Sampling macroinvertebrate communities from natural snags (large woody debr
is) is difficult, particularly in southeastern Australia where the predomin
ant river red gum snags are very large and heavy. This makes removal of nat
ural snags impractical. The Snag Bag is a new method for sampling macroinve
rtebrates from natural snags. It consists of a 250 mu m mesh bag with the s
ides closed by 'Velcro'. When opened out, one side of the Snag Bag can be p
assed under the snag and the bag is then closed around the snag, before bru
shing the surface down to remove invertebrates. The Snag Bag is removed fro
m the snag by loosening the 'Velcro' above the snag and closing it below, u
ntil the bag has almost been removed. The bag is then quickly pulled from t
he snag and out of the water. Snags with diameters of 5-30 cm at depths of
up to about 50 cm can be sampled. The Snag Bag was trialed in two lowland r
ivers in south-eastern Australia. Repeated sampling showed that its efficie
ncy was about 96%, comparing very favourably with other methods of collecti
ng macroinvertebrates in streams. Its efficiency was not affected by curren
t speed or the surface area sampled. A subsampling procedure was developed
and tested, as whole samples were often extremely large. A proportion of th
e sample equivalent to a set surface area (0.05 m(2)) was sorted. If fewer
than 300 animals were obtained, further known percentages of the original s
ample were processed fully, until 300 animals had been removed. The Snag Ba
g provided quantitative data that allowed discrimination between densities
of invertebrates on snags at two different sites on the same river.