The metacommunity concept incorporates space and dispersal into community ecology.  However, metacommunity theory does not explicitly consider linear, or simple-branching dispersal networks which are common in both natural and anthropogenically-modified systems.  Stream systems are obvious representatives.  We used large-scale survey data for fish and benthic macroinvertebrates (Maryland Biological Stream Survey) to address two questions related to stream metacommunities:  1)  Do the constraints imposed by linear-type dispersal networks affect metacommunity composition on a large scale?  and 2) Are network effects similar among different stream fauna (fish vs. macroinvertebrates)?  Our analysis of the MBSS data revealed that 1)  The linear-type dispersal network of streams did constrain community composition through dispersal limitation.  However, effects were most prominent in higher-order (3^rd) streams.  In headwater streams, our analysis indicated environmental filtering rather than dispersal limitation was the chief determinant of community structure.  2)  Surprisingly, effects were similar for both fish and macroinvertebrates, though the signal was stronger for fish, most likely because macroinvertebrates have the capacity for dispersal outside of stream networks.  We suggest that the metacommunity concept is a useful framework within which to consider stream communities, but researchers should take care to explicitly consider the specific constraints that streams impose on metacommunities.