News & Views | Published:

Sensory biology

Search for the compass needles

Nature volume 484, pages 320321 (19 April 2012) | Download Citation

The idea that bird orientation is guided by magnetic-sensing structures in the animals' beaks has been challenged by the suggestion that the iron-containing cells are macrophages, which have no link to the brain. See Letter p.367

Most humans are hopeless navigators unless they have a map or the Global Positioning System to hand. By contrast, it is well known that many animals, including birds1,2, use information from Earth's magnetic field to orient themselves, even over thousands of kilometres. But how do they sense the field? Iron-mineral-based structures within nerve cells in birds' upper beaks were previously proposed3 to act as one of two main magnetic sensors in birds, the other being a light-based system in the retina4,5. However, on page 367 of this issue, Treiber et al.6 report that these iron-mineral deposits are in fact located inside macrophages, a kind of immune cell that is also involved in iron homeostasis7.

When people are asked how they think birds can sense magnetic fields, most suggest that birds must have small 'compass needles' somewhere in their body. The iron-containing minerals magnetite (Fe3O4) and maghemite (Fe2O3), which are found in various parts of a bird's body, are particularly well suited to sensing magnetic fields. However, these minerals can be relevant to magnetic sensing only if they are associated with nerve tissue and found in the same location in all individuals of a given species. In a landmark paper in 2003, Fleissner et al.3 reported that structures containing magnetite 'spherules' and maghemite 'platelets' are consistently present at six specific locations along the upper beaks of pigeons (Fig. 1a). These researchers suggested3,8 that the structures are magnetic sensors located in dendrites — nerve-cell endings. The structures included a large substructure that the researchers described3,8 as a vesicle.

Figure 1: Ironing out magnetoreception.
Figure 1

a, Fleissner and colleagues3,8,11 previously proposed that birds' beaks contain dendrites (nerve-cell endings) harbouring iron-rich structures that can sense Earth's magnetic field and transmit this information to the brain, thereby helping the bird to navigate. By using electron and light microscopy to observe tissue sections stained with the iron-marking dye Prussian blue (blue colour), the authors identified magnetite-containing 'spherules', maghemite-containing 'platelets' (too small to depict) and a large substructure that they defined as a vesicle. They classified the structures as being clustered at six defined points along the upper beak. b, Treiber et al.6 now propose that these iron-mineral deposits are instead located in macrophages, immune cells that have a type of organelle (siderosomes) containing the iron-storage protein ferritin, and that the substructure is the cell's nucleus. The authors have found that these iron-containing cells are diffusely distributed throughout the beak and in other parts of the bird's body.

If these structures sense the magnetic field, this information must be transmitted to the brain via the ophthalmic branch of the trigeminal nerve, as this is the only nerve that enters a bird's upper beak. Indeed, when this nerve is cut, pigeons trained to detect a strong magnetic field can no longer perform this task9. Furthermore, in European robins, which are migratory birds, neurons in the brain regions that receive direct input from the trigeminal nerve are activated by a changing magnetic-field stimulus, and this activation disappears if the trigeminal nerve is cut or if the magnetic field is removed10. Therefore, even though there was no direct proof for their function, the iron-mineral structures described by Fleissner and colleagues3,8,11 have been generally assumed to be the primary magnetic receptors associated with the trigeminal magnetic sense in birds.

Treiber et al.6 challenge the interpretations of Fleissner and colleagues3,8,11. On the basis of data that they collected from almost 200 pigeons, Treiber et al. propose that the iron-mineral-containing structures reported by Fleissner et al.3 are macrophages (Fig. 1b), not magnetic receptor cells. They show that the structures are located in variable quantities at inconsistent locations along the upper beak, and also elsewhere in the pigeon's body, such as in the respiratory tract and skin, and that they are not associated with nerve tissue6. They also suggest that the spherules, which are strongly stained by the iron-marking dye Prussian blue, are in fact organelles called siderosomes, which contain the iron-storing protein ferritin, and that the large, round substructure is a nucleus, not a vesicle6.

Whose interpretation is correct? The answer to this question will be extremely important for magnetic-sense research. To form my own opinion, I examined many of the original microscope slides that make up the basis of Treiber and colleagues' report6, and compared them with my previous observations of some of the original slides that contributed to Fleissner and colleagues' studies3,11. In my view, a significant proportion of the iron-mineral-containing structures found by Treiber et al.6 seem to be identical to those previously reported3,8,11, even though some of the structures classified as such by Treiber et al. would, I believe, be regarded as artefacts by Fleissner and colleagues. However, even when I consider only those structures that look identical, these certainly seem to occur at many more than six specific locations along the beak. Indeed, Treiber and colleagues' laborious quantification shows that the occurrence of the iron-mineral-containing structures is highly variable between individual birds6. By contrast, Fleissner and colleagues' claim3,8 of six specific locations is not supported by quantitative data. This lack of consistent distribution makes it highly unlikely that the iron-rich cells are part of a magnetic sensory system.

The most defining difference between macrophages and dendrites is that macrophages have a nucleus and dendrites do not. Treiber and colleagues' data and original slides convincingly show that the iron-containing structures are nucleated and that there is no regular co-localization between Prussian- blue staining and nerve fibres. The structures, therefore, cannot be located in dendrites. Furthermore, the authors report6 95–99% co-localization of staining with Prussian blue and a macrophage marker, strongly indicating that most, but maybe not all, of the described structures are macrophages.

In conclusion, I find that serious doubt has been raised about the original proposals3,8,11 and that the burden of proof now lies with Fleissner and colleagues, if they still think that their structures are potential magnetic sensors. However, it is important to stress that Treiber and colleagues' results cannot exclude the possibility that there are iron-mineral-based sensors somewhere in the upper-beak region of pigeons. Only relatively few sensors may be needed. In fact, a magnetite-based sensory cell might contain only a few magnetite crystals12 and thus evade detection by Prussian-blue staining12, the method used by both groups.

The implications of this new report are dramatic. The evidence supporting a magnetic compass embedded in birds' visual systems4,5 remains unaffected, but we are left with only two studies9,10 that clearly implicate the ophthalmic branch of the trigeminal nerve, although not necessarily iron-related-structures, in magnetic sensing in birds. Therefore, for now, we have evidence supporting only a trigeminal-nerve-related magnetic sense in birds, not an iron-mineral-related magnetic sense. Both the biological function and basic sensory origin of the trigeminal magnetic sense are now unknown. Understanding the magnetic senses in animals indeed remains a formidable scientific problem.


  1. 1.

    & Science 176, 62–64 (1972).

  2. 2.

    , & Science 304, 405–408 (2004).

  3. 3.

    et al. J. Comp. Neurol. 458, 350–360 (2003).

  4. 4.

    , , , & Nature 429, 177–180 (2004).

  5. 5.

    et al. Nature 461, 1274–1277 (2009).

  6. 6.

    et al. Nature 484, 367–370 (2012).

  7. 7.

    & Biochem. J. 434, 365–381 (2011).

  8. 8.

    , , , & Naturwissenschaften 94, 631–642 (2007).

  9. 9.

    , , & Nature 432, 508–511 (2004).

  10. 10.

    , , , & Proc. Natl Acad. Sci. USA 107, 9394–9399 (2010).

  11. 11.

    et al. PLoS ONE 5, e9231 (2010).

  12. 12.

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

Download references

Author information


  1. Henrik Mouritsen is at the Research Center for Neurosensory Sciences and the Institut für Biologie und Umweltwissenschaften, Carl-von-Ossietzky Universität Oldenburg, 26111 Oldenburg, Germany.

    • Henrik Mouritsen


  1. Search for Henrik Mouritsen in:

Corresponding author

Correspondence to Henrik Mouritsen.

About this article

Publication history



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.

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