Landscape ecology is based on the premise that there are strong links
between ecological pattern and ecological function and process. Ecolog
ical systems are spatially heterogeneous, exhibiting considerable comp
lexity and variability in time and space. This variability is typicall
y represented by categorical maps or by a collection of samples taken
at specific spatial locations (point data). Categorical maps quantize
variability by identifying patches that are relatively homogeneous and
that exhibit a relatively abrupt transition to adjacent areas. Altern
atively, point-data analysis (geostatistics) assumes that the system p
roperty is spatially continuous, making fewer assumptions about the na
ture of spatial structure. Each data model provides capabilities that
the other does not, and they should be considered complementary. Altho
ugh the concept of patches is intuitive and consistent with much of ec
ological theory, point-data analysis can answer two of the most critic
al questions in spatial pattern analysis: what is the appropriate scal
e to conduct the analysis, and what is the nature of the spatial struc
ture? I review the techniques to evaluate categorical maps and spatial
point data, and make observations about the interpretation of spatial
pattern indices and the appropriate application of the techniques. Pa
ttern analysis techniques are most useful when applied and interpreted
in the context of the organism(s) and ecological processes of interes
t, and at appropriate scales, although some may be useful as coarse-fi
lter indicators of ecosystem function. I suggest several important nee
ds for future research, including continued investigation of scaling i
ssues, development of indices that measure specific components of spat
ial pattern, and efforts to make point-data analysis more compatible w
ith ecological theory.