Head and backbone of the Early Cambrian vertebrate Haikouichthys

Abstract

Agnathan fish hold a key position in vertebrate evolution, especially regarding the origin of the head and neural-crest-derived tissue. In contrast to amphioxus, lampreys and other vertebrates possess a complex brain and placodes that contribute to well-developed eyes, as well as auditory and olfactory systems. These sensory sytems were arguably a trigger to subsequent vertebrate diversifications. However, although they are known from skeletal impressions in younger Palaeozoic agnathans, information about the earliest records of these systems has been largely wanting. Here we report numerous specimens of the Lower Cambrian vertebrate Haikouichthys ercaicunensis, until now only known from the holotype. Haikouichthys shows significant differences from other fossil agnathans: key features include a small lobate extension to the head, with eyes and possible nasal sacs, as well as what may be otic capsules. A notochord with separate vertebral elements is also identifiable. Phylogenetic analysis indicates that this fish lies within the stem-group craniates. Although Haikouichthys somewhat resembles the ammocoete larva of modern lampreys, this is because of shared general craniate characters; adult lampreys and hagfishes (the cyclostomes if monophyletic) are probably derived in many respects.

Full text

inferred from fault-zone-guided waves on the scale of a few hundred
metres
4
.
Previously we showed that fault zones heal
5
, and here we have
shown that the combination of shaking and static stress reduces the
modulus of a recently broken fault zone. This probably indicates
that the strength of the fault was reduced, which could trigger
earthquakes. Our observation thus may provide a direct measure-
ment of the missing connection between shaking and the facilitation
of distant aftershocks
20,21
. Regional clustering of earthquakes is
another plausible result of one big earthquake weakening the
regional set of faults around it, either as an anomalous activation
of a region
22
or as the progression of ruptures along a fault, with the
shaking modulated by directivity
23
.
Another implication of our result is that existing friction laws
may need improvement. Currently, friction is modelled simply (or
not so simply) as a function of a state variable, which is the history of
sliding, and current sliding rate of a point on a fault plane
24
.If
shaking can significantly reduce strength, it may also help to explain
the puzzle of aftershock occurrence very near the mainshock fault
plane, which strikes where the Earth has been strongly shaken but
where the regional stress is reduced. Shaking-induced weakening
also may be involved in the progression of rupture during earth-
quakes, because strong shaking is likely to precede the arrival of the
rupture front. A
Received 22 August; accepted 29 November 2002; doi:10.1038/nature01354.
1. Reid, H. F.in The California Earthquake of April 18, 1906: Report of the State Earthquake Investigation
Commission (Carnegie Institution of Washington, Washington DC, 1910).
2. Li, Y.-G., Vidale,J. E., Day,S. M., Oglesby, D.D. & Cochran, E. Post-seismic fault healingon the rupture
zone of the 1999 M7.1 Hector Mine, California earthquake. Bull. Seismol. Soc. Am. (in the press).
3. Rubinstein, J. L., Beroza, G. C., Bokelmann, G. & Schaff, D. Near surface damage caused by the strong
ground motion of the M6.9 Loma Prieta and M5.4 Chittenden earthquakes. Eos 83, NG21B (2002).
4. Li, Y.-G., Vidale,J. E., Aki, K., Xu, F. & Burdette, T.Depth-dependence structure of the Landers fault
zone from trapped waves generated by aftershocks. J. Geophys. Res. 105, 6237–6254 (2000).
5. Li, Y.-G. & Vidale, J. E. Healing of the shallow fault zone from 1994–1998 after the M7.5 Landers,
California, earthquake. Geophys. Res. Lett. 28, 2999–3002 (2001).
6. Schaff, D. P. & Beroza, G. C. Postseismic response of repeating aftershocks. Geophys. Res. Lett. 25,
4549–4552 (1998).
7. Baisch, S. & Bokelmann, G. H. R. Seismic waveform attributes before and after the Loma Prieta
earthquake: scattering change near the earthquake and temporal recovery. J. Geophys. Res. 106,
16323–16337 (2001).
8. Ikuta, R., Yamaoka, K., Miyakawa, K., Kunitomo, T. & Kumazawa, M. Continuous monitoring of
propagation velocity of seismic wave using ACROSS. Geophys. Res. Lett. 29 5-1–5-4 (2002).
9. Fialko, Y. et al. Deformation on nearby faults induced by the 1999 Hector Mine ear thquake. Science
297, 1858–1862 (2002).
10. Richardson, E. & Marone, C. Effects of normal stress vibrations on frictional healing. J. Geophys. Res.
104, 28859–28878 (1999).
11. Cochran, E. & Vidale, J. E. Seismic anisotropy of the Hector Mine fault zone. Eos 82, S41A (2001).
12. Bokelmann, G. H. R. & Harjes, H.-P. Evidence for temporal variation of seismic velocity within the
upper continental crust. J. Geophys. Res. 105, 23879–23894 (2000).
13. Ratdomopurbo, A. & Poupinet, G. Monitoring a temporal change of seismic velocity in a volcano:
Application to the 1992 eruption of Mt. Merapi (Indonesia). Geophys. Res. Lett. 22, 775–778 (1995).
14. Hill, D. P. et al. Seismicity remotely triggered by the magnitude 7.3 Landers, California, earthquake.
Science 260, 1617–1623 (1993).
15. Brodsky, E., Karakostas, V. & Kanamori, H. A new observation of dynamically triggered regional
seismicity: earthquakes in Greece following the August, 1999 Izmit, Turkey earthquake. Geophys. Res.
Lett. 27, 2741–2744 (2000).
16. Knopoff, L., Levshina, T., Keilis-Borok, V. I. & Mattoni, C. Increased long-range intermediate-
magnitude earthquake activity prior to strong earthquakes in California. J. Geophys. Res. 101,
5779–5796 (1996).
17. Bowman, D. D. & King, G. G. C. P. Accelerating seismicity and stress accumulation before large
earthquakes. Geophys. Res. Lett. 28, 4039–4042 (2001).
18. Price, E. & Sandwell, D. T. Small-scale deformations associated with the 1992 Landers, California,
earthquake mapped by synthetic aperture radar interferometry. J. Geophys. Res. 103, 27001–27016
(1998).
19. Fialko, Y., Simons, M. & Agnew, D. The complete (3-D) surface displacement field in the epicentral
region of the 1999 Mw7.1 Hector Mine earthquake, southern California, from space geodetic
observations. Geophys. Res. Lett. 28, 3063–3066 (2001).
20. Kilb, D., Gomberg, J. & Bodin, P. Triggering of earthquake aftershocks by dynamic stresses. Nature
408, 570–574 (2000).
21. Gomberg, J., Reasenberg, P. A., Bodin, P. & Harris, R. A. Earthquake triggering by seismic waves
following the Landers and Hector Mine earthquakes. Nature 411, 462–466 (2001).
22. Rockwell, T. K. et al. Paleoseismology of the Johnson Valley, Kickapoo, and Homestead Valley faults:
clustering of earthquakes in the eastern California shear zone. Bull. Seismol. Soc. Am. 90, 1200–1236
(2000).
23. Stein, R. S. Progressive failure on the North Anatolian fault since 1939 by earthquake stress triggering.
Geophys. J. Int. 128, 594–604 (1997).
24. Marone, C. Laboratory-derived friction laws and their application to seismic faulting. Annu. Rev.
Earth Planet. Sci. 26, 643–696 (1998).
25. Sieh, K. et al. Near-field investigation of the Landers earthquake sequence, April to July 1992. Science
260, 171–176 (1993).
26. Scientists of the USGS, SCEC, and CDMG. Preliminary report on the 16 October 1999 M7.1 Hector
Mine, California, earthquake. Seismol. Res. Lett. 71, 11–23 (2000).
27. Hauksson, E., Jones, L. M. & Hutton, K. The 1999 Mw7.1 Hector Mine, California earthquake
sequence: complex conjugate strike-slip faulting. Bull. Seismol. Soc. Am. 92, 1154–1170 (2002).
28. Ji, C., Wald, D. J. & Helmberger, D. V. Source description of the 1999 Hector Mine, California
earthquake. Part II: complexity of slip history. Bull. Seismol. Soc. Am. 92, 1208–1226 (2002).
Acknowledgements We thank E. Hauksson for supplying relocations of regional seismicity;
H. Houston, P. Davis, E. Brodsky, C. Marone, R. Stein and L. Knopoff for comments; and
T. Burdette, F. Xu and E. Cochran for field help.
Competing interests statement The authors declare that they have no competing financial
interests.
Correspondence and requests for materials should be addressed to J.E.V.
(e-mail: vidale@ucla.edu).
..............................................................
Head and backbone of the Early
Cambrian vertebrate Haikouichthys
D.-G. Shu*†, S. Conway Morris‡, J. Han*, Z.-F. Zhang*, K. Yasui§,
P. Janvierk, L. Chen*, X.-L. Zhang*, J.-N. Liu*,Y.Li*& H.-Q. Liu*
*Early Life Institute and Department of Geology, Northwest University, Xi’an,
710069, China
†School of Earth Sciences and Resources, China University of Geosciences, Beijing,
100083, China
‡Department of Earth Sciences, University of Cambridge, Downing Street,
Cambridge CB2 3EQ, UK
§Institute of Molecular Embryology and Genetics, Division of Development and
Biohistory, Kumamoto University, 2-2-1 Honjo, Kumamoto 860-0811, Japan
kUMR 8569 du CNRS, Laboratoire de Pale
´ontologie, Muse
´um National
d’Histoire Naturelle, 8 rue Buffon, Paris 75005, France, and Department of
Palaeontology, Natural History Museum, London SW7 5BD, UK
.............................................................................................................................................................................
Agnathan fish hold a key position in vertebrate evolution,
especially regarding the origin of the head and neural-crest-
derived tissue
1
. In contrast to amphioxus
2
, lampreys and other
vertebrates possess a complex brain and placodes that contribute
to well-developed eyes, as well as auditory and olfactory systems
3
.
Figure 4 Model of velocity as a function of time owing to damage from the Landers
rupture, the Hector Mine shaking, and the combination of the two compared with
observations. Shown is healing as a logarithm of time
10
, although details just after each
event and extrapolating into the future are not well constrained. The velocity before the
Landers earthquake was not measured.
letters to nature
NATURE| VOL 421 | 30 JANUARY 2003 | www.nature.com/nature526 © 2003 Nature Publishing Group

These sensory sytems were arguably a trigger to subsequent
vertebrate diversifications. However, although they are known
from skeletal impressions in younger Palaeozoic agnathans
4
,
information about the earliest records of these systems has
been largely wanting. Here we report numerous specimens of
the Lower Cambrian vertebrate Haikouichthys ercaicunensis,
until now only known from the holotype
5
.Haikouichthys
shows significant differences from other fossil agnathans: key
features include a small lobate extension to the head, with eyes
and possible nasal sacs, as well as what may be otic capsules. A
notochord with separate vertebral elements is also identifiable.
Phylogenetic analysis indicates that this fish lies within the stem-
group craniates. Although Haikouichthys somewhat resembles
the ammocoete larva of modern lampreys, this is because of
shared general craniate characters; adult lampreys and hagfishes
(the cyclostomes if monophyletic
6,7
) are probably derived in
many respects.
The putative Early Cambrian agnathan Haikouichthys, from the
Chengjiang Lagersta
¨tte
8
near Kunming, Yunnan, was only known
from a single, incomplete specimen. Although its vertebrate affin-
ities have been widely accepted
9,10
, many of the conclusions about its
anatomy and thereby its phylogenetic position are regarded as
‘highly provisional’
11
. The discovery of more than 500 specimens,
from a locality near Haikou, reveals a series of new and unexpected
features that imply a major reconsideration of several features of
early agnathan evolution.
Most notable in this respect is the identification in at least 300
specimens of a small (usually less than 1 mm in length) lobate
extension at the anterior end of the animal (Fig. 1a–d). It is
separated from the rest of the head by a slight constriction. Its
most conspicuous feature is a pair of dark oval stains, interpreted as
eyes. Although an approximately circular area in some specimens
may indicate the lens, this feature is too inconsistent to be reliable.
In a few specimens (Fig. 1e, g) the position of an eye is marked as a
concave impression, consistent with it having a scleral layer. In life,
the eyes may have been relatively flat, although the quality of
preservation does not allow inferences on the presence of extra-
ocular muscles. The upward direction of the eyes and their position
on the anterior lobe suggests, however, that their mobility may have
been restricted. Between the eyes is a median structure, often paired
and preserved in a similar manner to the eye stains. This may
represent part of the olfactory organ, possibly the nasal sacs (Fig. 1a–
d). An alternative interpretation, that these structures represent the
pineal/parapineal complex
12
, is considered less likely. This is because
of their relative sizes, position with respect to the eyes, paired
nature, and style of preservation suggesting relatively tough tissue.
Along the anterior edge of this lobate extension to the head there is a
pair of arcuate structures, meeting about a notch that may define the
position of a median nostril (Fig. 1a–d). Typically the arcuate
structures have relief, and in life were apparently plate-like and
possibly composed of cartilage. Other than the plates, no other part
of the anterior lobe appears to have been reinforced. More posteri-
orly, however, the next region of the head is often heavily pervaded
by diagenetic mineralization, probably originally sulphides but now
oxidized. In the holotype an attempt was made to reconcile this
mineralization with the organization of the cranial cartilages seen in
the lampreys
5
. The distribution of this mineralization in the new
material shows considerable variation and makes this comparison
problematical. There is, however, some consistency in the
expression of a pair of sub-circular areas, and these may represent
the otic capsules (Fig. 1a, b), although no internal structure is
evident.
The mouth is not clearly visible, but a ventral recessed area
immediately behind the anterior lobe may indicate the position of
the oral opening. Another newly observed feature is a series of more
or less square-shaped structures extending posteriorly within the
main dorsal region of the head and anterior trunk. These are
sometimes connected by a broad, dark strand. This feature is
interpreted as a series of separated vertebral elements (arcualia),
associated with the notochord (Fig. 1e–l). The shape of the
individual vertebral elements is somewhat variable. They may be
bifid or arched, and in some cases appear to have encompassed the
entire notochord. Up to ten such vertebral elements have been
identified, and it is likely that the series extended further backwards,
but it is obscured by the overlying impressions of the trunk
myomeres. The original composition was possibly of cartilage,
and occasional diagenetic mineralization may reflect some degree
of calcification. The last noteworthy feature that is not apparent in
the holotype is lamellate regions located between the branchial
supports on the lower side of the head region. These most probably
represent the gills, and on occasion are delimited by boundaries that
suggest an original pouch-like structure (Fig. 1e, g, i–l). The entire
branchial area generally has a much rougher texture than the
remainder of the head, from which it is clearly delimited.
In other respects the new material of Haikouichthys largely
confirms the earlier observations
5
. The more complete specimens
suggest that the posterior part of the trunk tapered evenly (Fig. 1f, h,
j, l). Although preservation of the caudal region is poor, there is no
evidence for a caudal fin. The presence of a dorsal fin and what
appears to be a ventral fin-fold is confirmed, although unequivocal
evidence of the latter being paired is not available (Fig. 1f, h, j, l). In
the dorsal fin the associated fin-rays are only occasionally visible,
perhaps owing to the original thickness of the dorsal fin, but as in
the holotype they appear to be anteriorly tilted. The myosepta are
well defined, and at the posteriorly directed flexure of the myomeres
some specimens show repeated areas of positive relief. Additional
details are not visible, but if these correspond to a set of underlying
gonads
5
(Fig. 1j, l) then this may indicate a primitive metameric
arrangement reminiscent of amphioxus. A dark strand running
posteriorly on the lower side of the trunk may represent the
intestine. There is some evidence that the anus was sub-terminal,
thus defining a short post-anal region (Fig. 1f, h, j, l). In the head
region the branchial arches are confirmed as a series of posteriorly
recurved structures, each apparently composed of a single unit
(Fig. 1e, g, i, k). The exact total is equivocal because of variable
preservation, but it is most probably seven or eight.
Haikouichthys was evidently a swimmer, although its degree of
activity is conjectural. We note, however, co-occurrence of this fish
with other pelagic taxa such as Xidazoon, and the relative scarcity of
benthic forms. The specimens may have been buried alive, possibly
as a result of storm-induced burial. Most specimens were collected
from a series of graded beds consisting of a lower sand/siltstone unit
(about 3 mm thick) and an overlying mudstone, up to 50 mm thick.
The typical occurrence is close to the boundary between the two
units, and ‘shoals’ of specimens often show a preferred orientation.
Specimens tend to occur in bed-parallel orientation, but where
buried obliquely the portion in the mudstone is usually much better
preserved than that in the adjacent silty unit.
The new material of Haikouichthys shows that our knowledge of
the earliest agnathans has been incomplete, with the implication
that in several respects the cyclostomes, although almost unchanged
since the Carboniferous
13
, are in fact much derived. In particular,
the anterior lobe with its eyes (and possible olfactory organ) has no
direct counterpart in the other early craniates. There is, however, a
possible similarity in the anterior position of the prominent eyes of
conodonts
14
and the Ordovician arandaspids
15
, although in the
latter group this feature has been interpreted as a specialization
16
.
Anteriorly located eyes are also known in some of the ‘naked’ fossil
agnathans, most notably Jamoytius
17
and probably also the anaspid-
like fish Euphanerops
4
. If this condition of anterior eyes is primitive
to vertebrates, it may indicate derivation from the frontal eye of an
amphioxus-like ancestor
18
, although this would entail various
changes including their location on a stable platform and the
development of a balancing mechanism
19
, which in the form of a
letters to nature
NATURE| VOL 421 | 30 JANUARY 2003 | www.nature.com/nature 527
© 2003 Nature Publishing Group

correlated vestibulo-ocular system represents a key step in ver-
tebrate evolution
20
. It also suggests that the more usual posterior
position of the eyes is a result of rostral growth, perhaps in response
to the increase in size of the olfactory organ. The more tenuous
identification of the olfactory organ in Haikouichthys makes its
phylogenetic role more speculative, but if correctly identified its
position and size is more comparable to that seen in lampreys, and
to some extent hagfish, including fossil representatives
13
, than in
most other agnathans and the gnathostomes. The branchial arches
of Haikouichthys are unlike the intricate branchial basket of lam-
preys. Their apparently simple structure is somewhat reminiscent of
the arrangement seen in the gnathostomes, especially regarding
Figure 1 Haikouichthys ercaicunensis from Haikou, Kunming, Yunnan. a–d, Details of
the head, emphasizing its anterior structures; preservation is dorso-ventral.
a,b, ELI-0001003 (273). c,d, ELI-00010013 (323). e,g,i,k, Details of the anterior,
emphasizing the notochord with associated vertebral elements and branchial arches.
e,g, ELI-0001015 (12B), anterior to the right. i,k, ELI-0001020 (8), anterior end to the
left. f,h,j,l, Nearly complete specimens. f,h, ELI-0001002 (191), anterior end to the
right. j,l, ELI-0001001 (172), anterior end to the left. Ap, anterior plates; Ba, branchial
arches; Df, dorsal fin; Myo, myosepta; Nc, notochord; Nc and vert, notochord with
vertebral elements; Nos, nostril; Ns, nasal sacs; ?Oc, otic capsule; Oe, oesophagus; Pa,
post-anal tail; Vert, vertebral elements; Vf, ventral fin-fold. L, left; R, right.
letters to nature
NATURE| VOL 421 | 30 JANUARY 2003 | www.nature.com/nature528 © 2003 Nature Publishing Group

lower units known as the ceratobranchials and hypobranchials. This
suggests that the arrangement of articulated branchial elements in
gnathostomes retains some primitive characters, even though it is
clear that the origin of the jaw entailed developmental rearrange-
ments
21–23
. The widely spaced vertebral blocks probably acted as
supportive structures for the notochord, and as such are reminis-
cent of the arcualia of adult lampreys. In Haikouichthys, however,
the vertebral units are larger and more regularly spaced.
The possession of eyes (and probably nasal sacs) is consistent
with Haikouichthys being a craniate, indicating that vertebrate
evolution was well advanced by the Early Cambrian. Although
evidently a jawless fish, its precise phylogenetic position is still
speculative because this fish shows a puzzling mixture of characters
contrary to some previous expectations. Nevertheless, several fea-
tures of Haikouichthys, including what may be metamerically
arranged gonads and the anteriorly located eyes, suggest that ‘the
first fish’ may be best regarded as a stem-group craniate (Fig. 2a).
Specific connections to other agnathans include the vertebral
elements and probable nasal sacs. Despite some similarities to
lampreys, in certain respects these living agnathans are probably
highly derived.
Until now the other Chengjiang agnathan, Myllokunmingia,is
only known from a single specimen recovered from a separate
locality at a slightly higher stratigraphic level within the Qiongzhusi
Formation
5
. Its most obvious differences from Haikouichthys are
possession of fewer gill pouches (five or six), absence of curved
branchial supports, the presence of possible exhalant chambers, and
a more anterior extension of the dorsal fin that lacks obvious fin
rays. It is clear that the Chengjiang Lagersta
¨tte offers unique insights
into early deuterostome diversifications
5,24–26
, and continued
excavations are expected to extend further our knowledge of their
earliest evolution, including that of the vertebrates. A
Methods
Phylogenetic analysis
The analysis, based on the data of ref. 27, was done by using the phylogenetic package
HENNIG86 1.5 (ref. 28) and the matrix and tree editor TREE GARDENER (ref. 29) with a
data matrix of 17 taxa (4 extant, 13 fossil, cephalochordates as the outgroup) and 115
morphological and physiological characters (see Supplementary Information). All
characters are coded binarily as present/absent. Non-applicable characters are coded as 0
and missing data are coded as ‘?’. Equally, most parsimonious trees were obtained by using
the implicit enumeration* (ie*) command.
Received 18 June; accepted 18 October 2002; doi:10.1038/nature01264.
1. Northcutt, R. G. The origin of craniates: neural crest, neurogenic placodes, and homeobox genes. Isr.
J. Zool. 42, S273–S313 (1996).
2. Holland, L. Z. & Holland, N. D. Evolution of neural crest and placodes: amphioxus as a model for the
ancestral vertebrate? J. Anat. 199, 85–98 (2001).
3. Kleerekoper, H. The Biology of Lampreys Vol. 2 (eds Hardisty, M. W. & Potter, I. C.) 373–404
(Academic, London, 1972).
4. Janvier, P. Early Vertebrates (Clarendon, Oxford, 1996).
5. Shu, D.-G. et al. Lower Cambrian vertebrates from south China. Nature 402, 42–46 (1999).
6. Mallatt, J. & Sullivan, J. 28S and 18S rDNA sequences support the monophyly of lampreys and
hagfishes. Mol. Biol. Evol. 15, 1706–1718 (1998).
7. Kuraku,S. et al.Monophyly of lampreys and hagfishessupported by nuclearDNA-coded genes. J. Mol.
Evol. 49, 729–735 (1999).
8. Zhang, X.-l. et al. New sites of Chengjiang fossils: crucial windows on the Cambrian explosion. J. Geol.
Soc. Lond. 158, 211–218 (2001).
9. Janvier, P. Catching the first fish. Nature 402, 21–22 (1999).
10. Shimeld, S. M. & Holland, P. W. H. Vertebrate innovations. Proc. Natl Acad. Sci. USA 97, 4449–4452
(2000).
11. Holland, H. D. & Chen, J.-y. Origin and early evolution of the vertebrates: new insights from advances
in molecular biology, anatomy, and palaeontology. BioEssays 23, 142–151 (2001).
12. Cole, W. C. & Youson, J. H. Morphology of the pineal complexof the anadromous sea lamprey,
Petromyzon marinus L. Am. J. Anat. 165, 131–163 (1982).
13. Bardack, D. & Zangerl, R. The Biology of Lampreys Vol. 1 (eds Hardisty, M. W. & Potter, I. C.) 67–84
(Academic, London, 1971).
14. Gabbott, S. E., Aldridge, R. J. & Theron, J. N. A giant conodont with preserved muscle tissue from the
Upper Ordovician of South Africa. Nature 374, 800–803 (1995).
15. Gagnier, P.-Y. Sacabambaspis janvieri, verte
´bre
´Ordovicien de Bolivie: I: Analyse morphologique. Ann.
Pale
´ontol. 79, 19–51 (1993).
16. Gagnier, P.-Y. Sacabambaspis janvieri, verte
´bre
´Ordovicien de Bolivie 2: Analyse phyloge
´ne
´tique. Ann.
Pale
´ontol. 79, 119–166 (1993).
17. Ritchie, A. New evidence on Jamoytius kerwoodi White, an important ostracoderm from the Silurian
of Lanarkshire, Scotland. Palaeontology 11, 21–39 (1968).
18. Lacalli, T. C. Frontal eye circuitry, rostral sensory pathways and brain organization in amphioxus
larvae: evidence from 3D reconstructions. Phil. Trans. R. Soc. Lond. B 351, 243–263 (1996).
19. Lacalli, T. C. New perspectives on the evolution of protochordate sensory and locomotory systems,
and the origin of brains and heads. Phil. Trans. R. Soc. Lond. B 356, 1565–1572 (2001).
20. Fritzsch, B. Evolution of the vestibulo-ocular system. Otolaryngology Head Neck Surg. 119, 182–192
(1998).
21. Kuratani, S. et al. Embryology of the lamprey and evolution of the vertebrate jaw: insights from
molecular and developmental perspectives. Phil. Trans. R. Soc. Lond. B 356, 1615–1632 (2001).
22. Cohn, M. J. Lamprey Hox genes and the origin of jaws. Nature 416, 386–387 (2002).
23. Shigetani, Y. et al. Heterotopic shifts of epithelial–mesenchymal interactions in vertebrate jaw
evolution. Science 296, 1316–1319 (2002).
24. Shu, D.-G. et al. Primitive deuterostomes from the Chengjiang Lagersta
¨tte (Lower Cambrian, China).
Nature 414, 419–424 (2001).
25. Shu, D.-G. et al. An early Cambrian tunicate from China. Nature 411, 472–473 (2001).
26. Shu, D.-G., Conway Morris, S. & Zhang, X.-L. A Pikaia-like chordate from the Lower Cambrian of
China. Nature 384, 157–158 (1996).
27. Janvier, P. The dawn of the vertebrates: characters versus common ascent in the rise of current
vertebrate phylogenies. Palaeontology 39, 259–287 (1996).
28. Farris, J. S. HENNIG86, version 1.5. Program and user’s manual (published by the author, Port
Jefferson Station, New York, 1988).
29. Courrol Ramos, T. TREE GARDENER, version 1.0. khttp://www.icn.unal.edu.co/extensio/servicio/
servicio.htmll(1996).
Supplementary Information accompanies the paper on Nature’s website
(çhttp://www.nature.com/nature).
Acknowledgements This work was supported by the Ministry of Sciences and Technology of
China, the Natural Science Foundation of China, the Ministry of Education of China, the
National Geographic Society (USA), The Royal Society, and St John’s College, Cambridge. We
thank K. Kardong and B. J. Swalla. L. Guo, X. Cheng, M. Cheng and S. Last are thanked for
technical assistance, and Y. Ji and H. Guo for fieldwork help.
Competing interests statement The authors declare that they have no competing financial
interests.
Correspondence and requests for material should be addressed to D.-G.S.
(e-mail: elidgshu@nwu.edu.cn ).
Figure 2 Phylogenetic analysis. a, Strict consensus of 23 equally parsimonious trees
(length 177 steps; consistency index 0.64, retention index 0.64). Haikouichthys appears
here in a trichotomy with hagfishes and all other vertebrates (that is, one possibility is that it
is a stem craniate), but this is largely because of the inferred presence of metameric
gonads. b, Strict consensus of four equally parsimonious trees (length 175, consistency
index 0.64, retention index 0.64) obtained when the character ‘metameric gonads’ (114)
is inactivated. Haikouichthys appears here as the sister-group to all other vertebrates
except hagfishes, like Myllokunmingia in the analysis of ref. 5.
letters to nature
NATURE| VOL 421 | 30 JANUARY 2003 | www.nature.com/nature 529
© 2003 Nature Publishing Group