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.

Wandering nostrils

A small matter of head anatomy has long been a cause of controversy among those interested in vertebrate evolution. An answer that may prove generally palatable now emerges from an ancient fossil fish.

The structures known as choanae may seem obscure. But we've all got them; they are the ‘internal nostrils’ that form the passage between our nasal cavity and throat that we use for breathing when our mouth is closed. They have also been the subject of much argument among those studying comparative vertebrate anatomy — in particular, the question of how choanae originated in the tetrapods, or land vertebrates, a group that consists of amphibians, reptiles, birds and mammals.

This unique feature of the tetrapod palate was first regarded as an adaptation to breathing air. Now, however, we know that choanae occurred first in extinct fish relatives of the tetrapods at least 380 million years ago — that is, 30 million years before tetrapods developed limbs and took to the land. So choanae may not have been initially involved in breathing. But did they arise as a novelty within the ancient fish relatives of tetrapods, or were they derived from a pre-existing structure in other fishes? This is a debate that has lasted for about a century, but it is practically settled by new data presented on page 94 of this issue by Zhu and Ahlberg1. The subject of their studies is Kenichthys, a 395-million-year-old fossil fish from China. The background to the story, described below, is also outlined in Figure 1.

Figure 1: The hole picture.

These depictions of the head and palate in some living and fossil lobe-finned fishes and tetrapods show the apertures of the nasal capsule — the anterior (a) and posterior (p) nostrils and the choanae (c). The tree indicates the relationships of the groups, and the orange arrows indicate the two occurrences of the migration of the posterior nostril into the palate. In Kenichthys, the fossil studied by Zhu and Ahlberg1, the posterior nostril forms a notch in the outer dental arcade, and is intermediate in position between that of an external nostril and that of choanae (p/c). The diagrams (1–3) on the right are of the left nasal cavity, in ventral view, in the three living forms. In lungfishes, the path of the maxillary nerve (mx) is medial to the posterior nostril; in tetrapods it is lateral to the choana. t, tear duct.

Most living jawed fishes have two external nostrils on each side of the snout: an anterior one allowing entry of water into the nasal cavity, and a posterior one for its exit. In contrast, tetrapods have only one pair of external nostrils, but their nasal cavity communicates with the mouth cavity by means of the choanae. A single group of living fish, lungfishes, seemed to bridge this gap between fish and tetrapods, as they have ‘internal nostrils’ in their palate that superficially resemble the choanae. Early evolutionary anatomists thus considered that lungfishes constituted ideal intermediate forms, and were evidence for the piscine origin of tetrapods. Embryological studies then showed that these ‘internal nostrils’ in fact corresponded to the posterior nostrils of other fishes, but they were also taken to be homologous with — that is, the same as — the tetrapod choanae. The choanae were thus regarded as posterior nostrils that had migrated into the palate in the common ancestor of lungfishes and tetrapods.

Early in the twentieth century, however, that view was shaken by work on certain lobe-finned fishes from the Palaeozoic era (385–280 million years ago). These fishes, later referred to as ‘osteolepiforms’ and including such examples as Osteolepis and Eusthenopteron2, proved to have a single external nostril on either side of the snout, and choana-like openings in the palate. They became preferred as ‘ancestors’ to tetrapods because of their close overall resemblance to them and their antiquity, and it soon became received wisdom that the tetrapods had originated from osteolepiforms. The lungfish ‘internal nostrils’ became increasingly regarded as having merely paralleled the tetrapod choanae in evolution, because of major differences in their relationships to the surrounding nerves, bones and sensory-line canals3. This was later confirmed by the discovery of Diabolepis, a primitive, 400-million-year-old lungfish, which retains two pairs of external nostrils (and thus lacks ‘internal nostrils’)4.

Several theories have been proposed for the origin of the choanae of tetrapods and of their fossil osteolepiform relatives. The two main possibilities are that choanae are derived from the posterior external nostrils, but independently of development of the ‘internal nostrils’ of lungfishes, or that they are a novelty, and the posterior external nostrils have either disappeared or became the tear ducts. Debate on the matter rumbled on for decades. Rosen et al.5 even revived the theory that lungfishes have true choanae, and claimed that osteolepiforms provide no evidence for choanae, thereby creating a controversy6 that has had far-reaching consequences for the use of fossil data in reconstructions of evolutionary history7.

The current, morphology-based scheme of lobe-finned-fish and tetrapod evolution, first proposed by Ahlberg8 and shown in Figure 1, displays two major sister groups. These are the dipnomorphs (lungfishes, and the extinct Diabolepis and porolepiforms) and the tetrapodomorphs (tetrapods and their closest fossil fish relatives, including the osteolepiforms). If it is now clear, thanks to Diabolepis, that the migration of the posterior nostrils into the mouth occurred in the dipnomorphs within the lungfish lineage, there was no comparable fossil evidence for any precursor of the tetrapodomorph choanae.

That changes with Zhu and Ahlberg's report1 on Kenichthys. Kenichthys is clearly a tetrapodomorph, and it displays exactly the intermediate condition that would be predicted when assuming that the tetrapod choanae were posterior nostrils displaced into the palate, in the same way as in lungfishes. The posterior nostrils (or already choanae?) of Kenichthys are actually right at the margin of the upper jaw, interrupting the outer dental ‘arcade’ formed by the premaxillary and maxillary bones.

This remarkable discovery, however, does not answer a question that was raised long ago3 and that centres on other anatomical features. If the posterior nostrils passed into the palate to form the tetrapod choanae, why did they not leave indelible traces by interrupting the outer dental arcade and the sensory-line canal that runs below the eye sockets, and pushing the maxillary nerve inwards, as they did in lungfishes (Fig. 1)? Zhu and Ahlberg's analysis suggests that this canal and nerve, and the outer dental arcade, were restored laterally to the choanae in more advanced tetrapodomorphs. The restoration of the dental arcade may not be a problem, but that of the maxillary nerve and sensory-line canal certainly is. Further study of the nerve canals inside the snout of Kenichthys, or of tetrapod snout development, might offer a solution.

Nevertheless, Kenichthys clearly provides the first factual basis for the theory that tetrapod choanae actually are the posterior nostrils, as are the internal nostrils of lungfishes. In a way, the new analysis reconciles Rosen and colleagues' provocative theory5 and the classical interpretation of the fish members of the tetrapodomorphs — internal nostrils of lungfishes and tetrapod choanae are homologous, but their position in the palate is not.


  1. 1

    Zhu, M. & Ahlberg, P. E. Nature 432, 94–97 (2004).

    ADS  CAS  Article  Google Scholar 

  2. 2

    Ahlberg, P. E. & Johanson, Z. Nature 395, 792–794 (1998).

    ADS  Article  Google Scholar 

  3. 3

    Jarvik, E. Zool. Bidrag Uppsala 21, 235–675 (1942).

    Google Scholar 

  4. 4

    Chang, M. M. Bull. Mus. Natl Hist. Nat. 17, 235–268 (1995).

    Google Scholar 

  5. 5

    Rosen, D. E., Forey, P. L., Gardiner, B. G. & Patterson, C. Bull. Am. Mus. Nat. Hist. 167, 157–276 (1981).

    Google Scholar 

  6. 6

    Jarvik, E. Syst. Zool. 30, 378–384 (1981).

    Article  Google Scholar 

  7. 7

    Gee, H. In Search of Deep Time (Free Press, New York, 1999).

    Google Scholar 

  8. 8

    Ahlberg, P. E. Zool. J. Linn. Soc. 103, 241–287 (1991).

    Article  Google Scholar 

Download references

Author information



Rights and permissions

Reprints and Permissions

About this article

Cite this article

Janvier, P. Wandering nostrils. Nature 432, 23–24 (2004).

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