534 THE WILSON BULLETIN l Vol. 97, No. 4, December 1985

Transcription

1 534 THE WILSON BULLETIN l Vol. 97, No. 4, December 1985 Wilson Bull., 97(4), 1985, pp Differences in the feeding behavior of Little Egrets (Egretia garzetfu) in two habitats in the Camargue, France.-Relationships among predators, prey, and the habitats they use are of continued interest in behavioral ecology. Holling (Can. J. Entomol. 91: , 1959) distinguished between two types of predator responses to changes in prey abundance: numerical and functional. The first response occurs at the population level, and the second at the individual level. Studies of the functional response were pioneered by Gibb (Ibis 102: , 1960), L. Tinbergen (Arch. Neerl. de Zool. 13: , 1960), Royama (J. Anim. Ecol. 39: , 1970; Proc. Adv. Study Inst. Dynamics Numbers Popul. [Oosterbeek]: , 1970), and others on foraging theory, and their studies have become the stage upon which optimal foraging theory has developed. Recently, a number of studies have related prey density and distribution to predator behavior (e.g., J. Tinbergen [Ardea 69:1-67, 19811, Goss-Custard [J. Anim. Ecol. 39:91-113, 1970; J. Anim. Ecol. 46: , 19771, and Pienkowski [Anim. Behav. 31: , 1983; Mar. Ecol. Prog. Ser. 11: , 19831). In the Camargue, SouthernFrance, extensive work is being conducted on prey densities and the feeding behavior of Little Egrets (Egret& gurzetta) (Hafner, Terre et Vie 32: , 1978; Hafner et al., Ardea 70:45-54, 1982; Hafner and B&ton, Colonial Waterbirds 6:24-3 1, 1983). This study is part of that work. Here we (1) compare prey type and abundance and feeding behavior of birds at two different habitats, an artificial shallow marsh and a natural freshwater marsh; (2) compare feeding behavior of birds using different microhabitats; and (3) determine how extrinsic factors such as group size, time of day, and reduced water levels influence feeding behavior. Study area and methods. -The study was conducted in July 1984 on the Tour du Valat, a former 1900-ha farm located in the arid hypersaline eastern region of the Camargue Regional Park in the Rhone River delta. The Camargue includes an area of ca. 85,000 ha in Mediterranean France and is noted for its dry, windy climate and its mosaic of wetlands that include both temporary and permanent marshes (ranging from freshwater to hypersaline), and extensive active and abandoned rice fields with interconnecting canals. We used observation hides at two sites to watch feeding egrets that nested in the vicinity. St. Seren is a large natural, freshwater marsh that provides important feeding and roosting habitat for wading birds during the nesting season (April through July) and for waterfowl in winter. A large permanent hide constructed at the south end of the marsh has been used for a number of years to census aquatic birds (Station Biologique de la Tour du Valat, unpubl. reports). This hide served as one of our observation points. The portion of the marsh that could be observed from our hide (24 ha) was classified into three microhabitat types: open water (predominant), edges (5 1 m) of Scirpus spp. patches, and interior portions of Scirpus patches. St. Seren was subject to continuous drying during the observation period and was completely dry by 19 July. The other site, field C 13, is an artificial shallow marsh, one of a series of interconnected marshes whose water levels are maintained during the summer. This 1 -ha field was composed of a relatively uniform cover of grasses and sedges ca cm in height. Water levels remained at about 10-l 5 cm during July. Such marshes become important feeding sites for wading birds as natural marshes dry up (Hafner 1978). This marsh and several adjacent ones had been used by egrets throughout the nesting period, therefore a 3.5 m high hide had been erected adjacent to C 13 in June. The two sites were selected to represent the two most important feeding habitat types for wading birds in the eastern Camargue (Hafner 1978, Hafner et al. 1982). Observations offeeding birds were conducted between 3 and 26 July A prey sampling

2 GENERAL NOTES 535 TABLE 1 PREYABUNDANCE~FORTHROW-TRAPSAMPLESATST.SERENAND Cl3 MARSHES Prey St. seren No. of prey items per sample Cl3 Fish 1.0 * Tadpoles 0 7.9?Z 6.8 Hydrous piceus Other invertebrates 61.1 & Total prey 62.4? k 15.0 "Means *HE for%. Seren(N = 2l)andCl3(N = 15)samplestaken2 July and 12July,respectively. scheme was devised using one-m2 throw traps (Kushlan, Trans. Am. Fish. Sot. 103: , 1974). Fifteen samples were taken from field Cl 3 on 12 July, and 21 samples were taken at St. Seren on 2 July. Samples were taken 5-10 m apart along a randomly selected transect line at each site. Earthwatch participants (Earthwatch is a Massachusetts-based nonprofit conservation organization that supports scientific field research through the use of lay volunteers who provide their time and financial support) conducted the sampling using hand sieves to collect potential prey organisms in the water column and the top O cm of bottom sediments within the one-m2 trap perimeter. Hand sieving continued until three consecutive sweeps produced no organisms. Prey were removed with forceps and placed in labelled glass vials containing alcohol. They were later identified (to family or genus usually), counted, and measured in the laboratory. Feeding egrets were watched with binoculars and spotting scopes from the hides. Data were collected continuously using cassette tape recorders and blocked into 5-set increments on standard forms. Observation periods lasted from 30 min to 3 h, depending on numbers of birds using a site. Our goal was to observe at least 5 individuals for at least 5 min each during each observation period. The following data were collected: date, time, and numbers of feeding birds seen upon our arrival; feeding rates (FR, the number of pecks per min), success rate (SR, the number of captures per min), efficiency (EF, or SRFR), movement rate (MR, the number of steps per min), turning rate (TR, the number of changes in direction [N, E, S, W] per min while feeding), and stirring rate (STR, the number of foot-stirs per min). We also recorded the number of other feeding birds within 5 m of the focal individual each min. Focal individuals were chosen haphazardly and, whenever possible, followed for a minimum of 5 min during a feeding bout. Notes were taken of agonistic interactions. At St. Seren, the microhabitat of the feeding individual was recorded each minute as either open water, Scirpus patch, or edge of patch. Eight feeding areas of different sizes were designated on a map using microhabitat features as landmarks. These areas were denoted to monitor spatial and temporal changes of feeding birds and to monitor indirectly changes in water levels. Water level readings could not be measured directly for logistical reasons. Identification of prey items taken by egrets was usually not possible because of the small size of the prey (except for relatively large water beetle [Hydrous piceus] larvae, which were frequently 6-7 cm long). We conducted I-tests and least squares regression analyses using the Statistical Analysis System (Helwig and Council, eds., SAS User s Guide edition, SAS Institute, Cary, North Carolina, 1979). For most of the analyses, the individual bird was the sampling unit;

3

4 GENERAL NOTES 537 TABLE 3 RELATIONSHIP BETWEEN GROUP SIZE= AND FEEDING BEHAVIOR OF LITTLE EGRETS Group sizeb Behavior Solitary (N = 3 I) Few (N = 32) Many- (N = 43) Pecks/min 2.62? ? CD Captures/min 1.80 f * CD Captures/peck 0.76 t ?Z c Steps/min CD Changes in directiommin f CD Foot-stirs/min t * Group size based on average no. of birds within 5 m of focal bird; see text for size definitions. b Means + SE given. L Results of Duncan s and Scheffe s MR tests shown. A CT indicates a significant difference (P < 0.05) between column 3 and column I values; a CD indicates that column 3 value is significantly different from both column 1 and column 2 values. test); however, the other four variables (FR, SR, TR, STR) did not vary significantly with microhabitat (P > 0.05). To compare sites, we pooled the St. Seren sample, excluding the movement rate data that showed microhabitat variation. Because of the influence of time of day (above), we compared only morning observations. Nearly a twofold difference in rates was found between sites (Table 2). Little Egrets feed in a variety of social contexts, ranging from territorial (solitary) feeding to large groups (Hafner et al. 1982, P. Dugan, unpubl. data). We recognized three groupsize classes of neighboring feeding birds, based on the average number of neighbors (within 5 m) observed at the beginning of each I-min sample period. Birds were classified into solitary, few (mean >O, <2.0), and many (mean >2.0). For FR, SR, MR, and TR, birds with many neighbors had higher rates than solitary feeders or those with few neighbors (Table 3). Efficiency was higher for solitary birds than for those in larger groups. Discussion.-A twofold difference in overall prey density at the two sites was mirrored in nearly twofold differences in feeding, success, and turning rates and even larger differences in stirring rates. Caution must be exercised, however, in directly relating prey densities to feeding behavior. The sites also differed in microhabitats. St. Seren was primarily open water with mud bottom, while the rice field had a relatively uniform grass-sedge cover. Thus, prey availability, in addition to density, may also contribute to differences in feeding behavior. The results found in this study concerning movement rates are opposite to those reported by Hafner et al. (1982) in the same region. Whereas the egrets we watched moved nearly twice as rapidly in St. Seren as in C 13, Hafner et al. ( 1982) found that movement was greater in rice fields than marshes. The rice fields studied by Hafner et al. were different enough from our artificial marsh that a direct comparison is probably not justifiable. The influence of other birds feeding nearby resulted in major differences in feeding behavior. In general, egrets in large (>2) groups were much more active than those feeding alone or in pairs. This difference may reflect local enhancement, i.e., an aggregation of birds at localized sites after one or two birds discover a hot spot, begin feeding rapidly, and in turn attract others. We noted this behavior on several occasions at St. Seren. Hafner et al.

5 538 THE WILSON BULLETIN l Vol. 97, No. 4, December I985 (1982) found that the percentage of successful pecks in flocks of birds was greater than for solitary birds; however, our results showed that solitary birds were more efficient than those in larger groups. This effect could be caused by direct interference in larger groups, or by a change in prey behavior with increases in the number of predators (K. Bildstein, pers. comm.). Acknowledgments. -The senior author wishes to express his warmest appreciation for the opportunity to work at the Tour du Valat. The director, Dr. Luc Hoffman, his family, and the staff at the Station made the experience most rewarding in many ways. We also appreciate the support of all the valuable Earthwatch project participants. J. Wallace and R. Lansdown provided field assistance, logistical support, and identified the prey organisms throughout the project. K. Bildstein, S. Gregory, H. Ohlendorf, J. Rodgers, D. White, and an anonymous reviewer provided constructive comments on an earlier draft. C. Bunck and B. Dowel1 assisted in statistical analyses and K. Hall typed the manuscript.-r. MICHAEL ERWIN, U.S. Fish and Wildlife Service, Patuxent Wildlife Research Center, Laurel, MD 20708; HEINZ HAFNER AND PATRICK DUGAN, Station Biologique de la Tour du Valat, le Sambuc, Aries, France. (Present address of PD: International Union for the Conservation of Nature, CH- I I96 Gland, Switzerland). Accepted 14 May Wilson Bull., 97(4), 1985, pp Interactions between Great Blue Herons and gulls. - Previous reports of interactions between Great Blue Herons (Ardea herodias) or Gray Herons (A. cinerea) and gulls (Larus spp.) have been largely anecdotal, and they indicate that the interactions are generally immediately selfish. For instance, interactions have included food piracy (Lowe, The Heron, Collins, London, England, 1954; Dummigan, Br. Birds 70: 117, 1977; Willard, Condor 79: , 1977; Quinney et al., Can. Field-Nat. 95: , 1981), mutual cueing to each other s feeding flocks (Marshall, Br. Birds 54:202, 1961), or gulls mobbing predatory Gray Herons in gull colonies (Axell, Br. Birds 49: , 1956). There have, however, also been several accounts of gulls attacking Great Blue Herons (Imhof, Wilson Bull. 62:210, 1950) or Gray Herons (Lowe 1954:109-l 10; Birkhead, Br. Birds 66: , 1973), where an immediate purpose was not evident. Here, I examine the interactions of Great Blue Herons and various gulls, describe interaction types, determine interaction frequencies, and ascertain if all interactions are immediately purposeful. Study area and methods. -All observations were at the 15.8 km2 Yaquina Estuary (approximately N, W), on the midcoast of Oregon. From 1973 through 1981, I recorded over 1000 h of observations (including min watches) of herons in intertidal areas as they foraged in the 3-h interval before and after low tide. Only interactions during 60-min watches were used to determine relative interaction frequencies, but notes from supplementary observations were used for other analyses. Sample sizes differ for various aspects of interactions because all characteristics of each interaction were not always recorded. All statistical tests are two-tailed. Because of the frenzied activity generally accompanying interactions between herons and gulls, and because of the subtle plumage differences between Western Gulls (L. occidentalis), Glaucous-winged Gulls (L. glaucescens), Western x Glaucous-winged Gull hybrids (see Hoffman et al., Auk 95: , 1978), Herring Gulls (L. argentatus), and Thayer s Gulls (L. thayerz), I was unable to distinguish among these species of gulls. My impression, however, was that 90% or more of these large gulls were Western Gulls or hybrids. Other