SUR LES POISSONS FOSSILES DB l'ARGILE DE LONDRES. 281 



plupart des poissons se sont pourris dans ce iin limon, leurs os se sont de- 

 taches, et les parties moUes ont ete remplacees par du limon. Or comma ce 

 ne sont plus des Ganoides a corps cuirassee recouverts d'ecailles osseuses en- 

 chevetres, mais des Cyclo'ides, des Cteno'ides a ecailles minces, fragiles, leur 

 enveloppe n'a pas ete assez solide pour maintenir I'integrite de leur forme et 

 de leurs contours. Leur corps s'est decompose, leurs nageoires se sont de- 

 faites, leurs ecailles desagregees, et il n'est reste du plus grand nombre que 

 les boites craniennes qui se sont conservees en entier, grace a la soudure de 

 leurs pieces osseuses. Si au lieu d'appartenir a des poissons, ces cranes pro- 

 venaient de mammiferes ou de reptiles, il est a presumer qu'on en tirerait tout 

 le parti possible, et que le paleontologiste n'aurait pas de peine a les deter- 

 miner, car pour ces classes les materiaux de comparaison ne manquent pas, 

 les points de depart sont tixes ; on connait les traits caracteristiques des 

 cranes des mammiferes et des I'eptiles, on sait quelles sont les variations que 

 tel OS, telle crete, telle fosse' part subir dans telle ou telle famille, et du pre- 

 mier coup d'ceil deja on peut s'assurer, si I'animal qu'on a devant les yeux, est 

 un carnivore, un ruminant, ou un solipede. 



Mais rien n'est variable conime les formes du crane et de la tete des pois- 

 sons. Ces multitudes d'aretes et d'epines qui servent d'attaclie aux muscles, 

 cette infinie variete de formes dans les families elles-memes, donne aux cranes 

 des poissons une telle diversite que I'ichthyologiste desespere souvent de pou- 

 voir les ramener a leurs types respectifs, et en eiFet une craniologie compares 

 des poissons n'existe pas, et il n'est personne que je sache qui puisse dire 

 d'emblee si tel ou tel crane appartient a un Perco'ide, a un Sparoi'de, ou a un 

 Chetodonte, &c. 



La grande majorite des fossiles de Sheppy avons nous dit, consiste en ver- 



fectly soft. The greater number of the fish have rotted in this fine clay, 

 their bones have separated, and the soft parts have been replaced by clay. 

 Now, since it is no longer Ganoids with cuirassed body covered with in- 

 terlocked bony scales, but Cycloids and Ctenoids with thin fragile scales, 

 their coating has not been sufficiently solid to preserve the integrity of their 

 form and outline. Their body has become decomposed, their fins have be- 

 come detached, their scales disaggregated, and of the greater number only 

 the cranium has remained preserved entire, owing to the soldering of the 

 osseous pieces composing it. If, instead of belonging to fish, these skulls were 

 derived from Mammalia or reptiles, it is to be presumed that all possible ad- 

 vantage would be taken of them, and that the palaeontologist would have no 

 trouble in determining their relations, since for these classes the materials for 

 comparison are not wanting, the points of departure are fixed. The charac- 

 teristic features of the skulls of the Mammalia and Reptilia are known ; 

 the variations which such a bone, such a crest, such a groove may undergo 

 in such and such a family are understood, and already at the first glance it is 

 possible to ascertain whether the animal under consideration is carnivorous, 

 ruminant, or solipedal. 



But nothing is more variable than the forms of the cranium and of the 

 heads of fish. The multitude of bones and of spines which serve for the at- 

 tachment of the muscles, the infinite variety of forms in the families them- 

 selves, imparts such a diversity to the crania of fish, that the ichthyologist 

 frequently despairs of being able to reduce them to their respective types, 

 and in fact a comparative craniology of fish does not exist. There is no 

 one that I know who can tell at first sight whether such and such a cranium 

 belongs to a Percoid, to a Sparoid, or to a Chetodontal type. 



The great majority of the fossils of Sheppey consists, we have said, of de- 



