Table 1. The x-y coordinates for feeding and breeding of 
Steller’s jays in the mature-tree stage of ponderosa pine 
in the eastern ponderosa forest type. 


Feeding Breeding 
Consumer coordinate y# coordinate x 
Primary consumer a4 229 
a4 233 
56 229 
56 233 
pi 229 
57 233 
58 229 
58 233 
60 229 
60 233 
85 229 
85 233 
Secondary consumer 159 229 
159 233 
181 229 
181 233 

aA ppendix I. 
bAppendix II. 
1979:72; Everitt 1974:17, group average method; Sneath 
and Sokol 1973:124, 230-234). 
The cluster analysis program used also computes a “co- 
phenetic” correlation coefficient, R, which is a measure of 
how faithfully the phenogram (i.e., results of the cluster 
analysis) represents the information in the original data. It 
is a way of objectively comparing different clustering re- 
sults. Sneath and Sokol (1973:278-280) provided informa- 
tion about the cophenetic correlation coefficient. 
The y-axis of the species-habitat matrix separates pri- 
mary and secondary consumers so that guilds from pheno- 
grams represent only a single role. About 25% of the 
guilds (12 of 52) produced when the y-axis was not split 
into primary and secondary consumer roles contained 
both primary and secondary consumers (Table 2). This 
was counter to our intended purpose which was to group 
into guilds only those species that used the resources of a 
habitat in approximately the same manner. The occupied 
cells within the species-habitat matrix, when the y-axis 
was split, represented sufficiently unique patterns so that 
wildlife species were aggregated into ecologically mean- 
ingful groups. The quality of these guilds (aggregation of 
species) was not improved (judging from the cophenetic 
correlation, R) by weighting or distorting average 
Euclidean distances between groups of species within the 
matrix by arbitrarily increasing the distance along the 
x-axis of the matrix by a factor of 2. Calculations with raw 
or unstandardized data (simple means and standard devia- 
tions determined directly from data like that in Table 1) 
produced guilds whose members shared similar strata 
blocks with better fidelity than did guilds developed from 
Secondary consumer 
aa 
Tree ca i | \ 
GS Ow ee a ge Oo Oa en ees. 
180 | i \ 
| \ 
| \ 
Se eo en, ey) ee ee. 



| 
| 
1704 Shrub + 
A a eae. a 
Ty! | 
| | 
Terrestrial surface \ ! 
See oe Ss oe ee J j————-—- 
160 PD 
\ / | 
‘ 
| 
150 | | 
90 | | 
Primary consumer 
Actual data points 
e~< | listed in Table 1 
Feeding strata 
Tree canopy 
ptt ee frente 
Tree bole | | 
| 1 SD of Y 
a | 
| ‘ Outline of ellipse 
Shrub 4 | I tata 
mee ia - eee ee 
e+ if 1 SB at ix 
oe 
60 | 
cae Pe he iy 
Tree canopy | 
226 237 
Breeding strata 

Fig. 3. The x and y coordinate values for Steller’s jay (listed in 
Table 1) and the x, y, SD,, SD, values for these data are plotted 
on the species-habitat matrix. The elliptical plots indicating 
the relative areas occupied by the data within the matrix are 
also plotted in this figure. 
data that were standardized by a mathematical transfor- 
mation that produced unit variances for the x and y co- 
ordinates for each species. The phenograms and guilds 
represented in the Results section were therefore de- 
veloped with unstandardized and unweighted x and y 
values in the species-habitat matrix. 
The x and y coordinate values for wildlife species can be 
compiled for both the climax vegetation in a potential 
vegetation type and for the different actual vegetative 
cover types and their seral stages within a potential vege- 
tative type. The data set that is produced when the x and y 
coordinate values are compiled for the wildlife species 
occupying the climax vegetation can be analyzed to de- 
scribe a potential wildlife community that can then be 
used as a standard of comparison in habitat evaluation 
studies. The data sets that are produced when the x and y 
coordinate values are compiled for actual vegetation types 
