It has been suggested that collections of affected sib pairs, or their nuclear families, may be an efficient method for screening for genetic linkages in schizophrenia. We present the data collected in five years from 15 hospitals in the state of Maryland in an effort to determine if such a collection scheme will be feasible. Probands in our sample were eligible for inclusion in the sample if they were white, were age 16 years or older, and carried a research diagnosis of schizophrenia. Family data are reported for 258 probands. Using the most stringent category of affected (RDC schizophrenia) revealed ten families with two or more affected sibs. The broadest category of affected (any psychotic disorder or psychiatric hospitalization) identified only 36 families with two or more affected sibs. We conclude that, if schizophrenia is a heterogeneous disorder with decreased penetrance, an effort to collect multiplex nuclear families is unlikely to provide enough data to identify genetic linkage. Alternatively, an effort to seek out and collect larger multiplex, multigenerational families rather than a collection of affected sib pairs may be more efficacious.