Annales des Sciences Natttrelles. 255 



principle is followed up at some detail in the enumeration of the 

 genera according to the regions in which they are found. Con- 

 trary however to an opinion very frequently entertained, MM. 

 Qijoy and Gaimard declare that fishes do not swarm in the ocean, 

 ■which possesses, like the land, its solitudes and deserts, inhabited 

 only by those species which chiefly or entirely exist by prey. 



In the " Sketch of a general distribution of the Mollusca, by M. 

 Latreille," we are presented with the first fruits of the studies of 

 this well-known Entomologist in those departments of natural 

 science which have devolved upon him in consequence of the 

 severe affliction of the veteran zoologist Lamarck. In this ex- 

 tract from an unpublished work on the " Natural Families of 

 Animals," M. Latreille, proceeding on a principle frequently 

 adopted by Cuvier, divides the Mollusca into Phanerogama, in 

 which a coitus is necessary, and Jgama, in which the individuals 

 are capable of self-fecundation. The first of these is composed 

 of the Pteri/giens (comprehending the Cephalopoda and Ptero- 

 poda)f and the Apterygiens, which include the whole of the 

 Gasteropodut with the exception of the Scutibranches and the 

 Ci/clob ranches. The principal sections of the Apterygiens are 

 also formed on a principle deduced from the generative faculty, 

 being either Hermaphrodite, or Dioicous ; each of which embraces 

 various orders breathing free air, or furnished with branchiae. In 

 the formation of these the method of Lamarck is in general fol- 

 lowed, although frequent variations are introduced in the mode of 

 their arrangement, which concludes with the genus Sigaretus, the 

 succeeding division commencing with llalioiis. This division, or 

 the Agama, also forms two sections, Exoccphala^ including the 

 Scutibranches and Cyclobranches^ and Endoccphala^ which is 

 composed of two classes, the Brachiopoda and the Cunchifera of 

 Lamarck. The sections of the latter follow the anatomical me- 

 thod, and are the Paiulipalla, BiforipaUa^Tr/foripalla, and Tubi- 

 palla, the three former of which correspond with (hose of M. 

 Cuviec, whose two other families are included ii\ the las*. 



This distribution appears in many instances to succeed in point- 

 ing out afiBnities existing between the groups, but the principles 

 on which it proceeds are too limited to lead to a truly natural 



