(26-41 g m?) (Table 12), and the predicted 
apparent dry matter digestibility of the live 
biomass was high at 59% (Table 19). This com- 
munity also contains considerable amounts of D. 
fisheri, a grass which, until flowering, is poten- 
tially highly digestible (ca. 79%, Table 18), plus 
a wide range of willows and herbs which provide 
considerable variation in the diet. 
In comparison with the Dupontia brook 
meadow, the most used community, the 
Eriophorum polygon marsh, was of only moder- 
ate nutritive value. The live biomass was moder- 
ate (34-41 g m?), which might limit eating rate 
and cause some compensatory increase in graz- 
ing time (Figs. 10a, 10b, 11), and the average 
digestibility of live material of the community 
was only 50% (Table 19). Selection within the 
community could lead to a higher digestibility 
of forage, but this must be balanced against 
increased searching time. 
The apparent lack of preference for the 
Carex aquatilis marsh may be attributable to 
several factors: It is a prime mosquito habitat, is 
unavailable due to high water levels in early 
summer, is of low live biomass (24-26 g a), 
and is relatively high in dead material (Table 
12). In compensation, the predicted apparent 
dry matter digestibility of the live plant material 
was over 60% (Table 19). Thus, when mosquitos 
are not a problem, this community is potentially 
useful for grazing by caribou in late July and 
August. 
Care should be taken in translating informa- 
tion based on the original vegetation analysis to 
the more detailed vegetation analysis of Webber 
and Walker. In Fig. 7 a comparison is shown 
between our present scheme and the more 
detailed scheme outline by Webber and Walker. 
In general, good agreement was noted between 
the schemes. However, because the cryptogams 
were not used in defining the vegetation and 
because Carex aquatilis, Eriophorum angusti- 
folium, and Dupontia fisheri are found in most 
vegetation types (Appendix Table 10), some 
inconsistencies may arise in defining the Carex 
aquatilis and Eriophorum angustifolium 
marshes. The key used by Webber and Walker is 
shown in Fig. 7, and it is clear that, initially, the 
absence or presence of moss and, finally, the 
type of moss dictates the allocation of vegeta- 
183 
tion to types 4, 5, and 6. These vegetation types 
were placed in the Carex marsh type in Skog- 
land’s scheme. 
A summary of the above characteristics of 
each community is shown in Fig. 14. Future 
gaming runs with model GRAZE may help to 
show the relative importance of each factor in 
affecting food intake by caribou. The present 
results are limited to the observation that above 
maintenance, an 18% increase in apparent dry 
matter digestibility can lead to a 200% increase 
in energy retention. The relative importance of 
small changes in other factors (e.g., community 
availability, live biomass, or the ratio of live/ 
dead biomass) are not known. It can also be 
shown that caribou have the potential to denude 
the range of live biomass at a rate of 11 min m°?. 
Thus, any inadvertent restriction of animals 
onto a small area could result in rapid removal 
of live plant material. Effects of trampling may 
also be important in destroying the habitat. 
From the observed available live biomass 
(Table 12) and relationships between eating rate 
and biomass (Figs. 9 and 10), it is clear that 
available biomass limits forage intake on most 
plant communities. However, when not harassed 
by insects, caribou can apparently graze for 
extended periods of time. They spend up to 53% 
of the day eating, which is equivalent to a 
grazing period (eating plus walking) of 60-65% 
of the day (Tables 2a, 2b). Predictions based on 
grazing behavior (Tables 2 and 14) indicate that 
by virtue of their higher rates of grazing inten- 
sity, lactating females have a higher rate of 
ingestion of food than non-lactating females and 
males. Although this observation agrees with 
field studies on domestic sheep (Arnold and 
Dudzinski 1967), future work is required to 
verify this critical point in reindeer. Esophageal 
collections using lactating females should be 
compared with those for adult males and the 
current estimates of eating rate of non-lactating 
females. This type of experiment should be 
extended to weaned calves and yearlings, as we 
currently have no estimates on the relative eat- 
ing activities of these cohorts. Similarly, more 
information is required on the interaction of 
grazing and eating times with available biomass 
and canopy structure (Fig. 10). More informa- 
tion on these factors will allow refinement on 
the current estimates of food intake. 
