Recognizing that a further fundamental feature of stream ecosystems is their
flow-related disturbance regimes, Townsend & Hildrew (1994 ) developed predic-
tions about how disturbance might be linked to species traits related to resistance
(the capacity to withstand the disturbance) and resilience (the capacity to recover
rapidly after a disturbance). There have now been many tests of these predictions
(e.g. Reshet al., 1994 ), with support, for example, for the hypotheses that more
frequent disturbance should be associated with traits that confer resilience (adult
mobility, habitat generalism, rapid population growth) or resistance (clinging
behaviour, streamlined/flattened morphology, several life stages outside the
stream) (Townsend, Dole ́dec & Scarsbrook,1997 ). Others have sought to match
further habitat features to the representation of traits, including the size and
heterogeneity of the substratum (Bourassa & Morin,1995 ; Gayraud & Phillipe,
2001 ; Schmid, Tokeshi & Schmid-Araya,2002 ) and hydraulic variables (Statzner,
Gore & Resh,1988 ;Hart&Finelli,1999 ; Merigoux & Dole ́dec, 2004 ).
Among the diversity of physiological, morphological and behavioural traits
(related to gathering resources, surviving in the face of various threats, and
reproduction), body size has a pivotal position. The size structure of communities
can be directly influenced by ecological drivers including size-selective predation
(particularly by fish), size-related risk of dislodgement by hydraulic forces, and
the provision of streambed size-related refugia that may counter predation or
dislodgement risks. But we can also expect indirect links between size and habitat
dimensions because size is related to many other traits. For example carnivores
and shredders are over-represented in bigger size classes, reflecting the larger
size of their food particles. Moreover, species that are resilient because of short
life cycles, rapid growth and high reproductive rates, are also usually small
(Townsendet al., 1997 ). Furthermore, size may be positively correlated with
competitive status (Bourassa & Morin,1995 ) and negatively correlated with the
possession of an aquatic imago stage and ovoviviparity (Statzneret al., 2004 )as
well as vulnerability to environmental contaminants (Hendriks & Heikens,2001 ).
The focus of this chapter is the match between body size and influential
abiotic and biotic environmental gradients. We explore relationships with dis-
turbance, substratum particle size and hydraulic stress, and with primary
productivity and fish predation. We also test hypotheses in the context of
human-influenced habitat gradients, comparing streams in forested, native
grassland and pastoral settings. In thenext section, we review previous stream
studies of body-size patterns before proceeding to our analysis of a new database
for tributaries in the south of New Zealand.Different approaches for matching body size with
environmental gradients
The advantage of experimental over descriptive studies is their comparative
simplicity, controlling (ideally) for all factors except the environmental gradient78 C. R. TOWNSEND AND R. M. THOMPSON