A new technique is offered that combines information on food supply, feeding rates and vegetative cover to predict the feeding sites of several sparrow species. New research will determine to what extent this technique which has been successful on a very local scale, can be used to predict the abundance andhabitat distributions of these same species in a complex mosaic of many habitats. An "iso-preference" technique will be employed at the local scale of measure habitat preferences quantitatively and to predict how the distribution of several species will change as a function of food depletion and habitat manipulations. The iso- preference technique ranks habitats according to preference of each species based on food and cover characteristics of the habitats. To predict patterns of habitat utilization, information on habitat preference and habitat availability must be combined. Therefore, it is proposed to create a mosiac of experimental habitats in replicate old-field landscapes. The food and cover characteristics of both the experimental habitats and the old-field matrix surrounding them will be manipulated. Since limitations of this approach of "scaling up" from the local to the landscape scale are recognized, simultaneous "scaling down" will be done by initiating a study in larger, more natural landscapes. The scaling down will determine how characteristics of the matrix in which a study site is embedded influence the abundance and distribution of sparrow species on the focal study site. With this hierarchial approach, the extent to which variation in sparrow abundance and community structure is controlled in the resources on the study sites versus differences in the characteristics of the surrounding landscapes will be determined.
