of interest. Although beset by complexity, descriptive studies, on the other
hand, have the advantage of naturalness: but to be detectable, matches between
body size and environmental gradients must be powerful enough to outweigh
background noise. In the descriptive and experimental studies reviewed below,
body size has been dealt with in two ways. Some researchers record the size of
every individual present, while others use the shortcut of ascribing to all indivi-
duals in a given taxon the full size normally attained (‘maximal size’ below).
The size range of invertebrates also varies, with a few studies including the
meiofauna (42–500mm; Fenchel, 1978 ) while most have focused on macroinver-
tebrates alone. Invertebrates are invariably sampled to include both those living
on the surface and those in the top 5–10 cm of the substratum.
Descriptive studies of abiotic gradients
Townsend and Hildrew ( 1994 ) predicted that small size would be more prevalent
in more frequently disturbed habitats, but that this pattern would be muted
where spatial heterogeneity was high. The conceptual basis for the first predic-
tion is that resilience in the face of frequent and intense disturbances would be
provided by young age at reproduction, many descendants per reproductive
cycle and many reproductive cycles per year, features that are themselves linked
to small size. The second prediction derives from the idea that some streams
possess a greater variety and quantity of physical refugia, such as large substra-
tum particles and interstitial spaces in the bed. In these cases, larger species may
still do well even in the face of disturbance. These hypotheses were tested in a
suite of 54 stream sites scattered throughout a large river catchment in New
Zealand. Disturbance regime was quantified by recording the movement on
several occasions through a year of painted particles representative of those
present on the bed of each stream. A refugium index was calculated from a set of
component scores for the proportion of the streambed where shear stresses
were low (<0.771 dynes cm^2 ), the percentage of bed particles that were large
(>128mm wide), and the depth of interstitial space below the streambed. The
proportion of the community made up of smaller individuals (maximal body
size<10 mm long) was indeed greater in streams with more disturbed beds and
low refugium scores (Fig.5.1a), a relationship that was not evident in the subset
of streams with higher refugium availability (Fig.5.1b).
High discharge events often produce a mosaic of patches, where substratum
has variously been scoured out, deposited or left unaffected (Matthaei, Peacock &
Townsend,1999 ). After bed patches have been disturbed, faunal recovery is
sometimes size related. Thomson (2002 ) noted after moderate floods in an
Australian river that invertebrates tended to be larger (later instars) than before
the event, indicating recolonization mainly by larger invertebrates that sur-
vived in refugia in the stream. After a large flood that disturbed much more of
the streambed, on the other hand, invertebrates were smaller (early instars)
BODY SIZE IN STREAMS 79