Study site and sampling
The study was conducted from 2012 to 2017 in a free-living population of jackdaws (Corvus monedul a) breeding in nest boxes located south of Groningen, The Netherlands (53° 14’N, 6°64’E), which has been routinely monitored since 2005. Nestboxes were routinely checked to assess egg laying and, after incubation commenced, both pair-members were identified by their unique colour ring combination. Brood size manipulations are described in detail in . In brief, broods were manipulated when the oldest nestling was 4 days old (hatching date = day 0, hereafter, day 5). Dyads of broods that were manipulated were matched by clutch size and laying date (±1 day). Nestlings were cross fostered so that the net manipulation was ±2 nestlings, creating enlarged and reduced broods with both manipulated broods containing nestlings hatched in both broods in the dyad . Biometric measurements and weight (to the nearest 0.1 g) were taken on days 5 and 30, and a blood sample was taken from the brachial vein. Upon collection, all blood samples were stored in 2% EDTA buffer at 4– 7°C and within 3 weeks they underwent snap freezing in a 40% glycerol buffer for long-term storage at −80°C.
Parasite counts (hereafter Carnus counts) were recorded for each nestling on every visit to the nest (due to regular nest checks and the experimental manipulation, data on parasite infestation were available for days 0, 1, 5, 10, 11, 20 and 30). Parasites were located by visual inspection of the nestlings’ body, with particular attention to areas where either the parasite or bite marks are common (under the wings and legs and belly). From 2012 to 2014 both parasite counts and bite marks on the belly were recorded (n = 229), while from 2015 to 2017 only parasite presence was recorded. Bite marks were recorded as a ‘belly score’ index from 0-4 (where 0 = no marks, 1 = under 10 marks, 2 = 10-20 marks, 3 = 20-50 marks, 4 = 50 or more marks).
We validated the belly score index as indicator of parasite presence in the nest (for cases when no Carnus flies were seen but belly scores was > 0) using a subset of individuals for which both parasite counts and belly score data were available. Belly score and Carnus counts predict the same outcome (belly score = 0Carnus counts = 0 vs. belly score > 0 Carnuscounts > 0) in 94% of cases (χ23 = 180.84, p-value < 0.0001). We further validated that belly score is a good indicator of parasite abundance (Table S1, Fig. S1) with a generalised linear mixed model with Poisson error distribution with Carnus counts as dependent variable and belly score index, age, their interaction, and year as fixed effects. This model included foster nest and colony as random effects to control for repeated measures and spatial variation. Because parasites often move from one sibling to another within the nest (, we scored the whole brood as ‘infested’ when at least one nestling from a brood was infested at any sampling point.
We recognize that variation in parasite infestation would ideally have been experimental in origin but in this study we relied on natural variation. This may have biased our results when Carnus infection varies systematically between or within colonies, or when jackdaws could detect Carnus infection when selecting a nest site, but neither appears to be the case. That the flies are widespread and first appear long after (egg) incubation has led us to assume for the present study that jackdaws have little control over whether they breed in a nest box with or without Carnus infection.