Small body size is predicted to be beneficial in disturbed hydraulic conditions
because it permits invertebrates to remain within the boundary layer. Snook
and Milner (2002) found that small maximal size (<10 mm) was possessed
by almost all individuals where hydraulic stress was highest in streams in
the Pyre ́ne ́es (Boundary Reynolds number>30 000), whereas a mix of small
and larger invertebrates occurred in less turbulent conditions. Merigoux
and Dole ́dec (2004) obtained contradictory results in their hydraulic study of
invertebrates in contrasting patches in a French river. In this case maximal
body size was actually larger in the more extreme conditions of high hydraulic
stress (estimated as shear stress on the bed). They argued that morphological
and behavioural traits, such as the ability to cling to the substratum or the
breadth of gills that aid in anchoring, may increase with individual size and
outweigh the predicted disadvantage of large size under hydraulically stressed
conditions.
Flash floods in desert streams are particularly intense disturbances. They can
cause more than 90% mortality in the juvenile stages of stream insects, but
recolonization is rapid via oviposition of aerial adults (Lytle, 2002 ). Thus, timing
of emergence may play a critical role in persistence of taxa in flood-prone
systems. Modelling the joint action of a seasonal limit for reproduction by the
caddisflyPhylloicus aeneusin the absence of a flood, and of the probability that
one or more floods might occur before this limit is up, led to two size-related
predictions (Lytle, 2001 ). First, because individuals can use the adult stage as a
refuge from floods but, by emerging early, pay an opportunity cost in lost
growth (and fecundity), larger juveniles are expected to emerge early while
smaller juveniles should risk more of the flood season in order to continue
growing. Second, because no floods occur in some seasons, smaller individuals
should risk all and emerge towards the end of the reproductive season, produc-
ing a second decline in body size in years where floods do not occur before the
end of the season. In his study of three Chihuahuan desert streams over two
years, Lytle ( 2002 ) indeed found that in most years there were two periods of
emergence and two consecutive declines in body mass at emergence. To test
whether these were associated with increasing flood probability and end of the
reproductive season, respectively, he used maximum-likelihood methods to
compare a null model with one incorporating a reproductive time constraint
and several others that added flood disturbances into the mix. The latter
performed best, indicating that flood dynamics played an important role in
determining body-size patterns, a pattern consistent with natural history obser-
vations in similar systems (Scarsbrook & Townsend,1993).
Descriptive studies of biotic gradients
Bourassa and Morin (1995; see above) extended their work on substratum grain
size in Canadian streams to study the relationship between invertebrate body
BODY SIZE IN STREAMS 81