Folia Zoologia 2003-03_str 275
Transkript
Folia Zoologia 2003-03_str 275
Folia Zool. – 52(3): 275–286 (2003) Song repertoire and microgeographic variation in song types distribution in the corn bunting Miliaria calandra from Poland Tomasz S. OSIEJUK1 and Katarzyna RATY¡SKA2 Department of Animal Morphology, Institute of Environmental Biology, Adam Mickiewicz University, 28 Czerwca 1956/198, 61-485 Poznaƒ, Poland; 1e-mail: [email protected], 2e-mail: [email protected] Received 20 June 2002; Accepted 28 May 2003 A b s t r a c t . Corn buntings in the Wielkopolska region (W Poland) show a clear local dialect pattern of microgeographic song variation only in relatively dense and stable populations, which inhabited typical farmland landscape. In less preferred habitats, where males were much more dispersed, or in sites that where colonized recently, we found no such a pattern of song type sharing. The between-individual song type variation was higher in such sites and males from such locations did not sing any common dialect. The pattern of full and shortened song variant usage was rather inconsistent. Shortened song variants were used more frequently during counter-singing than solo-singing, but some song types were shortened more often than others regardless of the context. Key words: song variation, dialect, song variants, passerines Introduction Passerine songs undoubtedly play an important role in mate attraction and stimulation as well as male-male competition (C a t c h p o l e & S l a t e r 1995). As a sexually selected trait, songs vary substantially between species with respect to acoustic structure and temporal pattern. Many hypotheses involving intra- and inter-sexual selection and several putative constraints have been presented in an attempt to explain this variation (K r o o d s m a & B y e r s 1991, P o d o s 1997, G i l & G a h r 2002). Comparisons within and between species reveal that our knowledge is limited. In particular, influence of various factors on evolutionary trajectories in song complexity require investigation (K r o o d s m a & B y e r s 1991, K r o o d s m a 1996, V e h r e n c a m p 2000). Members of the family Emberizidae have been subjects of numerous bioacoustic investigations, which have employed a wide range of methods, from simple observations (A n d r e w 1987, M ø l l e r 1983) to comparative studies (C a t c h p o l e & M c G r e g o r 1985) and advanced playback experiments (S t o d d a r d et al. 1991, B e e c h e r et al. 2000). The corn bunting M. calandra (L., 1758), the only representative of the genus Miliaria possesses many similarities to its closest relatives (i.e. Emberiza spp.) but diverges from them in some traits, such as sexual dimorphism, mating system and song complexity (C a t c h p o l e & M c G r e g o r 1985, G r a p u t t o et al. 2001). The corn bunting thus offer a unique opportunity to compare relationships between song and other traits within the bunting family. The most characteristic features of the corn bunting are: the polygynous mating system and sexual dimorphism expressed only in size (males are ca. 8% larger than females). The majority of well-studied Emberiza spp., including yellowhammer E. citrinella, cirl bunting E. cirlus, ortolan bunting E. hortulana and reed bunting 275 E. schoeniclus, are monogamous (at least socially) and have more or less distinct plumage sexual dimorphism (C a t c h p o l e & M c G r e g o r 1985, C r a m p & P e r r i n s 1994). The corn bunting seems to be clearly different from its relatives also in terms of song. Admittedly, the corn bunting is also a typical discontinuous singer with a small song repertoire (typically 2–3 song types), like many other buntings, but it is characterized by a mosaic pattern of geographical variation, reflected in the formation of so-called “local dialects”. In such a pattern each male within a local dialect population sings only several song types characteristic of the population. Very few males sing song types of two dialects. Except for those unusual individuals, all song types within a local population of the corn bunting are shared (M c G r e g o r 1980, 1986, H o l l a n d et al. 1996). By contrast, other Emberiza buntings typically share only some song types within a local population (C a t c h p o l e & M c G r e g o r 1985, O s i e j u k et al. 2003). Unfortunately, no study has provided an unambiguous answer to the question of whether local dialects are functional (C a t c h p o l e & S l a t e r 1995). Local dialects sometimes persist over a decade, but the mechanism of their maintenance is also unknown (T r a i n e r 1983, C a t c h p o l e & S l a t e r 1995, H o l l a n d et al. 1996). In the case of the corn bunting, C z i k e l i (1982) suggested that usage of different song types is density-dependent and related to habitats occupied by the birds. However, there are no direct proofs of such a mechanism working and there are some disagreements between C z i k e l i (1982) and other authors in understanding the concept of song type and song dialect in the corn bunting (C r a m p & P e r r i n s 1994). The main aims of this study were (1) to describe song variation of the corn bunting from Poland (no other data on this subject are available from this part of Europe), (2) compare the pattern of microgeographical song types distribution between local populations of different density, and (3) compare these data with other populations studied earlier in Europe. Material and Methods Study area and population The study was carried out in three open areas located south of the city of Poznaƒ – in the Wielkopolska National Park (Plot A) and its vicinity (Plots B and C) - (W Poland, 52°17’N, 16°56E’, Fig. 1). The study area is typical for this region of Poland, dominated by farmland with a mosaic of fields, meadows, rough ground and wasteland. Some parts of Plot A are overgrown with young oak (Quercus sp.), beech (Fagus silvatica) and birch (Betula sp.) trees (< 10 years, 1-3 m high). The area between plots is generally unsuitable for corn buntings (woodlands, wetlands) or for recordings (fields along a major road). Territories of corn buntings are often placed linearly and at the geographical scale chosen for the study birds generally tend to occur in clusters. Therefore, the density estimations for study plots (A - 3 males ⋅ 10 km-2, B - 6 males ⋅ 10 km-2, C - 11 males ⋅ 10 km-2) are low and did not reflect real differences in distribution. These values much more depend on the size of the study area and how many aggregations were found within it. So we assigned males to one of two simple categories: (1) those with ≤ 1 neighbour, or (2) those with ≥ 2 neighbours within hearing range. Base on this assumption, we can indicate one high density population on Plot A, two on Plot B and one on Plot C (Fig. 1). There was also one crucial difference between study plots. Plot A was know to be relatively unsuitable for corn bunting, which generally avoid 276 Fig. 1. Map of the study area, showing Plots A, B, C and location of male corn bunting , Wielkopolska Region, Poland. Black arrows indicate area of high corn bunting density. A two examples of low and high density area are enlarged for detailed comparison of song types distribution. forest vicinity (K u ê n i a k 2000). Detailed census in the entire Wielkopolska NP in 1992 revealed occurrence of only five males, but subsequently the number increased to 16 in 1995 (B e d n o r z 1997). Birds from the only high density group on Plot A inhabited an area, which was overgrown by small trees and became suitable for them a few years ago (area accessible since 1991). More than five males was observed here firstly in 2000. On the other hand, local populations on Plot B and C were connected with stable agricultural or wasteland habitat, which remained unchanged for at least 20 years. These population were known to be rich and stable at least since 1996, when we started recording of other buntings in this area (O s i e j u k unpublished). Basic methods The study was conducted between March and June 2001. Birds were recorded in the morning (between 0500 and 1100) using a HHB PDR 1000 Professional DAT recorder with a Telinga 277 V Pro Science parabola, and a SONY TCD-D8 recorder with a Sennheiser ME 67 shotgun microphone. All recorded males were positioned on a map with the use of Garmin 12CX GPS with at least ± 10 m accuracy. We estimate that we recorded over 90% of males present in the field. For identification we also used recordings from playback experiments, which were conducted simultaneously and will be a subject of another paper. In sites of higher density of males, neighbours were recorded simultaneously by two observers, which also ensured that identification of different males was unequivocal. Although the birds were not marked individually, it was usually easy to identify males because each had a few permanent song posts within the territory. Sonogram analysis and bioacoustic terminology used All recordings were digitally transferred from the above-mentioned HHB DAT recorder via a SPDIF cable to a PC workstation with SoundBlaster Live! 5.1 (full version) using 48 kHz / 16 bit sampling and analysed using Avisoft SASLab Pro 3.9 software (S p e c h t 2002). All sonogram measurements were calculated using the following settings of SASLab: 1024 FFTlength, Frame [%] = 25, Window = Hamming and Temporal Overlap = 87.5%. This gave 244 Hz bandwidth with 42 Hz frequency and 2.9 ms time resolution. We used the same terminology as in previous studies of corn bunting vocalisation and tried to identify each strophe according to a level of its completeness (henceforth called variant), song type and finally a dialect (e.g. M c G r e g o r 1980, H o l l a n d et al. 1996 or L a t r u f f e et al. 2000). First, regardless of belonging to a particular song type and dialect, each song represented a full or incomplete strophe. Detailed comparison of sonograms revealed that predominantly males shorten strophes in fixed positions, which enabled categorization into different song lengths variants: (a) full song phrase, (b) song phrase lacking a portion of the final part, (c) song phrase lacking the final part and a portion of the middle part, and (d) only the initial part of song phrase (Fig. 2). Second, each male had a finite repertoire of strophes similar in structure, which were called song types (Fig. 3). Third, we hypothesized that neighbouring males (within so called local dialect) would share the same song repertoire types and that in their song types the back part of a strophe-end would be similar (M c G r e g o r 1980, M c G r e g o r & T h o m p s o n 1988, H o l l a n d et al. 1996). Song types that shared the same strophe end were assigned to the same dialect (Fig. 3). However, many song types identified, even within a repertoire of particular male, did not share any part of a strophe. Therefore, we were unable to assign such a song types to any dialect. Different song type strophes (regardless of their completeness) were named with a capital letter (e.g. A, B, etc.) or letters (AA, AB, AC, etc.). If two or more song types shared strophe ends, their were designated as belonging to the same dialect named by Arabic numerals (1, 2 and so on). The sonograms were categorized through a three-step process. First, the second author assigned song strophes to different types and variants within each type by visual inspection. At this stage, each new type and variant for each male was printed and described. In the second step, the first author checked the coherence of categorization at random examples of each song category for each male. In cases of any doubt, more examples where compared visually and with respect to the measured parameters: frequency range, frequency of maximal amplitude and cross-correlation coefficients. In the third step, song types sharing strophe-ends were grouped into dialects. 278 Fig. 2. Sonograms of examples of the four song variants (a, b, c, d) of song type A, in corn bunting of the Wielkopolska Region, Poland. The following sonograms illustrate the method of distinguishing complete and incomplete song versions. We also tested geographical pattern of song type sharing. Within the studied group of 86 males we randomly generated 43 pairs and repeated such drawing 10 000 times (M a n l y 1991). Then we tested the relation between male to male distance and the probability that randomly paired males had at least one common song type in their repertoires. Results A total of 145 recordings of 86 male corn buntings (Plot A – 15 males, Plot B – 49 males, Plot C – 22 males) were made, containing 4,149 song phrases and last over 10 h. Most males were recorded only once, but some males were recorded twice or more times, also in different contexts (e.g. solo- or counter-singing). 279 Repertoire composition The visual inspection of sonograms enabled us to distinguish 32 song types. The most common song types were: A (25 males; 29.1%), M (20 males; 23.3%), B (19 males; 22.1%), I (19 males; 22.1%), H (18 males; 21.0%) and AC (13 males; 15.1%). The next 10 popular song types were sung by 2 to 6 males (i.e. 2.3–7.0%) and the other 16 types were sung only by single males from the studied population. Altogether we found 36 combinations of one or more song types building individual repertoires. Only 4 of them were common: H + I (13 males), A + B (12 males), M (8 males) and M + AC (7 males), and another 4 occurred in 2-4 males: A (4 males), H (3 males), F + G and G (2 males). The remaining 28 repertoire combinations were specific to a single male. We were able to distinguish only four clear dialects: dialect 1 with five song types (A, B, C, D, Q); dialect 2 with two song types (S, T), dialect 3 with two song types (M, AC) and dialect 4 with two song types (H, I) (Fig. 3). In all cases, song types within dialect shared a large portion of the-end strophe. All of the other 21 song types were much more different from each other and did not share either of the initial or final parts of a strophe. We also found three examples of evident mixing of song types within a repertoire. Male 23, which sang types B and C from dialect 1, occasionally (3 times per 92 strophes Fig. 3. Sonograms of song types from dialects in corn bunting of the Wielkopolska Region, Poland: 1 (A, B), 2 (S, T), 3 (M, AC) and 4 (H, I). Joined rectangles indicate parts of song strophes, which are common for a particular dialects. Sonograms of strophes F, N, O and Z are examples of song types, which were not find to share final part with any other type. Sonograms marked as pla1 – pla4 are examples of plastic strophes sung successively by one male early in the season. 280 recorded) uttered phrases with mixed B–C initial part of song. Such mixed versions were always preceded by song type B and followed by song type C. The second example, male 153, was similar. Only once in the 64 strophes recorded for this individual was a mixed D-A song version recognized; thus still matched dialect 1 characteristics. In the third example, male 137 mixed song phrases that did not belong to one dialect, i.e. AC and AI, and this behaviour was sporadic (2 times per 130 strophes recorded). Two males sang a number of variable strophes, which differed slightly in consecutive performance (Fig. 3), in addition to songs that could be assigned to particular types. However, in both cases the recordings were made rather early in the season and the reason for this variation was probably an unfinished process of repertoire crystallization, rather than exceptional repertoire size. Therefore, we did not analyse these recordings in the context of repertoire size. Repertoire size Reliable data on individual repertoire size depends upon a sufficiently large sample of recorded phrases with which to estimate full repertoire size. The situation of the corn bunting is ambiguous, as we found that some males switched song types very rarely. In an extreme example, a male switched to the second song type after 120 phrases in a bout. If we consider only bouts of at least 50 song phrases, then the second song type appeared on average (± SE) after 24 ± 3.7 phrases (min-max: 2–121, n = 37). The third type appeared after 25 ± 5.5 phrases (min-max: 4–45, n = 8) and the fourth type after 55 ± 13.0 phrases (min-max: 39–81, n = 3). As we had recordings varying from 1 to 200 phrases in a bout, the possibility of underestimating repertoire size for some males was high. Therefore, we present data on repertoire size just as they are (Table 1), and note that the proportion of males having more than one song type in their repertoire is probably higher. On the other hand, there is no doubt that repertoires composed of two song types predominated in the studied population. Table 1. Song type repertoire size of corn bunting males from Wielkopolska. (*Including one male with two typical song types and one mixed song type and one male with two typical song types and several plastic song phrases; **Including two males singing three typical song types and one mixed song type; ***Including one individual singing 4 typical song types and several plastic song phrases). Repertoire size 1 2 3 4 Total n % 24 48* 10** 4*** 27.9 55.8 11.6 14.7 86 100.0 Microgeographical distribution of song types and dialects The distribution of males in the study area was not even (Fig. 1). In some locations, several males had territories close to each other, but in other locations territories were separated by hundreds of meters. We also found some males that were acoustically isolated from other individuals. The large gaps between aggregations of males are natural to a large extent, 281 because the gaps are dominated by forest stands, wetlands or urban areas, which are unsuitable for the corn bunting. On the other hand, even in some seemingly suitable habitats we found no singing males. In places of highest male density, we found a high withinpopulation homogeneity of repertoires. In Plot C, all males sang one or both song types from dialect 3 (i.e. M and AC) We found only one male singing song type M outside this area (ca. 10 km away). Song type AC was exclusive to this population (Fig. 1). In Plot B, we found two high-density populations separated by a distance of a few kilometers. In the northern plot, we found only males singing song types from dialect 4 (H, I), whereas the southern plot was dominated by males singing song types A and B from dialect 1. Along the road linking the two plots, we found males inhabiting roadside fields and rough lands but in a much lower density; this distribution resembled the shape of hourglass, with a few linearly distributed males linking the two plots. In the southern plot, we found a few males singing the northern dialect as well as males singing songs from both the northern and the southern dialect. A completely different situation was observed in areas where territories were distributed in small groups of one or a few males (northern part of Plot A and the southern part of Plot B). In such cases, the repertoire of the population was much more variable, but still neighbouring males shared some song types (Fig. 1). Males did not exhibit local dialect pattern on relatively recently inhabited area in SW part of the Plot A, where birds were local density was at similar level to that of Plots B and C. This distribution of song types reflects to some extent the mosaic pattern of local dialects. The randomization procedure showed that the probability of sharing song types decreases with growing distance between males (Fig. 4). However, boundaries between males singing different dialects were not always sharp. Fig. 4. Probability (mean ± SE) of sharing at least one song type (P) for randomly chosen pairs of male corn bunting in the Wielkopolska Region, Poland. The procedure of choosing 43 pairs was repeated 10 000 times. 282 Song versions Most males sang a few versions of each song type from their own repertoires, but full and slightly shortened versions predominated (Table 2). Males sang shorter versions more frequently when counter-singing (31.9% of strophes) than when singing solo (25.6%). To avoid pseudo-replication, we randomly chosen one recording per male. The differences in proportion of shortened strophes sung during solo and counter-singing context were not significant (Mann-Whitney U-test, Z = -0.043, n = 86, P = 0.966). We also tested differences in proportion of singing incomplete strophes sung by the same male recorded during solo (median = 0.17, 95% CI = 0.02 – 0.53) and counter-singing (median = 0.33, 95% CI = 0.09 – 0.75) context (Wilcoxon matched pairs test Z = -1.572, n = 8, P = 0.156). The last test indicate that males may shorten song strophes in aggressive context but we had to small sample size of “paired” recordings to reach significance level. The last method of testing seems to be the most valuable as it eliminates an influence of between males differences in song type repertoires. On the other hand, our results indicate also that shortening of strophes may be connected with a character of a particular song type. Among most popular song types (i.e. A, B, H, I, M and AC) for which we have a representative number of recordings (i.e. both recordings of different males and in different context), all types except AC where shortened in ca. 30% of cases. Song type AC was shortened in about 75% of all recorded performances and simultaneously had one of the longest complete variant, lasting over 3 seconds. We also found that there was a general positive relationship between strophe duration (complete song) and the frequency of shortening in particular song types (r = 0.49, n = 15, P = 0.066; for the most common 15 song types). Table 2. Length (sec) and frequency singing complete (a) and incomplete (b-d) variants of song strophes in corn bunting from Wielkopolska Region, Poland. Variant Mean SE n % a b c d 2.02 1.73 1.47 0.61 0.004 0.006 0.012 0.055 2884 953 301 11 69.5 23.0 7.2 0.3 Total 1.91 0.004 4149 100 Discussion Song type repertoire and local dialects If we consider only areas with high density of corn buntings, then the pattern of song type variation and the system of local dialects found in Wielkopolska corresponds in general with results obtained in Britain (M c G r e g o r 1980,1986, H o l l a n d et al. 1996, M c G r e g o r et al. 1997). However, when all studied sites in Wielkopolska are considered, song variation is higher in Poland than in the UK (M c G r e g o r 1980, 1986, M c G r e g o r et al. 1988). First, the UK populations studied by those authors were dominated by males singing exactly the same number and combination of song types, with the rare exception of mixed dialect singers. Although our study area was also dominated by males with 2 song 283 types in their repertoire, over 16% of individuals had 3 or 4 song types. Furthermore, many males inhabiting less preferred sites sang strophes typical for more than one dialect, or song types, which did not share final parts (i.e. could not be assigned to the same dialect).The term local song dialects is typically used to refer to clear microgeographic song variation with a mosaic of different dialects occurring within the species’ dispersal capacities (M c G r e g o r 1980). Our findings and those of some earlier studies indicate that local dialects may have slightly different character if viewed from a larger geographical perspective. For example in Portugal, song types were found to be restricted to sub-groups of males within local dialect rather than all males singing all song types as in Britain (L a t r u f f e et al. 2000). Whereas no local dialects have been reported for Azerbaydzhan (S u l t a n o v & G u m b a t o v a 1989). The local dialect pattern also seems to be a density-dependent phenomenon, in which formation and maintenance depend on habitat changes, duration and stability of local population. In Poland, clear local dialects were typical only for three high-density populations inhabiting Plots B and C (Fig. 1). Outside these highdensity areas, the pattern of song type distribution was much more random, and neighbouring males often did not share any song type. We also found that borders between dialects were not sharp under some conditions. For example, males were observed singing songs typical for one dialect in an area dominated by another dialect. Moreover, some males had song types of different dialects in their repertoires or they even mixed song types from two dialects in one song phrase. M c G r e g o r & T h o m p s o n (1988) reported similar results. Similar to H o l l a n d et al. (1996), we found males singing a song characteristic of a dialect population several km away from the centre of that population, which indicates that birds can disperse to recolonise. Furthermore, in this study we found an evident example of a disturbance of a local dialect pattern and this disturbance was found on less preferred or relatively new sites (B e d n o r z 1997, own unpubl. data). Thus, it might be an effect of, for example, offspring overproduction, increased winter survival, or other reasons that caused some males to fail to settle within the local dialect area. Such males try to recolonise or colonise less suitable areas, which is manifested in a higher song variation in such areas (multiple founder effect, see H o l l a n d et al. 1996). Formation of a new local dialect in such less suitable areas is rather unlikely. First, corn bunting males in less suitable areas have larger territories, often separated by unoccupied areas (C r a m p & P e r r i n s 1994, this paper), which due to reduced social interaction lower the chance of creating a common dialect. Second, in corn buntings, females tend to be mated to males singing the same local dialect as their father (M c G r e g o r et al. 1997), which should acts as a stabilizing factor for the existing local dialect population and at the same time negatively affects mating success on recolonised areas outside (see M c G r e g o r & T h o m p s o n 1988, H o l l a n d et al. 1996). Song variants The usage of full and shortened song versions gives ambiguous results. Generally, males sang shorter versions during counter-singing, but some song types were shortened more frequently regardless of the context. C o n s t a b l e (1989) and S h e p h e r d (1992) (both after C r a m p & P e r r i n s 1994) showed similar results, namely some song types in populations studied by them were more likely to be sung as incomplete versions. S h e p h e r d (1992) found that mated males sang more incomplete phrases of one song type from their two-type 284 repertoire. The pattern found in Wielkopolska indicates that longer complete versions of a song type are more likely to be abbreviated during song performance. It is difficult to determine now whether shortening songs should be considered as an independent level of within-song type variation, which may serve as, for example, an aggressive signal. A few other explanations are still possible, e.g. variant switching might be a signal of submissiveness, tendency to shortening may be culturally transmitted or even represent only a production error, and should be tested experimentally (S e a r c y & N o w i c k i 1999, S e a r c y et al. 2000). Acknowledgements We would like to thank J. P. C y g a n , A. S t a s z y ƒ s k a and J. M. R u t k o w s k a for their assistance in the field and two anonymous referees for their valuable comments. The financial support was provided by the Polish Committee for Scientific Research (KBN grant no. 6/P04C/038/17). L I T E R AT U R E ANDREW R. J. 1987: The aggressive and courtship behaviour of certain emberizines. Behaviour 10: 255–308. BEDNORZ J. 1997: Birds of the Wielkopolska National Park. Bogucki Wydawnictwo Naukowe, Poznaƒ (in Polish with English summary). BEECHER M. D., CAMPBELL S. E., BURT J. M., HILL C. E. & NORDBY J. C. 2000: Song-type matching between neighbouring song sparrows. Anim. Behav. 59: 21–27. CATCHPOLE C. K. & MCGREGOR P. K. 1985: Sexual selection, song complexity and plumage dimorphism in European Buntings of the genus Emberiza. Anim. Behav. 33: 1378–1379. CATCHPOLE C. K. & SLATER P. J. B. 1995: Bird song. Biological themes and variations. Cambridge University Press, Cambridge. CRAMP S. & PERRINS C. M. 1994: Handbook of the Birds of Europe, the Middle East and North Africa. Vol. IX Buntings and New World Warblers. Oxford University Press, Oxford: 323–338. CZIKELI H. 1982: Dichtebedingter Zusammenhang zwischen Gesangstyp und Ökotyp bei der Grauammer (Emberiza calandra)? J. Orn. 123: 95–100. GIL D. & GAHR M. 2002: The honesty of bird song: multiple constraints for multiple traits. TREE 17: 133–141. GRAPPUTO A., PILASTRO A., BAKER A. J. & MARIN G. 2001: Molecular evidence for phylogenetic relationships among buntings and American sparrows (Emberizidae). J. Avian Biol. 32: 95–101. HOLLAND J., MCGREGOR P. K. & ROWE C. L. 1996: Changes in microgeographic song variation of corn bunting Miliaria calandra. J. Avian Biol. 27: 47–55. KROODSMA D. E. 1996: Ecology of passerine song development. In: Kroodsma D. E. & Miller E. H. (eds), Ecology and Evolution of Acoustic Communication in Birds. Cornell University Press, Ithaca and London: 3–19. KROODSMA D. E. & BYERS B. E. 1991: The function(s) of bird song. Am. Zool. 31: 318–328. KUèNIAK S. 2000: Miliaria calandra (L., 1758) – corn bunting. In: Bednorz J. et al. (eds), Birds of Wielkopolska. Faunistic monography. Bogucki Wydawnictwo Naukowe, Poznaƒ: 569–570 (in Polish with English summary). LATRUFFE C., MCGREGOR P. K., TAVARES J. P. & MOTA P. G. 2000: Microgeographic variation in corn bunting (Miliaria calandra) song: quantitative and discrimination aspects. Behaviour 137: 1241–1255. MANLY B. F. J. 1991: Randomization and Monte Carlo methods in biology. Chapman and Hall, London. MCGREGOR P. K. 1980: Song dialects in the corn bunting (Emberiza calandra). Z. Tierpsychol. 54: 285–297. MCGREGOR P. K. 1986: Song types in the corn bunting Emberiza calandra: matching and discrimination. J. Orn. 127: 37–42. MCGREGOR P. K., HOLLAND J. & SHEPHERD M. 1997: The ecology of corn bunting Miliaria calandra song dialects and their potential use in conservation. UK Nat. Conserv. 13: 76–87. 285 MCGREGOR P. K. & THOMPSON D. B. A. 1988: Constancy and change in local dialects of the corn bunting. Ornis Scand. 19: 153–159. MCGREGOR P. K., WALFORD V. R. & HARPER D. G . C. 1988: Song inheritance and mating in a songbird with local dialects. Bioacoustics 1: 107–129. MØLLER A. P. 1983: Song activity and territory quality in the corn bunting Miliaria calandra; with comments on mate selection. Ornis Scand. 14: 81–89. OSIEJUK T. S., RATY¡SKA K., CYGAN J. P. & DALE S. 2003: Song structure and repertoire variation in ortolan bunting (Emberiza hortulana L.) from isolated Norwegian population. Ann. Zool. Fennici 40: 3–16. PODOS J. 1997: A performance constraint on the evolution of trilled vocalizations in a songbird family (Passeriformes: Emberizidae). Evolution 51: 537–551. SEARCY W. A. & NOWICKI S. 1999: Song types as fundamental units in vocal repertoires. Anim. Behav. 58: 37–44. SEARCY W. A., NOWICKI S. & HOGAN C. 2000: Song type variants and aggressive context. Behav. Ecol. Sociobiol. 48: 358–363. SPECHT R. 2002: Avisoft-SASLab Pro Sound Analysis and Synthesis Laboratory. A PC-software for MSWindows 95/98/ME/NT/2000/XP, 144. STODDARD P. K., BEECHER M. D., HORNING C.L. & CAMPBELL S. E. 1991: Recognition of individual neighbors by song in the song sparrow, a species with song repertoires. Behav. Ecol. Sociobiol. 29: 211–215. SULTANOV E. H. & GUMBATOVA S. E. 1989. The composition and variability of the song of the corn bunting (Emberiza calandra L.) in Azerbaijan. Byulleten Moskovskogo Obshchestva Ispytatelei Prirody Otdel Biologicheskii 94: 3–13 (in Russian with English summary). TRAINER J. M. 1983: Changes in song dialect distributions and microgeographic variation in song of whitecrowned sparrows (Zonotrichia leucophrys nuttalli). Auk 100: 568–582. VEHRENCAMP S. L. 2000: Handicap, Index, and Conventional Signal Elements of Bird Song. In: Espmark Y., Amundsen T. & Resenqvist G. (eds), Animal Signals: Signalling and Signal Design in Animal Communication. Tapir Academic Press, Trondheim: 277–300. 286 Folia Zool. – 52(3): 287–298 (2003) Spring migration of birds in relation to North Atlantic Oscillation Zdenûk HUBÁLEK Institute of Vertebrate Biology, Academy of Sciences of the Czech Republic, Kvûtná 8, CZ-60365 Brno, Czech Republic; e-mail: [email protected] Received 21 August 2002; Accepted 22 May 2003 A b s t r a c t. Long-term spring phenological instants of 57 migratory bird species, i.e. arrival in summer visitors and departure in winter visitors, were recorded in South Moravia (Czech Republic) from 1952 through 2001 and evaluated for annual correspondence with the North Atlantic Oscillation (NAO) weather system. The migration instants occurred significantly earlier following positive winter/spring NAO index values (causing periods warmer than normal in Europe) in a number of short-distance migrants with a European winter range (e.g., Alauda arvensis, Columba palumbus, Corvus frugilegus, Motacilla alba, Phoenicurus ochruros, Phylloscopus collybita, Serinus serinus, Sturnus vulgaris, Vanellus vanellus), whereas they did not correlate with NAO in most long-distance migrants having a sub-Saharan winter range (e.g., Acrocephalus spp., Anthus trivialis, Apus apus, Cuculus canorus, Delichon urbica, Ficedula albicollis, Hippolais icterina, Hirundo rustica, Jynx torquilla, Lanius collurio, Locustella spp., Muscicapa striata, Oriolus oriolus, Phylloscopus sibilatrix, Riparia riparia, Streptopelia turtur, Sylvia spp.). The winter/spring (especially February and March) NAO conditions thus affect the migration timing of short-distance migrants that winter in western or southern Europe, and could explain their earlier than normal arrival that had been observed in Europe since the 1980s. Key words: phenology, migration, weather, climate, temperature, NAO Introduction Recent climate warming is often regarded as the cause of the advanced spring arrival and egg-laying dates of migratory birds that has been recorded in Europe and North America since the 1980’s (e.g., F o r c h h a m m e r et al. 1998, W i n k l e r et al. 1999, B o t h & V i s s e r 2001, S o k o l o v 2001, Z a l a k e v i c i u s & Z a l a k e v i c i u t e 2001). An approach to test this hypothesis could be to evaluate whether the bird arrival data correlate with some global measure of the weather/climate change. It has been demonstrated that El Niƒo/Southern Oscillation (ENSO) system of ocean and atmosphere circulation in the tropical Pacific affects the biota in the Americas, Australia, southern Asia and eastern Africa either directly, through extreme precipitation or temperature, or indirectly, through ecosystem changes (J a c o b s et al. 1994, K a r l et al. 1995, S i l l e t t et al. 2000, N o t t et al. 2002, S t e n s e t h et al. 2002). Another major mode of the world climate variability is the North Atlantic Oscillation (NAO) that profoundly affects the weather in continental Europe and eastern North America (W a l l a c e & G u t z l e r 1981, B a r n s t o n & L i v e z e y 1987, H u r r e l l 1995, Y o o & D ’ O d o r i c o 2002). The NAO system can be expressed quantitatively using simple monthly, seasonal or annual indices that make a correlation analysis possible and relatively easy. A number of papers have reported a correspondence between the NAO fluctuation and plant phenology (P o s t & S t e n s e t h 1999, C h m i e l e w s k i & R o t z e r 2001, P o s t et al. 2001), cephalopod migration (S i m s et al. 2001), marine fish migration (A l h e i t & H a g e n 1997, R e i d et al. 2001), 287 mammalian population dynamics (P o s t & S t e n s e t h 1999, T k a d l e c 2000, P o s t & F o r c h h a m m e r 2002, S t e n s e t h et al. 2002) or the avian breeding phenology and productivity (F o r c h h a m m e r et al. 1998, P r z y b y l o et al. 2000, S a e t h e r et al. 2000, B o t h & V i s s e r 2001, S o k o l o v 2001, M ø l l e r 2002, N o t t et al. 2002, S a n z 2002). However, the effect of NAO on the variability of bird migration instants in Europe has not been studied until very recently (F o r c h h a m m e r et al. 2002, J o n z é n et al. 2002, T r y j a n o w s k i et al. 2002, H ü p p o p & H ü p p o p 2003). Material and Methods Study area Observations of birds were carried out in the Bfieclav area (48°40’–48°50’N, 16°30’–17°00’E) of South Moravia, Czech Republic. The relief is a mostly flat to slightly undulating lowland (150 to 200 m a.s.l.), formed largely by accumulations of Pleistocene loess and riverine sand deposits on Pliocene sedimentaries. The climate is relatively dry and warm (annual precipitation 550 mm; annual mean air temperature 9.5 °C ( January -1.8 °C, July 19.1 °C). There are agroecosystems (mostly arable fields), forest ecosystems (floodplain forests, subxerophilic mixed deciduous oak stands), fishponds and big water reservoirs (Nové Ml˘ny) in the wider area. Local community is composed of about 150 species of breeding terrestrial and water birds (H á j e k 1992, · È a s t n ˘ et al. 1996, H u b á l e k 1997, Z u n a - K r a t k y et al. 2000), and many migrants use the area for stopovers. Bird phenology data Longitudinal records of migratory birds were documented from 1969 through 2001. Spring phenological instants recorded for each species involved the first occurrence (spring arrival) in summer visitors, and the last occurrence (spring departure) in winter visitors. In addition to the author’s observations, data of H á j e k (1992) were used for the years 1952 to 1969. Total length of the period thus covered 50 years (1952–2001), and 57 common bird species with a distinct pattern of spring phenology and a sufficient number of annual records were selected for the correlation and regression analysis. North Atlantic Oscillation data Monthly (January; February; March; April) and seasonal (December to March, DJFM; January to March, JFM; February to April, FMA) NAO indices for particular years 1952 through 2001 were extracted from the Internet at http://www.cgd.ucar.edu/~jhurrell/nao.html (H u r r e l l 1995: NAOH) and http://www.cpc.ncep.noaa.gov/data/teledoc/nao.html (B a r n s t o n & L i v e z e y 1987: NAOBL; W a l l a c e & G u t z l e r 1981: NAOEA, the East Atlantic Pattern). The indices are based on normalized sea-level pressure differences between two points or areas: (1) Ponta Delgada (Azores) and Stykkisholmur/Reykjavik (Iceland) in the eastern sector of North Atlantic (NAOH); (2) the central latitudes (35°N–40°N) and Greenland (NAOBL, reflecting thus spatially a more broad-scale atmospheric pressure pattern in the western North Atlantic sector); and (3) NAOEA covers the eastern North Atlantic sector. The positive NAO index generally shows that the atmospheric pressure over the subtropical part of the North Atlantic is higher than normal while that over 288 the northern sector of the North Atlantic is lower than normal; this increased pressure difference between the two sectors results in more and stronger storms crossing the Atlantic Ocean and, in turn, causing warm and wet weather (especially in winter) in northern and central Europe. The negative NAO index reflects an opposite pattern of height and pressure anomalies over these sectors; this reduced pressure gradient results in fewer and weaker storms crossing the Ocean, bringing cold air to northern (and central) Europe and moist, often cold air into the Mediterranean (H u r r e l l 1995). Statistical analysis Calendar data of phenological instants were transformed into sequential numbers (e.g., 1 for 1st January, 32 for 1st February, 60 for 1st March, 91 for 1st April, 121 for 1st May 1; in leapyears, the sequential numbers were corrected by adding 1, starting from 1st March). Linear correlation and regression models were then used to examine relationships between avian spring phenological instants of particular bird species and the three NAO indices (NAOH, NAOBL, NAOEA); Pearson’s correlation coefficient values were calculated and statistical test done for all comparisons using SOLO 4.0 (BMDP Statistical Software, Los Angeles, CA). Results The late-winter and early-spring NAO weather system conditions affected significantly the spring arrival (in summer visitors) or departure (in winter visitors) instants of some bird species (Tables 1 and 2, Fig. 1). The species significantly inversely correlated with both the ‘JFM’ and ‘FMA’ seasonal NAO indices (i.e., the phenological instants occurred earlier following the January-February-March and February-March-April seasonal air pressure difference in the North Atlantic higher than normal) were Alauda arvensis, Anas crecca, Aythya ferina, Bucephala clangula, Columba palumbus, Corvus frugilegus, Gallinago gallinago, Mergus merganser, Motacilla alba, Phoenicurus ochruros, Phylloscopus collybita, P. trochilus, Pyrrhula pyrrhula, Regulus regulus, Remiz pendulinus, Saxicola torquata, Serinus serinus, Sterna hirundo, Sturnus vulgaris, Sylvia atricapilla, Tringa glareola, Turdus philomelos and Vanellus vanellus. In some species, there was a single statistically significant seasonal (either ‘JFM’ or ‘FMA’) NAO index (Acrocephalus scirpaceus, Anas querquedula, Charadrius dubius, Hirundo rustica, Luscinia megarhynchos, Phylloscopus sibilatrix, Streptopelia turtur, Sylvia communis, Tringa nebularia, T. totanus), whereas the remaining species did not correlate with NAO at p<0.05: Acrocephalus arundinaceus, A. palustris, A. schoenobaenus, Actitis hypoleucos, Anthus trivialis, Apus apus, Cuculus canorus, Delichon urbica, Ficedula albicollis, Hippolais icterina, Jynx torquilla, Lanius collurio, Locustella fluviatilis, L. luscinioides, L. naevia, Motacilla flava, Muscicapa striata, Nycticorax nycticorax, Oriolus oriolus, Riparia riparia, Sylvia borin, S. curruca, S. nisoria and Upupa epops. The winter NAOH index (DJFM data, not given in Table 1) correlated significantly (p<0.01) with phenological instants of Columba palumbus (-0.45), Mergus merganser (-0.68), Phylloscopus collybita (-0.49), Sturnus vulgaris (-0.44), Tringa nebularia (-0.51) and Vanellus vanellus (-0.40). According to monthly NAO indices (data not given in Table 1), the number of avian species significantly (p<0.05) correlated in their phenology with NAOH/NAOEA during January, February, March and April were 8/8, 12/12, 9/13 and 3/3, respectively. Therefore, the NAO conditions especially in February and March have 289 Table 1. Spring phenology instants of 57 migratory bird species in South Moravia, 1952–2001, with the number of years (n), the mean date of arrival or departure (mm/dd), principal wintering area (W), and Pearson’s correlation coefficient value (r) between the avian phenology data and the mean seasonal indices (JFM, January to March; FMA, February to April) of NAOH (H), NAOBL (BL) and NAOEA (EA); r values in bold are significant (p<0.05). The bird species are arranged according to the mean arrival/departure date. JFM Bird species n Mean date Alauda arvensis A Sturnus vulgaris A Mergus merganser D Turdus philomelos A Corvus frugilegus D1/2 Motacilla alba A Vanellus vanellus A Aythya ferina A Anas crecca A Pyrrhula pyrrhula D Bucephala clangula D Columba palumbus A Phylloscopus collybita A Phoenicurus ochruros A Saxicola torquata A Regulus regulus D Anas querquedula A Gallinago gallinago A Serinus serinus A Remiz pendulinus A Tringa totanus A Sylvia atricapilla A Phylloscopus trochilus A Hirundo rustica A Charadrius dubius A Anthus trivialis A Motacilla flava A Jynx torquilla A Nycticorax nycticorax A Upupa epops A Sylvia curruca A Ficedula albicollis A Acrocephalus schoenobaenus A Actitis hypoleucos A Delichon urbica A Phylloscopus sibilatrix A Cuculus canorus A Luscinia megarhynchos A Tringa nebularia A Locustella luscinioides A Streptopelia turtur A Sylvia communis A Riparia riparia A Tringa glareola A Sterna hirundo A 49 50 18 39 35 47 47 24 23 27 30 38 36 38 35 23 29 26 38 37 27 33 37 46 33 45 35 42 18 38 44 45 32 41 40 30 50 39 29 28 34 44 23 31 29 02/28 03/01 03/05 03/08 03/09 03/10 03/11 03/12 03/15 03/19 03/20 03/23 03/23 03/24 03/24 03/26 03/26 03/26 03/30 03/30 04/01 04/03 04/06 04/09 04/10 04/12 04/12 04/13 04/13 04/14 04/15 04/16 04/19 04/20 04/21 04/21 04/22 04/23 04/23 04/24 04/24 04/24 04/25 04/26 04/27 290 W E E E E E E E E E E E E E E E E A E E E E E A A E A A A A A A A A A A A A A E A A A A E E H FMA BL EA H -.47 -.43 -.47 -.33 -.62 -.42 -.39 -.50 -.52 -.39 -.52 -.38 -.40 -.54 -.02 -.26 -.36 -.29 -.23 -.58 -.33 -.53 -.44 -.03 -.07 +.13 -.12 -.12 -.22 -.31 -.14 -.03 -.07 -.13 +.06 -.36 -.14 +.02 -.56 -.06 -.39 -.12 -.28 -.21 -.01 -.35 -.53 -.59 -.06 -.55 -.31 -.42 -.42 -.42 -.18 -.26 -.58 -.54 -.34 -.47 -.49 -.19 -.43 -.34 -.49 -.24 -.10 -.20 -.20 -.15 +.11 -.19 -.15 -.21 -.13 -.05 +.15 -.17 -.08 +.29 +.14 -.09 -.26 -.27 -.30 -.08 -.26 +.08 -.45 -.53 -.16 -.09 -.51 -.44 -.61 -.60 +.06 +.25 -.52 -.41 -.11 -.08 -.34 -.28 -.46 -.45 -.42 -.51 -.14 -.15 -.10 -.20 -.55 -.55 -.60 -.52 -.38 -.33 -.46 -.40 -.44 -.55 -.28 -.30 -.56 -.48 -.44 -.36 -.38 -.26 -.46 -.39 -.12 -.15 -.18 -.19 -.17 -.26 -.48 -.38 +.18 +.08 -.30 -.15 -.08 -.16 -.34 -.23 -.21 -.12 -.12 -.03 +.10 +.17 -.26 -.13 -.00 -.02 +.24 +.23 +.00 -.06 -.13 -.21 -.36 -.31 -.45 -.42 -.16 -.23 -.22 -.25 -.33 -.33 -.13 +.13 -.39 -.33 -.43 -.52 BL EA -.19 -.45 -.54 +.15 -.38 -.22 -.27 -.40 -.51 -.27 -.21 -.56 -.44 -.32 -.46 -.62 -.22 -.30 -.29 -.39 -.27 -.24 -.25 -.34 -.25 +.07 -.24 -.07 -.21 -.14 +.09 +.13 -.25 -.17 +.13 +.01 -.23 -.23 -.30 -.37 -.18 -.21 +.12 -.33 -.36 -.31 -.28 -.33 -.44 -.54 -.45 -.29 -.42 -.34 -.52 -.57 -.19 -.20 -.42 +.14 -.21 -.23 +.04 -.06 -.41 -.13 -.56 -.42 -.01 +.11 -.11 +.13 -.29 -.15 -.27 -.03 -.08 +.03 -.26 +.17 -.41 -.10 +.24 -.23 -.14 -.23 -.04 -.35 +.03 +.01 Table 1 – continued. Sylvia borin A Apus apus A Acrocephalus arundinaceus A Acrocephalus scirpaceus A Locustella naevia A Oriolus oriolus A Hippolais icterina A Locustella fluviatilis A Muscicapa striata A Sylvia nisoria A Lanius collurio A Acrocephalus palustris A 37 37 31 22 26 42 44 29 33 41 31 24 04/29 04/29 04/30 05/02 05/04 05/05 05/07 05/07 05/07 05/07 05/09 05/10 A A A A A A A A A A A A -.26 -.17 +.24 -.34 -.12 +.12 -.04 +.23 +.08 -.05 +.01 +.05 -.29 -.20 +.20 -.44 -.21 +.15 -.07 +.26 -.07 +.05 -.16 -.10 -.11 -.25 -.02 -.20 -.03 -.08 +.15 -.03 -.03 -.12 -.06 -.03 -.19 -.12 +.10 -.45 +.03 +.13 +.07 +.16 +.06 -.04 -.15 -.38 -.20 -.01 +.03 -.45 -.04 +.17 +.08 +.25 -.02 -.02 -.22 -.38 -.04 -.27 -.12 -.06 +.01 -.15 +.14 -.13 -.12 -.10 -.10 -.24 A arrival of the first bird; D departure (the last observed bird in the spring); D1/2 departure of about half of the wintering population. Principal wintering area (W): E, southern, western or central Europe, Mediterranean; A, sub-Saharan Africa (or southwestern Asia). Table 2. Regression (y = a + b.x) of 18 selected bird species on seasonal NAO index (see Fig. 1). NAO index Bird species W Type Alauda arvensis Sturnus vulgaris Corvus frugilegus Columba palumbus Phoenicurus ochruros Remiz pendulinus Phylloscopus collybita Phylloscopus sibilatrix Sylvia atricapilla Sylvia curruca Charadrius dubius Actitis hypoleucos Hirundo rustica Delichon urbica Ficedula albicollis Muscicapa striata Hippolais icterina Cuculus canorus E E E E E E E A E A E A A A A A A A EA H H H EA EA H H EA EA H H EA EA H H H H Season Regression slope (b) Intercept (a) R2 JFM JFM JFM JFM JFM JFM JFM JFM JFM JFM JFM JFM JFM JFM JFM JFM FMA FMA -8.5757*** -2.4273*** -1.4594** -2.8513*** -5.7958*** -8.5900*** -2.2751*** +0.0008 -4.9794** -1.1621 -2.4511** -0.0652 -0.7587 +0.2758 +0.2707 +0.1748 +0.2002 -0.2552 58.71 60.26 69.87 83.05 83.44 89.97 84.23 111.31 93.40 105.36 100.26 110.08 99.47 111.03 105.87 127.38 127.19 112.19 0.216 0.261 0.275 0.308 0.306 0.338 0.356 0.000 0.282 0.018 0.231 0.000 0.006 0.001 0.011 0.006 0.004 0.009 W, principal wintering area. NAO index type: EA, NAOEA; H, NAOH ). Other abbrevations as for Table 1. **, significant at p<0.01; ***, significant at p<0.001. a profound effect on many early-spring migrants that usually winter in western or southern Europe (Mediterranean). On the other hand, NAO does not seem to affect significantly the timing of long-distance migrants arriving in Central Europe later, having wintering grounds largely in sub-Saharan Africa. The seasonal (JFM) NAOH index explained on the average 18% (range 1–37%) variability in the phenological instants of the 25 short-distance migrants, whereas only 4% (range, 0–13%) variability in the 32 long-distance migrants; the difference between the two categories of migratory birds is highly significant (p<0.001) as found with either t-test or nonparametric Mann-Whitney two-sample test. 291 ALA.ARV. STU.VUL. COR.FRU. COL.PAL. PHO.OCH. REM.PEN. Fig. 1. Scatter plot diagrams of relations between seasonal (JFM or FMA) NAO indices (NAOH or NAOEA) and spring migration instants in 18 selected bird species (acronymes) in South Moravia, 1952–2001. Y-axis shows calendar data of the phenological instants transformed into sequential numbers. For the regression analysis, see Table 2. 292 Fig. 1 – continued. 293 ACT.HYP. CHA.DUB. SYL.CUR. SYL.ATR. PHY.SIB. PHY.COL. Fig. 1 – continued. 294 HIR.RUS. DEL.URB. FIC.ALB. MUS.STR. HIP.ICT. CUC.CAN. Discussion T r y j a n o w s k i & S p a r k s (2001) suggested that the relationship between arrival date and population size should be considered in particular bird species: an increasing population of Lanius collurio showed an advanced arrival in western Poland during 1983–2000. In the present study, usually common and abundant avian species were taken into account whose population sizes did not change considerably with time. The exceptions with the population steadily low (l), on decline (d) or increasing (i) were Acrocephalus spp. (d), Anas spp. (d), Ficedula albicollis (i), Gallinago gallinago (d), Lanius collurio (i), Motacilla flava (d), Nycticorax nycticorax (l), Sylvia nisoria (l), Tringa nebularia (l), T. totanus (d), Upupa epops (d) and Vanellus vanellus (d) (H á j e k 1992, · È a s t n ˘ et al. 1996, H u b á l e k 1997, etc.). However, the results yielded in these species do not seem to contradict the general conclusion that long-distance migrants are not affected significantly with the NAO activity, in contrast to short-distance migratory species. The positive NAO index values mean generally a stronger than usual subtropical high pressure center and a deeper than usual subarctic atmospheric low. The increased pressure difference causes intense atmospheric mass circulation resulting in more winter storms crossing the Atlantic Ocean on a more northerly track, a milder winter and the whole year in Europe, and specifically a warmer (but drier) than normal weather over central Europe (H u r r e l l 1995). For the Czech Republic, a significant correlation was found between the winter (DJFM) NAOH index and local winter air temperature (r = 0.78), but that between this index and winter precipitation (r = -0.30) was insignificant (T k a d l e c 2000). The milder temperatures associated with the positive NAO phase lead to higher invertebrate loads in spring which might benefit short-distance migrants more than long-distance migratory species (N o t t et al. 2002). Three indices that characterize NAO were used in this study. Although NAOH and NAOBL correlated well (r value for JFM season was 0.87), NAO BL seemed to be less sensitive in the phenology of European migratory birds (in terms of the number of significant correlations with avian species) than NAO H, probably because it describes weather conditions in the western sector of the North Atlantic and may thus be more relevant for the North American continent than for Europe. The eastern Atlantic measures NAOH and NAOEA were comparable in their phenological sensitivity although they did not correlate well (r value for JFM season was very low, 0.17): in contrast to NAOH, NAOEA revealed a marked correspondence between NAO activity and phenological instants in Alauda arvensis, Bucephala clangula, Motacilla alba, Pyrrhula pyrrhula, Sylvia atricapilla and Turdus philomelos, whereas NAOH, contrary to NAOEA, yielded significant correlations with phenological data of Charadrius dubius, Gallinago gallinago, Larus ridibundus, Regulus regulus, Saxicola torquata, Serinus serinus, Sterna hirundo, Tringa totanus and T. glareola. At present, it is difficult to evaluate which of the two indices is better for the use in the European phenology; they might be used as complementary measures. In an earlier study, a cluster analysis of temporal spring migration patterns of birds in Moravia between 1881 and 1960 revealed several groups (called ‘migrons’) of co-related avian species: the first (‘Mediterranean’) migron consisted almost exclusively of shortdistance migrants wintering in southern or western Europe (Alauda arvensis, Sturnus vulgaris, Fringilla coelebs, Columba palumbus, Motacilla alba, Vanellus vanellus, Turdus philomelos, Erithacus rubecula, Scolopax rusticola), whereas the remaining migrons had been called ‘African’ in that they involved long-distance migrants Hirundo rustica, Jynx 295 torquilla, Delichon urbica, Cuculus canorus, Luscinia megarhynchos, Streptopelia turtur, Apus apus, Oriolus oriolus and Coturnix coturnix, i.e. species having their winter quarters in sub-Saharan Africa (H u b á l e k 1985). In the present study, the seasonal winter/spring NAO index did not correlate significantly with the arrival of a vast majority of long-distance migratory bird species wintering in tropical and southern Africa (only 2 of 32 those species tested revealed significant both ‘JFM’ and ‘FMA’ seasonal measures). Timing of their departure from the winter grounds must therefore be based on principles other than the weather system fluctuation at northern latitudes. However, significant inverse relationship was found between the arrival of nearly all (22) short-distance migrants wintering in western and southern Europe (25 species tested) and the seasonal winter/spring NAO index, indicating that a higher than normal air pressure difference over the North Atlantic during the winter/spring (especially in February and March) determines an earlier than normal arrival of these birds in Central Europe. This result is not in accord with a very recently published study of F o r c h h a m m e r et al. (2002), who found no significant difference in the effect of NAO on spring arrival between long-distance and short-distance migrants (but only three species of each group were tested) breeding in Norway, although it corresponds well with the results of N o t t et al. (2002). Positive values of the winter NAO index prevailed in 11 out of 15 years between 1987 and 2001 (H u r r e l l 1995), as did the bird phenological data in South Moravia during the same period: earlier than normal early-spring bird instants occurred in 12 out of 15 years, whereas those markedly later than normal in one year only (H u b á l e k 1997, H u d e c et al. 1999, unpublished data for 1998–2001). The present study revealed that the NAO weather system affects spring phenological instants in a number of bird species and that this effect could explain the earlier than normal arrival of common migratory species (e.g., Alauda arvensis, Sturnus vulgaris, Vanellus vanellus, Columba palumbus, Motacilla alba, Phoenicurus ochruros, Phylloscopus collybita and Serinus serinus) that has been observed in Central Europe since the 1980s. When we accept that the NAO system is an indicator (or even the proximate causative mechanism) of the climate change, then the earlier than normal spring arrival of migratory birds should be considered as both a result of, and evidence for, the climate warming in the last two decades. Acknowledgements I am grateful to Emil T k a d l e c for advice about the NAO index, and to M. Philip N o t t for valuable suggestions on the manuscript. L I T E R AT U R E ALHEIT J. & HAGEN E. 1997: Long-term climate forcing of European herring and sardine populations. Fish. Oceanogr. 6: 130–139. BARNSTON A. G. & LIVEZEY R. E. 1987: Classification, seasonality and persistence of low-frequency atmospheric circulation patterns. Monthly Weath. Rev. 115: 1083–1126. BOTH C. & VISSER M. E. 2001: Adjustment to climate change is constrained by arrival date in a long-distance migrant bird. Nature 411: 296–298. CHMIELEWSKI F. M. & ROTZER T. 2001: Response of tree phenology to climate change across Europe. Agricul. Forest Meteorol. 108: 101–112. 296 FORCHHAMMER M. C., POST E. & STENSETH N. C. 1998: Breeding phenology and climate. Nature 391: 29–30. FORCHHAMMER M. C., POST E. & STENSETH N. C. 2002: North Atlantic Oscillation timing of long- and short-distance migration. J. Anim. Ecol. 71: 1002–1014. HÁJEK V. 1992: [Bird fauna of South Moravia, 1952–1991]. Manuscript, 324 pp. (in Czech). HUBÁLEK Z. 1985: Cluster analysis of the spring migration of birds in Moravia (Czechoslovakia), 1881–1960. Vûstník âeskosl. Spol. Zool. 49: 81–86. HUBÁLEK Z. 1997: Trends of bird populations in a managed lowland riverine ecosystem. Folia Zool. 46: 289–302. HUDEC K., HUBÁLEK Z. & VAâKA¤ J. 1999: [Spring arrival of birds in the South Moravian region in 1997]. Zprávy MOS, 57: 117–128 (in Czech, with English summary). HÜPPOP O. & HÜPPOP K. 2003: North Atlantic Oscillation and timing of spring migration in birds. Proc. Roy. Soc. Lond., B 270: 233–240. HURRELL J. W. 1995: Decadal trends in the North Atlantic Oscillation: regional temperatures and precipitation. Science, 269: 676–679. JACOBS G. A., HURLBURT H. E., KINDLE J. C., METZGER E. J., MITCHELL J. L., TEAGUE W. E. & WALLCRAFT A. J. 1994: Decade-scale trans-Pacific propagation and warming effects of an El NiÀo anomaly. Nature 370: 360–363. JONZÉN N., HEDENSTRÖM A., HJORT C., LINDSTRÖM Å., LUNDBERG P. & ANDERSSON A. 2002: Climate patterns and the stochastic dynamics of migratory birds. Oikos 97: 329–336. KARL D. M., LETELIER R., HEBEL D., TUPAS L., DORE J., CHRISTIAN J. & WINN C. 1995: Ecosystem changes in the North Pacific subtropical gyre attributed to the 1991–92 El NiÀo. Nature 373: 230–234. MØLLER A. P. 2002: North Atlantic Oscillation (NAO) effects of climate on the relative importance of first and second clutches in a migratory passerine bird. J. Anim. Ecol. 71: 201–210. NOTT M. P., DESANTE D. F., SIEGEL R. B. & PYLE P. 2002: Influences of the El Niƒo/Southern Oscillation and the North Atlantic Oscillation on avian productivity in forests of the Pacific Northwest of North America. Global Ecol. Biogeogr. 11: 333–342. POST E. & FORCHHAMMER M. C. 2002: Synchronization of animal population dynamics by large-scale climate. Nature 420: 168–171. POST E., FORCHHAMMER M. C., STENSETH N. C. & CALLAGHAN T.V. 2001: The timing of life-history events in a changing climate. Proc. Roy. Soc. Lond., B 268: 15–23. POST E. & STENSETH N. C. 1999: Climatic variability, plant phenology, and northern ungulates. Ecology 80: 1322–1339. PRZYBYLO R., SHELDON B. C. & MERILA J. 2000: Climatic effects on breeding and morphology: evidence for phenotypic plasticity. J. Anim. Ecol. 69: 395–403. REID P. C., BORGESS M. D. & SVENDSEN E. 2001: A regime shift in the North Sea circa 1988 linked to changes in the North Sea horse mackerel fishery. Fish. Res. 50: 163–171. SAETHER B. E., TUFTO J., ENGEN S., JERSTAD K., ROSTAD O. W. & SKATAN J. E. 2000: Population dynamical consequences of climate change for a small temperate songbird. Science 287: 854–856. SANZ J. J. 2002: Climate change and breeding parameters of great and blue tits throughout the western Palaearctic. Global Change Biol. 8: 409–422. SILLETT T. S., HOLMES R. T. & SHERRY T. W. 2000: Impacts of a global climate cycle on population dynamics of a migratory songbird. Science 288: 2040–2042. SIMS D. W., GENNER M. J., SOUTHWARD A. J. & HAWKINS S. J. 2001: Timing of squid migration reflects North Atlantic climate variability. Proc. Roy. Soc. Lond., B 268: 3607–3611. SOKOLOV L. V. 2001: Spring ambient temperature as an important factor controlling timing of arrival, breeding, post-fledging dispersal and breeding success of Pied Flycatcher Ficedula hypoleuca in Eastern Baltic. Avian Ecol. Behav. 5: 79–104. ·ËASTN¯ K, BEJâEK V. & HUDEC K. 1996: [Atlas of Breeding Birds in the Czech Republic, 1985–1989]. H&H, Jinoãany (in Czech). STENSETH N. C., MYSTERUD A., OTTERSEN G., HURRELL J. W., CHAN K.-S. & LIMA M. 2002: Ecological effects of climate fluctuations. Science 297: 1292–1296. TKADLEC E. 2000: Brown hare population dynamics in the Prostûjov district. Pfiírodovûdné Studie Muzea Prostûjovska 3: 133–149 (in Czech with English summary). TRYJANOWSKI P., KUZNIAK S. & SPARKS T. 2002: Earlier arrival of some farmland migrants in western Poland. Ibis 144: 62–68. 297 TRYJANOWSKI P. & SPARKS T. H. 2001: Is the detection of the first arrival date of migrating birds influenced by population size? A case study of the Red-Backed Shrike Lanius collurio. Internat. J. Biometeorol. 45: 217–219. WALLACE J. M. & GUTZLER D. S. 1981: Teleconnections in the geopotential height field during the Northern Hemisphere winter. Monthly Weath. Rev. 109: 784–812. WINKLER D. W., DUNN P. O. & McCULLOCH C. E. 2002: Predicting the effects of climate change on avian life-history traits. Proc. Nat. Acad. Sci. 99: 13595–13599. YOO J. C. & D’ODORICO P. 2002: Trends and fluctuations in the dates of ice break-up of lakes and rivers in Northern Europe: the effect of the North Atlantic Oscillation. J. Hydrol. 268: 100–112. ZALAKEVICIUS M. & ZALAKEVICIUTE R. 2001: Global climate change impact on birds: a review of research in Lithuania. Folia Zool. 50: 1–17. ZUNA-KRATKY T., KALIVODOVÁ E., KÜRTHY A., HORAL D. & HORÁK P. 2000: Die Vögel der MarchThaya-Auen im österreichisch-slowakisch-tschechischen Raum. Distelverein, Deutsch Wagram. 298 Folia Zool. – 52(3): 299–308 (2003) Postmating sexual selection in house sparrows: can females estimate “good fathers” according to their early paternal effort? Herbert HOI1*, Radovan VÁCLAV2 and Du‰ana SLOBODOVÁ3 1 Konrad Lorenz Institut für vergleichende Verhaltensforschung, Österreichische Akademie der Wissenschaften, Savoyenstraße 1a, A-1160 Vienna, Austria; e-mail: [email protected] 2 Institute of Zoology, Slovak Academy of Sciences, Dúbravská cesta 9, SK-842 06 Bratislava, Slovakia; e-mail: [email protected] 3 Department of Animal Physiology and Ethology, Comenius University, Mlynská dolina, SK-842 15 Bratislava, Slovakia Received 30 September 2002; Accepted 6 March 2003 A b s t r a c t . Recently, several studies suggested that in species with biparental care, male parental effort (for instance in terms of nest building) serves as a sexual signal to females. In this study on free-living house sparrows Passer domesticus, we investigated how male contribution to parental care varies between breeding stages and whether early parental care of a male reflects his paternal effort in later stages. We found that nest building was performed predominantly by males. However, hatching success was not related to male participation in nest building or early nest guarding. The contribution of males to incubation and chick brooding was lower than in females. Investment in chick feeding did not differ between partners, but varied considerably between males. Only the male effort in chick feeding was related to the number of young at fledging age, suggesting that male assistance is essential to maximise reproductive success in house sparrows. Except for the positive correlation between male nest building and male incubation during egg laying, we found no relationship between early (nest building and nest guarding) and later paternal effort (late incubation, brooding and provisioning rates). Consequently, intensity of male nest building and early nest guarding do not seem to be honest indicators of later paternal effort in the house sparrow. Instead, we speculate that high early paternal effort might be a strategy of some males to manipulate the reproductive effort of their partners. Key words: paternal care, nest building, sexual selection, house sparrows, breeding success Introduction One important mechanism of sexual selection, female mate choice, operates through females preferring males that display certain morphological and behavioural traits (Z a h a v i 1975, W e s t - E b e r h a r d 1979, C a t c h p o l e 1980, H a m i l t o n & Z u k 1982, N u r & H a s s o n 1984, R y a n et al. 1990). Females are supposed to benefit from their choice indirectly by acquiring “good genes” for their offspring (Z a h a v i 1975, W e a t h e r h e a d & R o b e r t s o n 1979, P o m i a n k o w s k i 1987, I w a s a et al. 1991), or directly by a greater assistance of their mates in chick rearing or territory defence (T r i v e r s 1972, B o r g i a 1979, B u r l e y 1981, K i r k p a t r i c k 1985, H o e z l e r 1989, P r i c e et al. 1993). However, sexual selection can operate also after mate choice (see S o l e r et al. 1998a,b). Hereby, some males can produce more or higher quality offspring as their mates invest in reproduction differentially depending on male physical (e.g. B u r l e y 1988, M ø l l e r 1994) or parental quality (G i b b o n s 1987, d e L o p e & M ø l l e r 1993). The latter case should especially account for the species with biparental care, for which male * Corresponding author 299 parental quality is often crucial to maximise female reproductive success (e.g. M o r e n o et al. 1999, see M ø l l e r 2000). In order to stimulate their mates to a higher reproductive effort, males should therefore try to advertise their ability and/or willingness to parental care. Females should in turn pay attention to the reliability of such advertisements. In fact, some studies showed that male song display can be an indicator about his parental quality (e.g. H o i - L e i t n e r et al. 1995, W e l l i n g et al. 1997). Other studies recently demonstrated that the number of advertised nests, nest size or male nest building activity in wrens Troglodytes troglodytes, black wheatears Oenanthe leucura, barn swallows Hirundo rustica and magpies Pica pica might represent a sexual signal involved in postmating sexual selection (E v a n s & B u r n 1996, S o l e r et al. 1996, S o l e r et al. 1998a, S o l e r et al. 2001). However, studies examining the reliability of early paternal advertisement and its potential to manipulate female reproductive decisions are lacking. The house sparrow Passer domesticus is a suitable species to study postmating sexual selection with regard to paternal quality because: (i) there is no clear female preference for males with larger sexual ornament, a melanin based throat patch (K i m b a l l 1996), (ii) early reproductive decisions of females (date of egg laying, clutch size) were showed to be related to early male parental effort (V á c l a v & H o i 2002), whereas (iii) chick feeding as well as fledging success were reported to increase with size of male ornament (M ø l l e r 1988, V á c l a v & H o i 2002). In this study, we examined different male and female parental activities, including nest building, nest guarding, chick brooding and chick feeding during (i) pre-laying and laying period, (ii) incubation, (iii) chick brooding and (iv) chick feeding periods. We wanted to answer the following questions: (i) How does the contribution of male parental care vary between breeding stages? (ii) How important is male parental care for hatching and fledging success? (iii) Is early paternal effort a reliable predictor of male parental effort later on during the breeding cycle? Material and Methods Study area and population House sparrows were studied in a nest-box colony in the Schönbrunn Zoological Garden, Vienna in 2001. During the breeding season, 72 of 84 nest boxes were occupied. Nest-boxes were installed on the opposite walls of 7 barns. Egg laying started at the beginning of April. From then the contents of all nest-boxes were inspected at least every third day throughout the breeding season. We recorded the progress in clutch size, number of hatched and fledged young. In the analyses, we use data collected from 31 different pairs. Only the data from first breeding attempts for these pairs were used in our analyses, since the largest part of the nests are constructed prior to the first breeding attempt and afterwards the nest volume changes only slightly (see C r a m p 1994, R. V á c l a v , unpublished data). Furthermore, female investment in parental care may later be affected by male performance in the whole first breeding attempt as well as achieved breeding success. We did not succeed in collecting sufficient data of nest building behaviour for six pairs; hence the sample size for the corresponding analyses is only 25 pairs. Hatching success refers to the proportion of hatched eggs relative to clutch size. Fledgling numbers were estimated according to the number of chicks found in nest-boxes at the age of 13–15 days. Fledging success corresponds to the proportion of fledglings relative to the number of hatched chicks. 300 Behaviour of house sparrows was monitored from around three weeks before the first egg was laid until fledging of the last chick. The behavioural observations were carried out from 7:00 to 12:00 noon. Throughout the breeding cycle, each pair was observed on average for 4.5 hrs. The behaviour at each nesting site was recorded daily during 15 min protocols. In order to avoid an influence of daytime on behaviour, nesting sites were visited following a rotating scheme. We recorded: (i) nest guarding (time spent at or around the nest-box < 5 m), (ii) nest building (the number of arrivals with a nest material to the nest-box), (iii) time at incubation (time spent in the nest-box > 1 min), (iv) chick brooding (time spent in the nest-box after hatching > 1 min), and (v) the number of chick feeding visits. Since behavioural protocols of the studied pairs covered variable periods of the breeding cycle, we restricted our analyses to comparable time intervals of nest building, egg laying, incubation and chick feeding phases (see below). Parental care was classified as early or late, depending on whether it took place before or after clutch completion. The early parental care corresponds to the pre-laying and laying period from 5 days prior to the first egg until the last egg was laid (i.e., late nest building, and nest guarding and incubation during egg laying). This is not only the period when the presumed fertile period of females occurs in house sparrows (see M ø l l e r 1987, B i r k h e a d et al. 1994), but also the time when nest building culminates (S u m m e r s - S m i t h 1963, see C r a m p 1994). Chick feeding was restricted to the period of the first 10 days after hatching, i.e., approximately 70% of the chick feeding period (see S u m m e r - S m i t h 1963). Statistical analyses Some parametric tests were used after normalising data by a log transformation. The frequency of male chick feeding visits was related positively, though not significantly, to brood size (R2 = 0.07, β = 0.27, n = 31, p = 0.14). Because the result would become significant after removing an outlier, to control for brood size, when examining the effect of male provisioning rates, we used residual values of male chick feeding. Relative to the opposite sex or absolute values of parental care activities were used in analyses. To avoid type-I errors in cases when the same variables would have to be examined more times, we used multiple regression tests. For examination of the intensity of nest guarding of identical birds during three breeding stages, we used repeated measures ANOVA. In order to minimise confounding effects of multiple parental behaviours on the parameters of reproductive success, we examined the effect of early paternal effort on hatching success and the effect of later paternal effort on the number of fledglings and fledging success. Results Male contribution to parental care throughout the breeding cycle Examining male parental effort over the course of the breeding cycle, we found that male share in parental care was highest during nest building (Fig. 1). In addition, this form of paternal care varied most dramatically (range: 0–100%, SD = 30.36). Even so, male contribution to nest building was significantly higher than that of their mates (Wilcoxon matched pairs, Z = 2.93, n = 25, p = 0.003; Fig. 1). In contrast to nest building, males contributed the least to incubation and chick brooding (Fig. 1). Again, between-male 301 variation of the parental contribution to these activities was considerable (incubation – range: 0–62.6%, SD = 18.7, n = 31; chick brooding – range: 0–72.5%, SD = 18.0, n = 31). The average contribution to incubation and chick brooding was significantly lower in males than females (contribution to incubation: paired t-test, t = -5.09, n = 31, p < 0.001; contribution to chick brooding: t = -4.67, n = 31, p < 0.001; Fig. 1). In terms of chick feeding, however, male contribution was in average as high as in females (chick feeding: paired t-test, t = -1.64, n = 31, p = 0.11). Male feeding contribution ranged from 14.7–93.3% (SD = 15.6, n = 31). Fig. 1. Male (open bars) and female (filled bars) contributions to different categories of parental care (error bars are SD). Two asterisks (**) indicate p < 0.01, three asterisks (***) p < 0.001. Relationship between paternal care, female parental effort and reproductive success More intensive nest building and constructing bigger nests may decrease incubation or brooding bouts and increase hatching success. However, incubation and brooding time of neither parent seemed to be influenced by male nest building activity (male nest building vs. incubation time of females: R2 = 0.04, β = -0.21, n = 25, p = 0.32, and males: R2 = 0.02, β = 0.13, n = 25, p = 0.53;male nest building vs. brooding time of females: R2 = 0.08, β = 0.28, n = 25, p = 0.18 and males: R2 = 0.001, β = 0.04, n = 25, p = 0.87). Moreover, hatching success was also not significantly related to male nest building activity (R2 = 0.05, β = 0.22, n = 25, p = 0.29). To evaluate the effect of later paternal effort on reproductive success, stepwise multiple regression analyses were performed with male and female brooding time and feeding frequencies as independent variables and the number of fledglings as a dependent variable. Only residual male feeding rates (see method) entered a stepwise regression model at the significance level α < 0.05 (R2 = 0.32, β = 0.57, n = 31, p < 0.001; Fig. 2). Likewise, variation in fledging success was also explained only by residual male feeding of chicks (R2 = 0.30, β = 0.55, n = 31, p < 0.01). 302 Fig. 2. The relationship between residual male feeding of chicks and number of fledglings. Residual values represent feeding rates after controlling for the confounding effect of brood size. Number of fledglings are estimates based on the number of chicks in the nest-box at the age of 13-15 days. Nest guarding, especially during pre-laying and laying period, made up a considerable part of the daily time budgets of males and females (Fig. 3). A 2-way ANOVA revealed a significant effect of the sex and the stage of breeding cycle on the time parents invested in nest guarding (2-way repeated measures ANOVA: sex effect – F1,60 = 32.9, p < 0.001, period effect – F2,120 = 77.3, p < 0.001, interaction – F2,120 = 3.0, p = 0.06; Fig. 3). Both sexes invested in nest guarding most heavily in the onset of the breeding cycle, whereas the activity decreased towards the end of breeding cycle. A high initial nest guarding may assure a better protection of the clutch from intra- or inter-specific brood destruction. However, we found that clutch size was not correlated with the time parents guarded their nests during the period of egg laying (only female nest guarding entered a stepwise multiple regression: R2 = 0.04, β = 0.19, n = 31, p > 0.3). Similarly, nest guarding during incubation period was not able to explain the number of hatchlings (only female nest guarding entered the regression model: R2 = 0.10, β = 0.32, n = 31, p = 0.08), or hatching success (no sex entered the model). Finally, nest guarding during the period of chick feeding could not explain the number of young at fledging age or fledging success (no sex entered the models). Does early male parental contribution reflect his investment in later breeding stages? We were unable to detect that the contribution to early male parental activities would have explained the contribution of male parental care in later stages of the breeding cycle (regressions between male contribution to nest building vs. (i) incubation: R2 = 0.02, β = 0.15, 303 Fig. 3. Time males (solid line) and females (broken line) spent at nest guarding throughout the three periods of the breeding cycle (error bars are SD). n = 25, p = 0.46, (ii) chick feeding: R2 = 0.03, β = 0.17, n = 25, p = 0.40, and (iii) chick brooding: R2 = 0.001, β = -0.03, n = 25, p = 0.88). The early nest guarding likewise proved to be a weak predictor of later paternal effort (early nest guarding vs. (i) incubation: R2 = 0.002, β = -0.04, n = 31, p = 0.83, (ii) chick feeding: R2 = 0.04, β = -0.21, n = 31, p = 0.26, and (iii) chick brooding: R2 = 0.02, β = 0.12, n = 31, p = 0.50). However, after dividing incubation into early (incubation of incomplete clutches) and late (incubation after clutch Fig. 4. The relationship between male contribution to nest building and early incubation. Male contribution to nest building and incubation refers to the relative activity of males to that of their partners. The early incubation represents the period of time from onset to termination of egg-laying. 304 completions), we found that the male contribution to nest building, but not to early nest guarding, significantly explained the variation of his contribution to early incubation (nest building and early incubation: R2 = 0.21, β = 0.46, n = 25, p = 0.022; Fig. 4; early nest guarding vs. early incubation: R2 = 0.001, n = 25, p = 0.87). Discussion Our results revealed variation in paternal care between individuals as well as between breeding periods. We show that male parental share was highest during nest building, decreased during incubation and increased again during chick feeding period. Only chick provisioning turned out to be comparable between partners, whereas nest building was more intense for males, and incubation and brooding for females. Provided that male parental care in house sparrows is essential to maximise reproductive success, females may benefit from correctly assessing male parental ability. Because we found variation of parental effort among male house sparrows, females should invest in reproduction according to some male cues that honestly indicate his condition and perhaps also his future parental effort. Although house sparrows are distinctly sexually dimorphic in their plumage, female assessment of male condition based on epidermal sexual characters such as plumage may not be accurate. This is because plumage does not inform about the current male condition, but about the condition during his previous moult (O w e n s & S h o r t 1995). That is why the size of sexual ornaments of collared flycatchers Ficedula albicollis or house sparrows was suggested as informing females mate’s about a male previous rather than future reproductive effort (G u s t a f s s o n et al. 1995, G r i f f i t h 2000). Hence, in order to estimate male condition and his paternal ability, females may pay attention to specific male activities happening prior to pairing and/or copulations. In our study, the fact that male contribution to nest building was highest of all parental activities suggests that this behaviour could be in the house sparrow a candidate for sexual signal. Male nest building was indeed demonstrated to be involved in sexual signalling in several bird species (G a r s o n 1980, M ø l l e r 1982, B o r g i a 1985, E v a n s & B u r n 1996, S o l e r et al. 1998b), but also in invertebrates (e.g. B a c k w e l l et al. 1995). For instance, the number of advertised nest cups, nest volume or intensity of nest building were shown experimentally to affect mating success in wrens (E v a n s & B u r n 1996), laying date and clutch size in barn swallows (S o l e r et al. 1998a) and magpies (S o l e r et al. 2001). Similarly, V á c l a v & H o i (2002), who studied the same house sparrow population, found that females started to breed earlier and laid larger clutches when mated with those males investing in early paternal care more heavily than other males. Larger nests in penduline tits Remiz pendulinus provide a better insulation and increase hatching success (G r u b b a u e r & H o i 1996). In collared flycatchers, the time of incubation bouts shortens with increasing nest size, whereas hatching success is not compromised (M o r e n o et al. 1991). However, we were unable to find any significant relationships between the intensity of male nest building and hatching success or time at incubation. Because we did not measure any physical parameters of the nest, we cannot rule out the possibility that intensity of nest building during the observed period did not reflect, for example, a nest size. However, R. V á c l a v (unpublished results) found for a Viennese population of house sparrows a positive correlation between the intensity of early parental care and nest weight. Overall, our results suggest that instead of constructing a well insulated nest to protect the clutch and chicks from cold, exaggerated male nest building in the house sparrow may 305 represent a sexual display. Nest guarding may also increase reproductive success, because it lowers the likelihood of egg or brood destruction (S u m m e r s - S m i t h 1963, V e i g a 1990, for house sparrows). Nevertheless, nest guarding that was maintained in our house sparrows high until clutch completion did not explain the variation in clutch size. Male assistance in parental care was lowest in terms of incubation and brooding. This has likely to do with the fact that, as in many other passerines, male house sparrows have lower capacity to incubate than females (S e l a n d e r & Y a n g 1966). Yet, we detected that the intensity of male incubation of incomplete and complete clutches was not consistently correlated with the intensity of nest building. The loss of the relationship between male nest building and the incubation of complete clutches might be explained by a reduction of the circulating testosterone of aggressive males as the fertile period of their partners finished. This is a typical image for males in many species with biparental care (W i n g f i e l d & M o o r e 1987, W i n g f i e l d et al. 1990). However, such a result could also be a consequence of less aggressive males withdrawing to further advertise their parental ability after clutch completion. The highest egg losses that were found in the nests of males investing heaviest in the early forms of parental care (V á c l a v & H o i 2002) could in reality be the consequence of such a reduction of paternal care. Importantly, we found that the intensity of male nest building was not only unrelated to late incubation, but also to male feeding visits. This finding is important because we have shown that male parental assistance is critical during the period of chick feeding. Namely, as in several other studies (e.g. B a r t & T o r n e s 1989, M a r k m a n et al. 1996, D i c k i n s o n & W e a t h e r s 1999), we found that solely male provisioning was able to explain the variation in fledgling numbers and fledging success. Similarly as for nest building, early nest guarding was related neither to male incubation, nor to chick feeding. Moreover, intensity of nest guarding itself was not consistent throughout the breeding cycle, declining with time of the breeding cycle. No matter how intensively they are displayed, male nest building and early nest guarding do not always seem to reliably indicate later paternal effort, say during chick feeding. Some other behaviours, such as courtship feeding in common terns Sterna hirundo, might serve as more reliable signals of paternal quality (N i s b e t 1973). Nevertheless, as V á c l a v & H o i (2002) reported, female house sparrows do allocate their reproductive effort in relation to the intensity of early parental activity of their partners. Hence, males are potentially able to manipulate (stimulate) females to start breeding earlier and lay bigger clutches by displaying a high initial paternal effort or, alternatively, just by providing them a complete nest earlier than other males. The latter option is however less likely, because females were often found to lay eggs into almost empty nest-boxes (pers. obs.). As a result of elevated investment in early parental activities, males can benefit through stimulating females to lay larger clutches and potentially produce more offspring. However, as such males are occupied with parental care and leave their partners unguarded just when they are fertile, the likelihood that they will be cuckolded may likewise increase (V á c l a v et al. 2002). Acknowledgement We thank the staff of the Vienna Zoo for allowing us to carry out this study on their property. Authors are grateful to R. M a n n and anonymous referees and to all who contributed to the discussion and improved the manuscript. R.V. and D. S. were supported by the exchange programme between Austrian and Slovak Academy of Sciences and by the Konrad Lorenz Institute in Vienna. 306 L I T E R AT U R E BACKWELL P. R. Y., JENNIONS M. D., CHRISTY J. H. & SCHOBER, U. 1995: Pillar building in the fiddlercrab Uca beebei: evidence for a condition-dependent ornament. Behavioral Ecology and Sociobiology 36: 185-192. BART J. & TORNES A. 1989: Importance of monogamous male birds in determining reproductive success. Evidence for house wrens and a review of male-removal studies. Behavioral Ecology and Sociobiology 24: 109-116. BIRKHEAD T. R., VEIGA J. P. & MØLLER A. P. 1994: Male sperm reserves and copulation behaviour in the house sparrow, Passer domesticus. Proceedings of the Royal Society London, Series B 256: 247-251. BORGIA G. 1979: Sexual selection and the evolution of mating systems. In: Blum S.M. & Blum N.A. (eds), Sexual Selection and Reproductive Competition. Academic Press, New York: 19-80. BORGIA G. 1985: Bowers as markers of male quality. Test of a hypothesis. Anim. Behav. 35: 266-271. BURLEY N. 1981: Mate choice by multiple criteria in a monogamous species. American Naturalist 117: 515–528. BURLEY N. 1988: The differential-allocation hypothesis: An experimental test. American Naturalist 132: 611628. CATCHPOLE C. K. 1980: Sexual selection and the evolution of complex song among European warbles of the genus Acrocephalus. Behaviour 74: 149-166. CRAMP S. 1994: The birds of the Western Palearctic. Vol. IX. Oxford University Press, Oxford. DE LOPE F. & MØLLER A. P. 1993: Female reproductive effort depends on the degree of ornamentation of their mates. Evolution 47: 1152-1160. DICKINSON J. L. & WEATHERS W. W. 1999: Replacement males in the western bluebird: opportunity for paternity, chick-feeding rules, and fitness consequences of male parental care. Behavioral Ecology and Sociobiology 45: 201-209. EVANS M. R. & BURN J. L. 1996: An experimental analysis of mate choice in the wren: a monomorphic, polygynous passerine. Behav. Ecol. 7: 101-107. GARSON P. J. 1980: Male behaviour and female choice: mate selection in the wren? Anim. Behav. 28: 491–502. GIBBONS D. W. 1987: Hatching asynchrony reduces parental investment in the jackdaw. J. Anim. Ecol. 56: 403-414. GRIFFITH S. C. 2000: A trade-off between reproduction and a condition-dependent sexually selected ornament in the house sparrow Passer domesticus. Proceedings of the Royal Society London, Series B 267: 1115-1119. GRUBBAUER P. & HOI H. 1996: Female penduline tits (Remiz pendulinus) choosing high quality nests benefit by decreased incubation effort and increased hatching success. Ecoscience 3: 274-279. GUSTAFSSON L., QVARNSTRÖM A. & SHELDON B. C. 1995: Trade-offs between life-history and a secondary sexual character in male collared flycatchers. Nature 375: 311-313. HAMILTON W. D. & ZUK M. 1982: Heritable true fitness and bright birds: a role for parasites? Science 218: 384-387. HOEZLER G. A. 1989: The good parent process and sexual selection. Anim. Behav. 8: 1067-1078. HOI-LEITNER M., NECHTELBERGER H. & HOI H. 1995: Song rate as a signal for nest site quality in blackcaps (Sylvia atricapilla). Behavioral Ecology and Sociobiology 37: 399-405. IWASA Y., POMIANKOWSKI A. & NEE S. 1991: The evolution of costly mate preferences. II. The ‘handicap’ principle. Evolution 45: 1431-1442. KIMBALL R. T. 1996: Female choice for male morphological traits in house sparrows, Passer domesticus. Ethology 102: 639-648. KIRKPATRICK M. 1985: Evolution of female mate choice and male parental investment in polygynous species: the demise of the “sexy son”. American Naturalist 125: 788-810. MARKMAN S., YOM-TOV Y. & WRIGHT J. 1996: The effect of male removal on female parental care in the orange-tufted sunbird. Anim. Behav. 52: 437-444. MØLLER A. P. 1982: Clutch size in relation to nest size in the swallow Hirundo rustica. Ibis 124: 339-343. MØLLER A. P. 1987: House sparrow, Passer domesticus, communal displays. Anim. Behav. 35: 203-210. MØLLER A. P. 1988: Female choice selects for male sexual tail ornaments in the monogamous swallow. Nature 322: 640-642. MØLLER A. P. 1994: Sexual selection and the barn swallow. Oxford University Press, Oxford. MØLLER A. P. 1998: Sperm competition and sexual selection. In: Birkhead T.R. & Møller A.P. (eds), Sperm Competition and Sexual Selection. Academic Press, London: 55-90. MØLLER A. P. 2000: Male parental care, female reproductive success, and extrapair paternity. Behav. Ecol. 11: 161-168. 307 MORENO J., GUSTAFSSON L., CARLSON A. & PÄRT T. 1991: The cost of incubation in relation to clutchsize in the collared flycatcher, Ficedula albicollis. Ibis 133: 186-193. MORENO J., VEIGA J. P., CORDERO P. J. & MINGUEZ E. 1999: Effects of paternal care on reproductive success in the polygynous spotless starling Sturnus unicolor. Behavioral Ecology and Sociobiology 47: 4753. NISBET I. C. T. 1973: Courtship feeding, egg size and breeding success in common terns. Nature 241: 141-142. NUR N. & HASSON O. 1984: Phenotypic plasticity and the handicap principle. Journal of Theoretical Biology 110: 275-297. OWENS I. P. F. & SHORT R. V. 1995: Hormonal basis of sexual dimorphism in birds: implications for new theories of sexual selection. Trends in Ecology and Evolution 10: 44-47. POMIANKOWSKI A. 1987: The ‘handicap principle’ does work - sometimes. Proceedings of the Royal Society London, Series B 127: 123-145. PRICE T. D., SCHLUTER D. & HECKMAN N. E. 1993: Sexual selection when the female directly benefits. Biological Journal of Linnaean Society 48: 187-211. RYAN M. J., FOX J. H., WILCZYNSKI W. & RAND A. S. 1990: Sexual selection for sensory exploitation in the frog Physalaemus pustulosus. Nature 343: 66-67. SELANDER R. K. & YANG S. Y. 1966: The incubation patch of the House Sparrow. General and Comparative Endocrinology 6: 325–333. SOLER J. J., CUERVO J. J., MØLLER A. P. & DE LOPE F. 1998a: Nest building is a sexually selected behaviour in the barn swallow. Anim. Behav. 56: 1435–1442. SOLER J. J., MØLLER A. P. & SOLER M. 1998b: Nest building, sexual selection and parental investment. Evolutionary Ecology 12: 427-441. SOLER J. J., DE NEVE L., MARTINEZ J. G. & SOLER M. 2001: Nest size affects clutch size and the start of incubation in magpies: an experimental study. Behav. Ecol. 12: 301-307. SOLER M., SOLER J. J., MØLLER A. P., MORENO J. & LINDÉN M. 1996: An experimental analysis of the functional significance of an extreme sexual display: Stone-carrying in the black wheatear Oenanthe leucura. Anim. Behav. 51: 247–284. SUMMERS-SMITH J.D. 1963: The House Sparrow. Collins Clear-Type Press, London. TRIVERS R. L. 1972: Parental investment and sexual selection. In: Campbell B. (ed.), Sexual selection and descent of man. Aldine, Chicago: 136-179. VÁCLAV R. & HOI H. 2002: Different reproductive tactics in house sparrows signalled by badge size: is there a benefit to being average? Ethology 108: 569-582. VÁCLAV R., HOI H. & BLOMQVIST D. 2002: Badge size, paternity assurance behaviours and paternity losses in male house sparrows. J. Avian Biol. 33: 315-318. VEIGA J. P. 1990: Sexual conflict in the house sparrow: interference between polygynously mated females versus asymmetric male investment. Behavioral Ecology and Sociobiology 27: 345-350. WEATHERHEAD P. J. & ROBERTSON R. J. 1979: Offspring quality and the polygyny threshold: the “sexy son” hypothesis. American Naturalist 113: 201-208. WELLING P. P., RYTKONEN S. O., KOIVULA K. T. & ORELL M. I. 1997: Song rate correlates with paternal care and survival in willow tits: Advertisement of male quality? Behaviour 134: 891-904. WEST-EBERHARD M. J. 1979: Sexual selection, social competition and evolution. Proceedings of the American Philosophical Society 123: 222-234. WINGFIELD J. C., HEGNER R. E., DUFTY A. M. & BALL G. F. 1990: The “Challenge Hypothesis:” Theoretical implications for patterns of testosterone secretion, mating systems and breeding strategies. American Naturalist 136: 829-846. WINGFIELD J. C. & MOORE M. C. 1987: Hormonal, social, and environmental factors in the reproductive biology of free-living male birds. In: Crews D. (ed.), Psychobiology of Reproductive Behavior. An Evolutionary Perspective. Prentice Hall, International, Englewood Cliffs, New Jersey: 149-175. ZAHAVI A. 1975: Mate selection – a selection for a handicap. Journal of Theoretical Biology 53: 205-214. 308 Folia Zool. – 52(3): 309–316 (2003) Breeding ecology of Algerian woodchat shrikes Lanius senator: low breeding success Zahra BRAHIMIA1, Hamdi DZIRI1, Slim BENYACOUB1, Yassine CHABI1 & Jerzy BA¡BURA2* 1 Department of Biology, Faculty of Science, Badji Mokhtar University, B.P. 12, Annaba, Algeria 2 Department of Experimental Zoology and Evolutionary Biology, University of ¸ódê, Banacha 12/16, PL-90237 ¸ódê, Poland; e-mail: [email protected] Received 17 January 2002; Accepted 28 February 2003 A b s t r a c t . Breeding ecology of woodchat shrikes Lanius senator was investigated in the ELKala National Park in North-East Algeria (36°53’N; 8°30’E) in 1998–1999. Quercus suber was the tree species most frequently used as support for nests, which were constructed at a mean height of 5 m. First eggs in clutches were laid 7 May, and clutch size was 4.9. Approximately 42% eggs gave fledglings. Clutch size declined during the course of the breeding season but fledgling success did not. Fledging success was positively correlated with per clutch mean egg length and the height of the nest location above ground. We suggest that the major selective pressures that shape the life history of Algerian woodchat shrikes are relatively heavy predation and poor food availability. Key words: North Africa, passerine, biology, clutch size, nesting, egg size, nestling growth, fledging success Introduction Over most of the Western Palearctic area, populations of the woodchat shrike Lanius senator are in decline, and the species is now considered threatened in many countries (S n o w & P e r r i n s 1998) but remains relatively common in the Mediterranean area, including North Africa (S n o w & P e r r i n s 1998, I s e n m a n n & F r a d e t 1998, B e n y a c o u b & C h a b i 2000). The breeding biology of this species remains poorly known and many aspects require further study. Reproductive strategies of passerine birds show some consistent patterns, which most probably evolved under selection pressures generated by food and predation (M a r t i n 1995). Food of insectivorous birds is usually abundant and accessible for a rather short period during the breeding season (M a r t i n 1995). Predation on nests is an especially important factor of selection, and its rate is to a large extent associated with nesting sites and sometimes also with time during the breeding season (R i c k l e f s 1969, M a r t i n 1995). Open nests located in bushes or short trees experience particularly heavy predation (R i c k l e f s 1969). As such, a common pattern of decline in clutch size and breeding success has been observed over the course of the breeding season (P e r r i n s 1970, C r i c k et al. 1993), as has careful nest site selection by breeding pairs (M a r t i n 1995). We aim in this paper to analyse some characteristics of the breeding ecology of the woodchat shrike in Algeria, an area for which detailed data have been so far lacking. The main objectives of this study were to: 1) obtain information on nest location and size, clutch size, egg sizes, nestling growth and fledging success; 2) characterize timing of egg laying and its relation with other life history traits; and 3) investigate whether fledging success is * Corresponding author 309 related to other life history traits and nest location. Our study raised the important question of whether failure or success of a brood is influenced by bird quality related variables. If so, then we can predict that fledging success of a particular clutch should be positively influenced by nest and nesting characteristics or egg size (c.f. G o o d b u r n 1991). Material and Methods The study was carried out in the EL-Kala National Park in North-East Algeria (36°53’N; 8°30’E) over a two-year period (1998–1999). Climatic conditions during April and May of the study were largely similar, with relatively high mean temperature 14–19 °C and rainfall 21–53 mm per month. The study area was extensively cultivated farmland, with nonintensive livestock systems influencing the character of vegetation (B e n y a c o u b & C h a b i 2000). The study site was located amongst the mosaic of degraded forest habitats with Quercus suber dominating and concurring with shrubs of Myrtus comunis, Rubus ulmifolius, Pistacia lentiscus, Calycotome villosa, Daphne gnidum, Erica arborea, Genista tricuspidata, Lolium multiflorum, Lythrum hypossopifolia and others. Searching for nests started in February and was conducted at least once a week. Thirtynine initiated clutches of the woodchat shrike were monitored during the course of nesting (23 nests in 1998 and 16 in 1999) every 1–5 days, more frequently during the expected time of hatching and the nestling period. As the search for nests within the forest patches studied was detailed, it is unlikely that any initiated woodchat shrike clutch could remain nonrecorded in the study area. Consequently, we were able to consider a within-clutch proportion of eggs, which produced fledglings as a simple measure of fledging success. The nesting tree/shrub species were determined in 1998 (n = 22 records, 1 record lost). In 1998–1999, tree height and the height of nest location were measured (with n = 37 records of nest location height), with n = 27 of the nests measured to the nearest mm (external diameter, nest cup diameter and depth). The sample sizes differ for different variables due to subsampling or incompleteness of records. All of the initiated clutches were controlled to record the date of the onset of egg laying, clutch size, the number of fledglings and losses at different stages. Egg dimensions were measured in a subsample of clutches to the nearest 0.1 mm with calipers: egg length (L) and breadth (B). These measurements were used to calculate egg volume (V) and egg shape (ES) according to the formulas: V = 0.000507*L*B2 (H o y t 1979), and ES = B/L*100%. All characteristics related to egg dimensions are presented as means per clutch to avoid the non-independence of egg traits within clutches (B a ƒ b u r a & Z i e l i ƒ s k i 1990, 1998, Z i e l i ƒ s k i & B a ƒ b u r a 1998). For descriptive purposes, we also studied the growth of nestling woodchat shrikes. In 16 broods, the growth of nestlings was monitored by weighing a random nestling from a brood, from the day 1 to 13 of the nestling life (3–7 nestlings from different broods per day). The data were analysed by standard parametric and non-parametric methods (S o k a l & R o h l f 1995). In the analysis of the within-clutch proportion of eggs that produced fledglings (binomial error), the generalized linear model approach was applied (D o b s o n 1990). Original nest characteristics were used to calculate principal components for further analysis. Split-linear regression was conducted following K o v á ã et al. (1999). All analyses and calculations were carried out using MS Excel and STATISTICA 5.5 (S t a t S o f t 1997). 310 Results Sixteen or 69.57% ± 9.59 (SE), of the 22 nests analysed in 1998 were constructed on Quercus suber, 2 on Fraxinus angustifolia, 2 on Crategus monogyna, 1 on Olea oleaster and 1 on Pyrus communis. Woodchat shrikes tended to select trees rather than bushes on which to construct a nest; nest height varied greatly (Table 1). Table 1. Nest site and nest characteristics of woodchat shrikes in 1998–1999 in Algeria. Variable n Mean ± SD Median Min Max Nesting tree height (m) Nest site height (m) Nest external diameter (mm) Nest cup diameter (mm) Nest cup depth (mm) 35 37 27 27 27 6.4 ± 2.7 5.1 ± 1.6 119.6 ± 15.3 77.6 ± 5.7 54.4 ± 8.3 5.9 5.0 119.7 78.0 55.0 2.5 1.7 91.8 50.0 47.0 16.0 9.0 160.0 90.0 90.0 No significant difference in the distribution or mean value of the date of the onset of laying was recorded between 1998 and 1999 (Kolmogorov-Smirnov test: P > 0.1; Mann-Whitney test: P = 0.31). No between-year difference in clutch size, fledgling number or egg traits (length, breadth, volume, shape; 56 eggs in 11 clutches in 1998 and 40 eggs in 7 clutches in 1999; mean clutch size in the subsample: 5.3 ± 0.2 SE) was found (Hotelling T test: P = 0.60; individual t tests with separate variance estimation: 0.11 < P < 0.56). Consequently, pooled characteristics are presented hereafter (Tables 2 and 3). Incubation lasted 15–17 days, mean 15.7 ± 0.2 (SE), from the moment of having laid the last egg in the clutch. Clutch size was negatively correlated with laying date (r = -0.62, n = 35, P < 0.0001) (Fig. 1). Successful and non-successful broods did not differ in the date of laying (MannTable 2. The date of the onset of laying, clutch size and the number of fledglings in woodchat shrike broods in 1998–1999 (pooled) in Algeria. Variable n Mean ± SD Median Min Max Laying date Clutch size Fledgling number Fledgling number (non-0) 39 35 35 17 7 May ± 11.37 4.91 ± 1.01 2.14 ± 2.40 4.41 ± 1.28 4 May 5.00 0 4 13 April 3 0 2 11 June 6 6 6 Table 3. Characteristics of eggs of woodchat shrikes in 1998–1999. Per-clutch mean values for 17 complete clutches included. Variable Mean ± SD Median Min Max Egg length (mm) Egg breadth (mm) Egg volume (cm3) Egg shape 22.9 ± 0.65 16.7 ± 1.06 3.0 ± 0.40 72.9 ± 4.08 22.9 16.9 3.1 73.2 21.7 13.4 1.8 61.2 24.1 17.6 3.5 78.9 311 Whitney test: P = 0.46). A generalized linear model with laying date as the independent variable and fledging success (binomial error) as response variable showed no relation between these two variables (χ2: P = 0.30). Fig. 1. Relationship between clutch size and the onset of egg laying in the woodchat shrikes studied in 1998–1999. Seventy-five of 178 eggs gave fledglings (42.13% ± 3.71 SE), and 17 out of 39 clutches (43.59% ± 7.94 SE) gave at least 2 fledglings (Table 2). So, most breeding attempts failed (56.41 % ± 7.94 SE): 20.51% ± 6.46 (SE) of the 39 initiated clutches were abandoned at different stages of egg laying and incubation probably due to predation; 25.64% ± 6.99 (SE) of the clutches were completely destroyed by human or animal predators; 2.56% ± 2.53 (SE) were completely destroyed at the nestling stage; and 7.69% ± 4.27 (SE) were completely destroyed at the stage of nestlings after first having partially destroyed eggs (n = 39). Partial destruction occurred in 10.26% ± 4.86 (SE) at the stage of eggs and in 7.69% ± 4.27 (SE) at the stage of nestlings (n = 39). We found no breeding losses caused by bad weather. Mean egg size positively influenced fledging success (generalized linear model: χ21 = 39.34, P < 0.0001); mean egg length in successful clutches was 23.25 mm ± 0.21 (SE) as compared with 22.58 mm ± 0.18 (SE) for clutches that failed to produce any fledglings. A generalized linear model with nest and nest site characteristics as independent variables and fledging success as response variable showed that the height of the nest location above ground positively influenced fledging success (χ21 = 14.77, P = 0.0001), whereas neither nest characteristics nor the height of the nesting tree were related to fledging success (NS). Nests that produced no fledglings were generally constructed at a lower height than successful nests: 4.73 m ± 0.40 (SE) v. 5.68 m ± 0.25 (SE). Growth of nestlings was fast and roughly linear during the first 9 day period (slope = 2.83 ± 0.11 SE), then markedly slower reaching fledging mass (slope = 0.92 ± 0.38 SE), as shown by the split-linear regression (Fig. 2). Day 9 ± 0.59 SE is the breakpoint of the splitlinear regression line. The split-linear regression accounted for more variation (R2 = 0.95) 312 than a simple linear regression (R2 = 0.94) or a quadratic regression (R2 = 0.92) which were analysed for comparison (F tests: P < 0.001). Fig. 2. Growth of woodchat shrike chick body mass with a split regression curve shown. Discussion Woodchat shrikes nested mostly on Quercus suber in our study area. This corroborates suggestions that open Quercus forests are the species’ preferred habitats, where available, in the Mediterranean area (S n o w & P e r r i n s 1998). I s e n m a n n & F r a d e t (1998) suggest that woodchat shrikes generally use the most abundant tree or bush species, so that the selection of the nest site is rather opportunistic. The tree or shrub species that support shrike nests constrain the possible range of height at which the nests are constructed. As in most of the species’ breeding area, woodchat shrikes constructed nests at a mean height of about 5 m (G u s e v & B e d n y i 1961, A d a m i a n 1964, U l l r i c h 1971, H u d e c et al. 1983, C r a m p & P e r r i n s 1993, B o n a c c o r s i & I s e n m a n n 1994). However, this mean value is much higher than that of another Algerian population studied near Tizi Ouzou, 2.62 m (M o a l i et al. 1997). Nest external diameter was relatively low in our study area, whereas the diameter and depth of nest cup were similar to the values reported in other studies (H u d e c et al. 1983, C r a m p & P e r r i n s 1993). Also, egg dimensions in our woodchat shrike population were within the range reported for different parts of the species’ area (G u s e v & B e d n y i 1961, E t c h e c o p a r & H u e 1964, H u d e c et al. 1983, C r a m p & P e r r i n s 1993). In many areas of woodchat shrike distribution, a mean clutch size is > 5 eggs, but a mean clutch size of 4.75 has also been reported (I s e n m a n n & F r a d e t 1998, G u e r r i e r i & C a s t a l d i 2000), so that the value of about 4.9 recorded in our study is within this range. Typical of single–brooded passerines (P e r r i n s 1970, C r i c k et al. 1993), Algerian woodchat shrikes showed a seasonal decline in clutch size. However, the timing of breeding did not have any effect on nest or fledging success. In our study area, laying eggs started at similar time as in the Tizi Ouzou (M o a l i et al. 1997), but earlier than in southern France (I s e n m a n n & F r a d e t 1998). 313 Nest predation is known to cause relatively low breeding success in different woodchat shrike populations (C r a m p & P e r r i n s 1993; B e c h e t et al. 1998). Predation by animals and humans was also a main factor of breeding failure in our study population in which weather-related causes, important in many shrike populations (e.g. S t a u b e r & U l l r i c h 1970; T a k a g i 2001), seem to play a non-significant role. In French populations, B e c h e t et al. (1998) recorded an even lower breeding success (36.5%) than we observed. Time of the season does not appear to affect nest predation (B e c h e t et al. 1998). A high level of nest loss due to predatory animals is characteristic of the passerines constructing open nests in shrubs (R i c k l e f s 1969, M a r t i n 1995). A similar phenomenon may also occur in species that nest in short trees. Our observation that fledging success is positively correlated with the height at which a nest is located suggests that there may exist a pressure of natural selection to nest higher up within a tree. On the other hand, I s e n m a n n & F r a d e t (1998) and B e c h e t et al. (1998) did not record any relation between breeding success and nest site, including the nest height. This suggests that the existence and direction of the selection pressure to nest higher may be highly site-specific. Fledging success was also positively related to egg length, which is not reported very often (W i l l i a m s 1994). A possible explanation may be that woodchat shrike pairs of high quality, perhaps old ones, lay bigger eggs, are good at selecting a right location to nest and are better at nest defense against predators. Such effects were described in some avian species such as magpies Pica pica (G o o d b u r n 1991). Comparative data for nestling growth in woodchat shrikes are scarce. G u s e v & B e d n y i ’ s (1961) graph is similar to ours in general characteristics, but detailed comparison is not feasible. The shape of the woodchat shrike growth curve may be considered as a version of the typical curve of altricial passerines (c.f. O ’ C o n n o r 1984). The curve has one significant break point that divides it into two component lines reflecting two stages of the nestling development (Fig. 2). At the beginning of the first stage nestlings develop their thermoregulatory abilities, which reduces energy loss and thus favours mass gain, which results in the high growth rate. At the second stage, there occurs a decrease in growth rate at about 9 days old, resulting from energy expenditure to plumage development (O ’ C o n n o r 1984). The general pattern suggests that nestling shrikes grow as fast as possible to attain adult body mass (30–40 g, S n o w & P e r r i n s 1998). They do not show an excess weight before fledging, which combined with slow growth is very characteristic of many passerines nesting in safe places, such as hole-nesters, especially aerial insectivores (O ’ C o n n o r 1984, R e m e Ê & M a r t i n 2002). In summary, the woodchat shrike population studied seems to be exposed to relatively high predation, resulting in low breeding success (B e n y a c o u b & C h a b i 2000). If this pressure were consistent over a longer time range, then some changes in breeding behaviour would be expected (M a r t i n 1995). In general, a tendency to make ontogeny as short as possible would be expected (R i c k l e f s 1969, M a r t i n 1995, R e m e Ê & M a r t i n 2002). Nest site choice and nesting behaviour would be predicted to undergo a strong pressure of natural selection. The actual modifications of the breeding strategy, if the predation pressure continued to operate, would be constrained by some genetic and environmental factors. The major environmental constraint in the study area would be a lack of more suitable trees (accessibility and height). This may also explain some discrepancy between the results concerning nest site selection reported in this paper and those shown by I s e n m a n n & F r a d e t (1998) and B e c h e t et al. (1998). We would not predict the 314 occurrence of a selective predation pressure on the timing of breeding as long as the predation rate is not related to time during the season. Food is probably also an important factor, influencing both clutch size and breeding success in Algerian woodchat shrikes. In many European areas, woodchat shrikes prey upon lizards and frogs (C r a m p & P e r r i n s 1993). These animals are rarely available in our study site, the diet being composed mostly of insects belonging to Orthoptera and Coleoptera, whose abundance increases seasonally prior to breeding (unpublished observation). This may negatively influence the quality of food preferred by woodchat shrikes. Consequently, we suggest that the major selective pressures that shape the life history of Algerian woodchat shrikes are relatively heavy predation and poor food availability and quality (c.f. M a r t i n 1995). Acknowledgements We are grateful to the authorities of the El-Kala National Park for their help. We thank Paul I s e n m a n n for supplying some items of literature. Vladimír K o v á ã kindly helped us in calculating and analysing split-line regressions. We are grateful to anonymous referees and Gordon H. C o p p for their helpful remarks on the manuscript. L I T E R AT U R E ADAMIAN M. S. 1964: The ecology of Lanius senator niloticus BP. in Armenian SSR. Academy of Sciences of Armenian SSR Zoological Institute, Zoological Papers 13: 85–104 (in Russian with English summary). BA¡BURA J. & ZIELI¡SKI P. 1990: Within-clutch repeatability of egg dimensions in the black-headed gull Larus ridibundus. J. Ornith. 131: 305–310. BA¡BURA J. & ZIELI¡SKI P. 1998: An analysis of egg-size repeatability in barn swallows Hirundo rustica. Ardeola 45: 183–192. BECHET A., ISENMANN P. & GAUDIN R. 1998: Nest predation, temporal and spatial breeding strategy in the woodchat shrike Lanius senator in Mediterranean France. Acta Oecol. 19: 81–87. BENYACOUB S. & CHABI Y. 2000: Diagnose ecologique de l’avifaune du Parc National d’El-Kala. Synthese – Publication de l’Universite d’Anaba-Algerie, Annaba. BONACCORSI G. & ISENMANN P. 1994: Biologie de la reproduction et nourriture de la pie-grieche à tête rousse Lanius senator badius et de la pie-grieche ecourcheur Lanius collurio en Corse (France). Alauda 62: 269–274. CRAMP S. & PERRINS C. M. 1993: The Birds of the Western Palearctic. Vol. VII. Oxford University Press, Oxford. CRICK H. O. P., GIBBONS D. W. & MAGRATH R. D. 1993: Seasonal changes in clutch size in British birds. J. Anim. Ecol. 62: 263–273. DOBSON A. J. 1990: An Introduction to Generalized Linear Models. Chapman and Hall, London. ETCHECOPAR R.D. & HUE F. 1964: Les Oiseaux du Nord de l’Afrique. Editions N. Boubee & C., Paris. GOODBURN S.F. 1991: Territory quality or bird quality? Factors determining breeding success in the magpie Pica pica. Ibis 133: 85–90. GUERRIERI G. & CASTALDI A. 2000: Selezione di habitat e reproduzione dell’ averla capirossa, Lanius senator, nel Lazio – Italia Centrale. Avocetta 24: 85–93. GUSEV V. M. & BEDNYI S. N. 1961: Contribution to the ecology of the eastern woodchat shrike (Lanius senator niloticus Bonaparte) in eastern Georgia. Trudy Instituta Zoologii Akademii Nauk Gruzinskoi SSR 18: 41–51 (in Russian). HOYT D.F. 1979: Practical method of estimating volume and fresh weight of bird eggs. Auk 96: 73–77. HUDEC K. et al. 1983: Fauna of the CSSR. Birds. Vol. III/2. Academia, Praha (in Czech with German summary). ISENMANN P. & FRADET G. 1998: Nest site, laying period, and breeding success of the woodchat shrike (Lanius senator) in Mediterranean France. J. Ornith. 139: 49–54. 315 KOVÁâ V., COPP G. H. & FRANCIS M. P. 1999: Morphometry of the stone loach, Barbatula barbatula: do mensural characters reflect the species’ life history thresholds? Env. Biol. Fish. 56: 105–115. MARTIN T. E. 1995: Avian life history evolution in relation to nest sites, nest predation, and food. Ecol. Monogr. 65: 101–127. MOALI A., AIT SIDHOUM D. & ISENMANN P. 1997: Quelques donnes sur la reproduction de la pie-grieche à tête rousse Lanius senator en Algerie. Alauda 65: 205–207. O’CONNOR R. J. 1984: The Growth and Development of Birds. Wiley, Chichester. PERRINS C. M. 1970: The timing of birds’ breeding season. Ibis 112: 242–255. REME V. & MARTIN T. E. 2002: Environmental influences on the evolution of growth and developmental rates in passerines. Evolution 56: 2505–2518. RICKLEFS R. E. 1969: An analysis of nesting mortality in birds. Smithsonian Contrib. Zool. 9: 1–48. SNOW D.W. & PERRINS C.M. (eds) 1998: The Birds of the Western Palearctic. Concise Edition. Vol. II. Oxford University Press, Oxford. SOKAL R. R. & ROHLF F. J. 1995: Biometry. 3rd edition. Freeman, New York. STATSOFT 1997: STATISTICA for Windows Manual. StatSoft, Inc, Tulsa. STAUBER W. & ULLRICH B. 1970: Der Einfluss des nasskalten Frühjahres 1969 auf eine Population des Rotrückenwürgers (Lanius collurio) und Rotkopfwürgers (Lanius senator) in Südwestdeustchland. Die Vogelwelt 91: 213–222. TAKAGI M. 2001: Some effects of inclement weather conditions on the survival and condition of bull-headed shrike nestlings. Ecol. Res. 16: 55–63. ULLRICH B. 1971: Untersuchungen zur Ethologie und Ökologie des Rotkopfwürgers (Lanius senator) in Südwestdeustchland im vergleich zu Raubwürger (L. excubitor), Schwarzstirnwürger (L. minor) und Neuntöter (L. collurio). Die Vogelwarte 26: 1–77. WILLIAMS T. D. 1994: Intraspecific variation in egg size and egg composition in birds: effects on offspring fitness. Biol. Rev. 68: 35–59. ZIELI¡SKI P. & BA¡BURA J. 1998 Egg size variation in the Swallow Hirundo rustica Acta Ornith. 33: 191–196. 316 Folia Zool. – 52(3): 317–322 (2003) Production-growth model applied in eublepharid lizards (Eublepharidae, Squamata): accordance between growth and metabolic rates LukበKRATOCHVÍL and Daniel FRYNTA Department of Zoology, Faculty of Sciences, Charles University, Viniãná 7, CZ-128 44 Praha 2, Czech Republic; e-mail: [email protected]; [email protected] Received 13 February 2002; Accepted 20 February 2003 A b s t r a c t . We studied growth rates in four eublepharid species in which mass-metabolic allometry was available from the literature. The lizards were reared in a common garden environment and the growth data were analysed using the production/growth model (W e s t et al. 2001: Nature 413: 628–631). The model fits well our data in eublepharid geckos, and thus the applicability of this model to reptile growth was demonstrated. Estimated values of growthrate parameter (a) fell within the range known in other ectotherm animals. As expected from theoretical parameter derivation, species sharing the same mass-metabolic allometry (Eublepharis macularius, Coleonyx mitratus, Coleonyx elegans) show comparable a in spite of considerable interspecific differences in asymptotic body mass. Moreover, in accordance with theoretical predictions, the only examined species with the elevated metabolic scaling, i.e. Coleonyx brevis, showed higher a than the other species. Key words: growth curve, Eublepharis, mass-metabolic allometry Introduction Individual growth of animals has been modeled using a number of growth curves (e.g. H a i l e y & C o u l s o n 1999, W a n et al. 2001, M o s c a r e l l a et al. 2001). These empirical functions usually fit to each data set well, on the other hand, their parameters have been difficult or even impossible to interpret biologically (but see V o n B e r t a l a n f f y 1957, L ó p e z et al. 2000). The need to approach this problem in a new spirit, in particular to deduce general growth function from basic physiological mechanisms, has been evident. The production/growth model recently derived by W e s t et al. (2001) is based on the allocation of metabolic energy between maintenance of existing tissue and the production of new biomass. Generally, this model ascribes the slowing of growth as body size increases to limitation on the capacity of networks to supply adequate resources to support further increase in body mass. W e s t et al. (2001) derived a universal sigmoidal growth curve ⎧ m ⎫1/4 ⎡ ⎧ m0 ⎫1/4⎤ –at /4M 1/4 ⎜ ⎢ = 1 – ⎢1 – ⎜ ⎢ ⎢e ⎩M⎭ ⎢ ⎩M ⎭ ⎢ where m is the body mass at the time t, M is the asymptotic maximum body mass, m0 is the mass at birth, and a is the curve parameter calculable from fundamental cellular parameters. Parameter a ≡ B0mc/Ec, where mc is the mass of a cell, Ec is the metabolic energy required to create a cell and B0 is mass normalized metabolic rate (B0 = B/m3/4, where B is the average resting metabolic rate of the whole organism at time t), i.e. B0 is constant for a given taxon. 317 The growth-rate parameter (a) is thus expected to be similar for species with similar body temperature and similar metabolic scaling on body mass (C h a r n o v 2001). The curve shows good accordance with data on individual growth across a range of animal taxa (mammals, birds, fish, shrimp), however, the taxonomic spectrum of examined organisms has been limited, e.g. reptiles has been completely lacking (W e s t et al. 2001). This work summarizes growth data from a common-garden experiment carried out in four species of the gecko family Eublepharidae, which are good model organisms for a study investigating the new model. These lizards exhibit considerable diversity in body size and standard metabolic rate scaling (D i a l & G r i s m e r 1992, K r a t o c h v í l & F r y n t a 2002). In fact, few groups have such a broad range of body size within one small monophyletic assemblage. Large Eublepharis macularius Blyth, 1854, and much smaller Coleonyx elegans Gray, 1845 and C. mitratus (Peters, 1845) share the same (ancestral) allometric relationship between standard metabolic rate and body mass. The smallest species C. brevis Stejneger, 1893 is a member of the desert clade with derived high standard metabolic rates (D i a l & G r i s m e r 1992). Therefore, when the production/growth model is fitted on growth data taken at the same temperature, E. macularius, C. elegans and C. mitratus are expected to show comparable a despite the considerable among-species differences in asymptotic body mass, and C. brevis, the species with elevated metabolic scaling, should show relatively higher a. Material and Methods Animals born in the laboratory were weighed from hatching up to the age of approximately two years in intervals of one or two months with a digital balance to the nearest 0.1 g. Lizards were maintained individually in small cages with shelters and substrate (wet peat-moss or sand according to preferred humidity) in conditions that allowed no active thermoregulation. They were placed in a centrally heated room with a 12L:12D photoperiod at 26 ± 1 °C. This temperature is near the preferred body temperature in eublepharids and the temperature used in metabolic rate measurements (D i a l & G r i s m e r 1992). Water and food (vitaminised crickets and mealworms) were provided three times a week, i.e. both were provided ad libitum. In total, we obtained 624 measurements in 111 individuals of known age (in days) kept under these standard conditions. Data concerning a single individual (outlier) showing apparently retarded growth were excluded. Growth curve characteristics of these species in other measured trait (snout-vent length) were published elsewhere (K r a t o c h v í l & F r y n t a 2002). First, we applied a production/growth curve to the overall sample of a given species using the Levenberg-Marquardt algorithm, which minimized the sum of squares between predicted and observed values of growth. In this case, the mass at birth (parameter m0) was estimated as a mean mass of freshly hatched juveniles (weighted within three days after hatching) from our lab (E. m.: 3.36 g, n = 24; C. m.: 0.83 g, n = 16; C. e.: 1.11 g, n = 30; C. b.: 0.30 g, n = 8). Because the intersexual differences were much smaller than interspecific, we pooled data of both sexes in a given species sample. Second, to exclude pseudoreplicates we selected well-represented individuals, i.e. individuals with more than seven measurements and measured long enough after the inflection point of growth trajectory, and fitted the production/growth curve separately to each individual growth trajectory of these individuals (11 individuals of E. m., 14 ind. of C. m., 7 ind. of C. e., and 3 318 ind. of C. b.). Interspecific differences in growth curve parameters were tested with KruskalWallis nonparametric ANOVA. All calculations were performed using STATISTICA, version 6.0 (S t a t S o f t Inc. 2001). Results The production/growth model fits well our longitudinal growth data as obvious from the plot of the dimension-less mass ratio versus the dimension-less time variable (Fig. 1). The model explained 86 %, 85 %, 90 %, 92 % of total variance in E. macularius, C. elegans, C. mitratus, and C. brevis, respectively. As expected, species differ significantly in asymptotic mass (Table 1). The parameter a was similar in E. macularius, C. mitratus, and C. elegans. Confidence intervals (95%) of these values showed extensive overlap, except in the case of Fig. 1. A plot of dimension-less mass ratio, r = (m/M)1/4, versus the dimension-less time variable, τ = (at/4M1/4)-ln[1(m0/M)1/4], for four species of eublepharids. When plotted in this way, the production/growth model predicts that growth curves should fall on the same universal parameterless curve 1 – e –τ (shown as solid line). According to the model, parameter r represents the proportion of total lifetime metabolic power for maintenance and other activities. Table 1. The number of pseudoreplicates (n), the estimated values (mean ± SE) of the asymptotic mass in g (M) and the growth rate (a) for the examined species of eublepharid geckos. Species E. macularius C. mitratus C. elegans C. brevis n 279 192 120 33 a M 0.0289 ± 0.0012 0.0247 ± 0.0010 0.0220 ± 0.0014 0.0421 ± 0.0048 51.53 ± 1.81 10.06 ± 0.29 12.64 ± 0.59 2.35 ± 0.11 319 Fig. 2. Asymptotic body masses (M) estimated from individual growth curves in four species of eublepharid geckos. Explanations: Em = Eublepharis macularius, Cm = Coleonyx mitratus, Ce = C. elegans, Cb = C. brevis. Fig. 3. Growth rate parameters (a) estimated from individual growth curves in four species of eublepharid geckos. Abbreviations as in legend to Fig. 2. 320 E. macularius – C. elegans comparison. However, confidence intervals were systematically underestimated by inclusion of pseudoreplicates (multiple values obtained from each individual), and the comparison was not significant when this inflation was avoided (see below). On the contrary, although the data set concerning C. brevis was small, its parameter a was significantly higher than in the other species (Table 1). These observations are supported by non-parametric statistics applied on parameters estimated from individual growth trajectories of well-represented individuals. KruskalWallis ANOVA confirmed considerable interspecific differences in asymptotic mass (H = 28.12; P < 0.0001; Fig. 2). More remarkably, Kruskal-Wallis ANOVA found interspecific differences in growth parameter a (H = 11.70; P < 0.01; Fig. 3). This significance is clearly the result of the higher a in C. brevis (Kruskal-Wallis ANOVA testing for differences in a among the other three species: H = 4.35; P = 0.11). Discussion The production/growth model shows good accordance with our data in eublepharid geckos, and thus the applicability of this model to reptilian growth in non-fluctuating thermal environment was clearly demonstrated (Fig. 1). Lizards fit well universal growth curve (compare with Fig. 2 in W e s t et al. 2001). Estimated values of growth-rate parameter (a) for the examined lizards fall within the range (0.017 – 0.1) known in other ectotherm animals (several fish species, shrimp), but are substantially lower than those approximated for mammals and birds (0.21 – 1.9). This sharp distinction between ectotherms and endotherms clearly reflects both the differences in body temperature and in fundamental metabolic parameters of the cell across taxa (W e s t et al. 2001). A recently published paper (G i l l o o l y et al. 2001) shows that temperature-compensated resting metabolic rates of ectoterms are slightly lower than those of birds and mammals. It is important to note that the values of growth-rate parameter a for eublepharids are in accordance with the general prediction derived from the production/growth model. Although the geckos differed considerably in body mass, mass cannot explain the observed differences in their growth rate. Growth rates apparently follow the size-adjusted metabolic values. E. macularius, C. elegans and C. mitratus share simultaneously the similar metabolism allometry and growth parameter values. In contrast, C. brevis demonstrated an elevated level of metabolic scaling and a correspondingly higher a parameter. Cases such as this, in which closely related species show different metabolic scaling, may be very informative when evaluating theoretically derived growth relationships. Our study is presumably the first to demonstrate that reptilian species also follow the general production-growth model (at least in non-fluctuating thermal environment). Reptiles seem to be good model for getting a fundamental understanding of animal growth: unlike birds and mammals, neonate reptiles show similar metabolic rates to those of adults after correction for body mass differences (see review by N a g y 2000). Therefore, parameter a is more likely to be constant during ontogeny. Moreover, reptiles could be easily reared under different thermal conditions (both constant and fluctuating), which would be very informative. Such simple manipulation affects many parameters of the production/growth model, and it is not only the case of the mass-normalized metabolic rate (parameter B0). But also the mass of a single cell (mc) and the asymptotic maximum body mass (M) in ectotherms are affected by rearing temperature, 321 both latter generally increase at lower temperatures (A t k i n s o n 1994, V a n V o o r h i e s 1996). Manipulations of thermal environment in ectothermic animals and comparison between theoretical predictions and empirical results in future studies can help us to judge unequivocally the accuracy and universality of this new model. Acknowledgements We thank J. P o l e c h o v á , P. M i k u l o v á and two anonymous referees for their useful comments. This project was supported by Grant Agency of Charles University Project No. B-BIO-121/2001 and the Institutional Grant given by the Ministry of Education, Youth and Sports of the Czech Republic (No. J13-8113100004). L I T E R AT U R E ATKINSON D. 1994: Temperature and organism size – A biological law for ectotherms? Adv. Ecol. Res. 25: 1–58. CHARNOV E. L. 2001: Evolution of mammal life histories. Evol. Ecol. Res. 3: 521–535. DIAL B. E. & GRISMER L. L. 1992: A phylogenetic analysis of physiological character evolution in the lizard genus Coleonyx and its implications for historical biogeographic reconstruction. Syst. Biol. 41: 178–195. GILLOOLY J. F., BROWN J. H., WEST G. B., SAVAGE VAN M. & CHARNOV E. L. 2001: Effects of size and temperature on metabolic rate. Science 293: 2248–2251. HAILEY A. & COULSON I. M. 1999: The growth pattern of the African tortoise Geochelone pardalis and other chelonians. Can. J. Zool. 77: 181–193. LÓPEZ S., FRANCE J., GERRITS W. J. J., DHANOA M. S., HUMPHRIES D. J. & DIJKSTRA J. 2000: A generalized Michaelis-Menten equation for the analysis of growth. J. Anim. Sci. 78: 1816–1828. KRATOCHVÍL L. & FRYNTA D. 2002: Body size, male combat and the evolution of sexual dimorphism in eublepharid geckos (Squamata: Eublepharidae). Biol. J. Linn. Soc. 76: 303–314. MOSCARELLA R. A., BENADO M. & AGUILERA M. 2001: A comparative assessment of growth curves as estimators of male and female ontogeny in Oryzomys albigularis. J. Mammal. 82: 520–526. NAGY K. A. 2000: Energy costs of growth in neonate reptiles. Herpetol. Monogr. 14: 378–387. StatSoft, Inc. 2001: STATISTICA (data analysis software system), vers. 6. www.statsoft.com. VAN VOORHIES W. A. 1996: Bergmann size clines: a simple explanation for their occurrence in ectotherms. Evolution 50: 1259–1264. VON BERTALANFFY L. 1957: Quantitative laws for metabolism and growth. Q. Rev. Biol. 32: 217–231. WAN X., WANG M., WANG G. & ZHONG W. 2000: A new four-parameter, generalized logistic equation and its applications to mammalian somatic growth. Acta Theriol. 45: 145–153. WEST G. B., BROWN J. H. & ENQUIST B. J. 2001: A general model for ontogenetic growth. Nature 413: 628–631. 322 Folia Zool. – 52(3): 323–328 (2003) Reproductive biology of Capoeta tinca in Gelingüllü Reservoir, Turkey Fitnat Güler EKMEKÇİ and Saniye Cevher ÖZEREN Hacettepe University, Faculty of Science, Biology Department, Beytepe Campus, 06532 Ankara, Turkey; e-mail: [email protected] Received 5 November 2001; Accepted 5 June 2003 A b s t r a c t . The breeding and sexual maturation properties of a cyprinid fish Capoeta tinca were studied in Gelingüllü Reservoir, a recently impounded dam in Central Anatolia. Ripening of gonads commenced in early spring, whereas spawning occurred between May and June. Sexual maturity age was 2+ for males and 3+ for females. The results obtained from this study were compared with those of other populations of C. tinca in Turkey. Key words: Central Anatolia, spawning, sexual maturity age, gonadosomatic index Introduction The Cyprinid fish, Capoeta tinca, has a wide distribution in Anatolia, including the Sakarya, Kızılırmak and Çoruh river basins as well as some streams in the Marmara region that are hydrologically connected to the Black Sea. But it is not present in the Mediterranean Basin and in Thrace, the European part of Turkey (E r k ’ a k a n 1983). Capoeta tinca can adapt very easily to changes in water regime, it occurs both in lotic and lentic habitats so this species has economic value as a commercial fish from natural and man-made lakes. In the meantime, C. tinca has tolerance to salinity and pollution, and lives in streams with salinity of 10.5 g.l-1 and conductivity of 22000 µS.cm-1 (E k m e k ç i 2002). As an herbivorous fish, it is not a high value food like Esox lucius, Sander lucioperca and Silurus glanis. However, owing to its wide distribution and tolerance to different habitats, and its size (max. standard length=43cm, max weight=1178g; Y ı l m a z & G ü l 1999a), this species is consumed as food by people. We examined the breeding and sexual maturation properties up on which minimum fishing size is to be based for the sustainable management of Gelingüllü Reservoir, a recently impounded dam in Central Anatolia. There have been a few studies of C. tinca reproduction (A k g ü l 1986, 1988, Y ı l m a z & G ü l 1996, Y ı l m a z & G ü l 1999b), though only one in a reservoir (E k m e k ç i 1996). Whereas its growth properties are welldocumented (e.g. S o l a k 1982, E r k ’ a k a n & A k g ü l 1985, A k g ü l 1986,1988, C e n g i z l e r & E r d e m 1994, Y ı l m a z et al. 1996, E k m e k ç i 1996, 2002, Y ı l m a z & G ü l 1999a). Study Area, Material and Methods The Gelingüllü Reservoir is located to the southeast of the city of Yozgat, in the region of Central Anatolia (39°36’30’’N, 35°03’20’’E). The reservoir has an area of 2.4 km2 at the maximum water level. The dam was constructed on the Delice stream, a tributary of Kızılırmak River (Fig.1). The altitude of the reservoir is 1000 m. above sea level. Typical continental climate prevails in the study area, maximum and minimum air temperatures were 323 recorded as 37.1 °C (in July) and -23.7 °C (in January) respectively. Surface water temperature ranges between 22 °C (July) and 2 °C (December), whereas the dissolved oxygen content varies between 8.9 mg.l-1 (June) and 11.8 mg.l-1 (April). Water pH ranges between 7.47 and 9.39, and the conductivity ranges between 310 and 430 µS.cm-1. Secchi disc transparency is 1.11 m in May 1996 and 4.42 m in November 1996. Fig. 1. Map of the study area. Beside C. tinca, native fish species such as Capoeta capoeta sieboldi, chub (Leuciscus cephalus), nase (Chondrostoma regium), bleak (Alburnus orontis), barbel (Barbus tauricus), loach (Orthrias sp.) and introduced species such as goldfish (Carassius auratus), topmouth gudgeon (Pseudoraspora parva) and common carp (Cyprinus carpio) live in the Gelingüllü Reservoir. A total of 220 fish were sampled on a monthly basis between December 1995 to December 1996, except in January and February 1996, when the lake was frozen. In the second week of every month samples were collected with the aid of fishermen, using gill nets of 25 to 110 mm mesh size. During our field studies in Gelingüllü Reservoir, we had some doubts about the hybrids among C. tinca and C. capoeta. In this study, specimens suspected of being hybrids were excluded. 324 Fork length (FL) were measured with an accuracy of 1 cm and weighed to nearest 1 g in the field. Scales were used for age determination. They were firstly cleaned by 4% NaOH and dry mounts held between microscope slides as given in L a g l e r (1966). Males differ from females morphologically by the presence of breeding tubercules formed on the head during the spawning period. The sex of immature fish was confirmed by observation of gonads with the aid of a binocular. For sex determination of mature specimens, the gonads of fresh samples were observed macroscopically, and then the ovaries were removed and weighed. Body weights were determined to the nearest gram, where as the gonads of females were weighed with an accuracy of 0.1 g. and preserved in 4% formaldehyde for subsequent assignment of gonad development, 29 ovaries could not be weighed as they were less than 0.1g. Gonadosomatic index (GSI) was calculated as: GSI = Gonad weight x 100 / total body weight. a) b) Fig. 2 (a). Monthly variation of gonadosomatic index and (b) relative fecundity (eggs.g-1) for female Capoeta tinca. 325 Seasonal changes in GSI, egg diameter and number of eggs per 1 gram of ovary were used to elucidate the reproductive biology of this species. Egg diameter and number of eggs could be counted from 50 females whose ovaries contain visible eggs. Age at maturity was determined from fish collected in April and June based on gonad development. Fecundity was estimated from counts of yolky eggs in 1 g samples from three different parts (proximal, middle and distal) of a preserved ovary and multiplying the mean value by the total weight of both ovaries (B a g e n a l & B r a u m 1971). Mean egg diameter was calculated from the average of a total of 30 eggs, 10 from the proximal, middle and distal parts of the ovary, which were measured to the nearest 0.01 mm. The students’ t-test was used to test for the differences in egg number and diameter between different parts of ovary. Results and Discussion Age at maturity in Capoeta tinca of Gelingüllü Reservoir is at age 3. Only 7 % of females <3 years were immature, whereas the remaining females of ≥ 3 years were all mature. It is apparent that the females became mature at the age of 3. The males, on the other hand, become mature at age 2. Although age at maturity of both males and females in C. tinca populations of other river basins differs from those reported here, the one-year difference between males and females is characteristic of all populations (A k g ü l 1986, 1988, E k m e k ç i 1996, Y ı l m a z & G ü l 1996, Y ı l m a z & G ü l 1999b). The females were dominant in the population to the males by a ratio of 1:4.23 during the sampling period. In other populations of Capoeta tinca in Turkey, the ratio of males to females fluctuates between 0.64–1.15:1 (A k g ü l 1986, 1988, E k m e k ç i 1996, Y ı l m a z & G ü l 1996, Y ı l m a z & G ü l 1999b). Gonad weight and GSI in females peaked in May (Fig.2a) as did egg diameter (Fig.3), suggesting that eggs were laid between May and July. In July, all the ovaries were empty and only a few remaining eggs were observed. However, GSI values from C. tinca populations in Fig 3. Monthly variation of egg diameter for Capoeta tinca. 326 Fig. 4. Length-fecundity (Fa) relation for Capoeta tinca. Fig. 5. Relation between age and fecundity (Fa) for Capoeta tinca. other regions of Turkey indicate that reproduction starts in April and lasts until September, depending on the water temperature and altitude, with spawning periods in Anatolia such as, May to June (S o l a k 1982, Y ı l m a z 1994), May–July (E k m e k ç i 1996, Y ı l m a z & G ü l 1996), July to September (E r k ’ a k a n & A k g ü l 1985), and June–July (A k g ü l 1988). Altitude, climate and the ecological differences of stagnant and running water have great effects on the spawning period as stated by N i k o l s k i i (1963) and B e n n e t t (1970). The water temperature in the Gelingüllü Reservoir during the spawning period was about 19 °C. The recorded water temperatures during the spawning period of C. tinca vary between 21–22.5 °C (S o l a k 1982; E r k ’ a k a n & A k g ü l 1985, A k g ü l 1988, Y ı l m a z 1994, E k m e k ç i 1996, Y ı l m a z & G ü l 1996, Y ı l m a z & G ü l 1999b). Relative fecundity can be used to identify the start of the breeding season, because the number of eggs in a constant weight decreases when they become ripe. This occurred in May and June (Fig. 2b). There were no significant differences in means of the number of 327 eggs per g. and egg diameter between the parts of ovary (student’s t-test). The maximum mean egg diameter (1.88 mm) was measured in June. Y ı l m a z & G ü l (1996, 1999b) stated the maximum of mean egg diameter as 1.8 mm and 1.62 mm respectively. The mean number of eggs (absolute fecundity) in ovaries was significantly related to the mean fork length with (Fig. 4) (F=45.93 df=1,4 p<0.05). A significant relation (F=27.23, df=1,4 and p<0.05) was also established between age and the mean number of eggs in ovaries (Fig. 5). The maximum absolute fecundity was found as 58 896 eggs for a 7 year old female with a weight of 720g and 39.9cm total length (36.6cm fork length) caught in December 1995. Acknowledgements The authors extent their sincere thanks to two anonymous referees and G. H. C o p p for both linguistic and scientific contributions and Ali Celal H o ş for his assistance in field works. L I T E R AT U R E AKGÜL M. 1986: [Investigations on the bio-ecology of Capoeta tinca living in Kızılırmak Basin]. VIIIth National Biology Congress, İzmir: 599–61 (in Turkish). AKGÜL M. 1988: [An investigation on the growth, condition factor, spawning period of C. tinca (Heckel, 1843) living in Kelkit Stream]. IXth National Biology Congress, Sivas (in Turkish). BAGENAL T.B. & BRAUM E. 1971: Eggs and early life history. In Bagenal T.B. (ed.), Methods for Assessment of Fish Production in Fresh Waters. IBP Handbook no:3 Oxford: Blackwell Scientific Publications: 166–198. BENNET G. W. 1970: Management of lakes and ponds. Van Nostrant Reinhold Company. CENGİZLER I. & ERDEM Ü. 1994: [Growth of Barbus plebejus, Bonaparte, 1832, Capoeta tinca (Heckel, 1843) living in Almus Dam Lake]. XIIth National Biology Congress: 36–42 (in Turkish). EKMEKÇİ (ATALAY) F. G. 1996: Some growth and reproduction properties of Capoeta tinca (Heckel, 1843) living in Sarıyar Dam Lake (Ankara). Tr. J. of Zool. 20: 117–127 (in Turkish with English summary). EKMEKÇİ F. G. 2002: The effects of high salinity on the production of Capoeta tinca in a naturally contaminated river. Tr. J. Zool. 26: 265–270. ERK’AKAN F. 1983: The fishes of Thrace region. Hacettepe Bull. of Nat. Sci. and Eng. 12: 39–48. ERK’AKAN F. & AKGÜL M. 1985: Investigation of economical fish stock in Kızılırmak Basin. The Scientific and Technical Research Council of Turkey, Project Report No:VHAG-584, Ankara (in Turkish with English summary). LAGLER K. F. 1966: Freshwater Fishery Biology, W.M.C. Brown Company, Iowa. NIKOLSKII G. V. 1963: The ecology of fishes (Translated by L. Birkett). Academic Press, London. SOKAL R. R. & ROHLF F.J. 1997: Biometry: the principles and practice of statistics in biological research. W. H. Freeman and Company, New York. SOLAK K. 1982: Investigations on relations with biology and ecological parameters of Capoeta tinca species living in Çoruh and Aras Basin. Habiltation Thesis, Erzurum (in Turkish with English summary). YILMAZ M. 1994: Bio-ecological properties of carp (Cyprinus carpio L., 1758) and In Balığı (Capoeta tinca (Heckel, 1843) living in Kapulukaya Dam Lake (Kırıkkale). Gazi University, PhD Thesis (in Turkish with English summary). YILMAZ M. & GÜL A. 1996: The reproduction properties of In Balığı (Capoeta tinca (Heckel, 1843)) living in Kirmir Stream of Sakarya River (Ankara, Turkey). G.U. Gazi Eğitim Dergisi 4: 84–97 (in Turkish with English summary). YILMAZ M. & GÜL A. 1999a: Growth properties of Capoeta tinca (Heckel, 1843) living in Devres Stream of Kızılırmak River. G.U. Gazi Eğitim Dergisi 19 (1): 11–26 (in Turkish with English summary). YILMAZ M. & GÜL A. 1999b: The reproduction properties of In Balığı (Capoeta tinca (Heckel, 1843)) living in Devres Stream of Kızılırmak River. G.U. Gazi Eğitim Dergisi 19 (2): 57–72 (in Turkish with English summary). YILMAZ M., GÜL A. & SOLAK K. 1996: Investigation of some biological properties of Capoeta tinca (Heckel, 1843) living in Kirmir stream of Sakarya River. Tr. J. of Zool. 20: 177–187 (in Turkish with English summary). 328 Folia Zool. – 52(3): 329–336 (2003) The expansion and occurrence of the Amur sleeper (Perccottus glenii) in eastern Slovakia Ján KO·âO1, Stanislav LUSK2, Karel HALAâKA2 and Vûra LUSKOVÁ2 1 Department of Ecology, Faculty of Human and Natural Sciences, Univesity of Pre‰ov, 17. novembra 1, 081 16 Pre‰ov, Slovakia; e-mail: [email protected] 2 Institute of Vertebrate Biology, Academy of Sciences of the Czech Republic, Kvûtná 8, 603 65 Brno, Czech Republic; e-mail: [email protected] Received 14 April 2003; Accepted 21 July 2003 A b s t r a c t . The Amur sleeper, Perccottus glenii Dybowski, 1877, is indigenous in eastern Asia. During the second half of the 20th century, with the aid of man, it spread over the eastern part of Europe as well as in central Asia. In the course of 50 years of its dispersal in the western direction the species already reached the Vistula drainage area (the Baltic Sea basin) and the Danube drainage area (the Black Sea basin). In the latter basin, its occurrence was ascertained in the drainage area of the Tisza river in Hungary in 1997. In eastern Slovakia, the Amur sleeper was first recorded in 1998 in the Latorica drainage area. In the course of subsequent years it has become a common species in the streams in the basins of the Latorica, Bodrog and Tisza rivers. In shallow lentic waters densely grown with aquatic plants the species becomes a superdominant or even exclusive species in the local fish communities. It has no marketable value but presents a serious threat to the existence of native fish species with similar identical microhabitat requirements. Key words: Perccottus glenii, exotic species, invasion, Tisza river basin Introduction In an absolute majority of cases, both intentional and unintentional introduction or invasion of an exotic fish species causes a risky contamination of indigenous ichthyofauna. As a rule, the presence of one or several exotic taxa exerts a permanent, direct or indirect, negative influence on the native species, as has been widely demonstrated for fishes (A l l e n d o r f 1991, R o s s 1991, E f f o r d et al. 1997, C r i v e l l i 1995, H o l ã í k 1991). In most cases, the introduction of an exotic species will not remain limited to the initial hydrological area into which it had been introduced. Often the species will gradually spread uncontrolled over a large territory. Any attempt at preventing the exotic element from spreading or even removing it from the biota it had invaded has failed. After the non-indigenous form, Carassius auratus, had expanded over the Danube river basin in the second part of the 20th century (H o l ã í k & Î i t À a n 1978, L u s k et al. 1998), we are now witnessing the expansion of several species of the genus Neogobius (A h n e l t et al. 1998, Z w e i m ü l l e r et al. 1996, K a u t m a n 2001, S t r á À a i & A n d r e j i 2001, H o l ã í k 2002, etc.). In recent years, Ictalurus melas (K o ‰ ã o & K o ‰ u t h 2002) and Perccottus glenii Dybowski, 1877, a species hitherto unknown in central Europe, have recently occurred and spread in eastern Slovakia (the Tisza drainage area). Perccottus glenii is indigenous to the Russian Far East, north-eastern China, and the northern part of the Korean Peninsula (B e r g 1949, N i k o l s k y 1956, K i r p i c h n i k o v 329 1945, E l o v e n k o 1981, B o g u t s k a y a & N a s e k a 2002). In the course of the 20th century, two introductions into the European part of Russia were recorded. The first introduction took place in St. Petersburg in 1912. There, after having been temporarily kept in aquaria, the species was released into small ponds and occurred in free waters in their environs. Subsequently, the species gradually spread over the drainage area of the Gulf of Finland (D m i t r i e v 1971, P a n o v et al. 1999). The second introduction took place in Moscow in 1948 when the species was imported by the participants of the Amur Expedition. After having been kept in aquaria and having become popular among the aquarium keepers, the Amur sleeper was released into nature. Then it began spreading over the hydrological system of the Moscow river and the upper part of the Volga river basin (S p a n o v s k a y a et al. 1964). Apparently, this second introduction was the onset of a gradual occupation of the range in the European part of Russia and its dispersal westwards. Its further spread over both the Asian part of Russia, the Baikal area, Kazakhstan, Uzbekistan, Turkmenistan and the eastern part of Europe and, in recent years, even the White Sea and the Arctic Ocean basins is connected with transports of stocking materials of various fish species in which the Amur sleeper was an undesired admixture (L i t v i n o v & O ’ G o r m a n 1996, E l o v e n k o 1981, B o g u t s k a y a & N a s e k a 2002). Around 1980 the European range of the Amur sleeper was limited to the environs of St. Petersburg, the drainage area of the Oka river (Moscow region) and the middle part of the Volga river basin (E l o v e n k o 1981). At present the species is dispersed over practically the whole European part of the former Soviet Union, and it is spreading westwards. In Poland the Amur sleeper was first recorded in the Vistula river (the Baltic Sea basin) in 1993 (A n t y c h o w i c z 1994). Over of subsequent years it has spread into the middle reaches of the Vistula, including its floodplains and tributaries (T e r l e c k i & P a l k a 1999). K o z l o v (1993) reports the species from the Don river, B o g u t s k a y a & N a s e k a (2002) from the Dnieper river. K o r t e et al. (1999) found the Amur sleeper to be a dominant fish species in the upper reaches of the Dniester river basin in Carpathian Ukraine. The first finds of the species in the Danube river system come from the Tisza river basin in Hungary (H a r k a 1998), in Carpathian Ukraine from the Latorica drainage area (M o s h u & G u z u n 2002), and in Slovakia from the basins of the Latorica and Bodrog rivers (K o ‰ ã o et al. 1999, K a u t m a n 1999, our own data). Material and Methods In 1999–2002 years the occurrence of the Amur sleeper was studied by means of electrofishing, using a gear producing pulsating electrical current, 170–250 V, 0.5–3.5 A. The fish caught in the localities under study were determined down to species and released. For later examinations, samples of fish were preserved in 5 % formaldehyde (food, growth), 80 % ethyl alcohol (growth, biometrics), frozen (genetic analyses) or kept live (karyology analysis). The species diversity index (H’) was calculated according S h a n n o n & W e a v e r (1949) using binary logarithms. The documentary material is deposited in the collections of the Department of Ecology, Pre‰ov University, Slovakia. Results and Discussion In Slovakia the Amur sleeper was first found for in August 1998 in a pool near Kamenná Moºva in the floodplain of the Latorica river (K o ‰ ã o et al. 1999). K a u t m a n (1999) 330 found the species in April 1999 in the floodplain of the Latorica river, viz., in the Vel’ké plytãiny and Brestovisko channels near the village of BoÈany, and in the Lelesk˘ channel near the village of Leles. In July 1999, we caught the Amur sleeper in several places in the floodplain of the Latorica river. During our subsequent investigations carried out in August 1999, we found the species even in the dead river branches of the Bodrog river. Fig. 1. Localities of evidenced occurrence (■) of Perccottus glenii in the Latorica, Bodrog and Tisza drainage areas, eastern Slovakia. Intense investigations implemented in 2001 and 2002 in the Bodrog, Latorica and Tisza drainage areas, eastern Slovakia, have shown that the Amur sleeper had become a common species in the hydrological systems of those rivers (Fig. 1). In shallow lentic waters of various habitats (gravel pits, channels, backwaters), densely grown with aquatic vegetation (Stratiotes aloides, Myriophyllum spicatum, Ceratophyllum demersum, Elodea canadensis, Nuphar luteum, Typha latifolia), the species was dominant in the local fish communities in a number of places. Such microhabitat types harbour minimum numbers of fish species (Esox lucius, Perca fluviatilis) reported as important predators of the Amur sleeper (B o g u t s k a y a & N a s e k a 2002). This is fully confirmed by our own observations: in localities with the greatest abundance of the Amur sleeper we would usually find only occasional individuals of E. lucius, less than 150 mm in standard length. The share of the Amur sleeper in samples of fish taken in the localities under study increases significantly with increasing surface of the habitat grown with aquatic vegetation. Habitats with high dominance of the Amur sleeper show rather low species richness and low species diversity (H’) of the fish community (Table 1). The representation of the Amur sleeper in the fish community in a locality is significantly dependent upon the degree to which it is grown with aquatic vegetation (R=0.97, F=177.7, p<0). Likewise, the number of fish species present decreases significantly with increasing grown-up area (R=0.86, F=34.5, p < 8-5), the same as their species diversity H’ (R=0.87, F= 36.2, p< 6-5). The number of fish species in the locality examined decreases with increasing share of the Amur sleeper (R=0.81, F=22.1, p<51-4). 331 Table 1. Occurrence of the Amur sleeper in various localities of eastern Slovakia in 2001–2002. Locality (year) Habitat type Grown-up Species surface richness (%) No. fish captured (n) Share of P. glenii (%) H’ E Svätá Mária (2002) channel backwater 80 % 5 469 95.3 0.35 0.06 Karãa-V. Kamenec (2001) river arm 0.5% 17 363 4.1 3.18 0.78 Karãa-V. Kamenec (2002) river arm 0.5% 16 484 0.6 2.96 0.74 Hranice-Streda n.B. (2001) channel 35% 12 363 19,0 2.63 0.73 Hranice-Streda n.B. (2002) channel 40% 11 153 39.2 2.55 0.74 Somotor (2001) channel 15 % 10 229 0.9 2.27 0.68 Somotor (2002) channel 15 % 17 284 2.5 3.52 0.86 Kamenná Moºva (2002) channel 70 % 9 74 71.6 1.60 0.505 Kamenná Moºva (2002) 3 gravel pits 5% 17 677 0.9 2.36 0.59 Tisa-M.Trakany (2001) lakes 10 % 14 197 2.5 2.74 0.68 Tice – Leles (2002) old backwater 95% 2 205 82.9 0.66 0.66 KapoÀa (2001) backwater 10 % 9 93 2.1 2.35 0.74 KapoÀa (2001) gravel pit 45 % 10 147 51,0 2.15 0.65 N.Vieska-Somotor (2001) old backwater 40% 9 295 33.9 2.03 0.64 The highly probable source of the Amur sleeper in eastern Slovakia is the upper part of the Latorica drainage area, lying in the territory of Ukraine. The occurrence of this species in the Latorica basin near the town of Chop, Carpathian Ukraine, is reported in an editorial attached to the article by M o s h u & G u z u n (2002) as well as by L i t v i n c h u k & B o r k i n (2002). Likewise, K a u t m a n (1999) refers to oral information on the occurrence of the Amur sleeper in the same region. Considering the ecological characteristics of this species, which is not a good swimmer, it may be expected that it will spread mainly from localities lying upstream to those lying in lower parts of the river basin. Usually, the occurrence of a new species will be observed with some time delay from its first penetration into the drainage area, the species having already reproduced and occurring in greater numbers. The rather abundant occurrence of this species which, moreover, shows a widespread occurrence in the Latorica and Bodrog drainage areas, as demonstrated in 1999 and 2000, would suggest that it had invaded the Latorica drainage area prior to 1998. One may assume that it was the 1998 floods that contributed to the rapid occupation of the 332 wide areas of the Latorica and Bodrog drainage areas. In that way the species got even to the original floodplain lying behind the levees. Its rapid spread over the Latorica, Bodrog and Tisza floodplains was also facilitated by the local amelioration channel system. General observations on a rapid spread of this species over the Vistula drainage area have been reported by T e r l e c k i & P a l k a (1999), and over the Volga and Don by B o g u t s k a y a & N a s e k a (2002). The later spread of the species over the Tisza river basin shows an identical course, the common occurrence of which has also been reported from Serbia, besides that in Hungary (G e r g e l y 2002, H a r k a & F a r k a s 2001). The occupation of a river basin is significantly accelerated by high water levels and the connected floods, as also reported by E l o v e n k o (1981) and as was the case of the Bodrog and Tisza drainage areas in eastern Slovakia. According to our observations, the Amur sleeper in the Tisza riverine system, eastern Slovakia, is currently confined, to the Latorica, Bodrog and Tisza drainage areas, including their floodplains and the connected amelioration channel systems (Fig. 1). So far, we have not ascertained the presence of this species in the lower reaches of the Ondava and Laborec rivers (tributaries to the Bodrog river). This fact also tends to support the hypothesis that the species will spread chiefly in a downstream direction within a river drainage area. With regard to the amelioration channel network in the original floodplains along the lower reaches of the Laborec and Ondava rivers, one may expect that the Amur sleeper may penetrate even those areas in the next few years. The causes why the Amur sleeper is spreading beyond its range are unambiguously connected with man’s activities. At first, there were intentional introductions of this interesting species (St. Petersburg 1912, Moscow 1948). Some of the introductions were connected with utilising the Amur sleeper as a predator of mosquito larvae. A considerable number of introductions were unintentional, the Amur sleeper being introduced together with stocking materials of other fish species as their undesired admixture. This consideration even includes the transport and release of this species by amateur fishermen, and one cannot overlook even the probable releases of this species from aquarium cultures. At any rate, the fact that the species has overcome the barriers between river basins is connected with man’s activities (H a r k a & S a l l a i 1999). For a review of literature on the various aspects of dispersal of the Amur sleeper, see B o g u t s k a y a & N a s e k a (2002). What will be the further occurrence of this species in Slovakia? It may be supposed that through natural migration the Amur sleeper will gradually invade the drainage areas of the lower reaches of the Ondava and Laborec rivers. Its further spread to additional streams in the Tisza river basin (the Slaná and Hornád rivers) and further on, eventually to the Danube basin (western Slovakia) through natural migration, is problematical. A much greater probability is seen in the species spreading through unintentional introduction with the stocking material of other fish species, as was the case in the past, e.g. with the dispersal of Carassius auratus or Pseudorasbora parva. Nor can a transfer by intentional activity of fishermen be ruled out (bait fish). The Amur sleeper is also suitable for aquarium culture (M a c h l i n 1957, S c h e n k e & G r a m b o w 1965) and therefore one cannot exclude the possibility of the species being released into nature from aquarium cultures. The rapid colonization of the Latorica, Bodrog and Tisza river basins in the territory of Slovakia by the Amur Sleeper in the course of the past 4–5 years was invasive in character. The Amur Sleeper shows several biological and ecological properties that are typical of the so-called invasive fish species. Its reproduction capacity is high (portional spawning, the male 333 guarding the spawned eggs), it is resistant to high water eutrophication, including lack of oxygen. It is even capable of surviving when frozen in ice, or near drying up. It ingests animal food of all kinds, including smaller fish and amphibian larvae (K i r p i c h n i k o v 1945, S o k o l o v 2001, B o g u t s k a y a & N a s e k a 2002). Amur sleeper is of no positive economic importance. It attains a maximum of 250 mm in total length and up to 250 g in weight but the vast majority of populations consists of individuals less than 120 mm in total length (B e r g 1949, R e s h e t n i k o v 2000). In some places it is the object of sport angling (D m i t r i e v 1971, V e r i n 1978, S h l y a p k i n & T i k h o n o v 2001, N a u m e n k o 2002). The initial intention to utilise this species in the control of mosquitoes did not bring any significant results (K i r p i c h n i k o v 1945). The Amur sleeper presents a serious threat to the existence of a number of native fish species that show identical microhabitat requirements (stagnant waters densely grown with aquatic plants). As regards food, the Amur sleeper is characterised as a potential predator (S p a n o v s k a y a et al. 1964). The diet of the Amur sleeper consists of aquatic invertebrates of all sizes, besides larvae and smaller-sized fish and even amphibians. The predation of the Amur sleeper on other fish species, including individuals of its own species as well as fish eggs, is described on a general level by a number of authors (K i r p i c h n i k o v 1945, N i k o l s k y 1956). S h l y a p k i n & T i k h o n o v (2001) reported on a total disappearance of Leucaspius delineatus by this species in smaller reservoirs. Likewise, introduced into reservoirs harbouring a single-species fish stock of Carassius carassius, the Amur sleeper liquidated all individuals smaller than 40 mm in size (R e s h e t n i k o v 2000). In the diet of Amur sleepers older than one year, living in the Baikal Lake basin as well as in the Volgograd Dam Lake, fish eggs and small-sized fish formed a significant share (L i t v i n o v & O ’ G o r m a n 1996, N a u m e n k o 2002). Likewise, the trophic competition of the Amur sleeper, considering its voraciousness and capability to produce very numerous populations, is a negative aspect in relation to native fish species (L i t v i n o v & O ’ G o r m a n 1996). Here it is necessary to point out that a diet composition similar to that of the Amur sleeper is also found in Umbra krameri (L i b o s v á r s k ˘ & K u x 1998). For this species showing similar microhabitat requirements, the Amur sleeper is a deadly threat even as a predator. The devastation action of the Amur sleeper has been demonstrated even against amphibian larvae and adults (M a n t e i f e l & B a s t a k o v 1986, R e s h e t n i k o v 2000, 2001, K a u t m a n 1999). We have observed that being captured, individual Amur sleepers regurgitated swallowed individuals of their own species up to 40 mm in size. In aquarium environment, individual Amur sleepers 70–90 mm in total length were observed to prey on goldfish up to 45 mm in total length. For some fish species native to central Europe (Umbra krameri, Leucaspius delineatus, Carassius carassius, Rhodeus amarus and the larvae of other species), the Amur sleeper presents a real threat both as regards trophic competition and as a predator. Pains should be taken to prevent incidental transport of the Amur sleeper with stocking materials of other fish species to further river basins. Acknowledgements The investigations were implemented and financially supported within the research grant No. 1/9209/02 of the Grant Agency VEGA of the Slovak Academy of Sciences and the Ministry of Education of the Slovak Republic, and the research project reg. No. S6093007 of the Academy of Sciences of the Czech Republic. 334 L I T E R AT U R E ANTYCHOWICZ J. 1994: [Perccottus glehni in our waters]. Komunikaty rybackie 2: 21–22 (in Polish). AHNELT H., BĂNĂRESCU P., SPOLWIND R., HARKA Á. & WAIDBACHER H. 1998: Occurrence and distribution of three gobiid species (Pisces: Gobiidae) in the middle and upper Danube region – examples of different dispersal patterns? Biologia, Bratislava 53: 665–678. ALLENDORF F.W. 1991: Ecological and genetic effects of fish introductions: synthesis and recommendations. Can. J. Fish.Aquat. Sci. 48 (Suppl. 1): 178–181. BERG L.S.1949: [Freshwater Fishes of the U. S. S. R. and adjacent countries]. Akad. Nauk SSSR, Moscow – Leningrad, III.: 929–1 382 (in Russian). BOGUTSKAYA N. G. & NASEKA A. M. 2002: Perccottus glenii Dybowski, 1877. Freshwater Fishes of Russia, Zoological Institute RAS. INTERNET: http: //www.zin.ru/Animalia/Pisces/eng/taxbase e/species_e/perccottus_el.htm. CRIVELLI A.J. 1995: Are fish introductions a threat to endemic freshwater fishes in the northern Mediterranean Region? Biological Conservation 72: 311–319. DMITRIEV M. A. 1971: [Look out, Rotan]. Rybovodstvo i rybolovstvo 1: 26–27 (in Russian). EFFORD I. E., GARCIA C. M. & WILLIAMS J. D. 1997: Facing the challenges of invasive alien species in North America. Global Biodiversity 7: 25–30. ELOVENKO V.N. 1981: Taxonomic status and geographical distribution of the fishes (Gobioidei, Perciformes, Eleotridae) introduced in water bodies of the European part of the USSR, Kazakhstan and Middle Asia. Zoologicheski Zhurnal 60 (10): 1517–1521 (in Russian with English summary). GERGELY J. 2002: [We introduce the ichthyologist Akos Harka]. Magyar Szó (Novi Sad), 2002, október 30. szám 14. oldal. (in Hungarian). HARKA Á. 1998: New fish species in the fauna of Hungary: Perccottus glehni Dybowski, 1877]. Halászat 91: 32–33 (in Hungarian with English summary). HARKA Á. & FARKAS J. 2001: Situationsbericht über die Fischfauna der Theiss. Österreichs Fischerei 54: 122–128. HARKA Á. & SALLAI Z. 1999: Morphological features of Amur Sleeper (Perccottus glehni Dybowski, 1877), its habitat and spreading in Hungary. Halászat 92: 33–36 (in Hungarian with English summary). HOLâÍK J. 1991: Fish introductions in Europe with particular reference to its central and eastern part. Can. J. Fish. Aquat. Sci., 48 (Suppl. 1): 13–23. HOLâÍK J. 2002: Our gobies and how to distinguish them. Biodiversity of Fishes in the Czech Republic (IV), Brno: 73–78 (in Slovak with English summary). HOLâÍK J. & ÎIT≈AN R. 1978: On the expansion and origin of Carassius auratus on Czechoslovakia. Folia Zool. 27: 279–288. KAUTMAN J. 1999: [Perccottus glenii Dybowski, 1877 from East Slovakian water bodies]. Chránené územia Slovenska, SAÎP Banská Bystrica, 40: 20–22 (in Slovak). KAUTMAN J. 2001: The first occurrence of Neogobius gymnotrachelus in the Slovak Danube. Folia Zool. 50: 79–80. KIRPICHNIKOV V.S. 1945: [Biology of Percottus glehni Dyb. (Eleotridae) and possibilities of its utilization in the control of encephalitis and malaria]. Byull. MOIP 50 (5-6): 14–27 (in Russian). KORTE E., LESNIK V., LELEK A. & SONDERMANN W. 1999: Impact of overexploitation on fish community structure in upper River Dniester (Ukraine). Folia Zool. 48: 137–142. KO·âO J. & KO·UTH P. 2002: On the occurrence of Perccottus glenii Dybowski, 1877 and Ictalurus melas Rafinesque, 1820 from the East Slovakia. Biodiversity of Fishes in the Czech Republic (IV): 105–108 (in Slovak with English summary). KO·âO J., KO·UTH P. & HRTAN E. 1999: [Further new fish element of fishes in Slovakia: Amur sleeper]. Poºovníctvo a rybárstvo 51: 33 (in Slovak). KOZLOV V. I. 1993: [Ecological prediction of ichthyofauna of fresh waters (by the example of Ponto-Caspian Region)]. Moskva, VNIRO, 252 pp. (in Russian). LIBOSVÁRSK¯ J. & KUX Z. 1958: Pfiíspûvek k poznání bionomie a potravy blatÀáka tmavého Umbra krameri krameri (Walbaum). (Contribution to the knowledge of bionomy and food in Umbra krameri krameri (Walbaum)). Zool. listy 7: 235–248 (in Czech with German summary). LITVINCHUK S. N. & BORKIN L. YA. 2002: [Distribution, ecology and conservation status of Triturus dobrogicus (Amphibia, Salamandridae) in the Ukraine and Moldovia]. Vestnik zologi, Kiev 36: 35–44 (in Russian). LITVINOV A.G. & O’GORMAN R. 1996: Biology of Amur sleeper (Perccottus glehni) in the delta of the Selenga River, Buryatia, Russia. J. Great Lakes Res. 22 (2): 370–378. 335 LUSK S., LUSKOVÁ V. & HALAâKA K. 1998: Prussian carp – 25 years since its natural introduction. Proc. 3th Czech conference of ichthyology, VodÀany: 135–140 (in Czech with English summary). MACHLIN M. 1957: Der Amur-Eleotridae. Aquarien und Terrarien 4 (9): 291–293. MANTEIFEL Y.B. & BASTAKOV V.A. 1986: Percottus glehni Dybowski – a new colonizer in the ichthyofauna of Lake Glubokoe. Hydrobiologia 141: 133–134. MOSHU A. YA. & GUZUN A. A. 2002: [The first discovery of rotan Perccottus glenii (Perciformes, Odontobutidae) in Dniester River]. Vestnik zoologii 36: 98 (in Russian). NAUMENKO A. N. 2002: [New species in the ichthyofauna structure and of higher Crustacea in the Volgogradskoye reservoir]. In: Rybokhozyastvennye issledovaniya v bassejne Volgo-Donskogo mezhdurechiya v sovremennom etape. GosNIORKH, S.Peterburg: 184–190 (in Russian). NIKOLSKY G. V. 1956: [Fishes of Amur River basin. Results of Amur ichthyological expedition of 1944–1949]. Moskva, 551 pp. (in Russian). PANOV V., LEPPAKOSKI E. & OJAVEER H. 1999: Introduction of alien species into the Gulf of Finland – an increasing environmental problem. In: Panov V., Dianov M. & Lobanov A. (eds), Regional Biological Invasions Center. INTERNET: http://www.zin.ru(projects/invasions/gaas/aliens 3r. htm. RESHETNIKOV A.N. 2000: Species diversity of aquatic animals (Invertebrates, Amphibians and Fishes) in ponds and the effect of predation by the introduced fish rotan (Perccottus glenii). Biodiversity and dynamics of ecosystems in North Eurasia, Novosibirsk 3 (2): 279–281. RESHETNIKOV A.N. 2001: Influence of introduced fish Perccottus glenii (Odontobutidae) on amphibians in small waterbodies of Moscow region. Zhurnal obshchej biologii 62: 352–361 (in Russian with English summary). ROSS S.T. 1991: Mechanisms structuring stream fish assemblages: are there lessons from introduced species? Environ. Biol. Fish. 30: 359–368. SHANNON C. E. & WEAVER W. 1949: The mathematical theory of communication. Univ. Illinois Press, Urbana. SCHENKE G. & GRAMBOW A. 1965: Deutsche Erstzucht der Amurgrundel. Aquarien-Terrarien 12 (9): 292–295. SHLYAPKIN I.V. & TIKHONOV S.V. 2001: Distribution and biological features of Amur sleeper Perccottus glenii Dybowski in the upper Volga reservoirs. U.S.-Russia invasive species workshop Borok, Book of Abstracts: 201–202. SPANOVSKAYA V.D., SAVVAITOVA K.A. & POTAPOVA T.L. 1964: [Variation of rotan (Percottus glehni Dyb., fam. Eleotridae) in acclimatization]. Voprosy Ikhtiologii 4 (4): 632–643 (in Russian). SOKOLOV L.I. 2001: [On wintering of rotan in Amur water bodies]. Voprosy Ikhtiologii 41 (4): 572–573 (in Russian). STRÁ≈AI I. & ANDREJI J. 2001: [Sand goby – (till now) last invasive gobiid species]. Poºovníctvo a rybárstvo 53: 44–45 (in Slovak). TERLECKI J. & PALKA R. 1999: Occurrence of Perccottus glenii Dybowski 1877 (Perciformes, Odontobutidae) in the middle stretch of the Vistula river, Poland. Archieves of Polish Fisheries 7 (1): 141–150. VERIN I. 1978: [ I advocate rotans]. Rybovodstvo i rybolovstvo 5: 26 (in Russian). ZWEIMÜLLER I., MOIDL S. & NIMMERVOLL H. 1996: A new species for the Austrian Danube - Neogobius kessleri. Acta Univ. Carol. Biol. 40: 213–218. 336 instrukce.jpg (660x982x24b jpeg)