To relocate nests quickly and reduce the chance of attracting
predators (see Salek & Smilauer 2002), nests were marked by a stone
placed on a brick within 1.5 m. To minimise disturbance we did not
spend more than 10 seconds near the nest during inspection. When
a nest was found empty, the contents were carefully scrutinize and
recorded. Nests were recorded as successful when at least one of the
following criteria was apparent: small fragments of eggshell were
present in the nest lining; at least one chick was seen; behaviour of
the adults indicated presence of a brood. A nest was classed as
successful if at least one egg hatched. A nest was assumed to have
failed if it was found to be empty before the expected hatch date
(and did not comply with the above criteria), or if there was evidence
of predation (i.e. large egg fragments, disturbed nest lining, etc.)
(Galbraith 1988).During each visit, nests, eggs and chicks were counted and sorted
by nest-type (ground or roof). In a number of nests, asynchronous
hatching was observed, i.e. all eggs did not hatch simultaneously
and it took 20–43 hours until the complete clutch hatched. In those
nests, the young started moving out of nests within a couple of
hours and concealed themselves in nearby vegetative cover.Such
nests were observed at either midday or dawn because parents were always observed sitting over the eggs and young during these periods.
Along with two local inhabitants, we observed individual nests for
longer continuous periods (up to four hours) from a hide or vehicle
in order to spot fleeing chicks, and we searched vegetation for hiding
chicks. Roofs provided less cover for chicks than ground sites, thus
offering better opportunities to locate the chicks. In both nest-types
we observed most chicks before they left the nest.
Hatching success was calculated with the Mayfield method
(Mayfield 1975) as well as with the traditional method (% of eggs
that hatched successfully out of total eggs laid). Numbers of eggs
and chicks that hatched in ground- and roof-nests were compared
using two-tailed t-test (Zar 1984). The mean values were presented
with the standard deviation (±SD).