No job name

U ϭ 4.0, P ϭ 0.028), when birds are only moving away from the shelf area during the first day of foraging. The difference wasonly due to the higher sinuosity of tracksduring Henri Weimerskirch,1,2* Francesco Bonadonna,1 (0.294 Ϯ 0.084), where birds are known to Fre´de´ric Bailleul,1 Ge´raldine Mabille,1 Giacomo Dell’Omo,3 catch most prey (9), compared with when birds were over the shelf itself (0.693 Ϯ0.182; Wilcoxon paired test, Z ϭ 2.42, P ϭ0.015) or over oceanic waters (0.648 Ϯ Developments in satellite telemetry have recent- straight-line course). The ratio was not affected 0.09). Thus, birds increase the sinuosity of ly allowed considerable progress in the study of by wind direction with respect to overall their flight only over a specific area, the long-range movements of large animals in the route direction because birds always have a wild (1), but the study of the detailed patterns of When foraging, birds landed regularly on the their foraging behavior on a small to medium sea surface (Fig. 1, A and B), on average every scale is not possible because of the imprecision P ϭ 0.429]. Predators foraging in a hetero- 1.8 Ϯ 0.9 hours, and drifted when sitting on the of satellite telemetry systems (2). We used a geneous environment are expected to adjust water. The overall direction of the drift was miniaturized Global Position System (GPS) that their search pattern (e.g., the straightness of partly due to wind direction, but marine currents intervals (3) to examine the exact breeding birds (3) either starting a turbulence such as small gyres (Fig.
4.5 km hourϪ1) when birds are inflight. When in flight, birds fre- References and Notes
1. B. A. Block et al., Science 293, 1310
2. G. C. Hays et al., Anim. Behav. 61,
Fig. 1. Movements of wandering albatrosses moving from Possession Island,
Crozet Archipelago: (A) over oceanic waters during the incubation period and
able on Science Online at www.
(B) during the brooding period in the vicinity of the island. Part of tracks
(4, 5). Small-scale flight paths show showing (C) the movement in relation to ground speed and (D) the drifting
movement of a bird over the shelf break. Upward triangles indicate take-offs 4. T. Alerstam, G. M. Gudmundsson, B.
and downward triangles landings on water, the stars are conjunctions of Larsson, Philos. Trans. R. Soc. London landings and take-offs, and large bold arrows represent the wind direction and Ser. B 340, 55 (1993).
speed. Dashed lines indicate bathymetric contours; the shelf area is consid- 5. C. J. Pennycuick, Philos. Trans. R. Soc. ered at depths shallower than –500 m, the shelf edge at depths between London Ser. B 300, 75 (1982).
6. H. Weimerskirch et al., Proc R. Soc. –500 and –2000 m, and oceanic waters over waters deeper than –2000 m.
London Ser. B 267, 1869 (2000).
costs (4 – 6), wandering albatrosses 7. E. Batschelet, Circular Statistics in Bi- have to adjust their searching behavior according their route, the flight speed, and/or turning to wind conditions, but at the same time they rate) to increase the probability of encoun- 8. G. H. Pyke, Annu. Rev. Ecol. Syst. 15, 523 (1984).
must adjust their foraging movements to in- tering prey (8), but this prediction is gener- 9. H. Weimerskirch, P. Doncaster, F. Cue´not-Chaillet, crease the probability of encountering prey. The ally impossible to test on marine animals.
Proc. R. Soc. London Ser. B 255, 91 (1994).
zigzagging small-scale movements added to the 1Centre d’Etudes Biologiques de Chize´, Centre National de la larger scale changes in overall direction affect straightness of their movements according Recherche Scientifique, 79360 Villiers en Bois, France. 2Insti- overall the sinuosity of the track. The straight- to the season or the marine habitat visited.
tut Franc¸ais pour la Recherche et la Technologie Polaire, ness index of the path, as measured by the ratio The straightness index of the track was low- 29280 Plouzane´, France. 3Division of Neuroanatomy and of straight-line distance between the initial and er during the brooding period (0.41 Ϯ 0.1), Behavior, Institute of Anatomy, University of Zurich, Winter- final positions of two consecutive landings when birds are searching for food close to thurerstrasse 190, CH-8057 Zurich, Switzerland.
relative to the actual path (7), was on average the colonies (9), compared with the incuba- *To whom correspondence should be addressed. E- 0.512 (range ϭ 0.72 to 0.280, with 1.0 being a tion period (0.588 Ϯ 0.09; Kruskal-Wallis, www.sciencemag.org SCIENCE VOL 295 15 FEBRUARY 2002 Henri Weimerskirch, Francesco Bonadonna, Frédéric Bailleul, Géraldine Mabille, Giacomo Dell'Omo, and Hans-Peter Lipp
Supplementary Material

Methods: The study was carried out in January to April 2001, on the Crozet Islands. We used a fully self-
contained GPS-MS1 receiver with an onboard non-volatile memory that stores up to 100,000 positions. Details of the GPS are given by I. Steiner et al. [Physiol. Behavior 71, 1-8 (2000)]. The GPS used has a circular error probability of 4 m for horizontal position. Accuracy for altitude was lower and was not used, especially because albatrosses rarely fly over the sea at altitudes higher than 20 m. The GPS devices, weighing 105 g (1-1.3% of birds mass) including batteries and waterproof packaging, were taped to back feathers on the bird leaving the nest after a change over by its partner. Eight GPS units were deployed during the incubation period (average trip duration 8.26 ± 1.88 days) and nine units were deployed during brooding (trip duration 2.99 ± 1.42 days). Wind speed and wind directions were derived from meteorological models that estimated twice daily the wind strength and direction from NOAA/NESDIS, based on near real-time data collected by NASA/JPL's SeaWinds Scatterometer aboard the QuikSCAT satellite. Wind direction was mainly from the west. In order to acquire the precise movements of birds compared to obtaining movements over longer periods with lower resolution, the loggers were programmed to run in continuous mode by measuring one fix every second for a period of 27.7 hours from the time the GPS was started before the memory was full. This study was supported by Institut Français pour la Recherche et la Technologie Polaire and by a grant from Swiss National Science Foundation SNF 31-58822.99. We thank G. Merlet for help with the packaging of the GPS, Ralf Lashefski-Sievers (GFT) for help with the GPS, Armel and Charles for help in the field, and Scott Shaffer and Yves Cherel for helpful comments on earlier drafts of the manuscript. Additional legend to Fig. 1. In (A), the movement of a male was recorded for 20.2 hours at sea with the bird
spending 68.8% of its time in flight and covering a total distance of 1014 km (i.e., 996 km in flight and the rest drifting). The average flight speed was 71.6 km hour-1 and the overall straightness ratio of the track between two landings was 0.63. In (B), a female's movement was recorded for 23.4 hours at sea during which she spent 38.9% in flight, flew a total distance of 625 km (i.e., 580 km in flight and the rest drifting on the water), traveled at an average flight speed of 63.7 km hour-1, and had a straightness ratio of 0.36. (C) Smaller scale part of a flight bout with cross-head winds; ground speeds progressively decreased when birds soar against the wind or with side winds, while speeds increased abruptly after birds has oriented from a head to tail wind. (D) Drifting movement on the sea surface after landing over the shelf edge.

Source: http://www.cebc.cnrs.fr/publipdf/2002/WSci295.pdf