178 
In an analagous manner to that used to 
estimate digestibility of esophageal and rumen 
samples, average digestibilities were estimated 
for each of the vegetation types at Prudhoe Bay 
(Appendix Tables 5-10). These estimates, shown 
in Table 19, are the predicted digestibilities if 
caribou were to graze these communities, con- 
suming the various vascular plant species in 
proportion to their availability. The estimates 
refer only to the time period 6-25 July. If 
caribou were to eat mainly the live plant mate- 
rial, the average digestibilities would be in the 
range of 44-62%, with the higher values noted 
for the wetter communities (the Dupontia 
fisheri brook bank/meadow and the Carex 
aquatilis marsh). The wetter communities also 
contain a substantially higher proportion of 
dead plant material. I!f such material is assumed 
to be 25% digestible, then the average digest- 
ibility for the community as a whole is reduced 
substantially (e.g., the Eriophorum angustifoli- 
um marsh, from 50 to 38%, and the Carex 
aquatilis marsh, from 62 to 51%) (Table 19). 
Also shown in Table 19 are the average /n 
vitro digestibilities of esophageal egesta collected 
from reindeer grazed on these communities, plus 
estimates of the digestibilities of the esophageal 
samples based on their botanical composition. In 
the drier habitats, it was apparent that digest- 
ibilities of selected material approximated that 
which was available as live material. However, in 
two of the wetter habitats, Dupontia fisheri 
brook meadow and Carex aquatilis marsh, the 
selected material was of a digestibility similar to 
that of the habitat as a whole (i.e., live + dead 
material). The suggestion that reindeer were 
consuming a significant proportion of dead plant 
material is not substantiated by the botanical 
analysis of the esophageal fistula samples, which 
indicated a low intake of dead plant material 
and litter (Fig. 7). We conclude that the /n vitro 
digestibility estimates for these communities 
(D. fisheri brook meadow, C. aquatilis marsh), 
may be underestimated. 
No evidence was found for selection of 
highly digestible material from within the com- 
munity. However, the pattern of selection of 
vegetation types described in Table 9 suggests 
some selection for digestibility, based on prefer- 
ence for the D. fisheri brook meadow vegeta- 
tion type. The lack of preference for the C. 
aquatilis marsh (Table 9), despite its predicted 
high quality (Table 19), suggests that its prefer- 
ability may have been lowered by other factors, 
such as water ponding in early July, being a 
prime mosquito habitat, and containing a high 
proportion of dead plant material. 
Chemical composition of plant material 
To date, a limited number of analyses have 
been made on plant samples collected at Prud- 
hoe Bay (Table 20), plus samples of lichen and 
moss collected at Nome and Cantwell, Alaska. 
These same samples were used in the /n vitro 
digestion studies (Table 18). Analyses have been 
confined to the determination of acid detergent 
fiber (ADF) and neutral detergent fiber (NDF), 
further subdivided into hemicellulose, cellulose, 
and lignin. Future determinations will be made 
of energy, total N, P, Ca, and K. 
The amount of cell contents in shrubs and 
herbs was high relative to hemicellulose. In 
grasses and sedges these components were simi- 
lar, and in some lichens the hemicellulose com- 
ponent was high relative to cell contents. The 
cellulose and lignin components of all vascular 
plants were variable. Lichens were virtually free 
of cellulose, although they contained a small 
amount [3.4 g (100 g)'] of acid resistant 
material which was allotted to the lignin compo- 
nent pending identification (Table 20). Table 20 
indicates that ADF is quite variable in the 
vascular plants, but tended to be higher in the 
less digestible species. Lichens are generally 
thought to be high in crude fiber; the present 
analyses indicate that the ADF is low, and that 
the bulk of the cell wall constituents of lichen is 
a material which is extracted by neutral deter- 
gent in a manner similar to hemicellulose. 
The prepared diets given the reindeer were 
either based on lichen or were of a medium to 
high crude fiber commercial pellet. The former 
food contained 78% lichen, and the high content 
of hemicellulose reflects this component in the 
ration. 
Relationships between chemical composition 
and jn vitro digestibility of vascular plants. 
As stated above, it was apparent that /n vitro 
digestibility was high for plants of low crude 
fiber (ligno-cellulose) content. A significant rela- 
tionship between /n vitro digestibility (D, %) and 
