Microtus brandti, and Me. unguiculatus (P<0.05); S. daurica had the longest stomach and Me.

unguiculatus had the shortest one, but no differences were found between S. daurica, C. barabensis, and C. triton (P>0.05), and between P. robovskii, M. brandti, and Me. unguiculatus (P>0.05). S. daurica, M. brandti, and C. triton had longer large intestines than did C. barabensis, P. robovskii and Me. unguiculatus (P<0.05). M. brandti had the longest large intestine and P. robovskii had the shortest one, but no significant differences were found between S. daurica, M. brandti, and C. triton (P>0.05), and between C. barabensis, P. robovskii and Me. unguiculatus(P>0.05). The cecum was also the longest in M. brandti and the shortest in P. robovskii. No differences were found between S. daurica, P. robovskii, Me. unguiculatus, and C. triton (P>0.05).

Discussion
Of the six rodent species studied M. brandti had the longest caecum and large intestine, and the length of their small intestine does not reflect the fiber content of their diets. In some rodent species, the estimated volume and surface area of the caecum and colon were greater in the herbivorous species than that in the granivorous species (B a r r y 1977). And food habit may be closely related to gut morphology and structure (W i l c z y n s k a 1998, 1999). S c h i e c k & M i l l a r (1985) compared the digestive tract morphology for 35 species of rodents and also found that the masses and lengths were greater in herbivores than in granivores, and stated that small intestine lengths in small mammals did not reflect the amount of fiber in the diet of each species. But significant differences were not found in measurements of the small intestine in white-footed mouse (Peromyscus leucopus) - an omnivorous species, and meadow vole (Microtus pennsylvanicus) - a herbivorous species, the hind gut measurements revealed diet-specific anatomical differences (D e r t i n g & N o a k e s 1995). Generally, because of their high quality and nonfibrous diets, omnivorous and granivorous species do not depend upon post-gastric fermentation chambers for the storage of digesta and for nutrient extraction; the protein and easily digestible carbohydrates can be processed and the nutrients absorbed in the foregut, and consequently, it can feed intermittently using its relatively large stomach as a storage chamber for food and still meet its daily energy needs (D e r t i n g & N o a k e s 1995). In the present study, the lengths of stomach were longer in S. daurica and C. triton than in other species. The herbivorous M. brandti has similar stomach length to the granivorous-herbivorous Me. unguiculatus and granivorousomnivorous P. robovskii . This is somewhat not in consistent with the general trend. However, D e r t i n g & N o a k e s (1995) found that P. leucopus has a longer stomach than M. pennsylvanicus. In contrast to granivorous and omnivorous species, herbivorous species rely upon a more continual flow of their low quality, highly fibrous diet throughout the anterior portion of the gut prior to retention, fermentation, and assimilation of nutrients in the cecum and colon (D e m m e n t & V a n S o e s t 1985, H u m e 1994). Therefore, with a more continual feeding pattern and the use of postgasric fermentation, the capacity of the caecum is likely to be of greater importance than that of other gut organs in herbivorous rodent (D e r t i n g & N o a k e s 1995). Larger hind guts were associated with increased capacity of those organs that contribute the most to nutrient and energy extraction from a particular diet type. The structure of the gastrointestinal tract (GI) is fairly homogeneous among different orders of mammals (C h i v e r s & H l a d i k 1980), and the development of different parts of the GI generally reflect adaptations to different foods (B r u o r t o n & P e r r i n 1991,
53

W i l c z y n s k a 1998, 1999). V o r o n t s o v (1962) argued that the evolutionary transitional trends of gut morphology for rodent species indicated the changes from feeding on energy-dense but hard to get food, such as invertebrates and seeds, to less energy-dense but easier-to-get food in the form of the vegetative parts of plants. E l l i s et al. (1994) extended Vorontsovs hypothesis and stated that it could be predicted that diet and alimentary tract morphology are reflected by their life histories of animals. Species that eat high-energy foods and display digestive tract adaptations to a high quality diet will possess life-history traits similar to those of relatively opportunistic species such as high motional ability, high fecundity, and the ability to exploit temporarily ideal but relatively unstable habitats, whereas those with adaptations to a herbivorous diet should possess the converse life history traits. In our study, the only strict herbivore M. brandti is diurnal and living in family groups. Their habitats are relative stable and, however, they have a relative high fecundity and high activities, which is not consistent with the prediction. For the other five species, P. robovskii is nocturnal, mainly feeding on plant seeds and invertebrates, but they live in relative stable habitat and have a relative low fecundity and activity. S. daurica, which is diurnal and mainly feeding on grass and plus a little insects, live in a relative stable habit and have relative high fecundity and high activity, which is in agreement with the prediction. C. barabensis, C. triton, and Me. unguiculatus have similar food habits to P. robovskii. They all have relative high fecundity and high activities, and C. barabensis lives in a relative stable habitats (not consist with prediction) whereas C. triton (nocturnal) and Me. unguiculatus (diurnal) live in relative unstable habitats (consist with the prediction). In conclusion: 1) the strict herbivore, M.brandti, showed the longest large intestine and ceacum, which is consistent with our prediction; 2) the small intestine does not reflect the dietary fiber contents and thus not to be a good indicator for food habits; 3) hind gut is more important for herbivorous than for omnivorous rodents, and could be a relative reliable indicator for food habits; 4) it seems that there is no direct cause and effect relationship between some life history traits and gut morphology for rodent species in this study. It should be noted that the gut sizes of rodents change with season (food quality and temperature), reproductive status, and other internal and /or external factors. These conclusions are only limited to the summer season.
Appendix. The absolute mean measurements of digestive tract for six species of rodents (SD). Species Stomach (cm) (cm) 2.50.4 5.10.5 4.60.4 3.20.5 2.30.2 2.20.2 Small intestine (cm) 25.72.7 78.36.5 49.84.3 32.05.7 34.35.1 21.30.4 Large intestine (cm) 21.72.3 24.73.2 22.61.7 12.61.4 13.22.9 9.50.0 Cecum (cm) 12.11.7 8.21.3 7.21.1 5.60.5 4.90.6 2.70.7 Total GI (cm) 62.05.2 116.39.2 84.24.9 53.45.8 54.78.2 35.61.3

Microtus brandti Spermophilus daurica Cricetulus triton Cricetulus barabensis Meriones unguiculatus Phodopus robovskii

Acknowledgements
We are grateful to Y. W a n g and G. W a n g for their kind help for field work. This study was supported in part by the National Natural Science Foundation of China (No.39730090, No.39770122 and No.39970128), the Chinese Academy of Sciences (No. KSCX2-SW-103 and No. STZ-01-06), the CAS innovation program, and the Ministry of Science and Technology of China (No.FS2000-009)

54

L I T E R AT U R E
BAO W.D. 1998: [Comparative study on the energy and water metabolism of four sympatric small mammals in Ordos semi-arid environment]. Ph.D. Thesis, Institute of Zoology, Chinese Academy of Sciences, Beijing (in Chinese). BARRY R. E. 1977: Length and absorptive surface area apportionment of segments of the hidgut for eight species of small mammals. J. Mamm. 57: 273290. BRUORTON M. R. & PERRIN M. R. 1991: Comparative gut morphometrics of vervet (Cercopithecus aethiops) and samango (C.mitis erythrarchus) monkeys. Z. Sugetierkd. 56: 6571. CHIVERS D.M. & HLADIK C. M. 1980: Morphology of the gastrointestinal tract in primates: comparisons with other mammals in relation to diet. J. Morphol. 116: 337386. DEMMENT M. W. & VON SOEST P. J. 1985: A nutritional explanation for body size patterns of ruminant and nonruminant herbivores. Amer. Nat. 125: 641672. DERTING T.L. & NOAKES III.E.B. 1995: Seasonal changes in gut capacity in the white-footed mouse (Peromyscus leucopus) and meadow vole (Microtus pennsylvanicus). Can. J. Zool. 73: 243252. DERTING T. L. & BOGUE B.A. 1993: Responses of the gut to moderate energy demands in a small herbivore (Microtus pennsylvanicus). J. Mamm. 74: 5968. ELLIS B.A., MILLS J.N., KENNEDY E.J.T, MAIZTEGUI J.I. & CHILDS E. 1994: The relationship among diet, alimentary tract morphology, and life history for five species of rodents from the central Argentine pampa. Acta Theriol. 39: 345355. HUME I.D. 1994: Gut morphology, body size and digestive performance in rodents. In: Chivers D.J. & Langer P. (eds), The digestive system in mammals: food, form and function. Cambridge University Press, New York: 315323. KARASOV W.H. 1986: Energetics, physiology and vertebrate ecology. Trends Ecol. Evol. 1: 101104. SCHIECK J. O. & MILLAR J. S. 1985: Alimentary tract measurements as indicators of diets of small mammal. Mammalia 49: 93104. VORONTSOV N.N. 1962: The ways of food specialization and evolution of the alimentary system in Muroidea. In: Kratochvl J. & Pelikn J. (eds), Symposium Theriological Proceedings of the International Symposium on Methods of Mammalogical Investigation, Brno. Publ. House Academia Praha: 360377. WILCZYNSKA B. 1998: Anatomical structure and size of large intestinal mucosa in selected species of shrews and rodents. Acta Theriol. 43: 363370. WILCZYNSKA B. 1999: Histometry and surface area of the alimentary canal of Apodemus flavicollis. Acta Theriol. 44: 2936. ZHANG Z.B. & WANG Z.W. 1998: Ecology and Management of Rodent Pests in Agriculture. Ocean Press, Beijing: 209220. ZHOU Q. 1990: Zoogeographical characteristics of mammals in Baiyinxile Area. In : Inner Mongolia Grassland Ecosystem Research Station, Academia Sinica, (ed.), Research on Grassland Ecosystem. Science Press, Beijing, China: 269275.

55

56

Folia Zool. 52(1): 5766 (2003)

Evaluation of genotoxic damage in wild rodents from a polluted area in the Czech Republic
Luisa Anna IERADI1, Jan ZIMA2,3, Fabio ALLEGRA4, Eva KOTLNOV2,5, Luigi CAMPANELLA6, Rossella GROSSI6 and Mauro CRISTALDI4
1

Nucleic Acid Centre, CNR, Department of Genetic and Molecular Biology, University La Sapienza, 00185 Rome, Italy; e-mail: ieradi@axrma.uniroma1.it Institute of Vertebrate Biology, Academy of Sciences of the Czech Republic, Kvtn 8, 603 65 Brno, Czech Republic; e-mail: jzima@brno.cas.cz Department of Zoology, Faculty of Science, Charles University, Vinin 7, 128 44 Praha 2, Czech Republic Department of Animal and Human Biology, University La Sapienza, 00161 Rome, Italy; e-mail: cristaldi@axrma.uniroma1.it Department of Zoology and Ecology, Faculty of Science, Masaryk University, Kotlsk 2, 611 37 Brno, Czech Republic; e-mail: kotlanova@cizp.cz Department of Chemistry, University La Sapienza, 00185 Rome, Italy

Received 11 December 2001; Accepted 22 July 2002

A b s t r a c t . An integrated system of evaluation of genotoxic damage was employed in two rodent species (Clethrionomys glareolus, Apodemus flavicollis) collected in a polluted and a control site in northern Bohemia. Elevated concentrations of heavy metals (Cd, Mn, Pb, Cu, Zn) were found in soil samples from the polluted site. Significantly higher concentrations of certain non-essential heavy metals were observed in kidney and liver of the animals studied in the polluted site, however, temporal variation between years and seasons was significant. Heavy metals originating from industrial aerial pollution may be not distinctly reflected in tissue concentrations recorded in small terrestrial rodents, in contrast to soil pollution resulting from mine works and/or from road traffic. In the same animal samples, micronucleus test and sperm abnormalities assays were used to study the correlation between environmental pollution and genetic damage. Higher genotoxic damage was demonstrated in the animals collected in the polluted site. Micronuclei frequencies appeared to be more suitable markers of mutagenicity than those of abnormal sperms. A. flavicollis seems to be a more suitable model in studies of environmental genotoxicity than C. glareolus. Key words: environmental pollution, heavy metals, rodents, central Europe

Introduction
Many investigations have been performed to reveal heavy metal contamination in mammals. Field studies have shown that high levels of heavy metals in the environment may appropriately be monitored by the assessment of their concentrations in target organs of freeliving small mammals. The increased levels have usually been considered to relate to environmental pollution, and significant relationships were found between residues of metals in soil and body organs or tissues (W r e n 1986, S h o r e 1995). Bioaccumulation of heavy metals in tissues of small terrestrial mammals was studied in different areas of several countries (e.g. J o h n s o n et al. 1978, A n d r e w s et al. 1984, H u n t e r et al. 1987, 1989, S a w i c k a - K a p u s t a et al. 1987, T i c e et al. 1987, W l o s t o w s k i 1987, C o o k e
57

et al. 1990, T a l m a g e & W a l t o n 1991, I e r a d i et al. 1992, 1996, T u l l - S i n g l e t o n et al. 1994, u m b e r a et al. 2003), and several species were proposed as promising indicators to estimate the hazards of heavy metal pollution. One of the most important sources of environmental contamination with heavy metals is the coal industry. The dust emitted contains zinc, copper, lead and cadmium, and this contamination may increase the content of the metals in the tissues of mammals inhabiting polluted areas. Among heavy metals, copper, zinc and manganese play an important role in metabolism as components or activators of enzymes and their tissue concentrations are effectively controlled over a wide range of metal intake (R o b e r t s & J o h n s o n 1978). Other heavy metals such as cadmium and lead are non-essential and xenobiotic and their concentrations are physiologically more poorly regulated (N e l s o n B e y e r et al. 1996). Therefore, these elements have to be included in the range applied in the environmental pollution assessment. Our previous results, made particularly in relation to the occurrence of supernumerary chromosomes in A. flavicollis, suggested the animals from the polluted area at the town of Litvnov in northern Bohemia actually experience an increased genotoxic damage (D e g r a s s i et al. 1999, Z i m a et al. 1999). Heavy metals may be among xenobiotic pollutants responsible for this genotoxic damage. Monitoring of genotoxic damage usually includes various tests, such as cytogenetic disturbances analysed in somatic cells, and/or analyses of sperm abnormalities. These endpoints are based on tests validated on laboratory mice (M a v o u r n i n et al. 1990, S a l a m o n e & M a v o u r n i n 1994). The micronucleus test in bone marrow (H e d d l e et al. 1983) or peripheral blood erythrocytes (M a c G r e g o r et al. 1987) is widely used as a short term assay for the detection of agents which induce chromosome aberrations in somatic cells. The sperm abnormality assay was proposed by W y r o b e k & B r u c e (1975) as a system, which uses the sperm morphology as an index of mutagens or carcinogens. It is particularly relevant for studying the genotoxic effects on germ cells because in this system transmissible genetic damage from one generation to another takes place. Sperm cells may be examined rapidly and in large numbers, and this is important because the sensitivity of sperm cells do not exactly parallel that of many types of somatic cells. The micronucleus test, chromosome aberrations and sperm abnormality assays have been successfully used on wild rodents to put in evidence the correlation between genetic damage and radioactive contamination (C r i s t a l d i et al. 1985, 1990, 1991) or heavy metal exposure in free-living animals (M c B e e & B i c k h a m 1988; I e r a d i et al.1992, 1996, T u l l - S i n g l e t o n et al. 1994). In this study, an integrated system of mutagenicity tests is proposed to evaluate the response to the damage induced by mutagens and carcinogens, produced by industrial pollution. The two mutagenicity tests (micronucleus test and sperm abnormality assay) were applied to animals collected in two areas with different degrees of pollution, located in the Czech Republic, and compared with the level of heavy metal concentration assessed in certain tissues.

Material and Methods

The field research was performed in two areas of northern Bohemia in October 1995 and July 1996. The first study site, Litvnov (1342E 5037N), was located in the Most Basin, an area

58

with an extensive concentration of petrochemical industries and thermal power plants. The yearly average concentrations of 68 mg/m3 of SO2, 69 mg/m3 of NOx, and 70 mg/m3 of dusty aerosols were recorded in the area by the C z e c h E c o l o g i c a l I n s t i t u t e (1997). The soil samples were collected and small mammals were trapped at the beginning of the umn dl Valley at the edges of an old beech forest. The control site was at a village Filipov near the town esk Kamenice (1423E; 5049N), situated within the Labe River Sandstone Protected Landscape Area. The soil samples were collected and small mammals were trapped in small woods and shrubs along a stream. The weight, sex and age were determined in all trapped animals. Heavy metals (cadmium, manganese, lead, copper, and zinc) were analysed in soil samples from both sites, and in kidney and liver. Liver and kidney were removed and stored at minimum -20 C prior to analysis. All instruments used in sample preparation were rinsed with metal-free deionized water. Tissues were then digested in suprapur nitric acid and hydrogen peroxide and the solution was brought to volume with deionized filtered water. Soil cores (5.5 cm in diameter, 10 cm depth) were dried at 60 C, processed by a standard sieve, and 500 mg was homogenized and digested with HNO3 with following addition of H2O2. All the samples were analysed by Sequential Inductively Coupled Plasma SpectroPhotometer. Detection limits were as follows: 0.3, 1.2, 1.5, 2, 0.9 g/l (ppb) for Mn, Pb, Cd, Cu and Zn, respectively. Transformed data (log x + 1) were analysed statistically using Student t-test for independent groups (p<0.05). Altogether, 27 A. flavicollis and 28 C. glareolus were examined in Filipov, and 40 A. flavicollis and 29 C. glareolus from Litvnov. The micronucleus (MN) test was performed in bone marrow and in peripheral blood. Marrow cells were flushed from a single femur of each animal with foetal calf serum. After centrifugation, the supernatant was removed and pellet was re-suspended in a small amount of serum and smeared on a clean glass slide. The slides, air dried and fixed with absolute methanol or alcohol, were stained with May-Grunwald (C o l e & A r l e t t 1979). Peripheral blood, taken from heart, was prepared using the same method. Altogether, 27 A. flavicollis and 32 C. glareolus were examined in Filipov, and 28 A. flavicollis and 30 C. glareolus from Litvnov. From each animal, 2000 erythrocytes were analysed in bone marrow and in peripheral blood respectively. The sperm abnormality assay was applied to mature males (totally 11 A. flavicollis and 11 C. glareolus from Filipov; 11 A. flavicollis and 4 C. glareolus from Litvnov). The number of mature males was low because many of them were affected by testicular regression. A suspension of sperm cells, obtained from cauda epididymis, was filtered through 80 m and mixed with 1% of eosin-Y solution (10:1). The air-dried smears were mounted in Permount and coded for microscopic examination. The frequency of abnormal sperm cells (ASC) was scored in 1500 sperm cells (SC) per individual. Only cells with tails were taken into consideration. In order to assess the abnormal sperm morphology, the basic criteria of the classification of abnormal sperm cells of laboratory mouse were used (W y r o b e k & B r u c e 1975). Data obtained were analysed with routine statistical methods, such as the analysis of variance (ANOVA), t-test and G test.

Results and Discussion

The surface soil layer in Litvnov contained higher concentrations of the heavy metals studied than that in Filipov (Table 1), and higher contents of Cd and Pb were observed in 1995 in

59

Table 1. Metal soil concentrations (g/g, wet weight) in the sites studied. Metal 1995 Cd Mn Pb Cu Zn 0.97 186.5 22.75 20 44.57 Filipov 1996 0.18 165 20 13 46 1995 34.33 279 47.13 18.6 72 Litvnov 1996 1.87 246 35 26 88

comparison with 1996. In data obtained for metal tissue concentrations, variation between the trapping periods, sites, tissues and species was found. No difference was found in the concentrations between the sexes. The age structure in the groups of animals from the two sites and years did not allow us to perform any exact statistical analysis. Nevertheless a slight increase of cadmium and lead was observed in adult animals in comparison to the young. In the pooled samples from the both trapping periods, significantly higher concentrations of certain metals were found in the kidney (Pb, p<0.05; Cu, p<0.001; Zn, p<0.001), and in liver (Cu, p<0.001; Zn, p<0.001) in A. flavicollis from Litvnov than from Filipov. In the kidneys of pooled samples of C. glareolus from Litvnov a significantly higher concentrations of cadmium (p<0.05) and manganese (p<0.01) were found than in the sample from Filipov, whereas no significant differences were observed in the liver. The comparison of metal levels between samples collected in different periods revealed marked differences between 1995 and 1996. Elevated metal concentrations of heavy metals (particularly of Cd and Pb) were found in the animals collected from Litvnov in 1995 compared to 1996 (Tables 2, 3). This difference was similar to that found in soil samples. Distinct differences between cadmium and lead contents were observed between kidney and liver in A. flavicollis. (Pb: t= 2.077, p= 0.043; Cd: t=2.928, p= 0.005). Significant differences between the two sites were found more frequently in A. flavicollis than in C. glareolus. In particular, a significant increase in the cadmium content was observed in the kidney of A. flavicollis. Our results thus demonstrated slightly elevated accumulation of heavy metals at the polluted site in Litvnov. In this site, the maximum concentrations of 20.00 g/g wwt of lead and of 19.42 g/g wwt of cadmium were found in the kidney of A. flavicollis. The concentration of cadmium exceeds the critical value (6.0 g/g wwt ) proposed by A u g h e y et al. (1984). The maximum concentration of lead did not reach the critical level which is indicative for adverse health effects in mammals (M a et al. 1991, S h o r e & D o u b e n 1994, N e l s o n B e y e r et al. 1996). Furthermore, our data indicate that temporal variation between years and seasons may be important in the monitoring of heavy metal pollution and the bioaccumulation assessment. Environmental industrial pollution in the Most Basin produce considerable acidity of forest soils which is known to increase solubility of many metals (B e r g v i s t 1986). Our data indicate that this supposed factor apparently did not markedly affect tissue concentrations in small rodents (F o l k e s o n et al. 1990). It could be supposed that the heavy metal pollution originated from industrial air pollution is not distinctly reflected in tissue concentrations recorded in small terrestrial rodents, in contrast to soil pollution resulting from mine works. K o r o l j e v a et al. (1992) studied heavy metal concentrations in small mammals, including C. glareolus and A. flavicollis in southern Bohemia, and found rather low levels of burden in animals trapped in relatively unpolluted habitats.
60

Table 2. Tissue metal concentrations (wet weight, standard deviation in italics) in the samples of C. glareolus from Litvnov and Filipov collected in 1995 and 1996. n, sample size; g/g, concentration. Differences between the sites indicated by * (p<0.05), ** (p<0.01), *** (p<0.001). kidney n Cd Mn Pb Cu Zn 5 7 11 12 12 n Cd Mn Pb Cu Zn liver n Cd Mn Pb Cu Zn 2 8 12 11 12 Filipov 1996 n Cd Mn Pb Cu Zn 11 16 1 16 16 8 8 2 15 15 Filipov 1995 Litvnov 1995 g/g 1.60 (1.57) 3.90 (1.43)** 2.92 (3.23) 10.63 (7.59) 26.33 (27.16) Litvnov 1996 g/g 0.36 1.82 1.76 1.54 7.33 Litvnov 1995 g/g 1.82 3.06 0.93 10.55 22.60 Litvnov 1996 n 12 22 5 22 19 g/g (1.51) (1.44) (0.19) (7.65)* (2.58) (0.27) (0.88) (0.20) (2.33) (8.33)

g/g

n 6 6 6 6 6 n (0.13) (0.77) (0.18) (1.02) (7.61) 13 17 14 20 20 n (1.12) (1.25) (0.27) (1.53) (5.21) 6 6 6 6 7

0.35 (0.25) 1.34 (1.13) 1.71 (2.33) 6.41 (4.20) 18.42 (19.11) Filipov 1996 g/g

0.13 0.96 0.27 3.71 18.34 Filipov 1995 g/g

0.80 2.34 0.82 4.83 18.49 g/g

0.12 (0.12) 1.61 (0.67) 0.13 3.89 (1.30) 17.96 (5.22)

0.32 (0.48) 1.98 (1.23) 0.40 (0.26) 3.91 (1.43) 15.49 (6.73)

The data obtained from the mutagenicity tests are presented in Table 4. The micronucleus test (Fig.1) in peripheral blood erythrocytes shows a significantly higher frequency of micronuclei in wood mice collected in Litvnov than those collected in Filipov (t=2.568, p=0.013). The frequency of micronuclei in the bone marrow cells was also higher in Litvnov, but the difference was not significant (t=1.663, p=0.102). A similar significant increase of MN frequency in peripheral blood was observed in bank voles from Litvnov in comparison with Filipov (t=2.787, p=0.007). No significant differences were observed in the bone marrow cells, however. The MN frequency observed in peripheral blood in =2.501.72) was significantly higher (t=2.997, p=0.004) than in C. glareolus A. flavicollis (x (x=1.41.09); the MN frequency in bone marrow was also significantly higher (t=3.248, =2.311.83) than in C. glareolus (x =0.910.57). The results p=0.002) in A. flavicollis (x obtained in peripheral blood micronucleus test in both species indicate possible absence of spleen function, demonstrated previously in Mus spretus (T a n z a r e l l a et al. 2001). Regarding the influence of the age, a small increase in the MN frequency was observed in adult wood mice in comparison with subadults , both in peripheral blood and in bone marrow.
61

Fig. 1. Frequencies of micronuclei (MN/1000) observed in bone marrow and in peripheral blood of A. flavicollis (A.f.) and C. glareolus (C.g.) collected in Filipov (Fi) and Litvnov (Litv).

The results were quite similar in bank voles, but the difference was not significant. No significant differences in MN frequency were observed between males and females in both species. The frequency of abnormal sperm cells in both species was relatively low (Fig. 2), however a higher incidence of the abnormal sperm cells was observed in wood mice from Litvnov in comparison with Filipov (t=2.0897, p=0.049), whereas the difference observed in the frequency of abnormal sperm cells between C. glareolus collected in Litvnov and Filipov was not significant. The level of sperm abnormalities was higher in A. flavicollis than in C. glareolus (t=2.3878, p=0.0328).

Conclusions
Environmental mutagenetic effects related to industrial and other pollution have affected more intensively the animals collected in Litvnov than those collected in Filipov at esk Kamenice. Therefore, according to the available data, this locality can be considered heavily polluted and exposed to a higher mutagenetic impact than Filipov. This conclusion is congruent with the previous results obtained in studies of CREST-staining of micronuclei and scoring of chromosome aberrations in the same sites (D e g r a s s i et al. 1999, Z i m a et al. 1999). Most of the mutagenicity tests performed indicated a significantly higher genotoxic impact in the samples from Litvnov than from Filipov. The micronucleus and chromosome aberration frequencies seem to be more suitable markers than abnormal sperm cells frequency to investigate the genetic damage produced in animals living in areas displaying environmental contamination. In particular, the micronucleus frequency in peripheral blood has demonstrated to be a more effective biomarker for detection of chronic exposition to environmental contaminants.
62

Table 3. Tissue metal concentrations (wet weight, standard deviation in italics) in the samples of A. flavicollis from Litvnov and Filipov collected in 1995 and 1996. For explanations see Table 2. kidney n Cd Mn Pb Cu Zn 6 15 13 15 15 n Cd Mn Pb Cu Zn liver liver Cd Mn Pb Cu Zn liver Cd Mn Pb Cu Zn 2 6 5 9 12 n 8 14 15 15 15 n 12 12 12 12 12 Filipov 1995 Litvnov 1995 g/g 6.73 (6.51)** 5.77 (1.27)* 5.57 (6.35)** 4.97 (1.09)* 46.28 (37.29)** Litvnov 1996 g/g 0.70 (0.66) 1.68 (0.99) 2.10 (1.47) 4.98 (1.38) 26.76 (10.26) Litvnov 1995 g/g 2.01 (2.67) 7.57 (0.91)*** 4.40 (6.42)* 12.35 (6.37)*** 33.94 (15.01)*** Litvnov 1996 g/g 0.41 (0.78) 2.53 (1.50) 0.80 (0.91) 4.77 (2.17) 31.90 (10.63)

g/g (0.24) (2.98) (0.43) (1.23) (3.80)

n 9 9 9 9 9 n (0.07) (0.96) (1.38) (1.23) (3.35) 12 16 19 23 23 n (0.78) (1.68) (0.60) (0.88) (4.09) 9 9 9 9 9 n 27 31 14 31 31

0.31 3.67 1.10 3.82 14.92 Filipov 1996 g/g

0.55 1.02 1.64 3.68 17.45 Filipov 1995 g/g

0.85 4.06 0.90 3.18 15.89 Filipov 1996 g/g

0.23 (0.30) 1.44 (0.32) 0.64 (0.38) 3.38 (1.62) 22.49 (13.46)

Fig. 2. Frequencies of abnormal sperm cells observed (ASC/1500 SC) in cauda epididymys of A. flavicollis (A.f.) and C. glareolus (C.g.) collected in Filipov (Fi) and Litvnov (Litv).

63

Table 4. Results of the tests of mutagenetic effects in samples from Litvnov and Filipov. MN/1000 = micronuclei per 1000 cells, ASC/1500 = abnormal sperm cells per 1500 cells, n = number of animals, x = mean, SD = standard deviation. site n MN/1000 peripheral blood x SD 1.54 2.50* 0.78 1.40* 1.32 1.72 0.50 1.09 MN/1000 bone marrow x SD 1.68 2.31 0.73 0.91 1.10 1.83 0.56 0.57 11 11 11 4 ASC/1500 x 8.45 14.0* 5.18 5.75 SD 4.56 8.37 5.28 3.86

A. flavicollis C. glareolus

Filipov Litvnov Filipov Litvnov

27 28 32 30

The differences between the polluted and control sites were mostly more pronounced in A. flavicollis compared to C. glareolus. Therefore, the former species appeared a more suitable indicator of environmental pollution and genotoxic damage in the area under study. Similar conclusion was proposed by A b r a m s s o n - Z e t t e r b e r g et al. (1997). Nevertheless, bank voles collected in 137 Cs contaminated sites showed an increased micronuclei frequency (C r i s t a l d i et al. 1991). The comparison of the results obtained in two different regions will allow an evaluation of the local situation as well as the achievement of the generalized conclusions. In fact, this research provides fundamental information on the selection of appropriate biomarkers of mutagens and carcinogens in the environment and may contribute significantly to the basic knowledge of industrial pollution.

Acknowledgements
We would like to thank Mgr. Milan o v e k , director of the SCHOLA HUMANITAS, Litvnov, for enabling us to use the laboratory and housing facilities. In particular, we would like to thank Mariangela C o r i a n d r i and Alessandra S a r t o r e t t i for their support in the field and laboratory work. A part of this work was supported by the Italian National Research Council (n. 96.00084.CT04 CNR, Italy). Work in the Czech Republic was partly supported by the Grant Agency of the Czech Republic (no. 206/97/0850) and the Ministry of Education, Youth and Sports of the Czech Republic (no. VS 97102).

L I T E R AT U R E
ABRAMSSON-ZETTERBERG L., GRAW J. & ZETTERBERG G. 1997: Spontaneous and radiation-induced micronuclei in erythrocytes from four species of wild rodents: a comparison with CBA mice. Mut. Res. 393: 5571. ANDREWS S. M., JOHNSON M. S. & COOKE J. A. 1984: Cadmium in small mammals from grassland established on metalliferous mine waste. Environ. Pollut. (Series A) 33: 153162. AUGHEY E., FELL G. S., SCOTT R. & BLACK M. 1984: Histopathology of early effects of oral cadmium in the rat kidney. Environ. Health Perspectives 54: 153161. BERGKVIST B. 1987: Soil solution chemistry and metal budgets of spruce forest ecosystems in S. Sweden. Water, Air Soil Pollut. 33: 131154. COLE J. & ARLETT C. F. 1979: Transplacental effects of chemical mutagens detected by micronucleus test. Nature 77: 317318.

64

COOKE J. A., ANDREWS S. M. & JOHNSON M. S. 1990: Lead, zinc, cadmium and fluoride in small mammals from contaminated grassland established on fluorpar tailing. Water, Air, and Soil Pollution 51: 4354. CRISTALDI M., DARCANGELO E., IERADI L. A., MASCANZONI D., MATTEI T. & VAN AXEL CASTELLI I. 1990: 137-Cs determination and mutagenicity tests in wild Mus musculus domesticus before and after the Chernobyl accident. Environ. Pollut. 64: 19. CRISTALDI M., IERADI L. A., LICASTRO E., LOMBARDI BOCCIA G. & SIMEONE G., 1985: Environmental impact of nuclear power plants on wild rodents. Acta Zool. Fennica 173: 205207. CRISTALDI M., IERADI L. A., MASCANZONI D. & MATTEI T. 1991: Environmental impact of the Chernobyl accident : mutagenesis in bank voles from Sweden. Int. J. Radiat Biol. 59: 3140. CZECH ECOLOGICAL INSTITUTE 1997: Statistical environmental yearbook of the Czech Republic 1997. The Ministry of the Environment of the Czech Republic and Czech Statistical Office. DEGRASSI F., TANZARELLA C., IERADI L.A., ZIMA J., CAPPAI A., LASCIALFARI A., ALLEGRA F. & CRISTALDI M. 1999: Crest-staining of micronuclei from free-living rodents to detect environmental contamination in situ. Mutagenesis 14: 391396. FOLKESON L., NYHOLM N. E. I. & TYLER G. 1990: Influence of acidity and other soil properties on metal concentrations in forest plants and animals. The Sci. Total Environment 96: 211233. HEDDLE J. A., HITE M., KIRKHART B., MAVOURNIN K., MACGREGOR J. T., NEWELL G.W. & SALAMONE M. F. 1983: The induction of micronuclei as measure of genotoxicity. A report of the U.S. Environmental Protection Agency Gene-Tox Program. Mut. Res. 123: 61118. HUNTER B. A., JOHNSON M. S. & THOMPSON D. J. 1987: Ecotoxicology of copper and cadmium in a contaminated grassland ecosystem. III. Small mammals. J. Appl. Ecol. 24: 601614. HUNTER B. A., JOHNSON M. S. & THOMPSON D. J. 1989: Ecotoxicology of copper and cadmium in a contaminated grassland ecosystem. IV. Tissue distribution and age accumulation in small mammals. J. Appl. Ecol. 26: 8999. IERADI L. A., CRISTALDI M., MASCANZONI D., CARDARELLI E., GROSSI R. & CAMPANELLA L. 1996: Genetic damage in urban mice exposed to traffic pollution. Environ. Pollut. 92: 323328. IERADI L. A., GIORDANO D. & AMARENA D. 1992: Heavy metal contamination in small rodents from urban areas. In: Boh J. (ed.), Bioindicatores Deteriorisationis Regionis. Proceedings of the 6th International Conference, esk Budjovice: 219227. JOHNSON M. S., ROBERTS R. D., HUTTON M. & INSKIP M.J. 1978: Distribution of lead, zinc and cadmium in small mammals from polluted environments. Oikos 30: 153159. KOROLJEVA E. G., MIHOLOVA D., MADER P., CEVELOVA J. & SEDLEK F. 1992: Heavy metals (Cd, Pb, Zn, Cu) in livers of mice and voles from the Temeln region in South Bohemia. In: Boh J. (ed.), Bioindicatores Deteriorisationis Regionis. Proceedings of the 6th International Conference, esk Budjovice: 430434. MA W.C., DENNEMAN W. & FABER J. 1991: Hazardous exposure of ground living small mammals to cadmium and lead in contaminated terrestrial ecosystems. Arch. Environ. Contam. Toxicol. 20: 266270. MACGREGOR J. T., HEDDLE J. A., HITE M., MARGOLIN B. H., RAMEL C., SALOMON M. F., TICE R. R. & WILD D. 1987: Guidelines for the conduct of micronucleus assay in mammalian bone marrow erythrocytes. Mut. Res. 189: 103112. MAVOURNIN K. H., BLAKEY D. H., CIMINO M. C., SALAMONE M. F. & HEDDLE J. A. 1990: The in vivo micronucleus assay in mammalian bone marrow and peripheral blood. A report of the U.S. Environmental Protection Agency Gene-Tox Program. Mut. Res. 239: 2980. McBEE K. & BICKHAM J. W. 1988: Petrol-chemical related DNA damage in wild rodents detected by flow cytometry. Bull. Environ. Contam. Toxicol. 40: 343349. NELSON-BEYER W., HEINZ G. H. & REDMON-NORWOOD A. W. (eds) 1996: Environmental contaminants in wildlife. SEATAC Special Publications Series. CRC Press, Boca Raton. ROBERTS R. D. & JOHNSON M. S. 1978: Dispersal of heavy metals from abandoned workings and their transference through terrestrial food chains. Environ. Pollut. 16: 293310. RUIZ-LAGUNA J., GARCIA-ALFONSO C., PEINADO J., MORENO S., IERADI L. A., CRISTALDI M. & LOPEZ-BAREA J. 2001: Biochemical biomarkers of pollution in Algerian mouse (Mus spretus) to assess the effects of the Aznalcollar disaster in Doana Park (Spain). Biomarkers 6: 146160. SALAMONE M. F. & MAVOURNIN K. H. 1994: Bone marrow micronucleus assay: a review of the mouse stock used and their published mean spontaneous micronucleus frequencies. Env. Mol. Mut. 23: 239273. SAWICKA-KAPUSTA K., GORECKI A. & LANGE R. 1987: Heavy metals in rodents from polluted forests in southern Poland. Ekol. Pol. 35: 345354.

65

SHORE R. F. 1995: Predicting cadmium, lead and fluoride levels in small mammals from soil residues and by species-species extrapolations. Environ. Pollut. 88: 333340. SHORE R. F. & DOUBEN P. E. T. 1994: The ecological significance of cadmium intake and residues in terrestrial small mammals. Environ. safety 29: 101112. UMBERA R., BARU V. & TENORA F. 2003: Heavy metals in the silvery mole-rat, Heliophobius argenteocinereus (Bathyergidae, Rodentia) from Malawi. Folia Zool. in press. TALMAGE S. S. & WALTON B. T. 1991: Small mammals as monitors of environmental contaminants. Reviews of Environmental Contamination and Toxicology vol. 119. Springer Verlag, New York. TANZARELLA C., DEGRASSI F., CRISTALDI M., MORENO S., LASCIALFARI A., CHIUCHIARELLI G. & IERADI L. A. 2001: Genotoxic damage in free-living Algerian Mouse (Mus spretus) after the Coto Doana ecological disaster. Environ. Pollut. 115: 4348. TICE R. R., ORMISTON B. G., BOUCHER R., LUKE C. A. & PAQUETTE D. E. 1987: Environmental monitoring with feral rodent species. Environ. Sci. Res. 36: 175179. TULL-SINGLETON S., KIMBALL S. & McBEE K. 1994: Correlative analysis of heavy metals bioconcentration and genetic damage in white-footed mice (Peromyscus leucopus) from a hazardous waste site. Bull. Environ. Contam. Toxicol. 52: 667672. WLOSTOWSKI T. 1987: Heavy metals in the liver of Clethrionomys glareolus (Schreber, 1780) and Apodemus sylvaticus (Pallas, 1771) from forests contaminated with coal-industry fumes. Ekologia Polska 35: 115129. WREN C. D. 1986: Mammals as biological monitors of environmental metal levels. Environ. Monit. Assess. 6: 127144. WYROBEK A. J. & BRUCE W. R. 1975: Chemical induction of sperm abnormalities in mice. Proc. Nat. Acad. Sci. USA 72: 44254429. ZIMA J., IERADI L. A., ALLEGRA F., SARTORETTI A., WLOSOKOV E. & CRISTALDI M. 1999: Frequencies of B chromosomes in Apodemus flavicollis are not directly related to mutagenetic environmental effects. Folia Zool. 48 (Suppl. 1): 115119.

66

Folia Zool. 52(1): 6776 (2003)

Influence of seasonality, temperature and rainfall on the winter diet of the long-eared owl, Asio otus
Diego RUBOLINI1*, Andrea PIROVANO2 and Sergio BORGHI3
1

Dipartimento di Biologia Animale, Universit degli Studi di Pavia, p.zza Botta 9, I-27100 Pavia, Italy; e-mail: diego.rubolini@unipv.it Wildlife Biology and Management Unit, Center for Life Sciences Weihenstephan, Technische Universitt Mnchen, Am Hochanger 13, D-85354 Freising, Germany Osservatorio Meteorologico di Milano-Duomo, piazza Duomo 21, I-20121 Milano, Italy

Received 22 January 2002; Accepted 16 August 2002

A b s t r a c t . In this study we analysed whether the diet composition of a wintering population (4070 individuals) of long-eared owls (Asio otus) in northern Italy showed within-season variation, and whether it was influenced by ambient temperature and rainfall. Diet composition was determined by pellet content, and over 5500 prey items were analysed; pellets were collected at 2-wk intervals over two consecutive winters (October to April), 19961998. Three out of five main prey categories showed a marked within-season variability in relative frequency in diet, both considering the number of prey items and prey biomass, whereas between-year variability was shown only by a single prey category (Savis pine vole). Although rainfall had no influence on diet composition, temperature affected negatively the prevalence of harvest mouse, a relatively unimportant prey category. Thus, the considered weather variables seem to have little influence on the winter diet composition (at the level of individual prey categories) of these owls. However, diet breadth (estimated by the Levins index of niche breadth) increased with increasing rainfall and decreasing temperature, when calculated on the proportion of prey items: hence it seems that the owls become more generalists under unfavourable weather conditions. Key words: weather conditions, nocturnal raptors, small mammals, niche breadth, seasonal variation

Introduction
Meteorological conditions can affect the diet and hunting success of raptors. For example, rain seems to have a negative impact on the hunting performance of some species, hampering flight or perceptive ability (B a r b i e r i et al. 1975, W i j n a n d t s 1984, M i c h e l a t & G i r a u d o u x 1992, O l s e n & O l s e n 1992, H e n r i o u x 1999). In the long-eared owl (Asio otus Linnaeus, 1758), a low ambient temperature and presence of rain seems to reduce flying activity, although the effect varied according to season (H e n r i o u x 1999). Also, surface activity patterns of small mammals (the commonest prey of the long-eared owl; M a r t i 1976) are influenced by the weather to varying degrees (e.g. S i d o r o w i c z 1960, G e t z 1961, M a g u i r e 1999). Further, diet composition may also depend on seasonal variation (hereafter defined as seasonality) in prey choice by the owls and/or prey activity, as shown by previous analyses of the owls diet (N i l s s o n 1981, P i r o v a n o et al. 2000a). The most comprehensive study so far is that of N i l s s o n (1981) in southern Sweden, which investigated changes in diet composition througout several year cycles. However, the data were grouped by month, and were mostly discussed in terms of betweenyear and between-season variations in prey and habitat choice, whereas no effort has been devoted to the analysis of within-season variability patterns in diet composition.
* Corresponding author 67

To our knowledge, few studies addressed whether there are relationships between weather conditions and diet composition in free-living raptors, while accounting for seasonal variation in diet (e.g. C a n o v a 1989). Here we analysed within-season variation in diet composition (determined by pellet content) in a population of long-eared owls studied over two consecutive winter seasons (October to April), 19961998. The diet and behaviour of this urban wintering population has been described elsewhere (P i r o v a n o et al. 2000a,b): the brown rat (Rattus norvegicus) was the most abundant prey in terms of biomass, and the monthly proportion of rats in the diet was inversely related to mean monthly rat weight (P i r o v a n o et al. 2000a), suggesting a selection on young rats (90100 g). The aim of the present study was to assess the effect of seasonality on diet and the relationships between diet composition and meteorological variables (mean temperature and amount of rainfall preceding pellet collection) and determine which of these factors explained the largest variation in diet.

Material and Methods

The roost site, occupied by 4070 owls during winter (October to April), was located in the southern tip of the city of Milan (northern Italy, 4528N 912E). Owls hunt in the suburbs and cultivated fields surrounding the city (P i r o v a n o et al. 1997, 2000a,b). The owls likely hunting range consisted mainly of winter stubbles (cereal crops, colza), poplar plantations, meadows, together with patches of network habitats (copses, hedgerows along field margins), which are actively selected by owls whilst searching for prey in our study region (G a l e o t t i et al. 1997). Mean monthly rainfall in the study period was 80 mm per month (ranging from 150 mm in December to 6 mm in March), while mean monthly temperatures ranged from 14.6C in October to 3.9C in January. The study was carried out over two consecutive winters (NovemberApril 199697 and OctoberApril 199798). Pellets were collected at 2 wk-intervals each month (first collection: 15th of each month; second collection: 30th/31st) and analysed following standard techniques (Y a l d e n & N o r r i s 1990). Possible biases associated with pellet collection were minimised by collecting all the intact pellets at each visit, always along the same trail below the owls perches. A preliminary cleaning-up of all pellet remains was undertaken 2 wk before the first collection for each of the two study years. As in other studies, mammals were determined to species level, and birds were considered as a single category (G a l e o t t i & C a n o v a 1994, P i r o v a n o et al. 2000a). Weather variables were recorded at Brera-Duomo Meteorological Observatory (3 km from the roost site) at hourly intervals. In the analyses, we concentrated only on main prey categories (those constituting at least 5% of prey items over the two years, following the definition given in P i r o v a n o et al. 2000a): they were wood mouse (Apodemus sylvaticus) (38.5 %), brown rat (21.7 %), Savis pine vole (Microtus savii) (17.3 %), birds (9.6 %), harvest mouse (Micromys minutus) (5.5 %). Main prey categories made up 93 % of prey items over the two years. The complete dataset used in the analyses is reported in the Appendix. Statistical analyses

Using mean hourly temperature data, we calculated the daily mean temperatures for 2-week periods, and rainfall was measured as the amount of rain (in mm) recorded in the fortnight preceding pellet collection, i.e. the timeframe during which pellets were produced. We used
68

both proportion of prey items (%N), which represents hunting acts, and proportion of prey biomass (%B), representing an index of energy intake by the owls. This latter index may be relevant in an energetic context, especially during winter, an energetically stressful period for raptors (e.g. O v e r s k a u g et al. 1997, N e w t o n 1998). Estimates of prey biomasses were derived from the literature (D i P a l m a & M a s s a 1981, G a l e o t t i & C a n o v a 1994) and from specimens collected in the study area. Birds were assigned a mean mass of 20 g each (G a l e o t t i & C a n o v a 1994, P i r o v a n o et al. 2000a). The body mass of predated brown rats was estimated by measuring mandible length and using the regression equation given in D i P a l m a & M a s s a (1981). In biomass calculations, rats were thus assigned the mean monthly weight. The proportions of prey categories (both as %N and %B) were calculated for each fortnight. The number of prey items for each fortnight varied between 29 and 580 (median = 198 prey items, see Appendix): overall, 25 fortnights were included in the analyses. All variables were normally distributed (Kolmogorov-Smirnov test, P > 0.18), so that no data transformation was performed before variables being used in parametric statistical analyses. For each fortnight we calculated diet breadth using the L e v i n s (1968) index of niche breadth (NB = 1 / pi2, where pi is the proportion of the ith prey category); compared to other diversity indices, this index gives more weight to dominant diet components (which can be hypothesized to have more biological relevance), and its value is less sensitive to stochastic variation in the occurrence of the least abundant prey categories. Diet breadth was calculated on all prey categories, including those not listed above. In order to separate seasonal variation in diet composition from the effect of weather (temperature and rainfall), we first built maximal models in which diet parameters were expressed as a function of year of pellet collection and date of sampling. Given that a preliminary look at the variation of prey categories in relation to date of sampling revealed mostly non-linear patterns, we also included in the maximal models the quadratic term of date, together with all the year by date interactions. Models were built using Type I sum of squares (hierarchical sum of squares), in which each term is corrected only for those preceding it in the model definition (N o r u s i s 1988). Thus, the factor year was always entered first in the model, followed by date and its quadratic term, whereas interactions (year x date, year x date2) were entered after main effects. Before testing the additive effects of temperature and rainfall, non-significant higher order terms and interactions were dropped from the model, and the model was run again with significant terms included only (this is termed Step 1 hereafter). Note that if higher order terms were significant, then the corresponding lower order terms were also left in the model, even if they did not reach statistical significance. Then we tested for the effects of temperature and rainfall, which were entered separately in the models (Step 2a and Step 2b, respectively). If the initial model (before Step 1) was non significant, then the effects of weather variables on diet parameters were tested alone in an analysis of covariance (ANCOVA). Owing to the strict non-linear covariation between temperature and date of sampling (temperature as a function of date: R2 = 0.73, 2nd order polymomial function, P < 0.0001), with temperature being lower in mid-winter months, effects of seasonality per se and temperature on diet could be difficult to disentangle. Hence, the same models as above (Step 2a) were also run entering temperature before the effects of date of sampling: to evaluate which model performed best, we compared the change in R2 (R2) between
69

model types (i.e. those with temperature entered after date vs. those with temperature entered before date) after entering the first term (note that seasonal terms, e.g. date and date2, are considered altogether in calculating the change in R2). Means are shown together with their standard deviations. Results In total, 5509 prey items were examined over the two years (see Appendix). The 2-wk proportions of prey items and prey biomasses were strictly correlated (mean rs = 0.93; range 0.850.96; N = 5 prey categories; all P < 0.0001). Three out of five prey categories showed a clear seasonal variation in prevalence in the diet (Table 1, Fig. 1), both in terms of number and biomass (i.e. wood mouse, brown rat and harvest mouse). Birds showed a seasonal variation in prevalence only when considering %N, whereas no variables were significant in the initial model for %B (Table 1). Savis pine vole was the only prey category showing a significant between-year variation in prevalence (Table 1). In both years, the dietary response over the winter season was non-linear, the prevalence of rats and birds decreasing around mid-winter being replaced by wood mice and birds. As mid-winter was the period with the lowest temperatures, the prevalence of rats, harvest mice and birds trend could also be modelled as a function of mean temperature (see significant effects at Step 2a in Table 1).

Table 1. Summary of ANCOVA models (type I sum of squares) describing the relationships between long-eared owl diet parameters (diet composition, expressed both as % number (%N) and % biomass (%B)), seasonal variation and weather variables during two consecutive study winters (October to April), 19961998. See Material and Methods for a detailed description of variables and model definition. Model steps Diet parameters Step 1 (year, season) Variables Apodemus sylvaticus Rattus norvegicus Microtus savii Micromys minutus Birds %N %B %N %B %N %B %N %B %N %B R2 (%) date, date2 date, date2 date, date2 date, date2 year year, date date2 date2 date, date2 Step 2a^ (temperature) P 70.7 61.9 66.8 48.7 43.9 38.1 47.5 45.3 59.6 R2 (%) 0.25 - 0.14 0.62 - 0.17 0.39 - 0.00* 0.49 - 0.15 0.22 - 0.19 0.07 - 0.17 0.07 - 0.00* 0.15 - 0.00* 0.29 - 0.02* 0.49 66.8 - 14.5 47.5 - 51.3 45.3 - 41.8 59.6 - 11.0 Step 2b^ (rainfall) P 0.68 0.72 0.10 0.24 0.88 0.47 0.39 0.32 0.76 0.10 R2 (%) -

Only significant terms are shown; associate model R2 is shown when overall models significant at P < 0.05. ^ P-values for additional effects of temperature and rainfall are shown; for temperature: left value: temperature entered after other variables (those selected with Step 1); right value: temperature entered before other variables. Significant P-values (P < 0.05) are marked with an asterisk (*). Corresponding changes in % R2 (R2) for the two model types (after entering the first term; seasonal terms are considered altogether in calculating the change in R2, see Materials and Methods for details) are also shown when P-values for temperature were significant at P < 0.05.

70

However, the only prey category that showed a consistent covariation in frequency with mean temperature upon the general effect of seasonality was the harvest mouse, which increased in proportion when temperature was lower, whereas seasonality seems to be far more important for other prey categories (Table 1). No correlation was found between the amount of rainfall and the percentages of any prey category (all P > 0.10, Table 1). Diet breadth calculated on %N was not correlated with the same index calculated on %B (rs = -0.11; N = 25; P = 0.59), suggesting that different factors could be affecting this index at these two levels of analysis (%N or %B). The maximal model was non-significant for %N diet breadth (F5,19 = 2.16, P = 0.10), whereas date (P = 0.004) and date2 (P = 0.04) were significant predictors of %B diet breadth after running Step 1, with breadth mostly increasing from winter to spring (model: F2,22 = 7.59, P = 0.003). %N diet breadth increased with decreasing temperature (F 1,23 = 9.29, P = 0.006) (Fig. 2a) and with increasing rainfall (F1,23 = 5.62, P = 0.026) (Fig. 2b), but there were no correlations between weather variables and %B diet breadth (both entering weather variables after and before date terms, all P > 0.05). The breadth indexes were not dependent on the total number of prey items included in each fortnight sample (%N: rs = 0.30; N = 25; P = 0.14; %B: rs = 0.08; N = 25; P = 0.70). Mean diet breadth calculated on %B was significantly lower than that calculated on %N (2.12 0.50 vs. 3.53 0.59, respectively; paired t-test: t24 = 8.57, P < 0.0001): this is because of the strong dominance of rats in biomass (mean fortnight proportion of rat biomass in diet is 64.9 12.9 %, range 36.888.5 %, see Fig. 2, Appendix). In fact, %B diet breadth was dependent on %B of rats in diet (rs = -0.98; N = 25; P < 0.0001), emphasizing that the brown rat is by far the most important prey and that the niche breadth in this apparently specialist predator is basically a function of the prevalence in the diet of the most important single prey type. However, in this particular situation, the Levins index calculated on %B, being heavily influenced by dominant prey categories, may not be able to provide an adequate representation of the seasonal variations in diet breadth. Discussion Owing to the proportional nature of dietary data, with one prey type increasing in prevalence as others decrease, it may be difficult to identify the reasons behind changes in diet composition (statistically as well as biologically). Indeed, from our investigations, it is clear that the brown rat is the primary prey type in this wintering owl population, in which birds behave as rat specialists (see also P i r o v a n o et al. 2000a). The prevalence of rats reaches up to 80% of prey biomass and is around 20% in number of prey items. As a consequence, the prevalence of other numerically important mammalian prey types (wood mouse and Savis pine vole) were negatively correlated to the percentage of brown rat and were thus seemingly alternative prey to the brown rat, given the dominance of rats by biomass (P i r o v a n o et al. 2000a). Overall, the most important prey categories showed clear within-season variations in prevalence throughout the winter. Weather variables had no or limited effects on the variation of individual prey categories: the only prey whose prevalence in diet was consistently affected by temperature was the harvest mouse, which, owing to its small size, makes a low contribution to the overall diet, and can thus be considered to have a low biological relevance (see Appendix). The increased prevalence of harvest mice with cold
71

Fig. 1. Seasonal variation in diet composition of long-eared owls, measured at 2-wk intervals. For clarity of presentation, only fitted lines for prey categories that showed significant seasonal variation in frequency are shown, according to patterns detected in Table 1; a) frequency of prey items (%N); b) frequency of prey biomass (%B).

weather may reflect a higher availability of this prey with lower temperatures, when a fraction of the population may be forced to leave cultivated fields by habitat deterioration, thus showing a more predictable occurrence in network habitats largely preferred by owls for hunting (G a l e o t t i et al. 1997). In general, the relationships between diet composition and weather conditions may have different origins: either (1) predation by the owls is influenced by weather conditions, given a uniform availability of prey species, or (2) activity/availability of prey
72

Fig. 2. Correlation between an index of long-eared owl diet breadth (Levins index), calculated for each fortnight on %N, and: a) 2-wk mean ambient temperature; b) total amount of fortnight rainfall (in mm); filled circles = winter 199697; open circles = winter 199798.

species depends on weather variations and the owls eat what is available. Of course, these possible processes are not mutually exclusive, and what is actually found in the owls diet may derive form their combination: a thorough investigation of the factors influencing diet should link dietary data to small mammal and owls activity pattern and habitat use. Future efforts should thus include the quantification of the above-mentioned variables, and perhaps simultaneous trapping of small mammals in different habitats and quantification of owls habitat use (by e.g. radio-tracking), while taking into account weather conditions, may provide adequate answers.
73

With regards to our results, possible reasons for the observed generalized lack of effect of weather variables on individual prey frequencies may include the relatively mild winter climate and the interval at which pellet were collected, that may have been too long and may have obscured the effect of weather on diet. Previous studies of resident raptors wintering at high latitudes have shown food availability, related to climatic factors, to be of major importance in determining the survival probability throughout winter (N e w t o n 1998, S u n d e 2002), but weather conditions are likely to have a limited importance in our study area on both owl activity and food abundance: hence, it would be interesting to investigate longer time series including harsh winter seasons, which may exacerbate the effect of weather on diet composition. Diet breadth (calculated on %N) was positively related to rainfall and negatively to temperature. This may simply reflect a decrease in diet breadth when many young rats are available, i.e. when temperatures are high (P e r r y 1946, P i r o v a n o et al. 2000a): however, this would imply a negative correlation between %N diet diversity and abundance of rats (both %N and %B), but this is not the case (Spearman Rank: P = 0.60 and P = 0.30, respectively). Hence, either overall average prey activity may be higher at lower temperatures and elevated soil moisture (M a g u i r e 1999), or owls may become more euryphagous under inclement weather conditions, which seems a likely explanation (C a n o v a 1989). In fact, lower prey selectivity under relatively unfavourable weather conditions, which may negatively affect the energy budget of the owls or the abundance of food resources, is to be a generally expected ecological response to climatic variability (e.g. M a c A r t h u r & P i a n k a 1966). When analysing diet diversity from an energetical point of view (%B diet breadth), given that rats dominate the diet, diversity increased when the prevalence of rats decreased during mid-winter months (probably in response to a lower availability of young rats), with owls shifting to other prey types (wood mouse, Savis pine vole). Thus, diet diversity seems to be influenced both by climatic factors and by the abundance of the preferred prey, depending on the measure used (%N or %B). To conclude, the diet composition of the long-eared owl in our study area appeared to be independent of rainfall and slightly influenced by temperature, while a measure of diet variability increased with decreasing temperature and with an increasing amount of rainfall. Taken together, these results suggest that further studies investigating the diet of owls should consider seasonal variation and weather conditions as potentially influential variables, and emphasize that the winter diet composition of the long-eared owl in Southern Europe show marked withinseason fluctuations.

Acknowledgements
The content of pellets was determined by S. B r a m b i l l a . We thank all the people who helped with fieldwork and G. B o g l i a n i , P. G a l e o t t i , C. M a g u i r e , C. M a r t i and an anonymous reviewer for suggestions on a previous draft of the manuscript. We are indebted with P. S u n d e for his insightful comments and for thorough discussion on data analysis. O. J a n n i kindly checked the English. We are also grateful to the Editor and Dr G. C o p p for comments on the final revision.

74

Appendix. Diet parameters (for both N and B) of long-eared owls in the suburbs of Milan, northern Italy, from October to April, according to year (Y) (1 = winter 1996/97; 2 = winter 1997/98) and date (D) of collection (1 = 1-15 October; 14 = 16-30 April). Diet breadth, number of prey items considered (N items) and weather variables (Temp = temperature; rain = rainfall) are also shown (see Material and Methods for details). Microtus savii %N 10.41 33.33 33.21 10.34 21.09 22.53 23.45 23.25 18.22 19.13 17.54 37.21 1.56 6.71 3.54 10.44 10.80 14.84 14.90 18.60 21.97 14.11 19.32 13.27 12.82 3.04 15.24 13.30 3.45 10.50 12.71 11.45 9.89 8.81 7.03 7.89 19.31 0.47 1.68 1.15 3.06 3.88 8.60 7.38 8.43 12.01 5.87 13.20 5.84 6.42 3.17 0.00 2.64 6.90 11.22 5.49 8.28 3.32 3.10 2.73 1.75 0.00 0.00 2.44 3.03 3.85 9.60 9.11 5.49 8.53 4.79 3.32 6.82 3.54 5.13 0.26 0.00 0.30 0.65 1.57 0.87 1.14 0.40 0.42 0.28 0.22 0.00 0.00 0.17 0.28 0.32 0.97 1.48 0.76 1.09 0.74 0.39 1.31 0.44 0.72 9.50 19.19 13.58 6.90 6.12 2.75 4.48 4.43 3.49 12.02 6.14 6.98 34.38 10.98 15.15 11.54 9.20 12.24 7.84 10.34 9.01 5.81 7.95 1.77 7.69 2.89 9.14 5.66 2.40 3.17 1.61 2.28 1.96 1.76 4.60 2.88 3.77 10.82 2.86 5.13 3.52 3.44 7.38 4.05 4.88 5.13 2.52 5.66 0.81 4.01 3.33 4.25 4.33 4.14 3.62 2.78 3.61 3.37 2.84 3.93 2.86 3.62 3.20 2.85 3.62 3.97 4.42 4.04 3.84 4.69 3.21 3.20 2.84 2.44 3.37 1.46 2.57 2.02 1.56 2.43 2.74 2.33 2.01 2.29 1.76 2.11 2.44 1.45 1.27 1.52 1.43 1.70 2.85 2.42 2.17 2.72 1.99 3.07 2.03 2.72 221 99 265 29 294 182 580 271 258 183 114 43 64 164 198 182 500 384 255 387 355 241 88 113 39 11.10 6.21 6.38 2.59 2.69 5.12 5.91 8.81 12.67 12.87 13.94 12.54 18.41 10.75 9.81 6.69 5.74 3.36 3.82 4.10 6.47 10.98 9.96 10.34 11.12 %B %N %B %N %B Micromys minutus Birds N items Temp (C) Rain (mm) 55.0 90.6 68.0 57.8 79.2 20.2 0.0 10.8 0.0 0.0 0.0 16.6 10.2 10.4 114.2 5.0 51.4 127.4 42.2 16.8 0.0 39.0 1.6 9.6 71.4

Y %N 42.53 18.18 22.26 31.03 13.27 9.89 13.79 21.03 17.44 28.96 21.05 16.28 42.19 55.49 38.89 39.56 26.20 10.42 16.08 17.31 12.39 20.33 6.82 22.12 17.95 81.74 58.24 68.00 78.98 58.07 47.44 59.13 66.03 58.11 73.57 62.42 55.69 82.04 88.50 80.39 83.16 75.10 51.16 58.76 64.79 51.42 66.85 36.79 64.01 53.48 %B

Apodemus sylvaticus Rattus norvegicus

%N

%B

Diet breadth N B

1 1 1 1 1 1 1 1 1 1 1 1 2 2 2 2 2 2 2 2 2 2 2 2 2

3 4 5 6 7 8 9 10 11 12 13 14 1 2 3 4 5 6 7 8 9 10 11 12 13

29.41 22.22 21.89 34.48 44.22 53.85 43.28 43.17 52.71 34.43 50.88 27.91 9.38 15.24 31.31 25.82 35.20 43.23 44.31 35.40 48.73 49.38 54.55 58.41 48.72

9.46 11.20 9.66 12.68 24.27 33.51 23.30 20.24 28.09 13.95 25.24 15.97 3.12 4.21 11.22 8.34 13.96 27.62 24.21 17.69 29.38 22.67 41.10 28.35 26.92

According to the Levins index of niche breadth (NB = 1 / pi2) on all prey categories determined to species level (including birds, that were considered as a single category).

75

L I T E R AT U R E
BARBIERI F., BOGLIANI G. & FASOLA M. 1975: Aspetti della predazione di un barbagianni (Tyto a. alba Scop.). In: Atti V Simposio Nazionale sulla Conservazione della Natura, Vol. I. Istituto di Zoologia, Universit di Bari: 293302. CANOVA L. 1989: Influence of snow cover on prey selection by Long-eared Owl Asio otus. Ethol. Ecol. Evol. 1: 367372. DI PALMA M. G. & MASSA B. 1981: Contributo metodologico per lo studio dellalimentazione dei rapaci. In: Farina A. (ed.), Atti I Convegno Italiano di Ornitologia. Mus. St. Nat. Lunigiana, Aulla: 6976. GALEOTTI P. & CANOVA L. 1994: Winter diet of Long-eared Owl (Asio otus) in the Po Plain (Northern Italy). J. Raptor Res. 28: 265268. GALEOTTI P., TAVECCHIA G. & BONETTI A. 1997: Home-range and habitat use of long-eared owls in open farmland (Po Plain, northern Italy), in relation to prey availability. J. Wildl. Res. 2: 137145. GETZ L. L. 1961: Responses of small mammals to live traps and weather conditions. Amer. Midl. Nat. 66: 160170. HENRIOUX F. 1999: cologie dune population de Hiboux moyen-ducs Asio otus en zone dagriculture intensive. Ph.D. dissertation, Universit de Neuchtel, Switzerland, 119 pp. LEVINS R. 1968: Evolution in changing environments. Princeton University Press, Princeton, 132 pp. MACARTHUR R. H. & PIANKA E. 1966: On optimal use of a patchy environment. Amer. Nat. 100: 603609. MAGUIRE C. C. 1999: Rainfall, ambient temperature, and Clethrionomys californicus capture frequency. Mammal Rev. 29: 135142. MARTI C. D. 1976: A review of prey selection by the Long-eared Owl. Condor 78: 331336. MICHELAT D. & GIRAUDOUX P. 1992: Activit nocturne et stratgie de recherche de nourriture de la chouette effraie Tyto alba partir du site de nidification. Alauda 60: 38. NEWTON I. 1998: Population Limitation in Birds. Academic Press, London, 597 pp. NILSSON I. N. 1981: Seasonal changes in food of the Long-eared Owl in southern Sweden. Ornis Scand. 12: 216223. NORUSIS M. J. 1988: SPSS Guide to Data Analysis for SPSS/PC+. SPSS Inc., Chicago, 416 pp. OLSEN P. & OLSEN J. 1992: Does rain hamper hunting by breeding raptors? Emu 92: 184187. OVERSKAUG K., SUNDE P. & KRISTIANSEN E. 1997: Subcutaneous fat accumulation in Norwegian owls and raptors. Ornis Fenn. 74: 2937. PERRY J. S. 1946: The reproduction of the wild brown rat (Rattus norvegicus Erxleben). Proc. Zool. Soc., London, 115: 1946. PIROVANO A., RUBOLINI D., DE MICHELIS S. & FERRARI N. 1997: Primi dati sullecologia di un roost di Gufo comune Asio otus in ambiente urbano. Avocetta 21: 89. PIROVANO A., RUBOLINI D., BRAMBILLA S. & FERRARI N. 2000a: Winter diet of urban roosting Longeared Owls Asio otus in Northern Italy: the importance of the Brown Rat Rattus norvegicus. Bird Study 47: 242244. PIROVANO A., RUBOLINI D. & DE MICHELIS S. 2000b: Winter roost occupancy and behaviour at evening departure of urban long-eared owls. Ital. J. Zool. 67: 6366. SIDOROWICZ J. 1960: Influence of the weather on capture of Micromammalia. I. Rodents (Rodentia). Acta Theriol. 4: 139158. SUNDE P. 2002: Starvation mortality and body condition of Goshawks Accipiter gentilis along a latitudinal gradient in Norway. Ibis 144: 301310. WIJNANDTS H. 1984: Ecological energetics of the Long-eared Owl (Asio otus). Ardea 72: 192. YALDEN D. W. & MORRIS P. A. 1990: The Analysis of Owl Pellets. Mammal Society Occasional Publication No. 13, London, 24 pp.

76

Folia Zool. 52(1): 7786 (2003)

Seasonal dynamics and breeding of amphibians in pristine forests (Biaowiea National Park, E Poland) in dry years
Bogumia JDRZEJEWSKA1, Marcin BRZEZISKI2 and Wodzimierz JDRZEJEWSKI1
1

Mammal Research Institute, Polish Academy of Sciences, 17-230 Biaowiea, Poland; e-mail: bjedrzej@bison.zbs.bialowieza.pl Department of Ecology, University of Warsaw, ul. Banacha 2, 02-097 Warszawa, Poland

Received 19 November 2001; Accepted 8 July 2002

A b s t r a c t . In summer 1992 through spring 1994, amphibian abundance and breeding was studied in the pristine temperate forests, typical of central European lowlands. The years 1991, 1992, and 1993 were among the driest in the recent decades, with the spring-summer precipitation 35% lower than the long-term average. In the primeval forests of Biaowiea National Park, common frogs Rana temporaria spawned in small (on average, 0.2 ha) ponds (postglacial melt-out hollows) devoid of wood cover and characterised by water pH 5.1-6.0 (as measured in April). Breeding success of frogs, monitored qualitatively in 1993, was rather poor due to pond desiccation. The capture of amphibians on forest grids revealed that densities and seasonal dynamics differed between wet and drier deciduous forests. No amphibians were captured in the mixed coniferous forests during the study. In the wet ash-alder forests, on average, 39 amphibians ha-1 were recorded in late April, 12 ind ha-1 in summer, and 195222 ind ha-1 in autumn (September). In those forests, 90% of captured amphibians were common frogs, 6% common toads Bufo bufo, and 4% moor frogs R. arvalis. In the drier oak-lime-hornbeam forests, amphibians appeared in May, and increased in numbers towards summer (1924 ind ha-1) and autumn (4571 ind ha-1). Of all amphibians caught in those forests, 43% were common frogs, 38% common toads, and 19% were moor frogs. A majority of amphibians captured in autumn were young of the year. By mid-October, all amphibians had left the forest for their hibernation sites. Comparison of our data collected in very dry years with other available data from Biaowiea Primeval Forest (various years between 1955 and 1998) revealed that summer indices of amphibian abundance were strongly positively correlated with rainfall in April-June of the census year and the previous year. Key words: common frog, moor frog, common toad, density, spawning ponds, water pH, ash-alder forest, oak-lime-hornbeam forest, rainfall

Introduction
Amphibian populations are characterised by wide fluctuations in numbers and heterogeneous spatial distribution due to the life history demands of amphibians (seasonal migrations from and to hibernation sites, spawning sites, and summer quarters) and their low resistance to unfavourable ambient conditions, especially water deficit (e.g. K u t e n k o v 1995, M e y e r et al. 1998, G r o v e r 2000, P i k u l i k et al. 2001). Hence, the temporal and spatial variation in amphibian abundance in the forests is much greater than that recorded in other groups of small vertebrates: birds, rodents, and insectivorous mammals (J d r z e j e w s k a & J d r z e j e w s k i 1998). Furthermore, there is growing concern that amphibian populations are declining worldwide (P e c h m a n n et al. 1991, H o u l a h a n et al. 2000), thus more data are needed on factors affecting their densities at local and regional scales.
77

This study was conducted in Biaowiea National Park (BNP), Eastern Poland, which represents one of Europes last pristine forest, typical of the lowland temperate zone (F a l i s k i 1986). The aims were to describe (1) spawning sites of frogs in a pristine woodland, and (2) seasonal dynamics of abundance of terrestrial amphibians (frogs and toads) in three major forest associations. The time of the study (from summer 1992 till spring 1994) was characterised by a strong deficit of precipitation as compared with the long-term average, so the results are representative of rather unfavourable environmental conditions for amphibians. We then compared our results with those from other surveys conducted in Biaowiea Primeval Forest between 1955 and 1998 in order to determine the factors affecting between-year variation in amphibian abundance.

Study Area
The strict reserve of Biaowiea National Park (E Poland, 2355E, 5245N) covers 47.5 km2. It is located in the centre of the Biaowiea Primeval Forest (about 1500 km2). The forest extends on both sides of the Polish-Belarussian border and preserves the remnants of the European temperate lowland forests of hemiboreal-nemoral type. In the BNP, the structure and character of the forest is typical of virgin forests unaltered by management, hunting, and tourism. The mosaic of fairly dry, swampy and river-flooded forest associations with many decaying uprooted trees makes the habitat extremely varied in micro- and macroscale. In the Park, hunting and timber exploitation are not allowed; tourism is restricted to a few pedestrian paths. The main forest associations of BNP are the rich, multilayered deciduous stands dominated by oak Quercus robur, hornbeam Carpinus betulus, and lime Tilia cordata, with admixtures of maple Acer platanoides and spruce Picea abies. In such forests, the ground vegetation is dominated by spring geophytes and it becomes meagre in summer and autumn due to overshadowing by tree canopies. Oak-hornbeam-lime forests cover about half of the BNPs area (Fig. 1). Mixed coniferous forests, composed of spruce, pine Pinus silvestris, and oak, with admixtures of aspen Populus tremula and birch Betula pendula grow on drier sites and cover about 25% of the area. The herbs layer, in addition to soft-tissue plants, contains dwarf shrubs, some evergreen species (such as Vaccinium spp.) and mosses. Two other associations, bog alderwoods and streamside forests, are shaped by water and cover 18% of the BNPs area. The streamside forests (dominated by black alder Alnus glutinosa, ash Fraxinus excelsior, with admixtures of elm Ulmus scabra) are situated along flowing waters and are flooded in spring. Bog alderwoods (composed of black alder with admixtures of birch B. pubescens and spruce) occur at the peripheries of river and stream valleys and in depressions of the denudation plains. High groundwater level and its stagnation over a large part of the year are typical of bog alderwoods. The forest floor is distinctly structured with hollows and hummocks. Streamside forests and bog alderwoods usually occur near each other. The herbs layer in both associations is very prolific and the peak of its abundance occurs in summer. A small stream, the Orwka (14 m wide, up to 1 m deep) flows through the Park (Fig. 1). The water level in the Orwka varies seasonally, reaching its maximum in spring. The Orwka often dries up in summer. In winter, it freezes but unfrozen air-holes can still be found, in its middle and lower course, at temperatures below -15 C. The upper part of the stream freezes completely at about -5 C. The BNP lies in the temperate subcontinental climatic zone. In 19931994, the mean daily temperature in January was 0.1 C, and in July 19.6 C. Snow cover persisted for an average of 193 days per year. More
78

Fig. 1. Map of the strict reserve of the Biaowiea National Park with localities of study sites. Spatial distribution of the main forest types after K w i a t k o w s k i (1994), simplified.

information on the vegetation of the BNP is given by F a l i s k i (1986) and J d r z e j e w s k a & J d r z e j e w s k i (1998).

Material and Methods

We applied the methods elaborated and tested by G o w a c i s k i & W i t k o w s k i (1970), and based on earlier methodological works by Z i p p i n (1956, 1958). We captured amphibians on 8 grids, 30 x 30 m (0.09 ha) each. Trapping (with a small net-bag and by hand) was usually done by 2 persons walking and carefully searching the forest floor on a grid. Two surveys were done daily: in early morning and before dusk. Capturing lasted 3 days in spring and summer, and 4 days in autumn. In 1992, two censuses were conducted: in July, on three grids in oak-hornbeam-lime forests (one of the grids was surrounded with cloth to prevent amphibian movements into and out of the grid during the census), and in September, on three grids in oak-lime-hornbeam forests and one grid located in the wet ash-alder forest. In 1993, all 8 grids (5 in the oak-hornbeam-lime forest, 2 in the ash-alder forest, and one in the mixed
79

coniferous forest) were operated 5 times: in late April, early June, mid July, late September, and mid October. Captured frogs and toads were identified to species, weighted, and put into a covered plastic bucket (bedded with moist litter and moss), where they stayed for 14 days, i.e. until the end of the census. Afterwards, they were released in the place of capture. Totally, 122 amphibians were caught in 19921993. Since the number of amphibians caught on the fenced grid did not differ from those captured in unfenced grids located in the same habitats, all data were treated in the same manner. For density calculations, we assumed that, during 3 to 4-day census, we caught 100% of amphibians inhabiting the grids. Owing to the low abundance of amphibians during the study, we could have checked this assumption for the pooled sample of data only. The numbers of amphibians captured on the grids declined on consecutive days of a census (r = -0.84), and the regression equation (y = 65 - 12.9x) indicated that on day 5, on average, only 0.5 amphibians would have been caught on all grids (see D e L u r y 1947, and H a y n e 1949, for the rationale of using linear regression to calculate densities from removal trapping). In April 1993, spawning sites of frogs were searched for on the area of about 15 km2 (Fig. 1). Seven ponds, where spawning frogs or the spawn had been found, were then monitored until April 1994. The presence of spawn and tadpoles were recorded, and water was sampled for pH measurements. Note that these seven ponds were not the only spawning places of amphibians in BNP. They represented breeding sites of frogs located deep in the forest interior. Other, unstudied spawning sites of frogs and toads were located in and along the two larger rivers (Narewka and Hwona), forming the borders of the strict reserve of BNP (see Fig. 1). Water pH was also measured in seven other sources (wooded bogs and ditches in the forest, and the Orwka River), where no breeding of frogs was recorded. Data on weather were obtained from the Biaowiea meteorological station.

Results
Abiotic conditions of frog reproduction Compared to the long-term mean conditions, spring and summer seasons of 19911993 were hot and dry. In 19481998 (51 years), the annual precipitation ranged from 426 to 939 mm (mean=637 mm, SD=123). In 1991, 1992, and 1993 (the study years and a preceding year), annual precipitation amounted 439, 525, and 531 mm, respectively (mean=498 mm, SD=51), i.e. 78% of the long-term mean value. Water deficit was especially severe in spring and early summer. In 19481998, AprilJune rainfall ranged from 94 to 338 mm (mean=189 mm, SD=58). By contrast, in 1991, 1992, and 1993, only 109, 129, and 127 mm of rainfall were recorded (mean=122 mm, SD=11), i.e. 65% of the long-term mean value. In 1994 and 1995, spring and summer seasons were again wet. In spring 1993, we located breeding individuals and the spawn of the common frogs Rana temporaria in six ponds and those of the moor frogs R. arvalis in one pond (Table 1). All ponds were postglacial melt-out hollows, devoid of tree and shrub cover, surrounded by oldgrowth forest, and as summer advanced, they became overgrown with herbaceous vegetation. The spawning ponds were small (mean size 0.2 ha in spring) and in general were located about 1 km from the nearest river or stream (Fig. 1, Table 1). Mean water pH in the ponds used by the common frogs was 5.5 (SD=0.3) in April, and increased to 6.0 (SD=0.6) in
80

May. The site used by moor frogs had lower pH, 4.1 and 4.4, respectively (Table 1). In 1994, common frogs were found lekking in five ponds (including one previously occupied by moor frogs) and their mean pH was 5.6 (SD=0.25). Generally, very little variability was recorded in the acidity of water used by spawning common frogs in spring (pH from 5.1 to 6.0; Table 1).
Table 1. Characteristics of frog spawning ponds in Biaowiea National Park, 1993-1994. a lekking moor frogs; in all other cases the common frogs were recorded. + yes, - no, (s) breeding success possible, ? no data. Mean values of water pH calculated for the common frogs spawning sites, only. See Fig. 1 for the spatial distribution of frogs spawning ponds. No. of pond Pond size Straight-line (ha) distance to the nearest river or stream (km) 0.20 0.22 0.20 0.18 0.11 0.27 0.20 0.20 (0.05) 1.57 1.37 0.83 1.33 0.57 0.67 1.02 1.05 (0.38) Water pH in 1993 April 4.1a 5.6 5.7 6.0 5.1 5.6 5.1 5.5 (0.35) May 4.4a ? 5.7 7.0 5.8 5.6 5.3 5.9 (0.65) +a + + + + + + Presence of spawn in spring 1993 Presence of tadpoles in MayJune 1993 +(s) + +(s) +(s) + Water pH in April 1994 Presence of spawn in spring 1994 + + + + +

1 2 3 4 5 6 7 Mean (SD)

5.4 6.0 ? 5.6 5.3 5.7 5.7 5.6 (0.25)

In spring 1993, we measured water pH in four sites with standing water (bog alderwoods and other places with tree cover) where no spawning frogs were recorded. It ranged from 5.5 to 7.1 in April (mean=6.4, SD=0.8) and from 6.3 to 7.1 in May (mean=6.6, SD=0.4). Flowing water of the Orwka stream had almost invariable pH of 7.6 in April and 7.7 in May in the lower and middle course of the stream. In the upper part (which dried out in May), pH of water in April was 5.6. Water pH in six spawning sites of the common frog differed significantly from the values recorded in the seven sites of standing and flowing waters not used by amphibians (Mann-Whitney U-test, Us = 35, P = 0.04). Breeding success and densities of amphibians In April 1993, spawn was laid in each of the seven ponds (Table 1). Reduced rainfall and hot weather caused rapid drying of the ponds, which were subsequently overgrown with tall herbaceous vegetation. Actually, two ponds dried nearly completely before the eggs hatched, and by July, all other ponds were very small and shallow (530% of their size and water capacity as estimated visually in April). In JuneJuly, large-sized tadpoles were not observed in any of the ponds, but given the amount of persisting water, some tadpoles could have metamorphosed in three ponds. In spring 1994, after 2 years with very dry springs and summers, fewer ponds were used for spawning (Table 1), despite the fact that abiotic conditions were favourable. It was only in 1995 (another wet year), that the spawning of common frogs was again abundant. A large spawning aggregation of frogs, many times exceeding those in each of the forest ponds, was found in the widely flooded springs of the Orwka stream. That place, characterised by
81

suitable acidity for common frogs (water pH 5.6), was very poor in water in 1993, when no lekking frogs were found. Censuses on grids in the oldgrowth stands of BNP revealed highest densities of amphibians in wet ash-alder forests and substantially lower densities in the drier deciduous forests. Except for a single specimen of the tree frog Hyla arborea caught in 1993, no amphibians were ever captured or seen in the mixed coniferous forests. The seasonal pattern of amphibian abundance markedly differed among the wet ash-alder forests and drier oakhornbeam-lime forests (Fig. 2). In wet forests, common frogs were fairly numerous in late April (on average, 39 ind ha-1 and 190 g ha-1 of crude biomass), declined in summer (1112 ind ha-1 and 50130 g ha-1), and rose again in autumn, when a cohort of young individuals appeared (195222 ind ha-1 and 82122 g ha-1). In the drier forests, amphibians appeared later (in May) and their abundance was highest in summer and/or early autumn (mean 1924 ind ha-1 and 193276 g ha-1 in July and 4571 ind ha-1 or 146304 g ha-1 in September; Fig. 2). By mid October, all amphibians had left the forest floor for their hibernation sites.

Fig. 2. Seasonal dynamics of amphibian abundance (in numbers of individuals and biomass per unit area) in the oldgrowth forests of BNP, in 19921993. Ash-alder forests data averaged for 2 trapping grids; oak-limehornbeam forests data averaged for 5 grids.

Not only the seasonal dynamics pattern but also the species composition of amphibians differed between the two habitats. Data averaged for all censuses in 1993 showed that among amphibians caught in ash-alder forests, the common frog strongly dominated (90% by numbers and 92% of biomass). The moor frog (4% by numbers and 6% of biomass) and the common toad Bufo bufo (6 and 2%, respectively) were rare. In oak-hornbeam-lime forests, the common frog was also the most numerous, but it formed less than half of amphibian numbers (43%) and biomass (49%). The common toad constituted 38% of amphibian numbers and 19% of biomass, and the moor frog 19 and 32%, respectively. The age structure (as approximated by body mass distributions) of common frogs inhabiting forests changed seasonally. In spring, only young and subadult non-breeding
82

frogs were captured (body mass 1.59.5 g, mean=5.2 g), as the breeding adults gathered at spawning sites. Reproductively mature adults (body mass 1838 g, mean=26 g) were found in forests in summer and autumn. In September, 59% of all individuals were in body weight class 13 g, 32% were in class 3.56 g, 3% in class 6.59 g, and 6% in class 2438 g. Thus, a majority of captured common frogs were young of the year. Moor frogs were captured in small numbers in spring through autumn with no clear differences in body mass (mean=10.4 g, minmax 120.5). Toads captured in summer were subadults (7.510.5 g) and all those captured in the autumn seasons were young of the year (body mass from 1 to 4.5 g, mean=2.2, SD=1.0). Worth mentioning is the fact that young, metamorphosed common frogs and common toads captured in September 1993 were significantly smaller than those captured in September 1992. In common frogs, mean body mass of a froglet was 3.7 g (SD=1.0, minmax 25.5 g) in 1992, and 2.5 g (SD=0.8, minmax 14.5 g) in 1993 (Mann-Whitney U-test, Us = 630.5, n1 = 23, n2 = 33, P < 0.0005). The respective values for the young toads were 3.3 g (SD=0.8, minmax 24.5 g) in 1992, and 1.6 g (SD=0.5, range 12.5 g) in 1993 (Us = 130.5, n1 = 8, n2 = 17, P < 0.0005).

Discussion
Our study represented the situation of amphibians typical of severely dry years. Undoubtedly water supply is decisive for the embryonic development and survival of tadpoles and metamorphosed amphibians. Z a m a c h o w s k i (1977) documented experimentally that, in frogs and toads, the resistance to water shortage increased from spring till autumn. They would survive out of water for 1.53 days in spring, and 35 days in autumn. Two factors seem essential for the survival of frog embryos: chemical properties and the amount of water where the spawn had been deposited. A s t o n et al. (1987) recorded in northern England that water pH in 61 ponds with common frogs spawn ranged from 4.2 to 8.9. In laboratory studies, the survival of common frog embryos was much lower in pH 4.5 than in pH 6.0 (B e a t t i e et al. 1992). In the BNP, common frogs were found spawning in waters of very narrow range of pH (5.16.0), but this might have been caused partly by the fact that, generally, very few suitable water sources were available in the dry years. In Belarus, P i k u l i k (1980) reported that spawn clumps of common frogs in wet years were found in a variety of both prospective and obviously short-lasting waters, such as ditches by forest roads, deep puddles on dirt roads, flooded river valleys, old river beds, and slowly flowing waters of forest streams and canals. Most likely, the acidity of such a wide array of places was also rather variable. Although we did not measure directly the breeding success of frogs in the monitored ponds, it seemed very poor as nearly all spawning sites dried by early summer. Breeding success of both R. temporaria and R. arvalis depends on the persistence of spawning waters until the tadpoles metamorphose. K u t e n k o v (1995) reported that the survival of common frog tadpoles in Karelia, Russia, was strongly correlated with the water balance (precipitation minus evapotranspiration) in MayJune. In the Oka River Reserve, Russia, from 0.1 to 8.9% of moor frog eggs deposited in 19711978 (mean=3%) gave rise to metamorphosed froglets (P a n c h e n k o 1980). As much as 95% of the variation in frogs breeding success was explained by year-to-year variation in water capacity in the studied reservoirs, which in turn was shaped by precipitation and flooding. P a n c h e n k o (1980) found that, in spring, the amount of spawn deposited per 1 m3 of water varied from 15 to
83

520 eggs (mean=127, CV=135%). In summer, however, the number of metamorphosed froglets per 1 m3 of water was rather stable, from 1 to 1.4 individuals (mean=1.2, CV=13%). Thus, the amount of water persisting throughout the full cycle of frog reproduction and metamorphosis is a crucial factor limiting the breeding success and autumn numbers of young frogs in the forests. The strong effect of prolonged drought has also been recorded in other species of amphibians. In north-central Florida, D o d d (1993, 1994, 1995) monitored the populations of the eastern narrow-mouthed toad Gastrophryne carolinensis, the striped newt Notophthalmus perstriatus, the oak toad Bufo quercinus, and the southern toad B. terrestris during 5 very dry years. In all four species, the deficit of rainfall and subsequent drying of a spawning pond severely limited or eliminated reproduction for up to 5 consecutive years, reduced the size of the local populations in some of the studied species, and affected the activity patterns of amphibians. On the contrary, M e y e r et al. (1998), who analysed the 2328-year long time-series of the common frog R. temporaria in the region of Bern, Switzerland, found only a weak effect of rainfall on the population fluctuations. In one of the three studied sites, for which data on precipitation were analysed, the number of breeding frogs correlated with rainfall of previous year (from the spawning date to end of October). Densities of amphibians recorded in the BNP in 19921993 were very low as compared to other data from Polish temperate forests collected by the same method (capturing on grids). In the Niepoomice Forest, S Poland, G o w a c i s k i & W i t k o w s k i (1970) estimated the densities of terrestrial amphibians in the oak-hornbeam forests in a wet year of 1967. In August 1967, when young of the year had already dispersed in forests, mean densities of amphibians were 2186 ind ha-1 (SD=429, minmax 16032552) and the mean biomass 11.56 kg ha-1 (SD=1.9, minmax 8.313.3). In July 1999, capturing of amphibians in the mixed coniferous and alder-ash forests near Lake uknajno (Masurian Lakeland, NE Poland) yielded 8290 ind ha-1 and 0.70.9 kg of total biomass ha-1 (M. B r z e z i s k i , unpubl. data). Thus, the abundance of amphibians recorded by us in oak-lime-hornbeam forests of the BNP in very dry years was 40 to 60 times lower than those reported from Niepoomice Forest, and 4 to 8 times lower than that found in the Masurian Lakeland. These three studies elucidate the range of variation in amphibian densities in the temperate forests of Central Europe. In Biaowiea Primeval Forest, this study has so far been the only attempt to estimate the densities of terrestrial amphibians by capturing them on quadrats. However, counts of amphibians on transects (one-meter wide) were conducted in BPF in the summer seasons of 1955, 1957, 19811985, 19951996, and 1998 (B a n n i k o v & B e l o v a 1956, B a k h a r e v 1986, P i k u l i k et al. 1987, 2001; in all cases only overwintering individuals, and not the youngest current-year cohort, were counted). The mean indices of amphibian abundance (weighted by proportional availability of various forest habitats) varied from 3.5 individuals per 1 km of the transect in 1957 to 24 individuals in very wet years of 1984 and 1985 (total precipitation in AprilJune 323 and 267 mm, respectively). Our data from quadrats can be compared with the above mentioned indices by recalculating the number of amphibians trapped during the first census on the first day in the summer trapping series, and the area of the quadrats into the length of a 1-m wide transect. Such rough estimation yielded an average of 0.6 amphibians km-1 in 1993 and 0.7 in 1994. Thus, the maximum numbers of terrestrial amphibians recorded in wet summer seasons (19841985) were again at least 40 times higher than the lowest numbers counted in very dry years (19931994). Indeed, in the entire data set available from the BPF, total AprilJune precipitation over two years (the year, when the census was done, and the previous year) was a crucial factor
84

Fig. 3. Indices of summer abundance of terrestrial amphibians (frogs and toads, the youngest current-year cohort not included) in the Biaowiea Primeval Forest in relation to the total rainfall in April-June (averaged for the year of census and the previous year). Data from: B a n n i k o v & B e l o v a (1956), B a k h a r e v (1986), P i k u l i k et al. (1987, 2001), and this paper.

shaping the abundance of terrestrial amphibians, and the mean values for AprilJune explained 84% of variation in their numbers from year to year (Fig. 3). The strong positive correlation between amphibians and two-year average rainfall most probably reflects the fact that precipitation acts through changes to both the density and activity of amphibians. Rainfall of the previous year, when majority of the censused frogs hatched and metamorphosed, affects the numbers of frogs. Rainfall in the year of the census changes the activity of frogs and, thus, their detectability by surveyors. As suggested by this study, the impact of precipitation on true numbers is heavier, because the larval stages of amphibians are not able to cope with pond desiccation in dry years.
Acknowledgements
We are grateful to E. B u j k o for his help in the field, K. Z u b and M. R a b c z u k - O r z e c h o w s k a for their preparing the map and figures. We sincerely thank A. B o b i e c (Biaowiea) for kindly lending us the equipment for water pH measurements. The study was financed by the grant KBN 4 4416 91 02 from the Committee for Scientific Research (Poland) and the budget of the Mammal Research Institute. Comments and criticisms by two anonymous reviewers are gratefully acknowledged.

L I T E R AT U R E
ASTON R. J., BEATTIE R. C. & MILNER A. G. P. 1987: Characteristics of spawning sites of the common frog (Rana temporaria) with particular reference to acidity. J. Zool., Lond. 213: 233242. BAKHAREV V. A. 1986: Sovremennoe sostayanie gerpetofauny Belovezhskoi Pushchi [The present state of herpetological fauna of Belovezha Primeval Forest]. Zapovedniki Belorusii 10: 8388 (in Russian).

85

BANNIKOV A. G. & BELOVA Z. V. 1956: Materialy k izucheniyu zemnovodnykh i presmykayushchikhsia Belovezhskoi Pushchi [Materials on amphibians and reptiles of Belovezha Primeval Forest]. Uchenye Zapiski Mosk. Gorodskovo Pedagog. Inst. im. V. P. Potemkina 61: 385402 (in Russian). BEATTIE R. C., TYLER-JONES R. & BAXTER M. J. 1992: The effects of pH, aluminium concentration and temperature on the embryonic development of the European common frog, Rana temporaria. J. Zool., Lond. 228: 557570. DE LURY D. B. 1947: On the estimation of biological populations. Biometrics 3: 145167. DODD C. K. 1993: Cost of living in an unpredictable environment: the ecology of striped newts Notophthalmus perstriatus during a prolonged drought. Copeia 1993(3): 605614. DODD C. K. 1994: The effects of drought on population structure, activity, and orientation of toads (Bufo quercinus and B. terrestris) at a temporary pond. Ethol. Ecol. & Evol. 6: 331349. DODD C. K. 1995: The ecology of a sandhill population of the eastern narrow-mouthed toad, Gastrophryne carolinensis, during a drought. Bull. Florida Mus. Hist., Biol. Sci. 38, Pt. I(1): 1141. FALISKI J. B. 1986: Vegetation Dynamics in Temperate Lowland Primeval Forests. Dr. W. Junk Publishers, Dordrecht, Holland. GOWACISKI Z. & WITKOWSKI Z. 1970: Numbers and biomass of amphibians estimated by the capture and removal method. Wiadomoci Ekologiczne 16(4): 328340 (in Polish with a summary in English). GROVER M. C. 2000: Determinants of salamander distribution along moisture gradients. Copeia 2000(1): 156168. HAYNE D. W. 1949: Two methods for estimating population from trapping records. J. Mamm. 30: 399411. HOULAHAN J. E., FINDLAY C. S., SCHMIDT B. R., MEYER A. H. & KUZMIN S. L. 2000: Quantitative evidence for global amphibian population declines. Nature 404: 752755. JDRZEJEWSKA B. & JDRZEJEWSKI W. 1998: Predation in Vertebrate Communities. The Biaowiea Primeval Forest as a Case Study. Berlin, Springer Verlag. KUTENKOV A. P. 1995: On the population dynamics of the common frog (Rana temporaria) in the northern part of its range. Memoranda Soc. Fauna Flora Fennica 71(34): 98102. KWIATKOWSKI W. 1994: Vegetation landscapes of Biaowiea Forest. Phytocoenosis, Supplementum Cartographiae Geobotanicae 6: 3587. MEYER A. H., SCHMIDT B. R. & GROSSENBACHER K. 1998: Analysis of three amphibians populations with quarter-century long time-series. Proc. R. Soc. Lond. B 265: 523528. PANCHENKO I. M. 1980: Effektivnost razmnozheniya ostromordoi lagushki yugo-vostoka Meshchery [Breeding success of the moor frog in the south-eastern Meshchera]. Ekologiya 1980(6): 9598 (in Russian). PECHMANN J. H. K., SCOTT D. E., SEMLITSCH R. D., CALDWELL J. P., VITT L. J. & GIBBONS J. W. 1991: Declining amphibian populations: the problem of separating human impacts from natural fluctuations. Science 253: 892895. PIKULIK M. M. 1980: Razmnazhenne travyanoi zhaby (Rana temporaria L.) u Belarusi [Reproduction of the common frog (Rana temporaria L.) in Belarus]. Vesci Akademii Navuk BSSR, Seryya Biyalagichnykh Navuk 1982(1): 9398 (in Belarussian). PIKULIK M. M., KOSOV S. V., DROBENKOV S. M., LETENCKII A. A. & KHANDOGII, A. V. 1987: Ocenka vliyaniya osushitelnoi melioracii nizinnykh bolot na sostoyanie amfibii i reptilii (na primere Belorusskogo Polesya, verkhove rek Yaseldy i Nareva) [Evaluation of the impact of lowland marsh drainage on the status of amphibians and reptiles (an example of the upper Yaselda and Narev rivers, Belarussian Polese)]. Izdatelstvo VINITI, Akademiya Nauk Belorusskoi SSR, Minsk (in Russian). PIKULIK M. M., SIDOROVICH V. E., JDRZEJEWSKA B. & JDRZEJEWSKI W. 2001: Summer abundance and habitat distribution of frogs (Rana temporaria, R. arvalis, and R. kl. esculenta) and toads (Bufo bufo) in the Biaowiea Primeval Forest, E Poland. Folia Zool. 50: 6573. ZAMACHOWSKI W. 1977: The water economy in some European species of anuran amphibians during the annual cycle. III. Resistance to water shortage. Acta Biol. Cracoviensia, Series: Zoologia 20: 207228. ZIPPIN C. 1956: An evaluation of the removal method of estimating animal populations. Biometrics 12: 163169. ZIPPIN C. 1958: The removal method of population estimation. J. Wildl. Manage. 22: 8290.

86

Folia Zool. 52(1): 8798 (2003)

Home range and habitat selection by Podarcis hispanica (Squamata, Lacertidae) in Western Spain
Francisco Javier DIEGO-RASILLA and Valentn PREZ-MELLADO Departamento de Biologa Animal, Facultad de Biologa, Universidad de Salamanca, Campus Miguel de Unamuno, Edificio de Farmacia 5a planta, 37007 Salamanca, Spain; e-mail: fjdiego@terra.es Received 2 January 2001; Accepted 10 June 2002
A b s t r a c t . A population of the lacertid lizard Podarcis hispanica was studied in the Sistema Central Mountains in Salamanca Province, Spain. Lizards confined their activities to stable home ranges, but size range varied enormously between individuals. Home ranges generally contained 1) a high proportion of rocks > 100 cm in height and 2) a significantly higher abundance of crevices than expected. The overall abundance of lizards increased with decreasing amounts of shrubs and herbaceous plants cover and densities were highest in bare and rocky areas. Adult females selected particular habitat types according to the season, whereas habitat selection of males and juveniles did not differ between spring and summer. Non-resident adult lizards and juveniles occupied habitats of marginal quality in relation to crevice abundance. Key words: lizards, microhabitat, electivity

Introduction
Animals are non-randomly distributed in space. In fact, heterogeneity in spatial resources determines habitat selection (L a w & B r a d l e y 1990). Animals respond to diversity in habitat resources by adopting selection strategies, which tend to increase their survivorship and opportunities for successful mating (W e r n e r et al. 1983, W e r n e r & H a l l 1988). Assuming equal availability (ratio of 1 to 1), disproportionate use of one resource over another is a demonstration of preference for one over the other, with this preference exercised through the process of selection. Therefore, it is necessary to identify the most significant features that are good habitat indicators, to determine which are used as cues by the animals, and to estimate their availabilities for the entire study habitat (K n i g h t & M o r r i s 1996). Models of habitat selection commonly assume that higher quality source habitats will be occupied at higher densities than sink habitats (F r e t w e l l 1972, C l o b e r t et al. 1994). What are the cues on which squamates base habitat selection? Several have been identified: food abundance and its distribution (S i m o n 1975, W a l d s c h m i d t 1983, H e w s 1993), abundance of refuges for avoiding capture (S t a m p s 1983, C l a r k & G i l l i n g h a m 1990, P o u g h et al. 1998), and places with suitable conditions for thermoregulation (L a w & B r a d l e y 1990, C a s t i l l a & B a u w e n s 1991, D a z et al. 1996, G r o v e r 1996). Podarcis hispanica (Steindachner, 1870) (Squamata, Lacertidae) is a small (adult snoutvent length 3770 mm) wall lizard whose distributional range is limited to the Iberian Peninsula, Southern France and North Africa. A complete description of this species and its geographical distribution can be found in P r e z - M e l l a d o (1998) and B a r b a d i l l o et al. (1999). These lizards are able to regulate their body temperature (Tb) within relatively narrow limits during their active periods (D a z et al. 1996). Behavioural
87

thermoregulation in P. hispanica is mainly characterised by the adoption of basking postures and shuttling between sun and shade (P r e z - M e l l a d o 1983, D a z et al. 1996). The preferred body temperature for our study species is 34.4 C (B a u w e n s et al. 1995). The aim of the present study was to assess the home range and habitat use of P. hispanica. We evaluated the habitat use of resident individuals in the study area as well as that of adult males introduced into the study area.

Material and Methods

The study was conducted at the western end of the Sistema Central of the Iberian Peninsula, at the Sierra de la Pea de Francia (1074 m a. s. l.), Salamanca, Spain (40 31 44 N, 06 07 10 W). This area comprised a granite afloration within an open mixed deciduous woodland dominated by oak Quercus pyrenaica. Vegetation in our study site was characterised by oakbush and other scrub vegetation, like Calluna vulgaris, Erica arborea, E. australis, E. umbellata, Lavandula stoechas, and Hallimium alyssoides. Agrostis castellana, Poa bulbosa, and Tuberaria guttata dominate grassy areas. The study was conducted during spring (25/0530/05, 14/623/06) and summer (2/078/07, 4/0913/09) 1991 in a 1500 m2 plot within the study area, and divided into 25 m2 quadrats. The National Institute of Meteorology (Territorial Meteorological Center of Castilla y Len) provided climatic data from two weather stations up to 10 km from the study site. The maximum and minimum shade mean temperatures recorded in spring were 7.3 C and 25.9 C. Corresponding summer values were 12.2 C and 30.1 C, respectively. Individual lizards were captured, marked, released and recaptured within the area. They were toe-clipped for permanent identification and painted with dorsal symbols to permit quick identification from a distance. Each specimen was determined for sex, measured for snout-vent-length (SVL, 1 mm), body mass ( 0.01 g), tail length and the presence of regenerated tail part was noted. Also, we measured body (cloacal = Tb), air (shaded bulb, 1 cm above substrate, = Ta) and substrate (shaded bulb, = Ts) temperatures to the nearest 0.1 C with a thermocouple connected to an electronic thermometer (Digitron). Time and behaviour were also recorded for each sighting. We considered four behaviour categories: basking (dorsoventral flattening of the body and orientation perpendicular to the suns rays), moving (locomotor activities), resting (voluntary inactivity periods in a protected and cool retreat) and other infrequent behaviours (prey capture, prey manipulation, and social behaviours), which were not included in the analyses (D i e g o - R a s i l l a & P r e z M e l l a d o 2000). Coordinates were determined to the nearest 0.1 m, using the point of initial sighting to the nearest grid marker. We established four categories of specimen: resident adult males, introduced adult males (see next paragraph), adult females, and juveniles (< 37 mm snout-vent-length). At the beginning of summer 1991 (24/06) we introduced 29 adult males into the centre of the study plot, where the entire study was undertaken. The individuals had been captured from neighbouring areas approximately 67 km from the study site (D i e g o - R a s i l l a 1999). Home range size was estimated as the area of the convex polygon that encloses all the capture locations for each individual (R o s e 1982). Polygonal home range area can only be calculated for lizards captured at least four times. The convex polygon method has a sample size bias when there is a low number of sightings (J e n n r i c h & T u r n e r 1969). In our study, polygonal home range estimates were based on an average of 7.5 sightings/individual during the spring and an average of 12.1 sightings/individual during the summer. Hence, we
88

corrected the estimates for sample size bias using the correction factors provided by J e n n r i c h & T u r n e r (1969). Fidelity to a home range, measured as the degree of area overlap between intervals (from spring to summer), was computed for all residents common to both periods. We computed the intersection of the two areas divided by the smaller of the areas (C o o p e r 1978): Overlap = [A1 A2 / min(A1, A2)] 100 K A1 = Home range area during the spring A2 = Home range area during the summer K = Criterion to determine the presence of home range. We assumed fidelity to a home range when the degree of areal overlap between spring and summer was greater than or equal to K= 50 %. Habitat use was characterised using one qualitative and one quantitative habitat variable: substrate type and height (bush, grass, sand, smooth rocks with no crevices, rock < 10 cm, rock 1050 cm, rock 50100 cm, rock > 100 cm), and number of crevices in every 25 m2 square in which the plot was divided (0, 15, 610, 1115, > 15 crevices). A map of the study area was constructed entering habitat shapes by keyboard in RANGES V software, an analysis system for biological location data (K e n w a r d & H o d d e r 1996). The map was used to give estimates of habitat use/availability by those lizards to whom home range size were estimated. Thus, RANGES V gave us the opportunity to assess the entire map and estimate habitat type over the study area. Also, it assessed habitat in home ranges and at fixes during the same run. After the analysis was run, the display showed the preference or avoidance of habitat at fixes using J a c o b s (1974) improved version of Ivlevs electivity index, which ranges from -1 (avoidance) to +1 (preference): D = (r - p) / (r + p - 2rp), where D is the electivity value, r is the used proportion of that microhabitat category and p is the available proportion of that microhabitat category. However, this design does not address the critical question of why a home range was selected. J o h n s o n (1980) cautioned that comparing habitat use within home ranges may be misleading because in choosing a home range, the animal has already made an important selection. Thus, the researcher may be studying a higher order of selection than that intended and possibly not the most important one for management purposes (T h o m a s & T a y l o r 1990). Animals may select a home range because one resource is abundant. Use of this resource may appear low when compared to availability within the home range, but higher when compared to the general area from which the home range was selected. So, we examined available habitat within a radius of 0.5 m around every animal recapture point and we also estimated habitat availability in the entire study area in order to increase accuracy. Food availability, absolute arthropod biomass, was measured using a 1 m3 cubic frame, covered with a fine nylon mesh but with one open side (B r o w n & P r e z - M e l l a d o 1994). During periods of lizard activity, the cube was placed at a randomly-selected point within each quadrat, which was selected pointing the finger in a random numbers table and counting 5 5 metre squares from 1 to the selected number from the table. This was repeated to provide a good coverage of the entire study area. An observer then inserted his head and upper body through a circular entry hole on one of the vertical sides of the cube and sealed the hole around the waist by means of an encircling cord. The nylon sides of the cube were trapped against the ground with sandbags to prevent arthropods from escaping. All arthropods trapped inside the cube were caught and frozen. Capture arthropods were later identified to order. Animals collected from each sample were dried to constant mass (at 65 C) and weighed. Results are expressed as arthropod availability per m2.
89

Statistical procedures Pearson product-moment correlation, Spearmans rank correlation, ANOVA, Student-NewmanKeulss a posteriori test, and the chi-square distribution were used to evaluate relationships, associations or differences among variables at 95 % statistical significance. Parametric statistics were used when the assumptions of the models could be reasonably met (S o k a l & R o h l f 1981). Test of homogeneity of variances (Levene test for homogeneity of variances) showed that in all cases variances were not significantly heterogeneous. Bias-corrected estimates of home range areas were log-transformed (base 10) to achieve normality (S o k a l & R o h l f 1981). A two-way ANOVA on log-transformed estimates of residents home range areas was performed to test significant differences between seasons and between lizard categories. Differences in home range areas between lizard categories (residents and introduced) during the summer, were evaluated by means of analyses of variance (ANOVA). We performed the analysis related to thermal properties of microhabitats used by lizards (Ts, Ta), Tb, and the types of behaviour lizards were engaged in, using just the first capture data of lizards, to avoid pseudoreplication. A multivariate analysis of variance (MANOVA) was performed on air and substrate temperatures at the capture point. We previously tested the assumptions associated with this analysis: multivariate normality, homogeneity of covariance matrices and independence. We checked whether the thermal properties of sand substrate and rocks microhabitats differed. Therefore, we tested the null hypothesis that the thermal properties of the habitat (i.e., air and substrate temperatures) did not change across seasons or across microhabitat types used by lizards. Next we tested the null hypothesis that time intervals had no effect on different microhabitat thermal properties. We carried out multivariate tests of the hypothesis that season interacts with time intervals. We used chi-square distribution to evaluate differences in microhabitat preferences of basking and moving lizards. The number of microhabitat categories was reduced (sand substrate, smooth rocks with no crevices, rocks < 50 cm, rocks > 50 cm) by pooling previous categories to achieve the assumptions associated with this analysis (S i e g e l 1956). Moreover, we used a chi-square distribution to assess microhabitat preferences of basking lizards. Only sand substrate and rocks microhabitats were included in this analysis because bush and grass microhabitats were too infrequent to be included. We examined, through a 3-way ANOVA, whether Tb differed among resident lizards caught in different microhabitat categories, seasons and time intervals. Microhabitat use may vary not only seasonally but also daily. Hence, we tested using a chi-square distribution, that the proportion of observed lizards at different time of day did not differ between spring and summer samples. All statistical analyses were done with SPSS for Windows (Ver. 9.0.1).

Results
Home ranges Bias-corrected estimates of home range areas were not significantly correlated with number of observations per lizard in either season (Pearson product-moment correlation; spring: n = 23, r = -0.085, P = 0.698; summer: n = 78, r = -0.082, P = 0.473). Observed home range area varied considerably among individuals in spring and summer (Table 1). Significant
90

differences in home range areas between seasons were found for resident animals, being greater during the summer (F[1, 88] = 4.260, P = 0.042), but no significant differences were found between resident lizards categories (F[2, 88] = 0.210, P = 0.811), nor interaction between these factors (F[2, 88] = 0.658, P = 0.521). We did not find significant differences in home range size between lizards categories, including introduced adult males, during the summer (one-way ANOVA; F[3, 74] = 1.158, P = 0.332). Resident animals showed a high degree of area stability (Table 2). Thus, the degree of home range overlap between spring and summer was significantly higher than 50 % (chisquare distribution; 2 = 8.067, df = 1, P = 0.005).
Table 1. Home range size (m2) for P. hispanica in the Sistema Central Mountains, Spain, in 1991. Spring n Resident adult males Adult females Juveniles Introduced adult males 7 13 3 Summer n 25 28 18 7

SE x
277.92 50.87 400.74 78.77 199.35 38.22

Minimum Maximum 113.45 467.88 53.13 1047.62 122.93 239.25

SE x

Minimum Maximum 1819.06 1260.56 1321.76 1807.61

482.08 77.88 80.24 448.43 61.09 112.96 495.11 79.16 100.29 891.09 242.72 99.78

Explanations: n, number of observations; x, mean; SE, standard error Table 2. Seasonal stability of area occupation by P. hispanica in the Sistema Central Mountains, Spain, in 1991. Percentage of home range overlap 50 was used as a criterion of fidelity to a home range. Adult males ID 61316 61317 61820 60911 60917 70911 % HRO 100.0 64.5 76.2 80.0 91.9 100.0 ID 61020 61318 61418 61419 61619 60918 71019 Adult females % HRO 28.2 52.1 74.0 56.7 100.0 68.6 86.5 ID 61516 70915 Juveniles % HRO 95.7 48.6

Explanations: ID, identification number of the specimen; % HRO, percentage of home range overlap.

Thermal properties of microhabitats used by lizards The thermal properties of habitat used did not differ between seasons (MANOVA; F[2,193] = 0.146, P = 0.865) nor between microhabitat types (MANOVA; F[10,386] = 1.760, P = 0.066). Thermal properties of microhabitat used by lizards varied significantly with respect to time of the day (MANOVA; F[4, 386] = 7.872, P < 0.001). Differences between thermal properties of microhabitat used by lizards did not depend on microhabitat types in conjunction with season (MANOVA; F[10, 386] = 1.140, P = 0.331) or time intervals (8:0010:00 / 10:0014:00 / 14:0020:00) (MANOVA; F[20, 386] = 1.174, P = 0.273). Thermal properties of microhabitat used by lizards across time intervals varied significantly according to season (F[4, 386] = 3.142, P = 0.015). Moreover, we found that the thermal properties of sand substrate and rocks microhabitats used did not differ among resident lizards engaged in different behaviours
91

(MANOVA; F[4, 432] = 1.727, P = 0.143). Lizard microhabitat use did not differ significantly when they were basking or moving (chi-square distribution; 2 = 1.678, df = 3, P = 0.642). Moreover, lizards basked more frequently on some microhabitats (i.e., rocks > 10 cm in height having a higher number of crevices) than on others (chi-square distribution; 2 = 84.193, df = 5, P < 0.001) (Table 3).
Table 3. Percentage of occurrence and frequency of resident lizards basking on different microhabitats of the Sistema Central Mountains, Spain, in 1991. Only the first capture data of lizards was used. Microhabitat Bush Grass Sand Smooth rocks with no crevices Rock < 10 cm Rock 10-50 cm Rock 50-100 cm Rock > 100 cm Frequency 2 1 20 10 12 68 22 39 Percentage 1.1 0.6 11.5 5.7 6.9 39.1 12.6 22.4

There was a significant difference between the animals cloacal temperature (Tb) in different types of microhabitats (F[5, 187] = 4.315, P = 0.001; Table 4). Tb was lowest on smooth rocks with no crevices and on rocks greater than 100 cm in height and highest on the other categories of rocky or sand substrates (Student-Newman-Keulss a posteriori test, P < 0.05). The mean Tb of P. hispanica in spring was higher than the mean Tb in summer (F[1, 187] = 8.516, P = 0.004; Table 4). There were also significant differences between time intervals (F[2, 187] = 8.313, P < 0.001; Table 4), Tb was lowest at onset and middle of the day and highest during the rest of the day (Student-Newman-Keulss a posteriori test, P < 0.05). Any differences between Tb do not reliably depend on microhabitat selected by lizards in conjunction with season (F[5, 187] = 0.346, P = 0.885) or time intervals (F[10, 187] = 0.778, P = 0.650), nor depend on season in conjunction with time intervals (F[2, 187] = 1.416, P = 0.245). Finally, there was no interaction of microhabitat type with season and time interval (F[7, 187] = 1.058, P = 0.392). Also, we found that Tb did not differ among resident lizards engaged in different behaviours (x basking = 30.11 0.20 C, n = 174; x moving = 31.01 0.36 C, n = 44; x resting = 31.66 0.63 C, n = 5), (one-way ANOVA, F[2, 220] = 2.755, P = 0.066). Habitat use and availability The proportion of observed lizards at different time of day did not differ between spring and summer samples (adult females: 2 = 4.898, d.f. = 2, P = 0.086; juveniles: 2 = 4.365, df = 2, P = 0.113), except marginally significantly in resident males (2 = 5.861, df = 2, P = 0.053). Substrate height was positively correlated with abundance of crevices (Spearmans rank correlation; rs = 0.737, P < 0.001). Electivity values for spring revealed that lizards were highly selective (Fig. 1). Interestingly, no relationships were found between these electivities and the availability of habitat variables (resident adult males: rs = -0.531, P = 0.062; adult females: rs = -0.358, P = 0.230; juveniles: rs = -0.113, P = 0.713). These data strongly suggest that the lizards were not selecting microhabitats because they were more abundant, but rather for some other reason.
92

Table 4. Summary statistics of lizard cloacal temperatures (Tb) (mean standard error; sample size in parentheses) in different microhabitat categories, seasons and time periods of the Sistema Central Mountains, Spain, in 1991. Variables Microhabitat Sand Smooth rocks with no crevices Rock < 10 cm Rock 10-50 cm Rock 50-100 cm Rock > 100 cm Season Time period Tb (C) 31.20 0.62 (24) 28.88 0.65 (12) 31.88 0.49 (18) 30.52 0.25 (87) 30.41 0.43 (34) 29.03 0.40 (45)

Spring 31.26 0.25 (103) Summer 29.45 0.22 (117) 8:00-10:00 29.81 0.35 (30) 10:00-14:00 29.19 0.28 (82) 14:00-20:00 31.27 0.24 (108)

Resident adult males, adult females and juveniles preferentially occupied sites with rocks > 100 cm in height characterised by providing the highest number of crevices (>10 crevices per 25 m2). Also, rocks from 50 to 100 cm in height were avoided by adult females and juveniles, whereas this type of microhabitat was preferred by resident adult males (Fig. 1). Electivity values for summer (Fig. 1) revealed that lizards were also highly selective. No relationship was found between electivities and availability of habitat variables, excepting adult females. Thus, resident adult males (rs = -0.534, P = 0.060), juveniles (rs = -0.416, P = 0.157), and introduced adult males (rs = -0.468, P = 0.107) were not actively selecting habitat in relation to habitat variables abundance. Adult females (rs = -0.630, P = 0.021) were actively selecting less abundant microhabitat variables. During the summer, resident adult males and adult females used similarly habitat resources, differing from introduced adult males and juveniles (Fig. 1). Mean arthropod density on the plot was 8.23 ind.m-2 (n = 73, 95% confidence interval: 6.56 - 9.90 ind.m-2). Mean dry arthropod biomass was 29.17 mg m-2 (n = 73, 95% confidence interval: 19.86 - 38.49 mg.m-2). Arthropods trapped (n = 588) were: Formicidae (22.62%), Homoptera (22.45%), Diptera (16.84%), Coleoptera (9.69%), Araneae (7.99%), Hymenoptera (6.46%), Heteroptera (5.78%), Orthoptera (5.10%), Lepidoptera (2.21%), Dyctioptera (0.85%).

Discussion
The overall abundance of lizards increased from areas with a dense cover of shrubs and herbaceous plants to bare and rocky areas in the Sistema Central mountains. Rocks were the most useful variable for distinguishing whether lizards were present or absent from a site. Rocks may be used as indicators of good basking conditions because they provide better thermal properties (thermally moderate and stable conditions) than grass or bush (B a u w e n s et al. 1996). However, only a narrow range of favourable basking sites was used, mainly rocks > 100 cm in height, which were more selected than smooth rocks with no crevices and rocks from 50 to 100 cm in height. Whereas other rocky microhabitats were not selected. The essential difference between these kinds of rocky substrata is the presence of crevices. Smooth rocks with no crevices and rocks from 50 to 100 cm in height provide fewer
93

refuges than rocks > 100 cm in height. Moreover, smooth rocks with no crevices could provide less refuge from predation since they lack shelters (G i l et al. 1988). This supports the habitat exploitation theory (T r a c y & C h r i s t i a n 1986), which predicts optimal patch selection when optimal and sub-optimal patches are equally accessible. Such a choice has a direct influence on an individuals fitness, therefore animals used sites having a higher number of crevices, possibly to reduce predation risk. Structural features clearly play a major role in habitat selection of this species, as has been described in other reptile species (H e a t w o l e 1977, A d o l p h 1990, B a l t o s s e r & B e s t 1990, F r i e n d & C e l l i e r 1990, S h e n b r o t & K r a s n o v 1997), and are related to environmental characteristics. Thus, taller rocks, which have abundant crevices, were not only optimal sites for avoiding predation, but favourable basking sites (P r e z - M e l l a d o 1983, G i l et al. 1988, D a z et al. 1996). We found that Tb and thermal properties of sand substrate and rocks microhabitats used by lizards did not differ among residents engaged in different behaviours. Also, lizard microhabitat use did not differ significantly when they were basking or moving. However, available data do not allow us to assess thermal properties of particular microhabitats, since we obtained temperatures at capture points (i.e., microhabitats selected by lizards). Although the thermal properties of basking sites are important, presumably they choice basking places influenced by other factors too. An animal basking to raise its body temperature, exposes itself to an increased risk of predation (H u e y 1974). Close access to a safe hiding place should be of great importance for survival (S c h a l l & P i a n k a 1980). Similarly, D o w n e s & S h i n e (1998) used habitat selection experiments with a rock-dwelling nocturnal gecko, Oedura lesuerii, to show that avoidance of predators was of greater priority than thermoregulation. Previous studies with P. hispanica from our study area revealed that the existence of rocky crevices appears to be the most important feature of its habitat (G i l et al. 1988). In fact, our population suffers from strong predation pressure (C a s t i l l a et al. 1999). Selecting the highest rocks, lizards can easily detect predators and rapidly seek cover, since these microhabitats offer the greatest number of crevices. Moreover, we frequently observed lizards basking a few centimetres from the entrance of the crevices, maintaining their bodies partially protected inside the hide. Also, crevices afford animals protection during inactive periods, when they may be vulnerable to predation (C l a r k & G i l l i n g h a m 1990). This resting behaviour is more frequent during the central hours of the day to compensate for the higher ambient temperatures (D i e g o - R a s i l l a & P r e z - M e l l a d o 2000). Some types of microhabitats were avoided, these electivities varying less than others, in particular zones covered by sand, bush or grass, that covered 43.3 % of the study site. Those areas with a dense cover of shrubs and herbaceous plants provided abundant trophic resources (D i e g o - R a s i l l a 1999), but food seems to be abundant throughout the site and would easily meet the low energetic needs of the lizards (P o u g h 1980, B r o w n et al. 1992, B r o w n & P r e z - M e l l a d o 1994). Thus, their spatial distribution is not influenced by food resources as has been described in other saurian populations in which habitat selection was influenced by habitat structure but not by food abundance (R u b y 1986, E d s m a n 1990, M c C l o s k e y et al. 1990). Moreover, we have to consider the fact that sandy areas, and those with a dense vegetation cover are preferred by some lizards predators, especially snakes (B e a & B r a a 1998, G a l n 1998, P l e g u e z u e l o s 1998). The habitat preferences of adult female lizards varied according to season, but those of males did not. During spring, adult females preferred rocky areas (rocks > 100 cm), but
94

Fig. 1. Habitat preference of P. hispanica in the Sistema Central Mountains, Spain, in 1991. Electivity was calculated using J a c o b s (1974) version of Ivlevs index; values range from -1 (avoidance) to +1 (preference).

95

during summer their use of other rocky substrates increased. This shift may be explained by behavioural modifications related to reproduction. Pregnancy adversely affects locomotor performance and it is associated with changes in antipredator tactics, tending toward crypsis (S h i n e 1980, B a u w e n s & T h o e n 1981, B r o d i e 1989, C o o p e r et al. 1990, S c h w a r z k o p f & S h i n e 1992). Behaviours such as movement influence susceptibility to predators (B u r g h a r d t & D e n n y 1983). Hence, they avoid detection by a predator because they are cryptic, but rapid flight is the option of choice when they have been detected by a predator. Rapid flight is often toward crevices, which are abundant in rocks > 100 cm in height. So, they are optimal places for avoiding predation during gestation (spring). Rocks > 100 cm in height were markedly selected by all lizard categories. Even those adult males brought into the plot from other areas preferred similar habitats to those of residents, as has been described in Anolis lizards (K i e s t e r et al. 1975). Levels of social interaction increased after introducing adult males (D i e g o - R a s i l l a & P r e z M e l l a d o 2000, 2001). High levels of social interaction might function to allow introduced adult males and juveniles to track resources by assessing the behaviour of neighbouring lizards (K i e s t e r & S l a t k i n 1974, S t a m p s 1987, 1988, G u y e r 1994). However, introduced adult males and juveniles occupied areas with fewer crevices than resident lizards and introduced adult males did not select rocks from 50 to 100 cm in height. They occupied habitats of marginal quality in relation to crevice abundance, perhaps because the best sites were already taken (C l o b e r t et al. 1994). Introduced adult males and juveniles were less competitive because they were not familiar with the habitat. Being familiar with the habitat, a lizard may also know the best places for basking, hiding and spending the night. Moreover, knowing the identity of other lizards in the area could save time, otherwise used in interactions, for activities such as feeding and basking (E d s m a n 1990, D i e g o - R a s i l l a & P r e z - M e l l a d o 2000). Competitive ability may be a major determinant of success in settling into a new environment, because established residents usually have an advantage in competition for sites with new immigrants (L e i m a r & E n q u i s t 1984, C o o p e r & V i t t 1987, E d s m a n 1990).

Acknowledgements
We thank R. M. L u e n g o , N. D i e g o and I. A u s n for help in the field. We also wish to thank R. P. B r o w n , P. B l a h k , L. G v o d k , G. H. C o p p and two anonymous reviewers for suggesting improvements in the manuscript. This research was partially supported by The Spanish Ministry of Education and Science with the Project PB98-0270.

L I T E R AT U R E
ADOLPH S. C. 1990: Perch height selection by juvenile Sceloporus lizards: Interspecific differences and relationship to habitat use. J. Herpetol. 24(1): 6975. BALTOSSER W. H. & BEST T. L. 1990: Seasonal occurrence and habitat utilization by lizards in southwestern New Mexico. Southwest. Natur. 35(4): 377384. BARBADILLO L. J., LACOMBA J. I., PREZ-MELLADO V., SANCHO V. & LPEZ-JURADO L. F. 1999: Anfibios y reptiles de la Pennsula Ibrica, Baleares y Canarias. Editorial GeoPlaneta, S. A. Barcelona. BAUWENS D., GARLAND T. Jr., CASTILLA A. M. & VAN DAMME R. 1995: Evolution of sprint speed in lacertid lizards: Morphological, physiological and behavioral covariation. Evolution 49(5): 848863. BAUWENS D., HERTZ P. E. & CASTILLA A. M. 1996: Thermoregulation in a lacertid lizard. The relative contributions of distinct behavioral mechanisms. Ecology 77(6): 18181830.

96

BAUWENS D. & THOEN C. 1981: Escape tactics and vulnerability to predation associated with reproduction in the lizard Lacerta vivipara. J. Anim. Ecol. 50: 733743. BEA A. & BRAA F. 1998: Vipera latasti (Bosc, 1878). In: Ramos M.A. et al. (eds), Reptiles. Fauna Ibrica, vol. 10. Museo Nacional de Ciencias Naturales, C.S.I.C., Madrid: 480488. BRODIE E. D. III 1989: Behavioral modification as a means of reducing the cost of reproduction. Am. Nat. 134: 225238. BROWN R. P. & PREZ-MELLADO V. 1994: Ecological energetics and food acquisition in dense Menorcan islet populations of the lizard Podarcis lilfordi. Funct. Ecol. 8: 427434. BROWN R. P., PREZ-MELLADO V., DIEGO-RASILLA F. J., GARCA J. A., NARANJO A. & SPEAKMAN J. R. 1992: Individual and population energetics of a lizard on a Mediterranean islet. Oecologia 91: 500504. BURGHARDT G. M. & DENNY D. 1983: Effects of prey movement and prey odor on feeding in garter snakes. Z. Tierpsychol. 62(4): 329347. CASTILLA A. M. & BAUWENS D. 1991: Thermal biology, microhabitat selection, and conservation of the insular lizard Podarcis hispanica atrata. Oecologia 85:366374. CASTILLA A. M., GOS A., GALN P. & PREZ-MELLADO V. 1999: Green tails in lizards of the genus Podarcis: Do they influence the intensity of predation?. Herpetologica 55(4):530537. CLARK D. L. & GILLINGHAM J. C. 1990: Sleep-site fidelity in two Puerto Rican lizards. Anim. Behav. 39: 11381148. CLOBERT J., MASSOT M., LECOMTE J., SORCI G., DE FRAIPONT M. & BARBAULT R. 1994: Determinants of dispersal behavior: the common lizard as a case study. In: Vitt L. J. & Pianka E. R. (eds), Lizard Ecology: Historical and Experimental Perspectives. Princeton University Press, Princeton, NJ: 183206. COOPER W. E. Jr. 1978: Home range criteria based on temporal stability of areal occupation. J. Theor. Biol. 73: 687695. COOPER W. E. Jr. & VITT L. J. 1987: Deferred agonistic behaviour in a long-lived scincid lizard (Eumeces laticeps). Oecologia 72: 321326. COOPER W. E. Jr., VITT L. J., HEDGES R. & HUEY R. B. 1990: Locomotor impairment and defense in gravid lizards (Eumeces laticeps): Behavioural shift in activity may offset costs of reproduction in an active forager. Behav. Ecol. Sociobiol. 27: 153157. DAZ J. A., DAZ-URIARTE R. & RODRGUEZ A. 1996: Influence of behavioral thermoregulation on the use of vertical surfaces by Iberian wall lizards Podarcis hispanica. J. Herpetol. 30(4): 548552. DIEGO-RASILLA F. J. 1999: Efectos de la densidad en la dinmica poblacional y la autoecologa de la lagartija ibrica (Podarcis hispanica) (Squamata, Lacertidae). Tesis Doctoral, Facultad de Biologa, Universidad de Salamanca, Salamanca, Spain. DIEGO-RASILLA F. J. & PREZ-MELLADO V. 2000: The effects of population density on time budgets of the Iberian wall lizard (Podarcis hispanica). Isr. J. Zool. 46(3): 215229. DIEGO-RASILLA F. J. & PREZ-MELLADO V. 2001: Density dependence of survival in the Iberian wall lizard (Podarcis hispanica). In: Prez-Mellado V. (ed.), Abstracts of the Fourth International Symposium on the Lacertids of the Mediterranean Basin. Govern de les Illes Balears, Menorca: 45. DOWNES S. & SHINE R. 1998: Heat, safety or solitude? Using habitat selection experiments to identify a lizards priorities. Anim. Behav. 55(5): 13871396. EDSMAN L. 1990: Territoriality and competition in wall lizards. Doctoral Thesis, University of Stockholm, Stockholm. FRETWELL S. D. 1972: Populations in a seasonal environment. Princenton University Press, Princenton, NJ. FRIEND G. R. & CELLIER K. M. 1990: Wetland herpetofauna of Kakadu National Park, Australia: seasonal richness trends, habitat preferences and the effects of feral ungulates. J. Trop. Ecol. 6: 131152. GIL M., PREZ-MELLADO V. & GUERRERO F. 1988: Estructura del hbitat y dominios vitales en Podarcis hispanica (Steindachner, 1870). Misc. Zool. 12: 273281. GALN P. 1998: Coronella girondica (Daudin, 1803). In: Ramos M.A. et al. (eds), Reptiles. Fauna Ibrica, vol. 10. Museo Nacional de Ciencias Naturales, C.S.I.C., Madrid: 75383. GROVER M. C. 1996: Microhabitat use and thermal ecology of 2 narrowly sympatric Sceloporus (Phrynosomatidae) lizards. J. Herpetol. 30(2): 152160. GUYER C. 1994: Mate limitation in male Norops humilis. In: Vitt L. J. & Pianka E. R. (eds), Lizard ecology: historical and experimental perspectives. Princeton University Press, Princeton, New Jersey: 145158. HEATWOLE H. 1977: Habitat selection in Reptiles. In: Gans C. & Tinkle D.W. (eds), Biology of the Reptilia, Vol. 7(A), Ecology and Behaviour. Academic Press, London: 137155.

97

HEWS D. K. 1993: Food resources affect female distribution and male mating opportunities in the iguanian lizards Uta palmeri. Anim. Behav. 46: 279291. HUEY R. B. 1974: Behavioral thermoregulation in lizards: importance of associated costs. Science 184: 10011003. JACOBS J. 1974: Quantitative measurements of food selection. Oecologia (Berl.) 14: 413417. JENNRICH R. J. & TURNER F. B. 1969: Measurement of non-circular home range. J. Theor. Biol. 22: 227237. JOHNSON D. H. 1980: The comparison of usage and availability mesasurements for evaluating resource preference. Ecology 61: 6571. KENWARD R. E. & HODDER K. H. 1996: RANGES V. An analysis system for biological location data. Institute of Terrestrial Ecology, Wareham. KIESTER A. R., GORMAN G. C. & ARROYA D. C. 1975: Habitat selection behavior of three species of Anolis lizards. Ecology 56: 220225. KIESTER A. R. & SLATKIN M. 1974: A strategy of movement and resource utilization. Theoret. Pop. Biol. 6: 120. KNIGHT T. W. & MORRIS D. W. 1996: How many habitats do landscapes contain?. Ecology 77(6): 17561764. LAW B. S. & BRADLEY R. A. 1990: Habitat use and basking site selection in the water skin, Eulamprus quoyii. J. Herpetol. 24(3): 235240. LEIMAR O. & ENQUIST M. 1984: Effects of asymetries in owner-intruder conflicts. J. Theor. Biol. 111: 475491. MCLOSKEY R. T., DESLIPPE R. J., SZPAK C. P. & BAIA K. A. 1990: Tree lizard distribution and mating system: the influence of habitat and food resources. Can. J. Zool. 68: 20832089. PREZ-MELLADO V., 1983: Activity and thermorregulation patterns in two species of lacertidae: Podarcis hispanica (Steindachner, 1870) and Podarcis bocagei (Seoane, 1884). Cinc. Biol. Ecol. Syst. (Portugal) 5: 512. PREZ-MELLADO V. 1998: Podarcis hispanica (Steindachner, 1870). In: Ramos M.A. et al. (eds), Reptiles. Fauna Ibrica, vol. 10. Museo Nacional de Ciencias Naturales, C.S.I.C., Madrid: 258272. PLEGUEZUELOS J. M. 1998: Malpolon monspessulanus (Hermann, 1804). In: Ramos M.A. et al. (eds), Reptiles. Fauna Ibrica, vol. 10. Museo Nacional de Ciencias Naturales, C.S.I.C., Madrid: 408427. POUGH F. H. 1980: The advantages of ectothermy for tetrapods. Am. Nat. 115: 92112. POUGH F. H., ANDREWS R. M., CADLE J. E., CRUMP M. L., SAVITZKY A. H. & WELLS K. D. 1998: Herpetology. Prentice-Hall Inc., Upper Saddle River, New Jersey. ROSE B. 1982: Lizard home ranges: Methodology and functions. J. Herpetol. 16(3): 253269. RUBY D. E. 1986: Selection of home range site in the female of the lizard, Sceloporus jarrovi. J. Herpetol. 20: 432435. SCHALL J. J. & PIANKA E. R. 1980: Evolution of escape behavior diversity. Am. Nat. 115: 551566. SCHWARZKOPF L. & SHINE R. 1992: Costs of reproduction in lizards: Escape tactics and susceptibility to predation. Behav. Ecol. Sociobiol. 31: 1725. SHENBROT G. & KRASNOV B. 1997: Habitat relationships of the lizard fauna in the Ramon erosion cirque, Negev Highlands (Israel). J. Zool. 241(3): 429440. SHINE R. 1980: Costs of reproduction in reptiles. Oecologia 46: 92100. SIEGEL S. 1956: Nonparametric statistics for the behavioral sciences. MacGraw-Hill Book Company. New York. SIMON C. A. 1975: The influence of food abundance on territory size in the iguanid lizard Sceloporus jarrovi. Ecology 56: 993998. SOKAL R. & ROHLF F. J. 1981: Biometry. W. H. Freeman and Company, San Francisco. STAMPS J. A. 1983: Territoriality and the defence of predator-refuges in juvenile lizards. Anim. Behav. 31: 857870. STAMPS J. A. 1987: Conspecifics as cues to territory quality: A preference of juvenile lizards (Anolis aeneus) for previously used territories. Amer. Nat. 129: 629642. STAMPS J. A. 1988: Conspecific attraction and aggregation in territorial species. Am. Nat. 131: 329347. THOMAS D. L. & TAYLOR E. J. 1990: Study designs and tests for comparing resource use and availability. J. Wildl. Manag. 54(2): 322330. TRACY C. R. & CHRISTIAN K. A. 1986: Ecological relations among space, time, and thermal niche axes. Ecology 67: 609615. WALDSCHMIDT S. 1983: The effect of supplemental feeding on home range size and activity patterns in the lizard Uta stansburiana. Oecologia 57: 15. WERNER E. E., GILLIAM J. F. & MITTELBACH G. C. 1983: An experimental test of the effects of predation risk on habitat use in fish. Ecology 64: 15401548. WERNER E. E. & HALL D. J. 1988: Ontogenetic habitat shifts in bluegill: the foraging rate-predation risk tradeoff. Ecology 69: 13521366.

98

Folia Zool. 52(1): 99108 (2003)

Could the small size of sunbleak, Leucaspius delineatus (Pisces, Cyprinidae) be an ecological advantage in invading British waterbodies?
Rodolphe E. GOZLAN1*, Adrian C. PINDER1, Sarah DURAND2 and Jonathan BASS1
1

Centre for Ecology and Hydrology Dorset, Winfrith Technology Centre, Winfrith Newburgh, Dorchester, Dorset DT2 8ZD, U.K; *e-mail: reg@ceh.ac.uk Purbeck Shool, Wareham, Dorset, U.K.

Received 21 January 2002; Accepted 12 August 2002

A b s t r a c t . Since the introduction of sunbleak (Leucaspius delineatus) to southern England in 1986, its life history characteristics (such as reproductive behaviour, early sexual maturity and an unusually small adult size) have contributed to its rapid dispersal. This study examines the length-weight relationships and age of this non-indigenous cyprinid to highlight the potential threat to native 0+ cyprinids. Sunbleak populations demonstrated an unusual growth pattern for a cyprinid, with an average of 42 % of its maximum growth occurring in the first year, followed by extremely low annual growth until death. Very few significant differences were found between the mean length of several sunbleak age groups and the length of native 0+ bream Abramis brama, roach Rutilus rutilus, bleak Alburnus alburnus and rudd Scardinius erythrophthalmus. We have also found that young-of-the-year of these cyprinids share the same food and habitat with all sunbleak year class, which in some places has had a detrimental impact on the recruitment of native species. Key words: bream Abramis brama, roach Rutilus rutilus, bleak Alburnus alburnus, rudd Scardinius erythrophthalmus, young-of-the-year, growth models, ageing, diet, invasive, ecological impact

Introduction
Sunbleak Leucaspius delineatus (Heckel, 1843) previously called the motherless minnow is of importance since it is a little known, invasive species that poses a potential ecological threat to many of the fish species native to Britain. The main feature of invasive species like the sunbleak is an ability to occupy an empty ecological niche within their new ecosystem (L v q u e 2000a), or to compete successfully with the native species already occupying a particular niche (W i t t e et al. 1992). The reason they are able to compete successfully with native species could either be that the ecosystem is already fragile (pollution, reduction of habitat etc.) (C a r l t o n & G e l l e r 1993, M a l a k o f f 1999, L v q u e 2000b) or that the new species competes during a particularly vulnerable aspect of native species life cycle (e.g. early life, reproduction, parasitism) (D a s z a k et al. 2000). Although sunbleak has been used as a biological model in experimental research (e.g. R a d a k o v & S i v e r t s e v 1972, S i e g m u n d & W o l f f 1972, D a r k o v 1975, G u l i d o v 1975, D o b l e r 1977, A n d r f e r 1980), very little is known about the sunbleak ecology, apart from its status in England (F a r r - C o x et al. 1996) as well as in Europe (L e l e k 1987) and some aspects of its reproductive biology (B i a o k o z et al. 1978, L e l e k 1987, C a s s o u & L e L o u a r n 1991).
* Corresponding author 99

Sunbleak were deliberately introduced at Two Lakes Fisheries in southern England in 1986 with other fish species from Europe. Since, it has spread rapidly and is now present in large numbers in several locations in southern England (F a r r - C o x et al. 1996). This success could partly be the result of sunbleak life history characteristics, including reproductive behaviour (batch spawner, nest guarder), early sexual maturity and an unusually small adult size for a cyprinid, which may favour its further dispersal in the British Isles. This study will examine length-weight and growth relationships in a population of sunbleak to identify potential overlaps in the use of habitat and food resources with native 0+ cyprinid species and to contribute to the knowledge of its life history. We discuss the potential ecological impacts on the native coarse fish populations and attempt to predict the habitat characteristics of watercourses, which should favour the sunbleaks future dispersal in Britain.

Materials and Methods

More than 10 000 sunbleak were sampled on 23rd April 2001 from Stoneham Lakes, Hampshire (National Grid Reference SU436 173, Fig. 1) using a 10m long seine net (mesh diameter of 5 mm). The three lakes at Stoneham were so overcrowded with sunbleak chasing food on the surface that only one netting was required to obtain our fish collection. A subsample of approximately 3000 fish was brought back to the laboratory alive for further research and 596 fish were anaesthetised and killed on capture using 2-Phenyl Ethanol and preserved immediately in 4 % formalin. Reading sunbleak scales proved to be particularly difficult due to their very small size (0.7 to 1 mm). Consequently, fish were aged using opercula, as their larger size facilitates annulus reading. At the laboratory, the fish fork length as well as wet weights were measured to the nearest mm and 0.01g (before removal of the opercula) respectively. Ageing was determined according to the number of annuli on the opercula and backcalculations were carried out according to F r a s e r (1916) using radii of annual rings measured from the most anterior part of the operculum. Back-calculations were made individually and then averaged in order to study growth by year-class and to remove the inter-annual variations (i.e. average of annual back-calculated growth for different age groups). Mean fork lengths were then plotted according to F o r d (1933) and W a l f o r d (1946) and parameters in Bertalanffys growth model (H o h e n d o r f 1966) were calculated in order to describe the asymptotic growth of sunbleak. In order to compare the growth of sunbleak with the growth of native young-of-the-year cyprinid species which share overlapping habitat needs in British watercourses, we have used the fork length at the end of the first winter of 104 0+ bream Abramis brama, 95 0+ roach Rutilus rutilus , 82 0+ bleak Alburnus alburnus and 69 0+ rudd Scardinius erythrophthalmus (unpublished data from Centre for Ecology and Hydrology database). An analysis of variance (ANOVA) was performed on the individual back-calculated data to compare average lengths at age for the sunbleak, with average length at the end of the first year for the native cyprinid species. The entire gut of 30 sunbleak was removed and the contents mounted on glass slides. Prey were identified to genus level only, due to the damage caused by pharyngeal teeth. Animal prey taxa were counted individually, but detrital material (mainly aufwuchs) was
100

Fig. 1. Sunbleak distribution map, based on a CEH 2001 questionnaire, fisheries survey results and F a r r - C o x et al. (1996). a- R. Test, R. Itchen, Two lake Fishery, Broadlands lake, Stoneham Lakes. b- Revels Fisheries. cEntire Bridgwater catchment including Kings Sedgmoor. Drain, Bridgwater & Taunton Canal, R. Parrett, Wych Lodge Lake, Combe Lake. d- The Huntspill, South Drain, North Drain, R. Brue, R. Alham, R. Whitelake. eSkegness, Lincolnshire.

assessed as a percentage of the gut fullness (see H y s l o p 1980). The relative importance of each animal prey taxon was indicated as a percentage of the total number of food items found in the guts. The Schoener Index was used as a diet overlap index (W a l l a c e 1981, W a l l a c e & R a m s a y 1982) between three different size classes
101

of adult sunbleak (FL<50mm; 50FL55; FL>55). It expresses the overlap between two compositional data sets as =1-0,5((|pxi-pxi|)
i=1 n

where pxi (pyi respectively) is the proportion of food category i in the diet of size classes x (y respectively): the weakest values of highlight little overlap.

Results
In Stoneham Lakes the majority of sunbleak (56%) ranged between 37 and 47 mm fork length following a plurimodal size distribution (Fig. 2). Sunbleak demonstrated an unusual

Fig. 2. Size distribution of sampled sunbleak populations (n = 368) from Stoneham Lakes (Hampshire). A moving average of period 2 has been calculated to highlight the various modes of the distribution (black line).

Fig. 3. Growth in fork length (mm) and standard deviation of sunbleak from Stoneham Lakes compared with UK of bream, roach, bleak and rudd.

102

growth pattern for a cyprinid, with very strong growth in the first year (average 42% of its maximum length) followed by extremely low annual growth thereafter (Fig. 3). When compared with other cyprinid species (roach, rudd, bream, bleak) living in the same type of habitat, sunbleak growth was significantly lower from year 2 to year 5 (Fig. 3) and few significant differences in length between the four 0+ cyprinid species and different sunbleak age groups were observed (Table 1). Variations in annual growth are derived from individual back-calculated data to remove the years effects (Table 1, Fig. 4). After the first year, unlike the four other cyprinid species, the sunbleak growth rate was low and remained low until death. This is confirmed by the exponential weight-length relationship (w = 0.0879e0.0576Fl, R2 = 0.95). The Walford plot (Fig. 5) revealed that a slow growing population (higher theoretical maximum length, L = 83 mm) as well as a high k value (k = 0.1536, t 0 = -0.404) was characteristic of sunbleak. Comparison of backcalculated standard lengths in successive age groups of sunbleak by different methods

Table 1. Mean fork length (MFL, in mm), standard error (SE), number of specimens (N) as well as range (MinMax) of four one year old native cyprinids and of several sunbleak age groups (1 to 5 years old) is given. An analysis of variance (ANOVA, Scheffe F-Test) between mean fork length of the four native cyprinid and the different sunbleak age groups was performed (ns indicates variance was not significantly different). Min-Max 26-43 29-52 31-64 33-60 21-49 28-56 34-52 38-61 44-65 N 69 82 104 95 70 70 67 57 20 SE 1.34 3.51 3.08 2.49 0.66 0.73 0.76 0.78 1.23 MFL 34 39 45 47 35 41 47 51 56 Species rudd bleak bream roach sunbleak 1 sunbleak 2 sunbleak 3 sunbleak 4 sunbleak 5 rudd ns ns p<0.05 ns ns p<0.05 p<0.05 p<0.05 bleak ns ns ns ns ns p<0.05 p<0.05 bream roach

ns p<0.05 ns ns ns ns

p<0.05 ns ns ns ns

Fig. 4. Back-calculated mean annual length growth rate (L) as well as mean annual weight growth rate (W) for sunbleak.

103

Fig. 5. Walford plots for sunbleak from Stoneham Lakes. Ultimate length l() and the equation of Walford line are given.

Fig. 6. Growth in fork length (mm) in successive age groups of sunbleak using different methods (Fraser, Rt = 2.4506 + 1.7931 Flt; Walford, Flt+1 = 0.8576Flt + 11.77 and Bertalanffys model, Flt = 83 * [1 - e-0.1536 (t - 0.404)] where Rt is scale radius at age & Flt is fork length at age).

(Fraser, Walford, Von Bertalanffy) highlighted a strong underestimation of growth calculated with the last two models (Fig. 6). The content of adult sunbleak guts highlight a large spectrum of prey types with a clear switch in the diet between April and May from chironomids (larvae and pupae) hydracarina and bryozoan toward zooplanktonic Cladocera (eg, Bosmina sp., Table 2). In May, different types of fish scales including those of sunbleak have also been found in large quantities in the guts of the three size-groups of sunbleak. The Shoenner Index highlights an increase in the diet overlap between the different size of adult sunbleak in May when compare to April (Fig. 7).
104

Table 2. Contents of the guts of three size-groups of sunbleak from Stoneham lakes caught during the spring 2001. Food items are expressed as percentage of the total number of items found in the guts and as percentage of gut fullness for the first four items (H y s l o p 1980). Date Size of fish examined (mm) Number of fish examined Benthic diatoms organic detritus inorganic detritus plant tissue Rotifera fish scales Bryozoan Statoblast Crustacean Bosmina Daphnidae Copepoda Ostracoda Argulus Oligochaeta Naididae Hydracarina Chironomidae larvae pupae adult Diptera adult Aphid 05/04/2001 50-55 n=4 11.3 29.9 5.0 12.5 20.6 0.7 5.3 0.7 29/05/2001 50-55 n =6 3.3 28.4 80.5 7.7 -

<50 n =8 6.9 17.4 1.3 34.8 1.6 0.0 3.2 7.9 3.2 1.6 14.0 3.2 6.3 1.6 11.0 -

>55 n=3 23.3 23.3 26.6 9.8 13.4 13.4 -

<50 n =5 8.5 0.1 88.7 1.4 -

>55 n=4 7.5 2.5 61 0.2 0.2 0.1 0.1

Fig. 7. Shoenner Index comparison between three classes of adult sunbleak size during April (black circle) and May (white circle).

Discussion
Sunbleak possess many important attributes to be an ideal invasive species. The short life span and early life reproduction aids the colonization process by increasing the turnover of the population (R o s e c c h i et al. 2001), the egg guarding behaviour of the males ensures high
105

survival of eggs and the batch spawning tactic (C a s s o u & L e L o u a r n 1991) provides the larvae with a better survival rate, with regard to temporal environmental variations. There is also a very important factor contributing to the rapid dispersal of sunbleak within a catchment, in that this species has the ability to spawn on any smooth objects such as branches, floating leaves, plastic debris, even the bottom of boats (C a s s o u & L e L o u a r n 1991, CEH unpublished data). This range of spawning substrate could aid dispersal of the eggs from upstream to downstream on floating objects but also on boats from downstream to upstream and even between different catchments via navigation canal connections. Sunbleak has already proved to be very well adapted to different watercourses in southern England since its introduction in 1987 (F a r r - C o x et al. 1996) and has since spread throughout the entire Bridgewater canal and River Brue catchment as well as mysteriously appearing in still waters such as Revels Fisheries (Dorset) and several unconnected lakes and ponds in the Somerset area. This unexplained dispersal highlights that sunbleak could spread even further if introduced to the neighbouring Bristol Avon catchment as this river provides access to the canal network of Britain. The unusual growth pattern of sunbleak is of particular importance when compared with native species, as some differences in length between 0+ roach, bleak, bream and rudd and several sunbleak age groups are not significant. Early life in fishes has always been described as a period of high mortality (B a l o n 1990, C o p p & K o v 1996, G o z l a n et al. 1998, 1999), but if food resources must be shared with juveniles and adults of sunbleak which are better adapted to competition (M a s u d a & T s u k a m o t o 1999), early life for these native cyprinids will be an even more vulnerable period (G a r d n e r 1996, B i s c h o f f & F r e y h o f 1999). After the first year, sunbleak are already participating in reproduction, which could explain the reduction in length growth rate, which appeared during the second year, as it will be related to the redistribution of their energetic reserves to gonad production (T a b o r s k y 1998, E s s i n g t o n et al. 2001). In terms of bioenergetics, resources allocated to one aspect of the life history will not be available for others (P a r t r i d g e & S i b l y 1991). According to the H e & S t e w a r t (2001) new predictive model based on von Bertalanffy coefficient (k, L), age at first reproduction in sunbleak should be between two and three years old. This discrepancy in the prediction of age at first maturity in sunbleak corresponds to an underestimation of sunbleak growth observed when calculated from Von Bertalanffy or Walford models. This underestimation is mainly due to the individual variability in growth, which can strongly influence the accuracy of estimates of population parameters (M o r i t a & M a t s u i s h i 2001, P i l l i n g et al. 2002). In the case of sunbleak estimates of growth, parameter variability among individuals is increased by the inevitable estimation errors of growth parameter for each individual associated with the sunbleak batch spawning strategy (see P i l l i n g et al. 2002). Young-of-the-year sunbleak have an extended hatching period from early May until the end of June allowing great individual variability in growth at the end of the first year (B i a o k o z et al. 1978). In the short term, spatial and temporal variations of food will increase in synergy with large populations of sunbleak, generating additional competition for food within native 0+ cyprinids in their feeding areas and accelerate following the development of several comparably sized sunbleak age groups. Although the life histories of roach, rudd, bream and bleak differ greatly from that of sunbleak, as juveniles they frequently share the same food
106

spectrum (B i a o k o z et al. 1978, M a n n et al. 1997). The increase of dietary overlap in May could explain in part the reduced growth in older sunbleak year classes highlighting a possible recruitment bottleneck. It is reasonable to speculate that, if sunbleak colonise new river catchments with suitable habitat (i.e. Great Ouse, Thames), it could potentially become locally dominant (i.e., as in the Bridgwater canal and River Brue catchments), with serious consequences for the native fish populations as other invasive species have demonstrated in Europe (i.e. topmouth gudgeon Pseudorasbora parva, nase Chondrostoma nasus) (R o s e c c h i et al. 2001).

Acknowledgements
The authors would like to thank Eastleigh and District Angling Club for allowing us access to Stoneham Lakes, the Nuffield Foundation as well as the Southern Science and Technology Forum and Southampton University for their financial support.

L I T E R AT U R E
ANDORFER B. 1980: The school behavior of Leucaspius delineatus (Heckel) in relation to ambient space and the presence of a pike (Esox lucius). Oecologia 47: 137140. BALON E. K. 1990: Epigenesis of an epigeneticist: the development of some alternative concepts of the early ontogeny and evolution of fishes. Guelph Ichthyological Reviews 1: 148. BIAOKOZ W., KRYZAWOSZ T. & ZACHWIEJA J. 1978: Rate of growth food composition and feeding coefficient for (Leucaspius delineatus Heckel) from lake Picek, Poland. Rocz. Nauk. Roln. Ser. H. Rybactwo. 98:924. BISCHOFF A. & FREYHOF J. 1999: Seasonal shifts in day-time resource use of 0+ barbel, Barbus barbus. Environ. Biol. of Fish. 56: 199212. CARLTON J. T. & GELLER J. B. 1993: Ecological roulette the global transport of nonindigenous marine organisms Science 261:BP 7882. CASSOU A-I. & LE LOUARN H. 1991: Etude de la reproduction de lable de Heckel (Leucaspius delineatus) dans un tang Breton. Bulletin Franais de la Pche et de la Pisciculture 322: 109120. COPP G. H. & KOV V. 1996: When do fish with indirect development become juveniles? Can. J. Fish. Aquat. Sci. 53: 746752. DARKOV A. A. 1975: Behavioral reactions of schooling fishes in individuals of the same species and the question of school formation. J. Ichthyol. 15: 691694. DASZAK P., CUNNINGHAM A. A. & HYATT A. D. 2000: Wildlife ecology Emerging infectious diseases of wildlife Threats to biodiversity and human health. Science 287: 443449. DOBLER E. 1977: Correlation between the feeding time of the pike (Esox lucius) and the dispersion of a school of leucaspius delineatus. Oecologia 27: 9396. ESSINGTON T. E., KITCHELL F. J. & WALTERS C. J. 2001: The von Bertalanffy growth function, bioenergetics, and the consumption rates of fish. Can. J. Fish. Aquat. Sci. 58: 21292138. FARR-COX F., LEONARD S. & WHEELER A. 1996: The status of the recently introduced fish Leucaspius delineatus (Cyprinidae) in Great Britain. Fisheries Management and Ecology 3: 193199. FRASER C. M. 1916: Growth of the spring salmon. Transaction of Pacific Society Seattle, Second annual meeting 1915: 2939. FORD E. 1933: An account of the herring investigations conducted at Plymouth during the years from 19241933. J. Mar. Biol. Assoc. 19: 305384. GARNER P. 1996: Microhabitat use and diet of 0+ cyprinid fishes in a lentic, regulated reach of the River Great Ouse, UK. J. Fish Biol. 48: 367382. GOZLAN R. E., MASTRORILLO S., DAUBA F., TOURENQ J-N. & COPP G. H. 1998: Multi-scale analysis of habitat use during late summer for 0+ fishes in the River Garonne (France). Aquatic Sciences 60: 99117.

107

GOZLAN R. E., MASTRORILLO S., COPP G. H. & LEK S. 1999: Predicting the structure and diversity of young-of-the year fish assemblages in large rivers. Freshw. Biol. 41: 809820. GULIDOV M. V. 1974: The effect of different oxygen conditions during incubation on the survival and some of developmental characteristics of Verkhovka (Leucaspius delineatus) in the embryonic period. J. Ichthyol. 14: 393397. HE X. Jr & STEWART D. J. 2001: Age and size at first reproduction of fishes: predictive models based only on growth trajectories. Ecology 82: 784791. HOHENDORF K. 1966: Eine Diskussion der Bertalanffy-Funktionen und ihre Anwendung zur Charakterisierung des Wachstums von Fischen. Kieler Meeresforschungen 22: 7097. HYSLOP E. 1980: Stomach content analysis-a review of methods and their application. J. Fish Biol. 17: 411429. LELEK A. 1987: The freshwater fishes of Europe. Threatened fishes of Europe. AULA-Verlag Wiesbaden 9, 343 pp. LVQUE C. 2000: Lhistoire des milieux modle les espces. La Recherche 333: 5657. LVQUE C. 2000: Symptomes de la mondialisation. La Recherche 333: 6367. MALAKOFF D. 1999: Biological control: Fighting fire with fire Australian biocontrol beats rabbits, but not rules. Science 285: 18411843. MANN R. H. K., BASS J. A. B., LEACH D. & PINDER A. C. 1997: Temporal and spatial variations in the diet of 0 group roach (Rutilus rutilus) larvae and juveniles in the river Great Ouse in relation to prey availability. Regulated Rivers: Research & Management 13: 287294. MASUDA R. & TSUKAMOTO K. 1999: School formation and concurrent developmental changes in carangid fish with reference to dietary conditions. Environ. Biol. of Fish. 56: 243252. MORITA K. & MATSUISHI T. 2001: A new model of growth back-calculation incorporating age effect based on otoliths. Can. J. Fish. Aquat. Sci. 58: 18051811. PARTRIDGE L. & SIBLY R. 1991: Constraints in the Evolution of Life Histories. Philosophical Transactions of the Royal Society of London Series B-Biological Sciences 332: 313. PILLING G. M., KIRKWOOD G. P. & WALKER S. G. 2002: An improved method for estimating individual growth variability in fish, and the correlation between von Bertalanffy growth parameters. Can. J. Fish. Aquat. Sci. 59: 424432. RADAKOV D. V. & SIVERTSEV M. A. 1972: Procedure for structural description of schooling movements in fish. J. Ichthyol. 12: 730733. ROSECCHI E., THOMAS F. & CRIVELLI A. J. 2001: Can life-history traits predict the fate of introduced species? A case study on two cyprinid fish in southern France. Freshw. Biol. 46: 845853. SIEGMUND R. & WOLF D. L. 1973: Circadian rhythm and group-suppression in Leucaspius delineatus (Pisces, Cyprinidae). Experientia 29: 5458. TABORSKY M. 1998: Sperm competition in fish: bourgeois males and parasitic spawning. Trends in Ecology and Evolution 13: 222227. WALFORD L. A. 1946: A new graphic method of describing the growth of animals. Biological Bulletin 90: 141147. WALLACE R. K. & RAMSAY J. S. 1982: Reliability in Measuring Diet Overlap. Canadian Journal of Fisheries and Aquatic Sciences 40: 347351. WALLACE R. K. 1981: An assessment of Diet-Overlap Indexes. Transactions of the American Fisheries Society 110: 7276. WITTE F., GOLDSCHMIDT T., WANINK J., VANOIJEN M., GOUDSWAARD K., WITTEMAAS E. & BOUTON N. 1992: The destruction of an endemic species flock quantitative data on the decline of the haplochromine cichlids of Lake Victoria. Environ. Biol. Fish. 34: 128.

108

NOTE
Folia Zool. 52(1): 109112 (2003)

Biometric relationships between body size and bone lengths in fish prey of the Eurasian otter Lutra lutra: chub Leuciscus cephalus and perch Perca fluviatilis
Gordon H. COPP1 and Vladimr KOV2
1

University of Hertfordshire, Hatfield, Hertfordshire and (current address) CEFAS, Pakefield Road, Lowestoft, Suffolk, NR33 0HT, UK; e-mail: g.h.copp@cefas.co.uk Comenius University, Faculty of Natural Sciences, Department of Ecology, Mlynsk dolina B2, 842 15 Bratislava, Slovakia; e-mail: kovac@fns.uniba.sk

Received 20 November 2001; Accepted 22 June 2002

Key words: dietary preferences, River Lee, cyprinids, percids A b s t r a c t . We determined regression relationships between head bone lengths and body size (standard length and weight) in chub Leuciscus cephalus and Eurasian perch Perca fluviatilis, two species taken frequently by the otter in the River Lee catchment. All relationships yielded significant linear equations, whereas those for bone vs. body weight were logarithmic. Regression slopes for head bones do not differ greatly within species but do across species of the same family.

Introduction
Investigations of prey size selectivity are important to understand the interactions between predators and their prey (M a n n & B e a u m o n t 1980, H a n s e l et al. 1988, M e h n e r 1990). Essential to dietary studies of fish predators, such as the European otter Lutra lutra (L.) (E r l i n g e 1967, C h a n i n 1981, K o e n et al. 1992, H a r n a 1993, J u r a j d a et al. 1996), are statistical relationships to estimate the size of the fish taken (W i s e 1980, P r e n d a & G r a n a d o - L o r e n c i o 1992, C a r s s & P a r k i n s o n 1996). Such studies provide a better understanding of both the composition and size of preferred fish prey, so as to assess how fish stocks can be managed to favour the maintenance and expansion of otter populations (e.g. K r u u k et al. 1993, L a n s z k i & K r m e n d i 1996). As part of investigations related to otter conservation in Hertfordshire, England (e.g. R o c h e et al. 1995), and in South Bohemia, Czech Republic (R o c h e 2001), the present study aimed to determine the biometric relationships between head bones and body size in two common species of fish prey. Cyprinid and percid fishes represent important elements of the otter diet throughout Europe, so we concentrated on the chub Leuciscus cephalus and the Eurasian perch Perca fluviatilis, two species of particular importance in the diet of otters in the River Lee catchment (G.H. C o p p & K. R o c h e , unpublished data) and for which relatively little biometric information exists.

Material and Methods

Specimens over a range of sizes, corresponding roughly to ages 1 to 5 i.e. those taken by otters in the River Lee catchment (G.H. C o p p & K. R o c h e , unpublished data), were collected from the River Lee in October 1995 (chub) at Woolmers Park (Nat. Grid Ref. TL 288 100) and in October 1996 (perch) from a quarry lake adjacent to the River Lee at Amwell (Nat. Grid Ref. TL 375 132). The specimens were captured at both sites using a Millstream,
109

generator-powered, DC electrofishing apparatus and additionally (at Amwell) a beach seine net. The specimens were killed using benzocaine and frozen. After defrosting, the fish were measured for standard (SL) length in mm and for weight in g. Chub were boiled until flesh was easily removed, but this was insufficient to loosen the flesh in perch, which were subsequently immersed in a solution comprised of one third 50% sodium hydroxide solution and two-thirds water; immersion was for 15 minutes, the solution heated from cold. In both species, the bones were left to air dry before measurement. Measurements of chub head bones followed P r e n d a & G r a n a d o - L o r e n c i o (1992) for cyprinids (dentaries, maxillae, premaxillae, pharyngeals), but also included the cleithra and opercula as per H a n s e l et al. (1988). For perch, we measured the four head bones suggested by L i b o i s et al. (1987): premaxillae, dentaries, opercula and preopercula. The dentaries were measured from the mandibular symphysis to the posterior ventral tip, the maxillae from the anterior edge to the posterior processus, the premaxillae from the maxillary symphysis to the anterior limb of the ascendant processus, the pharyngeals from the dorsal tip to the ventral tip (P r e n d a & G r a n a d o - L o r e n c i o 1992), the cleithra from the dorsal tip to the anterior tip and the opercula from the fulcrum tip to the primary ray tip (H a n s e l et al. 1988), the preopercula from the dorsal tip to the ventral tip (L i b o i s et al. 1987). Simple linear regressions were generated, as many relative growth characters in fish are related in this manner to their body length. Natural log (ln) transformation was used where necessary (e.g. body weight) to achieve a linear model. Other regression assumptions were examined to ensure the validity of the tests. The bone length data were regressed against SL, rather than the opposite, such as undertaken in some studies (e.g. P r e n d a & G r a n a d o - L o r e n c i o 1992) as we view the length of bones to be more dependent on the overall size of a fish than is overall size dependent on the length of any given bone. However, to permit comparison with some published equations, we also regressed SL or total length (TL) against bone length.

Results and Discussion

In both chub and perch, similar slope and intercept values were obtained in the regressions of bones from the left and right sides with respect to standard length (Table 1). The only exception was the pair of intercepts for left and right premaxillae in chub vs. body length. Otherwise, there appears to be little need in these species to distinguish the side of fish when back-calculating fish size from bone length. Our slope values (SL regressed against bone length) for chub are similar to those reported by P r e n d a & G r a n a d o - L o r e n c i o (1992) for the Iberian chub L. cephalus pyraenaicus (dentary b values common chub vs. Iberian chub, respectively = 11.76 vs. 12.57; maxilla b values = 12.36 vs. 12.18; premaxilla b values = 12.18 vs. 12.77; pharyngeal b values = 9.63 vs. 11.50). This suggests that the relationships between bone size and body length are relatively constant within species across geographical ranges. However, the slope and intercept values (TL regressed against bone length) within Cyprinidae (values from L i b o i s & H a l l e t - L i b o i s 1988) vary greatly, reflecting natural variability in ontogenies between the species compared. The onset of skeleton formation during early development of fish is species specific, and also strongly depends on environmental conditions, especially temperature and oxygen concentration (e.g. B a l o n 1981). L i b o i s et al. (1987) provided a bone length to body size relationship for only one of the four head bones they examined in perch: body length vs. preopercula length (pob) relationship of TL = 7.742 Lpob + 15.368 (r 2 = 0.93); their regression slope value
110

Table 1. Number of specimens, regression slope, intercept values, and coefficients of determination for linear and logarithmic relationships of bone sizes (in mm) regressed against standard length1 (bone size = bSL a) and body weight (bone size = b[lnWt] a) for the left (L) and right (R) sides of chub (mean SL = 231.7 mm, SE = 8.434, n = 59, min./max. = 124/375 mm) from the River Lee and perch (mean SL = 97.1 mm, SE = 4.815, n = 39, min./max. = 55/208 mm) from Amwell lake, Hertfordshire. In perch, the equation for bone size vs. weight is: ln(bone size) = b(lnWt) a. All models were significant at P 0.001. standard length (mm) species chub bone L-dentary R-dentary L-maxilla R-maxilla L-premaxilla R-premaxilla L-pharyngeal R-pharyngeal L-cleithrum R-cleithrum L-operculum R-operculum L-dentary R-dentary L-premaxilla R-premaxilla L-preoperculum R-preoperculum L-operculum R-operculum n 58 58 50 44 48 46 59 59 56 56 57 59 39 39 39 39 37 38 37 38 b 0.086 0.086 0.080 0.082 0.080 0.079 0.103 0.103 0.169 0.170 0.108 0.109 0.112 0.109 0.087 0.096 0.133 0.138 0.119 0.105 a -0.206 -0.425 +0.025 -0.329 -1.639 -7.781 -2.169 -2.412 -1.996 -2.333 -0.446 -0.746 -1.659 -1.283 -0.296 -1.097 -0.336 -0.172 -0.792 +0.496 r2 0.98 0.97 0.96 0.96 0.97 0.97 0.98 0.98 0.96 0.97 0.97 0.98 0.93 0.93 0.90 0.90 0.95 0.99 0.95 0.95 b 6.099 6.138 6.048 6.401 6.081 6.095 7.332 7.384 12.334 12.096 7.762 7.840 0.347 0.335 0.311 0.320 0.299 0.312 0.319 0.295 body weight (g) a -13.623 -13.941 -14.459 -16.412 -16.300 -16.601 -18.343 -18.695 -30.295 -28.885 -17.819 -18.138 +1.174 1.221 +1.162 1.139 -1.691 1.650 +1.426 1.494 r2 0.93 0.93 0.92 0.96 0.93 0.94 0.92 0.92 0.92 0.92 0.91 0.94 0.92 0.91 0.88 0.85 0.94 0.92 0.94 0.91

perch

1 Conversion equations between standard (SL), fork (FL) and total (TL) lengths are (G.H. C o p p , unpublished data): chub, FL = 1.118SL + 0.249; TL = 1.209SL + 0.297, TL = 1.081FL + 0.038; perch, SL = 1.09FL + 4.478, SL = 1.135TL + 6.323, FL = 1.039TL + 1.831 (all with r2 0.993).

corresponds well with our preopercula data (left side: 7.85; right side: 7.48) when plotted TL vs. bone length (Table 2). Head bones are particularly useful for identifying fish prey species in otter spraints, as they withstand digestion and are of good taxonomic value. But head bones
Table 2. Regression slope and intercept values for linear relationships (TL = bLbone a) between total length (TL in mm) and left-side head bone sizes (mm) for chub (see Table 1) and for bleak Alburnus alburnus (n = 31 and 27 for dentary and maxilla, respectively), carp Cyprinus carpio (n = 25, 25, 25), common bream Abramis brama (n = 15, 16, 16), gudgeon Gobio gobio (n = 15, 17), roach Rutilus rutilus (n = 88), rudd Scardinius erythrophthalmus (n = 17, 16), and tench Tinca tinca (n = 41, 41, 41) reported by L i b o i s et al. (1988). Coefficients of determination were all 0.95. species dentary b a maxilla b a pharyngeal b a chub 14.22 3.23 14.94 -0.08 11.64 23.54 bleak 18.07 -9.32 19.29 8.92 carp 14.44 -1.86 15.20 0.20 8.79 -2.13 c. bream 19.02 -1.89 19.52 7.04 17.75 7.21 19.60 -3.69 14.87 -1.52 gudgeon 16.68 6.26 roach rudd 19.31 -11.57 18.82 2.49 tench 23.52 -10.93 18.49 -4.95 13.81 -5.00

111

can occur in a low proportions, depending upon the size distribution of the prey population, e.g. 25 (dentaries, maxillae) to 40 % (pharyngeals) of otter spraints (P r e n d a & G r a n a d o - L o r e n c i o 1992). And when a larger fish prey is taken, the head is often discarded uneaten (E r l i n g e 1968), so other body bones such as vertebrae can be useful (W i s e 1980), as do scales (K. R o c h e , personal communication). Vertebrae also withstand digestion, but they are of less taxonomic value and, as with scales (K o e n et al. 1992), it is difficult to know their exact origin along the vertebral column.
Acknowledgements
VK was sponsored for visits to the UK by the EC and supported in part by the Slovak Scientific Grant Agency (project No. 1/6162/99. We thank J. L a t h e n and H. R i c h a r d s for technical assistance and K. R o c h e for comments to an earlier draft of the MS.

L I T E R AT U R E
BALON E.K. 1981: Saltatory processes and altricial to precocial forms in the ontogeny of fishes. Amer. Zool. 21: 573-596. CARSS D.N. & PARKINSON S.G. 1996: Errors associated with otter faecal analyses. Estimating prey size and distribution from bones. J. Zool. 238: 219332. CHANIN P. 1981: The diet of the otter and its relation with the feral mink in two areas of south-west England. Acta Theriol. 26: 8393. ERLINGE S. 1967: Food habits of the fish otter Lutra lutra L. in south Swedish habitats. Viltrevy 4: 371443. ERLINGE S. 1968: Food studies on captive otters Lutra lutra. Oikos 19: 259270. HANSEL H.C., DUKE S.D., LOFY P.T. & GRAY G.A. 1988: Use of diagnostic bones to identify and estimate original lengths of ingested prey fishes. Trans. Amer. Fish. Soc. 117: 5562. HARNA G. 1993: Diet composition of the otter Lutra lutra in the Bieszcady Mountains, south-east Poland. Acta Theriol. 38: 167174. JURAJDA P., PREK V. & ROCHE K. 1996. The autumnal diet of otters (Lutra lutra) inhabiting four streams in the Czech Republic. Folia Zool. 45: 916. KOEN I., URBAN P., TOURAOV I. & MAZUR I. 1992: The diet of the otter (Lutra lutra Linn.) in the Poana Protected Landscape Region. Folia Zool. 41: 107122. KRUUK H., CARSS D.N., CONROY J.W.H. & DURBIN L. 1993: Otter Lutra lutra numbers and fish productivity in rivers of North East Scotland. Symp. Zool. Soc. London 65: 171191. LANSZKI J. & KRMENDI S. 1996: Otter diet in relation to fish availability in a fish pond in Hungary. Acta Theriol. 41: 127136. LIBOIS R.M. HALLET-LIBOIS C. & ROSOUX R. 1987: lments pour lindentification des restes craniens des poissons dulaquicoles de Belgique et du nord de la France: 1 Anguilliformes, Gastrostiformes, Cyprinodontiformes et Perciformes. Fiches dostologie animale pour larchologie. Srie A : Poissons, No. 3. Juan-les-Pins: APCDA. LIBOIS. R.M. & HALLET-LIBOIS C. 1988: lments pour lindentification des restes craniens des poissons dulaquicoles de Belgique et du nord de la France. 2 Cypriniformes. Fiches dostologie animale pour larchologie. Srie A, No. 4. Juan-les-Pins: APCDA. MANN R.H.K. & BEAUMONT W.R.C. 1980: The collection, identification, and reconstruction of lengths of fish prey from their remains in pike stomachs. Fish. Manag. 11: 169172. MEHNER T. 1999: Zur Bestimmung der Beutefischarten aus Fragmenten der Wirbelsule bei der Nahrungsanalyse. Zool. Anz. 225 (3/4): 210222. PRENDA J. & GRANADO-LORENCIO C. 1992: Biometric analysis of some cyprinid bones of prey fishes to estimate the original lengths and weights. Folia Zool. 41: 175185. ROCHE K. 2001: Sprainting behaviour, diet, and foraging strategy of otters (Lutra lutra) in the Tebon Biosphere Reserve (Czech Republic). PhD Thesis, Institute of Vertebrate Biology, Academy of Sciences of the Czech Republic, Brno, 135 pp. ROCHE K., HARRIS R., WARRINGTON S. & COPP G.H. 1995: Home range and diet of re-introduced European otters Lutra lutra (L.) in Hertfordshire rivers. Aquat. Conserv. Mar. Freshwat. Ecosys. 5: 8796. WISE M.H. 1980: The use of fish vertebrae in scats for estimating prey size of otters and mink. J. Zool. (London)

112

instrukce.jpg (660x982x24b jpeg)