Air 

Tree canopy 
Tree bole 

Shrub 
Terr. surface 

Terr. subsurf. 

Water surface 
ll 
aE! A AR oe oes ha aE a 
—SS OEE SS el 

STE 
= 
aa 
was =e 
Ags 








Secondary consumers 
Water column 
Bottom water 
column 







Feeding strata 
Tree canopy 
Tree bole 
Shrub 
Terr. surface 



Terr. subsurf. 

Primary consumers 
Water surface 








Water column 

Water surface 
Terr. subsurf. 
Terr. surface 






Shrub 
Tree bole 
Tree canopy 
Breeds 
elsewhere 
Breeding strata 
Fig. 8. The positions occupied by individual vertebrate species within the super cells of the species-habitat matrix for upland coniferous 
woodland habitats. The two ellipses identified by arrows represent the area within the species-habitat matrix occupied by Steller’s 
jay (Table 1). 
tance of each other (empirically determined by us not to 
exceed 3.6 units of Euclidean distance on the x-axis for the 
phenograms in Figs. 9-11) generally share the same com- 
binations of super cells in fulfilling their breeding and 
feeding activities. 
The splitting of the y-axis in the species-habitat matrix 
(Fig. 5) causes the cluster analysis routine to produce 
guilds of species that are either primary or secondary con- 
sumers (Figs. 9-11). The use of 3.6 units of Euclidean dis- 
tance as a determinant for aggregating results from the 
cluster analysis into wildlife guilds produces groups of 
species that tend to feed and breed within the same combi- 
nation of super cells. The relationship between guilds and 
super cells that results from this mathematical model is 
strong: the regression equation (Table 3) is Y (number of 
guilds) = 1.13 times (the number of occupied super cells) — 
6.24 (r = 0.99). 
The actual development of wildlife guilds is illustrated 
in the following example. The ellipse plots developed for 
the primary consumers breeding in the tree canopy of a 
mature ponderosa pine community are represented in 
Fig. 12. These plots represent a subset of the ellipse plots 

