Tropical Forest Community Ecology

Resource Limitation of Insular Animals 325

assumes more normal levels (Davis 2001, Adler
unpublished data).
Clearly, frugivorous and granivorous mammals
regularly experience bottom-up limitation by sea-
sonal scarcity of resources, even in intact forests.
However, there is compelling evidence that fru-
givorous and granivorous mammals that live in
mainland tropical forests also are subjected to
strong top-down limitation (e.g., Terborghet al.
2001). In the absence of top predators, a trophic
cascade apparently results, leading to greatly
increased population densities of animal con-
sumers (Terborghet al. 2001), including rodents
(Lambertet al. 2003). Therefore, mammals iso-
lated on islands with depauperate predator com-
munities and where densities are much higher
thanonmainlandareasareexpectedtobeaffected
to an even greater degree by seasonal shortage
of resources and stronger bottom-up limitation.
To what extent are populations of herbivorous
mammals, when largely released from top-down
limitation, limited by bottom-up processes?

SPINY RATS AS A CASE STUDY

Introduction and methods

Since 1989, my students and I have been studying

the Central American spiny rat (Proechimys

semispinosus) on islands in the Panama Canal

as a model system to examine population pro-

cesses, including resource limitation. There are

over 200 islands available for study, and the spiny

rat is distributed widely on the islands (Adler and

Seamon 1991). Even tiny islands less than 0.1 ha

in area frequently contain rats if the islands are

close to larger landmasses, even though such

islands are too small to support persistent popu-

lations (Adler and Seamon 1991). Most islands of

more than 1 ha support persistent populations,

and the spiny rat is the only terrestrial mammal

that maintains such populations on those small

islands (Adler 1996). My students and I have

studied spiny rats on 76 islands ranging in size

from less than 0.1 to 1500 ha, and we selected

12 islands ranging in size from 1.7 to 3.9 ha

for long-term study to examine population-level

patterns and processes. We have conducted both

short-term studies (i.e., within a single season or

year) on geographical factors that influence spiny

rat distributions on islands, seed predation and

dispersal by spiny rats, mycorrhizal fungal spore

dispersal by spiny rats, and spacing patterns of

spiny rats, and long-term studies (i.e., spanning

several years) on demography and factors that

limit spiny rat populations, particularly resource

abundance. These long-term studies will be the

focus of this review.

Long-term study islands were selected because

they (1) were known or predicted to maintain

persistent populations, (2) were of a size that per-

mitted regular and thorough censuses, (3) were of

roughly similar size and isolation, and (4) differed

in tree species composition and forest structure

and therefore were likely to harbor populations

that differed in density and demography (Adler

2000). Because the long-term study islands main-

tain persistent and generally large populations of

spiny rats (often more than 100 or 200 individ-

uals), they (1) are amenable to both descriptive

studies and manipulative experiments to address

the question of population limitation by resource

abundance,(2)arelikelytoyieldlargesamplesizes

of individuals, and (3) are therefore appropriate

for robust statistical analysis. Because the islands

are isolated, they represent essentially closed sys-

tems from the perspective of spiny rat demog-

raphy and therefore ideal experimental systems.

Although there is circumstantial evidence that

spinyratsoccasionallyswimamongislands(Adler

and Seamon 1991), and we indeed recently docu-

mentedseveralinter-islandmovements(Lopezand

Adler unpublished data), we never recorded such

movements on the long-term study islands dur-

ing the study period. We therefore assume that

colonization events are rare along an ecological

time scale and have a negligible impact on the

demography of spiny rat populations.

We conducted monthly censuses of spiny rats

on each of the islands for 9 years (eight islands,

January 1991 through March 2000) or 7 years

(four islands, February 1993 through March

2000). For this purpose, we established perma-

nent sampling grids that covered the whole of

each island. Sampling points on the grids were

20 m apart, and each such point was occu-

pied by a live-trap. Live-trapping was conducted