The breeding biology of House Martins nesting in artificial boxes was examined in relation to aerial insect abundance and weather. The insect food supply was continuously monitored with a suction-trap at 40 ft above ground level. Aerial insect abundance started to rise from the winter level in April and reached a high stable state in June. Insects remained abundant until September; the subsequent decline continued at least until the end of October. The size of first clutches (mean size 3.87) was correlated with the abundance of aphids in spring, though not with temperature. It is suggested that this was an adaptive response because high food levels in May were associated with high levels in the main nesting period. Clutch-size declined through the season but was not matched by changes in food abundance. Second clutches (mean size 2–95) were most frequent following a high food level in June.
No correlation was found between egg-weight and preceding food conditions, nor with season, clutch-size or order of laying within clutches. If food was scarce on the day of laying of the first egg, these clutches suffered a suspension of laying of subsequent eggs. Some of the young in these nests showed a reduced growth rate. The duration of incubation (mean 14.6 ± 1.1 days) was not apparently correlated with food abundance, however it must have been potentially extended by extreme shortages because incubation duties were neglected at such times. Infertility was probably the main cause of hatching failure (14.2%).
A model was used to predict the date of onset of laying. It was based on the assumption that nestling tissue was produced with the same energetic efficiency as eggs. The model indicated that House Martins could collect enough food to lay eggs earlier than the observed date in most years. The high probability of food shortage before the laying period, however, appeared to discourage laying at this time. The observed onset of laying coincided with the appearance of aphids in the air, probably because they comprised an abundant and stable food source for egg formation. Breeding generally occurred in the period of highest food abundance. Pairs rearing only a single brood each year did so in July when food levels were highest.
The growth of first brood nestlings was more closely associated with food levels than any other factor investigated, while in second broods, rainfall was found to have the greatest influence. The total variance in growth explained by environmental factors in the first broods (41%) was greater than in second broods (22%). This relative independence of second broods from adverse environments probably arose from more abundant food and the feeding of some of these broods by the young of first broods. An exceptionally late brood showed an extended nestling period indicating deteriorating conditions later in the breeding season. It is proposed that the spread of laying may reflect differences in the feeding efficiency of nesting House Martins, because the growth of two early first broods was more rapid at given food levels than two late broods. Laying patterns conformed to a distribution based on a progressive threshold mechanism. The more efficient pairs may be able to attain breeding condition and lay at low levels of food abundance, and hence breed earlier in the season.
Subsequent layings are facilitated by the rise in food levels in early summer. Mortality within the nest was low (5.8%) and associated with food shortage. Nestling periods (mean 30.6 ± 2.3 days) were not shown to be correlated with food abundance; they were more dependent on brood-size.
Recruitment into House Martin populations is thus widely influenced by food supply, particularly through an influence on clutch-size, the occurrence of second clutches and nestling mortality.