Spatial variance of observed measures such as density is no longer vie
wed as a statistical annoyance. It is now treated as a biologically im
portant quantity that changes value depending on the scale of measurem
ent. Processes that generate spatial variance are often inferred by ma
tching scales of maximum biological spatial variance to dominant physi
cal processes It the same scale. Success in this approach has been lim
ited to patchiness of plant communities along environmental gradients
and to patchiness of passive aquatic organisms relative to physical fl
ow structures. Some progress in formalizing spatial variance has been
made using empirical models derived from quantitative descriptions of
pattern, bur further progress requires theoretical models of spatial v
ariance and processes that generate variance as a function of spatial
scale.