BUTEO 14 (2005): x-y Does urban structure explain shifts in the food niche of the Eurasian Kestrel (Falco tinnunculus)? Vysvětluje struktura městského prostředí posun potravní niky poštolky obecné (Falco tinnunculus)? BORATYŃSKI Z. (1), KASPRZYK K. (2) (1) Zbigniew Boratyński, Institute of Environmental Science, Jagiellonian University, ul. R. Ingardena 9, 30-060 Krakow, Poland; e-mail: bora@eko.uj.edu.pl (2) Krzysztof Kasprzyk, Institute of Ecology and Environment Protection, Nicolas Copernicus University, ul Gagarina 9, 87-100 Toruń, Poland ABSTRACT. We studied the diet composition of Eurasian Kestrel (Falco tinnunculus) in Ziemia Chełmińska (Kujawy, north-central Poland) from autumn 2000 to early spring 2002 to ask if differences in landscape exploitation impact this species food niche. We analyzed a total of 1068 pellets; 2092 prey individuals were found in three localities. A total of 32 prey categories were detected. Mammals represent a substantial part of the diet, approximately 76 % - by number (N) and 97 % - by biomass (B). Voles, primarily Common Vole (Microtus arvalis), were the main prey items: 65 % - N, 86 B. The remaining groups had a small impact on total diet biomass. However, insects were frequently detected: 23 % - N. Differences in food composition between sites reflects a higher frequency of insects and generally more categories in sites with a higher habitat heterogeneity - Toruń, compared to Chełmża and Papowo Biskupie (which were mostly large fields). Diet specialization vs. prey availability/abundance is discussed. INTRODUCTION The Eurasian Kestrel (Falco tinnunculus) is a frequent seasonal migrant or breeder in urban and suburban areas. It hunts mainly on small mammals, generally voles (e.g. RIJNSDORP et al. 1989, VIITALA et al. 1995, KOIVULA et al. 1999, THIRGOOD at al. 2003). This forces the raptor to optimize daily patterns of flight-hunting behaviour and site choice in relation to vole abundance. As has been noted, some vole-kestrel interactions can result from this type of monophagy, e.g. stabilization of vole population dynamics, as well as synchronization between prey and predator population dynamics (GALUSHIN 1974, VILLAGE 1982, IMS & ANDREASSEN 2000, HUITU at al. 2003). On the other hand the importance of other alternative types of prey in kestrel diet, such as birds and insects, has been suggested too (YALDEN & WARBURTON 1979, KORPIMÄKI 1985, RO- MANOWSKI 1996, PIATTELLA et al. 1999). Here we address the question of whether the Eurasian Kestrel is a true diet specialist, or whether prey choice reflects prey availability/abundance. We shed light on this important aspect of the autecology of this species by describing the diet composition between Kestrel localities with different landscape structure in Ziemia Chełmińska, Poland, verifying the importance of different constituents of Kestrel diet and describing variation between sites.
MATERIAL AND METHODS The study was carried in autumn through early spring during the 2000-2002 seasons in the Ziemia Chełmińska district (Kujawy, north-central Poland). Kestrels were detected by finding active roosts in monuments (churches, old buildings, town halls) and modern (apartment) buildings or ruins in Ziemia Chełmińska. These structures were used as night or bad weather shelters. Pellets were collected at last every third month in each collection place. Because pellet collections were mostly small, they were split into five groups for each locality separately. Comparisons were done for pellets from the autumn only, due to their similarity in collection time (Table 1). Three localities represented different types of anthropogenically modified environment. Toruń ( T, city: 206 thousand people) is surrounded by rural and agricultural areas with forested fragments. Chełmża ( C, town: 15.5 thousand people) and Papowo Biskupie ( P, village: 700 people) are oth surrounded by large fields. Chełmża is located near Chełmżyńskie Lake with a narrow belt of trees along the lakeside. Pellets in Papowo Biskupie were collected from the ruins of a castle located on a field border where birds could potentially hunt. A total of 1068 pellets were collected from Toruń, Chełmża and Papowo Biskupie (487, 303 and 278 respectively). Pellets and prey remains were dissected in water. Prey items were identified using diagrams and diagnostic keys (Yalden & Warburton 1979, Pucek 1984) and by comparison with specimens collected and deposited in the departments of Zoology and Animal Ecology, Nicolas Copernicus University, Toruń, Poland. Mean masses for prey categories were estimated using data from the literature (Yalden & Warburton 1979, Pucek 1984). When possible, a minimum number of prey items per pellet were estimated by summation of the most numerous elements of one body side (e.g. mandibles of mammals). In most cases, it was not possible to count the number of insects per pellet using this method. However, this allowed estimation of the frequency of occurrence of prey number (N) and prey biomass (B) in all collected material. Fig. 1 Percentage of variance explained by the following prey categories (extracted from PCA Component Score Coefficient Matrix). Obr. 1 Procenta variability vysvětlená jednotlivými kategoriemi potravy (výsledky PCA, Component Score Coefficient Matrix).
All analyses were performed using STATISTICA 5.1 and SPSS 11.5 software packages. Diversity indexes: Simpson; D = 1 Σ p 2 and Shannon-Wiener index; H = - Σ (p)*(log 2 p) were used to compare diet diversity from three localities during autumn (Table 1). The Chi-square method was used to compare the indexes as well as the differences in mean number of vertebrate specimens per pellet between localities. Principal Component Analysis (PCA-biplots, programme CANOCO 4.02, TER BRAAK & ŠMILAUER 1998) was used to estimate which of the seven obtained prey categories have the largest impact on variation in kestrels diet (Fig. 1) and for visualising the differences in diet composition at three localities (Fig. 2). a) b) Fig. 2 PCA-biplot diagrams for both number (a) and biomass (b) of main prey categories. X and Y axis refers to correlation matrix between prey categories. Obr. 2 Ordinační diagramy (PCA) pro počty (a) a biomasu (b) hlavních složek potravy. Osa X a Y je odvozena z korelační matice mezi kategoriemi. RESULTS A total of 2092 prey individuals were identified from three sites with a total biomass of 37055 g. In 996 pellets (those containing vertebrates), 1290 vertebrates were found (locality number of vertebrates/number of pellets: T 590/482, C 324/239, P 376/275). The mean number of vertebrate specimens per pellet differed significantly in Toruń from other localities (1.2 instead of 1.4; χ 2 = 19.97, df = 2, p < 0.01, n = 3) indicating a lower intake of these prey categories there. Because highly fragmented insect exoskeletons easily drop out of pellets, it was not reasonable to accurately estimate the number of invertebrates per pellet. Prey items were allocated into one of 32 taxonomic categories (18 for vertebrates and 14 for invertebrates) referred to as 13 vertebrate and 13 invertebrate separate species. Small mammals were the most frequently observed prey item (76 % by number, 97% by biomass). Voles, including the Common Vole (Microtus arvalis), were the most abundant mammalian preys (as calculated after splitting with Microtus sp.: 65% by number and 86% by biomass). The second most important prey categories were insects, although by number of specimens only (23%). Other prey categories were low in
abundance and had only a small impact on total biomass. However, there were significant differences between localities in both diversity indexes (Table 1; both: χ 2 = 174.75, df = 2, p < 0.001, n = 3). This resulted from different proportions of major groups of prey and absence or presence of uncommon prey in diet. Toruń most of the prey species were found (19 of 26). It was the most variable locality (D = 0.75 and H = 0.87). As shown by PCA analysis, prey categories that explained most of the variance in diet composition were: Soricidae and Ensifera; mostly because lack of them in Papowo Biskupie (Figure 1, Figure 2a and 2b). Altogether, insects explained 37.5% of variability. The lowest impact on differences between localities had the most frequent category: Microtinae, which explains 6.9% of variance only. This indicates stable and high proportion of he prey in Kestrel diet. DISCUSSION Kestrels are known to hunt in open areas such as farmland. The large proportion of mammals, especially voles (Microtus sp.), in the diet from Ziemia Chełmińska is in agreement with similar research from almost all of Europe (YALDEN & WARBUR- TON 1979, SHRUBB 1982, KORPIMÄKI 1985, ROMANOWSKI 1996, PIATTELLA et al. 1999). In open grassland (or farmland) the Common Vole is the most frequently found rodent (93-100%; JACOB & HEMPEL 2003) and the composition of predators diet reflects this abundance (e.g. GOSZCZYŃSKI 1977). Surprisingly, our results do not agree with the prediction that birds are important food source for Kestrels, particularly (but not only) in urban areas (KORPIMÄKI 1985, DAROLOVÁ 1989, ROMANOWSKI 1996, PIATTELLA et al. 1999). Large hunting areas of kestrels (up to 1.8 km), exceeding the expanse of urban areas probably cause this. On the other hand, a high frequency of insect occurrence previously noted by YALDEN & WARBURTON (1979) - % of diet composition is in concordance with our work (23%). But this is only supported by number, not by biomass (1%; maximum in Toruń = 3%), although this is probably highly underestimated. As noted by YALDEN & YALDEN (1985), the recovery of different rodent prey varies between 35 to 62%. Thus, the proportion of insects could be much higher as a consequence of digestion and loss of chitin fragments. Divergence between the three localities could be explained by differences in land structure surrounding the collecting sites. The more diverse landscape structure around Toruń (woodland, rural and grassland patches, riverbank) results in a greater variety of potential prey and therefore a diversified composition of kestrel food items (Table 1; Fig. 2a). Similarity between Chełmża and Papowo Biskupie seems to support this prediction. Since birds can be detected hunting within 1.8 km from their shelters (SHRUBB 1982), a varied landscape may better explain diversified diet rather than the expanse of urban areas. However, the expensed city may generate the scarcity of optimal hunting areas in Toruń and may lead to hunting in suboptimal areas with lower vole abundance. In this case, a greater variety of prey may be caught. This seems to be in accordance with general Optimal Foraging Theory predictions (EMLEN 1966, MACARTHUR & PIANKA 1966, SIH & CRISTENSEN 2001) that animals feed on items that maximize energy and/ or nutrient intake per unit of time. ACKNOWLEDGEMENTS. We are grateful to Dr. Michael A. Bogan (Museum of Southwestern Biology, University of New Mexico, Albuquerque, USA) for language revision, reviewer for con-
Table 1 Number, percentage and mass of prey specimens collected at three localities in Ziemia Chełmińska during autumns in 2000-2002, with summaries for main categories added and marked ( sum in parenthesis). Tab. 1 Počty, procentuální zastoupení a váha jednotlivých druhů kořistí, zjištěných na třech lokalitách v oblasti Ziemia Chełmińska v podzimním období 2000-2002, se sumarizovanými výsledky pro hlavní kategorie potravy (označené sum v závorce). Prey category Toruń Chełmża Papowo Biskupie N % mass g % N % mass g % N % mass g % Microtus arvalis 118 15.40 2596.0 25.20 116 24.10 2552.0 28.02 91 26.40 2002.0 27.70 Microtus sp. 269 35.10 5918.0 57.46 233 48.40 5126.0 56.27 213 61.70 4686.0 64.84 Microtus subterraneus 0 0.00 0.0 0.00 1 0.21 19.5 0.21 1 0.29 19.5 0.27 Microtus oeconomus 2 0.26 70.0 0.68 0 0.00 0.0 0.00 1 0.29 35.0 0.48 Clethrionomys glareolus 1 0.13 17.0 0.17 0 0.00 0.0 0.00 0 0.00 0.0 0.00 Microtinae (sum) 390 50.89 8601.0 83.50 350 72.71 7697.5 84.50 306 88.68 6742.5 93.30 Apodemus sp. 14 1.83 352.8 3.43 14 2.91 352.8 3.87 5 1.45 126.0 1.74 Apodemus agrarius 4 0.52 82.0 0.80 0 0.00 0.0 0.00 1 0.29 20.5 0.28 Apodemus (Sylvaemus) sp. 1 0.13 27.5 0.27 0 0.00 0.0 0.00 0 0.00 0.0 0.00 Apodemus flavicolis 1 0.13 31.0 0.30 0 0.00 0.0 0.00 0 0.00 0.0 0.00 Apodemus sylvaticus 0 0.00 0.0 0.00 0 0.00 0.0 0.00 1 0.29 24.0 0.33 Micromys minutus 11 1.43 88.0 0.85 15 3.12 120.0 1.32 5 1.45 40.0 0.55 Mus musculus 6 0.78 102.0 0.99 8 1.66 136.0 1.49 5 1.45 85.0 1.18 Rodentia unident. 24 3.13 528.0 5.13 28 5.82 616.0 6.76 3 0.87 66.0 0.91 Muridae (sum) 451 58.84 9812.3 95.26 415 86.22 8922.3 97.95 326 94.48 7104.0 98.30 Sorex minutus 3 0.39 12.0 0.12 1 0.21 4.0 0.04 0 0.00 0.0 0.00 Sorex araneus 4 0.52 32.0 0.31 0 0.00 0.0 0.00 0 0.00 0.0 0.00 Soricidae unident. 2 0.26 12.0 0.12 3 0.62 18.0 0.20 0 0.00 0.0 0.00 Soricidae (sum) 9 1.17 56.0 0.54 4 0.83 22.0 0.24 0 0.00 0.0 0.00 Lacerta agilis 5 0.65 50.0 0.49 0 0.00 0.0 0.00 0 0.00 0.0 0.00 Passeriformes 4 0.52 116.0 1.13 4 0.83 116.0 1.27 4 1.16 116.0 1.61 Acrididae 1 6 0.78 6.0 0.06 39 8.11 39.0 0.43 2 0.58 2.0 0.03 Acrididae 2 242 31.60 242.0 2.35 3 0.62 3.0 0.03 0 0.00 0.0 0.00 Ensifera (sum) 248 32.38 248.0 2.41 42 8.73 42.0 0.46 2 0.58 2.0 0.03 Geotrupes sp. 1 0.13 1.0 0.01 1 0.21 1.0 0.01 2 0.58 2.0 0.03 Dytictus sp. 1 0.13 2.5 0.02 1 0.21 2.5 0.03 0 0.00 0.0 0.00 Ditiscinae 11 1.43 11.0 0.11 2 0.42 2.0 0.02 1 0.29 1.0 0.01 Curculionidae 0 0.00 0.0 0.00 1 0.21 0.2 0.00 0 0.00 0.0 0.00 Carabus auratus 0 0.00 0.0 0.00 1 0.21 0.6 0.01 1 0.29 0.6 0.01 Carabidae 0 0.00 0.0 0.00 0 0.00 0.0 0.00 1 0.29 0.6 0.01 Elateridae 1 0.13 0.1 0.00 0 0.00 0.0 0.00 2 0.58 0.2 0.00 Staphylinidae 25 3.26 2.5 0.02 0 0.00 0.0 0.00 0 0.00 0.0 0.00 Sphecidae 0 0.00 0.0 0.00 1 0.21 0.1 0.00 0 0.00 0.0 0.00 Notonecta sp. 0 0.00 0.0 0.00 0 0.00 0.0 0.00 1 0.29 0.2 0.00 Coleoptera (sum) 39 5.08 17.1 0.17 7 1.47 6.4 0.07 8 2.32 4.6 0.06 Pallenia vagabunda 1 0.13 0.1 0.00 0 0.00 0.0 0.00 0 0.00 0.0 0.00 Insecta unident. 10 1.30 0.6 0.01 9 1.87 0.5 0.01 5 1.45 0.3 0.00 Total prey numbers 767 100.00 10300.0 100.00 481 100.00 9109.2 100.00 345 100.00 7226.9 100.00 Pellet numbers 487 303 278 Simpson index 0.75 0.69 0.55 Shannon0Wiener index 0.87 0.83 0.81
structive comments on earlier version of the manuscript as wall as editor, Dr. Ondřej Sedláček, for help and comments in statistics. Souhrn Složení potravy poštolky obecné (Falco tinnunculus) bylo zjišťováno od podzimu do časného jara v letech 2000-2002. Složení potravy jsme zjišťovali rozborem vývržků, sebraných na třech lokalitách v oblasti Ziemia Chełmińska. Vývržky byly sbírány pod nocovišti poštolek ve starší i nové zástavbě (kostely, starší domy, tovární haly). Vývržky byly preparovány ve vodní lázni a k určování nalezených částí byla použita srovnávací sbírka uložená na Univerzitě Mikoláše Koperníka, Toruń. Ráz krajiny je na jednotlivých lokalitách odlišný. Toruń ( T, velkoměsto: 206 tis. obyvatel) obklopuje ruderální a zemědělská krajina s remízy lesa a pastvinami, územím protéká řeka. V okolí sídel Chełmża ( C, město: 15,5 tis. obyvatel) a Papowo Biskupie ( P, vesnice: 700 obyvatel) se nachází zemědělsky využívaná krajina. Celkem bylo analyzováno 1086 vývržků, ve kterých bylo nalezeno 2092 položek kořisti. Jednotlivé taxony byly zařazeny do 32 kategorií (18 pro obratlovce a 14 pro bezobratlé). Průměrný počet obratlovců v jednom vývržku se pohyboval mezi 1,2 a 1,4 pro jednotlivé lokality, rozdíly mezi jednotlivými lokalitami byly statisticky průkazné (χ 2 = 9.97, d.f. = 2, n = 3). Savci představovali velkou část kořisti: 76 % kusů kořisti (N) a 97 % biomasy (B). Hraboši, především hraboš polní (Microtus arvalis), byly dominantní kořistí: 65 % (N), 85 % (B). Vysoké zastoupení hmyzu se projevilo pouze v počtech (23 %), v biomase představoval pouze 1 % (max. Toruń: 3 %). Výsledky potvrdily velký význam hlodavců v potravě poštolky, publikovaný v dřívějších studiích. Statisticky průkazné byly i rozdíly v indexu diverzity složení potravy na lokalitách (χ 2 = 174.75, d.f. = 2, n = 3; Tab. 1). Nejvyšší hodnoty dosahovaly indexy ve velkoměstě Toruń (D = 0.75, H = 0.87). Největší vliv na varianci v potravě měly kobylky (Ensifera) a rejskovití (Soricidae), 27,6 % respektive 23,4 % (Obr. 1 a Obr. 2). To může být vysvětleno vyšší heterogenitou prostředí a nedostatkem optimálních lovišť v okolí velkoměsta Toruń, v porovnání s lokalitami Chełmża a Papowo Biskupie. REFFERENCES DAROLOVÁ, A. 1989: [Falcon (Falco tinnunculus L., 1758) food in urban agglomeration of Bratislava]. Biológia (Bratislava), 44: 575-584. /In Slovak/. EMLEN, J. M. 1966: The role of time and energy in food preference. Am. Nat., 100: 611-7. GALUSHIN, V. M. 1974: Synchronous fluctuations in populations of some raptors and their prey. The Ibis, 116(2): 127-132. GOSZCZYŃSKI, J. 1977: Connections between predatory birds and mammals and their prey. Acta theriol., 22(30): 399-430. HUITU, O., NORRDAHL, K. & KORPIMÄKI E. 2003: Landscape effects on temporal and spatial properties of vole population fluctuations. Oecologia, 135: 209-220. IMS, R. A. & ANDREASSEN, H. P. 2000: Spatial synchronization of vole population dynamics by predatory birds. Nature, 408: 194-196. JACOB, J. & HEMPEL, N. 2003: Effects of farming practices on spatial behaviour of common voles. J. Ethol., 21: 45-50. KOIVULA, M., VIITALA, J. & KORPIMÄKI E. 1999: Kestrels prefer scant marks according to species and reproductive status of voles. Ecoscience, 6(3): 415-420. KORPIMÄKI, E. 1985: Diet of the Kestrel Falco tinnunculus in the breeding season. Ornis Fennica, 62: 130-137. MacARTHUR, R. H. & PIANKA, E. R. 1966: On optimal use of a patchy environment. Am. Nat., 100: 603-609. PIATTELLA, E., SALVATI, L., MANGANARO, A., FATTORINI, S. 1999: Spatial and temporal variation in the diet of the Common Kestrel (Falco tinnunculus) in urban Rome, Italy. J. Raptor Res., 33(2): 172-175. PUCEK, Z. 1984: [Identification key of Polish mammals]. PWN Warszawa. /In Polish/. RIJNSDORP, A., DAAN, S. & DIJKSTRA, C. 1989: Hunting in the Kestrel, Falco tinnunculus, and the adaptive significance of daily habits. Oecologia (Berl.), 50: 391-406. ROMANOWSKI, J. 1996: On the diet of urban kestrel (Falco tinnunculus) in Warsaw. Buteo, 8: 123-130.
SHRUBB, M. 1982: The hunting behaviour of some farmland Kestrels. Bird Study, 29: 121-128. SIH, A. & CRISTENSEN, B. 2001: Optimal diet theory: when does it work and when and why does it fail? Anim. Behav., 61: 379-390. THIRGOOD, S. T., REDPATH, S. M. & GRAHAM I. M. 2003: What determines the foraging distribution of raptors on heather moorland? Oikos, 100: 15-24. VIITALA, J., KORPIMÄKI, E., PALOKANGAS, P. & KOIVULA M. 1995: Attraction of kestrels to vole scent marks visible in ultraviolet light. Nature, 373: 425-427. VILLAGE, A. 1982: The diet of Kestrels in relation to vole abundance. Bird Study, 29: 129-138. YALDEN, D. W. & WARBURTON, A. B. 1979: The diet of the Kestrel in the Lake District. Bird Study, 26:163-170. YALDEN, D. W. & YALDEN, P. E. 1985: An experimental investigation of examining Kestrel diet by pellet analysis. Bird Study, 32: 50-55. (Received 28.3.2004, accepted 5.9.2004)