See: Dominguez, P., Jacobson, A. G. & Jefferies, R. P. S. 2002. Paired gill slits in a fossil with a calcite skeleton. Nature (London) 417:841-844.
By using high-resolution computed microtomography on the little Carboniferous mitrate Jaekelocarpus oklahomensis (about 300 million years old) we have shown that it had gill slits on left and right inside the head. Moreover, it was a tunicate because:
1) the gill slits were antero-posteriorly elongate;
2) the gill slits were arranged in a series one above the other;
3) the gill slits opened into paired atria which projected anteriorly right and left of the pharynx; and
4) these atria debouched through a pair of antero-dorsal openings like those of immediately post-larval Corella, for example.
Jaekelocarpus had an echinoderm-like calcite skeleton made of porous stereom with each plate crystallographically a single crystal and also a tail which would have flexed mainly in a vertical plane. Both these features were probably primitive for the chordates and, since parsimony suggests that they were absent in the latest common ancestor of living tunicates, they probably imply that Jaekelocarpus was a stem-group tunicate, not a crown-group tunicate.
The whole animal was built on the plan of a tunicate tadpole, with clearly distinct head and tail. (For simplicity of comparison with vertebrates, we prefer not to speak of "trunk" and tail.) Indeed this tunicate-tadpole-like body plan must have been primitive for chordates since it existed in all mitrates and throughout the bizarre "carpoid" group of fossils.
It is very important that there are right and left gill slits in Jaekelocarpus, as in a post-larval amphioxus, rather than left gill slits only, as in a larval amphioxus or a member of the fossil group Cornuta (e.g. Cothurnocystis).
This is not the first known fossil tunicate. Only last year, Degan Shu and his co-workers, of Xi'an University in China, reported the fossil tunicate Cheungkongella ancestralis. Since this was fixed down to the substrate rather than tunicate-tadpole-like, we think it probably belonged to the tunicate crown group rather than the stem group, although it came from much older rocks than Jaekelocarpus (it is from the Lower Cambrian of southern China, perhaps 560 million years old; see Shu et al., 2001).
In addition to Jaekelocarpus, there are several other known mitrates which, in our opinion, represent the tunicate stem group. None of them, however, shows the clear evidence for gill slits revealed by Jaekelocarpus.
Jaekelocarpus is not new. It was originally described, with a different interpretation, by Dennis Kolata, Terry Frest and Roy Mapes (Kolata et al., 1991). They were also the first to see the gill slits, although they did not call them by that name and their material was badly broken. The material was lent to us by Liz Nesbitt and Ron Eng of the Burke Museum, University of Washington. The x-ray tomography of the specimens was carried out by Timothy Rowe and his group at The University of Texas at Austin.
Part of the argument for recognising the slits as gill slits (or more precisely as stigmata) is based upon the monographs of Australian tunicates produced by Pat Kott. I hope she approves.
See "About the Scan" for original tiff data
Scans of a paratype specimen of Jaekelocarpus oklahomensis Kolata, Frest & Mapes (UWBM no. 74305) for Dr. Richard Jefferies, of the Department of Paleontology, The Natural History Museum, London.
Specimen fixed into place with wax for scanning. Scanned by Matthew Colbert and Richard Ketcham on 4 February, 2001.
16bit: original scanned images. II, 150 kV, 0.053 mA, no filter, air wedge, 0% offset, slice thickness 2 lines (= 0.016 mm), S.O.D. 14 mm, 1800 views, 2 samples per view, interslice spacing 2 lines (=0.016 mm), field of reconstruction 4.53 mm, reconstruction offset 600, reconstruction scale 1.6. Scanned in three-slice mode.
Zip file of original 16 bit tiff data files, 72 megabytes
Literature
Dominguez, P., A. G. Jacobson, and R. P. S. Jefferies. 2002. Paired gill slits in a fossil with a calcite skeleton. Nature 417:841-844.
Kolata, D. R., T. J. Frest, R. H. Mapes. 1991. The youngest carpoid: occurrence, affinities and life mode of a Pennsylvanian (Morrowan) mitrate from Oklahoma. Journal of Paleontology 65:844-855.
Shu, D.-G., L. Chen, J. Han, and X.-L. Zhang. 2001. An Early Cambrian tunicate from China. Nature 411:472-473.
, 
