Skip to main content

Thank you for visiting You are using a browser version with limited support for CSS. To obtain the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles and JavaScript.


Homing in on vertebrates

Imagine being set adrift in a canoe in the middle of an ocean. Which way would you paddle? Most humans would be as lost as lost can be, but creatures such as pigeons, turtles and whales have no difficulty navigating in such circumstances. How they do so remains one of the biggest mysteries in the behavioural sciences, at the centre of which is the question of how organisms might sense the Earth's magnetic field and use it for navigation and homing (a topic with a chequered history — see box, overleaf). The mystery, however, is gradually being solved, and the latest instalment in the story comes in Walker and colleagues' study of rainbow trout (page 371of this issue1).

All known sensory systems have specialized receptor cells designed to respond to the external stimulus, and these are always coupled to neurons to bring this information to the brain. In modern times the main objection to claims that magnetoreception is genuine was biophysical2 — that there was no evidence of appropriate receptors.

The first clue towards meeting this objection came in an obscure place, the teeth of small molluscs. The late Heinz A. Lowenstam discovered that the major lateral teeth of the chitons are capped by massive amounts of a hardening agent, the permanently magnetic mineral magnetite (Fe3O4) (Fig. 1). Later, with the development of ultrasensitive superconducting moment magnetometers, came the recognition that tiny (parts per billion) levels of biogenic magnetite are naturally present in animals as diverse as insects, birds, fish and even humans3,4. The final crack in the biophysical arguments against magnetoreception came with the discovery of the magnetotactic bacteria and protists (Fig. 2), which possess linear chains of either single-domain magnetite or greigite (Fe3S4). They provide unequivocal examples of biological activity being influenced by the geomagnetic field.

Figure 1: Magnetite-bearing teeth of the mollusc Cryptochiton stelleri .

Each tooth is about 1 mm in size, and capped with a layer of black magnetite (Fe3O4). (Photo, H. A. Lowenstam.)

Figure 2: A freshwater magnetotactic bacterium.

The chain of dark objects is composed of crystals of magnetite, which have the proper size and shape to behave as perfect, single magnetic domains. The largest crystals are about 70 nm in length. (Photo courtesy of A. Kobayashi.)

Magnetite seems made for magnetoreception. A typical magnetotactic bacterium is only a few micrometres in size, but contains en ough magnetite to make its rotational energy in the geomagnetic field exceed the thermal background energy by factors of 20 or more. The cells are very good, passive compasses, and will align with the Earth's magnetic field even when dead. The equivalent of only a single magnetotactic bacterium, connected to a single sensory neuron in a higher animal, could give that animal — ant, honeybee, trout or even whale — an extraordinarily good magnetic compass sense.

If this were all of the magnetite used for magnetoreception, finding and characterizing the receptor would be a needle-in-the-haystack operation. Fortunately, things are more complex. There are at least two types of response to the geomagnetic field — a simple compass, and another which is the result of sensing small fluctuations in the intensity of the background field. This latter sense has been implicated as a component of the navigational ‘map’ used by whales, turtles and birds5. Extensions of the biophysical analyses indicate that an array of a few thousand to a million magnetite-containing cells could yield responses to total intensity fluctuations of better than 0.1 per cent, as is observed behaviourally5, and that this entire receptor system could fit within a 1-mm cube and yet have a magnetite content of no more than 1 part per million (ref. 3). There is still no requirement for the receptors to be concentrated into such a small volume, but these calculations make the odds of finding them much better than previously thought.

Simple experiments using short but strong magnetic pulses (which exceed the coercive force of the magnetite) have shown that both the magnetic compass and intensity sensory systems involve the use of permanently magnetic materials such as magnetite6,7. Linear chains of single-domain biological magnetite crystals suitable for magnetoreception are easy to extract from animals and image8 (Fig. 3); and electrophysiological studies in birds have consistently identified fibres in the ophthalmic branch of the trigeminal nerve as the carriers of magnetic-field information7.

Figure 3: A linear chain of biogenic magnetite crystals, extracted from tissues in the frontal region of the sockeye salmon8, Oncorhynchus nerka, a close relative of the rainbow trout, O. mykiss .

These are also single magnetic domains, with crystal alignments similar to those in magnetotactic bacteria. (Photo courtesy of S. Mann.)

So to Walker et al.1, who have made two truly important advances. First, they have developed a simple laboratory conditioning regime for training rainbow trout to respond to magnetic cues. In principle, this could be extended to other vertebrates to tackle such questions as threshold sensitivities and frequency response, as has been done with similar conditioning experiments with honeybees6.

Second, and even more excitingly, they have traced the sensory nerves back to possible magnetoreceptor cells. After confirming that the ophthalmic branch of the trigeminal nerve in fish contains magnetically receptive fibres, as it does in birds7, they used a lipid-tracing dye to map the fibres back to the brain and to the location of putative receptor cells. Attempts to trace the avian magnetoreceptive nerves have failed because of the problem of identifying magnetite crystals in optical sections. Walker and colleagues' application of confocal laser microscopy provided the techniques for both tracing individual neurons back through a complex three-dimensional path and, by calibrating the confocal reflections with magnetotactic bacteria, for identifying possible magnetite crystals in the target cells.

Note, however, that the iron oxide mineral has not yet been identified conclusively. The cells containing the confocal reflections have distinctive shapes and always lie within a discrete sublayer of the olfactory lamellae (at the tips, near the distal terminals of fine branches of the trigeminal nerve), and the particles have similar size and shape to those extracted from salmon8 (Fig. 3). But follow-up studies with conventional transmission electron microscopy and electron diffraction are required to confirm that the iron oxide is indeed magnetite.

A huge range of organisms can sense magnetic fields5. Do humans remain an exception? We certainly have a trigeminal nerve, with an ophthalmic branch, and we can also make biogenic magnetite. At least one other vertebrate sensory system thought to have been lost in the final stages of human evolution — the sex-pheromone-sensing vomeronasal organ — has recently been found to be both present and functional9 (human vomeronasalins now form a booming perfume industry). Other effects, such as the ability of a 1-millitesla static field to elicit epileptiform activity in patients preparing for brain surgery10, have also emerged. Finally, some humans, particularly Polynesian navigators, seem able to judge direction in the absence of all obvious cues (Sun, Moon, stars, waves and so on)11. So there is hope for our lost canoeist — the final word on the existence of human magnetoreception has certainly not yet been written.


  1. 1

    Walker, M. M. et al. Nature 390, 371–376 (1997).

    ADS  CAS  Article  Google Scholar 

  2. 2

    Griffin, D. R. Q. Rev. Biol. 19, 15–31 (1944).

    Google Scholar 

  3. 3

    Kirschvink, J. L., Jones, D. S. & McFadden, B. J. Magnetite Biomineralization and Magnetoreception in Organisms: A New Biomagnetism (Plenum, New York, 1985).

  4. 4

    Kirschvink, J. L., Kobayashi-Kirschvink, A. & Woodford, B. J. Proc. Natl Acad. Sci. USA 89, 7683–7687 (1992).

    Google Scholar 

  5. 5

    Wiltschko, R. & Wiltschko, W. Magnetic Orientation in Animals (Springer, Berlin, 1995).

  6. 6

    Kirschvink, J. L., Padmanabha, S., Boyce, C. K. & Oglesby, J. J. Exp. Biol. 200, 1363–1368 (1997).

    Google Scholar 

  7. 7

    Beason, R. C. & Semm, P. J. Exp. Biol. 199, 1241–1244 (1996).

    Google Scholar 

  8. 8

    Mann, S., Sparks, N. H. C., Walker, M. M. & Kirschvink, J. L. J. Exp. Biol. 140, 35–49 (1988).

    Google Scholar 

  9. 9

    Berliner, D. L., Monti-Bloch, L., Jennings-White, C. & Diaz-Sanchez, V. J. Steroid Biochem. Mol. Biol. 58, 259–265 (1996).

    Google Scholar 

  10. 10

    Fuller, M., Dobson, J., Gregor-Wieser, H. & Moser, S. Brain Res. Bull. 36, 155–159 (1995).

    Google Scholar 

  11. 11

    Finny, B. Curr. Anthropol. 36, 500–505 (1995).

    Google Scholar 

Download references

Author information



Rights and permissions

Reprints and Permissions

About this article

Cite this article

Kirschvink, J. Homing in on vertebrates. Nature 390, 339–340 (1997).

Download citation

Further reading


By submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.


Quick links

Nature Briefing

Sign up for the Nature Briefing newsletter — what matters in science, free to your inbox daily.

Get the most important science stories of the day, free in your inbox. Sign up for Nature Briefing