
are less specialized than those advocated by Root (1967) 
which are associations of species with very similar forag- 
ing strategies and habitat requirements. Our guilds, on the 
other hand, are more specific and meaningful than are 
artificial classifications or groupings of species based on 
size, food habits, and foraging strategies (Severinghaus 
1981), and seem to provide a greater analytical and pre- 
dictive capability than does the life form concept of 
Thomas (1979). 
Our guilding technique allows habitat evaluations and 
assessments to be meaningfully based on a community 
principle rather than on selected indicator species. Advan- 
tages include those of thoroughness and even-handedness 
in the treatment of all segments of the wildlife community 
and an enhanced opportunity for interpreting wildlife 
data with data from other land uses in planning efforts. 
The potential applications cited in the Discussion section 
are presumed and not proven. Ongoing research will 
evaluate the usefulness of the present concepts in natural 
resource management. Although the analyses, examples, 
and discussions presented here are directed toward terres- 
trial communities, the statistical assessments should be 
equally relevant to aquatic systems. 
Six assumptions were made to develop the information 
presented in this paper: 
1, That complex biological requirements of animals can 
be structured into a two-dimensional matrix, and that 
energy sources and breeding requirements are so basic that 
they can meaningfully represent the axes of that matrix. 
2. That ordering vegetative cover types into a series of 
vertical strata is a natural organizational framework 
within the two-dimensional matrix. 
3. That major food sources “of similar importance” can 
be identified and that these food sources can be organized 
in a rational manner within the strata of the matrix. 
4, That major breeding niches “of similar importance” 
can be identified and that these breeding niches can be 
organized in a rational manner within the strata of the 
matrix, 
5. That the biological information about individual 
species is sufficiently detailed in the literature so that feed- 
ing and breeding requirements can be compiled for each 
wildlife species occurring within a cover type. 
6. That the abstract numerical designations given to the 
food source and breeding niche listings can be subjected to 
numerical and statistical analyses. 
Methods 
The development of a technique for translating a 
variety of descriptive data about wildlife species into a 
numerical framework was basic to our technology. The 
biology of any wildlife species is exceedingly complex and 
many of the biological requirements for any species are 
poorly known. Simplification of a complex, varied, and 
poorly known biological universe was obviously required. 

We selected energy sources and physical breeding require- 
ments as the basis for organizing wildlife information in 
our guilding analyses. These two criteria are obviously 
fundamental to the existence of species, are driving forces 
affecting the behavior of species, and are sufficiently gross 
parameters so that adequate information might be avail- 
able for most wildlife species. 
We then developed a simple two-dimensional species- 
habitat matrix with energy sources along the y-axis and 
physical features of the habitat required for breeding 
~along the x-axis. The y-axis was subdivided so that the 
lower half of the y-axis could provide loci for data about 
primary consumers, and the upper half of the y-axis could 
provide loci for data about secondary consumers. Both 
axes of the matrix are partitioned by physical strata, be- 
cause numerous authors have emphasized the importance 
of strata in describing the form and function of ecological 
communities. The food sources for animals comprising the 
y-axis of the matrix are organized as indicated in Fig. 1. 
The 191 food sources representing the rows of the matrix 
are identified in Appendix I. The physical features re- 
quired for breeding (the x-axis of the matrix) are organized 
as indicated in Fig. 2. The 238 columns representing 
breeding requirements are identified in Appendix II. The 
numbers assigned to rows and columns of the matrix are 
those listed in Appendices I and II. 
Feeding and breeding criteria for a species are located 
in the species-habitat matrix by listing the numeric x and y 
coordinate values which identify the cell or cells within 
the general matrix that best describe the habitat require- 
ments of that species. For example, the Steller’s jay (see 
Appendix III for scientific names of vertebrates) occurs in 
the mature-tree stage of ponderosa pine in the eastern pon- 
derosa pine type of woodland (Kiichler 1964, type 16). 
This jay consumes a variety of insects and herbaceous and 
pine seeds on the ground and a variety of insects and pine 
seeds in the pine canopy. Its nest is a bowl of coarse twigs, 
usually placed in the crotch of a horizontal tree branch 
within the pine canopy. The x and y coordinates (Ap- 
pendices I and I) that best describe the requirements of 
the Steller’s jay inhabiting mature ponderosa pine wood- 
lands are listed in Table 1. The data in Table 1 fill the cells 
in the matrix indicated in Fig. 3. 
We have used the concept of potential natural vegeta- 
tion as a convenient bound for our data sets. The potential 
vegetation types of Kiichler (1964) are areas that presum- 
ably produce vegetation of a common taxa and structure if 
left undisturbed for a sufficiently long period. Presumably 
the same wildlife community also will eventually develop 
throughout this type of climax vegetation. The structure of 
this wildlife community (guilds) could thus potentially be 
similar throughout this climax vegetation type. These 
potential guilds can be considered a standard of compari- 
son for the wildlife communities found in current cover 
types within the potential natural vegetation type. The 
potential vegetation type is also a useful bound because its 
area, and the numbers of cover types (current vegetative 
