RETURNING MATERIALS:
bV1ESI.} Place in book drop to
“saunas remove this checkout from

-—__ your record. FINES m]?

 

 

 

 

be charged if book is
returned after the date
stamped beIow.

 

 

 

 

 

 

 

 

 

 

 

 

REGROWTH OF CLEARCUT SUBTROPICAL DRY FOREST:
MECHANISMS OF RECOVERY

AND QUANTIFICATION OF RESILIENCE

BY

Vicki L. Dunevitz

A THESIS

Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

MASTER OF SCIENCE

Department of Botany and Plant Pathology

1985

ABSTRACT

REGROWTH OF CLEARCUT SUBTROPICAL DRY FOREST:
MECHANISMS OF RECOVERY
AND QUANTIFICATION OF RESILIENCE

BY

Vicki L. Dunevitz

The degree of recovery and mechanisms of regeneration
were investigated on a one ha lB-year-old cut-over dry
forest site near Guanica, Puerto Rico. Recovery was
relatively rapid, with leaf area index 64% and basal area
and mean plant height 40% of adjacent mature forest
values. Fifty-ninespercent of mature forest species
occurred in the Cut Site. Species diversity was higher in
the lB-year-old site than in one-year-old or mature forest
sites, with pioneer species contributing 46% of stem
density and 28% of species. Half the plants with stems 3
1 cm originated as sprouts from roots or cut stems, the
remainder as seedlings. Numbers of woody seedling origin
plants varied with distance from the forest edge, but not
significantly. Quantities of VA mycorrhizae did not_
differ in three different-aged dry forest sites.
Significant effects of 1969 recut/herbicide treatments

still persisted in 1982.

ACKNOWLEDGEMENTS

I would like to thank Dr. Peter Murphy for financial
support through NSF Grant DEB-8110208 and for his advice
and guidance. Thanks are extended to my other committee
members, Dr. Deborah Goldberg and Dr. John Beaman, and
to Dr. Ariel Lugo, for helpful suggestions and critical
comments. Alice Murphy provided invaluable assistance
with seed bank studies, species identification, and field
work. Baudilio Hernandez, Eusebio Santiago, and Roberto
Santiago provided inestimable help with field work. I am
grateful to Dr. Gene Safir for the use of his laboratory
equipment and for technical advice on mycorrhizal
analyses. I thank Dr. Robert Jansen and Dr. Cheryl
Crowder for taxonomic assistance, and Dr. Jay Harman for
advice and friendship.

Colleagues who have supplied unflagging support as
well as advice and critical comments include Maureen
Mulligan, Terese Hart, Martha Tilmann, Susan Gibbons, John
Hart, Leni Wilsmann, Dan Nepstad, Helen Kuhn, and Barbara
Wilber. This thesis would not have been possible without
the energetic support and enlightening analyses provided
by the Women in Science group (W888) and the Thursday

night volleyball league.

ii

TABLE OF CONTENTS

List Of TableSOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOiv
List Of FigureSOOOOOOOOOOOOOOO ...... OOOOOOOOOOOOOOOOOvi

Introduction..........................................l
Theoretical Background.............................2
Objectives and Hypotheses..... ...... ... ....... .....5
Literature Review

Forest Recovery.................................9
Dry Forest Ecology.............................12

The stUdY AreaOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO17

MethOdSOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO24
Vegetation Structure and Composition..............24
Seed Banks........................................31
MycorrhizanOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO34

ReSUItSOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO36
Site Description. O O O O O O O O O O O O O O O O O O O O O O O O O O O O O O O O O36
Vegetation Structure and Composition

speCies DivergitYO O O O O O O O O O O O O O O O O O O O O O O O O O O O O O39
structureOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO41
Mechanisms of Regeneration........................S4
MycorrhizanOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO66
Effects of Ewel's Recut/Herbicide Treatments......70

DiSCUSSiOnO O O O O O O O O O O O O O O O O O O O O O O O O O O O O O O O O O O O O O O O O O O72
Temporal Patterns of Succession...................72
Mechanisms of Regeneration........................79
Mycorrhizan O O O O O O O O O O O O O O O O O O O O O O O O O O O O O O O O O O O O O O83
Recovery and RegilienceO O O O O O O O O O O O O O O O O O O O O O O O O O O84

summarYOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOgl

AppendiXOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO93

Literature CitedOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO99

iii

Table

l.

10.

ll.

12.

13.

LIST OF TABLES

Page

Importance values for plant species with
DBH > 1 cm in Cut Site and adjacent forest

plotsOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO42

Structure of Cut Site and adjacent forest
vegetationOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO43

Leaf area index and percent cover in the
cut SiteOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO43

Structure of woody vegetation in Cut Site, in
adjacent forest, and on Cut Site before it

was cutOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOSO

Comparison of herbs in Cut Site and adjacent
forest between dry season (March) and wet
season (August)OOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO53

Effect of distance from forest edge on
seedling species and densities in Cut Site
2x2mplotSOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO60

Total seed densities in soil and litter
samples in September and March................62

Densities of seed species < 2.38 mm
germinating from soil and litter samples under
greenhouse conditions.........................63

Densities and species of seeds > 2.38 mm
collected from soil and litter samples........64

Viability and germinability of fresh seeds
c011eCtEd inAUQUSt1982.0.00000000000000.000067

Effects of scarification treatments on
Leucaena leucocephala seed germination........68

Mean mycorrhizal ratings of species in three
different-aged SiteSOOOOOOOOOOOOOOOOOOOOOOOOOO69

Structure of vegetation in the portion of

the Cut Site which had recut/herbicide
treatments in 1969 and in the portion

"hiCh was cut onIYOOOOOOOOOOOOOOOOOOOOOOOOOOOO71

iv

14.

Al.

A2.

A3.

A4.

Degree of recovery of structural parameters
in thirteen-year-old dry forest...............86

Characteristics of all species occurring
in plotSOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO93

Structural parameters of all stems > 5 cm
DBH on the study site before it was cut
in1969000OOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO96

Importance values for woody plant species
with DBH > 5 cm on the study site before
itwascut in1969000OOOOOOOOOOOOOOOOOOOOOOOOO97

Multivariate comparisons of structure in the
Cut Site and forest 2 x 10 m plots using
the HOtEllingTteStoooeeoeooooooooooooooooo0.98

LIST OF FIGURES

Figure Page
1. Location of the Guanica Commonwealth Forest...18

2. Monthly distribution of annual rainfall in
two Puerto Rican dry forest sites.............l9

3. Location of the thirteen-year-old Cut Site,
adjacent forest study area, and forest plots
studied by Murphy and Lugo....................22

4. Location of 2 x 10 m Cut Site and forest

plotSOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO26

5. Location of 2 x 2 m Cut Site and forest

plotSOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO28

6. Leaf area index sampling locations............30

7. Appearance of Cut Site vegetation in 1982,
13 years following cutting....................37

8. Appearance of adjacent forest vegetation in

1982000OOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO38

9. Species area curves for Cut Site and forest
2 x 2 m sampling plots........................40

10. Comparison of structure in Cut Site and
adjacent foreStOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO44

11. Leaf litter biomass in forest and Cut Site
in september and MarChOOOOOOOOOOOOOOO'OOOOOOOOO46

12. Numbers of plants in height interval classes
in cut Site and foreStooeooooooooeeo000000000047

13. Vertical stratification of Cut Site
vegetation using vegetation area index
readinQSOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO48

14. Size class distribution in the 2 x 10 m plots
in forest (N=30) and Cut Site (N=10)..........51

15. Percentages of woody stems > 1 cm DBH
originating as vegetative sprouts and as
seedlings in Cut Site and forest plots........56

vi

16.

17.

18.

19.

20.

Percentages of woody plants with stems >_l

cm DBH with stems originating as vegetative
sprouts and as seedlings in Cut Site and

forest plots..................................57

Height class distribution of woody plants
in Cut Site and forest originating as
vegetative sprouts and as seedlings...........58

Relationship between numbers of woody
seedling-origin plants in Cut Site and
distance to the forest edge...................61

Species diversity changes relative to
succession in the Guanica forest..............73

Patterns of structural change in long-term
dry foreSt succeSSionOOOOOOOOOOOOOOOOOOOOOOOOOBB

INTRODUCTION

The widespread human impact on many of the world's
forests has resulted in a great deal of interest in and
concern with the ability of forested ecosystems to recover
following disturbance. Because of the constraints on
obtaining data on long-term forest recovery, the most
formidable being the long time period involved (sometimes
up to several hundred years), rates and processes of
natural forest regeneration remain incompletely
understood.

Especially lacking are studies of succession in
tropical and subtropical dry forest (sensu Holdridge
1967). Since this community type constitutes an estimated
42% of the world's tropical and subtropical forests (Brown
and Lugo 1982), and is subjected to a myriad of land uses
causing various degrees of disturbance (Tosi 1980), it is
a topic that gravely needs addressing.

In 1981, a detailed study of dry forest dynamics and
succession in the Guanica Commonwealth Forest in Puerto
Rico was undertaken by P. Murphy and A. Lugo to begin to
address that need (Murphy and Kephart 1981, Murphy and
Lugo 1983,1985). A nearby one ha stand of forest which
had been clearcut in 1969 by John Ewel (Ewel 1971)
provided an ideal opportunity to study the degree of
longer-term recovery of the forest. This study was

1

undertaken in 1981 to attempt to elucidate the mechanisms
and measure the degree of recovery of a dry forest

thirteen years following clear-cutting.

Theoretical Background

The study of ecosystem recovery rates involves several
components of successional change. One that is of prime
importance is ecosystem resilience. Much confusion has
persisted regarding the use of the terms ”stability” and
”resilience” and their components (e.g. see discussions
in Margalef 1975 and Westman 1978). Stability is used
here to mean the pattern of fluctuation in a relatively
unimpacted ecosystem over time (sensu Westman 1978), with
low and/or predictable amounts of fluctuation equating
with high stability. Whittaker (1975) identified nine
aspects of stability. The first eight are described as
the application of four variables to both the community
and the abiotic environment: 1) relative amplitude of
regular fluctuation, 2) relative irregularity of
fluctuation, 3) presence of zero values in fluctuation,
and 4) the duration of this pattern in evolutionary time.
The ninth aspect is the one addressed in this study:
resilience, defined here as the ability of a system to

return to its initial, predisturbance condition after a

disturbance or stress period.

Degree of resilience can be expressed in empirical
studies as 1) the time required to restore a particular
characteristic of an ecosystem (e.g. biomass, basal area,
etc.) to a given percentage of pre-impact level, or 2)
the percent similarity of the disturbed ecosystem to the
undisturbed ecosystem after a particular period of time.

Cairns and Dickson (1977) developed a recovery index
for aquatic ecosystems based on subjective ratings of six
parameters. Ewel (1980) proposed a less subjective index
calculated with 12 structural measurements: stem density,
biomass, four measures of foliage density, three measures
of height, and three measures of species richness. He
compared relative resilience of five sites by expressing
each variable as a percentage of the maximum value
obtained at any of the five sites studied, and then
averaged the percentages for each site. This method is
limited to studies with comprehensive data from equal-aged
sites, and, more importantly, does not take into account
predisturbance structure.

There has been much discussion regarding potential
causes of variation in resilience among ecosystems.
Clapham (1971) suggested that resilience is maximized in
regions in which productivity, ”niche breadth”, and
environmental fluctuations are high, though he provided no

mechanisms for the effects of these factors on resilience.

He predicted that the tropical rainforest has a relatively
low resilience despite the small amplitude of
environmental fluctuation and high productivity because
organisms are adapted to narrow, highly specialized
niches, making them extremely vulnerable to perturbations.
He suggested that temperate deciduous forests, V
characterized by moderate fluctuation, productivity, and
"niche breadth", have relatively high resilience values.
Lacking in this theory are predictions of the relative
contribution of each of the 3 variables; it is doubtful
that they are weighted equally. Other hypothesized causes
of high relative resilience include low species diversity
(Loucks 1970) and small size of disturbed area (Hill
1975). Additional factors affecting resilience are
disturbance history (Holling 1975), abiotic properties
(Rolling 1973), and spatial organization (Hill l975).
Cairns and Dickson (1977) developed a recovery index
for aquatic ecosystems based on subjective ratings of 6
parameters. Ewel (1980) proposed a less subjective index
calculated with 12 structural measurements: stem density,
biomass, four measures of foliage density, three measures
of height, and three measures of species richness. He
compared relative resilience of 5 different sites by
expressing each variable as a percentage of the maximum
value obtained at any of the five sites studied, and then

averaged the percentages for each site. This method is

limited to studies with comprehensive data from equal-aged
sites, and does not take into account predisturbance

structure 0

Objectives and Hypotheses

The fundamental objectives were to assess the degree
and some of the mechanisms of recovery of 13-year-old dry
forest vegetation, and to quantify the forest's
resilience. To accomplish this, structural measurements
of the one ha site in Guanica, Puerto Rico are compared
with measurements made in adjacent older secondary forest,
representative of the most mature stands in Guanica.
Additionally, data collected on the site before and one
year after it was cut (Ewel 1971, 1977, and unpublished
data) are summarized to provide information on early site
recovery. Several hypotheses guided the research. The
first addresses resilience of dry forest vegetation in
relation to other forest ecosystems. The next four make
predictions regarding aspects of the dry forest recovery
process, and the last addresses the effects of multiple
disturbance on dry forest recovery. The hypotheses were

as follows:

A. Resilience

Dry forests are more resilient than tropical humid to

wet forests, and equivalent in resilience to temperate
dECldUOus forests. Clapham (1971) and Ewel (1980)
suggested that ecosystems in intermediate-stess
environments, such as temperate and tropical deciduous
forests, may be more resilient than those of more
benevolent (e.g. tropical rain forest) or more harsh
(e.g. montane rain forest or tundra) environments.
Clapham argued that intermediate-stress environments may
possess higher fluctuation and organisms with greater
phenotypic plasticity than benevolent environments, and
higher productivity than harsh environments. The lack of
recovery data from sites with age and disturbance history
comparable to that of the study site makes absolute
comparisons of relative ecosystem resilience impossible at
this time. In this study, dry forest resilience is
quantified and discussed in relation to short-term
recovery data from other forest types, and suggestions for

further study of resilience are made.

B. Diversity

Plant species diversity in successional dry forest
vegetation is higher than in mature forest stands.
Margalef (1963) postulated that diversity in ecosystems
tends to peak following a spurt of productivity such as
that occurring in young successional vegetation, and then
declines as the system matures and productivity lessens,

but he did not present empirical data to support this. In

this study, diversity of vegetation from 1-year-old,
lB-year-old, and mature dry forest sites are compared to
determine whether this pattern holds true during dry

forest recovery.

C. Mechanisms of Regeneration

(1) Vegetative sprouting, rather than seed
germination, is the dominant mode of regeneration in the
dry forest. Sprouting is known to be of importance in
cut-over dry forest recovery (Ewel 1977, Murphy and Lugo
1983). Seedling emergence and survival, on the other
hand, may be inhibited in dry forests due to lack of
sufficient soil moisture (Ewel 1977). Relative
proportions of sprouted stems and seedling-origin stems
are examined in the l3-year-old site and in mature forest
to determine whether there were more plants which had
orirginated as seedlings following the 1969 clearcut than
those which had originated as sprouts.

(2) Distance to the forest edge is inversely related
to species distribution and density of seedlings in a 1 ha
successional dry forest site. It appears from the few
studies available that the size of area disturbed is a
major determinant of the rate and pathway of succession.
When very large vegetated areas are disturbed, the
invasion of propagules must necessarily take place from
the edge of the plot, resulting in a possible delay of

full species reestablishment (Gomez-Pompa et a1. 1974).

The edge effect on species distribution of seedling-origin

plants is measured on the one ha site.

D. Mycorrhizae

Vesicular-arbuscular mycorrhizal associations are not
appreciably impacted by the cutting of a dry forest. All
of 33 plant species in the Guanica forest examined for
presence or absence of mycorrhizae were infected to some
degree (Dunevitz et al. 1983), indicating that VA
mycorrhizae are important components of the dry forest
ecosystem. Janos (1980) predicted that cutting a tropical
forest would not greatly affect mycorrhizal populations
when tree roots are left intact and soils remain in place,
since the primary impact is on the aboveground structures.
This hypothesis is tested by comparing mycorrhizal
quantities in l-year-old, lB-year-old, and mature dry

forest sites.

E) Multiple Disturbance

MUltiple cutting and disturbance treatments applied to
dry forest for one year following cutting will result in
less site recovery 12 years later than will cutting alone.
The findings of Isensee et al. (1973), Ewel (1971), and
Ewel et a1. (1981) supported the idea that increasing
intensity of disturbance generally results in slower rates
of recovery. A 6000 m2 portion of the Guanica study site

received a mosaic of repeated cutting and herbicide

treatments for one year following cutting; structural
variables there are compared to those on the remainder of

the site.

Literature Review

Forest Recovery

The essential components for estimating resilience
include predisturbance data, comprehensive measurements of
a number of structural variables following disturbance,
and documentation of the time and extent of disturbance.

A review of the forest succession literature reveals that
there are few published reports which include all of these
components.

Many forest succession studies have relied on the
reconstruction of historical changes using aerial
photographs, tree-ring analyses, fire scar analyses, or
comparisons of different-aged old-growth stands (Kittredge
1934, Martin 1959, Vogl 1969, Auclair and Goff 1971, Day
1972, Henry and Swan 1974, Taylor 1980, Sprugel and
Bormann 1981, Romme and Knight 1981, Romme 1982, Agee and
Smith 1984). These studies were limited by the
availability of historical information, and studies using
different-aged stands are further complicated by
site-to-site variability. A number of reports of tropical

forest succession are largely descriptive and lack

10

quantitative information (e.g. Symington 1933, Richards
1955, Nye and Greenland 1960, and Fox 1976). Studies of
post-agricultural succession (e.g. Grieg-Smith 1952,
Budowski 1963, 1965, Kellman 1970, Pickett 1982, Uhl and
Jordan 1984) demonstrate temporal sequences which are
likely to differ radically from those on cut-over forest
lands, as farming practices generally modify soils,
eliminate seed sources of woody vegetation, and kill
stumps and rootstocks which may otherwise resprout.
Forest recovery studies focusing on just one aspect of
ecosystem change, such as leaf production (Covington and
Aber 1980), species diversity (Shafi and Yarranton 1973,
Tomkins and Grant 1977), root biomass (Berish 1982), or
forest floor organic matter and nutrient content (Likens
et a1. 1978, Covington 1981, Binkley et a1. 1982) are
too limited to be useful in comparisons of ecosystem
resilience. Early studies of regeneration of American
northern hardwood stands following cutting (Leak and
Wilson 1958), fire (Larsen 1925), or hurricanes (Spurr
1956) focused only on measuring the restocking of
desirable tree species, while other floristic elements
were ignored.

More comprehensive studies (e.g. Ross 1954, Odum and
Pigeon 1970, Webb et a1. 1972, Malmer et al. 1978, Ewel
et al. 1981, Uhl 1982, Hibbs 1983) either lacked

information on pre-disturbance forests or measured only

11

short-term recovery. Perhaps the most detailed published
forest succession study to date (Bormann and Likens 1981)
reported only the first six years of regeneration
following cutting and herbicide application. Marks and
Bormann (1972) measured revegetation on 1 to 14-year-old
cut-over northern hardwood stands dominated by Prunus
pensylvanica, but reported only limited data on climax
associations.

While the studies cited above demonstrate a limited
amount of data on temperate and wet tropical forest.
recovery, even less is known about tropical dry forests,
despite their large areal extent. Opler et al. (1977)
compared the rate of recovery on a one-year-old cut and
burned dry forest to that of one to three-year-old wet
forest in Costa Rica, relying primarily on mean seed
weight data and making the assumption that lighter seeds
are found in younger sites. They predicted full recovery
times of 150 and 1000 years for dry and wet forests,
respectively. Van Riper (1980) suggested that in Hawaiian
dry forests, tree species composition and distribution
have been appreciably impacted by human disturbance,
browsing by feral mammals, and changes in seed dispersal
due to the presence of introduced avian species. Ewel
(1971, 1977, 1980) examined regrowth for one year using
various combinations of cutting and herbicide treatments

in five tropical forests ranging from the dry forest site

12

of this study (880 mm rainfall) to tropical wet forest
(4800 mm rainfall). Lowland wet forests grew the fastest,
exhibiting the maximal values of species richness, leaf
area index, biomass, and average height of the 3 tallest
plants per plot. The 2 dry forest sites showed lower
recovery rates, and a tropical montane rain forest was

least recovered (Ewel 1980).
Dry Forest Ecology

Old world dry forest formations include the African
miombo (reviewed and described by Malaisse 1978), dry
deciduous forests (Bourliere 1983), and Australian
semiarid woodlands (described by Pressland 1975). Water
relations of these and other areas were discussed by Doley
(1981).

Recent ecological studies demonstrate the variety of
species associations which fall under the category of New
World dry forest. Costa Rican dry forest (Hubbell 1979,
Opler et al. 1980) has structure and diversity similar to
Puerto Rican dry forest, but supports a very different
species association. Hubbell found that all dry forest
tree species were either clumped or randomly dispersed,
and that clumping was related to seed dispersal. Goldberg
(1982) found that soil factors strongly affected the
distribution of Lysiloma divaricata-dominated Mexican dry

forest and adjacent evergreen oak-woodland. Dry evergreen

13

' forests in Jamaica (described by Loveless and Asprey 1957)
had many species in common with Puerto Rican dry forests,
while Jamaican dry slope and dry limestone scrub forests
(studied by Grubb and Tanner 1976) were quite different.

Lugo et al. (1978) measured structure, productivity
and transpiration in the Guanica forest near the study
site of this investigation, and emphasized the role of low
soil moisture availability in determining structural and
functional aspects of the formation. Murphy and Lugo
(1983, 1985) are collecting comprehensive data on the
dynamics and succession of the Guanica forest. Their
study site includes a 1.4 ha plot located 1.0 km east of
the site used in this investigation. They found a total'
of 59 vascular plant species, including 33 species with a
minimum diameter at breast height (DBH) of 2.5 cm. Stem
density and basal area for stems 3 2.5 cm were 14,007/ha
and 19.8 mz/ha, respectively. Tree heights were small,
with fewer than 3% exceeding six m, and the tallest tree
sampled reaching only 8.5 m. The live-plant biomass was
89.9 t/ha, with about 50% occurring below ground.

Dry forest phenological studies have demonstrated a
distinct seasonality in leaf flushing, flowering, and
fruiting. In Costa Rican dry forests, where rainfall is
generally > 1500 mm/year, Daubenmire (1972) reported two
flowering seasons correlating with major and minor seasons

of drought. His work was substantiated by Frankie et al.

14

(1974), Janzen (1967), and Opler et al. (1980), who also
demonstrated a predominance of dry-season flowering in
Costa Rican dry forests. Malaisse (1974), in a study of
Zambezian miombo (mean rainfall 1270 mm/yr) found a
preponderance of flowering and leaf flushing in the warm
dry season just before the onset of the rainy season. In
areas with less rainfall than the Costa Rican and
Zambezian sites, phenological activity is apparently
largely limited to wetter seasons. Lieberman (1982) found
that flowering and fruiting in West African dry forest
(rainfall of 1100 mm/year) reached a peak during wet
periods, and the size of the fruit crops appeared to be
greatest following periods of abundant rainfall. Murphy
and Lugo (1983 and unpublished data) found that in the
Guanica forest, flowering tended to peak in May-July, a
period when precipitation is fairly high. Fruiting
followed a more sporadic, possibly bimodal pattern, with
peaks in January and in June-July. Of the 48 tree species
observed, there were never more than 32% flowering or 52%
fruiting at any one time. Leaf senescence in dry forests,
generally corresponding with the dry season, appears to be
triggered by drought stress more than by daylength
(Daubenmire 1972, Frankie et al. 1974, Lieberman and
Lieberman 1984, Murphy and Lugo 1985).

Data on seed dispersal, storage, and germination in

dry forests are sparse. Aristeguieta (1968) and

15

Daubenmire (1972) noted that animals and wind are
important vectors of seed dispersal in dry forests, but
did not quantify numbers of species which were gravity
dispersed. Cheke et al. (1979) found 40 seeds per .25
m2 and 16 species of seeds in 5 cm deep soil samples from
dry dipterocarp forest in Thailand. Hall and Swaine
(1980), in a study of seed banks from 3 different dry
forest types in Ghana, found that 107 to 696

seeds/m2 germinated from 4 cm deep soil samples. All
seeds were from secondary forest species except the 0.3 to
11.2% which were primary forest species.

Endomycorrhizae have been shown to be important
components of Tanzanian miombo (Hogberg 1982) and of the
Guanica forest (Dunevitz et al. 1983), corroborating
other reports of the prevalence of mycorrhizae in arid
zones such as deserts (e.g. Miller 1979, Diem et a1.
1981).

Herbivory patterns and rates are potentially important
factors in vegetative recovery. Herbivore damage to
leaves in Costa Rican dry forest was greater than that of
nearby riparian forest (Stanton 1975). Lieberman and
Lieberman (1984) reported that herbivory in African dry
forest was most common in newly flushed leaves and found
severe herbivore damage to leaves in 25% of the 59 species
studied. Forty-three per cent of the species had some

form of protection from herbivores. Janzen (1981) found

16

that folivory varied ”enormously" from year to year in

Costa Rican dry forest.

THE STUDY AREA

The 4000 ha Guanica Commonwealth Forest is located on
the southern coast of Puerto Rico at latitude 18 N (Figure
l). The low annual rainfall, ranging from about 600 to
1100 mm, is partially due to the forest's position in the
orographic rain shadow of Puerto Rico's Central
Cordillera. The monthly distribution of mean annual
rainfall in a dry forest site near Guanica (Figure 2)
illustrates a severe dry season, from January through
March, and a secondary moisture-limited period in mid to
late summer. The forest is classified as subtropical dry
forest by Holdridge's life zones classification system
(Holdridge 1967), based on mean annual precipitation,
biotemperature, and humidity.

The soil is derived from limestones of Miocene and
Oligocene age which surface in many places (Monroe 1976).
Soil types vary depending on the hardness and depth of the
limestone. Much of the Guanica forest is underlain by a
mollisol (Calciustoll: Aquilita series) which consists of
a 15-18 cm dark brown clay loam or clay surface layer
(bulk density a 0.7 g/cc), high in organic matter, which
grades into a soft, porous, mottled, white limestone (bulk

density a 1.3 g/cc) by 40-50 cm depthl. This soft

lUnpublished data collected by Daniel Nepstad, 1982.
17

18

.pmmcom ;p_mm3:oseou mowcoau on» Co cowuwuoE .H mc=m_m

Ex oH m o
. w i new among—Lou

mpoum

 

 

  

Baotou
mocoa . gupomzcoEEou
ouacoso

7

coac com

 

comuo owucopu<

IJoo mH

.l.om mH

MONTHLY RAINFALL (mm)

160

140

120

100

80

6O

4O

20

19

 

 

MONTH

Figure 2. Monthly distribution of annual rainfall in
dry forest at Santa Isabel, Puerto Rico (after Ewel and
Whitmore 1973).

2O

limestone has, in some areas, produced a layer of
indurated caliche above it. The remainder of the area is
underlain by shallower, redder soils, apparently derived
from harder limestone formations. Crevices in the rock in
the Guanica forest are sometimes filled by soils of 3 200
cm depth (Ewel 19171). Forest soils (measured at depths
of 5 and 15 cm) were found to be fertile in terms of total
nitrogen (9100 kg/ha), phosphorus (1820 kg/ha), and
potassium (7460 kg/ha) (Lugo and Murphy 1985). The high
exchangeable calcium (>4000 ppm) and high pH (7.3-7.6)
cause only 1.3% and 25% of total phosporous and potassium,
respectively, to be exchangeable.

The vegetation of the region was described by Gleason
and Cook (1926) and was labelled dry limestone forest by
Little and Wadsworth (1964). Wadsworth (1950) described
the climax dry forest of the southern coast of Puerto Rico
as being very dense, with tree heights from 40 to 60 feet,
and consisting of 2 tree layers, each with distinct
species. At present only old-growth secondary forest is
extant; all primary forests have been logged. Lugo et a1.
(1978) distinguished three upland plant associations in
the Guanica forest based on soil and floristic
differences; they included scrub, deciduous, and
semi-evergreen forest. The deciduous association, which
includes the sites of this study, occupies about 58% of

the Guanica forest.

21

The one ha site of this study is located 1.9 km from
the south coast of Puerto Rico, on a southwest-facing 5%
slope at 180 m elevation (Figure 3). The vegetation
occurring on the site before it was cut in 1969 had been
free from grazing and fire for about 40 years, but had
been subjected to some cutting (Ewel 1971). The
vegetation consisted of about half deciduous and half
evergreen species, with emergent trees on the site about
10-12 m tall (Ewel 1971).

Data collected on the study site previous to the 1969
cut were provided by J. Ewel. The sampling included
three 100 x_10 m transects across the site, in which the
species and DBH of each stem 3 5 cm and the total stem
density were recorded. The data are summarized in Tables
A1 and A2. Ewel identified 16 species with stems 3 5 cm
DBH and noted there were about 2 additional unidentified
species in each transect (Table A3), compared to 30

2 in that size class in the site studied by

species/1500 m
Murphy and Lugo (1983). The three most important species
in the Ewel site accounted for about half the total
importance value.

The mean density of 1127 stems Z 5 cm DBH per ha was
about half the density of 2307/ha in the site measured by
Murphy and Lugo (1983). Similarly, basal area on the 1 ha

site was about half the basal area in the site measured by

Murphy and Lugo: 4.9 mz/ha compared to 9.4 mz/ha. Though

22

 

 

 

 

 

 

l‘
N
PUERTO RICO
Scale
0 .25 .50 km_
Cut Site ‘
J -7
A /
Adjacent ”Q/ ( Forest / / 5‘“ '
. j .
Forest Site / \ Site .4 / I
/ /
/
I’— \ 1’ a”
,r"\ ‘\ /' \—’
9’
/ .. J \ \
’- fi. —- ‘— /
Paved Road "' "' Gravel or Dirt Road
...—.... Town
Figure 3.

Location of the thirteen-year-old Cut Site, adjacent

forest study area, and forest plots studied by Murphy and Lugo
(1983, 1985).

 

23

some density and growth increases probably occurred in the
12 year interim between the two sets of measurements, the
small annual dry forest tree diameter growth of 0.12 mm
(Murphy and Lugo 1983) and relatively low dry forest
seedling survival (Ewel 1980) would not account for the
differences between the two sites. Rather, environmental
site differences such as soil texture and structure,
elevation, and disturbance history probably account for

most of the differences in the vegetation.

METHODS

The one hectare study site was relocated and a trail
cut around the perimeter in July 1981 (Murphy and Lugo
1982). The field work for this study was conducted during
3 trips to the site in March 1982, August-September 1982,
and March-April 1983.

Statistical analyses were performed on the Michigan
State University Cyber 750 mainframe computer.

Statistical procedures used included SPSS BREAKDOWN (Nie,
Bent, and Hull 1975), which calculates means, sample size,
variance, standard deviation, and standard error for data
classified according to a number of variables; SPSS ONEWAY
and SPSS ANOVA (Kim and Kohout 1975), which calculate
one-way and more complex analyses of variance,
respectively, and provide a priori and a posteriori
contrasts; SPSS REGRESSION (Kim and Kohout 1975), which
calculates regression statistics; and BMDP3D ONE-SAMPLE
AND TWO-SAMPLE T-TESTS (Dixon, 1981), which performs a

Hotelling T Square multivariate test.
Vegetation Structure and Composition

Structural characteristics of woody vegetation in the
Cut Site were measured in twenty 2 x 10 m plots placed

randomly in the Cut Site in March 1982. Ten plots were in

24

25

the portion of the site which received a mosaic of cutting
and herbicide treatments following the 1969 clearcut
(hereafter referred to as the "recut/herbicided" portion),
and 10 were in the portion of the site which was merely
clearcut ("cut only” portion) (Figure 4)., For comparative
purposes, 30 plots were randomly placed in adjacent
forest. Species area curves and graphs of running mean of
basal area for each set of plots (in each site) indicated
that sample sizes were adequate. All stems 3 one cm DBH
were included in analyses, and epiphyte species and
numbers in the plots were tabulated. Measurements of
woody plants included density of stems per plant, plant
density, stem DBH, and origin of stems (stem sprout, root
sprout, or seedling). Stem origin was determined quite
easily in most cases by examining the stem bases.
Measurements of herbs and smaller woody plants as well
as larger diameter trees in the Cut Site were conducted in
2 x 2 m plots established in March 1982. Twenty-four
plots were placed at regular intervals along the diagonals
of the site (Figure 5) for an examination of the effect of
the distance from the forest edge on structural
parameters. Eight plots were in the "cut only" section of
the site, and the remaining 16 in the ”recut/herbicided"
section. Trails were cut parallel to and 2 m away from
the diagonals for access to the plots. 'Six additional 2 x

2 m plots were randomly placed in the "cut only" portion

26

 

Forest -——?

Adjacent 'lllllHl

 

 

 

 

 

 

 

 

 

 

 

 

 

Adjacent ///
Forest /<:>/ ///' //// ///’
__ TCut Only Area / / / /
L_________.
///’ l_ Recut-herbicided Area
_ / / ¢
-___- Cut Site
Scale '
O 10 2O 30 m Adjacent
‘ k | ‘ Forest
i

I ll

 

 

 

Figure A. Location of 2 x 10 m Cut Site and forest plots.

27

of the site in March 1983. Twenty 2 x 2 m plots were
randomly placed in adjacent forest and analyzed for
comparison in August 1982 (Figure 5). A11 plants with
.cover in a plot were identified and measurements of
height, crown area (computed from the average of two
diameter measurements), number of stems, origin of each
stem, and diameter at ground height (DGH) of each stem
were made. In the case of grasses, clumps (which were
generally < 0.2 m in diameter at ground level) were
tallied rather than stems. The plots were censused in
August 1982, and seasonal differences were examined by
remeasuring herbs in March 1983. When the field work was
completed, all plot stakes except those marking the 4
corners of the 1 ha site and the 24 Cut Site 2 x 2 m plots
were removed.

I identified all taxa to the species and variety
level. Nomenclature of trees follows Little and Wadsworth
(1964) and Little, Woodbury, and Wadsworth (1974); that
for all other species follows Britton and Wilson (1924),
except where noted. Species occurring primarily in
disturbed sites such as roadsides were considered pioneer
species, and those which were regular components of mature
forest were considered persistent species.

Importance values for all woody species in the 2 x 10

m Cut Site plots were determined using the equation

28

 

 

 

 

 

 

 

 

 

 

C) EJEJ
Adjacent U U D
Forest ’ U
C!
E]
E] Cut Only C!
Area
C) E! E]
Recut-herbicided E] D
B Area A C) U
E] [3
[II]
E] El E]
D D
[3
[3
Cut Site
El
Scale
6 16 2'0 50 m D U
U ‘ Adjacent
Forest
DD D

 

 

 

Figure 5. Location of 2 x 2 m Cut Site and forest plots.

29

IV = RF + RD + RDom

where IV: importance value, RF= relative frequency, RD=
relative density, and RDom= relative dominance (calculated
from basal area) (Brower and Zar 1977).

Vegetative cover on the Cut Site was measured with an
extendable pole which supported a taut vertical rope.
Measurements were made from trails so that the pole could
be easily maneuvered. Three cover indices were calculated
from 120 samples throughout the site (Figure 6), measured
in August 1982, March 1983, and May 1983. (A sample
consisted of one drop of the rope into the vegetation.)
The first, horizontal cover, was the number of ropes
encountering any vegetation, divided by the total number
of samples and expressed as a percentage. The second,
which I call vegetation area index (VAI), quantifies the
amount of plant tissue on a vertical plane over the site,
and is the number of times the rope intersected any plant
tissue (including herbaceous and woody) each time it was
dropped into the vegetation. The third, leaf area index
(LAI), was the number of times the rope encountered foliar
tissue each time it was dropped into the vegetation. In
August, VAI was assessed in 0.5 m vertical intervals to
quantify vegetation stratification.

Spatial and temporal (within-year) variability in leaf
litter biomass were investigated by removing all litter

2

from 10 round 0.25 m sampling areas in the Cut Site

30

 

 

 

 

o o o o
\ . /
\\ Cut Slte /‘
\ e/ '
. \ /.
\ 4 ‘
. \ /
2C -
/‘ \
/' \
o /. \ o
,/ \
o / \ o
/. \
ll . . . 0 O O
-— -Trail

 

0 Sampling Location

Figure 6. Leaf Area Index sampling
locations.

31

(located next to 2 x 2 m plots) and 10 in adjacent forest
(randomly located) on September 1, 1982 and March 24,
1983. Wood, insect, and rock components were removed from
the samples, and visible seeds were set aside for seed
bank studies. The foliage portion of each sample was oven
dried (105 C for 24 hours) and weighed.

Persistent treatment effects from Ewel's 1969 site
manipulations were estimated by comparing all measured
structural variables in the plots on the ”cut only” and

'recut/herbicided" sections of the site.

Seed Banks

Assessments of densities and species of seeds in the
Cut Site and adjacent forest were made by collecting soil
and litter samples on September 1, 1982, and on March 24,
1983. Seeds were collected from the litter biomass
samples and from the soil beneath each litter sample to a
depth of 5 cm. The soil samples were each divided into
two fractions using a 2.38 mm mesh screen. Seeds from the
large particle fraction were separated out by hand, and
each small particle fraction was sprinkled over sterilized
soil in trays which were placed in a greenhouse (with
constant 12 hour/day light regime, 15-29 C temperature,
watered 3-4 times/week) and monitored for seed

germination.

32

Seeds were removed from the litter collections by
different methods in August and March. In August, visible
seeds were removed by hand, and no quantification of small
seeds in litter was made. In March, litter samples were
divided into two size fractions using the 2.38 mm mesh
screen. Visible seeds were removed from the large
fraction, and the small fraction was weighed for addition
to litter biomass measurements, then sprinkled over
sterilized soil and germinated under the same greenhouse
conditions described above for soil samples.

Seed germination in all greenhouse trays was monitored
weekly for 9 months. Species were identified when they
were large enough to be recognizable, then seedlings were
tallied and pulled. If seedlings died before
identification was possible, they were recorded simply as
"dicot" or 'monocot”. There was some contamination in the
trays by local weed seeds, but these were easily
identified and removed.

The seeds separated by hand from soil and litter
samples were identified using plant specimens from the
Michigan State University herbarium and a reference seed
collection compiled with seeds from fruits collected in
the field. Several seeds from half the plots were
subjected to standard Tetrazolium tests to determine what
percentages were viable. Seeds were soaked in water

overnight, then covered with a 1% Tetrazolium solution (1

33

gm 2,3,5 Triphenyl Tetrazolium in 100 ml water, pH 6.7)
for six hours. Seeds were observed under a dissecting
microscope for stain indicating viability.

Fresh fruits were collected from several tree and
shrub species in August 1982 to determine whether fresh
seeds had viability and germinability similar to soil and
litter seeds. A subsample was tested for viability using
the standard Tetrazolium test, and subsamples of 200 seeds
of each of seven of the species were planted in sterile
soil in trays in the greenhouse. To determine whether
some seeds may require scarification to germinate, mature
fruits of Leucaena leucocephala were also collected in
March 1983, subjected to various scarification treatments,
and measured for germinability. Forty-five seeds were
used for each treatment; the treatments were: soaking in
sulphuric acid for 4 and 10 minutes, respectively, soaking
in boiling water for 2 minutes, and a control. Following
treatment, seeds were placed between moistened filter
papers in petri dishes in a growth chamber at 26 C with
76% relative humidity and 12 hours of light per day, for
21 days.

Mechanisms of seed dispersal of all species occurring
in the 2 x 10 m sampling plots were assessed with a
literature survey (using Little and Wadsworth 1964, Little
et al. 1974, and Tomlinson 1980) combined with field

observations. When dispersal agents were not known, it

34

was assumed that seeds surrounded by fleshy fruits were
animal-dispersed and winged seeds were wind-dispersed, and

all others were gravity-dispersed.
Mycorrhizae

The effect of age of site on VA mycorrhizal quantities
was investigated in August 1982. Root samples from 5
common dry forest species were collected from 3
different-aged sites: mature forest, the Cut Site, and
one-year-old ”regeneration plots" established by P.

Murphy and A. Lugo 1 km east of the Cut Site (Murphy and
Lugo 1983). The 5 species were Thouinia portoricensis,
Exostema caribea, Croton humilis, Gymnanthes lucida, and
Coccoloba diverslfolia (all canopy trees except Croton,
which is a shrub).

Field collections and laboratory procedures were
identical for the 2 sampling periods. Root samples were
collected from surface soils to a depth of 5 cm. The
entire root systems of seedlings and herbs were collected;
for larger woody plants, enough of the finest roots

3 vial.

available per plant were collected to fill a 25 cm
Samples were collected from 3 individuals per species when
possible, stored in distilled water, and refrigerated
until laboratory procedures were begun (within 3 weeks of

the collection date).

35

Laboratory methods for staining mycorrhizae were
modified from Philips and Hayman (1970). Root samples
were placed in test tubes with 10 percent KOH and heated
in a water bath at 85 C for 90 minutes. The samples were
then washed in distilled water to remove KOH and acidified
in 0.1 M HCl for one hour. Roots were stained in 0.1
percent acid fuchsin in lactophenol for 2 days and
destained in clear lactophenol (2 changes over a 24 hour
period), in which they were stored. Ten 1 cm root
segments per plant were mounted on a glass slide and
examined for infection. The samples were rated with a
score of 0 to 4, defined as follows: (0) no evidence of
infection, (1) light infection with occasional, isolated
infection points, (2) a few hyphae spread throughout, or
less than 1/3 of root with clumped hyphae, (3) continuous
hyphal strands throughout without clumping, or clumped
hyphae covering from 1/3 to 2/3 of the root, and (4)
heavily infected, with clumped hyphae throughout the root.
Vesicles and/or arbuscules were considered to be part of
the hyphae for quantification purposes, and their presence

was noted for each sample in which they occurred.

37

 

Figure 7. Appearance of Cut Site vegetation in 1982, 13 years
following cutting.

38

 

Figure 8. Appearance of adjacent forest vegetation in 1982.

RESULTS

Site Description

The thirteen-year-old site had a markedly different
appearance from the adjacent forest in 1981 (Figures 7 and
8). Prominent features in the Cut site included emergent
BUCida buceras trees up to 12 m in height scattered
throughout the area and a high density of the common
tropical weedy legume species, Leucaena leucocephala.

Also immediately obvious were the tangle of vines covering
much of the site and a predominance of grasses in much of
the understory, with Panicum maximum the tallest and most
prominent.

Cursory measurements of relative humidity and air
temperature in the site and in adjacent forest
(unpublished data) indicated that microenvironments were
similar in the 2 areas on at least the 4 days sampled.
Soil nutrient contents, pH, organic matter content, and
bulk density in the Cut Site were comparable to values of
those parameters for nearby forest (Murphy and Lugo,

unpublished data).

36

Vegetation Structure and Composition
Species Diversity

Species diversity was 36% higher in the Cut Site (76
species/80 m2) than in forest (49 species/80 m2), with 39
species common to the two sets of plots. The value of
Sorenson's coefficient of community (Brower and Zar 1977)
of 62.4% demonstrates a rather high similarity between the
two sites. When plots in the mature forest were randomly
divided into two groups, the coefficient of community
between the two groups was 76.1%, just 14% higher than
that between the Cut Site and forest.

The higher diversity of the younger site is largely
due to the presence of species classified as pioneer, many
of which were herbs or vines; 28% of Cut Site species were
pioneer species, compared to only 4% in forest plots
(Table A1). Pioneer species comprised 46% of the density
of stems in the Cut Site and only 3% of those in the
forest. The slow continual ascendance of species area
curves for Cut Site and forest plots (Figure 9) indicates
additional species occur in the sites. I recorded 10
species in the Cut Site and 13 in adjacent forest that did
not fall into sampling plots.

Species diversity of stems 3 1 cm DBH (including
trees, shrubs, cacti and vines) was similar in the two

2

sites (36 and 39 species/400 m in Cut Site and forest,

39

Cumulative Number of Species

80

7O

6O

50

40

3O

20

10

40

 

 

5 10 15 20 25

Number of 2 x 2 m plots

Figure 9. Species area curves for Cut Site and forest
2 x 2 m sampling plots.

Cut Site
i.
t ....
Forest
1 1 L 1 l 4
30

[,1

respectively), although species composition differed
somewhat (Sorenson's coefficient of 66.7%) (Table 1). Of
the 41 tree and shrub species in forest plots, 59% also
occurred in Cut Site plots. Several of the dominant
adjacent forest tree species also were dominants in the
Cut Site. Most notably, BUCida buceras, a fast-growing,
copious sprouter, was important due to its high relative
dominance in both sites, with percent of total importance
values of 20 (Cut Site) and 9 (forest). Exostema
caribaeum and Thouinia portoricensis were also among the
10 most important species in both sites (Table 1). A
major difference between the two sites is the relative
importance of Gymnanthes lucida, which comprises 11% of
the total importance value in the forest but less than 1%

in the Cut Site.
Structure

Relationships between structural variables in the two
sites are summarized in Table 2 and Figure 10. Stem
density in the Cut Site was nearly 3 times that in
adjacent forest; herbaceous, woody, and vine stems all
were significantly (p<.001, ANOVA) denser in the younger
site. Epiphytes were the only life form found in lesser
numbers in the Cut Site, although differences were not
significant due to a high variance between forest plots

(Table 2). Especially notable were the high Cut Site

Table l.

42

Site and adjacent forest plots .

plots.

Species

Gymnanthes lucida
Bucida buceras
Coccoloba diversifolia
Exostema caribaeum
Pictetia aculeata
Thouinia portoricensis
Amyris elemifera
Cephalocereus royenii
Eugenia foetida

Bursera simaruba

Samyda dodocandra
Pithecellobium unguis-cati
Guettarda krugii
Eugenia rhombea

Pisonia albida
Krugiodendron ferreum
Savia sessiliflora
Coccoloba microstachya
Coccoloba krugii
Leptocereus quadricostasus
*Helicteres jamaicensis
Guaicum sanctum
Tabebuia heterophylla
Jatropha hernandiifolia
*Sabinea florida
Antirhea acutata
Bumelia obovata

Cassine xylocarpa
Randia aculeata

Croton rigidus
Bernardia dichotoma
Colubrina arborescens
Guettarda elliptica
*Leucaena leucocephala
Plumeria alba

*Lantana involucrata
Unknown trees (5 species)
Unknown vines (2 species)
Reynosia guama

Clusia rosea
*Unknown shrub
Comocladia dodonea
Hedelia lanceolata
Unknown vine sp.

IData are from 2 x l0 m plots in Cut Site

(N-30).

FOREST

8 of
Total

.80
.l2
.l8
.SA
.l8
.3h
-59
.A9
.Ol
.08
.30
.46
.ll
.98
.89
-79
.70
.58
.36
.l6
.05
.99
.95
.83
-79
.65
.72
.52
.h6
.h0
.39
.2h
.21
.20
.20
.16
~55
.87

l

odooooooooooooooo—-_---—-—o---~ws-\n\n\na-l~loo\po

Rank

OWQNO‘WPUN—l

ll
12
)3
l4
15
l6
l7
l8
‘9
20
2l
22
23
25
25
26
27
28
29
30
BI
32
33
3h
35
36

37
38

IV

2.12
58-93
l.8k
29.2h
NP
l5.80
3-59
NP
2.02
8.k7
9.08
20.h2
b.5l
NP
20.89
l.k8
l2.65
ll.h0
NP
NP
2h.l7
NP
NP
NP
l0.6h
l.k8
NP
l.h3
5.06
5.10
NP
NP
NP
6.1h
l.h5
ll.00
NP
NP
l-SS
l.k8
l.k5
l.kh
l.hk
1.53

Importance values for plant species with DBH 3 l cm in Cut
NP - not present in sampling
* - species classified as pioneer.

CUT SITE

2 of
Total

0.71
19.6h
0.6l
9-75

5-27
1.20

.67
.82
.03
.8l
.50

~O‘WNO

6.96
0.k9
h.22
3.80

8.06

(Nclo) and adjacent forest

Rank
18

20

17

)9
12
ll

l6

2h

l0
22

29
15
lb

13
25

2)
23
26

27
28

30

43

Table 2. tructure of Cut Site and adjacent forest
vegetation . A11 stems included in analyses. Standard
errors are in parentheses.

Parameter Cut Site Forest p2
Species per 9 m2 17.0 (1.2) 9.0 (0.8) <.001
Plant height (m) 0.7 (0.1) 1.6 (0.1) <.001
Basal area (m2/ha) 19.7 (5.0) 52.8 (8.5) <.01
Stem density/m2 of:

Total stems 26.h (3.7) 9.8 (1.6) <.001

Pioneer species 12.7 (2.7) 0.3 (0.2) <.001

Woody 19.2 (3.0) 6.5 (0.8) <.001

Herbaceous 6.9 (1.6) 0.5 (0.2) <.001

Cacti 0.0h (0.02) 0.0h (0.02) NS

Epiphytes 0.3 (0.1) 2.8 (1.3) NS

Vines 5.7 (1.1) 0.8 (0.2) <.001
Mean crown area 2115.9 201.5 <.05
of vines (m2/ha) (5h7.6) (89.1)

1Data are from 2 x 2 m plots in Cut Site (N-20) and forest
(N‘ZO).

Significance in ANOVA. NS - not significant at p<.1.

Table 3. Leaf area index and percent cover in the Cut
Site. LAI is the Qeanznumber of foliar intersections per
vertical rope in m / m . Percent cover is the number of
times the rope encountered vegetation divided by the total
number of counts. (N-120).

Date Leaf Area Index Percent Cover
September 2-h, 1982 1.8 100.0
March 23-2h. 1983 2.2 99.2
May 30, 1983 3.9 100.0

26.4

17

\ \ V”
\\\l‘°

A4

 

/ ADJACENT FOREST

' CUT SITE

 

 

 

N
(D
H
01
01
N
on

19.7

\ \ \L‘”

/

\\1\l
\\\J

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

/ 1
Stem Species Density of LAI Mean Basal
Density Richness Stems 2 1 cm Height Area
(stems/m2) (species/ (stems/m2) (mZ/mz) (m) 2
4 m ) _ (m /ha)

Figure 10. Comparison of structure in Cut Site and adjacent forest.

45

densities of herbs and vines (14 and 7 times the densities
in adjacent forest, respectively). Vine cover was 10
times higher in the younger site.

Structural parameters which were significantly smaller
in the Cut Site included basal area (determined from
diameter at ground level measurements) and mean plant
height; both were 40% of adjacent forest values (Table 2).

Horizontal cover was 100% in the Cut Site (Table 3),
compared to values of 77 to 93% in nearby mature forest
(Murphy and Lugo 1983). Leaf area index in the Cut Site
(Table 3) was 57 to 65% of that measured by Murphy and
Lugo (1983) in mature forest. Leaf litter biomass did not
differ significantly between the Cut Site and adjacent
forest (p>.l, ANOVA) (Figure 11).

The stratification of plant heights in height interval
classes demonstrates that lower mean heights in the Cut
Site are due to the presence of more seedlings and fewer
tall shrubs and trees in the younger site (Figure 12).

The smallest height class (1-25 cm) made up 52% of Cut
Site woody plants and 30% of those in forest. Densities
in intermediate height classes were similar in the two
sites, while plants in height classes > 2 m contributed
larger percentages in forest plots. The vertical
stratification of vegetation area index (Figure 13)
demonstrates the concentration of plant tissues close to

the ground: 84% of VAI counts occurred below a height of l

—_

Mean Dry Weight of Leaf Litter (g/m2 + SE)

46

 

 

 

 

 

 

 

 

 

 

 

 

 

 

14001 1 I
1200
1000 . I I

. l
800 ‘
600 .

Forest Cut Site Forest Cut Site
SEPTEMBER MARCH

Figure 11. Leaf litter biomass in forest and
Cut Site in September and March.

NUMBER OF PLANTS PER 64 m2

337

80

70

60

50

40

30

20

 

 

47

[:3 CUT SITE

ADJACENT FOREST

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

0-25 25-50 50-100 100-150 150-200 200-500 500-1000

HEIGHT OF PLANT (cm)

Figure 12. Numbers of plants in height interval classes in the Cut
Site and forest plots. Data are from 2 x 2 m plots.

 

HEIGHT FROM GROUND (m)

48

5.0

4.5

3.5

 

 

 

3.0

1

2.0

 

 

 

1.5 ‘ l
1.0 f 1

- |

 

 

 

O V T U I 7

O 5 10 15 2O 25

PERCENT OF VEGETATION AREA INDEX READINGS

Figure 13. Vertical stratification of Cut Site vegetation using
vegetation area index readings.

49

m, with the remaining 16% between 1 and 5 m heights.

The Cut Site supported fewer plants with stems 3 1 cm
DBH than adjacent forest (8600/ha versus 10,452/ha). Cut
Site trees supported more stems/individual, however; this
was reflected in the comparative stem densities of
(27,000/ha versus 25,155/ha in forest). Many of the stems
in the younger site were small in diameter. When
distributed according to size class, 86% fell into the
smallest category (l-2.5 cm DBH) compared to 51% in that
category in adjacent forest (Figure 14). In all the
larger size classes, forest plots possessed a higher mean
number and higher percentage of stems than Cut Site plots.
Basal area of stems 3 1 cm DBH in the Cut Site (8.88
mZ/ha) was 29% of that in adjacent forest (30.75 mz/ha)
(Table 4). The two values differed significantly (ANOVA,
p<.001). Fewer than 1% of the total number of stems in
the Cut Site were dead, compared to 9.7% mortality in
forest plots (Table 4). An examination of composite
differences between the 2 sites using 4 variables measured
for stems 3 1 cm DBH revealed that differences were highly
significant (Hotelling T test, p<.001) (Table A4).

Summaries of density and basal area of stems 3 5 cm
DBH in the Cut Site and adjacent forest were compared to
pre-cut measurements made by Ewel on the one ha site in
1969 (consisting of diameter measurements of trees 3 5 cm

2

DBH in three 1000 m transects) (Table 4). The adjacent

50

Table A. Structure of woody vegetation in Cut Site in
adjacent forest , and on Cut Site before it was cut .
Values are means per ha. (*** - significantly different
from adjacent forest value in ANOVA at p<.001)3.

"Cut Site”
Parameter Cut Site Forest in 1969

Live stems > 1 cm DBH

Plant density 8600 10h52

Stem density 27000 25155

Basal area (m2) 8.877*** 30.751
Dead stems > 1 cm DBH

Stem density 150 2717
Live stems > 5 cm DBH

Stem density h00 3h33 1127

Basal area (m2) 1.237*** 18.574 b.950
Dead stems > 5 cm DBH

Stem density 0 283

IData are from 2 x 10 m plots in Cut Site (N-10) and in
2adjacent forest (N-30).

Data are summarized from data collected by J. Ewel in
1969 on the one ha site before it was cut.
3Where there are no asterisks, values did not differ
significantly between sites at p<.1.

1 cm DBH PER HECTARE

Z

DENSITY OF STEMS

51

86% ; FOREST

CUT SITE

 

23,100

 

 

 

 

UT
5.:
69

 

12,800 ‘

11,200 -

9600 _
35%

 

8000

6400 -

4800 -

13%

 

%
3200 q 12

 

1600 -

2% 2%

____1 0%

1.0-2.5 2.5-5.0 S.O-l0.0 >‘l0.0

 

 

 

 

 

\\\\\\\\\\\\

 

\\\\\\\\

 

 

\\\

 

 

 

 

 

DIAMETER CLASS (cm)

Figure 14. Size class distribution in the 2 x 10 m plots
in forest (N=30) and Cut Site (N=10).

52

forest in 1982 possessed stem densities and basal areas
which were 3.0 and 3.8 times higher, respectively, than
the pre-cut one ha site in 1969 (Table 4). This large
difference indicates the Cut Site probably began with less
structure than the adjacent forest. The Cut Site in 1982
supported only 400 stems 3 5 cm DBH/ha which contributed
1.24 mz/ha of basal area; these values were l/3 and 1/4 of

those values, respectively, for pre-cut vegetation.
Seasonal Variation

Comparisons of Cut Site structural parameters between
sampling periods revealed seasonal differences (although
not always significant ones) in leaf litter biomass, leaf
area index, and herbaceous density and cover (Figure 11
and Tables 3 and 5). Leaf litter biomass was
significantly (p<.05, ANOVA) higher in both sites in March
than in September (Figure 11).

Leaf area index values, summarized in Table 3, nearly
doubled in the interval between the March and May
measurements, reflecting the leafing out of many species
following the onset of the rainy season (Figure 2). LAI
in September was slightly lower than in March, probably as
a result of a brief dry spell in August 1982 that caused
many plants to wilt and/or lose their leaves (Murphy and
Lugo, unpublished rainfall data).

Herbs in both Cut Site and forest showed no

Table 5.

53

season (March) and wet season (August) .
Standard errors are in parentheses.

Cut Site
August

March

forest
August

March

Number of
species Density
2
1.0 (0.h) 23.5 (4.3)

5.52(0.5) 28.1 (5.2)

0.7 (0.2) 1.6 (0.9)

0.9 (0.3) 2.1 (1.0)

Comparison of herbs in Cut S'te and adjacent forest between dry
Values are means per plot.

Mean height Hean crown
(cm) area (cm2)

39-3 (7.7) “075.9 (1588.2)
26.3 (1.9) 3529.2 (807.4)

10.8 (3.7) 480.0 (202.2)
7.5 (2.1) 298.0 (144.3)

IData are from 2 x 2 m plots in Cut Site (N-Zh) and forest (N-19).
These two values differ significantly at p<.05 in ANOVA. The other
month-to-month comparisons do not differ significantly.

54

significant differences in density, mean height, or mean
crown area between the August and March sampling dates.
Mean number of herbaceous species in the Cut Site was
significantly higher in March (p<.05, ANOVA) than in
August, possibly reflecting seasonal patterns of more
species germinating in March. The trend of higher
densities coupled with lower mean heights (i.e. more
seedlings than taller plants) in both sites in March
supports this explanation. Mean herbaceous cover was

somewhat lower in both sites in March, the drier season.
Mechanisms of Regeneration

Proportions of sprouts and seedlings were analyzed
from the perspectives of both the number of plants and the
number of stems. The proportion of plants which were
seedlings yields information about the prevalence of seed
regeneration in the site. The examination of the number
of stems in each category allows one to analyze sproutiing
in such a way that decisions regarding how a particular
individual is defined (e. g. as a plant or as part of a
plant) are avoided. The latter method is especially
helpful when considering the case of root sprouts, since
without excavating root systems, it is impossible to
determine which plants root sprouts are connected to.

Seedlings constituted an important portion of plants

55

supporting stems 3 1 cm DBH in both sites. Fifty-one
percent of Cut Site plants and 68% of forest plants had
seedling origins (i.e. did not possess any sprouts)
(Figure 15).

From the standpoint of stems, 81% of Cut Site stems 3
1 cm DBH originated as sprouts from cut stumps, 3% as root
sprouts, and 16% as seedlings (Figure 16). The adjacent
forest possessed 49% stem sprouts, 23% root sprouts, and
28% seedling-origin stems. Stems 3 1 cm DBH in the
adjacent forest were primarily persistent species (98 to
99% of stem density) regardless of origin. In the Cut
Site, pioneer species comprised 32% of seedling origin
stems, 35% of root sprouts, and 15% of stem sprouts.
Every woody species occurring in the sampling plots in the
two sites produced vegetative sprouts, except the small
tree species Plumeria alba, which did produce sprouts in
another cut-over Guanica site (A. Murphy, pers. comm.).

A height class analysis of all woody stems in the Cut
Site plots demonstrates that stems originating as sprouts
were generally the taller plants in the site, while stems
of seedling origin were shorter (Figure 17). Of stems in
height classes 3 1.5 m, 54 to 100% were stem sprouts,
while in height classes < l m, 52 to 81% were of seedling
origin. Stems between 1 and 1.5 m heights were primarily
root sprouts (47%) or of seedling origin (41%). It

appears that either vegetative sprouts have faster height

56

51% SEEDLINGS

 

 

 

CUT
SITE
49% ROOT SPROUTS or
PLANTS WITH STEM SPROUTS
68% SEEDLINGS
ADJACENT
FOREST

32%
ROOT SPROUTS or

PLANTS WITH STEM SPROUTS

Figure 15. Percentages of woody plants with stems 2.1 cm
DBH with stems originating as vegetative sprouts and as
seedlings in Cut Site and forest plots.

57

3% ROOT SPROUTS

 

    
    
    

   

 

1.
16%
SEEDLINGS
81%
STEM SITE
SPROUTS
28%
ROOT SEEDLINGS
SPROUTS
ADJACENT
FOREST
49% STEM
SPROUTS

Figure 16» Percentages of woody stems 2. 1 cm DBH
originating as vegetative sprouts and as seedlings
in Cut Site and forest plots.

DENSITY OF WOODY PLANTS PER 56 m2

58

 

 

 

 

 

 

 

400 . .
1 V
375 q 3.}
;; CUT SITE
350 _ {f
325 - .. Stem Sprouts
) , ,L

  
    

Root Sprouts

 

75

 

 

 

Seedling-origin

50

25

100

75

50

25

 

Figure 17. Height class distribution of woody plants in
Cut Site and forest originating as vegetative sprouts and
as seedlings.

59

“growth rates than seedlings, or seedling regeneration was
at some point suppressed in the Cut Site, giving the
sprouts a temporal advantage.

An examination of the effect of distance from the
forest edge on seedling species and densities indicates
that seed dispersal has not been a limiting factor to
seedling success in the Cut Site. The 24 2 x 2 m plots
examined for an edge effect revealed no significant
differences (ANOVA, p > 0.5) in all parameters measured
(Table 6). There was a trend toward decreasing numbers of
woody seedling origin plants toward the center of the site
(Figure 18). Coefficients of determination ranged from
.002 to .115 (Table 6), indicating that distance from the
forest edge accounted for only 0 to 12% of variation in
plant density and species richness of herbs, woody plants,
persistent species densities, and pioneer species
densities.

The present-day seed bank in the Cut Site consists of
a broad array of species throughout the site, including
those classified both as pioneer and as persistent forest
species (Tables 7, 8, and 9). The combined densities of
species of seeds in September and March collections from
plots at varying distances from the forest edge
demonstrates that species were broadly distributed
throughout the sites (unpublished data).

Half of the 94 plant species identified in the Cut

60

Table 6. Effect of distance from forest edge on seedling species and
densities in Cut Site 2x2 m plots. N=4 plots per distance. Values are means
per plot.

Distance from forest edge (m)

2
1.4 10 20 30 40 48 *R **Sig.
Number of species
Total 12.0 10.8 13.0 8.8 11.5 9.3 .036 NS
Herbaceous 3.5 4.0 4.5 2.5 4.0 4.8 .010 NS
Woody 8.5 6.8 8.5 6.3 7.5 4.5 .103 NS
Density
Total 36.8 85.0 57.5 27.8 69.5 36.8 .008 NS
Herbaceous 8.8 38.5 22.0 5.0 43.0 24.3 .014 NS
Herb.
persistent sp. 4.8 5.5 5.3 1.3 16.5 9.5 .044 NS
Herb.
pioneer sp. 4.0 33.0 16.8 3.8 26.5 14.8 .002 NS
Woody 28.0 46.5 35.5 22.8 26.5 12.5 .115 NS
Woody
persistent sp. 15.0 18.3 20.3 6.0 9.5 5.3 .101 NS
Woody
pioneer sp. 13.0 28.3 15.3 16.8 17.0 7.3 .052 NS

*Coefficient of determination, calculated from regression.
**Significance in ANOVA. NS - not significant at p < .1.

MEAN DENSITY 0F WOODY SEEDLING ORIGIN PLANTS PER 4 m2

61

50

40 a

30 .

20 .

1O .

 

 

 

1
1

1.4 10 20 30 4O 48

DISTANCE FROM FOREST EDGE (m)

Figure 18. Relationship between numbers of woody seedling origin
plants in Cut Site and distance to the forest edge.

62

Table 7. Seed densities in soil and litter samples (N-10 .25 1112 samples per
site per sampling date). Pt-persistent species seeds. Pi-pioneer species
seeds. Total is the sum of Pt + Pi + unidentified species. Percent of total
for each fraction is in parentheses.

Seeds 2 2.38 mm Seeds < 2.38 mm

# seeds/m2 I seeds tested *percent ** I seeds
for viability viability germinated/m2

September collection

Cut Site
Pt 500.4 (64.0) 64 1.6 4.3 (8.8)
Pi 248.0 (31.7) 51 37.3 40.8 (83.3)
TOTAL 782.4 115 17.4 49.0
Forest
Pt 1249.6 (80.6) 168 1.2 3.8 (40.0)
Pi 208.8 (13.5) 32 62.5 3.9 (41.1)
TOTAL 1550.4 228 11.0 9.5
March collection
Cut Site
Pt 738.4 (58.3) 56 7.1 4.2 (3.3)
Pi 500.8 (39.6) 39 66.7 64.0 (49.9)
TOTAL 1265.6 95 17.4 128.3
Forest
Pt 1537.6 (88.8) 74 0 0
Pi 127.2 (7.3) 21 42.9 8.8 (16.2)
TOTAL 1732.4 95 9.5 54.2

*Viability tests were performed on May 21. 1983 for samples collected on
September 1, 1982, and on June 1. 1983 for samples collected on March 24.
Tests were based on non-randomly selected small samples, thus should be.

considered estimates.

**Seed germination was measured from soil and litter samples sprinkled over
sterilized soil and placed in a greenhouse. Germination experiments were
initiated 6 months following the September collection date and 2 months
following the March collection date, with samples stored at room temperature
in the interims.

63

Table 8. Densities of seed species < 2.38 mm germinating
from soil agd litter samples under greenhouse conditions
(N810 .25 m samples per site per sampling date).

Germinated seeds per m2

*September March

Species Cut Site Forest Cut Site Forest
Commelina erecta 0.7 0 0 0
Croton humilis 0 0 2.0 0
Echinochloa colonum 0 0 O 0.4
Eupatorium sinuatum 0 0 0.4 0
Galactia dubia 0 1.9 0 0
Helicteres jamaicensis 0 1.4 2.2 0
Jacquemontia pentantha 20.5 0 31.1 2.1
Lasiasis divaricata 0 1.9 0 0
Scleria lithosperma 3.6 0 2.2 0
Setaria setosa 9.3 0.6 6.7 0
Sida acuta 8.7 1.9 1.9 3.9
Spermacoce tenuoir 0 O 16.5 1.2
Stachytarpheta strigosa 0 O 3.6 0
Wedelia lanceolata 0 0 1.6 1.2
Unidentified 3.9 1.8 60.1 45.4
TOTAL 49.0 9.5 128.3 54.2

*In September, samples were from soil collections only
(not litter).

64

Table 9. Densities and species of seeds 3 2.38 mm collected from soil and
litter samples. N-lO 0.25 m samples per sampling date. Numbers in
parentheses are densities of viable seeds.*

September March

Species Cut Site Forest Cut Site Forest
Bucida buceras 358.0 (19.9) 210.4 (199.1) 284.0 (0) 708.8 (0)
Bursera simaruba 0 19.2 (0) 2.0 (0) 506.8 (0)
Coccoloba diversifolia 20.8 (0) 347.6 (0) 0 88.8 (0)
Coccoloba krugii 2.4 302.8 (0) 4.0 (0) 1.2
Coccoloba microstachya ‘ 0.8 214.0 329.6 (0) 27.2 (0)
Croton discolor 6.0 (0) 0 8.4 (O) 0.4 (0)
Croton humilis 32.4 4.8 0 0
Croton sp. 0 0 1.6 5.2
Erithalis fruticosa 0.4 0.4 (0.4) 0 0
Eugenia foetida 2.8 (0) 2.8 0 77.6
Exostema caribaea 1.2 31.2 (0) 4.0 17.6
Galactia striata 3.6 (3.6) 0.8 0 0
Gouania lupuloides 19.2 0 358.4 (150.9) 99.6 (8.3)
Guettarda elliptica 0.4 0 0.4 0
Guettarda krugii 2.0 (0) 8.8 0.4 2.4 (O)
Helicteres jamaicensis 69.2 (48.4) 0 0 0
Jacquinia barteri 0.8 (0) 1.2 (0) 0 0
Jacquemontia pentantha 121.6 (9.4) 11.2 13.2 2.4
Kramerii ixinia 0 0 0.4 0
Lantana involucrata 0.4 0 0.4 O
Lasiasis divaricata 5.2 3.2 92.8 (21.4) 22.0
Leucaena leucocephala 34.0 (20.4) 196.0 (123.5) 128.8 (115.9) 23.2 (20.6)
Passiflora sp. 0 0 0.4 0.4
Pisonia albida 1.6 (0) 0 0 0
Pithecellobium

unguis-cati 0 0.4 0 O

Reynosia sp. 0 1.6 O 0.4
Thouinia portoricensis 30.4 (0) 18.0 (0) 3.2 10.4 (0)
Waltheria indica 31.2 (0) 0 0 0
Wedelia lanceolata 0.4 0.8 0 2.0
Unknown species 34.0 92.0 26.4 67.6
TOTAL 782.4 1550.4 1265.6 1732.4

*Viability values are based on non-randomly selected small samples of seeds.
thus should be considered estimates. Where no numbers in parentheses are
given, viability tests were not done.

65

Site and forest occurred in seed bank samples, including
fifteen seed speCies in the < 2.38 mm size class, and 30
in the 3 2.38 mm group (Tables 8 and 9). Forty-nine
species either did not show up in the seed bank samples or
remained unidentified.

When seed densities were averaged over sampling
periods, the younger site contained 3.7 times more small
seeds (< 2.38 mm) per square m and 2.4 times more seeds of
pioneer species than adjacent forest. The forest
supported 1.6 times more large seeds ( 3 2.38 mm) and 2.3
times more seeds of persistent species (Table 7). Mean
seed densities in litter and soil samples averaged over
both collecting periods were 1112.5/m2 in Cut Site and
1650.9/m2 in forest.

Two-thirds of the 40 identified woody species in the
2x10 m plots showed adaptations for vector-assisted seed
dispersal. Fifty-three percent are dispersed by animal
consumption, particularly by birds; ten percent are wind
dispersed; 2% can stick onto moving vectors; and 35% show
no adaptations for vector dispersal.

Per cent viability of seeds from soil and litter
samples (per site, per sampling period) of pioneer species
was generally fairly high (21-64%) while that of
persistent species was low (0-7%) (Table 7). It is likely
that these figures would have been somewhat higher if

viability had been measured immediately after the seeds

66

were collected rather than 2 to 9 months later.

Viability of fresh seeds of 11 species collected in
September 1982 and March 1983 from trees and shrubs was
relatively high (12-100%/species), and germination rates
very low (Table 10). It is possible that scarification of
seeds or some dormancy-breaking environmental effect is
needed before some dry forest species of seeds can
germinate. Scarification of Leucaena seeds collected in
March 1983 by soaking in sulphuric acid (to simulate
passage through a bird's digestive system) increased

percent germination by as much as 80% (Table 11).
Mycorrhizae

A comparison of 5 tree species in the Cut Site,
forest, and in one-year-old "regeneration plots" yielded
no appreciable differences in degree of mycorrhizal
infection of tree roots in September 1982 in the
different-aged sites (Table 12). Standard errors were
rather high, indicating that larger sample sizes and more
precise methods of quantification may be in order to
elucidate small-scale patterns. Arbuscules were observed
in 4 out of the 5 species, indicating that the mycorrhizae
were functioning in their role of nutrient transfer in at

least those 4 species in September.

o o w.__
5 com 0.00—
o oo. n.3—
o o n.0m
o oow o.ow
o oo. o
o o o.mm
o oo~ o
o oo. o
o o 0.4»
o oo. o

7

,o pounc_etou couco_o »u__mnm_>

mooom § mooom x u

N—
0—
am—

su,__nn_>
toe uoumou
mooom x

um3u3< c_ peace—.00 muoom smote Co xu__wnmc_etou

~o_uc_

uumtuo_o>c_

m_tocu_mz
«.mzoooouoo_ ncomuooa

«canon;

mom_to o_ucmmm_to:
m_mcou_n5om «vacuum—u:
__motx notnuuooo
moumut_> mozuanmoo
«coomoLonto nc_Lna_ou
o__oe_mco>_o «no—ouoou
A.om crocxcov mouoau
motooon oo_uom

mo_ooom

ecu su,__na_>

.o—

.~mm_
o_noe

68

4.5 acoEumoLu oz

o.m Amouac_e NV Loam: uc___om

4.5m Amuu:c_e o_v o_un o_L::o_:m

m.nm Amouocwe 4v u_uo 0_Lo:o_:m
co_uac_ELou » unusuuotuota

.Aucosumotu Loo «ovum malzv co_umcmecou

ovum -u§amuouam~ nemuuauu co mucoEuoocu comuoo_e_tnum eo mucouum

.—— o_nmh

69

.Auxou oomv
no.3 Hoot mo n\_v to uoocoooccu unocom gone»; 30. ANV

.uoo. unocuooczu omen»: cones—u Agv

.oaznx: cones—o
.muc_oo co_uuoec_

.oco_mnuoo A_V .co_uoouc_ oc on 5a.: .a1o uo o_mua a co oomnn 0L0: muc_um¢«

3.9 II II
TN ....n ...~ ..m
3.8
s.~ m.. ..m ..~
.m.ov 111.
..m Tm mg ....m
...:
~.. ~.~ o m..
3.3 II
a.. 4.. m.~ 3..
n N .
can: * ace—m
umoaou

.momocucotmo c. «an meocao oauocmum

com:

o.. m.o o.~
..o ... mqw
... m.. m.m
5.. m.~ o
mqa ~.. o.~
n N .
* ac~.a
a..m ago

0L0: no.30m3ntm 50.53 c. mo_nE~m .0 menu:

.auu_m ooom1ucotoum_o m c. mu_uooa Co moc_uac .nu_cccouxee coo:

A~.ov m_mcoo_toucoo

m.~ wqm. m.~ w.~

3.3 II
o.n ~.~ m.~ ~.n
A~.ov.111 .111
m.~ n.~ 4.~ m.~
3.8

~.m m.m m.n o.m
3.8 l
m._ 1 n.— m.—

m N _

coo: * ecu—a

moo—n o_o1.moxuoco

~_c.:oge

Eoono_tmo

aeoumoxu

....eag

couotu

a..o...tu>.u

moo—oooou

~u.ua.

mozucmcExu

no.00om

.ooc__tooco can ou>tomno
.Aucm_o Loo mucoEuom uooc Eu _ o_uzv

.NH 0_nmk

Effects of Ewel's Recut/Herbicide Treatments

Vegetative recovery was consistently less on the
"recut/herbicided" portion of the Cut Site than on the
"cut only” portion (Table 13). Species richness, stem
density, woody plant density, percent of pioneer species
stems, and basal area were higher in the ”cut only"
portion of the site, the latter significantly so in the
smaller plots. There were nearly three times more
epiphytes in the "cut only” part of the site in both sets
of plots. While the total density of trees and the
density of trees supporting some stem sprouts were similar
in the two portions of the site, the number of stem
sprouts in the “cut only" area was almost twice that in
the "recut/herbicided" area (mean values significantly
different at p < .05). This indicates that trees which
were repeatedly cut or treated with herbicides responded
primarily by producing fewer vegetative sprouts.
Hotellings T statistic calculated with mean values of
species number, plant density, stem density, and DBH,
showed a significant composite difference between the two

sites at p < .1 (Table A4).

70

71
Table 13. Structure of vegetation in the portion of the Cut

Site which had recut/herbicide treatments in 19 9 and the
portion which was cut only. Values are means/m except where
noted. Standard errors in parentheses.

Parameter Cut Only Recut 8 Significance
Herbicided in ANOVA
*All stems included
Species/4 m2 4.3 (0.3) 3.6 (0.3) NS
Basal area 19.7 (5.0) 9.0 (1.6) p<.05
(cm2)
Height per plant 73.4 (6.0) 77.2 (13.1) NS
(cm)
Stem density:
Total 26.4 (3.7) 24.6 (4.8) NS
Woody 19.1 (3.0) 14.8 (1.3) NS
Herbaceous 6.9 (1.6) 4.7 (1.2) NS
Cacti .04 (.02) 5.0 (3.7) NS
Epiphytes 0.3 (0.1) 0.1 (0.1) NS
Vines ’ 5.7 5.1 NS
Pioneer 12.7 (2.7) 10.7 (1.9) NS
species
Persistent 13.7 (2.2) 13.9 (4.9) NS
species

**Woody stems > 1 cm DBH only

Number of 0.4 (.04) 0.3 (.04) p<.1
species
Number of 0.9 (0.1) 0.8 (0.2) NS
plants
Plants with 0.5 (.04) 0.5 (.07) NS
stem sprouts
Basal area 8.9 (1.7) 5.3 (1.8) NS
(cm2)
Stem density:
Total 2.7 (0.3) 1.6 (1.1) p<.05
Stem sprouts 2.2 (0.3) 1.2 (0.3) p<.05
Root sprouts .08 (.03) .03 (.02) NS
Seedling origin 0.4 (0.1) 0.4 (0.1) NS

* Data are from 2x2 m plots (N-14 in Cut Only portion,
and N-16 in Cut and Disturbed portion).

**Data are from 2x10 m plots (N-lO in each portion of
site).

DISCUSSION
Temporal patterns of succession

Margalef's (1963) view that species diversity peaks in
early to mid successional stages is supported by this
study (Figure 19). There were 31.6 vascular species per
18 m2 in control plots one year following cutting (data
summarized from Ewel 1971): this value increased to 39.7
per 16 m2 12 years later, compared to 24.0 per 16 m2 in
mature forest. The peak in diversity in the l3-year-old
site closely matches the pattern in temperate forests, in
which diversity increases rapidly in the earliest stage of
succession, then drops gradually after year 10 or 15
(Bormann and Likens 1979, Hibbs 1983).

The higher mean species richness in the younger site
is at least partially due to the greater density of stems.
When computed as number of species per 100 stems, the
mature forest has the highest value, with a nearly linear
decrease with decreasing age of site (Figure 19).

The importance of pioneer species in the
thirteen-year-old site is especially notable in light of
generalizations made by some earlier authors regarding
their relative unimportance in dry tropical forests. Ewel

(1980) stated that succession in the drier tropics is

floristically simpler and involves fewer seral stages than

72

MEAN NUMBER OF SPECIES PER 16 M2

 

 

73

 

 

 

45 -
SPECIES/AREA /
40 .1 / /
35 "
30 4
25 '-
/ x
O
/ SPECIES/100 STEMS
/
20 _ /
15
0 25 50
AGE OF SITE (YEARS)
Figure 19. Species diversity changes relative to succession

in the Guanica forest.

SWELS 001' 213d SHIOHdS .110 HEHNHN

74

in wetter sites. Loveless and Asprey (1957) noted that
there was little or no invasion of second-growth species
in disturbed Jamaican dry evergreen bushland owing to a
lack of sufficient soil for seedling colonization. While
it is undoubtedly true that wetter areas support more
seedling regeneration and greater successional species
diversity, there does exist a distinct dry site
successional flora consisting of both pioneer and
persistent species. Opler et a1. (1977) reported that
forbs and grasses followed by vines, shrubs, and small
trees dominated the first 3 years of succession on cut,
burned, and grazed Costa Rican dry forest. In the Guanica
study site, the proximity of roadsides and fields makes
pioneer species seed sources readily available. The
rather large size of the opening together with the
recut/herbicide treatments applied to a portion of the
site just after it was out probably created ideal
conditions for pioneer species to colonize and
proliferate.

Perhaps the most obvious and surprising aspect of the
Cut Site's species composition is the relative
unimportance of Gymnanthes, which was dominant on the site
prior to cutting and is important today in mature forest
adjacent to the site. It was also listed by Ewel as a
dominant in the first year of recovery. It can reproduce

by stem sprouting, root Sprouting, or seedling

75

germination, though sprouting is not common. The species
was not well represented in seed banks in the site and was
the only species in which an insect larva was found in
seeds from seed banks, so perhaps a species-specific seed
predator has limited seed survival. There has apparently
been plant mortality since Ewel's observations, indicating
the species may be particularly intolerant of vine
infestation, or requires a dense canopy cover to survive
for more than one year. Further study on this matter is
needed.

Patterns of temporal change in dry forest structural
parameters following clearcutting vary for different
vegetation components. Summaries of Ewel's data on
one-year-old regrowth on the Cut Site (from Ewel 1971),
this study's results for both Cut Site and forest, and
measurements made by Murphy and Lugo (1985) in mature
forest were combined to depict these patterns in Figure.
20.

All parameters increase sharply in the first year of
growth, demonstrating an immediate spurt of productivity
largely due to coppicing. Between ages 1 and 13 years,
only stem density and leaf area index decreased. The stem
mortality in this stage is representative of the thinning
stage in forest succession which follows the establishment
phase; it generally lasts until a subSequent disturbance

or natural thinning makes room for additional

76

establishment (Peet and Christensen 1980). Density
continues to decline in dry forest after year 13, with
most of the mortality in herbs, vines, and smaller woody
stems. The thinning stage in post-agricultural moist
evergreen forest begins after year 7 (Aweto 1981).

The decline in LAI in the mid-successional site is
unexplained, though Marks and Bormann (1972) reported a
similar pattern in temperate deciduous forest, with
l4-year-old sites exhibiting lower LAI than very young
sites. It is possible that the decline in the dry forest
simply reflects year-to-year variabiliy in timing of
deciduousness resulting from varying rainfall patterns
rather than a successional trend.

Horizontal cover in the 13-year-old site (100%)
represents a peak between the 90% cover value 1 year
following cutting and the 78% cover in mature forest.
There was a shift from the initial patchy multilayered
structure in the 1-year-old site (Ewel 1977) to a more
horizontally homogeneous structure consisting of a
profusion of vines, shrubs, coppiced trees, and spreading
herbs in the lB-year-old site.

Biomass continued to increase from year one to
thirteen (Murphy and Lugo 1985), and mature forest values
are higher still. This pattern is similar to that shown
for temperate forest biomass recovery (Kira and Shidei

1967, Woodwell 1967, Odum 1969), except a mid-successional

77

peak in biomass generally occurs and is followed by a
slight drop and subsequent stabilization in mature forest.
A similar peak may occur during dry forest recovery
sometime after age 13.

Mean plant height parallelled the successional pattern
of biomass. The tallest trees in the Cut Site and the
adjacent forest were emergent BUCida buceras individuals,
reaching maximum heights of about 12 m in both sites.

This was close to the canopy height mentioned by Wadsworth
(1950) as representing primary forest (12-18 m).

Vines are known to be an important component of
tropical dry forests, with densities of 24 to 81 lianas
with stems 3 2.5 cm per 1000 m2 occurring in 5 different
dry forest sites (Gentry 1983). The predominance of vines
in the Cut Site (22% of total stem density) undoubtedly
plays an important role in plant interactions on the site.
Putz (1980) stated that lianas generally interfere with
trees by competing for light, nutrients, and water and by
causing mechanical damage. Several vine species in
temperate forests have been shown to cause damage and even
death of trees by wrapping so tightly around stems that
the flow of assimilates in the phloem is inhibited (Lutz
1943, Trimble and Tryon 1974, Siccama et al. 1976).

There were several instances when I noted damage or death
of young trees in the Cut Site with abundant vines wrapped

around them. Vines may also have influenced tree species

78

composition in the Cut Site due to differing capacities of
trees to avoid them. Characteristics making trees less
susceptible to liana infestation include fast growth,
angles of sway differing from neighboring tree species to
prevent intercrown passage, and regular shedding of
branches and leaves (Putz 1980).

The very low density of epiphytes in the Cut Site is
typical of tropical early to mid-successional sequences
(Budowski 1963, 1965). Tiny seedlings of Tillandsfa
Pecurvata (too small to be included in analyses) apppeared
commonly in Cut Site trees, but reproductive-size plants
were quite rare. Benzing (1980) wrote that young forest
communities are relatively free of epiphytes because the
rapid growth and short life of many pioneer taxa render
the immature canopy too unstable for vascular epiphytes.
Davis and Richards (1933) suggested that a successional
sequence of epiphytes which varies with the age of the
host tree may exist, and noted that in rain forests,
pioneers are often of the Bromeliaceae ( Tillandsia spp),
the Araceae, and ferns, while orchids belong to later
stages. Observations in the Guanica forest indicate that
while Tillandsia recurvata and T. utriculata survive at
low densities and small sizes in successional forest, all
epiphyte species, including orchids and bromeliads, are
primarily mature forest species. Epiphytes may be an

accurate and easily quantifiable indicator of degree of

79

disturbance in dry forest. Their strong reliance on a
stable support system may make them especially sensitive
to the degree of stability, and therefore maturity, of a

community.
Mechanisms of Regeneration

Vegetative sprouting is an extremely important
regenerative mechanism in cut-over dry forest. All tree
and shrub species sampled in this study, including both
pioneeer and persistent species, can survive disturbance
by producing either root sprouts or stem sprouts. The
height dominance by coppiced plants indicates that
coppicing gives a competitive advantage to sprouted stems
over seedlings in forest recovery. The well-developed
root systems, complete with mycorrhizae, of cut trees
undoubtedly play an important role in this competitive
superiority both by their ability to reach and take up
water and by serving as a source of stored energy for new
Sprouts. Ewel (1977) noted in a visit to the Cut Site 4
years following cutting that the area was undergoing a
severe year-long drought, and that most of the plants in
the site were "dead and/or leafless": however, he
indicated that a few large (up to 6 m tall) clumps of
coppice appeared to be in relatively good condition.

The hypothesis of significanty more regeneration due

80

'to sprouting than to seed germination was not upheld,
however, since half of the plants with stems 3 1 cm DBH
had originated as seedlings since the plot was cut.
Despite harsh environmental conditions, dry forest species
have apparently developed adaptations to promote
successful dispersal, germination, and survivorship of
seeds and seedlings. The occurrence of high densities of
seeds of persistent forest species throughout the Cut Site
indicates that seed dispersal occurs effectively in dry
forest. This is not surprising, since 65% of the Species
assessed for dispersal type were vector-dispersed.

Another effective mechanism for dispersing seeds is the
rapid flowering and fruiting of coppiced individuals,
sometimes as quickly as one year following cutting (Ewel
1971 and pers. obs.). It is likely that seed banks
present in the soil prior to cutting also played a large
role in regeneration; numerous studies of tropical humid
to wet forests have demonstrated the importance of seed
banks to forest recovery (e.g. Keay 1960, Dubey and Mall
1975, Cheke et al. 1979).

The absence of a significant effect of the distance of
the forest edge on seedling densities in the Cut Site is
further indication that seed dispersal and seedling
survival occurs effectively throughout the site. It is
likely that a much larger clearcut (on the order of many

hectares) would show more of an edge effect.

81

Authors have generally postulated that dry forest
seeds can remain alive in a dormant condition for long
periods of time (e.g. Budowski 1965 and Gomez-Pompa et
al. 1972), but have not presented data to support this
assumption. Garwood (1983) reported mean length of
dormancy (time between sowing and germination) of seeds
from a semideciduous forest in Panama (which has few
species in common with the Guanica forest): values ranged
from 2 to 370 days. She watered the seeds continuously
throughout the study, however, so dormancy under field
conditions may be different. How long seeds can actually
survive harsh dry forest conditions such as heat and
desiccation, both of which generally cause high seed
mortality (Crocker 1938) has not been demonstrated. Many
dry forest species have seeds covered by harad seed coats
(pers. obs.), a character which generally contributes to
long sseed life spans (Crocker 1938). A few seeds from
litter and soil collections in this study germinated just
before the study was terminated one year after the
collection date, thus Guanica forest seeds may remain
germinable for 3 one year.

Although germinability was not measured in this study,
Murphy and Lugo (1983) reported that seeds 3 2.38 mm
collected from soil and litter in nearby mature forest
showed only 0-0.5% germination (0-1.6 seeds germinating

per m2) under greenhouse conditions. It is possible that

82

sterile potting soil and a greenhouse environment did not
provide ideal conditions for germination. It may also be
that dormancy-breaking requirements inhibit germination in
the greenhouse and in the field. Increased germination of
scarified Leucaena seeds (Table 11) and seeds of 4 other
species (Murphy and Lugo 1983), together with the low
germinability of viable fresh seeds (Table 10), indicates
that dormancy mechanisms are probably operable in at least
some species.

It is also possible that many dry forest persistent
species possess short-lived seeds, germinating only when
conditions are favorable. It is likely that at least some
species possess seeds which can survive drought, however,
since a number of trees produce fruits during the dry
season (Murphy and Lugo 1983).

The low viability, apparently low germinability, and
desiccated appearance of many seeds from litter and soil
seed banks indicate that seed mortality in the field is
high over the long run. However, with soil and litter
seed densities over 1000 seeds/m2 and a constant input of
new seeds, even very low germinability levels would be
sufficient to produce the 13.1 seedling-origin plants per

1112 which occur in the lB-year-old site.

Mycorrhizae

Janos' (1980) prediction that cutting a tropical
forest would not appreciably affect mycorrhizal
associations was supported by this study. Mycorrhizae may
be especially important in dry forests, Since they aid
plants in obtaining phosphorous and water in soils which
are low in available phosphorous (Safir et al. 1972, Read
et a1. 1976). Mycorrhizal plants have been shown to grow
significantly faster than non-mycorrhizal plants in a
number of tropical systems (see Janos 1980 and Bowen 1980
for reviews), rapid dry forest recovery may be partially
dependent on the maintenance of intact mycorrhizal
associations. In studies of severely disturbed arid
communities like old roadbeds (Reeves et a1. 1979),
revegetated mine spoils (Miller 1979), and freshly disked
farmlands (Ewel et al. 1976), where rhizospheres were
heavily disturbed, nonmycorrhizal secondary species were
the major invaders and mycorrhizal species became dominant
only in later seral stages. If the dry forest rhizosphere
were heavily disturbed by bulldozing, plowing, or heavy
herbicide application following a clearcut, mycorrhizal
systems would certainly be disrupted and a similar pattern
of nonmycorrhizal invaders might follow, slowing down the

recovery process considerably.

83

Recovery and Resilience

The long-lasting effects of Ewel's recut/herbicide
treatments underscores the importance of taking
disturbance history into account in studies of forest
recovery. The lower ratio of stem sprout to plant in the
”recut/herbicided“ portion of the site indicates that
herbicide and repeated-cutting treatments depleted the
stored energy reserves of woody plants so that fewer
sprouts were produced. Basal area differences after 13
years may indicate long-lasting reduced productivity.

Differences between the pre-cut structure of the 1 ha
site and the adjacent forest further emphasize the role of
disturbance in dry forests. The 1 ha site was
considerably more depauperate in species richness and
basal area and contained more individuals of the pioneer
species Leucaena leucocephela than mature forest in 1982,
indicating that the area was probably more disturbed in
the past, either naturally (e.g. by fire) or by humans
(selective cutting, trampling, etc.). The latter seems
likely because of the accessibility of the site; it is
closer to roads and flatter than much of the surrounding
forest. Evidence of past fires in the form of charcoal
fragments in soil pits was found in the Cut Site as well
as in mature forest stands (Murphy and Lugo, unpublished

data), but no fire frequency data are available. Because

84

85

on comparisons to adjacent forest as representing the
maximal measurable degree of maturity of Guanica dry
forest. This inserts site-to-site variability into the
study, but the immediate proximity of the forest to the
Cut Site should minimize differences.

Depending on the parameter considered, recovery after
13 years was within 7 to 64% of mature forest levels
(Table 14). Leaf area index and species composition were
closest to mature forest values, with total basal area,
mean height, and total biomass next at 40% of forest
values. Basal area in larger stem size classes had
recovered less, demonstrating the large contribution of
smaller stems to the structure of the younger site.

The general picture of dry forest revegetation is one
of relatively rapid recovery of biomass and cover. Only
13 years subsequent to cutting, the study site was
completely covered by vegetation, including many woody
plant Species characteristic of adjacent mature forest.
It may take quite a number of years for the site to obtain
the structure and full species complement of mature
forest, but the availability of propagules and the
relatively rapid growth rate of vegetative sprouts in the
site ensure that in the absence of further disturbance,
the forest will continue to aggrade relatively quickly.
Compared to comparable-aged sites, dry forest has

recovered more rapidly than one post-agricultural northern

86

Table 14. Degree of recovery of structural parameters in
13-year-old dry forest.

Parameter Cut Site Forest Percent
recovery

1) Species composition:
Number of mature forest
tree species with 17 41 59
stems > 1 cm DBH

2) Basal area (m2/ha)

All stems (from DGH) 19.7 52.8 40
Stems > 1 cm DBH 8.9 30.8 29
Stems > 5 cm DBH 1.2 18.6 7
3) Mean height - all 0.7 1.6 40
plants (m)
2 1
4) Aboveground biomass 19.1 44.9 43
(mt/ha)
1
5) Leaf Area Index 1.8 2.8 64
(November, m2/m2)
6) Epiphyte density 0.3 2.8 11
(plants/m2)

1From Murphy and Lugo (1985)
From Murphy, Lugo, and Murphy (1983)

87

temperate forest and more slowly than a post-agricultural
tropical moist evergreen forest. New Hampshire
14-year-old sites supported 19% of biomass of mature
northern temperate forests (Marks and Bormann 1972,
Whittaker 1974), while moist evergreen sites had recovered
56% of tree height, 35% of tree diameter, and 50% of basal
cover 14 years following abandonment (Aweto 1981).

The measurement of relative resilience is complicated
by several factors. The first is the difficulty of
obtaining accurate data on primary or mature secondary
forest for estimation of recovery rates. In Guanica,
there are no extant examples of primary dry forest, and
the extent to which older secondary stands have reached
their climax is not known.

Secondly, the constancy and rate of vegetation growth
are not known for most ecosystems; the oft-made assumption
that recovery is constant and logarithmic is largely
unsubstantiated (Westman 1978). Patterns of recovery
differ widely according to the parameter measured. Growth
rates have variously been described as sigmoidal (Ewel
1983), asymptotic (Opler et al. 1977), and linear (Hibbs
1983). Dry forest recovery showed highly variable growth
patterns for different parameters, as is demonstrated by
the shapes of the curves in Figure 20.

Another difficulty in quantifying resilience is the

assumption that a climax condition identical to the

88

BIOMASS

 

 

STEM DENSITY

I

/

’
‘1'"

 

_—-_-_—-*
’
’O'
/.

~
.- --‘-~~~

HEIGHT

 

 

 

 

11 LEAF AREA INDEX
3,7”
I,
’4 /
. \ I
\ I

1 \r—”'
41'

1

1

l

1 T

0 25 50

 

PERCENT COVER

 

SPECIES DIVERSITY

#1

 

 

0 25 50

AGE OF SITE (YEARS)

Figure 20. Patterns of structural change in long-term

dry forest succession.

Circles mark data points, and

dotted lines are hypothetical projections of pattern.
Data are from this study, Ewel (1971), and Murphy and

Lugo (1983).

89

original will be obtained by a recovering ecosystem. Many
ecologists have noted that stochastic processes are
important in succession (e.g. Horn 1975), so the
probability of restoration of identical structure may be
small.

Finally, the degree of recovery is highly dependent on
the type and intensity of disturbance, the size of the
disturbed area, and the proximity of seed sources.
Comparable data from different ecosystems with these 4
factors held constant are essentially nonexistent. For
this reason, it is important that recovery of ecosystems
under a variety of disturbance regimes be assessed with
standardized measurements which can be used to compare to
recovery rates from other studies. The delimitation of
the most useful parameters to measure and the most
important site information to collect will make possible
the collection of data which can be used to make accurate
assessments of relative resilience.

The measurements listed in Table 14 all provide useful
information regarding resilience of dry forest. The
combination of all these values into an index of
resilience may simply mask the important details of
recovery, and may place artificial emphasis on some of the
values. For example, the exclusion of any of the three
basal area values listed would leave out information

regarding the sizes of stems in the site, but the

9O

inclusion of all three in an index would heavily weight
that parameter. Another example is epiphyte density,
which gives a different kind of information than any of
the other values, and therefore should be considered
separately. It is suggested, therefore, that recovery
values of each of these parameters be quantified in future
studies, and reported together with available information
on areal extent of the disturbance, intensity and
frequency of disturbance, and relative maturity of the
original forest; all of these factors considered together

give a composite picture of resilience.

SUMMARY

In thirteen-year-old dry forest, basal area of stems 3
1 cm DBH and mean plant height were 40% of mature forest
measurements. Leaf area index and epiphyte density were
64% and 11%, respectively, of values in mature forest.
Structure and species richness still differed
significantly from mature forest values, however
(Hotelling T test: ANOVA). It is suggested that in
studies of ecosystem recovery, all these parameters be
measured and considered as a whole to provide a composite
picture of resilience.

Cut Site species composition included a distinct
pioneer species component, with 46% of stem density and
28% of species classified as pioneer as compared to 3% and
4%, respectively, in mature forest. Mature forest species
were also well represented in the site, with 59% of all
tree species with stems 3 1 cm DBH in mature forest plots
also occuring in Cut Site plots.

Species diversity (species per unit area) was higher
in the l3-year-old site than in one-year-old or mature
forest sites, supporting the hypothesis of highest
diversity in mid-successional sites.

The hypothesis of dominance of vegetative sprouting
over seed germination was not supported. Of plants

supporting stems 3 1 cm DBH in the Cut Site, half

91

92

originated as seedlings since the 1969 clearcut, and half
as sprouts from roots or cut stems. High seed densities

in soil and litter (780 to 1265 per 1112

in Cut Site in
1982) may account for high seedling densities despite low
seed viability of persistent species (0-7% per species),
apparent low germinability, and the many desiccated seeds.

Numbers of woody seedling origin plants decreased with
distance from the forest edge in the 13-year-old site,
though not significantly.

Quantities of VA mycorrhizal associations did not
differ in plants occurring in one-year-old, 13-year-old,
and mature forest sites, indicating that cutting did not
appreciably affect mycorrhizae.

The effects of recut/herbicide treatments applied to a
portion of the Cut Site in 1969 persisted 13 years later,
evidenced by significant structural differences (Hotelling
T test) between the 'recut/herbicided" portion and the
"cut only” portion. Trees in the 'recut/herbicided" area
possessed nearly twice as many stems and had significantly
less basal area than those in the "cut only" portion.

In conclusion, recovery on the 13-year-old site has
been relatively rapid. Because of the relative simplicity
of mature West Indian dry forest and the potential of
component species to grow and reproduce under highly
disturbed conditions, the system's resilience is perhaps

greater than that of tropical rain forest.

93

Table A1. Characteristics of all species occurring in plots.
Nomenclature according to Little and Wadsworth (1964) and Little.
Woodbury, and Wadsworth (1974) for trees. and Britton and Wilson
(1924) for other species. except where noted. T-tree, s-shrub,
H(P)-perennial herb. H(A)-annual herb, c-cactus, Ep-epiphyte,
V(W)-woody vine, V(H)-herbaceous vine, Pt-persistent, Pr-pioneer,
D-deciduous. E-evergreen, N-native to Puerto Rico, Ex-exotic.
En-endemic to Puerto Rico.

‘5
4‘68 990%
85‘ g. <9.

Qfi' $9 Q‘F
Species o ‘0 ‘9‘ \9 vs" 9
Q 9~ @- 14 ‘\ S \'

<9 ‘45" (3‘ ‘0 $9 ‘9 s Q
Q. 99‘

CUT SITE 8
Amyris elemifera T Pt E N Y
Antirhea acutata T Pt E N Y
Argythamnia fasciculata S Pt - N No
Ayenia pusilla H(P) Pr - N No
Bernardia dichotoma T Pt 0 N Y
Bourreria succulents T Pt E N No
Bucida buceras T Pt 0 N Y
Bumelia obovata T Pt E N Y
Bursera simaruba T Pt 0 N Y
Calliandra caracasana T Pr 0 N No
Cassine xylocarpa T Pt E N Y
Cephalocereus royenii C Pt - N Y
Chiococca alba V(W) Pt - N No
Cissus trifoliata V(H) Pt - N Y
Citharexylum fruticosum T Pt E N No
Citharexylum spinosum T Pt E N No
Clusia rosea T Pt E N No
Coccoloba diversifolia T Pt E N Y
Coccoloba krugii T Pt E N Y
Coccoloba michrostachya T Pt E N Y
Commelina elegans H(P) Pt - N Y
Comocladia dodonea T Pt E N Y
Corchorus hirsutus S Pr - N No
Crossapetalum rhacoma T Pt E N Y
Croton betulinus S Pr - N N
Croton discolor 5 Pt - N N
Croton humilis S Pt - N Y
Croton rigidus S Pt - En Y

*Dalea carthagenensis 5 Pt - 7 No
Desmanthus virgatus S Pr - N No
Distictis lactiflora V(W) Pt - N Y
Eugenia foetida T Pt E N Y
Eugenia ligustrina T Pt E N No

Table A1 (Cont'd)

Eugenia rhombea
Eupatorium sinuatum
Exostema caribaeum
Galactia dubia
Galactia striata
Gouania lupuloides
Guaicum sanctum
Guettarda elliptica
Guettarda krugii
Gymnanthes lucida
Helicteres jamaicensis
*Herissantia crispa
Heteropterus purpurea
Hibiscus brasiliensis
Hypelate trifoliata
lpomea arenaria
Jacquemontia pentantha
Justicia sessilis
Krugiodendron ferreum
Lantana involucrata
Lasiacis divaricata

Leptocereus quadricostasus

Leucaena leucocephala
Opuntia repens
Panicum maximum
Passiflora sp.
Pictetia aculeata
Pisonia albida

Pithecellobium unguis-cati

Plumeria alba
Randia aculeata
Reynosia uncinata
*Rhyncheletrum repens
Sabinea florida
Samyda dodocandra
Savia sessiliflora
Scleria lithosperma
Setaria setosa
Setaria utowanaea
Sida acuminata

94

Stigmaphyllon periplocifolium V(W) Pt

Thouinia portoricensis
Tillandsia recurvata
Tillandsia utriculata

Tournefortia microphylla

Tragia volubilis
Turnera diffusa
Vanilla egersii

T Pt
S Pr
T Pt
V(H) Pr
V (W) Pr
V (W) Pr
T Pt
T Pt
T Pt
T Pt
T Pr
v (H) Pt
v (11) Pt
V (W) Pr
T Pt
V(H Pt
V (W) Pr
H (P) Pr
T Pt
S Pr
H (P) Pt
C Pt
T Pr
C Pt?
H (P) Pr
V (W) Pr
T Pt
T Pt
T Pt
T Pt
T Pt
T Pt
H Pr
T Pt
T Pt
T Pt
H (P) Pt
H (P) ' Pr
H (P) Pr
5 Pr
T Pt
E Pt
E Pt
V (W) Pr
V (H) Pr
5 Pr
v (H) . E Pt

IMMMMMIIIMIM

16°C I MUMODOI I

-<-<zzz-<z-<
000 O

z-<-<-<

.O

-<-<zz-<zz-<
00 OO

00

O

<<<<4€Z<~<<<<<ZZ<<<

No

95

Table A1 (Cont'd)

Wedelia lanceolata S Pr? -
Zamia portoricensis H(P) Pt -
Unknown #87 7 7 7
OCCURRING IN FOREST ONLY
Adelia ricinella T Pt 0
Colubrina arborescens T Pt E
Jatropha hernandiifolia T Pt E
Reynosia guama T Pt E
Tabebuia heterophylla T Pt E
Unknown tree species (5) T Pt 7
Unknown vine species (4) V(W) Pt -

*Nomenclature according to Long and Lakela (1971).

En
En

uwzzzzz

No

No

96

Table A2. Structural parameters of all stems 3 5 cm DBH on the
study site before it was cut in 1969. Data summarized from J.
Ewel (unpublished data).

Value per 10 x 100 m transect

Number of Total Total
Transect species DBH Basal Area Density
(cm) (cm2)
1 12 795.3 5175.7 116
2 8 731.8 4653.2 102
3 10 822.2 5020.8 120
Total 18-22 2349.3 14849.8 338
Mean 10.0 783.1 4949.9 112.7

Mean/ha - 7831 49499 1127

97

Table A3. Importance values for woody plant species with DBH 3 5 cm on the
study site before it was cut in 1969. Data summarized from three 100 x 10 m
transects measured by J. Ewel (unpublished data).

Species IV 3 of Rel. Rel. Rel.
Total Density Freq. Dominance

Gymnanthes lucida 70.45 23.48 28.99 9.7 31.76
Exostema caribaeum 57.17 19.06 30.47 9.7 17.00
Bucida buceras 24.96 8.32 8.88 3.2 12.88
Pisonia albida 20.16 6.72 5.92 6.5 7.74
Amyris elemifera 19.84 6.61 5.33 9.7 4.81
Bursera simaruba 19.66 6.55 2.37 9.7 7.59
Leptocereus quadricostasus 17.49 5.83 1.78 9.7 6.01
Leucaena leucocephala 10.63 3.54 2.07 6.5 2.06
Cephelocereus royenii 8.26 2.75 1.78 3.2 3.28
Krugiodendron ferreum 4.68 1.56 0.88 3.2 0.60
Tabebuia heterophylla 4.43 1.48 0.59 3.2 0.64
Pictetia aculeata 4.22 1.41 0.59 3.2 0.43
Jacquinia berterii 4.16 1.39 0.59 3.2 0.37
Colubrina arborescens 3.79 1.26 0.30 3.2 0.29
Comocladia dodonea 3.73 1.24 0.30 3.2 0.23
Reynosia uncinata 3.71 1.24 0.30 3.2 0.22
2-6 unknown species 22.67 7.56 8.88 9.7 4.09
TOTAL 300.01 100.00 100.02 100.0 100.00

*It is possible that other species were included in the count of Gymnanthes. as
some of the diameters recorded seem rather high. The similarity of the Spanish
common names (used to record data) of Gymnanthes and Bucida may have caused
some errors in recording.

98

Table A4. Multivariate comparisons of structure in the Cut
Site and forest 2 x 10 m plots using the Hotelling T square
test. Variables used were per plot values for species
richness, plant density, stem density, and total DBH.

Cut Site ("Cut only" area) versus Forest plots

Hotelling T square 64.1637
F value 14.7745
Significance (P value) .0000
Degrees of freedom 4. 35

Cut Site: "Cut only" area versus ”Recut/herbicide" area plots

Hotelling T square 12.2061
F value 2.5429
Significance (P value) .0830

Degrees of freedom 4, 15

99

LITERATURE CITED

Agee, J.K. and L. Smith. 1984. Subalpine tree
reestablishment after fire in the Olympic Mountains,
Washington. Ecology 65(3): 810-819.

Aristeguieta, L. 1968. El bosque cadicifolio seco de
los llanos altos centrales. Boln. Soc. Venez. Cienc.
Nat. 27: 395-438.

Auclair, A.N. and F.G. Goff. 1971. Diversity relations of
upland forests in the western Great Lakes area.
Am. Nat. 105: 499-528.

Aweto, A.O. 1981. Secondary succession and soil fertility
restoration in south-western Nigeria. II. Soil
fertility restoration. J. Ecol. 69: 609-614.

Bazzaz, F.A. 1975. Plant species diversity in old-field
successional ecosystems in southern Illinois.
Ecology 56(2): 485-488.

Benzing, D.H. 1980. The Biology of the Bromeliads.
Mad River Press, Eureka, Calif. 305 pp.

Berish, C.W. 1982. Root biomass and surface area in three
successional tropical forests. Can. J. For. Res.
12(3): 699-704.

Bormann, F.H. and G.E. Likens. 1981. Pattern and Process
in a Forested Ecosystem. Springer-Verlag, New York.
253 pp.

Bourliere, F. (ed.). 1983. Ecosystems of the World. 13.
Tropical Savannas. Elsevier Scientific Pub. Co.,
New York. 730 pp.

Bowen, G.D. 1980. Mycorrhizal roles in tropical plants and
ecosystems. pp. 165-189 In: Mikola, P. (ed. ,
Tropical Mycorrhiza Research, Oxford Univ. Press.

Britton, N.L. and P. Wilson. 1923-30. Botany of Puerto
Rico and the Virgin Islands. New York Academy of
Sciences, New York.

100

Brower, J.E. and J.H. Zar. 1977. Field and laboratory
methods for general ecology. Wm. C. Brown Co. Pub.,
Dubuque, Iowa. 194 pp.

Brown, S. and A.E. Lugo. 1982. The storage and production
of organic matter in tropical forests and their role
in the global carbon cycle. Biotropica '
14(3): 161-187.

Budowski, G. 1963. Forest succession in tropical lowlands.
Turrialba 13(1): 42-44.

Budowski, G. 1965. Distribution of tropical American rain
forest species in the light of successional
processes. Turrialba 15(1): 40-42.

Cairns, J. Jr. and K.L. Dickson. 1977. Recovery of streams
and spills of hazardous materials. pp. 24-42 In:
J. Cairns, Jr., K.L. Dickson, and E.E. Herricks
(eds.), Recovery and Restoration of Damaged
Ecosystems. Univ. of Virginia Press, Charlottesville.

Cheke, A.S., W. Nanakorn, and C. Yankoses. 1979. Dormancy
and dispersal of seeds of secondary forest species
under the canopy of a primary tropical rain forest
in northern Thailand. Biotropica 11: 88-95.

Clapham, W.B. Jr. 1971. Natural ecosystems. Chapt. 7 In:
Resiliency and Fitness of Ecosystems. Macmillan
Publishing Co., New York.

Covington, W.W. and J.D. Aber. 1980. Leaf production
during secondary succession in northern hardwoods.
Ecology 61: 200-204.

Crocker, W. 1938. Life-span of seeds. Bot. Review
4(6): 235-274.

Daubenmire, R. 1972. Phenology and other characteristics
of tropical and semi-deciduous forest in north-
western Costa Rica. J. Ecol. 60: 147-170.

Davis, T.A.W. and P.W. Richards. 1933. The vegetation of
Moraballi Creek, British Guiana. J. Ecol.
21: 350-384.

Day, R.J. 1972. Stand structure, succession, and use of
southern Alberta's Rocky Mountain forest. Ecology
53: 472-478.

Diem, H.G., I. Gueye, V. Gianinaazzi-Pearson, J.A.
Fortin, and Y.R. Dommergues. 1981. Ecology of VA

101

mycorrhizae in the tropics: the semi-arid zone of
Senegal. Acta Oecologica/Oecologia Plantarum
2: 53-62.

Dixon, W.J. 1983. Comparison of two groups with t tests:
BMDP3D. pp. 94-103 In: Dixon, W.J. (ed.) BMDP
Statistical Software, 1983 Printing with Additions.
Univ. of Calif. Press, Berkeley. 733 pp.

Doley, D. 1981. Tropical and subtropical forests and
woodlands. p. 209-323 In: T.T. Kozlowski (ed.)
Water Deficits and Plant Growth. Academic Press,
New York.

Dubey, P.S. and L.P. Mall. 1975. Effect of herbicides on
germination, viability and seedling growth of weed
seeds. I. Digera alternifolia Aschers.

Trop. Ecol. 16: 39-42.

Dunevitz, V.L., M.F. Mulligan, P.G. Murphy, and A.E. Lugo.
1983. The occurrence of vesicular-arbuscular
mycorrizae in a subtropical dry forest. Abstract of
paper presented at Third International Symposium on
Microbial Ecology, Mich. State Univ., August
7-12, 1983.

Ewel, J.J. 1971. Experiments in arresting succession with
cutting and herbicides in five tropical environments.
PhD dissertation, Univ. of North Carolina, Chapel
Hill. 248 pp.

Ewel, J.J. 1977. Differences between wet and dry
successional tropical ecosystems. Geo-Eco-Trop
1(2): 103-117.

Ewel, J.J. 1980. Tropical succession: manifold routes
to maturity. Biotropica 12 (supplement): 2-7.

Ewel, J.J. 1983. Succession. pp. 217-223 In: F.B. Golley
(ed.) Tropical Rain Forest Ecosystems. A. Structure
and Function. Elsevier Scientific Pub. Co.,
Amsterdam.

Ewel, J., C. Berish, B. Brown, N. Price and J. Raich.
1981. Slash and burn impacts on a Costa Rican wet
forest site. Ecology 62(3): 816-829.

Ewel, J., R. Meador, R. Myers, L. Conde, and B. Sedlik.
1976. Studies of vegetation changes in south Florida.
Final report, subcontract on research agreement
#18-492 with USDA-Forest Service.

102

Ewel, J.J. and J.L. Whitmore. 1973. The ecological life
zones of Puerto Rico and the 0.5. Virgin Islands.
Forest Service Res. Paper ITF-lB, Institute of
Tropical Forestry, Rio Piedras, Puerto Rico. 72 pp.

Fox, J.E.D. 1976. Constraints on the natural regeneration
of tropical moist forest. Forest Ecol. Manage.
1: 37-65.

Frankie, C.W., H.G. Baker, and P.A. Opler. 1974.
Comparative phenological studies of trees in
tropical wet and dry forests in the lowlands of
Costa Rica. J. Ecol. 62: 881-919.

Garwood, N.C. 1981. Seed germination in a seasonal
tropical forest in Panama: a community study. Ecol.
Monog. 53(2): 159-181.

Gentry, A.H. 1983. Lianas and the "paradox" of contrasting
latitudinal gradients in wood and litter production.
Tropical Ecology 24(1): 63-67.

Gleason, H.A. and M.T. Cook. 1927. Plant ecology of Puerto
Rico. Sci. Surv. Puerto Rico and the Virgin Islands.
7(1-2): 1-173.

Goldberg, D.E. 1982. The distribution of evergreen and
deciduous trees relative to soil type: an example
from the Sierra Madre, Mexico, and a general model.
Ecology 63(4): 942-951.

Golley, F.B. and J.B. Gentry. 1966. A comparison of
variety and standing crop of vegetation on a
one-year and a twelve-year abandoned field.
Oikos 15(2): 185-199.

Gomez-Pompa, A. 1974. Recovery of tropical ecosystems.
pp. 113-138 In: F. Golley (ed.) Fragile Eccosystems.
Springer-Verlag, New York. 258 pp.

Gomez-Pompa, A., C. Vazquez-Yanes, and S. Guevara. 1972.
The tropical rain forest: a nonrenewable resource.
Science 177: 762-765.

Grieg-Smith, P. 1952. Ecological observations in degraded
and secndary forest in Trinidad, British West Indies.
J. of Ecol. 40: 283-330.

Grubb, P.J. and E.V.J. Tanner. 1976. The montane forests
and soils of Jamaica: a re-assessment. J. Arnold
Arboretum 57: 313-368.

103

Hall, J.B. and M.D. Swaine. 1980. Seed stocks in Ghanaian
forest soils. Biotropica 12(4): 256-263.

Henry, J.D. and J.M.A. Swan. 1974. Reconstructing forest
history from live and dead plant material - an
approach to the study of forest succession in
southwest New Hampshire. Ecology 55: 772-783.

Hibbs, D.E. 1983. Forty years of forest succession in
central New England. Ecology 64(6): 1394-1401.

Hill, A.R. 1975. Ecosystem stability in relation to
stresses caused by human activities. Canad. Geog.
19(3): 206-220.

Hogberg, P. 1982. Mycorrhizal associations in some
woodland and forest trees and shrubs in Tanzania.
New Phytol. 92: 407-415.

Holdridge, L.R. 1967. Life zone ecology. Tropical Science
Center, San Jose, Costa Rica. 206 pp.

Holling, C.S. 1973. Resilience and stability of ecological
systems. Ann. Rev. Ecol. Syst. 4: 1-24.

Horn, H.S. 1975. Markovian properties of forest
succession. pp. 196-211 In: M.L. Cody and J.M.
Diamond (eds.) Ecology and Evolution of Communities.
Belknap Press, Cambridge, Mass.

Hubbell, S.P. 1979. Tree dispersion, abundance, and
diversity in a tropical dry forest. Science
203: 1299-1309.

Isensee, A.R., W.C. Shaw, W.A Gentner, C.R. Swanson,
B.C. Turner, and E.A. Woolson. 1973. Revegetation
following massive application of selected herbicides.
Weed Science 21(5): 409-412.

Janos, D.P. 1980. Vesicular-arbuscular mycorrhizae affect
lowland tropical rain forest plant growth. Ecology
61(1): 151-162.

Janzen, D.H. 1967. Synchronization of sexual reproduction
of trees within the dry season in Central America.
Evolution 21: 620-637.

Janzen, D.H. 1981. Patterns of herbivory in a tropical
deciduous forest. Biotropica 13(4): 271-282.

Keay, R.W. 1960. Seeds in forest soil. Nigerian Forest
Infor. Bull., New Series 4: 1-4.

104

Kellman, M.C. 1970. Secondary plant succession in tropical
montane Mindanao. Dept. Biogeogr. Publ. BG/Z,
Australian National Univ., Canberra, Australia.

Kim, J. and F.J. Kohout. 1975. Analysis of variance and
covariance: subprograms ANOVA and ONEWAY. pp. 398-433
In: Nie, N.H., C.H. Hull, J.G. Jenkins, K.
Steinbrenner, and D.H. Bent (eds.) Statistical
Package for the Social Sciences, Second Edition.
McGraw-Hill, New York. 675 pp.

Kim, J. and F.J. Kohout. 1975. Multiple regression
analysis: subprogram regression. pp. 320-367 In:
Nie, N.H., C.H. Hull, J.G. Jenkins, K.
Steinbrenner, and D.H. Bent (eds.), Statistical
Package for the Social Sciences, Second Edition.
McGraw-Hill, New York. 675 pp.

Kira, T. and T. Shidei. 1967. Primary production and
turnover of organic matter in different forest
ecosystems of the western Pacific. Japan. J. Ecol.
17: 70-87.

Kittredge, J., Jr. 1934. Evidence of the rate of forest
succession on Star Island, Minnesota. Ecology
15: 24-35.

Larsen, J.A. 1925. Natural reproduction after forest
fires in northern Idaho. J. Agr. Res. 30: 1177-1197.

Leak, W.B. and R.W. Wilson, Jr. 1958. Regeneration after
cutting of old-growth northern hardwoods in New
Hampshire. USDA Northeastern Forest Exper. Station
Paper 103.

Leps, J., J. Osbornova-Kosinova and M. Rejmanek. 1982.
Community stability, complexity, and species life
history strategies. Vegetatio 50(1): 53-64.

Lieberman, D. 1982. Seasonality and phenology in a dry
tropical forest in Ghana. J. Ecology 70: 791-806.

Lieberman, D. and M. Lieberman. 1984. The causes and
consequences of synchronous flushing in dry tropical
forest. Biotropica 16(3)): 193-201.

Likens, G.E., F.B. Bormann, R.S. Pierce, and W.A. Reiners.
1978. Recovery of a deforested ecosystem. Science

Little, E.L. and F.B. Wadsworth. 1964. Common trees of

105

Puerto Rico and the Virgin Islands. USDA Forest
Service, Agr. Handbook 249.

Little, E.L., R. Woodbury, and F.H. Wadsworth. 1974. Trees
of Puerto Rico and the Virgin Islands. Vol. 2. USDA
Forest Service Agr. Handbook No. 499.

Long, R.W. and O. Lakela. 1971. A Flora of Tropical
Florida. A Manual of the Seed Plants and Ferns of
Southern Peninsular Florida. Univ. of Miami Press,
Coral Gables, Florida. 962 pp.

Loucks, O.L. 1970. Evolution of diversity, efficiency, and
community stability. Am. 2001. 10: 17-25.

Loveless, A.R. and G.F. Asprey. 1957. The dry evergreen
formations of Jamaica. 1. The limestone hills of the
south coast. J. Ecol. 45: 799-822.

Lugo, A.E., J.A. Gonzalez-Liboy, B. Cintron, and K.
Dugger. 1978. Structure, productivity, and
transpiration of a subtropical dry forest in Puerto
Rico. Biotropica 10(4): 278-291.

Lugo, A.E. and P.G. Murphy. 1985. Nutrient dynamics of a
Puerto Rican subtropical dry forest. Submitted.

Lutz, H.J. 1943. Injuries to trees caused by Celastrus
and Vitis. Bull. Torr. Bot. Club 70: 436-439.

Malaisse, F.P. 1974. Phenology of the Zambezian woodland
area with an emphasis on the Miombo ecosystem.
pp. 269-286 In: Lieth, H. (ed.) Phenology and
Seasonality Modeling. Springer-Verlag, New York.
444 pp.

Malaisse, F.P. 1978. The Miombo ecosystem. In: Tropical
Forest Ecosystems. UNESCO Nat. Resour. Res. 14,
UNESCO, Paris.

Malmer, N., L. Lindgren, and S. Persson. 1978.
Vegetational succession in a south Swedish wood.
Vegetatio 36(1): 17-29.

Margalef, R. 1963. On certain unifying principles in
ecology. Am. Nat. 97: 357-374.

Margalef, R. 1975. Diversity, stability and maturity in
natural ecosystems. pp. 151-160 In: W.H. van Dobben
and R.H. Lowe-McConnell (eds.) Unifying Concepts in
Ecology. Junk, The Hague, Netherlands.

106

Marks, P.L. and F.H. Bormann. 1972. Revegetation following
forest cutting: mechanisms for return to steady-state
nutrient cycling. Science 176: 914-915.

Martin, N.D. 1959. An analysis of forest succession in
Algonquin Park, Ontario. Ecol. Monogr. 29: 187-218.

Mellinger, M.V. and S.J. McNaughton. 1975. Structure and
function of successional vascular plant communities
in central New York. Ecol. Monogr. 45: 161-182.

Miller, R.M. 1979. Some occurrance of vesicular-arbuscular
mycorrhiza in natural and disturbed ecosystems of the
Red Desert. Can. J. Bot. 57: 619-623.

Monette, R. and S. Ware. 1983. Early forest succession in
the Virginia coastal plain. Bull. Torr. Bot. Club
110(1): 80-86.

Monk, C.D. 1974. Plant species diversity in old-field
succession on the Georgia piedmont. Ecology
55: 1075-1085.

Monroe, W.H. 1976. The karst landforms of Puerto Rico.
Geol. Surv. Professional Paper 899. U.S. Government
Printing Office, Washington. 69 pp.

Murphy, P.G. and S.R. Kephart. 1981. Seasonal dynamics and
succession in a subtropical dry forest. Proposal to
NSF, January 1981.

Murphy, P.G. and A.E. Lugo. 1982. Seasonal dynamics and
succession in a subtropical dry forest. Progress
report for the period 15 June 1981 - 15 February
1982, NSF Grant DEB-8110208.

Murphy, P.G. and A.E. Lugo. 1983. Seasonal dynamics and
succession in a subtropical dry forest. Progress
report for the period 1981-1983, NSF Grant
DEB-8110208.

Murphy, P.G. and A.E. Lugo. 1985. Structure and biomass
of a subtropical dry forest in Puerto Rico. In press,
Biotr0pica.

Murphy, P.G., A.E. Lugo, and A.J. Murphy. 1983. Seasonal
dynamics and recovery of a Caribbean dry forest.
Bull. Ecol. Soc. Am. 64(2):113.

Nicholson, S.A. and C.D. Monk. 1974. Plant species
diversity in old-field succession on the Georgia
Piedmont. Ecology 55: 1075-1085.

107

Nie, N.H., D.H. Bent, and C.H. Hull. 1975. Description of
subpopulations and mean difference testing: sub-
programs BREAKDOWN and T-TEST. pp. 249-275 In: Nie,
N.H., C.H. Hull, J.G. Jenkins, K. Steinbrenner, and
D.H. Bent (eds.) Statistical Package for the Social
Sciences, Second Edition. McGraw-Hill, New York.

675 pp.

Nye, P.H. and D.J. Greenland. 1960. The soil under
shifting cultivation. Tech. Comm. No. 51. Common-
wealth Bureau of Soils, Harpenden, England. 156 pp.

Odum, E.P. 1969. The strategy of ecosystem development.
Science 164: 262-270.

Opler, P.A., H.G. Baker, and G.W. Frankie. 1977. Recovery
of tropical lowland forest ecosystems. pp. 379-421
In: J. Cairns, Jr., K.L. Dickson, and E.E. Herricks
(eds.) Recovery and Restoration of Damaged Eco-
systems. Univ. Press of Virginia, Charlottesville.

Opler, P.A., G.W. Frankie, and H.G. Baker. 1980. Com-
parative phenological studies of treelet and shrub
species in tropical wet and dry forests in the‘
lowlands of Costa Rica. J. Ecol. 68: 167-188.

Peet, R.K. and N.L. Christensen. 1980. Succession: a
population process. Vegetatio 43: 131-140.

Phillips, J.M. and D.S. Hayman. 1970. Improved procedures
for clearing roots and staining parasitic and
vesicular-arbuscular mycorrhizal fungi for rapid
assessment of infection. Trans. Br. Mycol. Soc.
55(1): 158-163.

Pickett, S.T.A. 1982. Population patterns through 20 years
of oldfield succession. Vegetatio 49(1): 45-59.

Pressland, A.J. 1975. Productivity and management of mulga
in south-western Queensland in relation to tree
structure and density. Aust. J. Bot. 23: 965-976.

Putz, F.E. 1980. Lianas vs. trees. Biotropica 12: 224-225.

Read, D.J., J.K. Koucheki, and J. Hodgson. 1976.
Vesicular-arbuscular mycorrhiza in natural vegetation
systems. I. The occurrence of infection. New Phytol.
77: 641-653.

Reeves, F.B., D. Wagner, T. Moorman, and J. Kiel. 1979.
The role of endomycorrhizae in revegetation practices

108

in the semi-arid west. 1. A comparison of incidence
of mycorrhizae in severely disturbed vs. natural
environments. Am. J. Bot. 66(1): 6-13.

Richards, P.W. 1955. The secondary succession in the
tropical rainforest. Sci. Prog. 43: 45-57.

Romme, W.H. 1982. Fire and landscape diversity in
subalpine forests of Yellowstone National Park.
Ecol. Monog. 52(2): 199-221.

Romme, W.H., and D.H. Knight. 1981. Fire frequency and
subalpine forest succession along a topographic
gradient in Wyoming. Ecology 62: 319-326.

Ross, R. 1954. Ecological studies on the rain forest of
southern Nigeria. III. Secondary succession in the
Shasha Forest Reserve. J. Ecol. 42: 259-282.

Safir, G.R., J.S. Boyer, and J.W. Gerdemann. 1972.
Nutrient status and mycorrhizal enhancement of water
transport in soybean. Plant Physiol. 49: 700-703.

Shafi, M.I. and G.A. Yarranton. 1973. Diversity, floristic
richness, and species evenness during a secondary
(post-fire) succession. Ecology 54: 897-902.

Siccama, T.G., G. Weir, and K. Wallace. 1976. Ice damage
in a mixed hardwood forest in Connecticut in relation
to Vitis infestation. Bull. Torr. Bot. Club
103(4): 180-183.

Sprugel, D.G. and F.H. Bormann. 1981. Natural disturbance
and the steady state in high-altitude balsam fir
forests. Science 211: 390-393.

Spurr, S.H. 1956. Natural restocking of forests following
the 1938 hurricane in central New England. Ecology
37: 443-451.

Stanton, N. 1975. Herbivore pressure on two types of
tropical forests. Biotropica 7(1): 8-11.

Symington, C.F. 1933. The study of secondary growth on
rainforest sites in Malaya. Malay. Forester
2: 107-117.

Taylor, D.L. 1980. Biotic succession of lodgepole pine
forests of fire origin in Yellowstone National Park.
Nat. Geog. Soc. Res. Reports 12, 1971 Projects.

Tomkins, D.J. and W.F. Grant. 1977. Effects of herbicides

109

on species diversity of two plant communities.
Ecology 58(2): 398-406.

Tomlinson, F.B. 1980. The Biology of Trees Native to
Tropical Florida. Harvard University, Allston,
Mass. 480 pp.

Tosi, J. 1980. Life zones, land-use, and forest vegetation
in the tropical and subtropical regions. pp. 44-64
In: Carbon Dioxide Effects Research and Assessment
Program: The Role of Tropical Forests of the World
Carbon Cycle. Report of Symposium, Rio Piedras,
Puerto Rico. 156 pp.

Trimble, G.R., Jr., and S.H. Tryon. 1974. Grapevines a
serious obstacle to timber production on good hard-
wood sites in Appalacia. Northern Logger 23: 22-23.

Uhl, C. 1982. Recovery following disturbances of different
intensities in the Amazon rain forest of Venezuela.
Interciencia 7(1): 19-24.

Uhl, C. and G.F. Jordan. 1984. Succession and nutrient
dynamics following forest cutting and burning in
Amazonia. Ecology 65(5): 1476-1490.

van Riper, C. 1980. The phenology of the dryland forest
of Mauna Loa, Hawaii, and the impact of recent
environmental perturbations. Biotropica
12(4): 282-291.

Vogl, R.J. 1969. One-hundred and thirty years of plant
succession in a southeastern Wisconsin lowland.
Ecology 50: 248-255.

Wadsworth, F.H. 1950. Notes on the climax forests of
Puerto Rico and their destruction and conservation
prior to 1900. Carib. Forester 11(1): 38-47.

Webb, L.J., J.G. Tracey, and W.T. Williams. 1972.
Regeneration and pattern in the subtropical rain
forest. J. Ecol. 60: 675-695.

Westman, W.E. 1978. Measuring the inertia and resilience
of ecosystems. BioScience 28(11): 705-710.

Whittaker, R.H. 1975. The design and stability of plant
communities. pp. 169-181 In: W.H. von Dobbe and
R.H. Lowe-McConnell (eds.), Unifying Concepts in
Ecology. Junk, The Hague.

Whittaker, R.H., F.H. Bormann, G.E. Likens, and T.G.

110

Siccama. 1974. The Hubbard Brook ecosystem study:
forest biomass and production. Ecol. Monogr.
44: 233-254.

Woodwell, G.M. 1967. Radiation and the patterns of nature.
Science 156(3774): 461-470.

"I1111111111111111.1111“