Aquatic system - tidal 

(1.0) 
Aquatic system - nontidal 
High dissolved solids content 
(1. 
Moderate dissolved solids content 
Gas 
Bottom ~11) 





Column 




Surface 

1) (marine system) 


Bottom 

Column 


2) (estuarine system) Surface 
Temporary water 
(2.0) (2.1) 
Bottom (2.211) 
Cold water Column (2.212) 
(2.21) Surface (2,213) 
Bottom (2.221) 
Permanent water Cool water Column (2.222) 
(22) (2.22) Surface (2.223) 
(riverine, palustrine, 
and lacustrine systems) 
Bottom (2.231) 
Warm water Column (2.232) 
(2.23) Surface (2.233) 
Terrestrial system (34) Terrestrial subsurface 
(3.0) (3.2) Terrestrial surface 
(3.3) Shrub strata 
(3.4) Tree bole 
(3.5) Tree canopy 
Breeds elsewhere 
Seg ee ee ee eee ee a 
(4.0) 
Fig. 2, Organization of breeding habitats by strata along the x-axis of the species-habitat matrix. 
data in the species-habitat matrices are displayed in the 
Results section in a series of plots that graphically depict 
the increasing complexity of wildlife communities as verti- 
cal terrestrial strata are added. These graphics appear as 
ellipses developed from the summary X, 9, SD,, and SD, 
data for all species within a habitat. The ellipses are cen- 
tered on x, ¥ with semi-axis lengths of one standard devia- 
tion (SD, and SD,). Two ellipse plots for the data in 
Table 1 are listed in Fig. 3 to describe the position in the 
species-habitat matrix occupied by Steller’s jay. The data 
describe the jay as both a primary and secondary con- 
sumer. 
The ellipse plots were created by using CALCOMP graphics 
software. Only a small driver program (in FORTRAN) Was 
needed to read the summary data files and call the plotting 
routines, CALcomp plotting software is available for many 
computer systems. 
The ellipse plot is only one of the possible ways to asso- 
ciate the structure of wildlife communities with that of 
plant communities. Graphics that indicate the number of 
species occupying individual cells in the matrix (bar 
graphs or numerical values) or present this information as 
a response surface may also be useful in visually transmit- 
ting an impression of the structure of a wildlife commu- 
nity. 
The summary data (x, 7, SD,, and SD, for primary and 
secondary consumer roles) were further analyzed with 
cluster analysis routines to provide a statistical grouping of 
wildlife species that use similar resources within ecological 
communities (guilds). 
Cluster analysis provides an efficient grouping of species 
within habitat types and portrays this information as a 
phenogram (i.e., dendrogram). The formation of pheno- 
grams from x,y data is described in the Results section. 
Everitt (1974) and Sneath and Sokol (1973) provided an 
introduction to cluster analysis and computer programs 
for cluster analysis which are widely available. The soft- 
ware package Nr/sys used for these cluster analyses is a sys- 
tem of multivariate statistical programs that was de- 
veloped by Rohlf et al. (1979), although almost any set of 
routines for cluster analysis could be used. We used the un- 
weighted pair-group method with arithmetic averages 
(UPGMA) in conjunction with the euclidean distance be- 
tween data vectors (x, y, SD,, and SD,) as the dissimilarity 
measure, as recommended by Rohlf et al. (1979). This is 
perhaps the most common cluster method (see Rohlf et al. 
