28
A. MACFADYEN
of leys, not producing a dense turf, Raw’s modification of the Salt and
Hollick extractor is indicated for the small arthropods. The field as a
whole would be stratified and periodical samples taken on a random
basis. It is perhaps worth stressing here that considerable seasonal
variations in arthropod numbers do occur (Boness, 1953; Dhillon and
Gibson, 1962 ; Macfadyen, 1952) and that quite frequent sampling
occasions (perhaps bi-monthly ) would be necessary to avoid false conclusions. Once again the use of dominance order or of ranking coefficients would probably save much labour in a study of this kind,
although it must be admitted that such methods do not yet seem to
have been applied to this type of work. A preliminary investigation
would certainly be required to determine the optimal size and number
of the sample-units. The flotation method has the advantage that
quite a variety of sample sizes can be treated and it might well be
worth while to take two or more sets of sample-units of different sizes
and to treat them in flotation chambers with correspondingly different
mesh sizes. I n this way the larger but less common species, up to a t
least the smaller species of beetles, would be extracted in adequate
numbers. There would remain, however, the quite sparse species,
mostly predators, such as the large ground beetles and the spiders. In
a non-quantitative study of this kind it might be best to sample these
with the aid of pitfall traps provided these were operated almost continuously over quite long periods. This is necessary because the catches
i n such t-aps vary very much with weather and other conditions and
also because most of these animals have marked periods of seasonal
abundance (see for instance Van der Drift, 1959). Obviously the different
groups sampled by these separate methods would be treated in parallel
and the changes of species in each major group would be mainly of
significance within that group.
3. A trophic study of a single species or a limited group of species
in a particular soil fauna. No particular species specified.
This is a class of study which has been very little attempted so far;
work includes an investigation by Clark (1954) into the Amphipod
Talitrus sylvaticus in an Australian rain forest, and the Durham studies
on the Pennine peat fauna (see Cragg, 1961). The object of the work is
to gain insight into the trophic importance of particular species by
measuring the seasonal changes in abundance, age structure and metabolism. The demand for a truly quantitative census limits the range of
possible species but our ignorance is a t present so profound that the
investigator can afford t o choose in this respect and his attention is
likely t o be attracted by species which are shown, in a preliminary
survey, t o be abundant. The choice of size and number of samples will
be made in accordance with the principles discussed in Section 111.
A. MACFADYEN
of leys, not producing a dense turf, Raw’s modification of the Salt and
Hollick extractor is indicated for the small arthropods. The field as a
whole would be stratified and periodical samples taken on a random
basis. It is perhaps worth stressing here that considerable seasonal
variations in arthropod numbers do occur (Boness, 1953; Dhillon and
Gibson, 1962 ; Macfadyen, 1952) and that quite frequent sampling
occasions (perhaps bi-monthly ) would be necessary to avoid false conclusions. Once again the use of dominance order or of ranking coefficients would probably save much labour in a study of this kind,
although it must be admitted that such methods do not yet seem to
have been applied to this type of work. A preliminary investigation
would certainly be required to determine the optimal size and number
of the sample-units. The flotation method has the advantage that
quite a variety of sample sizes can be treated and it might well be
worth while to take two or more sets of sample-units of different sizes
and to treat them in flotation chambers with correspondingly different
mesh sizes. I n this way the larger but less common species, up to a t
least the smaller species of beetles, would be extracted in adequate
numbers. There would remain, however, the quite sparse species,
mostly predators, such as the large ground beetles and the spiders. In
a non-quantitative study of this kind it might be best to sample these
with the aid of pitfall traps provided these were operated almost continuously over quite long periods. This is necessary because the catches
i n such t-aps vary very much with weather and other conditions and
also because most of these animals have marked periods of seasonal
abundance (see for instance Van der Drift, 1959). Obviously the different
groups sampled by these separate methods would be treated in parallel
and the changes of species in each major group would be mainly of
significance within that group.
3. A trophic study of a single species or a limited group of species
in a particular soil fauna. No particular species specified.
This is a class of study which has been very little attempted so far;
work includes an investigation by Clark (1954) into the Amphipod
Talitrus sylvaticus in an Australian rain forest, and the Durham studies
on the Pennine peat fauna (see Cragg, 1961). The object of the work is
to gain insight into the trophic importance of particular species by
measuring the seasonal changes in abundance, age structure and metabolism. The demand for a truly quantitative census limits the range of
possible species but our ignorance is a t present so profound that the
investigator can afford t o choose in this respect and his attention is
likely t o be attracted by species which are shown, in a preliminary
survey, t o be abundant. The choice of size and number of samples will
be made in accordance with the principles discussed in Section 111.