size and phosphate concentration. They reasonably assumed that phosphate
availability was related to stream primary productivity, noting higher algal
chlorophyll a and total invertebrate abundance in more phosphorus-rich streams.
Average body size was marginally but significantly higher in nutrient-rich
(total phosphorus> 40 mgL^1 ) than nutrient-poor sites (< 25 mgL^1 ). It may be
that larger invertebrates are more competitive at gaining food or space in
productive environments.
In contrast, larger invertebrates might be at a disadvantage in the presence of
large predators such as fish. Size-selective predation may reduce average body
size in a community by reducing the mean size of prey species that occur there,
or by driving larger taxa extinct. In his study of two neighbouring New Zealand
streams that differed in terms of the presence of brown trout (Salmo trutta),
Huryn (1998) found that of the ten largest invertebrate taxa, in nine cases
individuals were on average significantly smaller in the stream with trout.
Primary productivity was considerably higher in this stream, so it seems that
size-selective predation has overridden any tendency towards larger size with
greater productivity. Note that the New Zealand crayfishParanephrops zealandicus,
a very large invertebrate, is rare or absent from streams where brown trout
occur (Usio & Townsend,2000 ). Moreover, crayfish themselves seem capable of
exerting size-selective predation, reducing the density of larger representatives
of carnivorous chironomids (Usio & Townsend,2004 ).Descriptive studies of complex gradients
Given the multiplicity of factors that can influence macroinvertebrate size-class
structure, it may be surprising that patterns are sometimes detectable when
complex gradients are the focus of attention. Dole ́decet al. (2006) classified
macroinvertebrates into five maximal size classes (<5to>40 mm), and com-
pared their representation in replicate New Zealand streams whose catchments
were in each of four land-use classes: ungrazed native tussock grass, grazed
tussock, ‘improved’ pasture (sheep or cattle), and heavily stocked pasture
(deer or dairy). Average invertebrate size increased with agricultural intensity,
perhaps reflecting higher stream productivity fuelled by increased nutrient
concentrations in more intensively farmed pasture, or higher levels of contam-
inants, which may influence small animals disproportionately (Hendriks &
Heikens,2001; and see below).
Altitude represents another complex gradient because hydraulic conditions,
disturbance regimes and substratum size often vary along the length of streams.
In a Europe-wide study, Statzneret al.(2004) found that the representation of
small individuals (maximal size 2.5–5mm) declined, and larger ones (10–20 mm)
increased with increasing altitude. The pattern was clear but, as with the land-
use study, its cause was not. The authors speculated that oxygen richness in the
cold, turbulent upstream areas might have facilitated respiration of larger82 C. R. TOWNSEND AND R. M. THOMPSON