HABITAT SELECTION BY AQUATIC INVERTEBRATES
337
from sieving the typical sediment of Corophium volutator (Meadows,
1 9 6 4 ~ ) . When offered paired choices Corophium always preferred the
finer of the two sediments, although again the mechanism by which the
animals distinguish between the sand grades is not known.
Landenberger (1968) has considered the changes in apparent
preference when qualitatively different choices, three species of
mussel, three species of gastropod and a chiton, are offered,
firstly in pairs, and then in a multi-choice experiment as food
to the starfish Pisaster. I n all but one pair of choices Pisaster showed
a significant preference for one alternative, and the order, or hierarchy
of preferences, was well defined and consistent among replicates. When
the seven prey species were presented together in a multichoice experiment, the preferences were in the same order but were in general
weaker-that is-not so clear cut as those when the prey were presented
in pairs. Similar studies of preferences for quantitatively different items
would be well worthwhile. Landenberger’s results can be extrapolated
to natural conditions as follows. Consider an environment in which
habitats A and B are randomly distributed at the same population
density in a uniform environment : then remove habitats A and B and
replace them by habitats A, B, C, and D, these to be distributed
randomly, but each at the same density, among the positions originally
occupied by A and B. If an animal encounters this environment and
prefers A to B to C to D, then the animal will aggregate in larger
numbers on A when only A and B are present than when A, B, C, and
D are present (Meadows and Campbell, 1972).
The difference between successive levels of a variable, where A, B,
C, and D represent the different levels, may be so small as to escape
detection, in which case an animal will not distinguish between A and B,
B and C or C and D, but might distinguish greater differences such as
those between A and C, B and D or A and D. We can regard this as the
slope of the variable (Harder, 1968, p. 159), and Amos (1969) and
Amos and Waterhouse (1 969) have demonstrated its ecological significance in the terrestrial environment during a study of the behaviour of
beetles along humidity gradients of different slopes. The concept could
be profitably applied to such variables as salinity, light and temperature
in aquatic environments, as indicated in Meadows and Campbell (1972).
So called “ noise ’’ or interference from other aspects of the environment might also obscure preferences (Landenberger, 1968). One might,
for example, class as noise the overriding of the preference for rough
over smooth surfaces by the preference for filmed over unfilmed surfaces
shown by larvae of Spirorbis rupestris (Gee, 1965) and also the influence
of other animals on habitat selection described by Clark (1956) and
337
from sieving the typical sediment of Corophium volutator (Meadows,
1 9 6 4 ~ ) . When offered paired choices Corophium always preferred the
finer of the two sediments, although again the mechanism by which the
animals distinguish between the sand grades is not known.
Landenberger (1968) has considered the changes in apparent
preference when qualitatively different choices, three species of
mussel, three species of gastropod and a chiton, are offered,
firstly in pairs, and then in a multi-choice experiment as food
to the starfish Pisaster. I n all but one pair of choices Pisaster showed
a significant preference for one alternative, and the order, or hierarchy
of preferences, was well defined and consistent among replicates. When
the seven prey species were presented together in a multichoice experiment, the preferences were in the same order but were in general
weaker-that is-not so clear cut as those when the prey were presented
in pairs. Similar studies of preferences for quantitatively different items
would be well worthwhile. Landenberger’s results can be extrapolated
to natural conditions as follows. Consider an environment in which
habitats A and B are randomly distributed at the same population
density in a uniform environment : then remove habitats A and B and
replace them by habitats A, B, C, and D, these to be distributed
randomly, but each at the same density, among the positions originally
occupied by A and B. If an animal encounters this environment and
prefers A to B to C to D, then the animal will aggregate in larger
numbers on A when only A and B are present than when A, B, C, and
D are present (Meadows and Campbell, 1972).
The difference between successive levels of a variable, where A, B,
C, and D represent the different levels, may be so small as to escape
detection, in which case an animal will not distinguish between A and B,
B and C or C and D, but might distinguish greater differences such as
those between A and C, B and D or A and D. We can regard this as the
slope of the variable (Harder, 1968, p. 159), and Amos (1969) and
Amos and Waterhouse (1 969) have demonstrated its ecological significance in the terrestrial environment during a study of the behaviour of
beetles along humidity gradients of different slopes. The concept could
be profitably applied to such variables as salinity, light and temperature
in aquatic environments, as indicated in Meadows and Campbell (1972).
So called “ noise ’’ or interference from other aspects of the environment might also obscure preferences (Landenberger, 1968). One might,
for example, class as noise the overriding of the preference for rough
over smooth surfaces by the preference for filmed over unfilmed surfaces
shown by larvae of Spirorbis rupestris (Gee, 1965) and also the influence
of other animals on habitat selection described by Clark (1956) and