Journal of Natural History, 2000, 34, 311–315
Caecilian viviparity and amniote origins: a reply to Wilkinson
and Nussbaum
MICHEL LAURIN² ³ , ROBERT R. REISZ§ and
MARC GIRONDOT³
² Institut fuÈr PalaÈontologie, Naturhistorisches Forschungsinstitut,
Museum fuÈr Naturkunde, Zentralinstitut der Humboldt-UniversitaÈ t zu
Berlin, Invalidenstraû e 43, D-10115 Berlin, Germany;
email: [email protected]. hu-berlin.de
 quipe de recherche `Formations squelettiques’ , UMR 8570, Case 7077,
³ E
CNRS, Universite Paris 7РDenis Diderot, 2, place Jussieu, F-75251
Paris Cedex 05, France; email: mgi@ccr. jussieu.fr
§ Departm ent of Zoology, Erindale Campus, University of Toronto,
Mississauga, ON L 5L 1C6, Canada;
email: [email protected]. ca
(Accepted 11 January 1999 )
An evaluation of the evolution of reproduction in amniotes by the authors has
been recently criticized. This criticism is shown to be based on a misrepresentation
of our work. The ancestral amniote was certainly oviparous, but this does not
necessarily imply that eggs were laid at an early stage of development. The
presence of a signi® cant amount of intrauterine development in the earliest
amniotes remains as parsimonious as the alternative, despite recent claims to the
contrary. The relative rarity of amniote groups that laid heavily mineralized egg
membranes in the Palaeozoic may explain why no amniote fossil egg has been
found in the Palaeozoic.
Keywords: Evolution, phylogeny, parsimony, Amniota, Gymnophiona, eggs,
fossil record.
Introduction
The evolution of the amniotic egg has fascinated many biologists, but its origin
is still inadequately understood. We know that the amniotic egg must have existed
at least 310 million years ago because amniotes had appeared at that time, and all
extant amniotes possess an amniotic egg (although it may be somewhat modi® ed in
viviparous groups) . In the absence of the necessary information from the fossil
record (no Palaeozoic tetrapod egg is known), the evolution of the amniotic egg
must be studied by comparing extant taxa in the context of amniote phylogeny.
Journal of Natural History
ISSN 0022-2933 print/ISSN 1464-5262 online Ñ 2000 Taylor & Francis Ltd
http://www.tandf.co.uk /journals/tf /00222933.html
312
M. Laurin et al.
Origin of the amniotic egg
The amniotic egg is usually interpreted as an adaptation for life on land (Carroll,
1988 ). However, Lombardi ( 1994 ) and Laurin and Reisz ( 1997 ) recently raised the
possibility that some of the extra-embryonic membranes evolved in an embryoretaining form, in which they facilitated embryo± mother interactions. Laurin and
Reisz ( 1997: ® gure 6 ) argued that an optimization of a binary character (absence or
presence of extended embryo retention) on a sarcopterygian phylogeny showed that
two hypotheses were equally parsimonious; extended embryo retention may or may
not have been present in the ® rst amniote.
All amniote eggs, as well as the eggs of all caecilians, some other lissamphibians,
and various other vertebrates ( the coelacanth, chondrichthyans, etc. ) are internally
fertilized. Internal fertilization allows the development of interactions between the
embryo and the mother, which could lead to viviparity ( but not always, as shown
by monotremes and many squamates, for example) . The ubiquitous presence of
internal fertilization in amniote eggs raises the possibility that the ® rst amniote
already retained embryos in utero for a signi® cant portion of their development, and
that the extra-embryonic membranes of the amniotic egg developed to facilitate
embryo± mother interactions.
This proposition was recently attacked by Wilkinson and Nussbaum ( 1998 ),
who misrepresented this issue by equating the state `extended embryo retention
present’ with `viviparity’ . Having thus reformulated the problem, Wilkinson and
Nussbaum then proceeded to show that the ancestral amniote would have been
oviparous. We concur with that view, and we have never argued the contrary.
When viviparity is observed in an organism, it is quite valid to argue that it
represents embryo retention. This is an obvious result of this particular reproductive
strategy, and we used this to argue for embryo retention in caecilians. However,
oviparity does not by itself show evidence to the contrary, i.e. absence of embryo
retention. Egg retention in tetrapods exhibits continuous variation, and viviparity is
only its extreme form. For example, in monotremes, there is an intrauterine phase
of development followed by oviposition and a short period of incubation (about 10
days) of the egg within a pouch. During intrauterine development, the volume of
the egg increases dramatically through absorption of uterine secretions (Stewart,
1997 ). Therefore, monotremes possess extended embryo retention while being oviparous. Similarly, a substantial amount of egg development precedes oviposition in
typical squamates ( Blackburn, 1993: 418 ). More precisely, numerous squamates
retain eggs in utero for about one-half of the total period of embryonic development
( Shine, 1985 ). Stewart ( 1993: 445 ) even suggested that in squamates, `Functionally,
the early stages of extraembryonic membrane formation may be equivalent in
oviparous and viviparous species, at least for viviparous species with telolecithal
eggs’ .
Wilkinson and Nussbaum ( 1998 ) based their contribution on a demonstration
that caecilians are primitively oviparous and did not provide any information or
references about the early embryology of egg-laying caecilians. A thorough literature
search revealed that information on this topic is scanty ( Duellman and Trueb, 1986 ).
Sarasin and Sarasin ( 1887± 1890 ) described embryos of Ichthyophis glutinosus of
various stages, but we could not, after careful reading, ® nd the mention of when
the eggs were laid. A few other embryos of Ichthyophis have been described
( Breckenridge and Jayasinghe, 1979; Balakrishna et al., 1983 ), but only from accidentally found eggs that had been laid for an unknown period of time. Only Brauer
Caecilian viviparity and amniote origins
313
( 1897 ) gives some information on the developmental stage at oviposition for another
caecilian species, Hypogeophis rostratus, that appears to lay its eggs in the blastula
stage. Unfortunately, no other study has con® rmed this observation, or provided
similar data for other caecilian species. In any other report (and any other caecilian),
the smallest observed caecilian embryos in laid eggs measured 6 mm and possessed
at least 21 somites ( Breckenridge and Jayasinghe, 1979; Balakrishna et al., 1983 ).
Therefore, the literature does not appear to provide decisive evidence regarding the
amount of intra-uterine development in most oviparous caecilians. Wilkinson and
Nussbaum themselves emphasized that the reproductive biology of caecilians is
poorly known, even though they were only interested in determining whether various
caecilian clades were oviparous or viviparous: `The most direct inferences on caecilian
reproductive modes come from observations of parturition, clutches of eggs, larvae,
and oviductal foetuses. Unfortunately, because caecilian amphibians are mostly
secretive, burrowing, tropical forms, such observations are unavailable for the great
majority of the approximately 160 caecilan species’ . If evidence about the early
development of all but one species of oviparous caecilians is lacking in the scienti® c
literature, the arguments presented by Wilkinson and Nussbaum are not valid, and
they have not falsi® ed our hypothesis. Indeed, whereas the viviparous caecilians
possess an extended embryo retention, the vast majority of oviparous caecilians for
which no data on early development are available may or may not share this
condition. The simultaneous occurrence of oviparity and viviparity within two
caecilian clades (scolecomorphids, caeciliids) suggests that the reproductive mode of
caecilians has evolved several times independently, and that the information about
the stage of embryonic development at oviposition for a single caeciliid species
(Hypogeophis rostratus; see Brauer, 1897 ) does not provide a safe basis to infer the
primitive reproductive pattern for caecilians. Therefore, until more data about the
early development of oviparous caecilians become available, it is still justi® able to
consider that caecilians could perform extended embryo retention.
Amniotic eggs in the fossil record
For vertebrate palaeontologists working on the early history of amniote evolution, the absence of fossil evidence of nests and eggs in the Palaeozoic is perplexing,
especially in light of the extensive fossil record of such things in the Mesozoic
( Mikhailov, 1991; Norell et al., 1994 ). It is clear that amniotes were present and
diversifying explosively during this time ( Laurin and Reisz, 1995 ), but nests and
fossilized eggs only appear in the Triassic ( Bonaparte and Vince, 1979 ), at least
80± 90 million years after the appearance of the oldest known amniotes (Carroll,
1964 ). Therefore, the fossil record is not inconsistent with the hypothesis that the
earliest amniotic eggs were poorly mineralized.
The distribution of various types of egg membranes in amniotes corroborates
the thesis that early amniotic eggs had an unmineralized shell membrane. Stewart
( 1997: 308± 309 ) concluded that the ancestral amniote eggshell was ¯ exible and
unmineralized, and that a mineralized eggshell membrane is a synapomorphy of
Reptilia ( the last common ancestor of turtles and diapsids, and all its descendants).
Finally, a heavily mineralized, rigid eggshell ( the only kind that would have a
reasonably good fossilization potential ) seems to have appeared independently
among archosaurs, some gekkonid lizards, and at least ® ve times among turtles
( Stewart, 1997 ).
314
M. Laurin et al.
The vast majority of the amniote diversi® cation in the Palaeozoic is represented
by synapsids, whereas reptiles were relatively rare members of the fossil record in
the Palaeozoic ( the only exception being the pareiasaurs, presumed to be extinct
relatives of turtles that were fairly common in parts of the Upper Permian). The
evolution of the mineralization in the amniote eggshell outlined by Stewart ( 1997 )
suggests that stem-anapsids (such as pareiasaurs) and most Palaeozoic diapsids
would have laid poorly mineralized eggs with a low preservation potential.
These conclusions about the evolution of the amniotic egg are compatible with
the hypothesis that the earliest amniotes performed embryonic retention, although
they do not prove it. Furthermore, they provide a compelling hypothesis for the
absence of fossil eggs for the ® rst 80 or 90 million years of the history of amniotes.
Acknowledgements
This research was funded by the Alexander von Humboldt Foundation and
NSERC (Canada) . We thank Ms P. Lai for improving the style of this manuscript.
References
Balakrishna, T. A., Gundappa, K. R. and Shakuntala, K. A., 1983, Observations on the
eggs and embryo of Ichthyophis malabarensis ( Taylor) ( Apoda: Amphibia), Current
Science, 52, 990± 991.
Blackburn, D. G., 1993, Chorioallantoic placentation in squamate reptiles: structure, function, development, and evolution, Journal of Experimental Zoology, 266, 414± 430.
Bonaparte, J. F. and Vince, M., 1979, El hallazgo del primer nido de dinosaurios Triasicos,
( Saurischia, Prosauropoda), Triasico Superior de Patagonia, Argentina, Ameghiniana.
Revista de la Asociocio n Paleontolo ngica Argentina, 16, 173± 182.
Brauer, A., 1897, I. BeitraÈge zur Kenntniss der Entwicklungsgeschichte und der Anatomie
der Gymnophionen, Zoologische JahrbuÈcher, Abteilung fuÈr Anatomie und Ontogenie der
T hiere, 10, 389± 472.
Breckenridge, W. R. and Jayasinghe, S., 1979, Observations on the eggs and larvae of
Ichthyophis glutinosus, Ceylon Journal of Science. Section B. Zoology and Geology.
Spoila Zeylanica, 13, 187± 202.
Carroll, R. L., 1964, The earliest reptiles, Zoological Journal of the L innean Society, 45,
61± 83.
Carroll, R. L., 1988, Vertebrate Paleontology and Evolution ( New York: W. H. Freeman
and Company), 698 pp.
Duellman, W. E. and Trueb, L., 1986, Biology of Amphibians ( New York: McGraw-Hill ),
670 pp.
Laurin, M. and Reisz, R. R., 1995, A reevaluation of early amniote phylogeny, Zoological
Journal of the L innean Society, 113, 165± 223.
Laurin, M. and Reisz, R. R., 1997, A new perspective on tetrapod phylogeny, in S. Sumida
and K. Martin (eds) Amniote Origins Ð Completing the Transition to L and ( London:
Academic Press), pp. 9± 59.
Lombardi, J., 1994, Embryo retention and evolution of the amniote condition, Journal of
Morphology, 220, 368.
Mikhailov, K. E., 1991, Classi® cation of fossil eggshells of amniotic vertebrates, Acta
Palaeontologica Polonica, 36, 193± 238.
Norell, M. A., Clark, J. M., Demberelyin, D., Rhinchen, B., Chiappe, L. M., Davidson,
A. R., McKenna, M. C., Altangerel, P. and Novacek, M. J., 1994, A theropod
dinosaur embryo and the a nities of the Flaming Cli€ s dinosaur eggs, Science, N.Y .,
266, 779± 782.
Sarasin, P. and Sarasin, F., 1887± 1890, Ergebnisse Naturwissenschaftlichen. Forschungen auf
Ceylon in den Jahren 1884± 1886. Zur Entwicklungsgeschichte u. Anat. der ceylonische
Blindwuhle Ichthyophis glutinosus ( Wiesbaden: Keidel’s Verlag), 263 pp.
Shine, R., 1985, The evolution of viviparity in reptiles: an ecological analysis, in C. Gans and
F. Billet (eds) Biology of the Reptilia, Vol. 15 ( New York: John Wiley), pp. 605± 694.
Caecilian viviparity and amniote origins
315
Stewart, J. R., 1993, Yolk sac placentation in reptiles: structural innovation in a fundamental
vertebrate fetal nutritional system, Journal of Experimental Zoology, 266, 431± 449.
Stewart, J. R., 1997, Morphology and evolution of the egg of oviparous amniotes, in S.
Sumida and K. Martin (eds) Amniote Origins Ð Completing the Transition to L and
( London: Academic Press), pp. 291± 326.
Wilkinson, M. and Nussbaum, R. A., 1998, Caecilian viviparity and amniote origins, Journal
of Natural History, 32, 1403± 1409.