3.1 Honey-buzzard´s diet composition
A total of 2381 wasp comb fragments were collected in the 24 honey-buzzard´s nests and surroundings, accounting for 61533 large cells and 150847 small cells analysed. We identified combs from 6 wasp species: three with small cells (Vespula vulgaris , Vespula germanica , Polistes sp.) and three with large cells (Vespa velutina, Vespa crabro , Dolichovespula media) . The two most abundant species were common-wasps (Vespula vulgaris) and Asian-hornets, accounting for the 99.2% of the small cells and the 99.0% of the large cells, respectively (Table 1 ). This allowed us to simplify our statistical analyses by assuming that all small cell combs could be associated to common-wasps and all large cell combs to be Asian-hornets with minimum error.
We obtained a total of 752678 pictures in the 24 nests from which 4611 prey items were recorded, accounting for 3909 vespid items, 607 vertebrate items, and 95 unidentifiable prey items. 81.9% of the prey items were vespids. Common-wasps were the most abundant species (57.0%), and Asian-hornets the second (24.2%) (Table 2 ).
The best-fitting model explaining comb size recorded by the cameras included the interaction between wasp species and number of nestlings (Supplementary material Table 1 ). Combs of common-wasps had more cells than those of Asian-hornets (t = 16.26; p < 0.001) and this difference was greater in nests with one nestling (t = -3.49; p < 0.001). There was not an isolated effect of the number of nestlings on comb size (Figure 1 ). Neither the year nor the nestling age influenced the size of the combs delivered to the nest, which means that the size of delivered combs remained stable during the study period.
Baited traps captured a total of 15803 wasps, 5709 in 2020 and 10094 in 2021. In both years, the most abundant species was Asian-hornets (68.1% of all vespids captured: 63.9% in 2020 and 74.3% in 2021). The second most abundant species was common-wasps (30.3%: 34.7% in 2020 and 23.5% in 2021). The other vespids species (V. germanica ,Vespa crabro, Dolichovespula media, D. sylvestris, Polistes dominula, P. nimpha ) represented the 1.6% (1.3% in 2020 and 2.1% in 2021). This resulted in a mean Ivlev´s selectivity index value (E) of -0.71 for Asian-hornets and 0.52 for common-wasps, which indicated a preference of honey-buzzards for the native species and a rejection for the exotic species (Table 3 ).
Vertebrates accounted for 16.7% of all prey items (Table 2 ). Reptile was the most important group of vertebrates (9%). Lacertidae (8.3%) were predominant compared to Anguidae (0.3%). The images did not allow us to identify up to the species level all vertebrate items. However, it was remarkable the abundance of Iberian emerald lizard (Lacerta schreiberi Bedriaga 1878) accounting for the 56.9% of all reptiles. Birds were the second most important group of vertebrates (4.9%). Chicks without feathers (4.2%) were predominant respect to feathered birds (adults, fledglings, and older nestlings) that accounted for 0.7%. In the case of avian prey, the common wood pigeon (Columba palumbus L. 1758) was the most abundant species (50.5% of all avian prey). Amphibians (all frogs) and mammals (mainly volves) were rare in the diet accounting for 0.3% and 0.1% of all prey items, respectively.
The proportions of the four main types of prey (common-wasp, Asian-hornet, reptiles, and birds) changed significantly with the age of the nestlings and the number of nestlings, but not with the study year. The best-fitting model included the interaction between the age and the number of nestlings and the additive effect of the study year (Supplementary material Table S1 ). Our results show an increase in the mean proportion of vespids with the age of nestlings (from 70.8% at 16 days of age to 91.9% at 45 days) due to the increase of common-wasp (from 38.4% to 74.6%) while Asian-hornet proportions decreased (from 32.4% to 17.4%). This caused an important reduction in the mean proportions of vertebrates with the age of nestlings, which constituted the 29.2% of the diet at 16 days of age and 8.1% at 45 days, showing a drastic reduction of reptiles (from 17.7% to 5.4%) and avian prey (from 11.5% to 2.7%). The interaction of the number of nestlings and age also showed a significant effect over the proportions of prey groups delivered (Likelihood ratio = 47.63, df = 2, p < 0.001), wherein the proportion of common-wasps increased more intensely with the age of the nestlings in nests with two nestlings (Figure 2 ). We observed a non-significant effect of year over the mean proportion of each prey group (Likelihood ratio = 7.14, df = 3, p = 0.067, Figure 3) .
3.2 Rates of prey delivery in the honey-buzzard´s nests
Honey-buzzards delivered, on average, 7.13 prey items per day: 5.5 were vespids, 1.57 were vertebrates, and 0.1 were unidentified prey (Table 4 ). The best-fitting model to explain variations in the daily rate of prey delivery included the number and age of nestlings (Supplementary material Table S1 ) but just the number of nestlings showed a significant effect on the response variable (z = 5.34; p < 0.001). Nests with two nestlings received a mean of 2.57 prey items per day more than nests with only one nestling (Figure 4 ). The daily rate of prey delivery did not vary significantly neither with the age of nestlings nor year.