and J. Selverstone, Contrib. Mineral. Petrol. 83, 348
(1982); (19)].
16. C. R. Vyhnal and C. P. Chamberlain, Am. J. Sci. 296,
394 (1996). At our field locality, there probably was
diagenetic exchange that caused the
87
Sr/
86
Sr of the
amphibolite protolith to be increased from about
0.703 (based on the Nd isotope data) to about
0.705 or 0.706. The pelite value may also have been
affected.
17. The Lebendun Nappe here contains three horizons
of garnet amphibolite (8 to 15 m thick) interlay-
ered with pelite. The “pelite” grades from a quartz-
rich (86% SiO
2
) psammite at the contact to a more
typical pelite (75% SiO
2
) within ⬃10 cm. The
amphibolite contains biotite, which decreases in
mode away from the contact, and garnet porphy-
roblasts up to 5 mm in diameter. At the contact,
garnets are more sparse and rarely exceed 500 m
in diameter. Pelitic garnets continue to decrease in
diameter and mode away from the contact, ⬍200
m near the contact and reaching zero after sev-
eral meters.
18. See www.sciencemag.org/feature/data/1049980.shl
for supplementary raw data and analytical methods.
19. D. Vance and R. K. O’Nions, Earth Planet. Sci. Lett.
114, 113 (1992).
20. This is because K
Nd
⬎⬎ K
Sr
such that D
Nd
*
is very
small.
21. N. S. Mancktelow, Tectonophysics 215, 295 (1992).
22. J. B. Brady, Am. J. Sci. 283A, 181 (1983).
23. A. R. Kotelnikov, I. V. Chernysheva, M. S. Parmuzina,
Geochem. Int. 36, 846 (1998).
24. E. H. Oelkers and H. C. Helgeson, Geochim. Cosmo-
chim. Acta 52, 63 (1988).
25. C. M. Graham, J. W. Valley, J. M. Eiler, H. Wada,
Contrib. Mineral. Petrol. 132, 371 (1998); J. L. M. van
Haren, J. J. Ague, D. M. Rye, Geochim. Cosmochim.
Acta 60, 3487 (1996); E. D. Young and D. Rumble,
Geochim. Cosmochim. Acta 57, 2585 (1993).
26. J. V. Walther, J. Geol. 102, 559 (1994).
27. We thank J. J. Ague, J. G. Bryce, and several anony-
mous reviewers for valuable reviews and discussions,
and J. Selverstone and T. Wawrzyniec for assisting in
field site selection. Supported by NSF grant EAR-
9805218 (D.J.D.) and a Berkeley Geochronology Cen-
ter Fellowship (E.F.B.).
11 February 2000; accepted 21 March 2000
Spatial Patterns in the
Distribution of Tropical Tree
Species
Richard Condit,
1
* Peter S. Ashton,
2
Patrick Baker,
3
Sarayudh Bunyavejchewin,
4
Savithri Gunatilleke,
5
Nimal Gunatilleke,
5
Stephen P. Hubbell,
6
Robin B. Foster,
7
Akira Itoh,
8
James V. LaFrankie,
9
Hua Seng Lee,
10
Elizabeth Losos,
1
N. Manokaran,
11
R. Sukumar,
12
Takuo Yamakura
8
Fully mapped tree census plots of large area, 25 to 52 hectares, have now been
completed at six different sites in tropical forests, including dry deciduous to
wet evergreen forest on two continents. One of the main goals of these plots
has been to evaluate spatial patterns in tropical tree populations. Here the
degree of aggregation in the distribution of 1768 tree species is examined based
on the average density of conspecific trees in circular neighborhoods around
each tree. When all individuals larger than 1 centimeter in stem diameter were
included, nearly every species was more aggregated than a random distribution.
Considering only larger trees (ⱖ 10 centimeters in diameter), the pattern
persisted, with most species being more aggregated than random. Rare species
were more aggregated than common species. All six forests were very similar
in all the particulars of these results.
The spatial dispersion of individuals in a
species is central in ecological theory (1, 2).
Patchiness, or the degree to which individuals
are aggregated or dispersed, is crucial to how
a species uses resources, to how it is used as
a resource, and to its reproductive biology.
Spatial patterns have been a particularly im-
portant theme in tropical ecology, because
high diversity in the tropics begets low den-
sities. Since Wallace (3) noted how difficult it
was to find two individuals of the same spe-
cies, the hyperdispersion of tropical trees has
focused much of theoretical tropical ecology.
In 1979, Hubbell (4) published a large
study of dispersion of trees in a dry forest in
Costa Rica. His results were contrary to Wal-
lace’s long-prevailing wisdom and the Jan-
zen–Connell prediction (5, 6 ) that wide dis-
persion is a defense against predators. Most
species were aggregated, so that near neigh-
borhoods of a tree had a higher than average
density of conspecifics. Since that study,
though, contradictory results have appeared,
particularly from Lieberman and Lieberman
(7), who found that most species in a wet
forest in Costa Rica, as well as from a liter-
ature survey, were not aggregated.
Over the past two decades, we have been
assembling a long-term, large-scale, global re-
search effort on spatial patterns and dynamics
of tropical forests (8, 9). An international team
has now fully censused six plots in five tropical
countries, mapping and identifying every indi-
vidual of ⱖ1 cm in stem diameter over 25 to 52
ha at each plot (Table 1). The large plot size is
necessary to encompass substantial populations
of most tree species in the community. Major
goals of this effort have been to examine Jan-
zen–Connell effects, density dependence, and
the spacing pattern of individual species.
The six sites represent a wide variety of
tropical forests (Table 1). At one extreme, the
two plots in Malaysia are in tall, evergreen
forest; have no regular dry season; and include
over 800 tree and treelet species each. The Sin-
haraja forest is also very wet and evergreen, but
its island setting reduces species diversity. The
site in India is in dry forest with a fairly open
canopy, grassy understory, and just 70 species;
the Thai site is also dry and low in diversity. The
single site in Central America is moist forest,
structurally quite like the Malaysian sites, but
intermediate in climate and diversity. The for-
ests also cover a wide taxonomic range. Four of
the Asian sites are dominated by the family
Dipterocarpaceae, but few species are shared
among them. The Indian and American sites are
distinct taxonomically; they are not dominated
by a single family and have few (India) or no
(Panama) dipterocarps.
We evaluate spatial patterning by examining
neighborhoods around individual trees. For each
individual, we tallied the number of conspecifics
between x and x ⫹⌬x meters for all x ⫹⌬x
inside the plot. We also calculated the area
inside the plot of each of these annuli. The
number of neighbors N
x
and the area A
x
in each
annulus at distance x were then summed over all
individuals of a given species. D
x
⫽⌺N
x
/ ⌺A
x
gives the density of neighboring conspecifics as
a function of distance from the average individ-
ual. This is a biologically meaningful measure
of clumping, because it evaluates the conspe-
cific population density in the neighborhood of
1
Center for Tropical Forest Science, Smithsonian
Tropical Research Institute, Unit 0948, APO AA
34002–0948, USA.
2
Center for Tropical Forest Sci-
ence, Harvard Institute for International Development
and Harvard University, Cambridge, MA 02138, USA.
3
Silviculture Laboratory, College of Forest Resources,
University of Washington, Seattle, WA 98195–2100,
USA.
4
Royal Thai Forest Department, Chatuchak,
Bangkok 10900, Thailand.
5
Department of Botany,
University of Peradeniya, Peradeniya, Sri Lanka.
6
Cen-
ter for Tropical Forest Science, University of Georgia,
Athens, GA 30602, USA.
7
Center for Tropical Forest
Science, Field Museum of Natural History, Chicago, IL
60605–2496, USA.
8
Osaka City University, Osaka
558-8585, Japan.
9
Center for Tropical Forest Science,
National Institute of Education, Singapore 1025.
10
Sarawak Forest Department, Kuching, Sarawak
93660, Malaysia.
11
Forest Research Institute of Ma-
laysia, Kepong 52109, Kuala Lumpar, Malaysia.
12
In-
dian Institute of Science, Bangalore 560012, India.
*To whom correspondence should be addressed. E-
mail: ctfs@tivoli.si.edu
R EPORTS
26 MAY 2000 VOL 288 SCIENCE www.sciencemag.org1414
each tree. It is closely related to Ripley’s K
statistic [called the correlation integral by as-
tronomers when applied, in three dimensions, to
the distribution of galaxies (10)], but K is a
cumulative distribution, whereas our neighbor-
hood density is a probability density function;
that is, K
x
refers to conspecifics of ⬍x meters
from the focal tree, and D
x
refers to an
annulus between x and x ⫹⌬x meters.
Although the K statistic is very popular
(11), our approach has the advantage of
isolating specific distance classes, whereas
K confounds effects at larger distances with
effects at shorter distances (12, 13).
To compare species with various population
densities, we standardized D
x
by dividing it by
the mean density of a given species across the
whole plot. We call this standardized index the
relative neighborhood density, or ⍀. In a per-
fectly random distribution, ⍀
x
⫽ 1 for all dis-
tances x. Aggregation is indicated when ⍀
x
⬎ 1
at short distances, whereas ⍀
x
⬍ 1 at short
distances indicates spacing at some scale, or
hyperdispersion. A great advantage of this stan-
dardized statistic is that it is sample-size inde-
pendent, which allowed us to directly compare
species and stem diameter classes and offered a
bootstrap method for estimating confidence
limits (14 ). Rare species, with N ⬍ 50 individ-
uals, had to be dealt with carefully, and most of
our statistics are based only on species with at
least one individual per hectare (15).
Nearly every species was aggregated when
all diameter classes of ⱖ1 cm were included. In
the six plots, 1768 species had at least one
individual per hectare: 1753 were aggregated at
0to10m(⍀
0 –10
⬎ 1), 1490 significantly so
(95% confidence limits around ⍀
0 –10
did not
include 1); 1759 were aggregated at 10 to 20 m,
and 1730 were aggregated at 20 to 30 m (16).
No plot had fewer than 96% of its species
aggregated at 0 to 10 m, and every one of 772
species at Lambir was aggregated. Relative
neighborhood density almost invariably de-
clined with distance (Fig. 1): ⍀
10 –20
⬍⍀
0 –10
in 1714 of 1768 species, and ⍀
20 –30
⬍⍀
10 –20
in 1581 species. Nevertheless, ⍀ values in near-
by distance classes were highly correlated with
one another. Thus, we can use ⍀
0 –10
—the
mean conspecific density within 10 m of a tree
(relative to the species’ overall density)—as a
simple measure of the intensity of aggregation
of a species.
There was an enormous range in ⍀
0 –10
.
Most species had values of ⬍10, but there
were a few species with much higher aggre-
gation. The highest of all among the species
with ⱖ1 individual per hectare, ⍀
0 –10
⫽ 906,
was in Lagerstroemia sp. (Lythraceae) in the
HKK plot; 51 of the 59 individuals of this
species occurred in a single clump in an area
of 400 m
2
. Nineteen species had aggregation
indices over 100, eight at Lambir and four
at HKK. These results are dominated by
saplings, because the vast majority of
trees ⱖ 1 cm diameter are ⬍10 cm diameter.
Results hold for larger trees, though. Nearly
all species were aggregated when only diameter
classes of ⱖ10 cm were considered. There were
543 species with more than 1 individual per
hectare at this size in all six plots, and 488 had
⍀
0 –10
⬎ 1; for 257 species, the aggregation
was significant (16). But aggregation intensity
weakened at greater sizes in most species: 321
of 543 had a higher ⍀
0 –10
for all trees than for
larger trees and of those with a significant dif-
ference, 84 of 102 were more aggregated at the
smaller than at the larger size. At BCI, Pasoh,
HKK, and Lambir, about two-thirds of the spe-
cies were more aggregated at the smaller diam-
eter class, but the pattern reversed at Sinharaja
and Mudumalai and most species became more
aggregated at the larger size. We repeated the
analysis for trees of ⱖ30 cm diameter and again
found most species aggregated (16 ).
Rare species were substantially more aggre-
gated than common species at all but the Mu-
dumalai site (17). Median ⍀
0 –10
was4to10
times higher in the rare abundance class than in
the commonest classes (Table 2). In the most
abundant species, ⍀
0 –10
⬍ 5, and at BCI and
Pasoh mostly ⍀
0 –10
⬍ 2, whereas in species
with fewer than 500 individuals, ⍀
0 –10
typical-
ly ranged from 5 to 30 or higher (Fig. 2). This
trend held for species as rare as 10 individuals
per 50 ha, even though many species had ⍀
0–
10
⫽ 0 when N ⬵ 10 (15). For N ⬍ 10
individuals, median scores were zero, but arith-
metic means were very high. Thus, the rarest
species were the most aggregated of all, and
across the entire range of abundances (four
orders of magnitude), the degree of clumping
correlated negatively with species density.
These results match Hubbell’s conclusion from
dry forest in Costa Rica (4) and two previous
analyses of the Pasoh plot (18, 19). All results
from the neighborhood density statistic ⍀ were
Fig. 1. Relative neighborhood density (⍀
x
as defined in the text) as a function of distance x in
sample species. Vertical bars give 95% confidence limits (14), sometimes too small to see.
R EPORTS
www.sciencemag.org SCIENCE VOL 288 26 MAY 2000 1415
confirmed by nearest-neighbor analysis (16 ).
The six forests were remarkably similar in
community-wide patterns of aggregation. The
range of values for ⍀
0 –10
was similar at all sites,
and the relationship between this aggregation
index and abundance was essentially identical
across sites. However, there were significant
differences (20). Pasoh and BCI had low aggre-
gation indices at a given abundance, whereas
Sinharaja and Lambir had high indices (Table
2). These results correspond with habitat varia-
tion within the plots in that Pasoh and BCI are
topographically uniform and Lambir and Sinha-
raja are relatively rugged (Table 1). Lambir also
has a sharp soil gradient (21). Mudumalai has
steep topography, but there is little indication
that species respond to it, whereas at Lambir and
Sinharaja many species’ distributions follow to-
pographic features (22). There are clear exam-
ples of habitat-related patchiness at most of the
plots, especially Sinharaja (Fig. 3). Species in
which larger trees were more densely aggregat-
ed than juveniles are also suggestive of habitat-
related patchiness, because adults might be col-
lected in sites most favorable for the species,
whereas juveniles are widely dispersed.
On the other hand, there are many species
in all the plots whose aggregated distributions
indicate dispersal limitation. These species
occur in circular clumps that do not corre-
spond with topography (Fig. 3), and they had
the highest values of ⍀
0 –10
. Their neighbor-
hood density functions ⍀
x
declined abruptly
with distance (Fig. 1). We tested the impor-
tance of dispersal limitation by comparing
aggregation intensity in poorly versus well-
dispersed trees. Species whose seeds are dis-
persed by animals were assumed to be better
dispersed than wind- or explosively dispersed
species, and canopy trees were assumed to
have well-dispersed seeds relative to under-
story treelets (23). We also considered the
Dipterocarpaceae, a family of trees with
poorly dispersed, winged seeds that domi-
nates Southeast Asian forests.
The prediction that better dispersal reduc-
es aggregation was partially borne out. There
was no significant difference in aggregation
intensity between canopy and understory spe-
cies at Pasoh, but at BCI there was (24). At
BCI, there was no significant difference in
aggregation for animal versus nonanimal dis-
persed species, but the difference was fairly
pronounced in the predicted direction (25).
Finally, at the two Malayasian plots, diptero-
carps were strikingly more aggregated than
nondipterocarps (26).
Table 1. Six tropical forest dynamics plots that have been fully censused at least once. Underlined census year is the one used in this report. Dry season months
gives the number per year with mean rainfall ⬍ 100 mm. dbh, diameter at breast height.
Plot
Plot size
(ha)
Fully
censused
Topographic
variation
(m)
Annual
rainfall
(mm)
Dry season
(months)
Species
ⱖ 1
cm dbh
Species
ⱖ 10
cm dbh
Individuals
ⱖ 1cm
dbh
Individuals
ⱖ 10 cm
dbh
Barro Colorado Island 50 1981–83 38 2500 4 299 226 229,071 21,459
(BCI), Panama 1985
1990
1995
Pasoh, Malaysia 50 1986–89 25 1800 0 818 673 320,382 28,997
(Peninsula) 1990
1995–96
Lambir, Malaysia 52 1991–94 140 2700 0 1174 996 366,121 32,962
(Borneo) 1997
Huai Kha Khaeng 50 1992–94 89 1450 6 248 211 81,145 21,957
(HKK), Thailand
Mudumalai, India 50 1988– 89 130 1200 6 72 63 25,306 14,922
1992
1996
Sinharaja, Sri Lanka 25 1996–98 151 5000 0 203 164 206,227 17,214
Table 2. Median aggregation index, ⍀
0–10
, across species in various abundance categories in the six large plots. Number of species within each abundance
category is listed under spp. Last row gives overall median ⍀
0–10
for all species with at least 50 individuals at each plot.
Abundance class
(per 50 ha)
Pasoh
Median
⍀
0–10
spp. Median
⍀
0–10
spp. Median
⍀
0–10
spp. Median
⍀
0–10
spp. Median
⍀
0–10
spp. Median
⍀
0–10
spp.
ⱖ5000 1.5 7 1.5 7 4.4 1 3.7 19 4.0 3 1.5 3
2000–4999 1.8 22 2.2 16 2.7 5 3.3 20 3.6 22 2.5 7
1500–1999 1.8 20 1.7 10 2.7 1 3.6 11 2.7 12 2.9 6
1000–1499 2.1 30 2.3 12 0 4.2 12 2.5 34 3.1 5
750–999 2.0 34 2.5 11 10.3 1 4.1 6 3.7 37 3.0 7
500–749 2.3 58 2.6 19 1.6 2 3.7 16 3.7 72 5.9 8
300–499 2.7 76 2.7 18 2.7 2 4.7 28 4.1 121 2.4 6
200–299 3.1 73 4.0 20 3.1 1 7.3 14 5.3 111 11.7 9
150–199 3.2 51 3.3 11 3.1 1 13.9 6 5.4 78 5.1 14
100–149 4.0 68 4.0 23 29.6 3 9.5 9 7.0 118 11.9 10
75–99 5.4 48 6.3 18 11.4 5 13.7 9 9.1 74 5.4 8
50–74 6.1 48 10.9 18 1.7 3 12.6 8 11.5 104 6.1 12
25–49 8.3 81 7.8 22 8.9 11 15.8 16 14.0 148 13.3 26
10–24 14.5 89 22.1 27 27.7 12 0.0 8 14.3 112 3.0 32
ⱖ50 3.0 535 3.1 183 3.4 25 4.7 161 5.6 772 4.6 95
BCI Mudumalai Sinharaja Lambir HKK
R EPORTS
26 MAY 2000 VOL 288 SCIENCE www.sciencemag.org1416
Finally, the observation that aggregation
is weaker in larger diameter classes supports
the notion that herbivores and plant diseases
play a role in reducing aggregation. Other
evidence from the BCI and Pasoh plots indi-
cates that pests have already substantially
weakened aggregation intensity by the time
trees enter the census at 1 cm diameter (27,
28). After 1 cm diameter, the pest effect is not
particularly dramatic, and we saw no indica-
tion for further loosening of aggregations be-
tween 10- and 30-cm diameters.
References and Notes
1. M. R. T. Dale, Spatial Patterns Analysis in Plant Ecol-
ogy (Cambridge Univ. Press, Cambridge, 1999).
2. C. L. Folt and C. W. Burns, Trends Ecol. Evol. 8, 300
(1999).
3. A. R. Wallace, A Narrative of Travels on the Amazon
and Rio Negro (Haskell House, New York, 1853).
4. S. P. Hubbell, Science 203, 1299 (1979).
5. D. H. Janzen, Am. Nat. 104, 501 (1970).
6. J. H. Connell, in Dynamics of Populations,P.J.den
Boer and G. R. Gradwell, Eds. (PUDOC, Waneningen,
Netherlands, 1971), pp. 298–312.
7. M. Lieberman and D. Lieberman, in La Selva: Ecology
and Natural History of a Neotropical Rain Forest,L.A.
McDade et al., Eds. (Univ. of Chicago Press, Chicago,
1994), pp. 106–119.
8. R. Condit, Trends Ecol. Evol. 10, 18 (1995).
9. P. S. Ashton, in Forest Biodiversity Research, Monitor-
ing and Modeling: Conceptual Background and Old
World Case Studies, F. Dallmeier and J. A. Comiskey,
Eds. (UNESCO and Parthenon Publishing, Paris, 1998),
pp. 47–62.
10. V. J. Martinez, Science 284, 445 (1999).
11. P. Haase, J. Veg. Sci. 6, 575 (1995).
12. A. Getis and J. Franklin, Ecology 68, 473 (1987)
13. A. Penttinen et al., For. Sci. 38, 806 (1992).
14. We proved sample-size independence by randomly se-
lecting 100 individuals, without replacement, from all
populations and observing no change in ⍀
x
, even for
species with as many as 40,000 individuals. There is
proof of sample-size independence of this statistic (29).
To estimate confidence limits for ⍀
x
, a sample of ex-
actly half the population of each species was drawn at
random 15 times, without replacement and, for each,
⍀
x
was calculated in all 10-m distance intervals. A
variance and 95% confidence limits—based on a t
statistic—were calculated from this sample; because
the sample was halved, the limits were divided by (2)
1/2
before being applied to the entire sample. We judged
statistical significance of aggregation or overdispersion
by checking whether confidence limits included 1 and,
to compare two different estimates, we checked wheth-
er confidence limits overlapped.
15. Even under complete spatial randomness, the probabil-
ity that ⍀
0–10
⫽ 0 is high when N ⬍ 50 in 50 ha. Across
all plots, no species with N ⱖ 100 had ⍀
0–10
⫽ 0,
whereas a few species with 50 to 100 individuals had
⍀
0–10
⫽ 0. With N ⬍ 10, ⍀
0–10
⫽ 0 in most species.
For this reason, most of our analyses refer to all species
with ⱖ50 individuals in the 50-ha plots (to standardize
density, the cutoff was set at 1 per hectare, or 25
individuals in the Sinharaja plot and 52 in the Lambir
plot).
16. Supplementary material is available at www.
sciencemag.org/feature/data/1048222.shl.
17. When all stems of ⱖ1 cm diameter were analyzed, a
Spearman rank correlation between ⍀
0–10
and a spe-
cies’ abundance was significant and negative at all plots
but Mudumalai and HKK for species with N ⱖ 50.
18. F. He, P. Legendre, J. V. LaFrankie, J. Veg. Sci. 8, 105
(1997).
19. T. Okuda et al., Plant Ecol. 131, 155 (1997);
20. We compared forests by calculating confidence limits
for the mean values of log(⍀
0–10
) within abundance
categories, each species being a single datum, using t
statistics. Tests were done separately on all catego-
ries listed in Table 2, and significant differences be-
tween plots were assumed when 95% confidence
limits did not overlap. Lambir, HKK, and Sinharaja had
significantly higher ⍀
0–10
than Pasoh and BCI in at
least two abundance categories, and Lambir and Sin-
haraja were significantly higher than Pasoh in most
abundance categories.
21. P. A. Palmiotto, thesis, Yale University, New Haven,
CT, 1998.
22. A. Itoh et al., Plant Ecol. 132, 121 (1997).
23. Dispersal mode (wind, animal, or explosive) was es-
timated from our own experience and from a pub-
lished account (30). Species were classified as canopy
or understory (at their largest size) at BCI and Pasoh
from our experience and from published floras (30,
31). Significance was tested with confidence limits
for log(⍀
0–10
)(20).
24. At Pasoh, the median ⍀
0–10
for canopy species was
Fig. 2. Aggregation index (⍀
0–10
, the relative density of conspecifics within 10 m of focal trees) for
all species with ⱖ100 individuals at three plots, as a function of the abundance of each species, on
a log-log scale.
Fig. 3. Distribution maps for species also used in Fig. 1. Small circles, trees of 1 to 9.9 cm diameter;
open circles, trees of ⱖ10 cm diameter. Grid squares ⫽ 1 ha. Vatica clumps follow ridges at Lambir.
Rinorea clumps at BCI do not correlate with any known canopy, topographic, or soil feature, and
the patches are probably due to limited seed dispersal (seeds disperse from exploding capsules).
Shorea follows ridge tops at Sinharaja, and Eugenia is very rare at Sinharaja, but most individuals
are close to several conspecifics. Additional maps published elsewhere (32, 33) illustrate many
cases of habitat and dispersal limited patchiness.
R EPORTS
www.sciencemag.org SCIENCE VOL 288 26 MAY 2000 1417
2.8, and for understory it was 3.1; the difference is
not significant. At BCI, the medians were 2.9 and 4.2,
and the difference is significant.
25. The median ⍀
0–10
for wind- or explosively dispersed
species at BCI was 4.5; for animal-dispersed species it
was 3.0, but the difference was not significant.
26. At Pasoh, the 24 dipterocarp species with N ⱖ 50
individuals had a median ⍀
0–10
of 5.1, compared with
2.8 for all other species. At Lambir, 65 dipterocarps
had ⍀
0–10
⫽ 21.8; for other species, ⍀
0–10
⫽ 5.3.
Both differences are significant. At Sinharaja, 12
dipterocarps were slightly but not significantly
more aggregated than nondipterocarps. HKK and
Mudumalai had just three and one dipterocarp
species, respectively.
27. C. Wills and R. Condit, Proc. R. Soc. London Ser. B
266, 1445 (1999).
28. K. E. Harms et al., Nature 404, 493 (2000).
29. H. M. Hastings and G. Sugihara, Fractals: A User’s
Guide for the Natural Sciences (Oxford Univ. Press,
Oxford, 1993).
30. T. R. Croat, Flora of Barro Colorado Island (Stanford
Univ. Press, Stanford, CA, 1978).
31. T. C. Whitmore, Ed., Tree Flora of Malaya: A Manual
for Foresters (Longman, London, 1972).
32. N. Manokaran et al., Stand Table and Distribution of
Species in the 50-ha Research Plot at Pasoh Forest
Reserve, Research Data (Forest Research Institute of
Malaysia, Kepong, Malaysia, 1992), vol. 1.
33. R. Condit, Tropical Forest Census Plots (Springer-
Verlag and R. G. Landes Co., Berlin, and Georgetown,
TX, 1998).
34. Supported by the Indian Institute of Science, the
University of Peradeniya (Sri Lanka), the Sarawak
Forest Department (Malaysia), the Forest Research
Institute of Malaysia, the Royal Thai Forest Depart-
ment, the Smithsonian Tropical Research Institute,
the Japanese National Institute of Environmental
Studies, the Japanese Ministry of Education and Sci-
ence, the National Science Foundation, and the John
D. and Catherine T. MacArthur Foundation. R.C.
thanks J. Franklin’s group in the College of Forest
Resources at the University of Washington for sup-
port during a sabbatical.
23 December 1999; accepted 28 March 2000
Mechanism of ATP-Dependent
Promoter Melting by
Transcription Factor IIH
Tae-Kyung Kim,
1
Richard H. Ebright,
2
Danny Reinberg
1
*
We show that transcription factor IIH ERCC3 subunit, the DNA helicase re-
sponsible for adenosine triphosphate (ATP)– dependent promoter melting dur-
ing transcription initiation, does not interact with the promoter region that
undergoes melting but instead interacts with DNA downstream of this region.
We show further that promoter melting does not change protein-DNA inter-
actions upstream of the region that undergoes melting but does change in-
teractions within and downstream of this region. Our results rule out the
proposal that IIH functions in promoter melting through a conventional DNA-
helicase mechanism. We propose that IIH functions as a molecular wrench:
rotating downstream DNA relative to fixed upstream protein-DNA interactions,
thereby generating torque on, and melting, the intervening DNA.
Human transcription factor IIH consists of nine
polypeptides with masses of 31 to 90 kD (1–3).
IIH is responsible for three critical functions in
transcription: phosphorylation of the COOH-
terminal domain (CTD) of the RPB1 subunit of
RNA polymerase II (RNAPII), promoter melt-
ing, and promoter clearance.
IIH-dependent CTD phosphorylation re-
quires ATP and is mediated by the IIH cdk7
subunit, which is a cyclin-dependent protein
kinase (1–3). The role of IIH in promoter melt-
ing is to melt about one turn of DNA encom-
passing the transcription start site to yield the
“transcription bubble” (4, 5). This process re-
quires ATP (4, 5) and is mediated by the IIH
ERCC3 subunit (6, 7) (also referred to as
XPB), which, in isolation, exhibits 3⬘-5⬘ DNA-
helicase activity (6, 8, 9). The role of IIH in
promoter clearance is to stimulate escape of
transcription elongation complexes stalled at
positions ⫹10 to ⫹17 (10 –12). Like promoter
melting, promoter escape requires ATP (10 –
13) and is mediated by the IIH ERCC3 subunit
(14).
The fact that promoter melting involves
generation of single-stranded DNA (ssDNA)
and that promoter escape involves a species
containing ssDNA, together with the fact that
the IIH subunit that mediates these processes
exhibits DNA-helicase activity, has led to the
proposal that IIH functions in these processes
through a conventional DNA-helicase mecha-
nism, with direct interactions between IIH and
ssDNA (1–3, 15). However, no direct evidence
has been obtained in support of this proposal.
As a first step to understand the mechanism
1
Howard Hughes Medical Institute, Division of Nucle-
ic Acids Enzymology, Department of Biochemistry,
University of Medicine and Dentistry of New Jersey,
Robert Wood Johnson Medical School, Piscataway, NJ
08854, USA.
2
Howard Hughes Medical Institute,
Waksman Institute, and Department of Chemistry,
Rutgers University, Piscataway, NJ 08854, USA
*To whom correspondence should be addressed. E-
mail: reinbedf@umdnj.edu
Fig. 1. Results of protein-DNA photo–cross-
linking experiments. (A) Representative data.
TBFR, transcription-complex intermediate con-
taining RNAPII, TBP, IIB, IIF, and promoter DNA
(18); TBFRE, TBFR plus IIE; TCC, transcription-
ally competent complex, consisting of TBFR
plus IIE and IIH; TCC ⫹ ATP, transcriptionally
competent complex after ATP-dependent CTD
phosphorylation and promoter melting. RPB1
and RPB1-Pn denote forms of the largest sub-
unit of RNAPII having unphosphorylated CTD
and phosphorylated CTD, respectively. Data are
shown for positions ⫺2, ⫹5, ⫹13, and ⫹21 of
the DNA nontemplate strand. (B) Representative data for experiments assessing
NTP specificity (positions ⫺2, ⫹5, and ⫹13 of DNA nontemplate strand). (C)
Representative data confirming identities of cross-linked IIE␣ [(left) parallel exper-
iment with IIE␣(1-394) (31, 49)] and IIH ERCC3 [(right) immunoprecipitation with
antibody to ERCC3 of cross-linked polypeptide (32)]. (D) Representative data
demonstrating increase in fraction of complexes competent for CTD phosphoryl-
ation (top; analysis of electrophoretic mobility of cross-linked RPB1) and transcrip-
tion [bottom; quantitation of transcription (arbitrary units) and RNAPII content (1.0
unit ⫽ 20 fmol RNAPII) upon Sarkosyl washing (24)].
R EPORTS
26 MAY 2000 VOL 288 SCIENCE www.sciencemag.org1418
