develop prediction equations between overstory density 
and the density of cover in the midstory and understory. 
Aerial photographs are then differentiated into polygons 
on the basis of the structure and density of the overstory so 
that, using the prediction equations, a representation of 
the vertical profile of each polygon is formed. A habitat 
characterization value is then developed for each polygon 
using the same structural parameters (number of guild 
blocks, vegetative density, equitability of cover between 
strata, etc.) used in predicting species-richness from on- 
site assessments. The habitat characterization value of a 
polygon multiplied by the area of that polygon and 
summed for all the polygons within the area being charac- 
terized represents the numerator of a ratio whose denomi- 
nator is the maximum characterization value that can be 
expected within that potential natural vegetation type. 
The greater the ratio, the greater the presumed species- 
richness in the area and the greater the presumed area X 
guilds value for the characterized area. 
Habitat evaluation using guilds is a measure of total 
guilds that have members which apparently find a piece of 
land to be adequate habitat. This type of analysis allows 
the animal to determine whether a habitat adequately 
provides for its life requirements. The analysis does not 
determine how adequate the habitat is for a particular 
species, nor does it determine whether nonhabitat-related 
factors are responsible for the absence of a species from an 
area. The assumption is that a member of a wildlife guild 
will be present in a habitat if that habitat provides suitable 
life requirements for a guild member. If the first of two 
pieces of land, with the same possible guild structure, sus- 
tains representative species of only 25% of the guilds that 
can occur in that vegetative community, then that land is 
of lesser quality to the total wildlife community than is the 
second piece of land sustaining representatives of 75% of 
the guilds that can occur in that vegetative community. 
Lists of guilds found on an area in a particular vegeta- 
tive type can be compared with lists of guilds that can 
theoretically occur in that potential vegetation type. The 
determination of why guilds are apparently missing from 
an area can be a diagnostic exercise. Is a guild missing 
because of some structural deficiency (affecting food 
sources and breeding niches) in a guild block? Can some 
remedial management practice be exercised to correct 
structural deficiencies so that habitat for particular guilds 
can be enhanced? 
The Habitat Evaluation Procedures (HEP; U.S. Fish 
and Wildlife Service 1980) will accept a measurement of 
habitat suitability that is a ratio of guilds observed in a 
cover type divided by the numbers of guilds that can occur 
in that potential vegetative type. The bound of potential 
vegetation thus provides a standard of comparison for all 
those cover types within a potential vegetation type. The 
habitat suitability value provides a measure of habitat 
quality for base-line conditions. For future conditions, 
predicted wildlife guilds can be simulated from data de- 
veloped from the species-habitat matrices for the vegeta- 
19 
tive cover types hypothesized present at some future time, 
t. The denominator to determine future habitat quality 
remains the same because the potential natural vegetation 
remains the same. The numerator at time ¢ is determined 
from the guilds predicted present at t. Predicted guilds are 
determined from the presumed vegetative cover type and 
its structure at time ¢. The simulation to determine the 
wildlife community present at t may include any modifica- 
tions to strata, and rows and columns within strata, for 
the species-habitat matrix describing the cover type pre- 
sumed present at time ¢t. Multiple future conditions can be 
simulated for the decision-maker, to determine the effects 
of different management options on habitat quality at 
time t. The ratio developed from predicted guilds and the 
standard of comparison produces a measure of future 
habitat quality. 
The advantages of using the guilding technique in HEP 
are that guilding provides a measure of habitat quality 
pertinent to the total wildlife community on an area, and 
that simple faunal surveys can provide the field data for 
determining base-line habitat quality. 
Conclusions 
We have described a technique for associating wildlife 
species with the structure of wildlife habitat so a variety of 
computerized analyses can occur to help predict impacts 
on wildlife caused by changes in the structure of habitat. 
The concepts developed in this paper were developed with 
a data base formulated for 275 wildlife species occurring 
in the eastern ponderosa pine type of forest (Kiichler 1964, 
type 16) in southeastern Montana and northeastern Wyo- 
ming. Our technique is based on the assumption that a 
wildlife species can be categorized as occupying a discrete 
area within a two-dimensional “species-habitat” matrix. 
Food sources and breeding niche requirements make up 
the two axes of the species-habitat matrix and numeric 
values identify the cells representing the food sources and 
breeding niche requirements necessary for each wildlife 
species. A variety of statistical analyses can be accom- 
plished on the areas (combinations of cells) occupied by 
different wildlife species so that a description of the struc- 
ture of the wildlife community within a habitat type can 
be developed. The assumptions made in the development 
of the species-habitat matrix are included in the text. 
The models produced in the present study indicate that 
(1) the species-habitat matrix provides a useful way to 
order wildlife information, (2) the statistical analyses pro- 
vide a structuring of the wildlife community that is bio- 
logically reasonable, and (3) a variety of management 
needs are served in a unique way because the analyses per- 
tain to the total wildlife community. In particular, the 
computer graphics illustrate, as expected, the complexity 
of the wildlife community that can occur as guild blocks 
are added to surface cover, and results from the cluster 
analyses describe, as expected, the increase in numbers of 
