2798
Ecology, 83(10), 2002, pp. 2798?2807
q 2002 by the Ecological Society of America
GERMINATION ECOLOGY OF NEOTROPICAL PIONEERS: INTERACTING
EFFECTS OF ENVIRONMENTAL CONDITIONS AND SEED SIZE
T. R. H. PEARSON,1 D. F. R. P. BURSLEM,1 C. E. MULLINS,1 AND J. W. DALLING2,3
1Department of Plant and Soil Science, University of Aberdeen, Cruickshank Building, St. Machar Drive,
Aberdeen AB24 3UU, UK
2Department of Plant Biology, University of Illinois, Urbana, Illinois 61801 USA
3Smithsonian Tropical Research Institute, Unit 0948, APO AA 34002 USA
Abstract. Germination provides many potentially unrecognized sources of variation in
the regeneration niche. In this study we relate germination requirements and seed size for
16 species of pioneer trees to microclimatic conditions present in gaps in semi-deciduous
rain forest in Panama. We found that, whereas increased duration of direct irradiance can
be an effective indicator of the presence of a canopy gap across all scales of canopy
openness, diel fluctuations in soil temperature effectively discriminate both understory sites
and small gaps (25 m2) from larger gaps. Germination response was significantly related
to seed size. Small-seeded species (seed mass ,2 mg) showed significantly greater ger-
mination in response to irradiance of 22.3 mmol?m22?s21 than in complete darkness. Their
germination was unaffected by an increasing magnitude of diel temperature fluctuation up
to a species-specific threshold, above which it declined. Large-seeded species (seed mass
.2 mg) germinated equally in light and darkness (with one exception) and either showed
a positive germination response to an increasing magnitude of temperature fluctuation (four
species) or no significant response (four species). The maximum seed burial depth from
which seedlings could emerge successfully was strongly positively associated with seed
mass. We conclude that photoblastic germination of tropical pioneer trees results in small-
seeded species germinating in gaps only when seeds are located in microsites that are
suitable for seedling emergence. A positive germination response to increasing temperature
fluctuation can stimulate germination of larger-seeded species in larger gaps and when they
are buried beneath an opaque soil or litter layer. Therefore, seed size differences constrain
the physiological mechanisms of canopy gap detection in tropical pioneer trees and might
contribute to observed differences in the distribution of adult plants in relation to canopy
gap size.
Key words: Barro Colorado Island, Panama; emergence potential; germination cue; niche dif-
ferentiation; pioneer species; tropical diversity.
INTRODUCTION
Many processes contribute to the maintenance of tree
diversity in tropical forests. Recently studied processes
include density and frequency-dependent recruitment
success (e.g., Harms et al. 2000), dispersal limitation
(e.g., Hubbell et al. 1999), and niche differentiation or
resource partitioning during early recruitment (Grubb
1996). Although considerable attention has been given
to seedling responses to resource availability, partic-
ularly irradiance (e.g., Metcalfe and Grubb 1997), the
roles of other potential differences between species
have not been examined in detail.
Here we examine the importance for pioneer species
of different germination responses to the environmental
signals that indicate the presence of a canopy gap. Trop-
ical pioneer trees have been defined as species requiring
high irradiance for establishment and growth to ma-
turity (Swaine and Whitmore 1988). Germination re-
quirements are likely to be under strong selection pres-
Manuscript received 30 October 2000; revised and accepted
15 February 2002; final version received 13 March 2002.
sure as early colonization leads to early growth in a
situation where larger plants have a disproportional ad-
vantage in competition for irradiance (Weiner 1990).
Studies of cues to germination in tropical pioneers
have examined irradiance, but largely neglected alter-
native cues and the role of seed size. Although many
pioneers are sensitive to irradiance, published results
are not representative of pioneers in general because
work has been biased toward small-seeded species
(Va?zquez-Yanes and Orozco-Segovia 1993). Temper-
ature has rarely been considered as a germination cue.
Va?zquez-Yanes and Orozco-Segovia (1982) showed
maximum germination in the pioneer tree Heliocarpus
donnell-smithii when seeds were exposed to a 158C
daily temperature fluctuation. Va?zquez-Yanes and
Pe?rez-Garc??a (1976) examined another pioneer, Och-
roma lagopus, and found that a marked temperature
fluctuation made the seed coat permeable; this allowed
imbibition and subsequent germination. None of this
work has been related to gap environments and micro-
climatic variation with gap size, gap position, or soil
depth.
October 2002 2799GERMINATION OF NEOTROPICAL PIONEERS
TABLE 1. Details of the 18 species of pioneer trees studied including seed properties and experimental treatments.
Species and family
Seed
mass (mg)
Seed coat
thickness?
(mm)
Dis-
persal?
Experi-
mental
treatment?
Seed
pretreatments
Piper peltatum, Piperaceae
Miconia argentea, Melastomataceae
Piper dilatatum, Piperaceae
Alseis blackiana, Rubiaceae
Cecropia peltata, Moraceae
Cecropia obtusifolia, Moraceae
Cecropia insignis, Moraceae
Luehea seemannii, Tiliaceae
0.04
0.08
0.10
0.12
0.58
0.59
0.68
1.90
9 6 0.1
33 6 1.1
18 6 1.5
???
84 6 4.0
83 6 4.3
67 6 4.0
83 6 3.4
A
A
A
W
A
A
A
W
I, T
I, T, SD
I, T
I
I, T
I, T
I, T, SD
I, T, SD
???
none
???
???
???
???
none
hot water (70 8C, 2 min)
Solanum hayesii, Solanaceae
Trichospermum mexicanum, Tiliaceae
Trema micrantha, Ulmaceae
Guazuma ulmifolia, Sterculiaceae
Apeiba tibourbou, Tiliaceae
2.40
2.70
3.90
4.60
6.10
59 6 2.8
76 6 3.8
???
105 6 1.7
90 6 3.4
A
W
A
A
A
I, T, SD
I, T, SD
I
I, T, SD
I, T
none
hot water (70 8C, 2 min)
???
hot water (70 8C, 10 min)
???
Ochroma pyramidale, Bombacaceae
Apeiba membranacea, Tiliaceae
Cochlospermum vitifolium, Bixaceae
Ceiba pentandra, Bombacaceae
Pseudobombax septenatum, Bombacaceae
6.60
14.20
25.40
70.00
88.30
160 6 5.2
95 6 5.9
???
???
202 6 4.5
W
A
W
W
W
I, T, SD
I, T, SD
I, T
T
I
water (100 8C, 30 s)
hot water (70 8C, 2 min)
???
???
???
? Treatments of 10 seeds per species (mean 6 1 SE).
? A, animal dispersed; W, wind dispersed.
? Abbreviations are: I, germination responses to irradiance; T, germination response to temperature; SD, soil depth and
seedling emergence.
Consideration must be given to the capacity of seeds
to emerge and the penetration of cues to germination
through different soil depths. In the South African fyn-
bos very small-seeded shrubs lack the necessary re-
sources or hydraulic capability to germinate and
emerge from more than superficial depths, while large
seeds germinate and emerge from greater depths in the
soil profile (Bond et al. 1999). Similar data for tropical
rain forest plants are not available. However, irradiance
is only able to penetrate 4?5 mm into the soil in phys-
iologically significant quantities (Tester and Morris
1987) and is attenuated by leaf litter (Va?zquez-Yanes
et al. 1990). Therefore we suggest that irradiance is
only an appropriate gap cue for very small-seeded spe-
cies. For larger seeded species additional cues are need-
ed if germination is to occur across the full range of
microsites that permit emergence and growth.
Temperature may provide an effective germination
cue for larger seeded pioneers with sufficient seed re-
sources to emerge from below more than a few mil-
limeters of soil. In contrast to irradiance, diel variation
in temperature in canopy gaps may be manifest at soil
depths of several centimeters (Raich and Gong 1990).
Consequently, we hypothesize that irradiance should
be an appropriate cue for germination of a pioneer only
if it has very small seeds, and larger seeded species
should germinate in response to an increased magni-
tude of temperature fluctuation. The few published
measurements of soil temperature in tropical forest
gaps and understory suggest that the magnitude of tem-
perature fluctuation increases with increasing gap size
for an equivalent soil depth (Va?zquez-Yanes and Oroz-
co-Segovia 1982, Raich and Gong 1990). Therefore,
species that require large gaps for seedling establish-
ment should be likely to show a positive germination
response to increasing magnitude of temperature fluc-
tuation.
The test species for this research were drawn from
the pioneer tree species of Barro Colorado Island
(BCI), Panama. On BCI, pioneers comprise a group of
30?40 species varying in seed size over four orders of
magnitude (Dalling et al. 1997). We used representative
species of pioneers to address the following questions:
(1) Does a shift from light-cued to temperature-cued
seed germination occur with increasing seed size
among pioneer species? (2) Are these results mirrored
in seedling emergence from different soil depths? (3)
What microsite conditions exist in different sized gaps
and therefore in what sites are species with different
cues most likely to germinate?
METHODS
Study site and species
Barro Colorado Island (BCI) lies in an area of sea-
sonally moist tropical forest, in the Republic of Pan-
ama. The mean annual rainfall is 2700 mm/yr, with a
distinct dry season between January and April (Rand
and Rand 1982). The flora of BCI is described by Croat
(1978) and by Foster and Brokaw (1982); nomenclature
follows Croat (1978), except for Piper peltatum L. (pre-
viously Pothomorphe peltata (L.) Miq.). Ceiba pen-
tandra did not produce viable seeds on BCI during our
study, so collections of seeds from populations of this
species in Africa were used instead.
Eighteen species of pioneer trees and shrubs were
used (Table 1). They include the most common pioneer
species on BCI, both as mature individuals and in the
2800 T. R. H. PEARSON ET AL. Ecology, Vol. 83, No. 10
seed bank (Putz 1983, Dalling et al. 1997). Species
selection was based both on prior knowledge of their
regeneration ecology (Garwood 1982, Putz 1983, Bro-
kaw 1987, Dalling et al. 1997, 1998a, b), and their
distributions in relation to canopy height on a 50-ha
forest dynamics plot on BCI. The sample includes the
maximum range of seed size for pioneer trees and was
deliberately biased toward common species and those
for which large numbers of seeds were available (see
Table 1). Generic names are used except for Apeiba,
Cecropia, and Piper, for which binomials are used.
Gap microclimates
Irradiance and temperature regimes were character-
ized in small (;25 m2), medium (;64 m2), and large
gaps (;225 m2) and understory sites within 40-yr-old
secondary forest (canopy 12?20 m). In mature forest
on BCI, Brokaw (1982) found median gap areas to be
57 m2 (mean 86 m2) with a range from 26 to 342 m2.
Gaps were created by clearing all vegetation down to
ground level. In each site a pit was dug taking care not
to disturb the adjacent soil. Thermistor temperature
sensors attached to data loggers (Datahog, Skye In-
struments, Powys, UK) were inserted horizontally at
10 cm depth and the pits refilled. Bead thermistors
(Skye Instruments, Powys, UK) were also placed on
the soil surface covered with a thin (,1 mm) layer of
soil to simulate the natural conditions of radiation ex-
change. Photosynthetically active radiation (PAR) sen-
sors (SKP 215, Skye Instruments, Powys, UK) were
placed on the soil surface in the center of each site.
Measurements made at 30-s intervals were averaged
over 10-min periods. There were two census periods
both during the transition from the dry to the wet season
(when most germination of pioneers occurs, Garwood
1982). Between 16 and 22 April 2000 a small and a
large gap, and an understory site were monitored, and
between 1 and 11 May 2000 a different small, two
medium sized and a different large gap and understory
site were monitored.
Germination responses to irradiance
Seeds were collected from intact fruits on or below
trees on BCI, stored in a dark air-conditioned labora-
tory and shipped within 12 wk, in a polystyrene con-
tainer, to Aberdeen, UK. All counting of seeds was
done under light with a red/far red ratio of ,0.01.
Six replicates of 25 seeds of each of 17 pioneer spe-
cies (Table 1) were placed on the surface of 1% agar
in 50 mm diameter petri dishes. Each dish was placed
in an open-topped aluminum container (19 3 12 3 3
cm) and arranged at random in a plant growth chamber
with a constant 308C temperature and 8-h day length.
Three replicates per species were assigned at random
to a zero irradiance treatment created by placing an
inverted aluminum container (as above) on top and
surrounding both containers with aluminum foil. The
second three replicates were covered by a deep-dyed
lexan polycarbonate filter (Chroma green, Supergel Fil-
ters, Rosco, UK), which transmitted irradiance with a
PAR of 22.3 mmol?m22?s21 at a red/far red ratio of 1.2.
Irradiance in both treatments was measured using PAR
(as above) and red/far red sensors (SKR 110, Skye
Instruments, Powys, UK). On days 6, 12, 20, 30, and
45 after the start of the experiment, germinated seeds
(defined as emergence of the radicle) were removed.
Germination responses to temperature
Seed germination responses to temperature were de-
termined on a two-way thermogradient plate (Model
GRD1, Grant Instruments, Cambridge, UK). Seeds
were collected and processed as described above. Ceiba
seeds were collected in Ghana and immediately trans-
ported to Aberdeen where they were stored as described
above. One batch of 20 (Ochroma and Piper dilatatum)
or 25 (all other species) seeds of each of 15 species
(Table 1) was placed in each of 49 germination cells
(13 for Ceiba) on the plate. This apparatus created
independent gradients of day (5 h at 27.5?44.48C) and
night (19 h at 15.3?33.58C) temperatures, which al-
lowed us to impose a broad range of treatments for the
magnitude of diel (day?night) temperature fluctuation
across the range 26.08C to 129.18C. These conditions
encompass the minimum and maximum temperatures
that were measured across a variety of gap microhab-
itats on BCI. Irradiance of 200 mmol?m22?s21 PAR with
a red/far red ratio of 1.8 (sensors as above) was pro-
vided for 5 h each day, and was excluded at other times
using a 250-mm layer of expanded polystyrene.
Seeds were placed on a double layer of moistened
absorbent paper and irrigated by capillary action using
a network of lamp wicks connected to a reservoir of
deionized water. This arrangement maintained nonlim-
iting water availability to seeds on the thermogradient
plate in preliminary trials. The experiments were
scored for germination every 5 d for 60 d, and ger-
minated seeds were removed when first encountered.
The 15 species were run in three separate experiments
to optimize the space available on the thermogradient
plate and minimize the storage time of the seeds.
Soil depth and seedling emergence
Soil, sieved to ,1 mm, from an understory site on
BCI was used. Subsamples were incubated along with
experimental treatments to check for seed contami-
nants. The remaining soil was compressed into 7 cm
deep, 8 cm diameter plastic pots to a bulk density ap-
proximately that of forest soil.
Three replicates of 20 seeds of each of eight pioneer
species (Table 1) were planted into the soil at seven
depths (0, 2.5, 5, 10, 20, 30, and 50 mm). Seeds of all
species were collected within 3 mo of the start of the
experiment, and in the interim period were stored in a
dark air-conditioned laboratory on BCI. Seed pretreat-
ments (Table 1) were used where necessary to break
dormancy (Acuna and Garwood 1987). All pretreat-
October 2002 2801GERMINATION OF NEOTROPICAL PIONEERS
FIG. 1. Daily photosynthetically active radiation (PAR,
mol/m2; mean 6 1 SE) and daily temperature fluctuation (8C;
6 1 SE) between 1 and 11 May 2000 for gaps of 25 m2, 64
m2, and 200 m2.
ments and treatment assignments were carried out un-
der light with a red/far red ratio of 0.05.
For each depth treatment, 0.25 mm mesh-size nylon
cloth was laid directly beneath the seeds to prevent
them from sinking to a greater soil depth. Soil was then
carefully added on top of the mesh to the required
depth. Finally, pots were covered with clear polyeth-
ylene to conserve moisture, and placed at random with-
in a plant growth chamber on BCI. Both temperature
and irradiance were regulated on a 24-h cycle with 12
h at 308C and 75 mmol?m22?s21 PAR and a red/far red
ratio of 1.65, and 12 h at 248C in the dark.
Seedling emergence was checked weekly during the
day, and pots were watered with a mist sprayer as re-
quired. In order not to disturb seedlings that had not
yet emerged, emerging seedlings were cut off at the
soil surface but not removed. The experiment was ter-
minated for individual species 2 wk after germination
ceased. Pots were then checked for seedlings that ger-
minated, but did not emerge.
Data analysis
The null hypothesis of no difference in the proba-
bility of germination between irradiance treatments was
examined using G tests (Sokal and Rohlf 1995). The
relationships between germination and the diel mag-
nitude of temperature fluctuation was examined graph-
ically to determine germination thresholds (after ex-
cluding treatments where night temperature exceeded
day temperature). Where a clear threshold existed the
data were partitioned; relationships between germina-
tion and the diel magnitude of temperature fluctuation
were obtained, for data points above and below the
threshold value, using logistic regression implemented
in SPSS version 9.0.0 (SPSS, Chicago, Illinois, USA).
To analyze effects of soil depth on seedling emer-
gence we used generalized linear modeling techniques
(McCullagh and Nelder 1989) implemented in the
GLIM statistical package (Crawley 1993). To avoid
problems of inconstant variance and negative seedling
counts that might be predicted with a normal errors
linear regression model, we carried out a regression on
count data using Poisson errors and a log link function.
Hypothesis testing was carried out using the x2 test on
differences in deviance. The appropriateness of the as-
sumption of Poisson errors was checked by comparing
the residual deviance with the residual degrees of free-
dom after fitting soil depth. To investigate the rela-
tionship between emergence potential and seed mass
we performed a normal errors regression of the slope
values of individual species against ln-transformed
seed mass.
RESULTS
Microclimatic variation in relation to canopy
openness and burial depth
Microclimatic conditions were sensitive to canopy
openness. In both census periods irradiance at ground
level and mean daily magnitude of temperature fluc-
tuation on the soil surface and at 10 cm increased with
gap size (Fig. 1). Minimum night temperature never
varied more than 61.58C from 258C, regardless of gap
size or soil depth. In contrast the maximum temperature
varied markedly dependent on gap size, weather con-
ditions, and soil depth (Fig. 1). For example, on four
days maximum temperatures between 458 and 528C
were recorded on the soil surface in the large gap sites,
giving temperature fluctuations of .208C.
Maximum daytime temperature decreased with in-
creasing soil depth, producing smaller magnitudes of
diel temperature fluctuation deeper in the profile. How-
ever, in the large gaps a magnitude of 5.28C fluctuation
was still recorded at 10 cm depth. In smaller gaps there
was very little diel change in temperature at 10 cm
depth (Fig. 1).
Effect of irradiance treatments on germination
The study species separated into small-seeded pho-
toblastic species and larger seeded species that are not
photoblastic. For eight species that germinated in the
light (Alseis, Cecropia insignis, C. obtusifolia, C. pel-
tata, Miconia, Piper dilatatum, P. peltatum, and So-
lanum) no germination was recorded in complete dark-
ness (Fig. 2). Six species showed no significant dif-
ference in germination between treatments (Apeiba
membranacea, Guazuma, Luehea, Ochroma, Pseudo-
bombax, and Trichospermum, Fig. 2b) while the re-
maining three species showed little or no germination
in either treatment. The seed mass of the genera show-
ing a significant positive germination response to ir-
radiance was 0.7 6 0.4 mg (mean 6 1 SE), which is
more than an order of magnitude less than the mean
seed mass of genera containing species that did not
respond (22.7 6 14.8 mg).
Effects of temperature treatments on germination
The species showed a wide variety of germination
responses to temperature fluctuation treatments (Fig.
2802 T. R. H. PEARSON ET AL. Ecology, Vol. 83, No. 10
FIG. 2. Final germination percentage in the light (open bars) and dark (closed bars) treatments for (a) seven species
with seed mass , 1 mg and (b) eight species with seed mass . 1 mg. Degree of significance (G test) is indicated as
follows: ** P , 0.01; *** P , 0.001.
3), although four general classes of response can be
identified. Final germination percentages were .20%
in at least one treatment for all species (except A. ti-
bourbou). The germination responses to day and night
temperature are presented in the form of contour dia-
grams in Appendix A.
The species in the first class displayed consistent
high germination until a high threshold temperature
fluctuation was exceeded and a steep decline in ger-
mination occurred. The seed mass of the three genera
in this class was 0.86 6 0.44 mg (mean 6 1 SE). For
Luehea, Cecropia obtusifolia, C. peltata, and Miconia
(Fig. 3a?d) the threshold temperature fluctuation was
in the range 20?278C, and above the species-specific
threshold, germination declined abruptly to very low
values. Below the threshold, germination was high and
not significantly related to the magnitude of tempera-
ture fluctuation (P 5 0.18?0.45), although Miconia
showed an additional negative effect of high daytime
temperature (.41.78C) irrespective of the fluctuation
treatment (open symbols on Fig. 3d).
The second group, comprising three species in two
genera (C. insignis, Piper peltatum, and P. dilatatum,
Fig. 3e?g), had a lower tolerance to large temperature
fluctuations and a less distinct threshold effect, with
germination declining gradually beyond the threshold.
The mean seed mass of the two genera in this class
was 0.38 6 0.22 mg. The Piper species had negative
linear germination responses to temperature fluctua-
tions of .108C (P. peltatum, Fig. 3f ) or 38C (P. di-
latatum, Fig. 3g), respectively, and zero or very low
germination in response to high night temperatures
(.31.38C, open symbols on Fig. 3f, g) at any value of
temperature fluctuation. C. insignis had a negative lin-
ear germination response to temperature fluctuation
above a value of 6.78C and an additional negative effect
of low night temperatures (,21.78C, open symbols on
Fig. 3e) at intermediate fluctuations.
The third class of germination behavior was a pos-
itive response to increasing magnitudes of temperature
fluctuation. The mean seed mass of the four species in
four genera in this class was 20.90 6 14.19 mg. So-
lanum, Guazuma, and Ochroma (Fig. 3h?j) showed a
positive linear germination response to temperature
fluctuations in the range 0?16.78C, while Ceiba showed
a positive response to temperature fluctuation when
night temperature was .21.78C (closed symbols on
Fig. 3k). For the first three of these species, germination
either declined (Solanum) or varied idiosyncratically
(Guazuma, Ochroma) above a temperature fluctuation
of 16.78C (open symbols in Fig. 3h?j).
The remaining species examined (Trichospermum,
A. membranacea, and Cochlospermum) did not respond
significantly to temperature fluctuations. The mean
seed mass of the three species in this fourth group was
12.75 6 5.45 mg.
Therefore the species that had an abrupt temperature
fluctuation threshold for maintenance of high germi-
nation (Fig. 3a?d) and those that responded negatively
to an increasing magnitude of temperature fluctuation
(Fig. 3e?g) all had a seed mass ,2 mg. In contrast the
species that responded positively to temperature fluc-
tuation across some or all of the treatments used (Fig.
3h?k) and those that showed no response to tempera-
October 2002 2803GERMINATION OF NEOTROPICAL PIONEERS
FIG. 3. Effect of diel temperature fluctuation on final germination percentage for day temperatures (dt) in the range 27.5?
44.48C and night temperatures (nt) in the range 15.3?33.58C (except where indicated by open symbols as described) in (a)
Luehea, (b) Cecropia obtusifolia, (c) Cecropia peltata, (d) Miconia (open circles, dt . 41.78C), (e) Cecropia insignis (open
circles, nt , 21.78C), (f ) Piper peltatum (open circles, nt . 31.38C), (g) Piper dilatatum (open circles, nt . 31.38C), (h)
Solanum, (i) Guazuma, ( j) Ochroma (open circles, diel temperature fluctuations .16.78C), and (k) Ceiba (open circles, nt
, 21.78C). Vertical lines in panels (a?d) represent the threshold values (see Methods: Data analysis for definition).
ture fluctuation all had a mean seed mass .2 mg (Table
1).
Effects of burial depth on seedling emergence
Seedling emergence was negatively related to burial
depth for all species except A. membranacea (Fig. 4).
Among species, there were large differences in the
amount of deviance explained by depth (range r2 5
0.20 for Ochroma to r2 5 0.97 for Miconia). The depth
at which 50% of maximum seedling emergence oc-
curred ranged between 1 and 6.5 mm for the small-
seeded species Miconia and Cecropia, and 40 and 82
mm for the larger seeded Trichospermum and Ochro-
ma.
Seed mass was significantly related to emergence
potential, determined as the slope of regressions be-
tween emergence success and soil depth (n 5 8 species,
r2 5 0.78, F 5 21.4, P , 0.01), so that the capacity
to emerge from depth increased with increasing seed
mass (Appendix B). Excluding the outlier value for
Miconia, this relationship was still significant (n 5 7,
r2 5 0.76, F 5 16.0, P , 0.05).
DISCUSSION
Variation in irradiance and temperature conditions
in the forest
Temperature and irradiance are both good indi-
cators of canopy openness, but differ in their ability
2804 T. R. H. PEARSON ET AL. Ecology, Vol. 83, No. 10
FIG. 4. Mean percentage seedling emergence (61 SD) as a function of seed burial depth. Curve-fitting is by regression
on emergence count data using Poisson errors and a log link function.
to indicate large and small gaps. Irradiance in trop-
ical forests has been well characterized (Chazdon and
Fetcher 1984).There are, however, fewer measure-
ments of soil temperatures in understory and gaps,
and these concur broadly with our findings (e.g.,
Va?zquez-Yanes and Orozco-Segovia 1982). Irradi-
ance can act as an index of canopy openness across
all spatial scales of disturbance, from a branch fall
to a multiple tree-fall gap (Chazdon and Fetcher
1984). In contrast, in the smallest gaps we studied,
temperature, especially beneath the soil surface, did
not differ greatly from that in the understory. In these
gaps, therefore, irradiance would be a more effective
cue for the germination of gap-demanding species.
Among the BCI pioneer flora, it is noticeable that
the photoblastic small-gap specialists (e.g., Miconia,
Piper, and Alseis) have the smallest seeds (mostly
,0.5 mg dry mass), while the large-gap specialists
October 2002 2805GERMINATION OF NEOTROPICAL PIONEERS
(e.g., Trema and Ochroma) tend to have larger seeds
(Brokaw 1987, Dalling et al. 2001).
Effects of irradiance on seed germination
Eight small-seeded pioneer tree species from five
genera (mean mass 0.7 mg) were photoblastic. In con-
trast, the only photoblastic species with a seed mass
.1 mg was Solanum. While photoblastic germination
has been reported in species from some of the genera
we studied (e.g., Cecropia, Piper, and Solanum species
[Baskin and Baskin 1998]), it has not been reported
previously for C. insignis, Miconia, P. dilatatum, P.
peltata, or S. hayesii. The mechanism of the photo-
blastic response has been studied extensively; phyto-
chrome B is sensitive to the spectral distribution of
incident radiation and stimulates germination when the
proportion of red wavelengths is high (Smith 2000).
It has been argued that photoblastic germination
evolved as a canopy gap detection mechanism in trop-
ical pioneer trees (Frankland 1976, Va?zquez-Yanes and
Orozco-Segovia 1990). Gap creation leads to a dra-
matic increase in the red/far red ratio of irradiance at
the forest floor, and these shifts are sufficient to initiate
germination in photoblastic species (Va?zquez-Yanes
and Orozco-Segovia 1993). Alternatively, photoblastic
germination may act to initiate germination when a
seed is close to the soil surface in a litter-free microsite.
Litter and soil decrease the quantity and the red/far red
ratio of irradiance (Tester and Morris 1987, Va?zquez-
Yanes et al. 1990) and are barriers to the establishment
of small-seeded species (Molofsky and Augspurger
1992, Va?zquez-Yanes and Orozco-Segovia 1992, Met-
calfe and Grubb 1997) with limited seed reserves for
emergence (Bond et al. 1999; Fig. 4 Cecropia and Mi-
conia). These factors may account for the high density
of pioneers in the pits and mounds formed by treefalls
on Barro Colorado Island, where high irradiance is
combined with an absence of litter (Putz 1983).
Effects of temperature and magnitude of temperature
fluctuation on germination
Germination response to increasing magnitudes of
temperature fluctuations was negative in the smaller
seeded species, but positive in four species with a seed
mass .2 mg. Previous studies have demonstrated the
positive response in Ochroma pyramidale and in the
Solanaceae (Baskin and Baskin 1998). The smaller
seeded species responded negatively, although respons-
es ranged from broad tolerance to temperature fluctu-
ation in Luehea and Cecropia obtusifolia to extreme
sensitivity even to small temperature fluctuation in Pip-
er. Thus, temperature fluctuation acts as a signal to the
presence of larger canopy gaps in a wide range of spe-
cies, but the response to the signal is associated with
differences in seed size.
Few previous studies exist on temperature effects on
germination although studies of temperate plants sup-
port our findings. Temperature has been shown to affect
germination of relatively few tropical pioneer tree spe-
cies (Baskin and Baskin 1998), possibly because most
researchers have chosen to work with the more abun-
dant tiny-seeded species that did not respond positively
to the magnitude of temperature fluctuation in our ex-
periments. However, among temperate-zone herbs a
positive effect of increased temperature fluctuations is
widespread among plants that require gaps in vegeta-
tion for seedling establishment (Thompson and Grime
1983). The mechanism for an effect of temperature
fluctuation on germination of neotropical pioneers and
in many temperate species is related to an increase in
seed coat permeability induced by alternating temper-
atures (Baskin and Baskin 1998).
The negative germination response to large temper-
ature fluctuations in small-seeded species and at ex-
treme values for some larger seeded species may rep-
resent a waiting strategy to avoid large-gap environ-
ments where water shortage would constrain seedling
establishment. Small-seeded species are especially sus-
ceptible to short periods of drought because of their
superficial initial rooting depth (Engelbrecht et al.
2001). Variation among small-seeded species in the
critical temperature fluctuation threshold that inhibited
germination may reflect differential selection for tol-
erance of conditions in different irradiance levels
linked to early establishment requirements of these spe-
cies.
Effects of burial depth on seedling emergence
Seedling emergence declined with depth of burial
for all species. This may be caused by a lack of seed
reserves to elongate the epicotyl sufficiently to reach
the soil surface, by buckling of the epicotyl or hypo-
cotyl as a result of insufficient stiffness to emerge, or
by the failure of a germination trigger to penetrate the
soil. For fynbos shrubs, Bond et al. (1999) found that
maximum emergence depth was similarly scaled al-
lometrically with seed mass.
Under natural conditions, emergence from the seed
bank might be determined by the detection of dormancy
breaking stimuli, rather than limitations imposed by
seed reserves. However, it seems implausible that seeds
??capable?? of emerging from deep in the soil would
not have evolved a gap detection mechanism to take
advantage of this ability. Thus it is significant that the
two species with the lowest emergence potential were
photoblastic (Cecropia and Miconia), while the only
species from the large-seeded group that possessed
photoblastic germination (Solanum), also responded
positively to temperature fluctuation.
Limitations and conclusions
Although we included most of the common pioneer
species at our study site, not all taxa may show a re-
lationship between germination cues and seed mass
(e.g., Carica papaya [Paz and Va?zquez-Yanes 1998]).
Nonetheless, the generality of our model is supported
2806 T. R. H. PEARSON ET AL. Ecology, Vol. 83, No. 10
by research on temperate zone plants demonstrating
that dependence on light for germination increases as
seed size decreases in interspecific comparisons (Mil-
berg et al. 2000).
One sense in which our sample of species may be
biased is that most of the relatively large-seeded spe-
cies possess thick, hard seed coats that might be
opaque. Thus the absence of photoblastism in these
species might be confounded by the need for large
seeds to be defended against seed predators and path-
ogens. In our sample of species, mean seed coat thick-
ness (excluding the wax layer) was significantly greater
in the larger seeded nonphotoblastic species (116 6 19
mm) than in the small-seeded photoblastic species (50
6 12 mm), and ln(seed coat thickness) increases lin-
early with ln(seed mass) (r2 5 0.77; P , 0.001). This
implies that larger seeded pioneer species might be
better defended against predation and disease than
smaller seeded species and provides a mechanistic ex-
planation for the greater persistence of larger seeded
species in soil (Dalling et al. 1997). According to this
evolutionary scenario, the mechanism of germination
response to gap signals has evolved as a consequence
of the seed size/seed number trade-off coupled with
selection for defense of large-seeded species (Baskin
et al. 2000).
Emergence potential may represent an important and
previously unrecognized element of the trade-off be-
tween seed size and the gap colonizing potential of
pioneer species. Larger seeded pioneers have the ad-
vantage in probability of establishment and in initial
competition. Small-seeded pioneers have a dispersal
advantage, since decreased size affords the production
of increased seed number, and a consequently higher
probability that one or more seeds will arrive at a suit-
able gap site (Metcalfe and Grubb 1996, Rees and Wes-
toby 1996). The pool of dormant seeds for small-seeded
species, however, may be relatively small for two rea-
sons. First, only a small proportion of the germinable
seed population in the soil may be physically capable
of germination because of their limited emergence po-
tential. Second, seed losses may be greater at the soil
surface than deeper in the soil (Dalling et al. 1997,
1998b). Thus seed banks of smaller seeded species may
be less persistent than those of larger seeded species.
ACKNOWLEDGMENTS
The manuscript was greatly improved by comments from
Peter Grubb, Bill Platt, and two anonymous referees. We are
grateful to Mike Swaine for supplying the seeds of Ceiba
pentandra. Funding for T. R. H. Pearson was provided by
The Natural Environment Research Council and The Lever-
hulme Trust.
LITERATURE CITED
Acun?a, P., and N. C. Garwood. 1987. Efecto de la luz y de
la escarificacio?n en la germinacio?n de las semillas de cinco
especies de arboles tropicales secundarios. Revista de Biol-
ogia Tropical 35:203?207.
Baskin, C. C., and J. M. Baskin. 1998. Seeds. Academic
Press, San Diego, California, USA.
Baskin, J. M., C. C. Baskin, and L. Xiaojie. 2000. Taxonomy,
anatomy and evolution of physical dormancy in seeds. Plant
Species Biology 15:139?152.
Bond, W. J., M. Honig, and K. E. Maze. 1999. Seed size and
seedling emergence: an allometric relationship and some
ecological implications. Oecologia 120:132?136.
Brokaw, N. V. L. 1982. Treefalls: frequency, timing, and
consequences. Pages 101?108 in E. Leigh, A. Rand, and
D. Windsor, editors. The ecology of a tropical forest. Smith-
sonian Institution Press, Washington, D.C., USA.
Brokaw, N. V. L. 1987. Gap-phase regeneration of three pi-
oneer tree species in a tropical forest. Journal of Ecology
75:9?19.
Chazdon, R. L., and N. Fetcher. 1984. Light environments
of tropical forests. Pages 27?36 in E. Medina, H. A. Moo-
ney, and C. Va?zquez-Yanes, editors. Physiological ecology
of plants of the wet tropics. Junk, Dordrecht, The Neth-
erlands.
Crawley, M. J. 1993. GLIM for ecologists. Blackwell Sci-
entific, Oxford, UK.
Croat, T. B. 1978. Flora of Barro Colorado Island. Stanford
University Press, Stanford, California, USA.
Dalling, J. W., S. P. Hubbell, and K. Silvera. 1998a. Seed
dispersal, seedling establishment and gap partitioning
among tropical pioneer trees. Journal of Ecology 86:674?
689.
Dalling, J. W., M. D. Swaine, and N. C. Garwood. 1997. Soil
seed bank community dynamics in seasonally moist low-
land tropical forest, Panama. Journal of Tropical Ecology
13:659?680.
Dalling, J. W., M. D. Swaine, and N. C. Garwood. 1998b.
Dispersal patterns and seed bank dynamics of pioneer trees
in moist tropical forest. Ecology 79:564?578.
Dalling, J. W., K. Winter, J. D. Nason, S. P. Hubbell, D. A.
Murawski, and J. L. Hamrick. 2001. The unusual life his-
tory of Alseis blackiana: a shade-persistent pioneer tree?
Ecology 82:933?945.
Engelbrecht, B. M. J., J. W. Dalling, T. R. H. Pearson, R. L.
Wolf, D. A. Galvez, T. Koehler, M. C. Ruiz, and T. A.
Kursar. 2001. Short dry spells in the wet season increase
mortality of tropical pioneer seedlings. Pages 665?669 in
K. N. Ganeshaiah, R. Uma Shaanker, and K. S. Bawa, ed-
itors. Tropical ecosystems: structure, diversity and human
welfare. Proceedings of the International Conference on
Tropical Ecosystems: structure, diversity and human wel-
fare. Oxford and IBH Publishing Company, New Delhi,
India.
Foster, R. B., and N. V. L. Brokaw. 1982. Structure and
history of the vegetation of Barro Colorado Island. Pages
67?81 in E. Leigh, A. Rand, and D. Windsor, editors. The
ecology of a tropical forest. Smithsonian Institution Press,
Washington, D.C., USA.
Frankland, B. 1976. Phytochrome control of seed germina-
tion in relation to the light environment. Pages 477?491
in H. Smith, editor. Light and plant development. Butter-
worths, London, UK.
Garwood, N. C. 1982. Seasonal rhythm of seed germination
in a semi-deciduous tropical forest. Pages 173?185 in E.
G. Leigh Jr., A. S. Rand, and D. M. Windsor, editors. The
ecology of a tropical forest: seasonal changes and long-
term changes. Smithsonian Institution Press, Washington,
D.C., USA.
Grubb, P. J. 1996. Rainforest dynamics: the need for new
paradigms. Pages 215?233 in D. S. Edwards, S. C. Choy,
and W. E. Booth, editors. Tropical rainforest research?
current issues. Kluwer, The Hague, The Netherlands.
Harms, K. E., S. J. Wright, O. Calderon, A. Hernandez, and
E. A. Herre. 2000. Pervasive density dependent recruit-
ment enhances seedling diversity in a tropical forest. Nature
404:493?495.
October 2002 2807GERMINATION OF NEOTROPICAL PIONEERS
Hubbell, S. P., R. B. Foster, S. T. O?Brien, K. E. Harms, R.
Condit, B. Wechsler, S. J. Wright, and L. Loo de Lao. 1999.
Light-gap disturbances, recruitment limitation, and tree di-
versity in a Neotropical forest. Science 283:554?557.
McCullagh, P., and J. A. Nelder. 1989. Generalized linear
models. Chapman and Hall, London, UK.
Metcalfe, D. J., and P. J. Grubb. 1996. Seed mass and light
requirements for regeneration in South-east Asian rain for-
est. Canadian Journal of Botany 73:817?826.
Metcalfe, D. J., and P. J. Grubb. 1997. The responses to shade
of seedlings of very small-seeded tree and shrub species
from tropical rain forest in Singapore. Functional Ecology
11:215?221.
Milberg, P., L. Andersson, and K. Thompson. 2000. Large-
seeded spp. are less dependent on light for germination
than small-seeded ones. Seed Science Research 10:99?104.
Molofsky, J., and C. K. Augspurger. 1992. The effect of leaf
litter on early seedling establishment in a tropical forest.
Ecology 73:68?77.
Paz, L., and C. Va?zquez-Yanes. 1998. Comparative seed eco-
physiology of wild and cultivated Carica papaya trees from
a tropical rain forest region in Mexico. Tree Physiology
18:277?280.
Putz, F. E. 1983. Treefall pits and mounds, buried seeds, and
the importance of soil disturbance to pioneer trees on Barro
Colorado Island. Ecology 64:1069?1074.
Raich, J. W., and W. K. Gong. 1990. Effects of canopy open-
ings on tree seed germination in a Malaysian dipterocarp
forest. Journal of Tropical Ecology 6:203?217.
Rand, A. S., and W. M. Rand. 1982. Variation in rainfall on
Barro Colorado Island. Pages 47?59 in E. Leigh, A. Rand,
and D. Windsor, editors. The ecology of a tropical forest.
Smithsonian Institution Press, Washington, D.C., USA.
Rees, M., and M. Westoby. 1996. Game-theoretical evolution
of seed mass in multi-species ecological models. Oikos 78:
116?126.
Smith, H. 2000. Phytochromes and light signal perception
by plants?an emerging synthesis. Nature 407:585?591.
Sokal, R. R., and F. J. Rohlf. 1995. Biometry. W. H. Freeman,
New York, New York, USA.
Swaine, M. D., and T. C. Whitmore. 1988. On the definition
of ecological species groups in tropical rain forest. Vege-
tatio 75:81?86.
Tester, M., and C. Morris. 1987. The penetration of light
through soil. Plant, Cell and Environment 10:281?286.
Thompson, K., and J. P. Grime. 1983. A comparative study
of germination responses to diurnally-fluctuating temper-
atures. Journal of Applied Ecology 20:141?156.
Va?zquez-Yanes, C., and A. Orozco-Segovia. 1982. Seed ger-
mination of a tropical rain forest tree Heliocarpus donnell-
smithii in response to diurnal fluctuations in temperature.
Physiologia Plantarum 56:295?298.
Va?zquez-Yanes, C., and A. Orozco-Segovia. 1990. Seed dor-
mancy in the tropical rain forest. Pages 247?259 in K. S.
Bawa and M. Hadley, editors. Reproductive ecology of
tropical forest plants. UNESCO, Paris, France.
Va?zquez-Yanes, C., and A. Orozco-Segovia. 1992. Effects
of litter from a tropical rain forest on tree seed germination
and establishment under controlled conditions. Tree Phys-
iology 11:391?400.
Va?zquez-Yanes, C., and A. Orozco-Segovia. 1993. Patterns
of seed longevity and germination in the tropical rainforest.
Annual Review of Ecology and Systematics 24:69?87.
Va?zquez-Yanes, C., A. Orozco-Segovia, E. Rincon, M. E.
Sanchez-Coronado, R. Huante, J. R. Toledo, and U. L. Bar-
radas. 1990. Light beneath the litter in a tropical forest:
effect on seed germination. Ecology 71:1952?1958.
Vazquez-Yanes, C., and B. Pe?rez-Garc??a. 1976. Notas sobre
la morfolog??a y la anatom??a de la testa de las semillas de
Ochroma lagopus Sw. Turrialba 26:310?311.
Weiner, J. 1990. Asymmetric competition in plant popula-
tions. Trends in Ecology and Evolution 5:360?364.
APPENDIX A
Contour maps of final germination percentages of 15 species across temperature treatments are available in ESA?s Electronic
Data Archive: Ecological Archives E083-054-A1.
APPENDIX B
A figure showing the effect of seed mass on emergence potential is available in ESA?s Electronic Data Archive: Ecological
Archives E083-054-A2.