Geoscience Reference
In-Depth Information
1975), upland birds (Eastman and Jenkins 1970), and raptors (Craighead and
Craighead 1956) as well as small (Maser et al. 1985) and large mammals
(Green et al. 1986). To distinguish fecal samples, field guides are available that
rely on size, shape, and color to identify the source (Webb 1943; Murie 1974).
If a less ambiguous approach is needed, especially where sympatric species have
feces or pellets that are similar, fecal pH or bile acid differences can be used to
distinguish samples (Howard 1967, Johnson et al. 1984). A newer approach
uses information on mitochondrial DNA from epithelial cells shed from the
intestines of the animal that defecated or regurgitated the pellet (Höss et al.
1992; Litvaitis and Litvaitis 1996; Paxinos et al. 1997).
Among herbivores, partially digested seeds and fruits can be identified
macroscopically. However, most of the analysis of herbivorous materials relies
on microhistological techniques to identify characteristic cells and structures
of foods consumed. A reference collection of potential food items is crucial.
Microscope slides must be prepared from all potential food plants in the same
manner as sample materials. In addition, a collection of local seeds and fruits
should also be made for reference. Generally, less than half of what appears on
a typical slide will be identifiable plant fragments. The cellular characteristics
used to identify plant fragments are those that survive the mastication and
digestive process and are generally composed of epidermal tissue (Storr 1961).
These include cuticle, stomata, cell walls, aperites, glands, trichomes, silica
cells, druses, crystals, starch grains, and silica-suberose couples as well as gen-
eral cellular configurations, size, and other structural characteristics.
Because the ratio of identifiable to nonidentifiable fragments changes dur-
ing digestion and sample preparation (Havstad and Donart 1978; Holechek
1982) and because certain browse species have a low proportion of epidermal
material in relation to their biomass (Westoby et al. 1976), correction factors
may be developed to improve the approximation of diet composition (Dear-
den et al. 1975). Several researchers have recommended that hand-mixed diets
be used to test the assumption that the actual diet matches the diet estimated
from microhistological analysis (Westoby et al. 1976; Vavra and Holechek
1980; Holechek et al. 1982). Others have suggested that the differences are too
small to justify developing correction factors or that correction factors do not
consistently improve the estimation of diet composition, particularly when the
diets contain a variety of grasses, forbs, and browse (Hansson 1970; Gill et al.
1983).
Investigators of herbivore diets that relied on fecal analysis have been criti-
cized for identifying fewer species than can be found in rumen samples. Gen-
erally, easily digested forbs are underestimated and the less digestible items are
