(Underwood and Roth 2002). However, to measure annual productivity it is not
necessary to find all attempts, only the successful ones. The breeding attempts of
many species become easier to detect as they proceed: for example, because adults
alarm-calling or carrying food to large nestlings and especially fledglings are quite
easy to detect. Hence, reliable estimates of productivity may be obtained, provided
that the study area is carefully and systematically searched for evidence of this kind
at intervals shorter than the duration of the period of within which the parents
show conspicuous behavior. What is needed is to be sure that at most a negligible
proportion of successful attempts have escaped detection.
3.6.5 Indices of productivity from surveys during
the breeding season
If an index of productivity, rather than an absolute estimate, is acceptable, then
it might be sufficient to map bird territories and then visit them regularly and
score evidence of reproductive activity, such as the presence of nestlings or fledg-
lings, on an ordinal scale (Vickery et al. 1992). Grant et al. (2000) showed that
counts of breeding pairs of Curlews Numenius arquata, followed by surveys of
pairs showing alarm-calling behavior characteristic of birds with chicks, could
provide an index of breeding success.
3.6.6Use of simulation models
If data are available on the success of individual breeding attempts, including those
from unmarked adults, then productivity can be estimated using simulation mod-
els of breeding. To do this, additional information is needed on the propensity of
females to make further breeding attempts after the success or failure of a previous
attempt, the time of the season when females cease making further attempts and
the interval between successive attempts. Data on nest success and replacement/
multiple nesting can be combined in a simulation model to produce estimates of
the number of young hatched or reared per female per season (Beintema and
Muskens 1987; Green 1988; Green et al. 1997; Powell et al. 1999). Ricklefs and
Bloom (1977) estimated productivity using a more general model which required
information on daily nest failure rates, the distribution of nest initiation dates
through the season and the durations of various stages of the breeding cycle.
3.7 Timing of breeding
Information on the timing of breeding is often an important component of
a simulation model of breeding (see above) and is valuable in its own right, for
example, in studies of the effects of weather on breeding biology. Studies that
76 |Breeding biology