Biologie van Wespendieven Pernis apivorus in het oerbos van Bialowieza

During 2003-06 Honey Buzzards were studied in Bialowieza Forest, eastern Poland, in an area of 131.6 km2 (52º41’N, 23º05’E). Most of study plot (84%) is covered by woodland, 13% consists of farmland (mostly meadows and marshes in river valleys and -­ abandoned -­ arable fields). About 2.5% of the area consists of villages (Fig. 1). The strict reserve of Bialowieza National Park lies within the study area. Tree species composition within and outside the strict reserve are shown in Table 1. Territories were mapped by scanning for birds from tops of tall Norway Spruces Picea abies that towered above the forest canopy. Duration of treetop sessions ranged from 30-­390 minutes per tree. In 2003-­06, observations were conducted from resp. 60, 42, 56 and 34 treetops totalling resp. 170, 118, 167 and 110 hours in May-­August (mostly July and August). When possible, birds were sexed and notes were made of the individually highly variable (male) plumage, moulting patterns and prey transportations. Prey-­carrying birds were followed as long as possible and locations where they ascended from or disappeared into the canopy were determined with help of a compass and by estimating distances. In case of food transportations, I attempted to find the nest, in most cases successfully, by triangulation and choosing observation posts closer to the presumed nest site. The majority of active nests were detected after several days of observing from treetops, and it is not likely that many were missed. I collected data on nesting tree species, nest height and position of nest in the canopy. Nests known from previous years were checked in June for occupation and clutch size. After chicks had hatched, nests were climbed every 4-­5 days. During nest visits, data were collected on presence of guarding or incubating parent(s), and biometrics of the chicks (maximum wing chord and weight). Food remains were identified to species (group) and diameter of wasp combs was recorded. Onset of laying was calculated from wing length of the oldest chick, using the growth curve in Bijlsma (1997). In order to get an impression of wasp abundance, I monitored the number of foraging wasps in the treetops from which I observed Honey Buzzards (Fig. 2). In many cases I was unable to identify wasp species, but Common Wasp Vespula vulgaris was by far the most common species visiting the tops, followed by Red Wasp V. rufa and Polistes spp. German Wasp V. germanica was common in the village of Bialowieza and surrounding fields, but was rarely observed in the woods. During 2003-­06, 26 territories were located on 131.6 km2 , of which 16 were occupied in all four years of study, 7 during three years and 3 during two years. Territories that were found in only one breeding season (3 in 2003 and 1 in 2006) are not regarded as separate territories, because they may represent misinterpretation of field observations. The number of 26 territories for the study area is slightly overestimated, as in some years some territories extended beyond the boundaries of the study area (Fig. 3). The density is therefore estimated at 25 pairs, or 16.7 pairs/100km2. The density of Honey Buzzards in the primeval forest of Bialowieza did not differ from Dutch populations breeding in fragmented landscapes with secondary, species-poor and partly non-­indigenous coniferous forests on poor soils (Table 4). Territories were distributed evenly across the study area, but with a tendency of avoiding the drier coniferous parts of the study area. There was no apparent difference between the density in the strict reserve (primaeval forest) and the more cultivated stands bordering the National Park. Nesting occurred in a wide range of habitats provided that woodland was in close range. Nearest-­neighbour distances of territories varied between 520 and 4080 m (mean=1680, SD=640, N=99). When restricted to nests, provided that no territories were in between, nearest-neighbour distances varied between 520 and 2760 m (mean=1430, SD=720, N=18). Including some nests outside the study area, 29 different nests were occupied by Honey Buzzards during 2003-­06. Most (17) were used only once during the four years of study, 11 were used twice and one nest was used three times. Nests may be used intermittently for longer periods of time, as shown in 2011 by a nest that had also been occupied in 2003. The majority of nests was originally built by a Honey Buzzard, one nest had been used previously by a Common Buzzard Buteo buteo, one was originally built by a Raven Corvus corax and six nests were evidently old (but original builder unknown). Honey Buzzards showed a preference for Linden Tilia cordata as nesting tree, as compared to the availability of tree species (Table 1). Maple Acer pseudoplatanus was more frequently used than expected, which was also true for Norway Spruce outside the strict reserve. Birch Betula spp., Ash Fraxinus excelsior and Alder Alnus glutinosa were used according to availability. European Aspen Populus tremulus, Scots Pine Pinus sylvestris, Hornbeam Carpinus betulus and Pedunculate Oak Quercus robur were not used as nesting tree. Nests were built in the crown of trees at heights between 14 and 37 m (mean=20.9, SD=4.7, N=29). In most cases, nests were built against the main trunk, six times in a fork of three or more branches and five times on a more or less horizontal branch at some distance from the trunk. Almost invariably, nests were well protected from the midday or afternoon sun by the canopy. Some nests were checked during incubation, but most nests were found during the nestling stage. Each of the ten nests checked during incubation contained two eggs. In nine of these (90%) egg hatched (8x 2 and 1x 1) and in eight (80%) young fledged (6x 2, 2x 1). The survey method based on following prey-­carrying adults in July likely underestimates the number of breeding attempts, as failures during incubation or during the early brood stage are easily missed. A nest was found in 42% of the territories, a clutch produced in 36%, eggs hatched in 35%, and young fledged in 28% of the territories. The single failure during incubation was due to desertion (1). Broods failed due to predation by Goshawk Accipiter gentilis (1), predation by unknown predator (1), breaking down of tree (huge Linden) during thunderstorm (1), and unknown reasons (3). Onset of laying, back-­calculated from wing length of the oldest chick, varied between 15 May and 27 June 2003-­06 (mean=25 May, SD=9.27, N=31)(Fig. 4). Earliest pairs in 2003-­06 started resp. on 15 May, 15 May, 17 May and 17 May, latest pairs resp. on 4 June, 27 May, 26 May and 27 June. The late clutch in 2006 was exceptional, but 2006 was an outlier in several other respects as well: fewer pairs were observed in the study area (Table 2), few pairs laid eggs and produced young, and of the six pairs with known laying dates, two were very late (22 and 27 June). This may have been a carry-­over effect from conditions encountered during the breeding season of 2005, when wasp numbers were poor, chick weights were 6% below average (corrected for age, see also below for chick conditions), nest attendance in the nestling period (chicks between 10 and 40 days old) of male and female was about 50% lower than average, and more breeding attempts failed during the nestling stage than in 2003-­04 and in 2006. All nests checked during incubation contained two eggs (N=10). Nineteen eggs measured on average 51.2 mm (SD=2.44) x 41.4 mm (SD=1.73). Brood size was 5x1 and 27x 2 (and 4x unknown), resulting in 1.8 young/nest. Reduction from clutch to brood size was caused by death of a chick during hatching. During this study unhatched eggs were not recorded. The number of fledglings amounted to 4x 1, 24x 2 and 2x unknown (mean 1.9). Partial brood reduction was recorded twice: once the youngest chick was killed by a Goshawk after the oldest chick had already fledged and once one of the chicks was killed by a Goshawk after fledging. The reproductive rate was 0.53 young/pair (including non-­laying and failed pairs). As found for density, reproductive values in Polish old growth and cultivated forests were very similar to those obtained in secondary production forests in The Netherlands (using the same field methods;; Table 4). Age difference between siblings was about 2.7 days (based on wing length), but speed of growth between oldest and youngest chicks were similar according to the steepness of growth rates in young 1 and 2 in Fig 5). Also, the ratio between wing length and weight, which could be regarded as a condition index, did not differ between oldest and youngest chicks (Fig. 6). The ratio weight/wing length expressed as residual of the mean (wing length >30 mm, mean set at 100) was in 2003 100.2 (SD=7.7, N=58 measurements of 12 young), in 2004 100.8 (SD=10.0, N=77 measurements of 15 young), in 2005 94.5 (SD=8.0, N=42 measurements of 14 young) and in 2006 102.1 (SD=11.1, N=66 measurements of 14 young). From hatching till 10 days old, almost invariably one or both parents were present on or nearby the nest (96% of cases, N=26 nest visits). Nest attendance gradually decreased from 81% at chick ages of 11-­20 days old (N=31), to 43% at chick ages of 21-­30 days old (N=49), 17% at chick ages of 31-­40 days old (N=29) and 0% at chick ages of 41-­50 days old (N=2). In 2003-­06, nest attendance during the nestling period (young between 10 and 40 days old) was respectively 47, 56, 25 and 57% (resp. N=30, 32, 24 and 23 nest visits). Of 1924 prey remains found on nests, the majority (96.8%) consisted of combs of social wasps (Table 3), mostly grey combs which could not be identified to species (except seven with identifiable pupae: 5x Vespula rufa, 2x Dolichovespula saxonica). Vespula vulgaris (yellowish combs) comprised a smaller part of the diet, but may have been underestimated since combs of this species are brittle and easily destroyed by trampling chicks. Vespa crabro, which normally nests in tree cavities, was rarely preyed upon, except in 2005 when wasp numbers were poor. The only other insect as prey of Honey Buzzards concerned the remains of a dung beetle Geotrupes spiniger. A small fraction of the prey remains consisted of birds, exclusively nestlings or recently fledged juveniles with feathers not yet fully developed: 21 Song Thrushes Turdus philomelos, 10 Blackbirds T. merula, 1 European Robin Erithacus rubecula and 3 unidentified passerines. Of 23 frogs, nine could be identified as Rana temporaria, the others as Rana spp. In the early breeding season the diet almost entirely consisted of wasp species that produce grey combs i.e. Red, Saxon and possibly German Wasp. European Hornets reached an early peak during the first half of July, but were not recorded as prey in August. The Common Wasp was an uncommon prey species in late June, but progressively increased in importance throughout the breeding cycle until almost half of all prey remains in August belonged to V. vulgaris (Fig. 7).