litter. For each canopy sample, a paired ground sample
was obtained. Invertebrates from sifted litter were ex-
tracted using a Winkler extractor (Ward 1987). Data
were tabulated in terms of number of individuals per
liter of sifted litter. For each category, ground versus
canopy differences in density were tested with a paired
ftest(P<.05).
The numerically dominant taxa in both canopy and
ground samples were mites, beetles, holometabolous
insect larvae, ants, Collembola, and Crustacea (amphi-
pods and isopods). Members of these taxa are the major
agents of fragmentation of organic matter (Leakey and
Proctor 1987). Relative proportions of these major taxa
were the same in the canopy and ground samples,
showing that the canopy shares a fundamentally simi-
lar invertebrate community (Fig. 9.14). Total inverte-
brate abundance was higher on the ground than in the
canopy for all sample pairs, with a mean density 2.6
times greater on the forest floor than in the canopy.
Canopy organic matter was clearly depauperate in
beetles and larvae. Ants were the only exception to this
pattern. Our results differ from another study at a simi-
lar site in Costa Rica (Atkin and Proctor 1988) in that
we found very few oligochaetes (earthworms). The
difference may be due to sampling techniques; the
hand-sorting methods of the latter study underesti-
mated meso- and microarthropods, and our sifting
techniques underestimated soft-bodied annelids.
What factors might explain the lower invertebrate
densities in the canopy? Differences in microclimate
could affect their composition and abundance. Tree
crowns are subject to greater amounts of insolation
and wind and more severe "dry-downs" than the for-
est floor (Bohlman et al. 1995). Invertebrates are sen-
sitive to microclimate regimes, particularly moisture
and temperature. Difficulties of dispersal to the can-
opy and/or within crowns of trees may exist for par-
ticular invertebrate groups, especially many of the
sedentary groups and those living in nest structures
(Moran and Southwood 1982). Elucidation of which
factors depress invertebrate density in the canopy,
and which canopy organic matter characteristics may
be caused by depressed invertebrate density await
further investigation. We showed that canopy-held
organic matter, an ecosystem component of lower
montane forests, differs from forest floor organic mat-
ter. This implies that organic matter in the canopy
should be treated as a separate component in ecosys-
tem studies and cannot be subsumed under studies
of organic matter on the forest floor.Acknowledgments We thank the MCFP and the
Monteverde community. John Campbell and William
Calvert provided access to study sites. Work was sup-
ported by a Faculty Career Development Grant from
the University of California, Santa Barbara, grants
from the National Geographic Society Committee on
Research and Exploration, the Whitehall Founda-
tion, and the National Science Foundation (BSR 86
14935).337 Ecosystem Ecology and Forest DynamicsFigure 9.14. Invertebrate density (number of individu-
als/liter of sifted litter) by invertebrate group for ground
and canopy samples. Ml = mites, BE = adult beetles, LA
= holometabolous insect larvae, AN = adult ants, CO =
collembola, CR = amphipods and isopods, OT = other
(Hemiptera, Homoptera, Orthoptera, Dictyoptera,
Dermaptera, adult Diptera, Thysanoptera, parasitic
Hymenoptera, Apterygota, Symphyla, millipedes,
centipedes, spiders, other Arachnida, Gastropoda
Annelida). Significance values above categories are results
of paired sample t tests. Error bars represent two
standard errors of the means. (From Nadkarni and
Longino 1990)