(Brittas and Karlbom 1990) or nests that go undetected during visual searches
(Hill 1998). There is much scope for improvements of this type in census work
(Chapter 2).
Beyond the speed and precision with which radio-tagging can supply data is the
huge amount of information on individual life-histories that can come from
long-life tags. Tags that last for months or years from fledging can show how
performance (e.g. survival, dispersal, productivity) relates to individual use of
habitats and other resources as well as to age. Such data are important components
for individual-based modeling. Although data for individual-based demographic
modeling has mainly come from visual observations (Goss-Custard 1996;
Sutherland 1996), the first demographic modeling based on functional responses
originated from radio-tagging (Kenward and Marcström 1988). Radio-tagging also
has huge potential in field experiments, not only by providing detailed information,
but also by enabling minimal samples (because there is minimal unexplained vari-
ance in tests if all outcomes are recorded).
However, the benefits of radio tagging may be lost if data are collected and ana-
lyzed in ways that are biased, or the sample of tagged animals is biased (e.g. by dif-
ferential capture) or there are adverse impacts of tagging. It can be hard to obtain
control data to demonstrate absence of bias from tagging, especially with the low
statistical power that results from small samples. Tag attachment must therefore
be planned with great care.
6.2.3Considerations for tag attachment
Radio tags have been attached externally to bird beaks, necks, backs, legs, tail feath-
ers, and patagia, or implanted, depending on species and study requirements
(Table 6.1). Implanting usually requires veterinary supervision and licensing and
may put bird health at risk. Implants also have low antenna efficiency (unless a
transcutaneous antenna further increases the health risk) and are therefore best
avoided unless required for measuring physiological parameters.
The choice of attachment also depends on requirements for tag mass, sensors,
and detachment. Tags on 1–2 tail feathers are shed at the molt, but tend to be
molted prematurely if a feather carries more than 1% of bird mass (i.e. more than
2% on 2 feathers). Tail-mounts (Figure 6.2(a)) are convenient for mounting
sensors that indicate behavior from tail-posture (Kenward et al. 1982) or, as they
are relatively remote from the skin, for measuring ambient temperature. Nasal
saddles have been used to indicate feeding activity from head posture (Swanson
and Keuchle 1976), but must be very small if they are not to affect behavior.
Sensors on necklaces (Figure 6.2(b)) and leg tags can indicate general activity or
mortality. However, antennas are apt to break on leg tags and both methods are
Choice of techniques| 145