez 1987). The bark is smooth, brownish-gray to
gray in color, and with abundant lenticels found in continuous horizontal bands
(Echavarria A. and Varon P. 1988). Cativo has no buttresses, and its compound leaves
generally have four opposite leaflets, elliptic-lanceolate in shape, with a swollen petiole.
The leaflets are 16 cm long and 8 cm wide, asymmetric, rounded at the base and with an
acuminate apex. The tiny, white flowers develop in panicles (Mufioz Valencia 1966).
The flowers have 10 stamens and no petals, but the sepals resemble petals (Gentry 1996).
The fruit is a flat indehiscent pod, with one side slightly convex and the opposite side
concave, 10 cm long, 7 cm wide, 3 cm thick, woody, with a single seed (Mahecha Vega
et al. 1984). Seeds are large (mean fresh weight = 48 g, Lopez 2001, to 96 g, Dalling et
al. 1997) and dispersed by water. In addition to the common name "cativo", P. copaifera
has been known as cautivo, kartiva, trementino, floresa, tabasara, amanza muj er, camibar,
murano, and Spanish walnut (Schmieg 1927, Anonymous 1933, Harrar 1941, Hess et al.
1950, Escobar and Vasquez 1987).
Del Valle [cited in Escobar and Vasquez (1987)] noted that cativo leaf-drop is
uniform throughout the year and the species is evergreen. Cooper (1928) found that the
species flowers generally in March and April and fruits in October and November, but
also found a tree with flowers and immature fruits in February. Linares Prieto (1988)
reported that flowering typically begins in June and peaks in August and September.
Fruiting then begins in September and October and peaks in April and May. In
Colombia, seedling recruitment peaks at the beginning of the rainy season from April to
June (Linares and Martinez Higuera 1991). Although cativo produces seeds twice a year,
large seed crops seem to be produced once every 2 years (Pizano SA 1995, Grauel in
Holdridge (1964) reported that throughout its range from eastern Nicaragua into
northwestern South America, cativo is found in the Tropical Moist, Wet, and Rain Forest
lifezones. According to Tosi (1976), the cativo forests of Colombia are found in the
Tropical Moist Forest and Tropical Wet Forest lifezones of the Holdridge classification
system; annual precipitation ranges from 2000 to 8000 mm with average annual
temperature of 24o to 28o C (Gonzalez Perez et al. 1991), while in Darien, Panama, cativo
is found in the Tropical Moist Forest lifezone (Holdridge and Budowski 1956).
Cativo is found in four distinct habitats:
* On the Atlantic coast of Costa Rica, Panama, and Colombia, it is found just inland
from mangrove forests where salt water does not intrude.
* Along the many rivers of southern Central America and northwest Colombia,
cativo is found in alluvial valleys that are flooded periodically, generally in the
* Away from maj or rivers cativo is found in low-lying areas that are inundated for
extended periods, up to the entire 9-month rainy season in Darien, Panama.
* Cativo is also found in upland forests, but never as abundant and dominant as in
flooded habitats. Although Consultores Ambientales LTDA (1995) declared that
cativo cannot germinate or develop in well-drained soil, it is a common component
of the upland, mixed-species forest on Barro Colorado Island in Panama (Condit et
al. 1993b, 1995a, Sheil and Burslem 2003).
Colombian researchers have classified several types of"cativales" based on
landscape position and duration of inundation. Linares Prieto (1988) used landscape
position to classify cativo forests as low, medium, and high alluvial plains. The National
Corporation for Forest Research and Promotion (CONIF, Corporaci6n Nacional de
Investigaci6n y Fomento Forestal, cited in Echavarria A. and Varon P. 1988) classified
different types of cativo forest according to length of inundation as greater than 6 months,
3 to 6 months, and less than 3 months. Gonzalez Perez et al. (1991), referring to the
Institute Geografico Agustin Codazzi cited in G6mez (1990), defined the inundation
periods as permanent, 6 to 8 months, 3 to 6 months, and less than three months. In a
study of early natural cativo regeneration, Martinez Higuera (1989) defined their two
study sites as low alluvial plain frequently flooded and low alluvial plain infrequently
Linares Prieto (1987b) offered various criteria for the classification of cativo-
* three types of forest distinguished by landscape position (plain, terrace, and alluvial
* six types of cativales defined by the combination of life zone and landscape
* number of cativo stems per hectare; median cativo basal area, average distance
between trees, number of trees of other species, the number of other species apart
from cativo, median basal area of other species than cativo, and median total basal
area per hectare.
* phytosociology of Prioria in terms of abundance, frequency, dominance,
importance, and productivity.
* common associates being Cynonsetra spp, Pterocarpus officinalis, Gustavia spp,
Carapa guianensis, Anacardium excelsunt, Escinveilera spp, nuanamo
(Myristicaceae), Castilla elastica, and Lecythis spp., the number of species
increasing with elevation and with better drained soils.
In a study that used life zone and landscape position as classification criteria, Escobar and
Vasquez (1987) proposed nine types of cativo forest but concluded that their
mathematical analysis failed to support the categories.
A defining ecological characteristic of cativo-dominated forests is their
monodominance or low species diversity within stands. Maximum homogeneity is found
in the riverine forests inundated by high Pacific Ocean tides in Darien, Panama. In these
sites, cativo can comprise more than 95% of the stems > 1 cm dbh and the forest contains
less than ten woody species per hectare (Grauel and Kursar 1999, Grauel and Putz 2004).
Along the Marea River in Darien, cativo had a relative abundance of 91% in a 10 x 1000
m transect that included five other tree species and one palm (2 10 cm dbh, Mayo
Melendez 1965). Away from the influence of tides, floristic diversity increases. Linares
Prieto (1988), citing the thesis of Escobar and Vasquez (1987), reported the relative basal
area dominance of cativo as 50 to 92%. In Colombia, Gonzalez Perez et al. (1991) found
that cativo-dominated forests contain approximately 60 tree species, 15 of which
comprise 95% of the individuals and are in the families Fabaceae, Bonabacaceae, and
Sterculiaceae, although the minimum diameter of the study was not stated. In cativo-
dominated forests in the area of Domingodo-Truando in the Colombian Department of
Choco, Consultores Ambientales LTDA (1995) found 86 tree species per hectare (>10
cm dbh). In Darien, Panama, Golley et al. (1975) found 44 species of trees per hectare in
a cativo-dominated forest along the Chucunaque River. Elsewhere in Darien, Holdridge
(1964) found Carapa guianensis to be the only other large tree associated with cativo,
although he noted that Pterocalpus officinalis was restricted to the edges of small streams
in the same forest.
The dominance of cativo in seasonally flooded habitats was suggested as being
attributed to the better competitive ability conferred by ectomycorrhizae (EM) compared
to vesicular-arbuscular mycorrhizae (VAM, Connell and Lowman 1989); but Torti et al.
(1997) showed the existence of VAM in cativo. Lopez and Kursar (1999) demonstrated
that 3 tierra firme species survived inundation as well as "flood-tolerant" cativo, and
suggested that a cycle of inundation and drought-induced water stress may better explain
patterns of tree diversity than inundation alone.
An inventory of animal diversity in a cativo forest in Colombia reported 24 species
of mammals, 16 species of birds (principally in the families Sittacidae, Cracidae, and
Ramphastidae), 6 species of fish, and 7 species of reptiles (Martinez Higuera 1989).
Several of these mammal, bird, and reptile species are becoming increasingly rare due to
hunting pressure, including Taxpirus bairdii, Mazamna amnericana reperticia, Penelope
pwrpura~scens wagler, Crax rubra, Lutra longicandis, and Caiman sclerops (Linares
Prieto 1988). Some other studies on wildlife in cativo forests include those of
MondragC~n et al. (1994), Ospina Torres (1994, 1995a, 1995b, 1996, 1997), and Orozco
Holdridge (1964) described how cativo forests alternate with monospecific forests
dominated by M~ora oleifera along the Tuira and Tucuti (Balsas) Rivers. M~ora-
dominated forests are typically found at slightly lower elevations where the effects of
brackish water from high tides are greater, but cativo trees can be found in M~ora-
dominated forests (Porter 1973). Cativo forests are characterized by a distinct
microtopography where adult trees are surrounded by mounds 20 to 30 cm in height and
5 m in diameter (Duke 1964). Duke (1964) also mentioned that fallen cativo trees were
very common. Although adult cativo trees are shallow-rooted, in a comparison of cativo-
dominated "gallery" forests with tropical moist, premontane, and mangrove forests.
Golley et al. (1969) found belowground biomass to be greatest in mangrove forests, and
similar among the other forest types. Overstory biomass, however, was notably higher in
gallery (cativo) forests, with over 100 Mg/ha dry weight in stems alone. Holdridge
(1964) estimated a leaf area index of 6. 1 in a cativo forest along the Tuira River in
In Colombia, soil fertility of cativo forests varies from very low to moderate
(Martinez Higuera 1989) with fine to medium texture, pH from 5.1 to 6.0, and poor
drainage in general (Linares Prieto 1987c). Soils of riverine cativo forests in Panama are
composed of clay to loamy clay with pH 5.2 to 6.8 (Mariscal et al. 1999, Tapia 1999)
while soils of inland swamps are more clayey and acidic (pH 4.4-6.0, Tapia 1999). The
Colombian Institute of Hydrology and Soil Use, cited in Linares Prieto (1988) and
Martinez Higuera (1989), classified soils of cativo forests as Inceptisols (63%) and
Wood Uses and Properties
The value of cativo wood lies in its historic abundance and accessibility, not
necessarily in its inherent properties. Early descriptions noted cativo for its cylindrical
form and general abundance (Kluge 1926, Cooper 1928) as well as its potential for
supplying raw material for architects and interior decorators (Schmieg 1927), although
Pittier and Mell (1931) considered the wood to be of little or no use.
Later, cativo was included in a series of studies before (Kynoch and Norton 193 8)
and during WWII (Harrar 1941, 1942a, b) that sought to provide technical information on
the physical and mechanical properties of foreign and domestic woods. Further research
beginning in 1947, funded by the U. S. Office of Naval Research, resulted in
recommendations of cativo for plywood, cabinetry, and furniture (Hess et al. 1950).
Similar studies of potential applications of cativo occurred later in Colombia (Hoheisel
and L6pez G. 1972, Universidad Nacional de Colombia 1984).
Several authors have mentioned the abundant resin that bleeds from freshly cut logs
and can make sawing difficult (Cooper 1928, Hess et al. 1950, Barbour 1952, Del Valle
1972). The copious resin of cativo was used by indigenous groups for such diverse uses
as repairing boats and for medicine (Cooper 1928, Duke 1986). By using high
temperatures during kiln-drying, appreciable amounts of resin can be removed from the
lumber with the additional benefit of relieving some of the internal stresses that are
caused by the presence of tension wood (Kukachka 1965).
Table 1-1. Cativo wood properties
Author .1 Shrmnkage %
Harrar (1941) 0.48 9.87
Hess et al. (1950) 0.40 8.9
0.41 sapwood 9.2 sapwood
Barbur (952) 0.50 heartwood 22.9 heartwood
Kukachka (1965) 0.40 8.8
Cativo wood properties have been investigated extensively (Kynoch and Norton
1938, Hernandez Hurtado 1984, Jaramillo Gallego and Velasquez Salazar 1992, Escobar
C. and Rodriguez 1993). It is moderately light in weight (Table 1-1) and although it is
relatively nondurable with respect to both fungal decay and insect attack, cativo has good
dimensional stability and was used as a base for piano keyboards for that reason
Cativo veneer from Panama was marketed in Canada and the United States in 1933
(Anonymous 1933). Large-scale imports of cativo to the United States occurred in the
mid to late 1940s from Costa Rica (Hess and Record 1950). By 1952, almost 9500 m3/y
were exported from Colombia and Costa Rica to the US, the Eigure increasing to over
47,000 m3/yr by 1958 (Kukachka 1965).
While some cativo from the Caribbean side of western Panama was exported, it
eventually supplied 90% of the raw material for the domestic plywood industry and 50%
of sawn-wood production in the country (FAO, Food and Agriculture Organization
1982). The Panamanian Institute of Renewable Natural Resources recommended using
cativo for furniture, packing crates, and cabinetry (INRENARE 1987a).
Diseases and Insects
Cativo was classified as moderately to non-durable in its resistance to the white-rot
fungus Polyporus versicolor (Tramnetes versicolor) and durable to non-durable for the
brown rot fungus Poria monticola (Hess et al. 1950). Ferrer (1999 ) collected 615 fungi
associated with living and dead cativo trees in five different forests, and found 58%
Ascomycetes and 42% Basidiomycetes. Apparently, it is not known how many of these
are pathogens, saprohytes, or mutualists. In a cativo forest along the Sambu River in
Darien, Panama, Ferrer (1999) found that 27% of the Basidiomycetes belong to the genus
Phellinus, one of which is among the most important tree pathogens of temperate forests
Hess et al. (1950), Barbour (1952), and Kukachka (1965) mention the susceptibility
of the boles of recently felled trees to attack by ambrosia beetles. Insects that perforate
recently cut logs belong principally to the families Scolytidae and Platypodidae, and
occasionally Brentidae and Tenebrionidae (Romero 1982). The species most commonly
found on recently cut logs, but not specific to Prioria, are Platy~pus para~lllleh Fabricius
and Xyleborus affinis Eichhoff (Estrada L6pez and G6mez Quiceno 1988). Some
protection from attack is rendered by direct sunlight, immersion of the logs in water, and
the presence of bark; one application of insecticides may prevent attack for 6 to 15 days
(Romero 1982, Estrada L6pez and G6mez Quiceno 1988).
Early research on cativo as a timber source stressed the large sizes and clear boles
of the trees. Cooper (1928) noted an average size of 60 to 90 cm dbh. Barbour (1952)
found the commercial size range of the species to be 60 to 120 cm dbh, with maximum
sizes of 150 to 180 cm. Barbour (1952) emphasized the straight form of the trunks,
without branches for 12 m (and many times branchless up to 30 m in height).
From 195 1 to 1953 Bruce Lamb studied the forests of Darien for the Panama Forest
Products Company to develop log-supply sources and determine available timber
volumes for both upland and lowland forests. Lamb estimated cativo wood volume along
the Chucunaque, Tuira, Balsas, Sambu, Congo, and Cucunati Rivers and around the
Laguna de la Pita (today called Matusagarati). Along the Balsas River, Lamb found pure
stands of cativo for a distance of 20 km and up to a km in width on each side of the river.
He estimated an average volume of 71 m3/ha and a total of 141,600m3 for the watershed
(Lamb 1953). On the Tuira River, Lamb encountered cativo forests 25 km upriver from
the Tuira' s confluence with the Balsas River up to the mouth of the Chucunaque River.
In this area of approximately 4000 ha, volumes averaged 24 m3/ha. Although Lamb did
not examine the forests upriver from the mouth of the Chucunaque, there were reports of
cativo forests up to Boca Cupe, and he estimated a total of 23,600 m3 for the Tuira
watershed. According to Lamb, the highest-quality cativo wood came from the
Chucunaque River watershed, and he estimated a total volume of 47,200 m3 along the 80-
km course of the river. Along the Sambu River, Lamb found cativo 10 km from the
river' s mouth at the confluence of the Jesus River up to the Sambu' s confluence with a
small stream called Morobichi (8 km further upriver). This cativo forest extended up to
1500 m inland from the river, and Lamb estimated an average volume of 35 m3/ha and a
total of 23,600 m3 for the Sambu watershed. Lamb estimated a total of 1,180,000 m3 foT
the entire province (Lamb 1953).
Cativo forests are known to contain large wood volumes per hectare, but
comparison of different estimates is difficult where the minimum diameter is not
specified; in addition, some estimates are for commercial volume and others for total
volume. In a previously unlogged inland swamp in the Balsas River watershed (near a
small stream called Naranzati), 96 m3/ha of commercial (> 60 cm dbh) cativo wood was
measured in 1999 (Grauel, unpublished data). Also in 1999, a 100% inventory was
carried out of a 50-ha plot along the Sambu River in Darien where cativo comprises 95%
of the species diversity. Using a form factor specifically developed for Prioria, a mean
volume of 65 m3/ha was calculated for trees > 60 cm dbh; while across the river in a
series of smaller permanent plots, mean volume totaled only 40 m3/ha for the same forest
type. The difference can probably be attributed to different management histories under
different ownership regimes: the latter being found on open-access public land that is
subj ect to frequent, low-intensity timber harvesting by local loggers; while the 50-ha plot
is located on land belonging to Embera-Wounaan indigenous communities who harvest
cativo much less frequently. When considering a minimum diameter of 40 cm dbh and
the commercial height to the lowest branch, this 50-ha plot contains 190 m3/ha (Grauel,
unpublished data). Recent volume measurements in cativo forests along the Balsas River
ranged from 20 m3/ha (> 60 cm dbh, Grauel, unpublished data) to 25 m3/ha (total volume,
Mariscal et al. 1999).
In Colombia, Linares Prieto (1987b) stated that a cativo forest contained more than
150 m3/ha in commercial wood and a mean of 80 to 100 m3/ha for trees > 52 cm dbh. In
a cativo forest with 60 tree species where cativo comprises 60% of the basal area, Linares
Prieto (1988) measured a total volume of 123 m3/ha and 46 m3/ha in trees > 40 cm dbh.
In a cativo forest with 5 other commercial tree species in Podega, Colombia Escobar
Munera (1981) calculated a mean of 36 m3/ha for all species. In another forest inventory
of trees > 49 cm dbh of 15 tree species, a mean of 7.3 individuals and 27 m3/ha, cativo
comprised 36% of the total volume (Consultores Ambientales LTDA 1995).
Growth and Mortality
As with volume estimates, growth estimates vary and depend on the methodology,
age and size of the trees, management history of the forest, and site-specific biotic,
abiotic, and climatic factors. Comparisons of growth estimates are difficult where
different field methodologies and modeling approaches are used.
Like many tropical trees, cativo produces growth rings, but no dendrochronology
based on crossdating has been performed to show that the rings are produced annually.
Using the pinning technique (Kuroda and Shimaji 1984) where wood is wounded and
subsequent growth is measured with destructive harvesting, however, McKenzie (1972)
concluded that cativo produces annual rings. I strongly suspect that cativo may produce
one or more nings per year.
Cativo diameter growth is probably influenced by many variables. Londofio
Londofio and Gonzalez Perez (1993) reported that crown area, crown position, and
Hegyi's diameter-distance competition index had significant effects on growth of cativo
in less diverse forests but not in the more diverse sites. In an unlogged cativo forest with
50 to 60 tree species, Del Valle (1979) found that maximum diameter increment was
attained by trees approximately 70 cm dbh. Two studies in Panama found maximum
diameter increment in medium-sized trees, from 20 to 50 cm dbh depending on the site.
In an upland forest in Panama, Condit et al. (1993a) measured maximum annual diameter
increments of 2 to 4 cm, while cativo from flooded forests showed maximum annual
growth rates of 1.5 to 2.0 cm, with means of 0.6 to 1.0 cm (Grauel 1999).
Several modeling approaches have been used to estimate lifetime growth
trajectories based on short-term growth rates. Del Valle (1979) used a matrix modeling
approach to produce an estimate of 98 years for a 10 cm dbh tree to reach 60 cm.
Gonzalez Perez (1995) developed a von Bertalanffy growth model for cativo, and
produced an estimate of 90 years for a 14.5 cm dbh tree to reach 60 cm. In a comparative
study of two sites where diameter structure, floristic diversity, spatial distribution, and
growth were contrasted, Gonzalez Perez et al. (1991) found growth to be three times
greater in the more diverse forests. Although the authors admit to small sample sizes,
they estimated 168 years in less diverse forests and 77 years in more diverse forests for a
10 cm dbh tree to reach 60 cm (Gonzalez Perez et al. 1991). There was no difference in
growth between diverse and cativo-dominated forests in another study in Colombia,
where Linares Prieto (1987b) estimated that a tree would reach optimum harvest size (60
cm dbh) in 55 years in less diverse forests and in 60 years in the more diverse forests. In
a different study, the same author estimated that a tree could reach 60 cm dbh in only 38
years, and suggested that this time period could be reduced with "adequate silvicultural
techniques" (Linares Prieto 1988). In a study of growth and yield potential of cativo on
an upland site in Panama, Condit et al. (1995b) estimated a period of 130 years for a stem
to grow from 1 cm dbh to 60 cm, based on mean growth using data from 1982-1985, and
180 years based on data from 1985-1990. Using the same modeling approach and based
on data from 1997-2001, I found two sites with similar inundation regimes to vary
considerably; one requiring 315 years, and the other 179 years for a 1 cm dbh stem to
grow to 60 cm. At another forest farther upriver, only 80 years would be required for a 4
cm dbh tree to reach 60 cm. For an inland swamp, based on growth data from 1997-
2000, 157 years would be required for a 1 cm dbh tree to reach 60 cm, and 186 years to
reach 80 cm.
In a demographic study of cativo in Colombia, Montero (1996) estimated 30,490
seeds/ha were produced during the 6-month period from December to May. Montero
(1996) noted that trees with relatively low seed production tended to produce greater
numbers of established seedlings than trees that had greater seed production and
hypothesized that there was some optimal period for seedfall that resulted in higher
probability of survival and maximum recruitment rates.
Linares and Martinez Higuera (1991) found lower densities of cativo natural
regeneration in frequently inundated forests than in less frequently flooded forests,
although the tendency toward monodominance was greater where flooding was more
frequent. Linares and Martinez Higuera (1991) also found a strong correlation between
mean monthly precipitation and seedling density, and determined that 2% of the large
initial seedfall became established. Lopez (2002) followed a 1997 cohort of cativo
seedlings and found 2.5% survivorship after 3 years.
Dalling et al. (1997) found that less than 10% of cativo seeds were viable 2 months
after seedfall, and that 30% of the viable seeds had suffered damage by insects or
pathogens. Even with up to 60% of the seedmass damaged, however, there was no
reduction in the probability of germination; and seeds with up to eight insect larvae
germinated as often as seeds with no infestation (Dalling et al. 1997). Furthermore,
cativo seeds have the ability to produce an average of 2.1 additional, sequential resprouts
after the initial sprout is damaged or lost (Dalling et al. 1997).
Tamayo Velez (1991) determined that a germinating seed required 48 days to
develop into a 29-cm tall seedling, and declared that height growth of cativo plants
between 30 and 150 cm in height was 50-60 cm/month. Martinez Higuera (1989)
suggested an annual growth rate of 2.4 m for trees in the same height range. In two
strongly monodominant cativo forests in Darien, Panama, density of cativo natural
regeneration (trees < 1 cm dbh) varied widely, from less than 5,000/ha at one site to over
17,000/ha at the other. Mean annual height growth for trees between 30 and 150 cm tall
varied little, from 2-5 cm, with maximum annual growth rates of 15-25 cm (Grauel, in
In a study of artificial regeneration of cativo carried out in Uruba, Colombia,
Linares Prieto (1987a) tested five planting methods and two planting seasons, and
determined that average survival and height were greatest for bare root seedlings planted
during the dry season. After 4 years, these seedlings had reached 2.7 m in height
(Linares Prieto 1987a). Caycedo (1988), cited by Martinez Higuera (1989), measured
annual height growth in seedlings of 70 cm. Cativo seedlings of two ages were planted in
three habitats in Darien, Panama. Seedling mortality after 4 years was highest in the
natural habitat of the forest understory (89%) and lowest in partial sun on the edge of the
forest bordering a treeless marsh (54%, Grauel in review 2004b). Maximum height
growth was observed in the full sun, where seedlings of both ages grew approximately 50
Relatively few tropical tree species have been studied as extensively as cativo.
Because of its accessibility, abundance, form, and wood properties, cativo logging has
provided livelihoods for thousands of rural Latin Americans as well as for forest
industries in North, Central, and South America. Little research has been carried out on
the importance of the ecosystem services that cativo forests provide, however.
Unfortunately, the existence of technical, silvicultural, and ecological knowledge of
cativo and the forests where it is abundant has not resulted in sustainable management.
As is frequently the case in natural resource management, technical knowledge is
insufficient when social and economic forces can influence the strength of forest policies
that determine the quality of management carried out on the ground. Nevertheless,
technical knowledge can form the foundation for research that links ecological dynamics
with the social and economic policies that affect forest management, with the aim of
promoting socioeconomic development that conserves natural ecosystems.
EFFECTS OF LIANAS ON GROWTH AND REGENERATION OF Prioria copalfera
INT DARIEN, PANAMA
The abundance and ecological roles of lianas in tropical forests have long attracted
the attention of tropical silviculturists (Fox 1968, Appanah and Putz 1984, Chaplin 1985,
Putz 1991, Vidal et al. 1997, Carse et al. 2000). Because lianas are a major component of
woody plant diversity and provide important food sources for wildlife, they play critical
roles in maintaining biological diversity (Nabe-Nielsen 2001, Burnham 2002, Schnitzer
and Bongers 2002). Unfortunately, where sustainable forest management is the primary
tool for forest conservation and the primary obj ective is timber production, lianas can be
a maj or impediment. Given that the likelihood of forest conversion to more profitable
land uses than forestry is enhanced if prospects for subsequent timber harvests are not
economically competitive, liana proliferation can contribute indirectly to forest loss.
The large trees that provide the timber value of a forest are more likely than smaller
trees to be infested with lianas (Putz 1984, Putz and Chai 1987, Nabe-Nielsen 2001,
Perez-Salicrup et al. 2001), and lianas can have various silvicultural implications for
forest management. During harvesting operations for example, felling of liana-laden
trees can induce excessive stand damage, because their crowns are likely to be connected
to their neighbors (Putz 1984). Avoidance of this accessory damage has frequently, but
not always (Parren and Bongers 2001), been accomplished through pre-felling liana cutting
(Fox 1968, Appanah and Putz 1984, Johns et al. 1996). An additional benefit of pre-
felling liana cutting is the post-harvest reduction in liana proliferation in logging gaps
(Alvira et al. 2004, Gerwing and Vidal 2002). Reducing post-logging liana infestations is
desirable, because lianas can seriously impede succession in gaps (Schnitzer et al. 2000)
and diminish opportunities for rapid recruitment and growth of desirable timber species.
In addition to physically impeding establishment of seedlings and saplings of tree species
in logging gaps, lianas can reduce host tree fecundity (Stevens 1987), lowering the
reproductive output of valuable timber species in forests where natural regeneration is the
only cost-effective silvicultural option for stand perpetuation. Heavy liana infestations
can also substantially reduce diameter growth of adult trees (Whigham 1984, Gerwing
2001, Clark and Clark 1990), which lowers the net present value of future timber yields
by prolonging cutting cycles.
Prioria copaifera (hereafter "cativo," Fabaceae), a canopy tree found in freshwater
wetland forests from Nicaragua to Colombia, has been exploited for timber for decades
(Barbour 1952), with little apparent concern for long-term management. Today,
commercial stands are found principally in eastern Panama and northwest Colombia.
Repeated logging of monodominant cativo stands during 40 years of exploitation testifies
to the regenerative capacity of the species. Nevertheless, large areas of cativo forest have
been converted to agricultural production or to mixed-species secondary forest and liana
tangles as a result of overharvesting. Of the original 363,000 ha of cativo in Colombia
for example, less than 90,000 ha remain (Linares Prieto 1987b). Similarly, extensive
stands of cativo were once found in western as well as eastern Panama, but today
commercial stands are found only in Darien Province. Of 30,000 ha of cativo-dominated
forest in Darien in 1987 (INRENARE 1987a), an estimated 15,000 ha remained in 1999
(ANAM 1999a). Increasingly, Panamanian foresters as well as local Darien community
members desire to promote sustainable logging of the remaining cativo forests.
The stands of almost pure cativo that are found as bands along the principal rivers
of Darien vary between 100 m and 1 km in width. Behind the forest, treeless wetlands
composed of the palms Elaeis oleifera and Oenocarpus nzapora and various lianas
i ncludi ng Dalbergia brownei, Conabre tunt sa~n buensis, Elachyptera floribunda,
Tetrapteris nzacrocalpa, AllllllllllIllllllla~nzad cathartica, Phrygan2ocydia corynabosa, Cydista
diversifolia, Snzilax spinose, Banisteriopsis spp., and Heteropteris spp. often dominate
the landscape. In the absence of silvicultural interventions other than logging, high-
statured riverine forests are likely to be converted into palm- and liana-dominated
Present day stand structure of many riverine cativo forests in Darien is a result of
traditional logging methods that do not employ heavy machinery (Grauel and Pineda M.
2001). Instead, logs are levered or rolled by hand towards the river on roads constructed
from 15 to 30 cm dbh (diameter at breast height, 1.3 m) cativo trunks cut and laid end to
end to form two parallel rails. In many riverine cativo forests, the combination of
removing all harvestable-size trees as well as many subcanopy individuals for rail
building has left a very discontinuous canopy and large multiple-tree gaps, which are
habitats favorable for liana proliferation.
The leguminous liana Dalbergia brownei proliferates abundantly in disturbed
cativo forests. A principal component of the treeless wetlands found behind the natural
river levees where cativo dominates, this liana uses cativo forest edges to climb into the
forest canopy. Although this species does not establish in the deep shade of the cativo
forest understory, large stems (up to 20 cm diameter) are commonly found hanging from
the 30-40 m high canopy in many of the cativo forests of the lower Balsas, Sambu, and
Tuira Rivers (Grauel and Pineda M. 2001). Areas with high liana densities seem to have
developed in large logging gaps created 20-30 years ago. Many mature cativo trees in
heavily infested areas are visibly deformed, apparently from having developed while
carrying large liana loads or from having been damaged during logging. In other areas
that have been continually and recently subj ected to small scale harvesting, D. brownei is
proliferating on the ground in large canopy gaps and appears to delay cativo regeneration.
In the present study I measured, by observation and experimental liana removal, the
effect of lianas on cativo adult stem growth as well as on seedling height growth,
recruitment, and mortality.
The study was conducted in a riverine cativo forest along the Balsas River in
eastern Panama (8o 07' N, 77o 52' W). Mean annual precipitation at Camoganti, the
nearest town (approximately 8 km from the study site), is 2457 mm (based on
Government of Panama published reports for 1978-1982, 1984, 1986, and 1988-1994)
while rainfall measured at the study site in 1998 and 1999 totaled 2970 mm and 2758
mm, respectively. The forest is inundated periodically with rainwater during the 9-month
wet season from April to December. In addition, it is flooded twice per day for about
five days during the monthly spring tides known locally as the 'aguaj e.' The freshwater
backup caused by the Pacific spring tides affects the riverine forests as far as Camoganti,
73 km from the mouth of the Tuira River at the Gulf of San Miguel. Although at the
study site the tidal flooding is mostly the freshwater backup, soil samples show a slight
brackishness (electrical conductivity 5.0 mmhos/cm) and mangrove forests are found
only 7 km downriver from the study site. Soils at the study site are heavy clays classified
in the suborders fluvent and aquept, are acidic to slightly acidic, and poorly drained
The study site is on private land owned by a logger and is next to an operating
sawmill. The owner, who has been logging cativo in Darien since 1960, is currently
logging further upriver and has protected the forest where the study took place because he
values it for hunting and aesthetics, although he told us that he had harvested a few
scattered trees about ten years prior to the study. This cativo forest is composed of about
95% Prioria copaifera of all size classes (Grauel and Kursar 1999). Other tree species
include Pterocarpus officinalis, M~ora oleifera, and Caurapa guianensis. Results from a 1
ha permanent plot show 10 cativo trees per hectare > 60 cm dbh, the legal cutting limit,
but the maj ority of these were left due to bad form or hollowness. Regeneration of cativo
of all sizes is abundant.
In September 1997 six 25 x 25 m plots were installed in a line at 50-75 m intervals
in areas with intact canopies but with relatively high densities of lianas compared to the
forest overall. Each plot was subdivided into twenty-five 25-m2 Subplots to facilitate
stem mapping. Inside the plots I measured all trees > 4 cm dbh as well as the diameters
of all ascending liana stems > 1 cm at breast height. I did not attempt to differentiate
genetically distinct lianas; every stem encountered at 1.3 m above the ground was
measured. To increase the sample size for the growth analyses of cativo, additional trees
were measured up to 5 m outside of each plot, but no lianas were measured outside the
plots. Every plant was tagged and mapped and subsequent censuses were carried out in
1998, 1999, and 2001. For the growth analysis, diameter classes for trees were selected
based on relative canopy position; 15 cm dbh was used as the cutoff between canopy and
understory individuals. Due to the low canopy of the forest where lianas are abundant,
even trees 15-30 cm dbh may receive substantial direct illumination, while trees < 15 cm
dbh are generally in the understory.
During the initial measurements, each tree was classified as severely or lightly
infested by lianas. Severely infested trees had at least five individual liana stems hanging
from the crown and some stems or branches apparently deformed by lianas. Lightly
infested trees had fewer than five liana stems hanging from the crown and no visible
deformations. For the growth analyses, growth rates of liana-free trees were included in
the lightly infested category. All lianas were cut with a machete inside and up to 10 m
outside of three randomly chosen plots.
In 10 randomly chosen 25m2 Subplots in each plot, all natural regeneration of
cativo from seedlings to small trees 1 cm in diameter were counted, tagged, and measured
(height) before the vine cutting treatment and two years later. Where necessary to reduce
heteroskedasticity, seedling frequency data were natural log-transformed. Mean relative
height growth for seedlings in treated and control plots was compared with a two-sample
t-test and the difference of mean absolute height growth was tested by ANOVA using
initial height as a covariate. Mortality of these seedlings and small trees in treatment and
control plots was also compared. Two years after the initial census, the same subplots
were surveyed for new cativo regeneration. Treatment differences in mean density of
seedlings recruited per plot was compared with t tests.
For several reasons, including the observation that increases in cross sectional area
of lianas are associated with much larger increases in leaf area than in trees (Putz 1983), it
is desirable to estimate diameter growth rates of lianas. In 2001, four years after the
initial measurements, 56 Dalbergia brownei lianas in the control plots were again
measured to estimate stem diameter growth. Individuals < 6 cm dbh were measured with
dial calipers; a mean diameter was calculated from measurements of the long and short
axes. Lianas > 6 cm dbh were measured with a diameter tape. Wood density was
estimated using ten bark-free stem samples, to allow comparisons with other studies.
While growth rates of trees are often negatively correlated with wood density, this pattern
may not hold for lianas that do not produce structural wood for support.
Canopy openness above 2 m was measured immediately before and two months
after liana cutting in all plots with a vertical densitometer (Stumpf 1993). Both
measurements were made during the rainy season. This instrument proj ects a point
vertically upward that encounters either canopy or open sky at each evenly spaced sample
point along a linear transect. Canopy openness is estimated as the proportion of points of
open sky along three transects in each plot.
To compare rates of cativo growth and regeneration in heavily vine-infested areas
with forest with low liana infestation, data from the six plots of the present study were
compared with data from plots selected at random for a demographic study of cativo in
the same forest. The demographic study was based on five 20 x 20 m and five 40 x 40 m
plots established in March 1997. All trees > 10 cm dbh were tagged, mapped, and
measured (dbh), while trees > 1 cm dbh were measured in all five 20 x 20 m plots and in
five randomly chosen 20 x 20 m subplots in each of the 40 x 40 m plots. All trees were
measured annually from 1997 to 2001. In addition, all trees < 1 cm dbh in eight
randomly chosen 5 x 5 m subplots of each plot were tagged and mapped and were
measured (height only) in November 1997. This population of seedlings and saplings
was censused approximately every two months for two years whereas height was
Two months after cutting lianas, significant but modest increases in canopy
openness were observed in the treated plots. There was no difference in the before and
after canopy coverage in the three control plots, while the three treated plots showed a
mean increase of 7% (p < 0.01) in canopy openness.
Mean annual diameter growth of Dalbergia brownei was 1.3 mm yr- (n = 56, sd =
1.4, range = -0.8 to 5.5 mm). Mean wood density of D. brownei (dry weight/fresh
volume) based on ten samples was 0.38 g cm-3 (Sd = 0.047).
Based on the mean number of stems > 4 cm from the six 25 x 25 m plots, cativo
dominated the forest with 1320 stems ha-l (sd = 212), virtually identical to nearby areas
of riverine forest with lower liana densities (133 8 stems ha l, Grauel and Kursar 1999).
Pterocarpus officinalis, the only other abundant tree species, was represented by 51 stems
ha-l > 4 cm dbh. The 35.1 m2 ha-l of cativo basal area represents 96.6% of the total basal
area of trees > 4 cm dbh.
For all cativo trees > 4 cm dbh, 71% had lianas hanging from the crowns, while
93% of mid- and upperstory trees (> 15 cm dbh) had lianas. There were 1757 ascending
liana stems ha-l > 1 cm dbh (sd = 270), with a mean liana basal area of 3.40 m2 ha-l (sd =
0.8). Of the two liana species found, Dalbergia brownei comprised 96.9% of the basal
area. The only other liana encountered, Elachyptera floribunda (Hippocrateaceae), was
mainly represented by small stems, with 75% of the ascending stems < 3 cm in diameter,
while over 80% of the D. brownei stems were > 3 cm in diameter (Figure 2-1).
Prior to liana cutting, in the six heavily liana-infested plots the mean density of
cativo seedlings and saplings (< 1 cm dbh) was 707 ha-l (sd = 1154). In contrast, in the
ten randomly located plots for the demographic study of cativo at the same site, the mean
density of cativo seedlings and saplings was 6350 ha-l (sd = 12882, Figure 2-2).
Although this was almost an order of magnitude difference and is plainly discernible in
the forest, variability was large due to the clumped distributions of seedlings, but the
difference was significant (t = 3.12, df = 14, p = 0.008).
For cativo regeneration present at the beginning of the study, relative and absolute
height growth over two years did not differ between liana-cut and control plots.
Although mean initial height for seedlings and saplings happened to be significantly
greater in the three control than in the three treatment plots, there was no difference in
initial height of only those trees that survived to produce growth records. Cativo seedling
and sapling mortality was nearly double in treated than in control plots (63% vs. 36%,
Pearson X = 6.2, p = 0.01).
Cativo seedling recruitment during two years after liana cutting was more than
three times greater in the treated than in the control subplots but, due to large variability,
was not statistically significant (t = 1.30, df = 4, p = 0.26, Figure 2-3). On a per ha basis,
over 7700 cativo seedlings recruited during two years after lianas were cut compared to
just over 2200 seedlings for the control plots.
Mean annual diameter growth of cativo trees during 1997-2001 was about twice as
rapid in the liana-cut compared to the control plots (Figure 2-4). For trees > 15 cm dbh
the difference was significant (t = 3.41, df = 4, p = 0.03), while for trees between 4 and
15 cm the difference was not statistically significant (t = 2.61, df = 4, p = 0.06).
Surprisingly, severity of liana infestation had little apparent effect (no significant
differences found) on cativo diameter growth for either control or treated plots (Figure 2-
5). The largest difference was for canopy trees in the control plots, where severely
infested trees grew slightly slower than lightly infested trees.
Despite their abundance, liana cutting had only a slight (7%) but statistically
significant (p < 0.01) effect on canopy openness of the cativo forest, because most of the
liana foliage is displayed on the tops of tree crowns. Liana-infested cativo forests look
"feathery" at the top of the canopy, due to the abundant emergent branches of small-
leaved D. brownei searching for higher trellises. Two years after liana cutting, canopy
openness was similar for all plots, perhaps because the cativo canopies increased leaf
production after liana cutting. In contrast, Gerwing (2001) found that increases in canopy
light transmittance persisted for two years following vine cutting, and Perez-Salicrup
(2001) measured no change in canopy openness four months after vine cutting in a
Bolivian lowland forest but an increase in openness two years later. Both the Bolivian
and Brazilian studies took place in much drier forests and probably on less fertile soils
than the present study; perhaps the cativo trees were better able to take advantage of the
removal of lianas and produce foliage rapidly.
The low diameter growth rate of Dalbergia brownei is similar to the growth rate
found by Putz (1990, 1.4 mm yr ) for fifteen liana species from a tropical moist forest in
Panama. In a lowland wet forest in Ecuador the most abundant liana, Machaerium
cuspidatum, had an average annual growth rate of 1.4 mm yrl for stems between 30 and
50 mm (Nabe-Nielsen 2002). In contrast to these low diameter growth rates, a single shoot
of D. brownei was observed to grow 1.24 m in length in 71 days.
Increased mortality of cativo seedlings in liana-cut plots compared to control plots
appeared to be due to numerous large liana stems falling from the canopy, most within
the first year following cutting. In addition, floodwaters commonly move coarse woody
material around on the forest floor, which frequently results in the bending and breakage
of seedlings. This additional impact may explain why tree seedling mortality increased
significantly following liana-cutting here but not in a tropical tierra firme forest in
Bolivia with higher liana densities (Perez-Salicrup 2001).
Enhanced seedling recruitment in liana-cut plots (Figure 2-3) more than
compensated for increased mortality of the initially scarce regeneration in these heavily
liana-infested areas. Large cativo seed crops were produced in 1997 and 1999, and the
germinated seedlings from the May-June 1999 seedfall were captured in the plot census
in November 1999 before dry season (January-April) mortality occurred. Three possible
mechanisms for this increased seedling recruitment following liana removal include: 1)
reduced seed movement out of the liana-cut plots; 2) enhanced seedling survival; and, 3)
increased seed production.
Reduced seed movement could result because cativo seeds are large (mean fresh wt
= 48 g, Lopez 2001) and mainly dispersed by water. Seeds that fall and that are not
partially buried in the heavy wet soil may be transported by the monthly spring tides or
the periodic flooding caused by wet season rains. Fallen liana stems could have acted as
small dams inhibiting cativo seed movement during inundation.
If seed production and retention as well as capture of water dispersed seeds were
equal in all plots, high early mortality rates in the control plots would explain the
difference in seedling densities between control and treatment plots. But most cativo
seedling mortality occurs during the short dry season from January to April (Lopez 2002),
and the plots were censused after seedfall but before the onset of the dry season of the
second year after liana cutting.
Cativo trees newly liberated from lianas may have produced more seeds. A
possible mechanism that could help explain an increased production of seeds by cativo is
that lianas interfere with cativo flower or seed production, either physically or through
competition for light. D. brownei produces long recurved spines that wrap around small
diameter objects that they encounter, such as flower-bearing branches. A drain on host
resources caused by the constriction of vascular elements in branches or twigs was
proposed by Stevens (1987) to explain the negative effect of lianas on the fecundity of
Bursera simaruba trees in Costa Rica, and a similar mechanism may be at work in the
Another possible mechanism for the proposed increased seed production in the
treated plots is the removal of belowground competition after the death of the lianas,
resulting in increased nutrient availability. Increased water availability may also have
been a factor, since seasonally flooded cativo forests can experience severe short-term
annual droughts. Putz (1991) suggests that because lianas do not need to produce large
diameter structural roots, root systems of lianas may be more efficient in water and
nutrient uptake. Lianas were also experimentally shown to be effective belowground
competitors in a study with north temperate vine species (Dillenburg et al. 1993).
Although liana cutting resulted in a notable positive increase in Prioria copalfera
stem diameter growth, lianas may not be solely responsible for low forest-wide growth
rates, as proposed by Gerwing (2001) for a seasonal Amazonian forest in Brasil. In the
riverine cativo swamps in Panama discussed in this study, mean annual growth of trees in
the control plots was not significantly different than similar sized trees in the nearby
permanent plots where lianas were generally less abundant (data not shown).
Mean annual growth of cativo trees of all sizes in both treatment and control plots
was greatest during the second year following liana cutting and then declined (Table 2-
1). Although the liana leaves fell during the first two months after cutting and the
hanging stems fell within the first year following treatment, this pattern of increased
growth followed by decline is probably not due to a "fertilizer effect" from the fallen
liana material; forest-wide growth rates for cativo, based on permanent plots at this site
and three others in different watersheds in Darien Province, showed the same pattern,
suggesting a correlation with climate.
The observation that the growth rates of liana-infested control plot trees > 15 cm
dbh did not differ from growth rates of trees of the same dbh in adj acent permanent plots
with few lianas (data not shown) could be attributed to the lower overall canopy height in
the area of the liana-cutting plots. With fewer large trees in the liana-abundant areas of
the forest, trees that otherwise would be subcanopy individuals may receive more light
than a similar sized tree in adj acent non-liana forest. This interpretation suggests that
these liana-abundant areas are old canopy gaps in the process of recovery, similar to the
"vine-dominated disclimax" of Whigham (1984) or the "stalled gap" of Schnitzer et al.
The profound negative effect of lianas on cativo growth and reproduction probably
results from a combination of aboveground and belowground influences. Lianas were
estimated to occupy 3 1% of the forest canopy surface area during the wet season in a
seasonally dry tierra firme forest in central Panama (Avalos and Mulkey 1999) and liana
leaves might significantly reduce light availability for cativo leaves. Belowground
competition for nutrients and water in the dry season could also be a factor. Given that
rooting depth in cativo forests is limited by the high water table (Lopez 2002),
belowground competition for water may be severe during the dry season because root
systems die back during wet season flooding (Lopez 2002). In a semi-deciduous lowland
forest in Bolivia, Perez-Salicrup and Barker (2000) found significantly less negative
water potentials in Senna multifuga trees where lianas were cut as well as increased tree
diameter growth. In contrast, in the same forest after liana cutting, Barker and Perez-
Salicrup (2000) found no difference in water status of mahogany trees with and without
lianas and concluded that lianas and trees had access to different sources of water due to
different rooting depths.
Two issues pertinent to considerations of liana cutting as a silvicultural tool are cost
of implementation and biodiversity impacts. Based on the experimental plots, it would
require 16 person-hours to treat one hectare of forest. Although the experiment took
place on private land, the maj ority of these degraded forests are on state land managed by
the Panamanian Environmental Ministry (ANAM). Currently, ANAM is working with
several communities in Darien to develop a partnership whereby forest areas are
identified for potential timber production and legal tenure is transferred to local
community groups. The government intends to train community members in mapping,
inventory, and other management activities; liana cutting could be one of the
recommended silvicultural treatments.
Conservation of biodiversity is a critical issue in our study area, especially because
of the presence of Darien National Park and the current improvement of the Pan-
American highway that will doubtlessly increase colonization rates and forest clearing.
Liana-cutting in seasonally flooded cativo forests is not expected to have severe effects
on species diversity for several reasons. Tree species diversity is extremely low in
tidally-flooded cativo forests (Grauel and Kursar 1999) and only one species of liana
seems to have proliferated excessively as a result of several decades of uncontrolled
timber exploitation. Because this species, Dalbergia brownei, is also a common
component of nearby treeless wetlands, attempts at controlling its proliferation in cativo
forests suited for timber production probably will never eliminate it. Promoting
sustainable management of cativo forests for timber production could actually serve as a
biodiversity conservation tool. By providing local communities with viable economic
activities in the areas surrounding Darien National Park, pressure to exploit the natural
resources of the park could be reduced.
Table 2-1. Mean (+1 SE) annual diameter growth (mm) of Prioria copalfera one, two,
and two to four years after liana cutting.
1997-1998 1998-1999 1999-2001
N Increment (mm)
171 1.5 (0.2)
204 1.6 (0.2)
166 2.2 (0.2)
180 3.3 (0.3)
N Increment (mm)
168 2.2 (0.3)
196 3.1 (0.3)
164 4.1 (0.4)
172 6.6 (0.4)
N Increment (mm)
152 1.6 (0.2)
181 1.8 (0.2)
144 2.7 (0.3)
154 4.5 (0.3)
0 Elachyptera floribunda
5001 1 Dalberqia browlnei
O Elachyptera floribunda
0.51 g Dalberqia browvlei
1-2 2-3 3-4 4-5 5-6 6-7 7-8 8-9 9-10 10-11 11-12 >12
Diameter Class (cm)
Figure 2-1. Diameter distributions of ascending lianas in six 25 x 25 m plots in heavily
infested riverine Prioria copaifera forest degraded by repeated entry logging.
a) Number of stems per hectare. b) Basal area in square meters per hectare.
Figure 2-2. Mean (-E 1 SE) density of Prioria copalfera regeneration (< 1 cm dbh) in
areas of high (N = 10) and low (N = 6) liana densities.
301 I i, 12000
25- 10000 (D
~f20- __, 8000
S15- 6000 v,
v, 5- 2000
Control Vines Cut
Figure 2-3. Mean (-E 1 SE) Prioria copaifera seedling recruitment censused two years
after liana cutting in three control and three treatment plots.
I I I Treatment
<( 1 -
01 m Lianas Cut
Figure 2-4. Mean (-E 1 SE) annual Prioria copaifera diameter growth based on five
annual censuses of all trees > 4 cm dbh in three control plots and three plots
in which all lianas were cut at the beginning of the study.
2 6 Control Lianas cut
0 m Severe
4-15 >15 4 -15 >15
DBH (cm) DBH (cm)
Figure 2-5. Mean (-E 1 SE) annual Prioria copaifera diameter growth of cativo based on
five annual censuses according to liana infestation level in control and
GROWTH AND SURVIVAL OF Prioria copaifera SEEDLINGS PLANTED ALONG
A HABITAT GRADIENT IN A PANAMANIAN SWAMP
Until relatively recently, native tropical species have been little utilized for
reforestation programs (Evans 1992). For plantation forestry in the past, a general lack of
information led to a reliance on few familiar exotic species with high growth rates.
Concern over the loss of biodiversity as well as recognition of other production systems
besides monospecific plantations has increased attention on native species for
reforestation (e.g., Butterfield 1995, Haggar et al. 1998). Furthermore, much effort has
now been expended in acquiring knowledge about growth and mortality rates, as well as
propagation techniques, of native tree species in the tropics, often with the specific aim of
identifying promising candidates for reforestation (Condit 1995, Foroughbakhch et al.
2001, Wightman et al. 2001, Moulaert et al. 2002).
Reforestation with native tropical species may be carried out for timber production
(Keenan et al. 1999), site restoration (Parrotta and Knowles 1999, Engel and Parrotta
2001, Montagnini 2001), fuelwood production (Kataki and Konwer 2002), carbon
sequestration (Silver et al. 2000), biodiversity conservation (Blakesley et al. 2002a,
Blakesley et al. 2002b), or other reasons. For a given end use, species choice should not
be based soley on site characteristics or growth rates because the purposes for which trees
are planted may vary not only with the type of tree but also with the type of 'user'
(Raintree 1991). Certain landowners may be more amenable to reforestation with native
species where opportunity costs for labor are low or motivation consists of a wider range
of benefits than only financial profitability (Putz 2000).
In this study, a valuable native timber tree, Prioria copaifera (hereafter 'cativo'),
was planted to evaluate growth and mortality rates in different habitats in swamp forests
in eastern Panama. I hypothesized that both mortality and growth would be greatest in
the high-light environment and lowest in the forest understory where cativo natural
regeneration was abundant.
Cativo is a large Caesalpinoid timber tree that has been harvested commercially for
decades (Anonymous 1933, Hess and Record 1950, Kukachka 1965). Originally
distributed from Nicaragua to Colombia, logging and conversion to banana plantations
have severely reduced the area of cativo forests in Costa Rica and western Panama
(Veiman 1982, Jimenez Madrigal 1995). Today, commercial stands are restricted to
eastern Panama and northwest Colombia. Cativo is known principally from single
species or monodominant stands in swamp and riparian habitats, but is also found, less
frequently, in upland soils where it sometimes dominates (Condit et al. 1993b).
In Darien, Panama, the area of cativo-dominated forests has been reduced to less
than 15,000 ha from an original coverage of 60,000 ha (Grauel and Pineda M. 2001).
Many riverine cativo forests are adj acent to extensive marshes composed of palms, dense
liana tangles, and occasional "vine towers" where trees still stand from the remnant
forest. In the early 1950s Lamb (1953) described a belt of commercial cativo forests
along the Balsas River that averaged 1km in depth on each bank of the river, but in 2000
the belt was only 100m deep in some areas. Decades of logging has left many of these
cativo forests badly liana-infested and otherwise severely degraded (Grauel and Putz
2004). Hence, an important obj ective of the study was to determine the feasibility of
restoring treeless marshes to cativo forests.
The study was conducted in a forest on private land along the banks of a tributary
of the Tuira River, the Balsas River, in Darien Province in eastern Panama (8o 07' N, 77o
52' W), 48 km from the mouth of the Tuira River at the Gulf of San Miguel. The
landowner has been logging cativo in Darien for forty years, and although he has cut only
a few scattered trees at the study site within the last ten years, it is likely that this forest
was subjected to an initial selective cutting around thirty years ago. In an early survey of
cativo-dominated forests Barbour (1952) noted that commercial-sized trees ranged from
60 to 120 cm in dbh (diameter at breast height, 1.3 m), with occasional specimens of 150
to 180 cm dbh. Today, few large trees are found in these easily accessible riverine forests
(Grauel and Kursar 1999).
The arboreal component of the study site is composed of 95% cativo (basal area
and stems ha l) of all sizes (Grauel and Kursar 1999). Other tree species that occur in the
stands include Pterocarpus officinalis, M~ora oleifera, and Carapa guianensis. Marshes
are found adj acent to the cativo forest and are composed of scattered palms (Elaeis
oleifera), liana tangles, and occasional "vine towers", suggesting that the tall forest has
been displaced. The lianas of the marsh, principally Dalbergia brownei, climb the tree
canopy at the well-defined edge of the forest where there is abundant light, but seldom
colonize the understory of the high-statured forest where little light penetrates.
Mean annual precipitation at Camoganti, the nearest town (approximately 8 km
from the study site), is 2457 mm (based on Government of Panama published reports for
1978-1982, 1984, 1986, and 1988-1994) while rainfall measured at the study site in 1998
and 1999 totaled 2970 mm and 2758 mm, respectively. The study site is subjected to
periodic flooding from rain events during the 9-month wet season (April-December) and
is also flooded when river water is backed up by high monthly tides. The slightly
brackish soils at the study site (electrical conductivity 5.0 mmhos/cm) are heavy clays
classified in the suborders Fluvents and Aquepts, are acidic to slightly acidic, and are
poorly drained (Tapia 1999).
In August and September 1997, cativo wildings of two ages were dug up in the
forest and planted in three different habitats with distinct light environments. First year
seedlings, seeds that matured in May-June of 1997, were identified by having an attached
seed and little woody tissue. Individuals referred to as "older seedlings" had no attached
seed and had woody stems. These latter seedlings may have been 3 years or older given
that cativo seedlings can survive for a number of years with little or no growth (Lopez
2002). These latter seedlings were likely 2-4 years old.
I tested the effects of 3 levels of canopy cover on two ages of seedlings in four sites
(= blocks) separated by 50-200 m along forest-marsh interfaces. In each of four blocks,
side-by-side pairs of 7 x 7 m plots were installed in the shade of the forest, on the edge
between the forest and treeless marsh, and in the marsh in full sunlight. Site preparation
in the marsh involved extensive liana cutting to establish plots and facilitate access but
only moderate liana cutting in the edge habitat where liana density was not as abundant
due to the partial shade of the adj acent forest canopy. Seedlings were carefully dug up
and bare root planted at 1 x 1 m spacing. For each plot pair in a given light environment,
seedling age was randomly assigned and 49 seedlings of a given age were planted in each
plot. Therefore, for each habitat seedling age combination, 196 seedlings were planted
for a total of 1176 seedlings. In the first plot that was planted, the seedlings (older,
shade) were transplanted with entire blocks of soil to protect the roots. This practice
proved to be excessively laborious and was abandoned. All subsequent results exclude
these 49 seedlings that had high survivorship but little height growth at the end of the
Canopy openness is defined as the proportion of the sky hemisphere that is not
obscured by vegetation when viewed from a single point (Jennings et al. 1999). At the
center of each plot, canopy openness at 1.3 m above the ground was measured with a
spherical densiometer (Lemmon 1957) by averaging four readings taken in the cardinal
After planting in August and September 1997, seedlings were tagged, each stem
was marked at 20 cm above the ground, and two perpendicular diameter measurements
were taken with Vernier calipers and averaged. Total height of each seedling was also
measured. Both height and diameter were recorded at approximately six month intervals
for two years; Einal measurements were made after an additional 20 months had elapsed
(July 2001). Repeated measures analysis of covariance was used with either initial height
or diameter as the covariate to test the significance of seedling age (between-subj ects),
habitat (within-subjects), and their interaction. In addition, final mean and maximum
height and diameter were each compared between seedling ages within habitats. Height
and diameter data were natural log-transformed to comply with the ANOVA assumption
of normality. For tests on seedling sizes after the initial measurements, Geisser-
Greenhouse adjusted p-values were used because the assumption of homogeneity of
variances of treatment-differences was not met (Maxwell and Delaney 1999).
Seedling survival was censused Hyve times in the first two years (1997-1999) with a
Einal tally in July 2001. Survival among the three habitats was compared for each
seedling age with one-way ANOVA, then Bonferroni post-hoc tests were used to
determine differences among means. All data were arcsine square root-transformed to
reduce the unequal variances found in a few plots.
Annual rates of absolute and relative growth in height and diameter were calculated
for one year, two years, and four years after planting. Growth was first examined for
each seedling age by comparing performance among the three habitats with ANOVA and
Bonferroni comparisons. Younger and older seedling growth was then compared with t-
tests for each habitat. For growth over the entire four years of the study, as well as Einal
seedling size, comparisons could only be made between the edge and sun habitats
because of high seedling mortality in the shade.
Canopy openness among the three habitats ranged from 10-85% (Table 3-1). The
high variability for the readings in the edge environment resulted from one reading
capturing the forest canopy, one capturing the open sky of the marsh, and the other two
including both canopy and sky.
Younger and older seedlings differed significantly in both height and diameter at
the time of planting (Table 3-2). Older seedlings were twice as large in diameter and
about 25% taller than first year seedlings.
The repeated measures ANCOVA revealed that only habitat, not seedling age, had
significant effects on height and diameter (both G-G adjusted p < 0.001) during the six
measurements, and there was no age-habitat interaction (Figure 3-1). Maximum heights
for both seedling ages were found in the full sun, where mean maximum height among
the four blocks for younger seedlings was 382 cm and for older seedlings 350 cm.
Within the edge and sun habitats, however, there was no significant difference in
maximum or mean height attained between seedling ages at the end of the four year
study. Because of high mortality of first-year seedlings in the shade, no comparisons
could be made regarding final size attained in that habitat.
The planted seedlings were also largest in diameter in the full sun habitat. The only
significant difference in all comparisons of diameter and DBH, however, was found in
mean stem diameter, where older seedlings were slightly larger than younger seedlings in
the sun (38.6 mm vs 32.2 mm, t = 2.5, df = 6, p = 0.047).
For the edge and sun habitats, 88% 99% of total four year mortality occurred in
the first seven months after planting (Figure 3-2). Seedlings initially survived better in
the shade but at the end of the study mortality was highest for both seedling ages in the
shaded understory. Results from the ANOVAs showed that survival was consistently
highest for both seedling ages in the edge habitat, but the Bonferroni tests revealed no
significant differences in survival at the end of the study between the edge and full sun
habitats for either seedling age. Furthermore, for the older seedlings there was no
significant difference between survival in the shade and full sun habitats. Variation
among plots was fairly modest (Figure 3-2).
Maximum mean annual height and diameter growth for both seedling ages occurred
in the sun habitat during 1999-2001, when seedlings grew about 1 cm in diameter (at 20
cm stem height) and from 80-90 cm in height (Table 3-3). For each seedling age,
absolute and relative height and diameter growth were significantly different in most
habitat comparisons. The few exceptions occurred when growth in the edge habitat was
similar to growth in either the full sun or the shade habitat for a few growth intervals. In
comparisons of seedling ages within each habitat, annual relative height growth was
never significantly different for any growth period. During the 12 months after planting,
younger seedlings grew relatively faster in diameter than older seedlings in the shade and
edge habitats. In the second year after planting, younger seedlings' relative diameter
growth was faster only in the shade. Annual relative diameter growth rates for the period
1997-2001 were significantly greater for younger seedlings than older seedlings in both
the edge and sun habitats.
This experiment was motivated by the continuing decrease in the area of cativo-
dominated swamp forests due to degradation from repeated entry logging. Planting
cativo wildings in the understory of the forest provided a baseline for comparison of
growth and survival with wildings planted in two habitats that might be considered when
the restoration of degraded swamps is a land management obj ective. There is a tradeoff
in terms of growth and survival and the labor requirements of site preparation and
maintenance. Initial clearing of dense, extensive liana tangles by hand proved to be much
more difficult than clearing areas in the edge habitat, and periodic cleaning was also
much easier in the edge habitat because lianas did not resprout as vigorously there.
Cativo has often been described as shade tolerant (Linares Prieto 1987a, 1988,
Tamayo Velez 1991) or as shade tolerant early in life but requiring high light levels for
further development (Linares Prieto et al. 1997). Younger seedlings had much higher
mortality in the shade of the forest understory than in the high light habitat, while older
seedlings had similar mortality rates in these two habitats, suggesting that seedling
establishment requires more than a single year.
Mortality rates observed in this study were generally higher than in a cativo
reforestation experiment in Colombia that examined the effects of different types of
planting techniques and timing of planting (Linares Prieto 1987a). That study occurred
on abandoned agricultural soils while the study site in Darien is less suited for agriculture
due to a slight brackishness in the soil and long hydroperiods (Tapia 1999).
Although mortality was highest in the shaded understory of the forest, seedling
deaths were not soley attributable to shade or root competition. During periodic
inundations, fallen branches float around and damage small cativo seedlings. Several
planted seedlings in the understory were found bent over completely by coarse woody
debris. Liana tangles in open habitats, in contrast, prevent this woody debris from
moving. Another agent of mortality present in the forest but not likely in liana-infested
areas was trampling by people, particularly hunters.
These findings add to the already highly variable published growth rates of cativo
seedlings and saplings. Maximum mean annual height growth after four years in the
present study was observed in the high light environment, where younger and older
seedlings grew 48 and 54 cm yr- respectively. These growth rates are similar but
predictably lower than those from a reforestation experiment in Colombia on less acidic
soils and in high light, where Linares Prieto (1987) reported annual rates of height growth
ranging from 67-76 cm yr- In that study, slow plantation height growth was attributed
to a lack of suitable mycorrhizae, infertile soils, or poor plantation management. Still,
plantation growth greatly exceeds seedling growth in natural forest. In natural cativo-
dominated forest in Darien, Panama, a large population (~ 1300 trees) of natural
regeneration, from newly germinated seedlings to small saplings 150 cm tall, revealed
essentially zero mean height growth for trees between 60 cm and 150 cm tall. Mean
annual height growth of seedlings less than 30 cm tall was 15 cm yrl and for seedlings
between 30 and 60 cm tall mean annual height growth was 5 cm yrl (Grauel 1999).
Cativo's fastest height growth seems to occur after germination until seed reserves are
exhausted (Tamayo Velez 1991).
In certain cases cativo may have the desirable socioeconomic and biophysical
attributes that would make it the opportune choice for reforestation. Small landholders or
cooperatives may have more reasons than just timber production for reforestation.
During the course of this study the author was asked by a member of a small-scale
loggers' cooperative about the feasibility of restoring degraded swamps by planting
cativo. A comment was also made regarding the lack of wildlife in the increasingly
extensive teak plantations in Panama, implying that native species attract wildlife. Cativo
seeds are consumed by agoutis (Dasyprocta aguti) and its leaves by land crabs (possibly
Gecarcinus spp), a species that also is harvested by local people.
Cativo is the principal raw material for the domestic plywood industry in Panama
but is also used locally for furniture and construction; its value lies in its relative
abundance and accessibility. Native timber species such as Tabebuia rosea, Dalbergia
retusa, Astronium graveolens, and Pachira quinata are currently being planted on a
commercial scale in Panama (Mariscal et al. 2002, Wishnie et al. 2002a, Wishnie et al.
2002b), but it remains to be seen if cativo can compete with other, more valuable
hardwoods, both native and exotic, for planting on upland soils. This study demonstrates
that cativo is particularly suited to reforesting severely degraded sites that are unsuited
for agriculture due to flooding and that previously were in cativo.
Table 3-1. Mean canopy openness at 1.3 m above the ground as estimated with a
spherical densiometer, arranged by seedling age and habitat.
Seedling Age Habitat % Openness
shade 10.7 0.7
Younger edge 38.7 12.3
sun 83.9 4.1
shade 11.1 1.6
Older edge 27.8 13.4
sun 85.2 9.2
Table 3-2. Initial mean seedling height and diameter (at 20 cm above the ground;
standard errors noted parenthetically). All differences between seedling ages
are significant (p < 0.01) based on t-tests. N = 4 plots in blocks 2-4 and N = 3
plots in block 1.
Age Younger Older Younger Older
Block Height (cm) Diameter (mm)
1 41.4 (0.75) 56.9 (1.56) 2.7 (0.05) 5.6 (0.25)
2 40.0 (0.82) 55.1 (1.94) 2.6 (0.08) 5.8 (0.41)
3 43.3 (0.39) 60.5 (2.60) 2.7 (0.02) 5.9 (0.21)
4 50.4 (2.73) 65.8 (1.33) 3.0 (0.19) 6.5 (0.45)
Table 3-3. Mean annual growth rates. Units of absolute growth are millimeters for
diameter and centimeters for height. Comparisons are among habitats within
each seedling age for different time periods throughout the study. All
comparisons are significantly different (P < 0.05) except where noted, "a"
indicates no difference between edge and sun, and "b" indicates no difference
between edge and shade.
Diameter Growth Year
1997-1998 1998-1999 1999-2001 1997-2001
Age Habitat Relative Absolute Relative Absolute Relative Absolute Relative Absolute
Sun 1.27 4.0 1.11 8.0 0.62 9.6 2.44 7.5
Younger Edge 0.83" 2.6ab 0.54 3.0b 0.28b 3.1 1.03 3.3b
Shade 0.27 0.7 0.14 0.6 0.04 0.1 0.05 0.2
Sun 0.54 3.5 0.98 9.7 0.51 10.4 1.30 8.3
Older Edge 0.33 1.9 0.46 3.7b 0.24 3.2 0.53 3.1
Shade 0.02 0.1 0.04 0.2 0.05 0.3 0.05 0.3
Height Growth Year
1997-1998 1998-1999 1999-2001 1997-2001
Age Habitat Relative Absolute Relative Absolute Relative Absolute Relative Absolute
Seedling Diameter Sun
Seedling Height Sun
-*- Younger Seedlings
-o Older Seedlings
Seedling Diameter Edge
Seedling Height Edge
Seedling Diameter Shade
Seedling Height Shade
I I I
Aug 97 Apr 98 Nov 98 May 99Nov 99
Jul01 Aug 97 Apr 98 Nov 98May 99Nov 99
Figure 3-1. Mean diameters and heights (+ 1 SE) of planted Prioria copaifera (cativo)
seedlings. Repeated measures analysis showed significant differences in size
among habitats but not between seedling ages when initial size (at time of
planting) was included as a covariate.
Younger Seedlings Sun
Older Seedlings Sun
Younger Seedlings Edge
40 -0 -0- -
Older Seedlings Edge
Younger Seedlings Shade
Older Seedlings Shade
Nov 97Apr 98 Nov 98 May 99 Nov 99 Jul01 Nov 97Apr 98 Nov 98 May 99 Nov 99
Figure 3-2. Percent seedling survival, beginning with the first census (November 1997)
after planting (September 1997). Each point pertains to percent survival of
the original 49 seedlings planted. Plot letters refer to pairs within habitats (A
and B shade, C and D edge, E and F sun), plot numbers refer to blocks.
STRUCTURE, COMPOSITION, AND DYNAMICS OF Prioria copalfera-
DOMINATED SWAMP FORESTS IN DARIEN, PANAMA
Unplanned selective timber harvesting over time results in a pattern of chronic
disturbance that strongly shapes forest structure and composition (Kittredge et al. 2003).
Timber harvesting is a discrete event that, alone, does not necessarily lead to forest
degradation. But where logging is poorly done or is too frequent, forests may become
susceptible to fires and liana infestations (Nepstad et al. 1999, Pinard et al. 1999,
Gerwing 2002). Given the unfortunate commonness of this disturbance regime, an
understanding of forest stand development in response to chronic degradation is critical
to the pursuit of sustainable forest management because the diverse values of forests
depend largely on forest structure and species composition (Oliver and Larson 1996).
The disturbance history of a site substantially influences present day forest stand
structure and productivity. Unfortunately, detailed site histories are usually unavailable
and evidence of past disturbances may not be obvious. Given that stand structure alone is
insufficient to indicate population trends in natural forests (Condit et al. 1998),
understanding patterns of recovery in degraded forests requires demographic information
as well. Without knowledge of how species, cohorts, and even individual trees respond
to disturbances such as logging, predictions cannot be made regarding the likely
responses of forest stands to further management interventions.
Neotropical swamp forests dominated by Prioria copaifera, a Caesalpinoid legume,
were long ago noted for their theoretical ease of management due to their low diversity
and the ability of this species to regenerate naturally (Barbour 1952, Holdridge 1964).
These traits have apparently allowed Prioria-dominated forests to be especially resilient
to repeated-entry logging. While potentially easy to manage, managers of Prioria-
dominated forests suffer from a lack of demographic information. In more diverse forests
where trees of most species are scarce, predictions are often based on small sample sizes,
although recent innovations such as large plots (Condit et al. 1999) and landscape-scale
sampling (Clark and Clark 1994, 1999) are addressing this limitation. But with few
exceptions (but see Favrichon 1998, Finegan and Camacho 1999, Fredericksen and
Mostacedo 2000), most information has been derived from unlogged forest preserves that
bear little resemblance to the actual working forests where harvesting occurs. And rarely
have studies of a particular forest type been replicated at different sites.
Although forest stand structure can reveal clues about disturbance regimes and past
uses, simple size distributions of stems often obscure underlying dynamic processes. In
particular, resprouting from snapped or partially uprooted trees can greatly influence
demographic parameters such as growth, recruitment, and mortality. Although
resprouting is increasingly recognized as an important component of forest dynamics,
many forest dynamic models omit this mechanism of regeneration and consequently
overestimate forest recovery rates (Paciorek et al. 2000).
Foresters typically focus on growth rates and on the adequacy of regeneration,
overlooking the importance of mortality rates of target species in the development of
management plans, often because they lack data. Although correlations have been made
between mortality and stand density (Lugo and Scatena 1996), light environment (Davies
2001), forest fragmentation (Mesquita et al. 1999), and climate (Condit et al. 1995b, Aiba
and Kitayama 2002), the estimation of a given species' mortality rate is usually difficult
due to small sample sizes and limited study periods. Although my study is based on only
Hyve annual censuses, I had the advantage of fairly large sample sizes that permitted
examination of mortality by stem type (fallen, inclined, resprout from erect stem, stem
sprout from prostrate or inclined stem) as well as evidence of growth-dependent
mortality. Although growth-dependent mortality has been reported for temperate forests
using growth estimates derived from tree rings (Kobe et al. 1995, Pacala et al. 1996,
Kobe and Coates 1997, Wyckoff and Clark 2000, Caspersen and Kobe 2001, Lin et al.
2001, Bigler and Bugmann 2003, van Mantgem et al. 2003), a frequent lack of tree rings,
inadequate sample sizes, and insufficient numbers of censuses have precluded
investigations of growth-dependent mortality in tropical forests (but see Finegan and
Camacho 1999 for a stand-level analysis).
Studying forests dominated by Prioria copaifera allowed collection of a large
dataset for a single, commercially important species and its main associates. In this paper
I describe the stand structure of five P. copaifera dominated forests in eastern Panama
and report on stand dynamics of four of those five sites based on Hyve years of monitoring
data; I also supply tree growth data from an additional six sites. All trees were described
on the basis of evident lean, breakage, and resprouting. Because P. copaifera sprouts
from erect broken stems as well as from inclined and partially uprooted trees, I
differentiate between these sprout types in presentations of data on growth, recruitment,
and mortality rates.
Distributed from Nicaragua to Colombia and also found in Jamaica, Prioria
copaifera (hereafter "cativo") was recognized as a potential source of commercial timber
in the first half of the twentieth century (Kluge 1926, Schmieg 1927, Anonymous 1933).
During and after WWII, interest in cativo wood increased (Harrar 1941, 1942a, b, Hess
and Record 1950, Hess et al. 1950, Barbour 1952). Much of the cativo forest in Costa
Rica was exploited to the point of current scarcity (Veiman 1982, Jimenez Madrigal
1995), while in the more remote parts of Panama and Colombia cativo forests were and
continue to be subj ected to silvicultural experimentation and intensive timber harvesting
(Lamb 1953, Mayo Melendez 1965, Christiansen 1980, FAO 1982, INRENARE 1987,
Linares Prieto 1988, FAO 1990, Alvarado Q. et al. 1996, CONIF 1997, Mariscal et al.
1999, Grauel and Pineda M. 2001). Although limited in extent, wetland forests
dominated by cativo are valued as sources of timber because of their large commercial
volumes (Golley et al. 1969, Grauel and Pineda M. 2001) and their ready accessibility by
Permanent plots were installed in 1997 at four sites in three watersheds in Darien,
Panama (Figure 4-1). Small populations of cativo trees at an additional six sites
provided additional information about growth of this species under a range of inundation
regimes. Eight of the ten sites were previously logged at various intensities and
frequencies, most recently three years before this study began. Plots were also installed
at a remote site near the Colombian border in Darien National Park where there was no
evidence of cativo logging.
Casarete is located along the banks of the Balsas River (8o 07' 1 1-14" N, 77o 52' 19-
47"Wi), 19 km upriver from its confluence with the Tuira River and 48 km from the
Tuira' s mouth at the Gulf of San Miguel (Figure 4-1). Soils are heavy clays classified in
the suborders Fluvent and Aquept, are acidic to slightly acidic (pH 5.2-6.5), poorly
drained, and slightly brackish (electrical conductivity 5.0 mmhos/cm, Tapia 1999);
mangrove forests are found only 7 km downriver. Rainfall measured at the study site in
1998 and 1999 was 2970 mm and 2758 mm, respectively. Mean annual rainfall at
Camoganti, the nearest town, (8.5 km from the study site), is 2457 mm (Reputblica de
Panama 1995). Because the forest owner values the area for hunting and aesthetics he
protected it from logging for approximately 25 years (but did harvest a few large trees
approximately ten years before the study).
The Sambu River site (8o 03' 49"-04' 06" N, 78o 13' 16-33" W) at the mouth of the
small Chunga River, is 17 km upstream from the Sambu's mouth at the Gulf of San
Miguel. This site is approximately 4.5 km from Boca de Sabalo, also on the Sambu
River, and 9.5 km from the Wounaan village of Taimati, on the coast of the Gulf of San
Miguel, where mean annual precipitation is 1342 mm and 1592 mm, respectively
(Reputblica de Panama 1995). Soils are similar in texture but slightly less acidic than
Casarete (pH 5.7-6.8), and although it is nearer to the ocean, there was no evidence of
salinity (Tapia 1999). The forest is located on open-access, public land and was
repeatedly logged, the most recent entry three years before the study began.
Juanacati is 65 km from the mouth of the Tuira River near the town of El Real (8o
04' 38-49" N, 77o 46' 37-48" W). Mean annual rainfall at El Real (5 km from the site) is
2096 mm (Reputblica de Panama 1995). Although Juanacati, like the previous riverine
sites, is flooded by monthly spring tides, no soil salinity is evident and soil pH is higher
than the other sites (pH 6.2-6.3). There is no evidence of recent logging, but anecdotal
evidence and the forest' s proximity to El Real suggest that the site was repeatedly logged
in the past. Indeed, L.R. Holdridge noted the presence of stumps only a few kilometers
downriver from this site in 1962 (Holdridge 1964).
The fourth site is an inland location near a small intermittent stream called
Naranzati (8. _02' 58"-4' 26" N, 770 -55' 40"-58' 02" W), approximately 7 km west of the
town of Camoganti. Unlike the three riverine study sites, this forest is flooded for the
entire nine month wet season (April to December). Due to its inaccessibility, this site and
other inland swamps were subj ect to logging only recently as riverine cativo forests
became increasingly depleted of large trees.
Figure 4-1. Principal study sites. a) Casarete, b) Sambu, c) Juanacati, and d) Naranzati.
To compare cativo growth in different landscape positions and flooding regimes, I
chose three sites adj acent to mangrove forests and subj ect to brackish water inundation.
Bajo Grande (80 22' 23" N, 780 09' 24-31" W) is an area near La Palma on the coast of the
Gulf of San Miguel; the study site is a few hundred meters inland from the coast, behind
the coastal Rhizophora mangle forests. The other two tidal sites are slightly upriver from
the transition zone of red mangrove to cativo forest, one along the Tuira River (8o 10' 13-
33" N, 77o 50' 18-23" W) and the other along the Balsas River (8o 09' 03-17" N, 77o 53' 04-
10" W). All three tidal sites are monodominant, uneven-aged cativo stands.
Additionally, I chose three sites that are unaffected by brackish water for
monitoring cativo growth. Two of these freshwater sites are along the Balsas River,
between the principal Balsas River site (Casarete) and the community of Camoganti
(Bongales: 8o 05' 30" N, 770 51' 35" W and Limon: 8o 04' 13" N, 770-53' 16" W). Although
these sites are occasionally inundated by high tides, the floodwaters are comprised of
freshwater tidal backup and are not brackish. The final freshwater site is an inland
swamp far up the Amarraderro River, a small tributary of the Balsas River (8o 00' 48-55"
N, 77o 49' 38-41" W). All six secondary sites, except possibly the latter inland swamp,
were subjected to intermittent logging during the last decades of the 20th Century. A
seventh site is found near the headwaters of the Balsas River inside Darien National Park
near the Colombian border (7o 34' 23"-35' 32" N, 77o 47' 02-11" W), 130 km upriver from
the mouth of the Tuira River. This site was probably never logged for cativo due to its
remoteness, a condition that limited my access to a single visit.
Because of the low tree species diversity of cativo-dominated forests, I was able to
use small plots and still gather sufficient demographic information for cativo and several
common associates. To capture landscape heterogeneity, and because some cativo
forests are limited to narrow bands along rivers, many small plots were installed at each
site instead of single large plots. Casarete has Hyve 40 x 40 m and Hyve 20 x 20 m plots,
Sambu has Hyve 40 x 40 m plots, and at Naranzati there are Hyve 40 x 40 m and three 40 x
20 m plots. At Juanacati, all trees > 4 cm dbh were measured in six 50 x 50 m plots. At
the other three sites, all trees > 10 cm dbh were measured in all plots and trees > 1 cm
dbh were measured in a randomly chosen 20 x 20 m quadrant of each plot, or in the entire
plot if it was 20 x 20 m (Table 4-1).
Most plots were installed in early to mid-1997, but the work was interrupted at
Naranzati and Juanacati and was completed in early 1998. All trees were tagged and
mapped at the time of plot installation and all plots were subdivided into 5 x 5 m subplots
to facilitate mapping. Trees > 7 cm dbh were measured with a diameter tape to the
nearest millimeter while smaller trees were measured with calipers, with two
perpendicular measurements being averaged.
Sampling and Analyses
To increase the sample size for large cativo trees, additional trees outside the
permanent plots were tagged and measured at Casarete and Naranzati. This approach
was not feasible at Juanacati and Sambu due to the overall scarcity of large cativo trees.
Three of the eight plots at Naranzati were mistakenly logged two months before the
1999 census. Analyses after that time were done using data from only the unlogged
In late 1997 and early 1998 all cativo stems < 1 cm dbh in 80 randomly chosen 5 x
5 m subplots within the larger plots at Casarete and 28 subplots at Sambu were tagged,
mapped, and measured (height); dbh was also measured for those saplings > 1.5 m tall.
These two populations of seedlings as well as new cativo recruits at these sites were
tagged, measured, and mapped approximately every two months for two years and
marked seedlings were measured annually.
The permanent plots were censused annually until 2001, except the Naranzati site
which was last censused in 2000. At the secondary sites trees 2 20 cm dbh were
measured in 1997, 1998, and 1999, except the Amarradero site, which was logged after
the 1998 census. Inter-census intervals were always nearly annual to minimize seasonal
effects on growth.
Methods of censusing and measuring trees as well as the approach to data checking
generally followed Condit (1998). During each census, in addition to recording each
tree's status (alive, dead, recruit), diameter (trees > 1 cm dbh), and height (trees < 1 cm
dbh), codes were assigned to denote if a stem was prostrate, inclined > 450 from vertical
but not lying on the ground, broken above or below the point of measurement (POM),
resprouting from a broken stem, or sprouting from a prostrate or inclined stem. I report
the incidences of these stem types, but for growth analyses I exclude fallen stems and
combine erect and inclined stems. I refer to stems that show evidence of previous
breakage and subsequent resprouting as broken stems. Although I separately coded trees
that had broken depending on if the break was above or below the POM, in this paper I
combine the two because there were only a few cases where a substantial recorded loss in
diameter was the result of stem breakage and subsequent resprouting. Living fallen stems
were measured if possible but were always tagged and mapped because they served as
hosts for vertical sprouts. References to prostrate stems refer to uprooted, not snapped,
stems. I measured all vertical sprouts from prostrate stems and from inclined stems if
they emerged from < 1.3 m from the ground.
Figure 4-2. Stem types: a) prostrate and inclined and b) vertical sprouts from a fallen
Two professional foresters performed all the ~ 22,000 dbh measurements in the
permanent plots. One forester measured trees only in 1997 and 1998, while the other (the
author) measured half the trees in 1997 and 1998 and all trees from 1999-2001, overall
measuring 80% of the trees in the study. I was also responsible for all error checking and
All data were independently entered into a computerized database by two different
people immediately upon return from the Hield. Any discrepancies between the dbh
measurements that could not be corrected in the office were noted, and follow-up field
trips a few weeks after the principal census were carried out to remeasure trees or
Within sites, I first examined variability among plots of each year' s annual cativo
growth of erect, broken, and vertical sprouts with ANOVA and Tukey post-hoc tests.
Data were natural log-transformed if variances were substantially unequal. For these
comparisons, the longest annual growth record available was used. Of 3065 cativo trees
in the four principal sites, 80% of the annual growth records were for the period 1997-
2001, 14% were for 1998-2001, and 6% were for 1997-2000. For cativo, growth was
then compared among normal stems, broken stems, and sprouts from inclined/prostrate
stems within five diameter classes using ANOVA and Bonferroni post-hoc tests. To
explore inter-annual growth of cativo I used Hyve diameter classes and I combined all
stem types but excluded those that were prostrate on the ground. I also report annual
diameter growth of cativo' s three principal arboreal associates.
I report annual recruitment and mortality rates for cativo at the four principal sites
using four diameter classes for trees > 1 cm dbh and Hyve height classes for smaller trees
at the two sites where cativo regeneration was studied. In each census I used the totals of
stems from the previous census, that is, I do not calculate demographic parameters using
only the originally-tagged 1997 population (see Sheil and May 1996). The monitored
large trees outside the plots at Casarete and Naranzati were included in calculations of
mortality rates. Furthermore, in the few cases where trees were not found in a particular
year' s census, those trees were subtracted from the previous year' s total number of trees
to exclude them from calculations of recruitment rates, instead of assuming the trees had
I evaluated the extent to which cativo mortality varied with recent growth rates.
With Hyve censuses at most sites, I was able to compare annual growth for a maximum of
three years between trees that were alive at the end of the study and those that died during
the study. I used two size classes (< 10 and > 10 cm dbh) and performed one-tailed t-
tests on populations with at least Hyve dead stems to test the hypothesis that slower
growing trees suffered a higher probability of mortality.
I report species diversity using Fisher' s alpha and the Shannon-Weiner (S-W)
index in two ways, first with only all stems > 4 cm dbh so that all four sites can be
compared. Then I also calculated the S-W index for the three sites with minimum dbh of
1 cm. Voucher specimens were collected from the four principal sites in 2000;
unidentified species were not separated into morphospecies; I noted the number of
unidentified species at each site and considered all unknown trees as a single species in
the diversity calculations.
Tree Species Diversity and Stand Structure
Although cativo dominated all four principal sites, species diversity varied
substantially. The riverine sites on the Sambu and Balsas (Casarete site) Rivers showed
the greatest cativo dominance, with cativo comprising 95 and 96% of the stems of all size
classes, respectively (Table 4-2). Diversity indices grouped the four sites into pairs, with
Casarete and Sambu being strongly monodominant and Juanacati and Naranzati being
relatively more diverse. Only seven tree species were tallied at each of the former two
sites, while 48 and 54 species were identified at Juanacati and Naranzati, respectively.
For trees > 10 cm dbh at the five sites with plots, cativo comprised from 46-96% of the
stems and from 33-96% of the basal area (Table 4-3). Pterocarpus officinalis was
cativo' s principal associate common to all four sites, making up 2-10% of the stems.
Other common overstory trees at Juanacati were Pentaclethra macroloba (10.2%),
Calrapaguianensis (7.2%), Licania platyipus (3.3%), and M~ora oleifera (1.3%). At the
Naranzati inland swamp, cativo tended to dominate the overstory with P. officinalis_and
P. nzacroloba, but understory associates included Andira inernzis_(1.3%), Escinveilera
integrifolia (4.7%), Gustavia nana (2.0%), and Brownea rosa-de-nzonte (2.5%). The
palms Oenocarpus nzapora and Astrocalyunt standlyanunt were common understory
species at the two more diverse sites as well, comprising 2-8% of the stems.
Stand density and basal area varied considerably among sites (Table 4-3). Stand
density of trees > 10 cm dbh ranged from 328 trees/ha (Darien NP) to 757 trees/ha
(Casarete). Stand basal area varied less markedly, but most cativo basal area was
represented by trees > 60 cm dbh at the inland swamp (Naranzati) and in Darien NP, and
by stems 10-60 cm dbh in the riverine forests.
Cativo made up ~ 95% of stand basal area at the riverine Sambu and Casarete sites
and ~ 83% at the inland Naranzati swamp (Table 4-3). The Juanacati site was visually
dominated by 14 huge, emergent M~ora oleifera trees per hectare that made up 21% of
stand basal area, the same percentage as Pterocarpus officinalis, while cativo comprised
33% of basal area and Pentaclethra nzacroloba 14%.
For all species, substantially more large prostrate and severely inclined trees were
found at the three riverine sites than at the inland swamp or the remote Darien NP site.
No prostrate, living trees were noted at either Naranzati or Darien NP. In contrast, large
prostrate stems were fairly common at the three riverine sites, where 1-5% of live stems
>10 cm dbh were on the ground (Table 4-4). Large inclined stems were also more
prevalent in the riverine forests, where 4-8% of all stems were partially uprooted and
leaning > 450
The rates at which trees partially uprooted and fell to the ground or leaned > 450
were generally higher at all sites for trees > 10 cm dbh than for smaller trees (Table 4-5).
Consistent with the proportion of live fallen and inclined stems recorded at the beginning
of the study, rates of falling and inclination were higher at the riverine sites than the
inland swamp, but rates varied greatly among years within sites.
At the time of plot installation a modest proportion of stems showed signs of
previous stem breakage at most sites. Where trees < 10 cm dbh were measured, between
6-11% were broken, and approximately 3-6% of large stems had suffered but recovered
from stem breakage. No broken and resprouted stems were noted in the plots at Darien
Sprouts from prostrate and inclined stems were much more prevalent in the riverine
forests than inland swamps. Less than 2% of small stems (< 10 cm dbh) at the inland
Naranzati swamp consisted of these sprouts, and no larger sprouts were found. In
contrast, 6-17% of smaller stems were sprouts at the riverine sites. Among the riverine
sites, Casarete stood out by having more large than small stems classified as sprouts,
where 12% of all stems > 10 cm dbh were sprouts from prostrate or inclined parent trees.
The 23.5 m long stem of a 33 cm dbh cativo that partially uprooted and fell to the ground
at Casarete between the 1999 and 2000 censuses produced 175 vertical sprouts > 1.5 m
tall by the 2001 census. Fifteen of these shoots were inside the plot (within 5.7 m from
the root system of the parent tree) and were tallied as recruits in 2001, with mean and
maximum diameters of 1.7 and 3.2 cm dbh, respectively.
Cativo Growth, Mortality, and Recruitment
Relative rates of cativo growth of three stem types (undamaged,
broken/resprouting, or sprouting from inclined and prostrate trunks) are unique to each
site. Sprouts grew faster than either undamaged or broken stems at Casarete for stems
<40 cm dbh, while at Sambu the smallest sprouts grew slower than undamaged stems and
only sprouts 4-10 cm dbh from inclined or prostrate stems grew faster than undamaged
stems of the same size. Undamaged stems grew faster than broken stems at Juanacati but
growth rates of normal stems and sprouts from prostrate/inclined trunks were similar
Cativo annual diameter growth varied considerably among both sites and years, but
a few patterns were apparent. Growth was slowest, with some exceptions, during the El
Nifio year of 1997-1998 and fastest for the following census period (1998-1999). For
trees < 10 cm dbh, the Casarete and Naranzati sites were both characterized by very slow
growth rates, while stems of this size class grew significantly faster in almost all years at
Sambu and Juanacati (Table 4-7). Mean diameter growth rates within the larger size
classes (2 10 cm dbh) varied notably among the 9 sites (10 sites for 1997), ranging from
only 0.5 mm to > 8 mm per year.
Sprouts from prostrate or inclined stems made up a significant portion of forest-
wide recruitment in many cases (Table 4-8). About 50% of saplings entering the 1 cm
dbh size class at Casarete were sprouts of this type. Also at Casarete, these sprouts
consistently comprised substantial portions of the recruitment into the 10 cm dbh size
Recruitment and mortality rates for cativo show similar inter-site patterns to those
observed for growth. Mortality equaled or exceeded recruitment at Casarete and
Naranzati for trees <10 cm dbh in the first two years of the study (Table 4-9). In general,
recruitment of cativo at Sambu and Juanacati greatly surpassed mortality for all size
classes and years.
Mortality rates of different cativo stem types varied among sites, but in general,
sprouts from prostrate/inclined trunks and undamaged stems showed the highest mortality
rates at Casarete and Sambu (Table 4-10). No sprouts from prostrate/inclined trunks died
at either Juanacati or Naranzati; at these two sites either broken or prostrate stems had
mortality rates that approached or occasionally exceeded those observed for undamaged
Growth of other Tree Species
Pterocarpus officinalis was the fastest growing tree species at all four sites, with
mean annual diameter growth of 9-1 1 mm for trees > 10 cm dbh in three of the four
forests (Figure 4-2b). The other two most common associates of cativo, Calrapa
guianensis and Pentaclethra macroloba had mean annual growth of large trees > 10 cm
dbh that approached 5 mm at the two more diverse sites where they were found.
Growth-dependent Mortality of Cativo Trees
Surviving trees grew significantly faster (p < 0.05) than trees that died for nine of
ten t-test comparisons (Table 4-11). The non-significant case revealed significance
when the size class was further divided. For 1997-1998 growth, survivors at Casarete
<10 cm dbh grew equally slowly as those trees that subsequently died, but when these
smaller trees were analyzed as two size classes, surviving trees between 1-4 cm dbh grew
significantly faster than those trees that died (mean 0. 1 vs. -0. 1 mm, df = 90, t = 3.1, p =
0.002), while trees 4-10 cm dbh in both groups had essentially zero growth.
Regeneration was much more abundant at the more recently logged Sambu River
site than at Casarete (Table 4-12). Small seedlings (< 30 cm tall) were not abundant at
either site, suggesting that they grow rapidly after germination. Seedlings 30-60 cm tall
were most abundant at Casarete, while at Sambu seedlings 60-90 cm tall were the most
common. Due to a large seedfall in April and May 1999, annual cativo seedling
recruitment rates based on the period November 1998 November 1999 for Casarete and
Sambu were 136.6% and 29. 1%, respectively. Annual mortality for the same period was
similarly high for seedlings < 30 cm tall but differed substantially for taller trees between
the two sites (Table 4-13). Mean annual height growth was generally < 5 cm at each site
(Table 4-14). Exceptions were the relatively small number of seedlings < 30 cm tall at
both sites and saplings > 90 cm tall at Sambu, both of which grew rapidly.
Forests dominated by cativo in Darien, Panama vary substantially in structure,
species composition, and stand dynamics. Flooding regimes and management histories
appear to be the main determinants of present-day structure and dynamics. Slight soil
salinity, in particular, seems to favor cativo dominance (Mayo Melendez 1965). Cativo
forests near mangrove forests exhibit almost total dominance by cativo, whereas inland
swamps and riverine forests that escape tidal flooding with brackish water contain
relatively high tree species diversity.
Cativo dominance can probably be attributed to a variety of mechanisms. Flood
tolerance alone is an insufficient explanation because other flood tolerant species are
generally rare in cativo forests (Lopez and Kursar 1999). Cativo's root system dies back
to a much lesser extent than other flood tolerant species and gives the species competitive
advantage in seasonally flooded forests that are subj ect to short but severe annual
droughts (Lopez 2002). High leaf area index, as shown by Holdridge (1964), may
modify the understory environment to be more favorable for cativo seedling survival than
for other species. Such a modification was attributed to Gilbertiodendron dewevrei, a
tropical tree species that forms monodominant stands in West Africa. Although the crab
species common to riverine cativo forests have not been studied in Darien, the land crab
Gecarcinus quadratus was shown to affect species diversity in a coastal Costa Rican
forest by selective seedling consumption (Sherman 2002).
The degree of cativo dominance varies widely in Panama. Cativo is a locally
common species in the 50 ha forest dynamics plot on Barro Colorado Island, with a mean
of 27-29 trees hal > 1 cm dbh and a maximum of 223 (Condit et al. 1993b). The three
sites in Darien for which I have comparable data have densities of cativo 5-15 times
greater than on BCI. Similar to cativo swamps in Darien, the relative basal area
dominance of cativo in Colombia is 50-92% (Escobar and Vasquez 1987). Anecdotal
evidence suggests that Caurapa guianensis, a species that produces wood of a similar
quality as mahogany, was probably much more common in the past in some cativo
forests and L.R. Holdridge (1964) identified C. guianensis as cativo's sole associate in a
1962 transect very near the Juanacati site on the Tuira River.
The histories of timber harvesting of these forests undoubtedly influence present
day stand structure and dynamics, but details on logging frequencies and intensities are
largely unknown. It is likely that all the riverine forests were repeatedly logged since the
1950s (Reputblica de Panama 1978). Although I characterized the inland swamps and the
Darien NP site as intact forests, it is possible that even these remote forests were logged
for mahogany several decades ago.
Stand structure analyses of cativo forests reveal high timber volumes or at least the
potential for high volume production. The forest at Darien NP stands apart from all other
sites with the lowest density of stems but the largest stand basal area. The other sites
(except Juanacati) contain higher basal areas than most other tropical forests (Leigh
1999). This is notable because the riverine cativo forests of Darien were identified for
their timber potential in the 1950s (Lamb 1953) and L.R. Holdridge noted stumps from
harvesting activities in the early 1960s near the Juanacati site (Holdridge 1964). The
history of logging is undoubtedly responsible for the paucity of trees >60 cm dbh (the
legal cutting limit), but these riverine forests still contain a similar number of stems >10
cm dbh as most other lowland tropical forests (Leigh 1999).
High stem density at Casarete may be a result of the site' s management history.
Having been protected by its owner from the repeated logging that typically occurs on
open access state land, this forest may have passed through a period of enhanced
recruitment after the first wave(s) of logging in the 1950s and 1960s. Increased
recruitment and growth of undamaged residual trees after logging is a well-documented
phenomenon (e.g. Magnusson et al. 1999, Parrotta et al. 2002). The Casarete forest 40
years ago may have been similar to the present-day Sambu forest that was recently
logged and exhibits high seedling and sapling densities, fast sapling growth, and low
Sprouting is a well-recognized regeneration strategy in tropical forest (Putz and
Brokaw 1989, Rijks et al. 1998, Gavin and Peart 1999, Kammesheidt 1999, Negreros-
Castillo and Hall 2000, Yamada et al. 2001) but sprout density varies from being
common (Paciorek et al. 2000) to absent in mature forests (Kammesheidt 1998).
Vegetative sprouts from various sources may be the principal colonizers of gaps (Putz
and Brokaw 1989, Negrelle 1995), but mortality rates of resprouted broken stems are
generally higher than non-sprouts (Guariguata 1998, Paciorek et al. 2000, Ickes et al.
2003). In cativo dominated forests it is important to distinguish between resprouts
originating from erect, broken trees and those that emerge from fallen and inclined trees.
Sprouts from fallen or inclined trunks are frequently much more common, often grow
more rapidly than trees originating from seed but have the highest mortality rates of any
stem type, including trees that have uprooted and are lying on the ground.
The tree recruitment assemblage in newly formed treefall gaps in cativo forests
may not be dominated by true seedlings. In some cativo forests the principal gap
colonists were not newly germinated seedlings or established shade tolerant saplings;
instead, regeneration was dominated by sprouts from inclined or prostrate stems. For
example, at least half the recruitment of 1 cm dbh stems every year at Casarete was
comprised of sprouts from fallen stems, which also grew faster than their conventionally
rooted counterparts. Sprouts from prostrate stems develop their own root systems
composed of roots that emerge from the bottom of the parent stem. At Casarete, these
sprouts continued their superior growth rates at least into the subcanopy after which it
was difficult to determine their mode of regeneration.
The three riverine sites share a somewhat similar history of logging as well as a
higher proportion of live prostrate and inclined trees than the inland swamps that have not
been logged. Logging and inclined or prostrate trees may not necessarily be related,
however. Riverine cativo forests were characterized by a large number of fallen trees at
about the time that widespread commercial logging was beginning in Darien (Duke 1964,
Holdridge 1964), but recent logging has been sporadic due to scarcity of commercial-
sized trees. Cativo sawnwood and plywood production during the late 1990s was only a
quarter of its peak in the late 1960s (Romero M. et al. 1999). I conclude that fallen and
inclined cativo trees may be a common feature of riverine forests due to their shallow
root systems and saturated soils, with logging being a lesser factor.
Growth rates of the various stem types at different sites may vary with forest
structure but may also be limited by different factors depending on stem type. Although
sprouts from inclined/prostrate trunks at Sambu and Casarete grew at similar rates,
undamaged stems and broken stems at Sambu grew four to five times faster than their
counterparts at Casarete, presumably because the recent logging at Sambu left a more
Growth of all cativo stem types increased with increasing distance from tidal
influences but decreased with increasing hydroperiods. In general, cativo trees in the
inland swamps that are flooded continuously during the rainy season grew more slowly
than in the riverine forests, and the riverine sites further upriver (Juanacati, Bongales,
Lim6n) showed faster mean cativo growth than downriver sites.
Diameter increment of canopy trees varied by site and year, but the patterns of
variability differed among sites. Canopy tree growth at Casarete varied up to 2-fold, with
1997 as the slowest growing year. At the other three principal sites, canopy tree growth
varied less, and the census period that spanned the 1997-1998 El Nifio was not always
the slowest growing year. These findings stand in contrast to the strong reductions in tree
growth in Costa Rica during the 1997-1998 El Nifio, which were negatively correlated
with daily minimum temperatures (Clark et al. 2003).
The fastest growing tree species in my sample plots was Pterocarpus officinalis,
which is not considered a timber tree. Cativo's two other most common associates,
Calrapa guianensis and Pentaclethra macroloba, are harvested in both Panama and Costa
Rica (Webb 1997, Sitoe et al. 1999). C. guianensis in particular is valued by Darien
loggers, while P. macroloba produces less valuable wood. C. guianensis is more
abundant in riverine forests flooded only with freshwater, and has moderate growth rates
and relatively high densities in some Darien cativo forests. This species may have been
locally extirpated in some riverine forests but could be reintroduced by seed scattering as
was recommended by Webb (1997) for logging gaps in swamp forest in Costa Rica.
In the absence of spatially-explicit data that allow for the development of
competition indices and the construction of distance-dependent forest dynamic models,
higher survival of faster growing trees (growth-dependent mortality) should be taken into
account when projecting growth traj ectories. When mean growth of a large population of
small trees is low, modeling lifetime growth based on mean growth may result in
unrealistically long traj ectories (Grauel and Kursar 1999), especially if a large proportion
of slow growing trees die before reaching commercial size.
I noted heavy seedfall during plot installation in April 1997, but the regeneration
study began in September 1997 at one site and May 1998 at the other, so I measured
recruitment in 1998-2000 at one site and 1999-2000 at the other. Although cativo
produces some seeds twice a year, large seedfalls seem to occur once every two years
(Pizano SA 1995), and I also measured a large pulse of seedling recruitment from July to
November 1999, two years after the large observed seedfall in April-May 1997.
When stand structures and dynamics vary so markedly among forests, it seems
inadvisable to extrapolate results widely. Witness the difference in cativo dynamics
between these Darien cativo forests and the population of cativo on Barro Colorado
Island. With much lower mortality rates and generally higher mean and maximum
growth rates for cativo on BCI, use of their data for management tasks such as timber
harvest scheduling or yield proj sections would justify over-harvesting of most Darien
cativo forests. Mortality, more than growth, was shown to be a pivotal factor in the
simulated sustainability of cativo harvest potential based on BCI data (Condit et al.
1995a), but annual mortality rates of Darien cativo forests varied greatly and were
sometimes much higher than on BCI.
A critical aspect for understanding present day forest structure and dynamics is the
history of use that has resulted in what are now degraded forests. Increasingly, these
degraded forests will be a source for wood and non-wood products as the area of intact,
mature tropical forest declines. For example, the forest most recently logged in this study
(Sambu) is one of the most dynamic and resilient, with relatively high growth and
recruitment rates, low mortality, and sufficient densities of advance regeneration to
theoretically provide additional timber harvests. Forest history, although it may only be
inferred, can yield insights into today's forest and perhaps help to guide management
This study highlights the importance of examining stand development patterns at a
variety of sites, even when a single "forest type" is identified based on species
composition and landscape position. Only where variability is recognized can it be
considered when making forest management decisions. Anthropogenic disturbance may
have been the primary factor in shaping the structure and function of many cativo-
dominated forests in Darien, but the persistence of these forests after decades of
harvesting attests to their resilience and should serve as inducement for better
Table 4-1. Total plot area measured for different minimum tree diameters and number of
tree species found.
Plot Area (ha) Number of Tree Species
Site >1 cm >4 cm >10 cm >1 cm >4 cm >10 cm % unidentified
Casarete 0.4 0.4 1.0 8 7 6 < 0.1
Sambu 0.2 0.2 0.8 7 5 5 0.0
Juanacati -1.5 1.5 48 24 0.4
Naranzati 0.32 0.32 0.96 54 42 28 1.5
Table 4-2. Species diversity indices and relative dominance of cativo (Prioria
DBH > 1cm
Site Fisher's a S-W Evenness Simpson dominance
Casarete 1.10 0.20 0.09 0.93 0.96
Sambu 1.00 0.22 0.11 0.91 0.95
Naranzati 12.37 2.10 0.53 0.32 0.55
DBH > 4cm
Site Fisher's a S-W Evenness Simpson dominance
Casarete 1.01 0.22 0.10 0.92 0.96
Sambu 0.76 0.28 0.16 0.87 0.93
Juanacati 9.52 2.09 0.54 0.25 0.47
Naranzati 7.19 1.81 0.52 0.36 0.59
Table 4-3. Stem density and basal
area of all species (above) and cativo only (below).
Stems/ha BA/ha (nr)
>10 cm >10 cm
Cativo Cativo Cativo Cativo
Stems/ha BA/ha (nr') Stems/ha BA/ha (m )
>10 cm >10 cm >60 cm >60 cm
727 41.5 8 2.5
463 39.9 12 4.4
195 10.3 3 0.9
240 39.3 51 29.2
160 48.8 58 45.4
Table 4-4. Incidence (%) of prostrate, inclined, broken stems and sprouts from prostrate
trunks. Other stems showed no signs of earlier breakage and presumably
regenerated from seed.
Prostrate Stems Inclined Stems Broken Stems Prostrate Stems
<10 >10 >10 >10 >10
Site cm cm <10 cm cm <10 cm cm <10 cm cm
Casarete 0.0 5.0 3.0 7.7 8.2 4.5 8.3 12.0
Sambu 0.0 3.1 1.9 7.2 6.1 4.0 16.7 1.6
Juanacati 2.1 1.1 4.2 4.1 10.9 5.6 5.9 0.7
Naranzati 0.0 0.0 3.5 0.9 9.7 3.1 1.8 0.0
Darien NP 0.0 0.0 0.0 0.0
Table 4-5. Forest-wide annual treefall and tree incline rates (i.e., partial uprooting) for
small (above) and large (below) trees for four sites.
< 10 cm dbh 1997-98 1998-99 1999-2000 2000-01
Incline Fall Incline Fall Incline Fall Incline Fall
Casarete 0.40 0 0 0 3.22 0.57 0.23 0.23
Sambu 0.17 0 0 0 0.13 0.13 0.15 0
Juanacati 1.48 0.37 0.97 0 1.18 0.34 0.49 0
Naranzati 0.87 0 0.55 0 1.16 0.93 -
> 10 cm dbh
DBH cm Undamaged Broken DBH cm Undamaged Broken inclined trunks
1-4 0.9" (92) 2.7" (8)-
4-10 1.2" (79) 2.6" (5) -4-10 3.0" (288) 2.0b (34) 2.7ab (15)
10-20 2.7" (61) 1.7" (5) -10-20 5.3" (164) 2.8" (9) 5.7" (3)
20-40 4.7 (43) --20-40 8.5" (70) 2.5b 7
S40 4.5 (140) 40 7.1 (32)--
Table 4-6. Mean annual diameter growth (mm/year) of cativo trees of three stem types,
based on 1997-2001, 1998-2001, or 1997-2000 census periods. Within
diameter classes, different letters denote significant differences (p<0.01)
among stem types, (sample sizes noted parenthetically).
DBH cm Undamaged
1-4 0.35" (427)
4-10 0.7" (205)
10-20 1.80 (288)
20-40 3.5a (204)
S40 2.6 (120)
Table 4-7. Mean annual diameter growth (mm/year) of cativo trees. Statistical
comparisons are vertical, within diameter classes and among sites; different
letters denote significant differences (p<0.01; sample sizes noted
parenthetically). All stem types except prostrate are grouped.
1997-98 1-4 cm 4-10 cm 10-20 cm 20-40 cm > 40 cm
Casarete 0.1b (485) 0.0b (242) 1.2b (374) 2.3C (230) 1.7b (123)
Sambu 1.8a (396) 2.2a (138) 4.4a (153) 4.3b (108) 3.3ab (107)
Naranzati 0.3b (64) 0.2b (52) 1.0b (37) 3.0b,c,d (20) 3.5a (23)
Juanacati -3.1a (118) 4.7a (53) 5.4a.b (26) 7.5a (9)
Bajo Grande --2.0a,b (9) 1.7c~d (70) 1.9b (48)
Rio Amarradero .7a,b (11) 0.5c~d (51) 2.0b (92)
Rio Balsas1 .5d (96) -0.5b (21)
Bongales --3.1a,b (11) 5.8a.b (62) 5.9a (34)
Limon --5.0a,b (4) 8.4a (36) 8.1la (8)
Rio Tuira --3.1ab (11) 2.0cd (151) 1.2b (38)
1998-99 1-4 cm 4-10 cm 10-20 cm 20-40 cm 2 40 cm
Casarete 0.5b (491) 0.9b (237) 2.5C (376) 4.3C (227) 3.5C (126)
Sambu 1.6a (410) 3.7a (148) 6.4b (156) 5.6bc (105) 3.90 (104)
Naranzati 0.6b (185) 1.2b (117) 2.4C (97) 4.2C (67) 3.5C (176)
Juanacati -4.5a (332) 8.2a (179) 10.5a.b (78) 9.4a.b (31)
Bajo Grande --3.7abc (8) 8.0b (68) 11.0a (49)
Rio Balsas1 3.2" (98) 4.1b (20)
Bongales --4.5a,b,c (9) 8.7a.b (57) 7.2b (36)
Limon --6.8a,b,c (3) 12.0a (34) 12.7a (10)
Rio Tuira --4.6a,b,c (11) 7.2b (144) 5.7b.c (39)
1999-00 1-4 cm 4-10 cm 10-20 cm 20-40 cm 2 40 cm
Casarete 0.5b (483) 1.1b (238) 1.7b (372) 5.0b (226) 3.2C (128)
Sambu 1.3a (427) 2.6a"b (162) 4.2a.b (157) 3.8b (110) 2.3" (104)
Naranzati 1.6a (100) 1.7b (83) 2.6b (66) 5.4b (43) 5.1b (139)
Juanacati -3.0a (338) 5.1a (190) 8.4a (81) 7.8a (39)
2000-01 1-4 cm 4-10 cm 10-20 cm 20-40 cm 2 40 cm
Casarete 0.8a (479) 1.0b (243) 2.1b (355) 3.6b (224) 2.0b (126)
Sambu 1.1a (459) 1.9a (170) 4.0a (169) 3.3b (116) 1.6b (107)
Juanacati -1.3b (332) 2.6b (197) 5.1a (83) 5.2a (40)
Table 4-8. Ingrowth by stem type. Percentage of recruited individuals from broken stems, undamaged stems, or sprouts from
prostrate and inclined trees.
Table 4-8. Continued
Table 4-9. Cativo annual recruitment and mortality rates (%) by stem diameter class for
four census periods.
-1998- 1999- 2000-
99 00 01
4.6 2.8 2.1
1.6 1.7 0.6
2.8 0.4 0.7
0 0.5 0
0.4 0 0.7
Table 4-10. Annual mortality rates (%) of cativo trees by stem type and stature for four census periods.
Table 4-11. Mean annual growth (mm/year) of cativo trees that were alive at the end of
the study and those that died during the study for which there was one or more
years of growth data.
Alive 2001 Dead 2001 df
0.04 (683) -0.09 (44) 65
1.7 (712) 0.3 (17) 25
1.9 (519) 0.5 (14) 16
0.2 (107) 0.6 (9) 9
Site DBH cm
Site DBH cm
Site DBH cm
Dead 2001 df
0.2 (21) 33
1.6 (13) 13
.03 (12) 14
8.7 (294) 3.6 (6) 5 2.5 0.025
Dead 2001 df
-0.8 (7) 15
4.9 < 0.001
17.2 < 0.001