TABLE 1.—Mean and standard error of hits and composition for species or groups of species in 
five nonforest range types 
Agsp Agsp Agsp Arar Artr 
shallow deep steep agsp agsp 
Hits on: 
Poa secunda _____________- ee 745 9+5 8+1.5 4+3 5+2 
IMETONYTON 225222525522 - 5252 7+4 844 84 9+3 6+6 
Hestuca, feces eee See ae 242 5+3 4+3 3+3 T+6 
Composition of: 
Poasecunda —....-2--22---2+ 40+15 85414 19+12 838+14 Q5+7 
IASropyron 224. 222.52 -52550 24+12 22+14 86411 29+14 80427 
MeStUCap eet tet Seek Pe 8+9 28+13 22+14 13+13 28+23 
Hits on: 
Decreasers ___________--_---- 9+4 138+5 12+4 14+6 1644 
In¢reasers:.2.2246..-2222cee4-- 9+4 10+5 4+4 5+3 6+3 
Composition of: 
Decreasers ________________- 382+11 48+12 51+13 44+11 58+9 
Incredsersi] 2 22 ee 42+14 387+13 252415 388+13 380+9 
community groups. This table was used as a 
basis for formulating condition guides, based on 
hits and composition of species and species 
groups, for each plant community group. Stand- 
ard errors vary from 25 to 100 percent of the 
mean for hits and from 15 to 100 percent of the 
mean for composition. Lowest standard errors 
were obtained by combining plant species into 
classifications of decreasers and increasers. 
Wide standard errors within groups reflect 
vegetation variability resulting from contin- 
uum gradient sampling. These illustrate the 
problems encountered in classifying commun- 
ity groups within a gradient by attempting to 
apply the modal group concept to field condi- 
tions. Variability exceeds one condition class 
(+10 percent of the mean), suggesting that 
five condition classes cannot be justified statisti- 
cally. Therfore, four condition classes were uti- 
lized. Very poor condition was defined as one 
lacking in decreaser plants and so deteriorated 
that adjustment in livestock management is 
generally not an economically feasible method 
for improving range condition. Three condition 
classes were based upon presence of decreaser 
vegetation: Good represents 66 to 100 percent 
of “possible” hits and composition; fair, 33 to 
66 percent; and poor, 2 to 33 percent. 
Condition could be more accurately evalu- 
ated by division into finer groups. However, 
criteria for field identification of these finer 
groups in poor and very poor condition were 
difficult to describe. Vegetation indicators are 
generally obliterated, and no single site fac- 
tor could be used. Accurate site identification 
would require measuring several environmental 
factors. These factors, when entered into a mul- 
tiple-variate analysis formula, could be used 
with much greater accuracy in suggesting po- 
tential vegetation than as indicators of a plant 
community group or association. 
Multiple-variate analysis and the continuum. 
—Sampling data were analyzed by multiple- 
variate analysis. Plant community types, a 
name assigned to community groups based 
upon similar management characteristics, were 
evaluated individually. Also, community types 
with generally similar vegetation were 
grouped and analyzed for continuum gradients. 
Dependent variables included hits and compo- 
sition by species and by groups of species, for- 
age production, and surface stoniness. Thirty- 
two independent variables were tested. 
Table 2 lists environmental factors signifi- 
cantly associated (5 percent level) with hits 
and composition of Agropyron spicatum, Fes- 
tuca idahoensis, and Poa secuda; total her- 
bage production; and surface stoniness in the 
Agropyron/Poa community type. The analyses 
accounted for 40 to 71 percent of the variabil- 
ity. No single factor is predominantly asso- 
ciated with all items or with any particular 
item, suggesting multiple-factor continuum 
tendencies and independent ecologic ampli- 
tudes. In many cases, factor associated with a 
species’ hits is not always associated with that 
same species’ composition because composition 
is associated with other dependent variables. 
For example, elevation is negatively correlated 
with Agropyron hits but is not associated with 
Agropyron composition. In six of eight cases, 
factors other than soil characterists account 
for most of the variability. 
Table 3 shows factors associated with vegeta- 
tion when four nonforest bunchgrass commun- 
ity types were combined for analysis. In this 
case, the characteristics of soils account for 
more of the variability than do topographic or 
climatic factors, suggesting that soils were ap- 
partently reflected in community-type classifi- 
cation. Different factors are associated with a 
species’ hits and composition when community 
types are grouped. What, then, are the ‘real’ 
factors associated with a species’ distribution? 
213 
