Content uploaded by Peter J Morin
Author content
All content in this area was uploaded by Peter J Morin
Content may be subject to copyright.
grounded estimates of future biodiversity loss. Limitations on the reliability of this
method include the difficulty in estimating the proportion of contemporary biota in
Singapore that has not yet reached relaxation (that is, presently surviving, but committed
to future extinction)
13,25
, the unknown extent to which supplemental migration from
mainland Malaysia may have buffered local populations in Singapore from extinction, the
uncertainties in estimating the true pristine biodiversity of the island, given the likelihood
of past habitat-loss-related extinctions in Peninsular Malaysia, and the uneven
geographical distribution of endemic biodiversity ‘hotspots’
11
, which currently suffer from
higher rates of deforestation and degradation than the average of the entire region
10
.
Collectively, these potential biases suggest that our projected regional losses are likely to be
conservative.
Received 10 January; accepted 9 May 2003; doi:10.1038/nature01795.
1. Novacek, M. J. & Cleland, E. E. The current biodiversity extinction event: scenarios for mitigation and
recovery. Proc. Natl Acad. Sci. USA 98, 5466–5470 (2001).
2. Heywood, V. H., Mace, G. M., May, R. M. & Stuart, S. N. Uncertainties in extinction rates. Nature 368,
105 (1994).
3. Pitman, N. C. A. & Jørgensen, P. M. Estimating the size of the world’s threatened flora. Science 298, 989
(2002).
4. Heywood, V. H. & Stuart, S. N. in Tropical Deforestation and Species Extinction (eds Whitmore, T. C. &
Sauer, J. A.) 91–117 (Chapman and Hall, London, 1992).
5. Corlett, R. T. The ecological transformation of Singapore, 1819–1990. J. Biogeogr. 19, 411–420
(1992).
6. Turner, I. M. et al. A study of plant species extinction in Singapore: lessons for the conservation of
tropical biodiversity. Conserv. Biol. 8, 705–712 (1994).
7. Brooks, T. M., Pimm, S. L. & Collar, N. J. Deforestation predicts the number of threatened birds in
insular Southeast Asia. Conserv. Biol. 11, 382–394 (1997).
8. Brooks, T. M., Pimm, S. L. & Oyugi, J. O. Time lag between deforestation and bird extinction in
tropical forest fragments. Conserv. Biol. 13, 1140–1150 (1999).
9. May, R. M. & Stumpf, M. P. H. Species-area relations in tropical forests. Science 290, 2084–2086
(2000).
10. Achard, F. et al. Determination of deforestation rates of the world’s humid tropical forests. Science 297,
999–1002 (2002).
11. Myers, N., Mittermeier, C. G., Mittermeier, G. A., da Fonseca, G. A. B. & Kent, J. Biodiversity hotspots
for conservation priorities. Nature 403, 853–858 (2000).
12. Laurance, W. F. Reflections on the tropical deforestation crisis. Biol. Conserv. 91, 109–117 (1999).
13. Corlett, R. T. in The Ecological Consequences of Environmental Heterogeneity (eds Hutchings, M. J.,
John, E. A. & Stewart, A.) 333–355 (Blackwell Science, Oxford, 2000).
14. Milner-Gulland, E. J. & Akc¸akaya, H. R. Sustainability indices for exploited populations. Trends Ecol.
Evol. 16, 686–692 (2001).
15. Russell, G. J., Brooks, T. M., McKinney, M. M. & Anderson, C. G. Present and future taxonomic
selectivity in bird and mammal extinctions. Conserv. Biol. 12, 1365–1376 (1998).
16. Castelletta, M., Sodhi, N. S. & Subaraj, R. Heavy extinctions of forest avifauna in Singapore: lessons for
biodiversity conservation in Southeast Asia. Conserv. Biol. 14, 1870–1880 (2000).
17. Sodhi, N. S. & Liow, L. H. Improving conservation biology research in Southeast Asia. Conserv. Biol.
14, 1211–1212 (2000).
18. Droege, S., Cyr, A. & Larivee, J. Checklists: an under-used tool for the inventory and monitoring of
plants and animals. Conserv. Biol. 12, 1134–1138 (1998).
19. Williamson, M. Natural extinction on islands. Phil. Trans. R. Soc. Lond. B 325, 457–468 (1989).
20. Ceballos, G. & Ehrlich, P. R. Mammal population losses and the extinction crisis. Science 296, 904–907
(2002).
21. Diamond, J. M. Extant unless proven extinct? Or, extinct unless proven extant? Conserv. Biol. 1, 77–79
(1987).
22. Ng, P. K. L. & Wee, Y. C. (eds) The Singapore Red Data Book (The Nature Society, Singapore,
1994).
23. Bierregaard, R. O. J., Lovejoy, T. E., Kapos, V., dos Santos, A. A. & Hutchings, R. W. The biological
dynamics of tropical rainforest fragments. BioScience 42, 859–866 (1992).
24. Peters, R. H. The Ecological Implications of Body Size (Cambridge Univ. Press, New York, 1983).
25. Turner, I. M. & Corlett, R. T. The conservation value of small, isolated fragments of lowland tropical
rain forest. Trends Ecol. Evol. 11, 330–333 (1996).
26. Tilman, D., May, R. M., Lehman, C. L. & Nowak, M. A. Habitat destruction and the extinction debt.
Nature 371, 65–66 (1994).
27. Lim, K. S. & Gardner, D. Birds: An Illustrated Field Guide to the Birds of Singapore (Sun Tree Publishing
Limited, Singapore, 1997).
28. Ng, P. K. L. & Lim, K. K. P. The conservation status of the Nee Soon freshwater swamp forest of
Singapore. Aquat. Conserv. 2, 255–266 (1992).
29. Araujo, M. B. & Williams, P. H. Selecting areas for species persistence using occurrence data. Biol.
Conserv. 96, 331–345 (2000).
30. Ridley, H. N. The flora of Singapore. J. Straits’ Branch R. Asiatic Soc. 33, 27–196 (1900).
Supplementary Information accompanies the paper on www.nature.com/nature.
Acknowledgements We thank I. Turner, K. Lim and T. Leong for providing timely answers to
our taxonomic queries, and D. Bowman, R. Corlett, P. Whitehead, K. Winker and N. Yamamura
for comments on the manuscript. This research was funded by grants from the Raffles Museum of
Biodiversity Research, National University of Singapore, Australian Research Council and the
Center for Ecological Research Visiting Research Scholar Programme.
Competing interests statement The authors declare that they have no competing financial
interests.
Correspondence and requests for materials should be addressed to B.W.B.
(barry.brook@ntu.edu.au) or N.S.S. (dbsns@nus.edu.sg).
..............................................................
Productivity–biodiversity
relationships depend on the history
of community assembly
Tadashi Fukami* & Peter J. Morin†
* Department of Ecology and Evolutionary Biology, University of Tennessee,
Knoxville, Tennessee 37996-1610, USA
† Department of Ecology, Evolution, and Natural Resources, 14 College Farm
Road, Rutgers University, New Brunswick, New Jersey 08901-8551, USA
.............................................................................................................................................................................
Identification of the causes of productivity–species diversity
relationships remains a central topic of ecological research
1,2
.
Different relations have been attributed to the influence of
disturbance
3,4
, consumers
5,6
, niche specialization
7
and spatial
scale
8–14
. One unexplored cause is the history of community
assembly, the partly stochastic sequential arrival of species
from a regional pool of potential community members. The
sequence of species arrival can greatly affect community struc-
ture
15–19
. If assembly sequence interacts with productivity to
influence diversity, different sequences can contribute to vari-
ation in productivity–diversity relationships. Here we report a
test of this hypothesis by assembling aquatic microbial commu-
nities at five productivity levels using four assembly sequences.
About 30 generations after assembly, productivity–diversity
relationships took various forms, including a positive, a hump-
shaped, a U-shaped and a non-significant pattern, depending on
assembly sequence. This variation resulted from idiosyncratic
joint effects of assembly sequence, productivity and species
identity on species abundances. We suggest that the history of
community assembly should be added to the growing list of
factors that influence productivity–biodiversity patterns.
Productivity, the amount of energy available for ecosystem
development in a given location, has a major effect on species
diversity
20–22
. Until recently, hump-shaped relationships, in which
diversity peaks at intermediate productivity levels, were the most
widely observed pattern
23–25
. We now know that the relationship
takes many forms, including hump-shaped, U-shaped, positive,
negative and flat (non-significant) patterns, and that none of these
patterns predominates
1
. Possible causes of variation include the
influence of disturbance
3,4
, consumers
5,6
, niche specialization
7
and
spatial scale
8–14
, which can create variation between taxonomic
groups and habitat types
1,13
. A few studies suggest that productivity
might control the probability that alternative community states are
produced through assembly
14,26–28
.
Table 1 Introduction sequences used to assemble communities
Sequence
ABCD
.............................................................................................................................................................................
First introduction Set 1 Set 1 Set 2 Set 2
Second introduction Set 2 Set 3 Set 1 Set 3
Third introduction Set 3 Set 2 Set 3 Set 1
Set 1 Set 2 Set 3
.............................................................................................................................................................................
Blepharisma americanum*‡ Colpidium striatum* Aspidisca sp.*
Chilomonas sp.* Colpoda cucullus* Holosticha sp.*
Colpoda inflata* Euplotes sp.*†‡ Lepadella sp.* (r)
Loxocephalus sp.* Paramecium tetraurelia*† Rotaria sp.* (r)
Paramecium caudatum*† Tetrahymena vorax*‡ Spirostomum sp.*
Tetrahymena thermophila* Uronema sp.* Tillina magna*
.............................................................................................................................................................................
The natural history of these rotifers (marked with (r)) and protozoans (all others) indicates that they
consume bacteria and/or microflagellates (*), algae (†), and/or small ciliates (‡). Regardless of their
diets, all the species can sustain their population solely on bacteria and/or microflagellates and thus
potentially compete with one another.
letters to nature
NATURE | VOL 424 | 24 JULY 2003 | www.nature.com/nature 423
© 2003
Nature
Publishing
Group
We conducted a laboratory experiment using freshwater
microbial communities as a model system to test for interactive
effects of assembly and productivity on diversity. We manipulated
productivity by changing the nutrient concentration of the me-
dium. At each of the five productivity levels used, we first inoculated
the medium with bacteria, microflagellates and algae. After allowing
them to become abundant, we assembled communities by using
four different introduction sequences of 18 protozoan and rotifer
species, which consumed bacteria, microflagellates and algae, and
thus potentially competed for these shared resources (Table 1). We
used a total of 100 microcosms, namely 5 productivity levels £ 4
assembly sequences £ 5 replicates for each treatment.
Species diversity was expressed as the complement of Simpson’s
index
29
,12
l
¼ 1 2
P
p
i
ð1 2 p
i
Þ; where p
i
is the relative frequency
of species i, using data obtained 18 and 25 days after the last
introduction. We examined the relationship between productivity
and species diversity by fitting data to the following models:
model 1: d ¼ b
0
þ b
1
p
model 2: d ¼ b
0
þ b
1
p þ b
2
p
2
model 3: d ¼ b
0
þ b
1
log
10
p
model 4: d ¼ b
0
þ b
1
log
10
p þ b
2
ðlog
10
pÞ
2
where d is species diversity (1 2
l
), p is productivity (grams
protozoan pellet per litre), and b
0
, b
1
and b
2
are regression
parameters. Models 1 and 3 linearly relate species diversity to
productivity and to log-transformed productivity, respectively. A
quadratic term is added to models 1 and 3 to form models 2 and 4,
respectively, to test for curvilinearity.
The use of microcosms permitted rigorous control over assembly
history, productivity and other environmental conditions. It also
enabled us to observe patterns resulting from long-term community
dynamics. Use of an experimental duration spanning tens of gener-
ations limited the possibility that observed patterns were trivial over
ecologically important timescales. It would have been impossible to
ensure this level of experimental control in most other natural or
laboratory settings. Because all of our microcosms ultimately received
the same set of species, any differences in productivity–diversity
relations could be attributed to different assembly sequences.
The specific assembly sequence used to create communities
generated striking differences in productivity–diversity relation-
ships. Statistical analysis (see Methods) revealed positive
(F ¼ 11.23, P ¼ 0.0004, adjusted R
2
¼ 0.4601 for the best model
(model 2); Fig. 1a), non-significant (flat; F , 1.45, P . 0.2425,
adjusted R
2
, 0.0181 for all the models; Fig. 1b), hump-shaped
(F ¼ 6.31, P ¼ 0.0068, adjusted R
2
¼ 0.3066 for the best model
(model 2) and P , 0.05 for Mitchell-Olds & Shaw’s test
30
; Fig. 1c)
and U-shaped (F ¼ 19.96, P , 0.0001, adjusted R
2
¼ 0.124 for the
best model (model 2) and P , 0.05 for Mitchell-Olds & Shaw’s
test
30
; Fig. 1d) patterns under sequences A, B, C and D, respectively, 25
days after introduction of the last species (see also Supplementary
Information). This endpoint corresponded to about 30 complete
generations of the organisms involved in community dynamics.
These results also qualitatively hold for data collected 18 days
after the last introduction (Fig. 1e–h), confirming that the patterns
were temporally consistent and long-lived over ecologically import-
ant timescales (see also population dynamics in Supplementary
Information). On both days, the relationship was positive, flat
(non-significant) and hump-shaped under sequences A, B and C,
respectively. Under sequence D, the best model was positive linear
for day 18 (Fig. 1h), whereas it was U-shaped for day 25 (Fig. 1d).
However, model selection was relatively indecisive for this sequence
on day 18, and a U-shaped model (adjusted R
2
¼ 0.181, model 4)
explained data almost as well as the positive linear model selected
did (adjusted R
2
¼ 0.184, model 1).
We examined how each species responded to productivity and
Table 2 Response of species to productivity and assembly sequence
Productivity
1 2345
Species (highest) (lowest)
.............................................................................................................................................................................
Aspidisca sp. DCBA BDCA DBCA CBAD CABD
(2.89) (1.98) (1.82) (0.88) (9.22**)
Blepharisma americanum
ABDC ACDB CDAB ADCB ADCB
(13.83***) (1.69) (0.67) (3.30) (6.65**)
Chilomonas sp.
DCAB DACB DACB DACB –
(37.54***) (1.00) (1.00) (1.00) –
Colpoda cucullus
ABCD ABCD ABCD ABDC ADCB
(1.00) (0.83) (1.92) (2.63) (5.92**)
Euplotes sp.
ACBD CABD CDBACABD CABD
(12.13***) (6.78**) (4.21*) (19.50***) (5.11*)
Holosticha sp.
CADB CADB ABDC DBCA BDCA
(3.86) (24.19***) (4.07) (1.09) (0.28)
Lepadella sp.
ABCD BADC BDCA ABDC DCAB
(4.98) (7.42**) (2.20) (0.86) (1.79)
Paramecium tetraurelia
CDBA BADC ADCB ACDB ABDC
(7.75**) (4.25) (0.42) (0.82) (3.58)
Rotaria sp. BA
CD BADC BADC BDAC DBAC
(160.78***) (11.13***) (13.21***) (6.98**) (2.96)
Spirostomum sp.
CABD CABD–––
(1.03) (1.57) – – –
Uronema sp.
CBDA DBAC BDAC BDAC BDCA
(16.69***) (12.90***) (0.33) (2.37) (8.04**)
.............................................................................................................................................................................
For each species and productivity level, assembly sequences (A, B, C and D) are listed in order of
decreasing abundance on day 25. Underlined groups cannot be distinguished statistically with
Tukey’s studentized range tests. Numbers in parentheses indicate F-ratio (ANOVA). *P , 0.05;
**P , 0.005; ***P , 0.0005 (only those found to be significant after a sequential Bonferroni
correction to preserve a Type 1 error rate of 0.05 are asterisked). Dashes indicate that the
species went extinct in all replicates at the corresponding productivity level. Species not listed
here went extinct in all replicates at all productivity levels.
A
B
C
D
ae
bf
cg
dh
Day 25 Day 18
Assembly
sequence:
Species diversity (1 – )
1.0
0.5
0
1.0
0.5
0
1.0
0.5
0
1.0
0.5
0
0 0.4 0.8 0 0.4 0.8
Productivity (g food per litre)
λ
Figure 1 Response of species diversity to productivity. Data are fitted to the best
regression models. Productivity refers to protozoan pellet concentration. Some data points
are slightly moved vertically (no greater than ^0.005) from their original points so that
they can be distinguished more clearly from one another.
letters to nature
NATURE | VOL 424 | 24 JULY 2003 | www.nature.com/nature424
© 2003
Nature
Publishing
Group
assembly sequence to search for processes that might have created
different patterns. For many species, the effect of assembly sequence
on abundance depended on productivity. For example, the assembly
sequence significantly influenced the abundance of Uronema sp. at
three out of five productivity levels (Fig. 2). When significant, the
effect of assembly sequence on abundance also depended on
productivity. Uronema was more abundant in sequence C than in
sequence D at one productivity level (Fig. 2a), but the difference was
reversed at another level (Fig. 2b) and became non-significant at the
other levels (Fig. 2c–e). Thus, productivity determined whether a
certain sequence either facilitated or inhibited population growth.
Furthermore, the effect of sequence also depended on species
identity (Table 2). Joint effects of assembly sequence, productivity
and species identity were highly idiosyncratic, with no discernible
general trends across species (Table 2).
Our results show that community assembly potentially interacts
with productivity to create a remarkable variety of productivity–
diversity patterns. We emphasize that different assembly sequences
alone produced various productivity–diversity patterns, without
experimental imposition of variation in disturbance
3,4
,con-
sumers
5,6
, niche specialization
7
or spatial scale
8–14
. We suspect that
assembly sequence might have influenced the operation of these
unmanipulated proximate mechanisms
17
. For example, some
sequences might have caused consumers and niche specialization,
realized by diet differences between species (Table 1), to exert a
strong effect on productivity–diversity patterns, whereas other
sequences might have minimized their influence. The next step in
understanding this phenomenon would be to uncover general
mechanisms that explain how assembly produces such diverse
patterns. Our results caution against assuming that a single expla-
natory mechanism at the community level exists, because inter-
actions between species apparently depend in a complex way on
assembly sequence, productivity and species identity (Table 2).
Our interpretation of these data assumes that the spatial scale of
productivity variation is smaller than that for assembly sequences.
Some natural systems meet this assumption. Examples include
situations where regular seasonal phenology of species arrival drives
assembly or where species invade a region through biogeographic
events and spread before the next species invade. Other systems
contain local sites that vary in both productivity and assembly
sequence. However, we note that assembly has important impli-
cations even for these systems: it can make patterns sensitive to the
scale of observation
14
. We found that the productivity–diversity
relationship was positive linear at a local scale (P ¼ 0.0017 and
adjusted R
2
¼ 0.3255 for model 1, which was the best model),
whereas it was non-significant at a regional scale (P . 0.02 for all
models; note that a ¼ 0.0125 with Bonferroni correction; see
Supplementary Information for how we calculated local and
regional diversity).
We studied only four of the many possible assembly sequences
that could have been used with our species pool. This small sample
of assembly sequences makes it impossible to say whether any one of
the specific patterns that we observed might predominate in a larger
sample. Nevertheless, it is possible that the variety of productivity–
biodiversity relationships seen in nature reflects differences in
assembly as well as other factors. We suggest that the history of
community assembly should be added to the growing list of factors
that influence productivity–biodiversity patterns. Assembly history
does not preclude the importance of other factors. However,
because the detailed assembly history of natural communities is
seldom known, it might be difficult to deduce the proximal causes of
natural productivity–diversity patterns unambiguously. For this
reason, manipulative experiments remain an essential tool for
exploring the possible causes of productivity–diversity relationships
within the historical context of community assembly. A
Methods
Microcosms
Microcosms were covered sterile 118-ml polypropylene containers. These containers were
filled with 30 ml of medium (made from Carolina Biological Supply protozoan pellet,
Herpetivite powdered vitamin supplement, and soil in well water) and kept at 22 8Cwith
14 h light/10 h dark cycles. The medium was autoclaved before use and inoculated with
bacteria (Bacillus subtilis, Bacillus cereus, Proteus vulgaris, Serratia marcescens and other
unidentified bacteria filtered from the stock cultures of all the protozoan and rotifer
species used in the experiment), microflagellates and algae (Chlamydomonas spp.). These
inoculations were done before the medium was distributed to microcosms.
Manipulating productivity
We used five levels of productivity, evenly spaced on a logarithmic scale. The medium for
the lowest productivity level consisted of 0.007g of the protozoan pellet, 0.033 g of the soil
and 0.001 g of the vitamins in each litre of well water. The media for the other productivity
levels consisted of 0.023, 0.073, 0.232 and 0.733 g of the pellet, 0.106, 0.334, 1.055 and 3.335 g
of the soil, and 0.004, 0.013, 0.042 and 0.133 g of the vitamins per litre, respectively, from the
second lowest to the highest levels. Removal and replacement of 3 ml (that is, 10%) of the
medium of the corresponding nutrient concentration once a week renewed nutrients.
Manipulating assembly sequence
Protozoan and rotifer species were introduced to the microcosms sequentially in
accordance with predetermined schedules (Table 1). Stock cultures of six species were used
for each introduction. First species were introduced 3 days after the algal inoculation.
Subsequent introductions had 14-day intervals. Microcosms received a very small number
of individuals, namely less than 0.7% of carrying capacity, but at least 20 individuals per
species to preclude trivial extinction by chance at initial stages. To standardize the number
of individuals introduced across introduction occasions, population densities in stock
cultures were estimated and, if necessary, diluted before introductions. The age of the
stock cultures at the time of species introductions was also standardized to minimize
variation in physiological conditions of species between different introduction occasions.
Microcosms also contained three additional algal species. Because these algae were most
probably from cultures of a particular species (Euplotes sp.), they did not confound the
effect of assembly sequence.
Abundance
123 45
X
YZ
XY
Z
X
X
YY
NS NS
X
Y
Y
Y
a b c d e
4
3
2
1
YZ
XY
Z
X
YY
NS NS
Y
Y
Y
Productivity level
(highest) (lowest)
(log
10
((cells per ml)+1))
ABCD ABCD ABCD ABCD ABCD
Assembly sequence
Figure 2 Response of the abundance of Uronema sp. to productivity and assembly
sequence on day 25. Bars with the same letter above them did not differ in Tukey’s
studentized range tests. We applied a sequential Bonferroni correction to preserve a
Type I error rate of 0.05. NS, not significant (see Table 2). Error bars are s.e.m.
letters to nature
NATURE | VOL 424 | 24 JULY 2003 | www.nature.com/nature 425
© 2003
Nature
Publishing
Group
Measuring population abundances
The population abundance of each species in each microcosm was measured once a week
until 25 days after the last introduction. Twenty-five days corresponds to roughly 30
generations of the protozoa and rotifer species. Densities were estimated by counting
protozoa and rotifers in samples of known volume, typically 0.3 ml, from the 3 ml of
medium removed for nutrient replacement. When species were too abundant to count
reliably, the sample was diluted. When one or more species were absent from the 0.3-ml
sample, the entire 3 ml of medium (and the entire microcosm on the last sampling
occasion) was scanned and protozoa and rotifers were counted.
Productivity–diversity relationships
When P . 0.0125 (that is, 0.05/4, using a Bonferroni correction to retain a Type I error
rate of 0.05) for all models (see model description in text), we concluded that the
relationship between productivity and diversity was not significant. When P , 0.0125 for
more than one model, we selected as the best model the one that had the highest value of
adjusted R
2
. Adjusted R
2
values are adjusted for the number of parameters in the models.
When model 2 or 4 was selected, we determined whether the relationship was hump-
shaped or U-shaped by using Mitchell-Olds & Shaw’s test
30
(see Supplementary
Information). We focused on the diversity of protozoan and rotifer species; our diversity
index does not include bacteria, microflagellates or algae. Productivity–diversity
relationships depended on assembly history when species richness (number of species per
unit volume), rather than the complement of Simpson’s index, was also used to express
species diversity (see Supplementary Information).
Response of species
We conducted analyses of variance to test for effects of introduction sequence on species
abundance at each productivity level. Abundance was transformed as log
10
((individuals
per ml) þ 1) before analysis to minimize heteroscedasticity. For each species we used a
sequential Bonferroni correction for the five tests corresponding to the five productivity
levels to preserve a Type I error rate of 0.05. When analyses of variance found a significant
effect of sequence, Tukey’s studentized range tests were used to identify which treatments
differed.
Received 18 March; accepted 8 May 2003; doi:10.1038/nature01785.
1. Waide, R. B. et al. The relationship between productivity and species richness. Annu. Rev. Ecol. Syst.
30, 257–300 (1999).
2. Morin, P. J. Biodiversity’s ups and downs. Nature 406, 463–464 (2000).
3. Huston, M. A. Biological Diversity: The Coexistence of Species on Changing Landscapes (Cambridge
Univ. Press, 1994).
4. Kondoh, M. Unifying the relationships of species richness to productivity and disturbance. Proc.
R. Soc. Lond. B 268, 269–271 (2001).
5. Worm, B. et al. Consumer versus resource control of species diversity and ecosystem functioning.
Nature 417, 848–851 (2002).
6. Leibold, M. A. et al. Species turnover and the regulation of trophic structure. Annu. Rev. Ecol. Syst. 28,
467–494 (1997).
7. Kassen, R. et al. Diversity peaks at intermediate productivity in a laboratory microcosm. Nature 406,
508–512 (2000).
8. Currie, D. J. Energy and large-scale patterns of animal- and plant-species richness. Am. Nat. 137,
27–49 (1991).
9. Wright, D. H., Currie, D. J. & Maurer, B. A. in Species Diversity in Ecological Communities: Historical
and Geographical Perspectives (eds Ricklefs, R. & Schluter, D.) 66–74 (Univ. Chicago Press, 1993).
10. Abrams, P. A. Monotonic or unimodal diversity–productivity gradients: What does competition
theory predict? Ecology 76, 2019–2027 (1995).
11. Gross, K. L. et al. Patterns of species density and productivity at different spatial scales in herbaceous
plant communities. Oikos 89, 417–427 (2000).
12. Scheiner, S. M. et al. Species richness, species–area curves and Simpson’s paradox. Evol. Ecol. Res. 2,
791–802 (2000).
13. Mittelbach, G. G. et al. What is the observed relationship between productivity and diversity? Ecology
82, 2381–2396 (2001).
14. Chase, J. M. & Leibold, M. A. Spatial scale dictates the productivity–biodiversity relationship. Nature
416, 427–430 (2002).
15. Gilpin, M. E. & Case, T. J. Multiple domains of attraction in competition-communities. Nature 261,
40–42 (1976).
16. Post, W. M. & Pimm, S. L. Community assembly and food web stability. Math. Biosci. 64, 169–192
(1983).
17. Drake, J. A. Community–assembly mechanics and the structure of experimental species ensemble.
Am. Nat. 137, 1–26 (1991).
18. Wilson, D. S. Complex interactions in metacommunities, with implications for biodiversity and
higher levels of selection. Ecology 73, 1984–2000 (1992).
19. Law, R. & Morton, R. D. Alternative permanent states of ecological communities. Ecology 74,
1347–1361 (1993).
20. Connell, J. H. & Orias, E. The ecological regulation of species diversity. Am. Nat. 98, 399–414 (1964).
21. Leigh, E. G. Jr On the relationship between productivity, biomass, diversity and stability of a
community. Proc. Natl Acad. Sci. USA 53, 777–783 (1965).
22. Pianka, E. R. Latitudinal gradients in species diversity: a review of concepts. Am. Nat. 100, 33–46 (1966).
23. Rosenzweig, M. L. Species diversity gradients: We know more and less than we thought. J. Mammal.
73, 715–730 (1992).
24. Rosenzweig, M. L. Species Diversity in Space and Time (Cambridge Univ. Press, 1995).
25. Tilman, D. & Pacala, S. in Species Diversity in Ecological Communities: Historical and Geographical
Perspectives (eds Ricklefs, R. & Schluter, D.) 13–25 (Univ. Chicago Press, 1993).
26. Van de Koppel, J. et al. Patterns of herbivory along a productivity gradient: An empirical and
theoretical investigation. Ecology 77, 736–745 (1996).
27. Holt, R. D. & Polis, G. A. A theoretical framework for intraguild predation. Am. Nat. 149, 745–764
(1997).
28. Chase, J. M. Food web effects of prey size refugia: Variable interactions and alternative stable
equilibria. Am. Nat. 154, 559–570 (1999).
29. Simpson, E. H. Measurement of diversity. Nature 163, 688 (1949).
30. Mitchell-Olds, T. & Shaw, R. E. Regression analysis of natural selection: Statistical inference and
biological interpretation. Evolution 41, 1149–1161 (1987).
Supplementary Information accompanies the paper on www.nature.com/nature.
Acknowledgements We thank members of the Morin laboratory for discussion, and J. A. Drake,
C. M. K. Kaunzinger, M. A. Leibold, Z. T. Long, P. B. Rainey and D. Simberloff for comments. The
National Science Foundation and the Department of Ecology and Evolutionary Biology at the
University of Tennessee supported this research.
Competing interests statement The authors declare that they have no competing financial
interests.
Correspondence and requests for materials should be addressed to T.F. (tfukami@utk.edu).
..............................................................
Theroleofneuronalidentityin
synaptic competition
Narayanan Kasthuri & Jeff W. Lichtman
Department of Anatomy and Neurobiology, Washington University School of
Medicine, St Louis, Missouri 63110, USA
.............................................................................................................................................................................
In developing mammalian muscle, axon branches of several
motor neurons co-innervate the same muscle fibre. Competition
among them results in the strengthening of one and the with-
drawal of the rest
1,2
. It is not known why one particular axon
branch survives or why some competitions resolve sooner than
others
3
. Here we show that the fate of axonal branches is strictly
related to the identity of the axons with which they compete.
When two neurons co-innervate multiple target cells, the losing
axon branches in each contest belong to the same neuron and are
at nearly the same stage of withdrawal. The axonal arbor of one
neuron engages in multiple sets of competitions simultaneously.
Each set proceeds at a different rate and heads towards a common
outcome based on the identity of the competitor. Competitive
vigour at each of these sets of local competitions depends on a
globally distributed resource: neurons with larger arborizations
are at a competitive disadvantage when confronting neurons with
smaller arborizations. An accompanying paper tests the idea that
the amount of neurotransmitter released is this global resource
4
.
A central feature of mammalian neural development is the
reapportionment of synaptic contacts such that neurons progress-
ively innervate fewer postsynaptic cells but with more synapses
5–7
.
Synapse elimination at the skeletal neuromuscular junction is
currently the best studied of all such rearrangements and viewed
by some as a model for changes that occur in the developing brain.
In the neuromuscular system of neonatal rodents, the number of
muscle fibres contacted by one motor neuron decreases during the
first two postnatal weeks until each muscle fibre is innervated by
only one axon
8
. Within the arbor of a single motor axon, this branch
withdrawal is protracted and asynchronous; after some terminal
branches have definitively won or lost at some neuromuscular
junctions, other branches of the same neuron still share synaptic
sites with other innervating axons
3
. Several lines of evidence suggest
that competitions between axon branches underlie this process
9,10
.It
is not known, however, what properties of an axonal branch or its
environment determine its destiny in these competitions. Here, we
ask whether the competitive vigour of each axon branch is deter-
mined by local factors or rather is set by a global property of the
parent neuron. If axonal branches were acting as agents of their
letters to nature
NATURE | VOL 424 | 24 JULY 2003 | www.nature.com/nature426
© 2003
Nature
Publishing
Group
