Nocturnal passerine migrants possess at least two different compass systems for orientation based on the geomagnetic field and celestial cues (stars, sun and the related pattern of skylight polarisation; for reviews see, for example, Emlen 1975; Able 1980; Wiltschko & Wiltschko 1995). At high geomagnetic latitudes, the magnetic field lines are close to vertical (inclination +90° at the magnetic North Pole), and might be difficult to use for orientation (Wiltschko and Wiltschko 1995, but see also, Sandberg et al. 1991, 1998). In addition to this complication, stars are not visible in these areas for a large part of the Arctic summer. Therefore, the most prominent visual cues used for orientation are the sun and the related pattern of skylight polarisation. The availability of celestial cues  is related to the elevation of the sun and becomes particularly prominent during the twilight period (cf. Figure 1 in Åkesson et al. 1996). A description of the pattern of partially linearly polarised skylight is given in Wehner (1989). Nocturnal passerine migrants departing on migration flights during the twilight period are prone to make use of visual sunset cues for orientation (for references see Moore 1987, Åkesson and Bäckman 1999). However, it is not completely understood how visual cues at sunset are related to geomagnetic information for orientation. The relative importance of visual and geomagnetic cues becomes especially intriguing in areas where birds on migration experience changing angles of declination, for example, at high geomagnetic latitudes (cf. Alerstam et al. 1990).

In this study we carried out cage experiments with two species of autumn migrating sparrows to investigate the relative importance of geomagnetic and visual sunset cues (i.e. sun position and skylight polarisation pattern) for orientation at high geographic latitudes. The experiments were performed in manipulated magnetic field conditions under natural, clear and overcast skies and in the local geomagnetic field with clock-shifted birds.

Methods

At the end of the breeding period between 15 July and 8 August 1999, we captured savannah sparrows Passerculus sandwichensis and white-crowned sparrows Zonotrichia leucophrys in Inuvik (68°21’N, 133°43’W), Northwest Territories in Canada for orientation studies. Savannah sparrows are common breeders all across mainland Canada and winter in the southern United States, down to Mexico (Wheelwright and Rising 1993). White-crowned sparrows breed in the northern part of mainland Canada, from Alaska eastwards to Labrador (Chilton et al. 1995). Both species are expected to migrate in a geographic southeast to south direction from Inuvik in autumn.

We trapped birds in deciduous shrub and open grassland habitats in close vicinity of the town of Inuvik mainly with mist nets but also with snap traps. From the birds captured, we selected adult and juvenile white-crowned sparrows and juvenile Savannah sparrows for cage experiments. A group of 30 white-crowned sparrows (15 adults and 15 juveniles) were displaced by an icebreaker to the eastern Canadian Arctic, where they were released near Iqaluit 2 September 1999 (for description of displacement experiments see, Åkesson et al. 1999). All birds that we caught were aged, sexed, measured and ringed. For those birds that were changing their feathers (moulting), we recorded the degree of moult at capture. Birds that were not used for cage experiments were released at the site of capture. The age of the birds was identified according to Pyle et al. (1987). Savannah and white-crowned sparrows were used for orientation studies later during the autumn migration period.

During captivity we kept the birds in individual cages indoors in a room with large windows facing geographic west. The birds were exposed to the natural light regime, and artificial lights were used to increase light intensity inside the room during daytime. The birds were fed mealworms, seeds and fruit ad libitum, and water with vitamins. We measured the birds’ body mass and fat content (9-graded visual scale for fat classification, cf. Pettersson & Hasselquist 1985) during the complete study period. Thereby we could record how the experimental birds increased in body mass and put on large amounts of fat prior to migration.

Experimental procedure

In total we performed 404 cage experiments with white-crowned sparrows and 193 with savannah sparrows between 6 August and 7 September 1999 in the backyard of the Aurora Research Institute in Inuvik. During this period we experienced variable weather conditions, with long periods with overcast conditions and rain, resulting in reduced opportunities for the birds to experience cloudless natural clear skies.

Magnetic field manipulations

We used electromagnetic coils (modified Helmholz coils, 800×800 mm), powered by car batteries (12 V), arranged in pairs around the orientation cages to manipulate the horizontal component of the geomagnetic field. The coils were constructed to produce a homogenous field in the centre, where the orientation cages were placed (for technical specification of the coils see, Sandberg et al. 1988).

A group of 18 white-crowned sparrows were exposed to shifted magnetic fields (90° deflection of magnetic north counter-clockwise) under natural clear skies for one hour starting at local sunset. The test was repeated at least twice and the birds’ orientation was registered in circular cages, so-called Emlen funnels (Emlen and Emlen 1966). All experiments before and after the magnetic shifts were performed in the local geomagnetic field both under clear skies and in overcast conditions. A second group of 18 young white-crowned sparrows were exposed to the same magnetic field shift (90° to the left) for one hour in the afternoon when they were kept outdoors under the natural sky. These birds were later tested in the Emlen funnels under natural skies and in the local magnetic field around sunset. The magnetic shift experiment was repeated at least three times for each bird. As was the case for the first group of birds, the second group of birds was likewise tested in the local magnetic field and under natural sky conditions. Experiments were performed both before and after the days when they were exposed to the shifted magnetic field. By means of this procedure we will be able to record the birds’ compass course selected by:  l) information obtained only from the geomagnetic field, and 2) visual information obtained only from the natural sky. The results of these experiments will be essential for understanding the relative importance of magnetic field and visual cues for orientation at sunset. We will also be able to extract information on when and how fast the birds are able to calibrate and recalibrate their celestial and magnetic compasses. Thus, the experiments will help us to understand more about the calibration of compasses in migratory birds passing areas of varying angles of declination during migration.

Clock-shift experiments

For savannah sparrows and a group of the white-crowned sparrows as well, the internal time sense was shifted by exposing the experimental birds to a light-dark regime different from the natural one at the location of Inuvik in autumn. After initial control experiments the clock-shifted birds were transferred to a separate room without windows and the light-dark cycle was manipulated with artificial lights for at least three days simulating a 4 hour slow-shift. After this treatment the birds’ orientation was again recorded in the local magnetic field and under natural clear skies or in partly overcast conditions.

Release experiments

At the end of the experimental period we released the experimental birds individually at night with a light stick glued to one of their central tail feathers (cf. method described for example in, Ottosson et al. 1990). Releases were performed in an open area east of the town of Inuvik with the objective of recording the migratory flight direction at departure. The individual bird with light stick was released and observed with a pair of binoculars until vanishing from sight. As for the displaced birds (Åkesson et al. 1999), the white-crowned sparrows did not respond to this treatment by initiating migratory flights, but rather landed on the ground nearby. The savannah sparrows, on the other hand, to a larger extent disappeared from sight by circular flights and increasing flight altitude. For these birds we recorded the direction in which they departed using a hand-held compass.

Previous results

Previous experiments where birds have been exposed to conflicting information from magnetic and visual cues have resulted in different outcomes. Initially birds in many studies seem to primarily rely on visual cues, and do not clearly follow magnetic field shifts (Åkesson 1994, Åkesson and Bäckman 1999). However, in experiments where birds have been repeatedly exposed to magnetic field shifts under natural sunset or night skies, several studies have shown that after some time they do indeed follow the magnetic field shift (Wiltschko et al. 1997 and references therein). It will be important for the future to try to clarify the cause of these different outcomes. Experiments with savannah sparrows have revealed the importance of visual cues for orientation (Moore 1980). More recent experiments in the eastern United States show that young as well as adult birds do indeed calibrate their compasses during migration (cf. Able and Ab le 1995). In the light of these results it would be interesting to see how birds from breeding areas with completely different angles of declination respond to shifts of the magnetic field as well as shifts of the internal time sense (i.e. clock-shifts, e.g. Helbig 1991). Our experiments will be important for understanding compass calibration during migration and in particular the orientation system birds use at high latitudes.