This document summarizes evidence that the ant species Myrmelachista schumanni creates structures called "devil's gardens" in the Amazon rainforest by killing surrounding plants and galling, or creating swellings, on the trunks of canopy trees. The ant colonies live within these galls and the hollow stems of their host ant-plants, expanding their nesting space. Observations of swollen scars on non-ant trees surrounding devil's gardens contained networks of cavities inhabited by M. schumanni, supporting the hypothesis that this ant species can gall other trees to create additional housing for their large colonies.

1. vol. 174, no. 5 the american naturalist november 2009
Natural History Note
A Plant Needs Ants like a Dog Needs Fleas: Myrmelachista schumanni
Ants Gall Many Tree Species to Create Housing
David P. Edwards,1,* Megan E. Frederickson,2,3,* Glenn H. Shepard,4
and Douglas W. Yu5,6,†
1. Institute of Integrative and Comparative Biology, University of Leeds, Leeds LS2 9JT, United Kingdom; 2. Society of Fellows and
Museum of Comparative Zoology, Harvard University, Cambridge, Massachusetts 02138; 3. Department of Ecology and Evolutionary
Biology, University of Toronto, 25 Harbord Street, Toronto, Ontario M5S 3G5, Canada; 4. Museum of Archeology and Ethnology,
University of São Paulo, Avenida Prof. Almeida Prado 1466, São Paulo, SP 05508-900, Brazil; 5. State Key Laboratory of Genetic
Resources and Eolution; Ecology, Conservation and the Environment Center; Kunming Institute of Zoology, Chinese Academy of
Science, Kunming, Yunnan 650223, China; 6. Centre for Ecology, Evolution, and Conservation and School of Biological Sciences,
University of East Anglia, Norwich NR4 7TJ, United Kingdom
Submitted May 28, 2009; Accepted July 30, 2009; Electronically published October 2, 2009
abstract: Hundreds of tropical plant species house ant colonies in
specialized chambers called domatia. When, in 1873, Richard Spruce
likened plant-ants to fleas and asserted that domatia are ant-created
galls, he incited a debate that lasted almost a century. Although we
now know that domatia are not galls and that most ant-plant inter-
actions are mutualisms and not parasitisms, we revisit Spruce’s sug-
gestion that ants can gall in light of our observations of the plant-ant
Myrmelachista schumanni, which creates clearings in the Amazonian
rain forest called “supay-chakras,” or “devil’s gardens.” We observed
swollen scars on the trunks of nonmyrmecophytic canopy trees sur-
rounding supay-chakras, and within these swellings, we foundnetworks
of cavities inhabited by M. schumanni. Here, we summarize the evi-
dence supporting the hypothesis that M. schumanni ants make these
galls, and we hypothesize that the adaptive benefit of galling is to
increase the amount of nesting space available to M. schumanni
colonies.
Keywords: ant-plant interactions, galls, myrmecophytes, mutualism,
parasitism.
Hundreds of tropical plant species obligately host ant col-
onies within hollow branches, trunks, or leaves. The or-
igins of these plants, called ant-plants or myrmecophytes,
and the benefits of their associations with ants were de-
bated by naturalists for nearly a century (Webber et al.
2007). In a letter to Alfred Russell Wallace in 1873 (Wallace
1905, pp. 64–65), the botanist Richard Spruce proposed
that the leaf pouches and stem cavities of several tropical
* These two authors contributed equally to this publication.
†
Corresponding author; e-mail: dougwyu@gmail.com.
Am. Nat. 2009. Vol. 174, pp. 734–740. 䉷 2009 by The University of Chicago.
0003-0147/2009/17405-51303$15.00. All rights reserved.
DOI: 10.1086/606022
plant genera had resulted from the “unceasing operations
of ants” producing, via Lamarckian adaptation, inherited
“excrescence[s].” Spruce disputed the notion that trees
receive any benefits from ants, writing, “the ants cannot
be said to be useful to the plants, any more than fleas and
lice are to animals.” As Wallace (1905, p. 65) pointed out,
however, Spruce could not have known at the time of
Thomas Belt’s (1874) observations on the bull’s horn aca-
cia, which was observed to provide food rewards and hol-
low thorns for its “standing army [of ants] kept for the
protection of the plant.”
In this exchange, the competing hypotheses over ant-
plants were established: either the hollow plant structures
(subsequently called domatia) were galls created by ants
(R. Spruce, 1873, cited in Wallace 1905; Becarri 1886–1887
cited in Uphof 1942; Chodat and Carisso 1920; Wheeler
1942), in which case the relationship would be deemed
parasitic, or the domatia were a normal part of plant de-
velopment (e.g., Darwin 1877; Bequaert 1922; Bailey 1924)
and the relationship could be considered mutually benefi-
cial. However, it was not until 1966, with the publication
of Daniel Janzen’s (1966) experimental study of bull’s horn
acacia plants in Mexico, that the ants-as-parasites stance
was finally upended. It is now abundantly clear that doma-
tia are not galls and that most ant-plant relationships are
mutualistic (Davidson and McKey 1993; Heil and McKey
2003). However, not all ant-plant relationships are mu-
tualistic (e.g., Janzen 1975; Yu and Pierce 1998; Gaume
and McKey 1999; Gaume et al. 2005), and it is in this
context that we return to Spruce’s original hypothesis that
ants can gall plants to create housing.
We focus our attention on “devil’s gardens,” which are
clearings in the rain forest where only one, two, or at most
three tree species grow. Devil’s gardens occur throughout

2. Ants Can Gall Trees 735
the western Amazon and differ markedly from the sur-
rounding rain forest, which is hyperdiverse (Gentry 1988).
The term “devil’s garden” is a loose translation of the
Quechua word supay-chakra, the name given to these
clearings by the Andean peoples who have colonized the
lowland rain forests of Peru. It is widely believed by both
Andean colonists and many indigenous peoples living in
the region that supay-chakras are cultivated by an evil
forest spirit (M. P. Gilmore, S. Rı́os-Ochoa, and S. Rı́os-
Flores, unpublished manuscript), hence their name.
Supay-chakras are actually created by Myrmelachista
schumanni ants (Frederickson et al. 2005). The trees and
plants that do grow in supay-chakras are ant-plants, and
M. schumanni nests in their hollow stem swellings or leaf
pouches. Myrmelachista schumanni workers actively patrol
supay-chakras, and when they come across plants other
than their myrmecophytic hosts, they attack them. During
an attack, each of hundreds of M. schumanni workers bites
a small hole in a leaf or a stem with its mandibles and
then inserts the tip of its gaster into the hole and releases
droplets of formic acid (Frederickson et al. 2005). Shortly
thereafter, the plant begins to turn brown near the wound
sites, and the necrosis gradually spreads, usually along the
leaf veins. Eventually, the plant wilts, sheds its leaves, and
dies.
In different regions, supay-chakras are dominated by
different species of ant-plants, although they are always
inhabited by Myrmelachista ants (Frederickson and Gor-
don 2007). In southeastern Peru, supay-chakras consist
mostly of Cordia nodosa (Boraginaceae) and the occasional
Tococa guianensis (Melastomataceae). In northeastern Peru
and southeastern Ecuador, the most common ant-plant in
supay-chakras is Duroia hirsuta (Rubiaceae), although C.
nodosa is often also present (Olesen et al. 2002; Freder-
ickson 2005; Frederickson and Gordon 2007). At slightly
higher elevations, supay-chakras consist primarily of T.
guianensis (Morawetz et al. 1992) or of a mix of T. gui-
anensis and Clidemia heterophylla (Melastomataceae; Ren-
ner and Ricklefs 1998). Although it is possible that more
than one species of Myrmelachista makes supay-chakras,
we have collected M. schumanni from D. hirsuta, C. nodosa,
and T. guianensis trees growing in supay-chakras in both
northern and southern Peru, suggesting that M. schumanni
is the main supay-chakra ant species.
Each supay-chakra is inhabited by a single, polygynous
colony of M. schumanni that can have as many as 3 million
workers and 15,000 queens (Frederickson et al. 2005). Like
the colonies of many other plant-ants (Fonseca 1999; Ed-
wards et al. 2006), M. schumanni colonies appear to be
nest site limited (Frederickson and Gordon 2009), and by
killing non-ant-plants, M. schumanni colonies promote the
growth and establishment of their myrmecophytic hosts
and thus gain more housing (Frederickson et al. 2005;
Frederickson and Gordon 2007, 2009). Here, we describe
for the first time how M. schumanni ants sometimes also
excavate chambers in nonmyrmecophytic trees in order to
increase the nesting space available to their colonies.
We owe this discovery to the traditional ethnobiological
knowledge and folklore of the people living in the western
Amazon, particularly the Matsigenka indigenous people
from the native community of Yomybato, who first
brought it to our attention. Yomybato is located inside
Manu National Park in southeastern Peru (Terborgh 1990;
Shepard et al. 2001, 2009; 11.802625⬚S, 71.910933⬚W, ∼380
m asl). The habitat is moist-to-seasonal tropical rain forest
(2,000–2,600 mm rainfall per year), with a major distinc-
tion between recently formed alluvial plains (lowland for-
est) and older elevated terraces or hills (upland or terra
firme forest). In several locations in the upland forests
around Yomybato, M. schumanni ants inhabit C. nodosa
trees, allowing C. nodosa, and a few T. guianensis, to es-
tablish supay-chakras. Although aware of the ant-plant
mutualism at work, the Matsigenka interpret these for-
mations as “spirit clearings” and believe they represent
invisible villages inhabited by benevolent spirits who serve
as guides and helpers to shamans (Shepard 1998). In 1996,
Matsigenka research collaborators showed Yu and Shepard
the swollen, rugose trunks of several hardwood canopy
trees, none of them ant-plants, around the periphery of
the spirit clearing. They explained that the scars were evi-
dence of the fires set by the invisible spirits, who are be-
lieved to clear and burn swidden gardens in the forest
around their villages (fig. 1a–1c), much as the Matsigenka
themselves do. Cutting into the swollen trunks in fact
revealed a network of cavities that extended around the
circumference, inhabited by M. schumanni workers, brood,
and queens, plus their associated pseudococcids (fig. 1d).
Subsequently, we made cross sections, which revealed that
the chambers extend to the center of the boles and form
intricate passageways (fig. 1e, 1f ). In a few cases, we have
observed that trees with these chambers are weakened to
the extent that they collapse under their own weight, either
because the chambers have caused early mortality or be-
cause the chambers increase susceptibility to wind throw.
In a survey of all subcanopy and canopy trees along a
-m transect that crossed the center of one 20-m-
50 # 10
diameter supay-chakra at Yomybato, 63 individual trees
were recorded, of which 45 (71%) had chambers that were
inhabited by M. schumanni (table 1). Furthermore, the
trees belonged to 21 different plant families, of which 15
families had chambers (table 1). Thus, only a few tree
species lacked these structures, including all palms (Are-
caceae) and some trees with smooth (Capirona decorticans,
Rubiaceae) or peeling bark (Miconia alata, Melastoma-
taceae). Finally, the trunks of some of the C. nodosa and
T. guianensis plants themselves had chambers.

3. 736 The American Naturalist
Figure 1: a, Maximo Vicente-Zakaro, a Matsigenka native, standing by a swollen and scarred trunk in a “spirit clearing” near Yomybato Native
Community, Manu, Peru. b, Swollen trunk at Los Amigos Research Center, Peru. c, Swollen and scarred trunk, Los Amigos Research Center. d,
Myrmelachista schumanni ants and brood in a chamber within a swollen trunk. e, Cross section of a trunk reveals that chambers can extend to the
center. f, Lengthwise section of another swollen trunk, with passageways and chambers.
More recently, in 2005–2007, Frederickson surveyed
supay-chakras at the Los Amigos Research Center
(12.568611⬚S, 70.099167⬚W, elevation ∼230 m), which is
about 200 km southeast of Yomybato. Within 5 km of the
research center, Frederickson found a total of seven supay-
chakras, all in terra firme forest. Each was inhabited by a
colony of M. schumanni ants and had between 2 and 19
C. nodosa trees ( ) growing together in a
mean p 6.8
clump. Around the periphery of all seven patches, there
were several nonmyrmecophytic trees with noticeably
swollen, gnarled trunks (fig. 1b, 1c). As in Yomybato, closer
inspection revealed that M. schumanni workers, brood,
queens, and their associated scale insects were nesting in-
side small cavities in these trunks. In all cases, the cavities
were restricted to the swollen portions of the trunks, which
were about 60 cm to 1.4 m off the ground (fig. 1b, 1c).
And at Los Amigos, as in Manu, M. schumanni inhabited
the swollen trunks of many different nonmyrmecophytic
tree species, including Pourouma sp. (Urticaceae) and Vi-
rola sp. (Myristicaceae), but never any of the palms or tree

4. Ants Can Gall Trees 737
Table 1: Trees with chambers in a -m transect, Yomybato, Peru
50 # 10
No. species No. trees
No. trees with
chambers
Annonaceae 1 1 1
Apocynaceae 1 1 1
Arecaceae 3 4 0
Bignoniaceae 1 1 0
Bombacaceae 1 1 1
Boraginaceae 1 5 4
Chrysobalanaceae 1 1 1
Celastraceae 1 1 1
Dilleniaceae 2 2 1
Ebenaceae 1 1 1
Elaeocarpaceae 1 1 0
Euphorbiaceae 2 3 3
Fabaceae 4 9 8
Lauraceae 1 1 0
Lecythidaceae 1 1 0
Melastomataceae 3 12 9
Moraceae 2 6 4
Myristicaceae 1 4 3
Rubiaceae 2 2 1
Sapotaceae 1 1 1
Dead 2 5 5
Total 63 45
Note: Each tree was identified to species or morphospecies and scored for the
presence of chambers within the cambium.
ferns that grew nearby. In two of the seven gardens at Los
Amigos, M. schumanni ants were also found nesting in
chambers within C. nodosa trunks, in addition to nesting
in C. nodosa domatia.
Frederickson also observed similar cavities in the swol-
len trunks of several nonmyrmecophytic trees in four
supay-chakras at the Las Piedras Biodiversity Station
(12.057278⬚S, 69.543694⬚W, elevation ∼200 m), about 83
km northeast of Los Amigos. These chambers held not
only M. schumanni workers, brood, and queens but even
winged males. Such chambers also occur, but to a much
lesser extent, on trees in supay-chakras in Loreto, Peru,
some 1,000 km to the north. However, in Loreto, the
chambers are restricted mostly to the trunks of M. schu-
manni–occupied ant-plants (principally D. hirsuta) and
only very rarely occur on nonmyrmecophytic trees (M. E.
Frederickson, personal observation). Finally, despite sev-
eral person-decades of working in the rain forests of south-
ern and northern Peru, we have never observed these
chambers on trees outside of the immediate surroundings
of ant-plant patches occupied by M. schumanni colonies.
Combined, these observations strongly suggest that M.
schumanni ants are the causal agent of these abnormal
growths.
Many arthropods, including mites, midges, aphids,
wasps, and sawflies, are able to create galls in the cambium
layers of tree branches and stems (Taft and Bissing 1988;
Ronquist and Liljeblad 2001; McIntyre and Whitham 2003;
Price 2005; Sliva and Shorthouse 2006). Gall-forming in-
sects inject chemicals (possibly mimics of plant hormones;
Taft and Bissing 1988 and references therein) into and/or
mechanically damage the plant’s periderm (bark) or cor-
tex, and the resulting abnormal hollow outgrowths are
used to house larvae (Taft and Bissing 1988; Redfern and
Shirley 2002). We propose that M. schumanni ants create
chambers in a similar manner, although we can only spec-
ulate about the mechanism. Myrmelachista schumanni is
unique among ant species in using formic acid as an her-
bicide to kill plants in its gardens (Frederickson et al.
2005). Perhaps M. schumanni workers also use a combi-
nation of mechanical damage and chemical attack to pro-
duce the abnormal outgrowths and nest chambers that we
observed on nonmyrmecophytic tree trunks inhabited by
M. schumanni (fig. 1). We do not know why M. schumanni
poisons and kills some nonmyrmecophytic trees and galls
others, although stem size is likely one determining factor.
Although there are many gall-forming species among
the Hymenoptera, we know of only one other possible
example of galling by ants, in which workers of an un-
identified Pseudomyrmex ant species excavate pith from

5. 738 The American Naturalist
young twigs of the tree Vochysia vismiaefolia and the twigs
subsequently swell to form domatia (Blüthgen and Wes-
enberg 2001). Mechanical drilling by the experimenters
also induced swelling in new twigs, but unfortunately,
other tree species, including two sympatric congeners,
were not tested in the same way, so it is not known whether
to interpret the swellings as galls per se or as induced
domatia (in the same sense that an obligately myrmeco-
phytic ant Pheidole bicornis ant is known to induce food
production in its host plant Piper cenocladum [Risch and
Rickson 1981]).
Myrmelachista schumanni thus appears to be the first
ant species found to make galls sensu stricto and the only
one to gall multiple plant species. Other wood-dwelling
ants typically create their housing by boring into dead
wood only (e.g., Camponotus carpenter ants; Chen et al.
2002; see also Hölldobler and Wilson 1990). Some ant
species shelter their brood in galls, but they depend on
galls that were made by other insects (Bequaert 1922;
Araujo et al. 1995; Carver et al. 2003), while still others
feed on honeydew secreted by galls (Abe 1992; Fernandes
et al. 1999; Inouye and Agrawal 2004).
We hypothesize that the adaptive benefit of galling is to
increase the amount of nesting space available to M. schu-
manni colonies. Because the colonies are polygynous and
can have thousands of queens, egg production is probably
not limiting, and colonies can quickly outgrow their ex-
isting lodgings. Myrmelachista schumanni colonies occupy
virtually all of the domatia on every myrmecophytic tree
they inhabit, no matter how large the trees get (Freder-
ickson and Gordon 2009). Furthermore, colony fecundity
is known to be highly correlated with the number of ant-
plants inhabited by the ant colony, suggesting that colony
fitness is tied to nest space (Frederickson and Gordon
2009), a general feature of plant-ants (Fonseca 1999). Myr-
melachista schumanni colonies shelter both developing
brood and scale insects in the galled tree trunks we ob-
served (as they do in domatia), so in addition to providing
more space for the ants to rear broods, galling should also
result in more food for the ant colony.
For M. schumanni colonies, nest sites appear to be
scarcer in southern than in northern Peru, perhaps ex-
plaining why the galling of nonmyrmecophytic trees is
more commonly observed in the south. In southern Peru,
the ant-plant patches inhabited by M. schumanni colonies
are typically much smaller than in northern Peru; at Los
Amigos, an M. schumanni colony occupies an average of
6.8 C. nodosa trees, while in Loreto, a colony occupies an
average of 23 D. hirsuta trees (and can occupy as many
as 594 D. hirsuta trees; Frederickson and Gordon 2009).
In southern Peru, the growth of M. schumanni colonies
may be more rapid than the growth of the C. nodosa stands
they inhabit, creating a need for additional nesting space
by galling trees, whereas the growth rate of D. hirsuta–
dominated stands may not differ dramatically from that
of M. schumanni colonies. In Loreto, M. schumanni col-
onies occupying fewer than 22 D. hirsuta trees did not
produce any female alates (Frederickson and Gordon
2009), suggesting that in southern Peru, M. schumanni
may depend on galling nonmyrmecophytic trees in order
for colonies to grow to a large enough size that they pro-
duce virgin queens capable of founding new colonies.
In general, ants are excellent “ecosystem engineers.” For
example, it is well known that ants that make their nests
in the soil do so in such a way as to create favorable
conditions, such as the right temperature and humidity,
for the growth of their colony. It turns out that ants that
nest in plants are no different. Myrmelachista schumanni
workers also appear to behave so as to create the right
environment for the growth of their colonies. In effect,
M. schumanni ants grow their own nests. To speed the
growth of their myrmecophytic host trees, they protect
their host plants against insect herbivores (Frederickson
2005), and they poison their plants’ competitors with for-
mic acid (Frederickson et al. 2005). Here, we have provided
the first evidence to suggest that, when necessary, M. schu-
manni workers also gall nonmyrmecophytic trees in order
to provide food and shelter for their colony. On one oc-
casion, Yu has observed M. schumanni workers destroying
a floral bud of a C. nodosa plant. Such behavior, if applied
to many flowers, is known to increase vegetative growth
in plants inhabited by the parasitic ant Allomerus octoar-
ticulatus (Yu and Pierce 1998; Frederickson 2009). It is not
known, however, whether Myrmelachista ants castrate
widely, since abundant fruits are produced by their host
plants (Frederickson and Gordon 2009; D. W. Yu, personal
observation).
Since Janzen (1966), almost every ant-exclusion exper-
iment conducted on an ant-plant has confirmed that plant-
ants are protection-mutualists. Yet it appears that in the
case of M. schumanni, Spruce’s interpretation that the op-
erations of ants can produce excrescences, although not
in a Lamarckian sense, was in fact correct. Of course, much
as it took Janzen’s ant-exclusion experiment to show de-
finitively that ants benefit ant-acacias, it will take a ma-
nipulative experiment to demonstrate beyond a shadow
of a doubt that M. schumanni ants can gall nonmyrme-
cophytic trees. We just hope that, this time, the scientific
community will not have to wait another near-century for
proof. Finally, we urge scientists to pay attention to local
people’s rich and often underappreciated knowledge about
forest ecosystems: sometimes even elements of folklore
that appear quaint or unscientific can lead to the acqui-
sition of scientific knowledge (see also Sheil and Lawrence
2004).

6. Ants Can Gall Trees 739
Acknowledgments
We thank A. Moog and B. Webber for considerable help
in tracking down historical references and I. V. Shamoko
and M. Shamoko for introducing D.W.Y. and G.H.S. to
this phenomenon. Support for this work came from the
Yunnan government (20080A001), the Chinese Academy
of Sciences (0902281081), the A. L. Green Fund, the Wil-
liam F. Milton Fund, the National Geographic Society, and
the Leverhulme Trust. M.E.F. thanks the Society of Fellows
at Harvard University, D.P.E. thanks the Leverhulme Trust,
and G.H.S. thanks E. G. Neves and the Fundação de Am-
paro a Pesquisa do Estado de São Paulo (FAPESP).
Literature Cited
Abe, Y. 1992. The advantage of attending ants and gall aggregation
for the gall wasp Andricus symbioticus (Hymenoptera, Cynipidae).
Oecologia (Berlin) 89:166–167.
Araujo, L. M., A. C. F. Lara, and G. W. Fernandes. 1995. Utilization
of Apion sp. (Coleoptera: Apionidae) galls by an ant community
in southeastern Brazil. Tropical Zoology 8:319–324.
Bailey, I. W. 1924. Notes on Neotropical ant-plants. III. Cordia nodosa
Lam. Botanical Gazette 77:39–49.
Becarri, O. 1886–1887. Piante ospitatrici ossia piante formicarie della
Malesia e della Papuasia. Malesia 2:1–340.
Belt, T. 1874. The naturalist in Nicaragua. Dent, London.
Bequaert, J. 1922. Ants in their diverse relations to the natural world.
Bulletin of the American Museum of Natural History 45:333–583.
Blüthgen, N., and J. Wesenberg. 2001. Ants induce domatia in a rain
forest tree (Vochysia vismiaefolia). Biotropica 33:637–642.
Carver, M., N. Blüthgen, J. F. Grimshaw, and G. A. Bellis. 2003. Aphis
clerodendri Matsumura (Hemiptera: Aphididae), attendant ants
(Hymenoptera: Formicidae) and associates on Clerodendrum (Ver-
benaceae) in Australia. Australian Journal of Entomology 42:109–
113.
Chen, Y., L. D. Hansen, and J. J. Brown. 2002. Nesting sites of the
carpenter ant, Camponotus vicinus (Mayr) (Hymenoptera: For-
micidae) in northern Idaho. Environmental Entomology 31:1037–
1042.
Chodat, R., and L. Carisso. 1920. Une nouvelle théorie de la myr-
mécophilie. Comptes Rendus des Séances de la Société de Physique
et d’Histoire Naturelle de Genève 37:9–22.
Darwin, C. 1877. On the glandular bodies on Acacia sphaerocephala
and Cecropia peltata serving as food for ants. Journal of the Lin-
nean Society, London: Botany 15:398–409.
Davidson, D. W., and D. McKey. 1993. The evolutionary ecology of
symbiotic ant-plant relationships. Journal of Hymenoptera Re-
search 2:13–83.
Edwards, D. P., M. Hassall, W. J. Sutherland, and D. W. Yu. 2006.
Selection for protection in an ant-plant mutualism: host sanctions,
host modularity and the principal-agent game. Proceedings of the
Royal Society B: Biological Sciences 273:595–602.
Fernandes, G. W., M. Fagundes, R. L. Woodman, and P. W. Price.
1999. Ant effects on three-trophic level interactions: plant, galls,
and parasitoids. Ecological Entomology 24:411–415.
Fonseca, C. R. 1999. Amazonian ant-plant interactions and the nest-
ing space limitation hypothesis. Journal of Tropical Ecology 15:
807–825.
Frederickson, M. 2005. Ant species confer different partner benefits
on two Neotropical myrmecophytes. Oecologia (Berlin) 143:387–
395.
———. 2009. Conflict over reproduction in an ant-plant symbiosis:
why Allomerus octoarticulatus ants sterilize Cordia nodosa trees.
American Naturalist 173:675–681.
Frederickson, M., and D. M. Gordon. 2007. The devil to pay: a cost
of mutualism with Myrmelachista schumanni ants in “devil’s gar-
dens” is increased herbivory on Duroia hirsuta trees. Proceedings
of the Royal Society B: Biological Sciences 274:1117–1123.
———. 2009. The intertwined population biology of two Amazonian
myrmecophytes and their symbiotic ants. Ecology 90:1595–1607.
Frederickson, M., M. J. Greene, and D. M. Gordon. 2005. “Devil’s
gardens” bedevilled by ants. Nature 437:495–496.
Gaume, L., and D. McKey. 1999. An ant-plant mutualism and its
host-specific parasite: activity rhythms, young leaf patrolling, and
effects on herbivores of two specialist plant-ants inhabiting the
same myrmecophyte. Oikos 84:130–144.
Gaume, L., M. Zacharias, and R. M. Borges. 2005. Ant-plant conflicts
and a novel case of castration parasitism in a myrmecophyte.
Evolutionary Ecology Research 7:435–452.
Gentry, A. H. 1988. Tree species richness of upper Amazonian forests.
Proceedings of the National Academy of Sciences of the USA 85:
156–159.
Heil, M., and D. McKey. 2003. Protective ant-plant interactions as
model systems in ecological and evolutionary research. Annual
Review of Ecology, Evolution, and Systematics 34:425–453.
Hölldobler, B., and E. O. Wilson. 1990. The ants. Belknap Press,
Harvard University Press, Cambridge, MA.
Inouye, B. D., and A. A. Agrawal. 2004. Ant mutualists alter the
composition and attack rate of the parasitoid community for the
gall wasp Disholcaspis eldoradensis (Cynipidae). Ecological Ento-
mology 29:692–696.
Janzen, D. H. 1966. Coevolution of mutualism between ants and
acacias in Central America. Evolution 20:249–275.
———. 1975. Pseudomyrmex nigropilosa: parasite of a mutualism.
Science 188:936–937.
McIntyre, P. J., and T. G. Whitham. 2003. Plant genotype affects
long-term herbivore population dynamics and extinction: conser-
vation implications. Ecology 84:311–322.
Morawetz, W., M. Henzl, and B. Wallnöfer. 1992. Tree killing by
herbicide producing ants for the establishment of pure Tococa
occidentalis populations in the Peruvian Amazon. Biodiversity
Conservation 1:19–33.
Olesen, J. M., A. B. Lindberg, L. I. Eskildsen, J. C. Svenning, and R.
Lindberg. 2002. Plants in the devil’s garden: intruders in an ant-
plant mutualism. Ecotropica 8:81–86.
Price, P. W. 2005. Adaptive radiation of gall-inducing insects. Basic
and Applied Ecology 6:413–421.
Redfern, M., and P. Shirley. 2002. British plant galls: identification
of galls on plants and fungi. Field Studies 10:207–531.
Renner, S. S., and R. E. Ricklefs. 1998. Herbicidal activity of domatia-
inhabiting ants in patches of Tococa guianensis and Clidemia het-
erophylla. Biotropica 30:324–327.
Risch, S. J., and F. R. Rickson. 1981. Mutualism in which ants must
be present before plants produce food bodies. Nature 291:149–
150.
Ronquist, F., and J. Liljeblad. 2001. Evolution of the gall wasp-host
plant association. Evolution 55:2503–2522.
Sheil, D., and A. Lawrence. 2004. Tropical biologists, local people

7. 740 The American Naturalist
and conservation: new opportunities for collaboration. Trends in
Ecology & Evolution 19:634–638.
Shepard, G. H., Jr. 1998. Psychoactive plants and ethnopsychiatric
medicines of the Matsigenka. Journal of Psychoactive Drugs 30:
321–332.
Shepard, G. H., Jr., D. W. Yu, M. Lizarralde, and M. Italiano. 2001.
Rain forest habitat classification among the Matsigenka of the
Peruvian Amazon. Journal of Ethnobiology 21:1–38.
Shepard, G. H., Jr., K. Rummenhoeller, J. Ohl, and D. W. Yu. 2009.
Trouble in paradise: indigenous populations, anthropological pol-
icies, and biodiversity conservation in Manu National Park, Peru.
Journal of Sustainable Forestry (forthcoming).
Sliva, M. D., and J. D. Shorthouse. 2006. Comparison of the devel-
opment of stem galls induced by Aulacidea hieracii (Hymenoptera:
Cynipidae) on hawkweed and by Diplolepis spinosa (Hymenoptera:
Cynipidae) on rose. Canadian Journal of Botany 84:1052–1074.
Taft, J. B., and D. R. Bissing. 1988. Developmental anatomy of the
horned oak gall induced by Callirhytis cornigera on Quercus pal-
ustris (pin oak). American Journal of Botany 75:26–36.
Terborgh, J. 1990. An overview of research at Cocha Cashu Biological
Station. Pages 48–59 in A. H. Gentry, ed. Four Neotropical rain-
forests. Yale University Press, New Haven, CT.
Uphof, J. C. T. 1942. Ecological relations of plants with ants and
termites. Botanical Review 8:563–598.
Wallace, A. R. 1905. My life. Vol. 2. Chapman & Hall, London.
Webber, B. L., J. Moog, A. S. O. Curtis, and I. E. Woodrow. 2007.
The diversity of ant-plant interactions in the rainforest understorey
tree, Ryparosa (Achariaceae): food bodies, domatia, prostomata
and hemipteran trophobionts. Botanical Journal of the Linnean
Society 154:353–371.
Wheeler, W. M. 1942. Studies of Neotropical ant-plants and their
ants. Bulletin of the Museum of Comparative Zoology, Harvard
University, Cambridge, MA.
Yu, D. W., and N. E. Pierce. 1998. A castration parasite of an ant-
plant mutualism. Proceedings of the Royal Society B: Biological
Sciences 265:375–382.
Natural History Editor: Craig W. Benkman
“Last spring, Mr. J. A. Lintner noticed on the sandy hills west of Albany, N.Y., a number of holes about half an inch in diameter, each surrounded
by a ring of sticks and bits of leaves loosely fastened together by fine threads. … Before opening the holes we sounded them with straws and tried
to provoke the spiders to come out, but they took no notice of it.” From “The Lycosa at Home” by J. H. Emerton (American Naturalist, 1871, 4:
664–665).