MODERN ECOLOGY 
Basic and Applied Aspects 
Edited by 
G. Esser 
International Institute for Applied Systems Analysis (IIASA), 
Schlossplatz 1, A-2361 Laxenburg, Austria 
and 
D. Overdieck 
Institut fUr Okologie der Technischen Universitat, 
Koningin-Luise-Strasse 22, D-1000 Berlin 33, Germany 
Elsevier 
Amsterdam - London - New York - Tokyo 1991 
hapter 4 
Survival and growth 
beneath and near parents: 
The case of Myrcianthes 
fragrans (Myrtaceae) 
D. F. WHIGHAM and E. CABRERA CANO 
4.1 Introduction 
Discussions of the fate of seedlings in tropical forests have mostly focussed on 
whether or not they become established beneath the canopy of their parents and 
the importance of gap versus non-gap environments to their growth and survival 
(CLARK and CLARK 1984). In the early 1970's, CONNELL (1971) and JANZEN 
(1971) independently proposed that seedling establishment would be lower near 
parents because of disproportionately high mortality associated with seed preda?
tion, herbivory, and pathogens. CLARK and CLARK (1984) reviewed 24 studies on 
this topic published between 1970 and 1984 and found general support for either 
density-dependence or distance-dependence in progeny mortality. Additional sup?
port for the CONNELL-JANZEN hypothesis was provided by AUGSPURGER (1984a) 
who found that seedlings of eight out of nine wind-dispersed species in Panama 
61 
62 D. F. WHIGHAM and E. CABRERA CANO 
had higher survival rates away from parents. Others have found different results. 
KITAJIMA and AUGSPURGER (1989) found that, seedlings of Tachigalia versicolor 
had a higher survival rate with increasing distance from the parent, but that few 
seedlings persisted beyond 40 m from parents. SCHUPP (1988) found that seed 
survival of Faramea occidentalis was greatest beyond the canopy of the parent but 
that mortality of established seedlings was similar beneath and away from parents. 
HUBBELL (1980) also noted that juveniles can reach maturity immediately adjacent 
to an adult of the same species. 
A number of investigators have examined the role of tree gaps as sites for suc?
cessful establishment and growth of seedlings (SCHUPP 1988, AUGSPURGER 1983, 
AUGSPURGER 1984b, AUGSPURGER and KELLY 1984, BROKAW 1985, BROKAW 
1987, DENDLOW et al. 1990, DE STEVEN and PUTZ 1984, UHL et al. 1988). In 
most of these studies, seedling survival and growth rates were greater in tree gaps. 
SCHUPP (1988), however, found that predation was higher in gaps and concluded 
that long-term seedling survival might be best in the forest understory. 
Given the high diversity of trees in most tropical forests compared to the few 
species that have been studied, it is not surprising that a variety of relationships 
between adults and their offspring have been described. More studies are needed 
before generalizations can be made either about the JANZEN-CONNELL hypothesis 
or about the importance of tree gaps to the dynamics of humid and dry tropical 
forests. 
In this paper we address the issues of parent-offspring growth and survival by 
using results of a six year study of Myrcianthes fragrans (Sw.) MCVAUGH (Myr?
taceae) in a dry tropical forest in the Yucatan Peninsula of Mexico. Myrcianthes 
fragrans occurs from northern South America through the West Indies and into 
southern Florida (MCVAUGH 1963). We have three objectives: 1) to evaluate fac?
tors affecting seedling survival by: a. characterizing distance relationships between 
adults and between adults and smaller individuals, and by b. evaluating the growth 
and mortality of seedlings in long-term plots located at three distances from par?
ent trees, 2) to determine the relationship between seedling growth and soil depth. 
Soils at our study site are shallow, usually less than 20 cm, and we hypothesized 
that seedling growth would be negatively related to soil depth, and 3) we address 
the hypothesis that seedlings respond positively to canopy opening using measure?
ments of seedling growth and mortality prior to and following complete canopy 
defoliation caused by Hurricane Gilbert in September, 1988. 
4.2 Study site 
The studies were conducted in a dry tropical forest ("selva mediana subperennifo?
lia" (MIRANDA 1958) on the Yucatan Peninsula. The site, hereafter called Rancho 
San Felipe, is located approximately 10 km south of the village of Puerto Morelos 
(20 0 49' N, 87 0 7' W) in the Mexican state of Quintana Roo. Canopy height at the 
site varies between 15 and 25 meters. Twelve long-term study plots (40 x 40 m) 
Survival and growth beneath and near parents 63 
were established in February 1984. All trees with a diameter (DBH) greater than 
10 cm were tagged and identified in each plot. The twelve plots contained 1508 
individual trees that belonged to 79 species. Nine species (M anilkara zapota (L.) 
VAN ROYER, Talisia olivaeformia (HBK.) RADLK., Gymnanthes lucida (SWARTZ), 
Brosimum alicastrum SWARTZ, Drypetes laterifiora (SWARTZ) DRUG and URBAN, 
Sapindus saponaria L., Coccoloba diversifolia JACQ, Beaucarnea pliabilis (BAKER) 
LUNDELL), and Myrcianthes fragrans accounted for approximately 60% of all in?
dividuals. Thirty~eight Myrcianthes trees occurred in eleven of the plots and its 
density ranged from 1 to 6 per 1600 m 2 . DBH of MYl'cianthes ranged from 14.7 to 
69.2 cm (mean 36.4 cm; SD?14.0). 
There are no long~term climatic data from Rancho San Felipe but relevant 
data are available from the Cancun airport (25 km), Felipe Carillo Puerto (185 
km) (LOPEZ ORNAT 1983), and Valladolid (135 km) (WALTER and LIETH 1967). 
The climate is strongly seasonal with a distinct period of low precipitation that 
usually lasts from November through May (Fig. 4.1). Annual precipitation is ap?
proximately 1100 mm, but the amount and seasonal distribution are highly variable 
(WHIGHAM et al. (in press)). The average annual temperature at Valladolid and 
Felipe Carillo Puerto is 25?C and the monthly mean varies from approximately 26 
to 22?C. Maximum temperatures occur between May and August (ca. 40~33?C) 
and minimum temperatures between January and March (ca. 18~8?C). 
The shallow azonal soil, classified as a litosol~rendzina and derived from weath?
ered Miocene and Pliocene parent material (WEST 1964, BACK and HANSHAW 
1970), occurs in pockets between exposed limestone rocks. The soil has an almost 
neutral pH (7.1 ? 0.1) and the organic matter content is high (51.5 ? 5.1%). 
A verage light conditions in the forest understory were determined prior to Hur?
ricane Gilbert. Shaded areas in long~term seedling plots (see page 65) received an 
average of 1.0 ? .5% (1 standard error) of full sunlight (N = 34; PAR measurements 
made in a nearby field and at the soil surface in seedling plots during cloudless 
conditions on 16 February, 1985 between 11AM and 12PM using a LI~COR Model 
185B light meter). During the same sampling period, sunflecks (N= 14) in the 
seedling plots received an average of 12.3 ? 11.4% (Range: 40.0~1.7%) of full sun?
light. No light measurements were made immediately after Hurricane Gilbert but 
the canopy was completely defoliated and light intensities were accordingly quite 
high (Ingrid Olmsted, personal communication). 
4.3 Methods 
Distances between juveniles, saplings, and trees. We measured interplant 
distances between all Myrcianthes individuals> 10 cm DBH in the 1600 m 2 plots. 
We also measured the distances between each individual tree and all juveniles 
(plants less than 1 m in height), saplings (plants greater than 1 m in height and 
less than 10 cm DBH), and trees within a ten meter radius. A ten meter sampling 
radius was chosen to represent areas beneath (0~5 m) and beyond (5~10 m) the 
64 D. F. WHIGHAM and E. CABRERA CANO 
VALLADOLID (22m) 25.0C 1181mm 
5 100 
-
40 80 E 
-
E 
u 
-
- c 
(!) 30 60 0 ~ 
-
:J a +-
-a 0. ~ 20 40 Q) (J 
0. Q) E ~ 
Q) 20 Q.. I- 10 
J F M A M J JA SON DJ 
Figure 4.1: Climate diagram (WALTER and LIETH 1967) for Valladolid (elevation 
22 meters). The long-term mean for temperature and precipitation are given in 
the upper right corner of the diagram. The relationship between temperature and 
precipitation suggests periods of excess or deficit water conditions. The dotted area 
is a period of relative drought when the precipitation curve falls below the temper?
ature curve. The vertically shaded area is the humid season when the precipitation 
curve is above the temperature curve. The black area represents the part of the 
year when precipitation exceeds 100 mm per month and the scale is reduced to 
1/10. 
canopy of each tree, and to include the distribution of virtually all seedlings. Dif?
ferences between the density of Myrcianthes juveniles, shrubs, and trees within the 
10 m sampling radius were compared with ANOVA and the Tukey-Kramer test 
(WILKINSON 1989). 
Survival and growth beneath and near parerJts 65 
Seedlings in long-term plots. In February 1984, we established nine plots 
(1 x 1 m) around each of five trees. Plots, three per transect, were positioned 
between 1-2, 5-6, and 10-11 meters from the boles of the putative source trees. 
These plots will hereafter be referred to as 1, 5, and 10 meter distances. A cohort of 
Myrcianthes seedlings had germinated during the pervious 2-3 months. Cotyledons 
were still attached and most seedlings had produced the first. true leaves. We 
counted the number of seedlings in each plot and fitted their stems with small 
numbered aluminum tags. In most plots we tagged all seedlings but in plots with 
more than 100 individuals, only a subset (approximately 50%) of the seedlings 
were tagged. Since 1984, the plots have been censused yearly at approximately 
the same time to determine the number and height of live seedlings. Height and 
density data were analyzed with a nested ANOVA model (quadrat nested within 
tree and distance) and means compared with the Tukey test (JOYNER (ed.) 1985). 
Seedling survival at 1, 5, and 10 meters was compared with the Peto and Peto's 
Logrank Test (PYKE and THOMPSON 1986). 
Depth of the soil-litter layer. The potential rooting depth available to seed?
lings was evaluated by measuring the depth (em) of the litter-soil layer at 10 em 
from the trees and then at 50 em intervals along three randomly located 10 m 
transects per tree (Total = 15 transects). The litter-soil layer was defined as the 
dist.ance between the top of the litter layer and solid rock. Depth was measured 
by inserting a wooden stick into the soil and marking it at the top of the soil-litter 
layer after the stick had reached solid rock. The distance from the bottom of the 
stick to the mark was then measured. Statistical comparisons of the depth of the 
litter-soil layer at different distances from parent trees were made using ANOVA 
(quadrats nested within trees) and means were compared with the Tukey-Kramer 
test (WILKINSON 1989). 
Seedling growth and IIlicrohabitat characteristics. In February 1990, we 
harvested 300 seedlings from the 1984 cohort. Twenty seedlings were collected along 
each of three randomly located 10 m transects at each of the five trees described 
on page 65. The following data were collected in the field: 1. distance (m) between 
parent and seedling, 2. qualitative assessments of light conditions were made by 
evaluating the percent canopy above each seedling. Three categories were used: A. 
shade - seedling judged to be completely shaded by leaves of overhead vegetation, 
B. partial shade - seedlings mostly shaded by overhead vegetation but judged 
to receive some direct light during the day, C. sun - seedlings covered by little 
overhead vegetation and judged to receive large amounts of direct light during 
the day, 3. depth of soil (em) immediately beneath the seedling, 4. number of 
leaves and branches, 5. shoot height (em), 6. length (em) of the root divided into 
vertical and/or horizontal components. The latter division was necessary because 
Myrcianthes roots begin to grow horizontally after they have reached the bottom 
of the soil profile. Plants were dried to constant weight in the sun, and biomass (g) 
66 
10000 
NE 1000 
"-! 
w 
c 
0 100 Ci. 
"0 
~ 
w 10 c 
'" Cl
01 
D. F. WHIGHAM and E. CABRERA CANO 
300 
Eo 
! 200 
w 
c 
'" u 
0' 
C 
U 
'" 
'" (/) 
100 
b 
5 10 
Distance from parent (m) 
Figure 4.2: Density (log scale) of Myrcianthes fragrans seedlings, juveniles, 
saplings, and trees. Values are means ? 1 standard error. 
(a) Seedling data are from the long-term plots (see page 65). Juvenile, sapling, 
and tree data represent densities within 10 meters of individual trees (see page 63). 
(b) Mean density (? 1 standard error) of seedlings in long-term study plots. Means 
that are not different at P < .05 share the same letter. 
determined for leaves, shoots, branches, vertical roots, and horizontal roots. Using 
a nested (transect nested within tree) ANOVA model, the data were analyzed using 
the GLM procedure in SAS (JOYNER (ed.) 1985). 
4.4 Results 
Density and distance relationships The mean and standard deviation of the 
distance between trees in the 11 plots was 22.8? 11.2 meters (N = 62; max= 51.3 m, 
min= 1.9 m). Within 10 m of each tree, the average distances to juveniles (7.0? 1.9), 
shrub-sized individuals (6.4 ? 2.0), and tree-sized individuals (6.2 ? 1.5) were not 
statistically different. Densities were, however, significantly different (P< .0001) 
with juveniles> shrubs = trees (Fig. 4.2a). Within the long-term plots, seedling 
densities in 1984 differed significantly (P < 0.01) as a function of distance to the 
source tree, with 1> 5 = 10 m (Fig. 4.2b). 
Growth and mortality in long-term plots. The shape of the mortality curves 
were similar for all three distances but the slopes were significantly different (P 
< .001) and survivorship was greater at 5 and 10 m (Fig. 4.3). After almost 60 
Survival and growth beneath and near parents 67 
100 
80 
01 
C 
c 10m 
0 60 \ E Q) 
... 
c 40 Q) 
u 
... 
Q) 
0.. 
20 
10 20 30 40 50 60 
Months 
Figure 4.3: Survivorship curves for seedlings in long~term study plots at 1, 5, and 
10 meters from parent trees. The vertical dashed line indicates when Hurricane 
Gilbert struck the study area in September, 1988. 
months, only 4.5 ? 1.2% of the original cohort of seedlings were alive at 1 III, 
compared with 19.5 ? 3.8 and 19.2 ? 3.8% at 5 and 10 m respectively. At each 
census, we were able to locate almost all seedlings and it was apparent that few 
had suffered from herbivore damage. Most dead seedlings were upright and had 
dried intact leaves attached. Most seedlings appeared to have died from drought 
stress although a few had been physically damaged from falling debris and a few 
others appeared to have died from shading as a result of being covered by fallen 
bark. 
From April, 1985 to February, 1988, mean seedling height increased less than 1 
cm at 1 and 5 m and only about 1.5 cm at 10 m (Fig. 4.4). Average annual height 
growth prior to the hurricane did not differ significantly between the three distances 
(Fig. 4.5). Annual growth rates increased significantly following the hurricane but 
68 D. F. WHIGHAM and E. CABRERA CANO 
15 r 
5m---.. 
E 
10 u 
.J::: 
0' 
CI) 
.J::: 
0' 
c 
u 
CI) 5 I) 
CI) 
(f) 
10 20 30 40 50 60 
Months 
Figure 4.4: Height (cm ? 1 standard error) of seedlings at 1,5 and 10 meters from 
parent plots. The vertical dashed line indicates when Hurricane Gilbert struck the 
study area in September, 1988. 
the three distances were not significantly different even though the growth rate was 
less at 1 m (Fig. 4.5). 
Soil depth and distance from parent trees. Large amounts of bark exfoliates 
from the trunk and large branches of Myrcianthes and a mound of litter develops 
that extends outward from the base of the tree to about 2.5 m (Fig. 4.6). Depth 
of the litter~soil layer at 0.1 m was significantly more (P < 0.01) than greater 
distances. Depth of the litter~soillayer was also significantly greater at 0.5, and 1 
m but there were no significant differences beyond 1.5 m where the soil depth was 
approximately 10 cm. 
Survival and growth beneath and near parents 69 
6 
? Pre-hurricane b 
D Past-hurricane 
5 
E b 
-=: 
~ 4 
3 
0 
~ 
0> 
>. 3 
~ 
0 
Q) 
>. 
2 
c 
0 
Q) 
::?: 
5 10 
Distance (m) 
Figure 4.5: Average annual growth in height of Myrcianlhes seedlings before (black 
bars: March 1984-March 1987) and after (Open bars: September 1988-March 
1990) Hurricane Gilbert. Values are means ? 1 standard error. 
60 
E 50 
,S 
0 40 
'" 
2 30 
a 20 
L 
0. 
OJ 
0 10 
2 4 6 8 10 
Distance from parent (m) 
Figure 4.6: Depth of the litter-soil layer as a function of distance from trees. Values 
are means ? 1 standard error. Methods for sampling see page 65. 
70 D. F. WHIGHAM and E. CABRERA CANO 
20------------------~ 1.0 __ -------------, 
Number of leaves Leaf weight (9) 
0.8 
15 
0.6 
10 
OA 
5 0.2 
0.5.__---------------, 2.0 __ -------------, 
Shoot weight (g) Total weight (9) 
OA 1.5 
0.3 
1.0 
02 
05 
0.1 
20--------------------, 20--------------------, 
Vertical length of root (em) Plant height (em) 
15 15 
10 10 
5 5 
04 __ ------------------- 0.5 .-------------------, 
weight of vertical root (g) Root:Shoot ratio 
OA 
0.3 
0.3 
0.2 
0.2 
0.1 0.1 
o Open o Partially shaded ~ Shaded 
Figure 4.7: Growth parameters for Myrcianthes seedlings in shaded, partially 
shaded, and open microhabitats. Values are means ? 1 standard error. Micro?
habitat designations are as shown at the bottom of the figure. 
Survival and growth beneath and near parents 71 
Table 4.1: Results of linear regressIOn analysis of growth variables against soil 
depth. N= 300. 
VARIABLE R2 INTERCEPT SLOPE 
Length of vertical root (em) .358 5.678 .560 
Weight of vertical root (g) .090 0.074 .005 
Number of leaves .058 10.475 .188 
Length of horizontal root (em) .053 4.381 -.162 
Plant height (em) .036 10.990 .158 
N umber of branches .035 .144 .024 
Weight of horizontal root (em) .035 .050 -.002 
Weight of leaves (g) .030 .395 .013 
Weight of branches (g) .023 .157 .005 
Root:shoot ratio .004 .266 -.001 
Seedling growth and microhabitat characteristics. After 6 years of growth, 
soil depth showed almost no statistically significant relationship to any of the vari?
ables measured on harvested seedlings (Table 4.1). Length of the vertical root was 
the only variable for which soil depth accounted for more than 10 percent of the 
vanance. 
Although there were no significant differences in seedling growth rates with 
distance, there were significant microhabitat effects on growth following Hurricane 
Gilbert. In comparison with seedlings in open and partially shaded microhabitats, 
seedlings in shaded microhabitats were significantly shorter (P< .001), had fewer 
leaves (P< .008), less leaf biomass (P< .009), shorter and smaller vertical roots 
(P< .046), and less shoot and total biomass (P< .009) (Fig. 4.7). 
4.5 Discussion 
Initially, seedling densities were highest beneath parent trees. Seedling mortality 
was greatest near the parent tree as predicted by the JANZEN-CONNELL hypothesis. 
However, herbivory or pathogens probably were not responsible for the higher levels 
of mortality observed near parent trees. As indicated, almost all marked seedlings 
appeared to have died from physical damage (crushing or breaking of the stems), 
shading, drought stress or some combination of those factors. 
AIDE (1987) and CLARK and CLARK (1989) found that seedling mortality due 
to physical damage was common in moist tropical forests. Mortality due to direct 
physical damage was of minimal importance in this study, even during the most 
intense hurricane ever measured in the Western Hemisphere. We found very few 
72 D. F. WHIGHAM and E. CABRERA CANO 
seedlings that had died as a result of direct impact (i. e., broken stems) from fallen 
branches. A few seedlings were killed by direct impact from fallen objects before 
and after the hurricane but most appeared to die from shading that resulted from 
being covered by exfoliated bark (before the hurricane) or by woody debris (after 
the hurricane). 
Although we have no direct measurements of seasonal variation in available 
soil water, we believe that drought was the primary cause of mortality. The soils 
are very shallow (Fig. 4.6) and contain very little water, especially during the dry 
season when many plants appear to be water stressed. These results contrast with 
AIDE and ZIMMERMAN (1990) who suggested that seedling growth in the forest 
understory may be enhanced during dry periods because of increased light and soil 
moisture due to less evapotranspiration from leafless canopy trees. Other factors 
such as nutrient availability may also be important but soil depth is not as it was 
related to only one seedling growth parameter, length of the vertical root. 
Light also appeared to be an important factor that limited seedling growth and 
perhaps survivorship (AUGSPURGER 1984a, AUGSPURGER 1984b, BROKAW 1985, 
DENSLOW et al. 1990, UHL et al. 1988, CLARK and CLARK 1989, DENSLOW 1987). 
Prior to the hurricane, light levels were very low (mostly < 1%) in the forest un?
derstory and the average seedling height increased very little during the 40 months 
before the hurricane. Seedling growth measured after the hurricane indicated that 
seedling in shaded microhabitats grew significantly less than seedlings growing in 
open and partially shaded areas. These findings agree with those of others (e. g., 
UHL et al. 1988, AIDE and ZIMMERMAN 1990) who have found that variations 
in light conditions played an important role in controlling the growth of woody 
seedlings in moist tropical forests. 
Following the seedlings stage, patterns of mortality must remain spatially con?
stant for many years. We found surprisingly constant mean distances (ca. 7 m) 
from parent trees to juveniles and saplings. Survival to tree size near parent trees 
can occur (HUBBELL 1980) in tropical forests but is probably uncommon. The 
average distance between adults exceeded 20 meters and, within the 1600 m 2 plots, 
only three of the sixty-two nearest neighbor interplant distances were less than 5 
meters. 
We do not know the causes of mortality for individuals larger than seedlings 
but hurricanes may be more important than all other factors combined. Of 38 
Myrcianthes trees in the eleven 1600 m 2 plots, one (2.6%) died (cause of mortality 
unknown) between 1984-1988, compared with 12 (31.6%) as a result of Hurricane 
Gilbert. Trees of all species in the forest were defoliated and suffered varying 
degrees of canopy damage during the hurricane. For Myrcianthes, two large (DBH 
> 50 cm) individuals were uprooted during the hurricane, two completely defoliated 
trees never sprouted, and the other eight individuals died even though they had 
initially sprouted. 
Perhaps more damaging to trees, however, were widespread fires which occurred 
during the dry season after the hurricane. Although Rancho San Felipe itself was 
Survival and growth beneath and near parents 73 
mostly protected, two small areas burned when fire jumped the fire lane that had 
been cut around the ranch. In one of these areas, we established four 1600 m 2 
and sampled all trees greater than 10 cm DBII. The number of trees in the 
four plots averaged 93 ? 15 (1 standard error) and mortality averaged S9.3 ?3.9%, 
mortality level typical of other areas in northern Yucatan that burned during 
the same period (Ingrid Olmsted, personal communication). All 18 Myrcianthes 
trees in the plots were killed and we found only two surviving saplings in one of 
seventy-two plots (each plot was 4 x 4 m) sampled within the four larger plots. 
Clearly, the distribution and abundance of Myrcianthes can be influenced by both 
hurricanes and subsequent fires. 
In summary, we found evidence to support almost every hypothesis about the 
distribution, abundance, and growth of seedlings and trees. Data from long-term 
plots support the hypothesis that mortality is greatest near parents. Mortality of 
Myrcianthes seedlings was greatest within one meter of parent trees but the primary 
cause did not appear to be herbivory and disease but physical damage and shading 
caused by exfoliated bark. These findings support the hypothesis that direct and 
indirect damage caused by fallen objects may be one of the most important causes 
of seedling mortality (AIDE 1987, CLARK and CLARK 19S9). Between-tree distance 
data support the hypothesis that few plants survive beneath their parent but, like 
HUBBELL (1980), we found a few individuals growing close to each other. 
Our data suggest that Myrcianthes has adopted a "sit-and-wait" approach to 
reproduction and growth. Individuals do not produce seeds regularly (ca. only 
once in the six years of this study) and seedlings growth in the understory is 
very slow and appears to be primarily light limited. If seedlings can survive long 
enough, however, a canopy disturbance (i. e., single-tree death or large disturbances 
such as those caused by the hurricane) may occur and subsequent growth may no 
longer be as strongly light-limited. In the case of single-tree disturbances, we 
agree with UHL et al. (198S) who suggested that individuals that are recruited 
into gaps or are present as seedlings would only reach the canopy after multiple 
disturbances. Seedlings, juveniles, and saplings would have a greater probability of 
reaching the sub-canopy or canopy following a single large-scale disturbance that 
defoliates and causes extensive damage to trees. Even in this instance, however, 
it seems likely that only larger (i. e. small tree-sized individuals) would reach 
the canopy after a single disturbance event because the canopy reforms rather 
quickly (unpublished data). Perhaps most important, this study demonstrates that 
endogenous (e. g., seed dispersal, seedling predation and herbivory etc.), factors 
that control the distribution patterns of trees in dry tropical forests may be less 
important than exogenous factors (e. g., hurricanes, fires). Because most seedling 
studies have been conducted in moist tropical forests, many more species-specific 
comparisons of this sort are needed before we can begin to understand the dynamics 
of tree populations in dry tropical forests. 
74 D. F. WHIGHAM and E. CABRERA CANO 
Acknowledgments 
Jay O'Neill and Elizabeth Ley helped count and measure seedlings in the perma?
nent plots. Volunteers with the Smithsonian Expedition Program (Owen Lavin, 
Sol Courtman, Sherley Higuera, Margery Plymire, Shirley and James Virosco, and 
William McNiven) helped harvest seedlings and measure interplant distances. Jess 
Zimmerman, Jay O'Neill, Julie Denslow, and Jim Lynch provided comments on 
drafts of the manuscript. Financial support was provided by the World Wildlife 
Fund, Smithsonian Environmental Sciences Program, and Smithsonian Interna?
tional Environmental Sciences program. The logistical support of Patricia Zugasty 
Towle, Felipe Sanchez Roman, and Ingrid Olmsted are appreciated. 
Bibliography 
AIDE, T. M., 1987, Limbfalls: a major cause of sapling mortality for tropical forest 
plants. 
Biotropica 19, 284-285. 
AIDE, T. M.; ZIMMERMAN, J. K., 1990 Patterns of insect herbivory, growth, and 
survivorship in juveniles of a neotropicalliana. 
Ecology 71, 1412-1421. 
AUGSPURGER, C. K., 1983, Offspring recruitment around tropical trees: changes 
in cohort distance with time. 
Oikos 40, 189-196. 
AUGSPURGER, C. K., 1984a, Light requirements of neotropical tree seedlings: a 
comparative study of growth and survival. 
J. Ecol. 72, 777-796. 
AUGSPURGER, C. K., 1984b, Seedling survival of tropical tree species: interactions 
of dispersal distance, light-gaps, and pathogens. 
Ecology 65, 1705-1712. 
AUGSPURGER, C. K.; KELLY, C. K., 1984, Pathogen mortality of tropical tree 
seedlings: experimental studies of the effects of dispersal distance, seedling density, 
and light conditions. 
Oecologia 61,211-217. 
BACK, W.; HANSHAW, B. B., 1970, Comparison of chemical hydrogeology of the 
carbonate peninsulas of Florida and Yucatan . 
.1. Hydro!. 10, 330-368. 
BROKAW, N., 1985, Gap-phase regeneration in a tropical forest. 
Ecology 66, 682-687. 
Survival and growth beneath and near parents 75 
BROKAW, N., 1987, Gap-phase regeneration of three pIOneer tree speCies III a 
tropical forest. 
J. Ecol. 75,9-19. 
CLARK, D. A.; CLARK, D. B., 1984, Spacing dynamics of a tropical forest tree: 
evaluation of the Janzen-Connell model. 
Am. Nat. 124, 769-788. 
CLARK, D. A.; CLARK, D. B., 1989, The role of physical damage in the seedling 
mortality regime of a neotropical forest. 
Oikos 55, 225-230. 
CONNELL, J. H., 1971, On the role of natural enemies in preventing competitive 
exclusion in some marine animals and in rain forest trees. 
In: DEN BOER, P. J.; GRADWELL, G. R. (eds.), Dynamics of populations. Pro?
ceedings of the Advanced Study Institute on dynamics of numbers in populations, 
Oosterbeek, 1970, Centre for Agricultural Publishing and Documentation, Wa?
geningen, pp. 298-310. 
DENSLOW, J. S., 1987, Tropical rainforest gaps and tree species diversity. 
Ann. Rev. Ecol. Syst. 18,431-451. 
DENSLOW, J. S.; SCHULTZ, J. C.; VITOUSEK, P. M., STRAIN, R R., 1990, 
Growth responses of tropical shrubs to treefall gap environments. 
Ecology 71, 165-179. 
DE STEVEN, D.; PUTZ, F. E., 1984, Impact of mammals on early recruitment of 
a tropical tree, Dipteryx panamensis, in Panama. 
Oikos 43,207-216. 
HUBBELL, S. P., 1980, Tree dispersion, abundance, and diversity in a tropical dry 
forest. 
Science 203, 1299-1309. 
JANZEN, D. H., 1971, Herbivores and the number of tree species in tropical forests. 
Am. Nat. 104, 501-528. 
JOYNER, S. P. (ed.), 1985, SAS/STAT guide for personal computers. Version 6 
edition, SAS Institute Inc., Cary, NC. 
KITAJIMA, K.; AUGSPURGER, C. K., 1989, Seed and seedling ecology of a mono?
carpic tropical tree, Tachigalia versicolor. 
Ecology 70, 1102-1114. 
LOPEZ ORNAT, A., 1983, Localisacion y medio fisico, 
In: SIAN KA' AN, Estudios preliminares de una zona en Quintana Roo propuesta 
como Reserva de la Biosfera. Centro de Investigaciones de quintana Roo, A. C. 
Puerto Morelos, Quintana Roo, Mexico, pp. 20-49. 
76 D. F. WHIGHAM and E. CABRERA CANO 
MCVAUGH, R., 1963, Flora of Guatemala. 
Fieldiana Botany 24, 377~379. 
MIRANDA, F., 1958, Vegetaci6n de la peninsula Yucateco, 
In: BELTRAN, E. (ed.), Los recursos naturales del sureste y su aprovechamiento. 
Volume II, Instituto Mexicana Recursos Nautrales Ren. Mexico, D. F. , Mexico, 
pp. 215~227. 
PYKE, D. A.; THOMPSON, .1. N., 1986, Statistical analysis of survival and removal 
rate experiments. 
Ecology 67, 240~245. 
SCHUPP, E. W., 1988, Seed and early seedling predation in the forest understory 
and in treefall gaps. 
Gikos 51, 7l~78. 
UHL, C.; CLARK, K.; DEZZEO, N.; MAQUIRINO, P., 1988, Vegetation dynamics 
in Amazonian treefall gaps. 
Ecology 69, 751~763. 
WALTER, H.; LIETH, H., 1967, Klimadiagramm-Weltatlas, VEB. Gustav Fischer 
Verlag, Jena. 
WEST, R. C., 1964, Surface configuration and associated geology of Middle Amer-
lca, 
In: WEST, R. C. (ed.), Handbook of Middle American Indians Vol. 1, University 
of Texas Press, Austin, pp. 33~83. 
WHIGHAM, D. F.; ZUGASTY TOWLE, P.; CABRERA CANO, E.; O'NEILL, J.; 
LEY, E., The effect of annual variation in precipitation on growth and litter pro?
duction in a tropical dry forest in the Yucatan of Mexico. 
Tropical Ecology (in press). 
WILKINSON, L., 1989, SYSTAT: The system for statistics. 
Evanston, IL.