Natural and Induced Magnetization in the Bobolink, Dolichonyx Oryzivorus (Aves: Icteridae)

Many species exhibit behavioural orientation responses in natural and artificial magnetic fields. These species range from bacteria (Blakemore, 1975, 1982; Kirschvink, 1980; Blakemore & Frankel, 1981) and insects (Lindauer & Martin, 1968; Martin & Lindauer, 1977; Gould, Kirschvink, Deffeyes & Brines, 1980) to fish (shark: Kalmijin, 1978; tuna: Walker, Dizon & Kirschvink, 1982; salmon: Quinn, Merrill & Brannon, 1981), amphibians (Phillips & Adler, 1978), reptiles (Rodda, 1984) and mammals (Mather & Baker, 1981). Research has been directed especially towards analysing the behavioural responses of birds to magnetic field information. The homeward orientation of the homing pigeon (Columbidae: Columba livia) can be altered by natural variations in the geomagnetic field (Keeton, Larkin & Windsor, 1974; Walcott, 1978, 1980; Kiepenheuer, 1982), magnetic storms (Larkin & Keeton, 1976; Carr, Switzer & Hollander, 1982), magnetic coils (Walcott & Green, 1974) or bar magnets (Keeton, 1971, 1972; Larkin & Keeton, 1976). Investigations on several European (Wiltschko & Wiltschko, 1975; Beck & Wiltschko, 1982) and a few North American (Emlen et al. 1976; Bingman, 1981, 1983; Beason & Nichols, 1984; R. C. Beason, in preparation) species of migratory birds have demonstrated that they too are sensitive to earth-strength magnetic fields. Modification of the ambient magnetic field produces predictable changes in orientation headings (Wiltschko & Wiltschko, 1975; Bingman, 1983).

Although several terrestrial vertebrate species have shown orientation responses to the geomagnetic field, nothing is known about the receptor and transducer mechanisms involved. The presence of a ferrimagnetic material, probably magnetite (Kirschvink, 1983), has been reported in several species, some of which have shown a sensitivity to geomagnetic information (Gould, Kirschvink & Deffeyes, 1978; Zoeger, Dunn & Fuller, 1981; Kuterbach, Walcott, Reeder & Frankel, 1982; Jones & MacFadden, 1982; Lohmann, 1984). Because of the manner in which single domain (SD) and pseudo-single domain (PSD) magnetite particles attempt to align themselves with the ambient magnetic field, it seems plausible that magnetite particles are involved in the perception of magnetic field information.

The bobolink (Icteridae: Dolichonyx oryzvuorus) has the longest migratory pathway of any New World landbird. Its breeding range stretches across North America through the northern United States and southern Canada (AOU, 1983). The non-breeding season (austral summer) is spent from southern Brazil to northern Argentina (deSchauensee, 1966). Because of this long migratory path, the bobolink is unable to visualize the circumpolar stars of either hemisphere during its entire journey, and cannot use stellar cues for migratory orientation in the same manner as the indigo bunting (Emberizidae : Passerine cyanea) (Emlen, 1969, 1970). Instead, stellar cues appear to be used as part of an integrated navigation system which also involves the earth’s magnetic field (R. C. Beason, in preparation).

The magnetic properties of the heads of 28 adult bobolinks, one adult indigo bunting and one adult savannah sparrow (Emberizidae : Passerculus sandwichensis) were examined. Nine bobolinks were taken from the nest and hand-reared under laboratory conditions, and the other 19 were captured as adults.

Preliminary measurements of saturation isothermal remanent magnetization (sIRM) of 15 bobolink heads (including all nine hand-reared birds) were made on a PAR SM-2 spinner magnetometer in the Department of Geological Sciences, SUNY at Geneseo. All IS heads were skinned and fixed shortly after death in 0·1 moll⁻¹ phosphate-buffered glutaraldehyde (electron microscope grade). After fixing, seven heads were individually embedded in JB-4 methacrylate plastic using an organic catalyst. Since previous histological examination of tissues from various areas of the bobolink heads revealed concentrations of inorganic iron around the olfactory nerve and bulb, and in bristles which projected into the nasal cavity (Beason & Nichols, 1984), we used glass and brass knives to remove the portion of each head between the anterior nasal cavity and the middle of the orbit from the remaining eight birds. These sections were combined into a composite sample and tested in the same manner as the intact heads. Care was taken to align all sections parallel to one another.

The natural remanent magnetization (NRM) and sIRM of an additional 12 bobolink heads, one indigo bunting and one savannah sparrow were measured on a SQUID (Goree & Fuller, 1976) magnetometer manufactured by Superconducting Technologies, Inc. (SCT) at the Department of Geology and Planetary Science, University of Pittsburgh. The heads were prepared by anaesthetizing each bird and perfusing it with avian saline (0·75%) followed by 0·1 mol l⁻¹ phosphate-buffered 2·5% glutaraldehyde (e.m. grade). Each head was then removed and the skin (including the nostril areas) completely removed by hand, and kept in the buffered glutaraldehyde until tested. Only glass-double-distilled water was used in preparing the specimens. At the time of measurement, each head was placed in the same polycarbonate plastic cylinder which was not subjected to any artificial magnetic field. At no time was the remanent magnetic moment of the cylinder above 10⁻⁹e.m.u. (1 e.m.u. = 10³ Am⁻¹). The sIRM of the original seven plastic-embedded heads and a plastic blank, which had been subjected to the same treatment as the heads, were remeasured on the SQUID. The sIRM of the blank (2·8×10⁻⁶e.m.u.) was subtracted from that of each embedded head. All measurements reported here for individual heads were made on the SQUID magnetometer.

Several heads, including the composite sample and plastic blank, were subjected to strong d.c. magnetic fields (up to 0·7 T for the PAR measurements and up to 0·5 T for the SQUID measurements) to impart an IRM or sIRM. In each test, the sample was placed between the poles of an electromagnet with no current flowing. The current was then increased to a pre-established level to produce the magnetic field desired. The current was reduced to zero before the sample was removed. The samples were maintained in a consistent orientation so that the resulting IRM was acquired in a direction perpendicular to the NRM (which averaged 5 % of the sIRM). The IRM was thus acquired along a sample direction for which the NRM was zero.

Progressive a.f. demagnetization of the IRM was conducted on five bobolink and one indigo bunting heads. One additional bobolink head was cooled in liquid nitrogen and an sIRM imparted while the head was at 77 K. The sample was then allowed to warm to room temperature in the SQUID while the remanent magnetization was monitored to test for the presence of superparamagnetic magnetite.

The average NRM of 12 unembedded bobolink heads was 3·20 (± 1·7O)×10⁻⁷ e.m.u. The orientation of the NRM was similar for all but one of the birds (Fig. 1). In general, the magnetization is directed horizontally through the skull, from right to left, and points somewhat towards the anterior. The one exception showed nearly the reverse orientation of the other heads. Because of the consistent orientation of the NRM in the fixed but unembedded heads, it is unlikely to have been caused by contamination of the nasal passages by air-borne material, nor could it have been the result of alignment with the ambient magnetic field following fixation because the orientation of the stored heads was random.

The average saturation magnetization for 23 bobolink heads (seven embedded individually in plastic, eight in a composite sample and eight unembedded) was 2·49 (± 2·17)×10⁻⁵e.m.u. There was no difference between embedded and unembedded heads (F = 3·69, P = 1·00). The greatest gain in IRM occurred in the 0–0·1T range, although most samples gained additional IRM in the 0·1–0·3T range (Fig. 2). No sample gained IRM in magnetic fields stronger than 0·3T. The failure of any sample to gain IRM in fields stronger than 0·3 T is a good indication that there was no contamination during sample preparation. Most other non-magnetite contaminants continue to gain remanence above 0·3 T.

The AF demagnetization curve of the sIRM intersects the sIRM acquisition curve at about 20 % of the sIRM, and indicates a remanent coercive force of 35–40 mT. An intersection at the 50 % level and remanent coercive force of approximately 100 mT or higher indicates the presence of single domain grains with each grain uninfluenced by the others. An intersection at lower values, such as we found, indicates either an interaction between grains because of their close proximity to one another (Cisowski, 1981) or the presence of some multidomain material.

There was no loss of remanent moment as one head (which had been given an sIRM at 77 K) warmed from 77 K through the isotropic point of magnetite (130 K) to room temperature. Thus, none of the magnetic material in the bobolink appears to contain superparamagnetic magnetite.

The results from the indigo bunting (NRM = 3·20×10⁻⁷ e.m.u., sIRM = 7·48×10⁻⁶e.m.u.) and savannah sparrow (NRM = l·70×l0⁻⁷e.m.u., sIRM = l·23×10⁻⁵ e.m.u.) were similar to those from the bobolinks. The orientation of the NRM in both species was similar to that of the majority of the bobolink heads (indigo bunting, inclination = 16°, declination = 292°; savannah sparrow, inclination = 43°, declination = 272°). The indigo bunting head was also subjected to a.f. demagnetization with results similar to those from the bobolink heads (Fig. 2).

Because all our samples attained sIRM in magnetic fields of 0·3 T or less, we believe that the magnetic material in the bobolink is magnetite. The theoretical maximum coercivity for magnetite is 0·3 T, whereas other ferromagnetic materials continue to gain remanent magnetization in stronger fields (McElhinny, 1973). Although the IRM data and the colour (black) of the material do not prove that the magnetization is due to the presence of magnetite, they are compatible with such an explanation. Regardless of the magnetic material involved, the magnetization we observed is sufficient to allow a sensitivity to even slight changes in the geomagnetic field (Yorke, 1979, 1981)..

Since multiple domain magnetite has a maximum coercivity of 90 mT (Evans & McElhinny, 1969), any IRM gain in fields of between 0·1 and 0·3 T must be due to single domain magnetite (Evans, McElhinny & Gifford, 1968). Because the specimens gained appreciable IRM below 0·1 T and the remainder between 0·1 and 0·3 T, the magnetic material in the bobolink heads appears to be interacting single domain magnetite or possibly pseudo-single domain magnetite grains.

Because magnetite is the only known ferrimagnetic material of biogenic origin (Kirschvink, 1983) and our data support such an explanation, we believe that most, if not all, of the remanent magnetization in the bobolink is carried by single domain and possibly some pseudo-single domain magnetite. Magnetite has also been reported at the same locations in the homing pigeon (Walcott, Gould & Kirschvink, 1979) as in the bobolink (Beason & Nichols, 1984), and proposed for some other species of migratory birds (Presti & Pettigrew, 1980). On the other hand, Ueda et al. (1982) concluded that no magnetite was present in the species of migratory and non-migratory birds they tested. Because they applied an external field of only 0·1T, their samples probably did not reach saturation IRM. The values they report for NRM and IRM in a 0·1 T field are very similar to our results. Had they continued to apply stronger magnetic fields, it is plausible the heads would have continued to gain remanence.

Because magnetite has been reported or proposed (based on remanent magnetization) in many species using the geomagnetic field for orientation (see references in the Introduction), it may be a common feature across phyletic lines for magnetoreception (Walcott & Walcott, 1982). The mechanism by which magnetite may be used to transduce magnetic field information to the nervous system has not been established for any species. Although magnetic perception, apparently without the use of magnetite, has been reported in the homing pigeon (Semm, Schneider, Vollrath & Wiltschko, 1982; Semm, Nohr, Demaine & Wiltschko, 1984), such findings do not preclude the role of magnetite in magnetic sensory perception.

We thank V. Schmidt for graciously allowing use of the SQUID magnetometer at the Paleomagnetics Laboratory of the University of Pittsburgh. R. Selfridge provided assistance with the equipment.