and recording the activities, such as the number of items eaten, the number of
pecks, or the total amount of time spent foraging (use a second stopwatch). This
is repeated for other individuals. It is important to ensure that the samples are not
biased in a manner that affects the conclusions, for example, do not select active
individuals in preference to sleeping ones.
Scan sampling involves systematically scanning each individual in turn and
categorizing its behavior (e.g. sleeping, walking, or alert) at the instant when first
observed. Scan samples are repeated but usually leaving a time gap to increase the
independence of the records. Data can be collected using a Dictaphone, com-
puter, or tally counters. Individuals can also be classified at the same time into
categories (e.g. ages, sexes, habitats) to show differences in behavior between
categories. An example of a scan sample would be to compare the numbers of adult
and juvenile geese feeding in a flock. The method might involve starting from
one end of the flock and systematically scanning across the entire flock recording
the behavior and age of each bird in turn. Again the sampling should minimize
bias, for example, as a result of individuals at the edge of the flock differing in age,
status, intake rate, or vigilance from those in the center. Sampling across the flock
can minimize this source of bias.
Scan samples are useful for quickly determining the time budget of abundant
or flocking species. For example, a simple scan count of a hundred geese can be
done in 10 min but a 10-min focal watch of a single individual is clearly insuffi-
cient (the individual might sleep for the entire period). Focal watches are better
for dispersed species. They are also necessary when data are required on particular
individuals: for example, to relate interactions between age, sex, intake rate,
vigilance, walking rate, and aggression.
Data loggers have been used on Brünnichs Guillemots Uria lomviato determine
the time spent underwater, swimming, flying, and on the nest (Falk et al. 2000). The
loggers also provided data on diving duration and depth. Time budgets can also
be determined from radio-tracking if an activity sensor is fitted (see Chapter 6).
10.5.2 Time spent feeding per day
This is measured by assessing the mean number of hours spent away from the
roost per day and multiplying by the percentage of time spent feeding through
the day. Measures of percentage of time feeding have to sample through the day
to allow for diurnal patterns.
10.5.3 Night observations
Nocturnal observations of foraging behavior are difficult to obtain, although the
technology is improving. Recent, but expensive, equipment using a photocathode
Foraging behavior| 245