152
(i.e., digestibility), which, in tum, depends on attributes of the plants (i.e., lignin content) and the
animals (i.e., digestive system).
Measuring each of these compartments for a
given situation is hardly feasible, so researchers
must either take most of them from the literature or
make strong simplifications. For example, Milchunas and Lauenroth (1993) converted stocking rates
into animal consumption based on a figure proposed by the Society for Range Management indicating that a 454-kg cow consumes 12 kg day-l
of plant dry-matter. That approach implies assumptions for the whole sequence described above, from
animal performance through digestibility. This may
be satisfactory for studies aimed at describing broad
regional patterns in which consumption may vary
by orders of magnitude, but is too coarse for studies
aimed at evaluating environmental modifications
within a system. Essentially identical approaches
are used to estimate consumption in systems dominated by arthropod herbivores. For example,
Schowalter et al. (1981) estimated sap-feeder consumption in forest ecosystems from determinations
of abundance and the assumption, based on laboratory studies, that individual consumption was
2.5 mg dry sap mg - 1 dry insect day - 1. Similarly,
to estimate below ground herbivory by nematodes,
Ingham and Detling (1986) measured density and
estimated individual consumption from relationships between metabolic rate and temperature and
a fixed assimilation efficiency taken from the
literature.
What sort of potential error is associated with
this approach? This may be better visualized in the
series of equations that lead to final consumption
(Scott 1979):
C=DXF
(10.1)
where C is consumption, D is animal density, and
F is individual food demand.
F = (G + R)/A
(10.2)
where G is growth, R is respiration, and A is assimilation efficiency. Thus, estimating C will require
good estimates not only of animal density, but also
of performance and assimilation efficiency. This, in
tum, will depend on animal condition and physiology and on the digestibility and other nutritional
properties of the forage.
Martin Oesterheld and Samuel J. McNaughton
The magnitude of these potential errors depends
on the variation that researchers may leave unaccounted for at each step of the process. Regarding
assimilation efficiency (A), plant digestibility may
strongly vary because of seasonality and species
composition. In vitro dry-matter digestibility of
whole plants of Lotium rigidum decreased from
58% at anthesis to 36% 69 days later (Ballard et al.
1990). Even greater variations were observed
among species fed to black-tailed deer, which
ranged between 37 and 72% (Hanley et al. 1992).
Plant digestibility may also vary among animal species grazing or browsing a common forage. Different digestive systems, gut capacities, body sizes,
and glandular systems affect the digestibility of a
forage (Demment and Van Soest 1985). For example, tannins strongly decrease cell wall digestion
in domestic sheep, but do not affect mule deer,
probably due to different salivary secretions (Robbins et al. 1987). Regarding animal performance (G
and R), metabolic needs for maintenance, which depend on body weight raised to 0.75, increase as
body size decreases (in relative terms). Thus, metabolic rates per unit mass may vary by 40% among
differentially sized individuals of the same or different species (e.g., cattle from 300 to 700 kg of
live weight) (Bondi 1989). Maintenance is not the
only metabolic activity; similarly sized calves may
need twice as much energy if they are growing at
a rate of 1 kg day - 1 than if they are just meeting
maintenance requirements (Bondi 1989). Thus, animal density cannot be translated directly into consumption unless a series of variables related to the
plants and the animals are determined or assumed.
Plant-Based Methods
The second approach to estimate herbivore consumption is based on estimates of missing plant
biomass that was removed by herbivores. For most
systems dominated by herbaceous vegetation and
vertebrate herbivores, the difference between caged
and uncaged plots is used (McNaughton et al.
1996). The cages may be set in place for an entire
season and a single biomass determination made at
the end or at peak biomass, or they may be moved
with some frequency and biomass determinations
made at each move. Although this method seems
straightforward, it also implies many assumptions
and involves a reduction of spatial scale compared
(i.e., digestibility), which, in tum, depends on attributes of the plants (i.e., lignin content) and the
animals (i.e., digestive system).
Measuring each of these compartments for a
given situation is hardly feasible, so researchers
must either take most of them from the literature or
make strong simplifications. For example, Milchunas and Lauenroth (1993) converted stocking rates
into animal consumption based on a figure proposed by the Society for Range Management indicating that a 454-kg cow consumes 12 kg day-l
of plant dry-matter. That approach implies assumptions for the whole sequence described above, from
animal performance through digestibility. This may
be satisfactory for studies aimed at describing broad
regional patterns in which consumption may vary
by orders of magnitude, but is too coarse for studies
aimed at evaluating environmental modifications
within a system. Essentially identical approaches
are used to estimate consumption in systems dominated by arthropod herbivores. For example,
Schowalter et al. (1981) estimated sap-feeder consumption in forest ecosystems from determinations
of abundance and the assumption, based on laboratory studies, that individual consumption was
2.5 mg dry sap mg - 1 dry insect day - 1. Similarly,
to estimate below ground herbivory by nematodes,
Ingham and Detling (1986) measured density and
estimated individual consumption from relationships between metabolic rate and temperature and
a fixed assimilation efficiency taken from the
literature.
What sort of potential error is associated with
this approach? This may be better visualized in the
series of equations that lead to final consumption
(Scott 1979):
C=DXF
(10.1)
where C is consumption, D is animal density, and
F is individual food demand.
F = (G + R)/A
(10.2)
where G is growth, R is respiration, and A is assimilation efficiency. Thus, estimating C will require
good estimates not only of animal density, but also
of performance and assimilation efficiency. This, in
tum, will depend on animal condition and physiology and on the digestibility and other nutritional
properties of the forage.
Martin Oesterheld and Samuel J. McNaughton
The magnitude of these potential errors depends
on the variation that researchers may leave unaccounted for at each step of the process. Regarding
assimilation efficiency (A), plant digestibility may
strongly vary because of seasonality and species
composition. In vitro dry-matter digestibility of
whole plants of Lotium rigidum decreased from
58% at anthesis to 36% 69 days later (Ballard et al.
1990). Even greater variations were observed
among species fed to black-tailed deer, which
ranged between 37 and 72% (Hanley et al. 1992).
Plant digestibility may also vary among animal species grazing or browsing a common forage. Different digestive systems, gut capacities, body sizes,
and glandular systems affect the digestibility of a
forage (Demment and Van Soest 1985). For example, tannins strongly decrease cell wall digestion
in domestic sheep, but do not affect mule deer,
probably due to different salivary secretions (Robbins et al. 1987). Regarding animal performance (G
and R), metabolic needs for maintenance, which depend on body weight raised to 0.75, increase as
body size decreases (in relative terms). Thus, metabolic rates per unit mass may vary by 40% among
differentially sized individuals of the same or different species (e.g., cattle from 300 to 700 kg of
live weight) (Bondi 1989). Maintenance is not the
only metabolic activity; similarly sized calves may
need twice as much energy if they are growing at
a rate of 1 kg day - 1 than if they are just meeting
maintenance requirements (Bondi 1989). Thus, animal density cannot be translated directly into consumption unless a series of variables related to the
plants and the animals are determined or assumed.
Plant-Based Methods
The second approach to estimate herbivore consumption is based on estimates of missing plant
biomass that was removed by herbivores. For most
systems dominated by herbaceous vegetation and
vertebrate herbivores, the difference between caged
and uncaged plots is used (McNaughton et al.
1996). The cages may be set in place for an entire
season and a single biomass determination made at
the end or at peak biomass, or they may be moved
with some frequency and biomass determinations
made at each move. Although this method seems
straightforward, it also implies many assumptions
and involves a reduction of spatial scale compared