NEWS AND VIEWS -DAEDALUS------, Rolling foam Tulerpeton 363 Ventastega 1 THE hovercraft supports itself on a c ~ Acanthostf ga lcht'}fostega 1 cushion of air that leaks away so rapidly Ill that it has to be replaced all the time by '2 I ~-' c 'I I ,' ,,' continuous fast pumping. Daedalus now Gl . E Hynerpeton proposes a hovercraft pumped not with Metaxygnathus ' ' ~, ~ '\ ' ,' air, but with detergentfoam. ' ,,, I',, Foam, he reckons, acts like a gas that I .. I can sustain a pressure gradient. Every & \ I film wall in the foam has a pressure 368 Elglnerpetontldae \ I \ ,, difference across it; the sharper the film I \ ) curvature the greater this pressure drop. \ I A foam of many small bubbles is a \ I \I multitude of sharply curved films in Eusthenopteron Panderichthys ," series, and can contain quite a large pressure. Special detergentfoams have already been proposed to impede rioters 377 ' \ \\ I '~======~======~=== \ \ I and reduce the force of terrorist bombs. \ \ I With a little development, they should be \ \I able to support a hovercraft. ' \ I , Daedalus's foam craft will be \ I \ I essentially a flat platform with a hole in \ I the middle. Into the hole will be pumped, \I not pure air, but detergentfoam. Under 381 \ the craft, it will spread out into a soft FIG . 2 The fish-amphibian transition. Temporal distribution of advanced osteolepiform fish and supporting cushion which will leak out Devonian amphibians. Circles indicate time of occurrence of known fossils; dashed lines show round the edges only slowly. A very small possible relationships as discussed by Ahlberg\ and the numbers indicate ages in millions pump will keep the vehicle foamborne. of years. The foam craft could drive itself along in several ways. It could use the usual indicate that skeletal features that were The increase of information on the inefficientairscrews. A seagoing once thought to be unique to land verte­ fish-amphibian transition emphasizes the foam craft could project a propeller shaft brates appeared stepwise in a succession continued lack of knowledge of the next down through the foam into the water. On of aquatic, semiaquatic and semiterrest­ crucial step in the evolution of land verte­ land it might use wheels or a caterpillar rial vertebrates, while features commonly brates. In contrast to the variety of well­ track, touching the ground but not attributed to fis h were retained in animals known Upper Devonian tetrapods, no loading it with much weight. While fitting classified as amphibians. This point is amphibians have yet been described from the craft for its most obvious use­ highlighted by evidence, provided by the succeeding 20 million years, during moving heavy loads over rough or Coates and Clack8 , that important fea­ which time all the main groups of later marshy country-this would destroy tures of the tetrapod hands and feet in land vertebrates diverged. 0 much of its elegance. But Daedalus Acanthostega may have evolved among hopes that the foam itself can be made to primarily aquatic animals. The distribu­ Robert Carroll is in the Department of exert a tractive force. By angling the jets tion of primitive and derived characters Biology, McGill University, 1205 Dr Penfield delivering the foam, or spinning small differs from lineage to lineage, showing Avenue, Montreal, Quebec, Canada H3A paddle-wheels within it, it might be given that many features were evolved or lost 181. a rotatory swirl which would turn it into a convergently. As in the case of other sortoffluid caterpillar track. The more major transitions in vertebrates, such as 1. Ahlberg, P E. Nature 373, 420- 425 (1995). 2. Ahlberg, P. E. Nature 354, 298- 301 (1991). viscous the foam, the greater traction it 12 13 the origin of birds and mammals • , the 3. Vorobyeva . E. I. Trudypaleont. lnst. 163, 1-239 could exert, but at the cost of higher convergent origin of derived features in (1977) 4. Vorobyeva. E. l. & Schultze. H.-P. in Orig!n ofHigher viscous losses. Careful optimization wil different lineages makes it difficult to Groups of Tetrapods: Controversy and Consensus (eds be needed. establish specific relationships, or to agree Schultze, H.-P. & Trueb, L.) 68-109 (Cornell Univ. Press, Atfirst Daedalus imagined that his Ithaca, 1991). on objective criteria to differentiate tetra­ 5. Clack,J. A. Nature 369, 392- 394 (1994). foam craft would leave a glistening trail of pods from their fish ancestors. 6. Coates, M.l. &Clack. J. A. Nature 347, 66-69 foam behind it, like some giant mollusc. Although all these new data make it (1990). 7. Coates. M. I. & Clack. J. A. Nature 352, 234- 235 But he is now designing a simple foam­ more difficult to classify species involved (1991). sucker atthe back to scoop up this trail in the transition between fish and amphi­ 8. Coates. M. l. &Clack, J. A. 1nProc. VII• Symposium International Etude des Vertebres lnff!fieurs (in the and recycle it. This would almost bians, they greatly illuminate the evol­ press). eliminate losses of water and detergent, utionary processes involved. They dem­ 9. Daeschler. E. B.. Shubin, N.H., Thomson, making the craft nearly self-contained. K. s. T. &Amaral, W. W. Science 265, 639--642 onstrate that the complete transition, (1994). Calmly and smoothly, with no blasts of air from early osteolepiform sarcopterygians 10. Lebedev. 0 . A. &Clack, J. A. Palaeontology36. 721- 734 and no roaring, foamcraft will glide to primarily terrestrial amphibians may (1993). 11. Thomson, K. S. T. in MajorFeaturesofVertebrate across water, along beaches, over marsh have extended over at least 15 million Evolution (ed . Spencer, R.) 85-107 (Paleontological and moor and even along ordinary roads. years. The fossil record is not yet well Society, Dept Geol. Sci .. Univ. Tennessee. Knoxville, 1994). Their cushion of churning detergent enough known to establish rates of evol­ 12. Ostrom,J. H. in MajorFeaturesofVertebrateEvolution foam could make them ideal for street ution, but the observed changes may be (ed . Spencer. R. ).160- 177 (Pa leontological cleaning duties. Pedestrians run over by Society. Dept Geol. Sci. , Univ. Tennessee, Knoxville, attributed to progressive adaptation to 1994). a foamcraftwould not be injured, but increasingly shallow water, accompanied 13. Hopson, J. A. in Major Features ofVertebrate Evolution merely subjected to a brisk Jacuzzi by repeated cladogenesis (branching of (ed. Spencer. R. ) 190-219 (Paleontological Society, DeptGeol. Sci., Univ. Tennessee, Knoxville, experience. David Jones evolutionary lineages). 1994). 390 NATURE · VOL 373 · 2 FEBRUARY 1995