Tropical Forest Community Ecology

Resource Limitation of Insular Animals 329

Climatic

seasonality

Fruiting

phenology

Resource (

abundance (

Reproductive

effort

) Density

)

Figure 19.2 Conceptual diagram showing the putative influence of climatic seasonality on spiny rat population
density (from Adler and Beatty 1997).

experimental provisioning studies and renders the
utility of such studies dubious because, although
densities increase in response to provisioning, the
increase is often due largely to nearby individuals
temporarily visiting the provisioned area rather
than to any organic response such as increased
survival or reproductive success. In the case of
relatively closed systems such as islands, any
substantive increase in density would have to
be promulgated by increased survival or repro-
ductive output because the rate of immigration
over water would be too low to explain such an
increase. We also could census each spiny rat
population regularly to obtain reliable estimates
of density, reproductive output, and survival. If
the first hypothesis were supported by the exper-
iment, it would seem that the second hypothesis
was unnecessary to test. However, we were inter-
ested not only in the effect on density but also in
the processes (e.g., increased survival, increased
reproductive output, or both) by which density
might change.
To test the first hypothesis, we selected four
islands as controls, whereby we simply censused
spiny rats and fruiting trees and lianas every
month, and four islands as experimentals. Criteria
for island designations as controls and experimen-
tals are given in Adler (1998). The experimental
populations were provisioned with fresh native
fruit from four species of trees known to be eaten
by spiny rats (Adler 1995; the palmsAstrocaryum
standleyanum,Attalea butyracea, andBactris major,
and the figFicus insipida) from May through
October1992,andspinyratpopulationsandfruit-
ing trees and lianas were censused every month.
Each experimental population was provisioned
with 315.2 kg ha−^1 over the 6-month experi-
mental period. We provisioned the populations
every week, except during the week in which
rat censuses were conducted. For this purpose,
we placed fruit into semipermeable exclosures to
exclude larger frugivorous rodents such as agoutis

and pacas, if such rodents were present (both

species were occasionally seen on two experi-

mental islands for several months but did not

establish populations). Exclosures were situated

permanently and spaced evenly across each island

at a density of 10 per hectare. Each month, spiny

rats had access to 52.5 kg ha−^1 of extra food,

or roughly 1 to 3 kg per spiny rat (depending on

the island and time of year), which is approxi-

mately 3–9 times the mean body weight of a spiny

rat. Thus, all spiny rats had access to reliable and

predictable additional resources throughout the

provisioning period.

Densitiesof spinyratsandof knownbirthswere

analyzed by repeated measures analysis of covari-

ance, with the density of fruiting trees and lianas

as the covariate. Thus, we controlled for the avail-

ability of natural fruit, which certainly had an

impactonpopulationdensitiesevenintheabsence

of the provisioned fruit (Adler and Beatty 1997).

Both density (Figure 19.3) and density of known

births (Figure 19.4) increased on the experimen-

talislandsrelativetothecontrolswhilecontrolling

for the availability of natural fruits. The experi-

mental increase in density could have been due

to improved survival or increased reproductive

output, and the increase in the density of births

could have been due to either increased per capita

reproductive output or simply an increase in the

number of adults and therefore in the number

of young that they produced. We therefore ana-

lyzed monthly survival rates and the number of

known births per adult female using linear anal-

ysis of count data (Lindsey 1995). Survival rates

did not differ between control and experimental

populations, but per capita reproductive output

was greater within the experimental populations.

Thus, the increase in density was due to increased

per capita reproductive output by females rather

than to improved survival, and spiny rat popula-

tions were food-limited even during the season of

greatest resource abundance.