The oral cirri of Siluroids Htid the oiiKin of the head in Vertebrates. 417 



the most primitive and archaic forms that possess the tentacles most 

 fiilly developed. 



All parts of an organ may not revert to the ancestral condition, 

 or in other words the reversion may be only partial. Such is the 

 case in the tentacles of Cobitidae where the skeletal axis is not 

 developed. In the language of Weismann the reversion in this case 

 is due to determinants in the skin, the skeletal determinants not 

 being evolved. 



When once a structure has arisen by reversion and been rendered 

 constant by natural selection, it will develope ontogenetically direct 

 to the adult condition, and therefore it is useless to seek for infor- 

 mation as to its ancestral history in its embryological history. 



It will I think be obvious, to anyone fully acquainted with the 

 writings of Darwin and Weismann, that reversion may occur at any 

 free living stage. Larval forms are often supposed to represent an- 

 cestral or existing adult forms. The resemblance has no doubt been 

 greatly exaggerated. For instance I am not aware that Ammocoetes 

 shows any approach in positive characters to Myxine or Amphioxus. 



Nevertheless such characters as the prepalatine piece of tadpoles, 

 and the maxillo-coronoid tentacles of the larva of Bactylethra^ at its 

 fancied "Siluruid'' stage, have to be accounted for. Tentacles do not 

 occur so far as I know in other tadpoles nor in the larvae of Uro- 

 deles. Therefore probably we have here a case of larval reversion, 

 but it is only an exceedingly incomplete reversion. 



The Zoological Position of Siluroids. 



The Siluroids are mostly freshwater fish with an extraordinary 

 diversity of habits and structure and with a remarkably wide geo- 

 graphical distribution. While however the group as a whole occurs 

 in all the zoogeographical regions, yet certain families are confined 

 to one. Thus, for instance, the Loricarina with peculiar anatomical 

 features appear to be confined to the rivers of S. America. (Certain 

 forms from the oriental region have been allied with them by some 

 ichthyologists.) They are freshwater forms without any means of 

 passing across seas, being heavily armoured with feeble powers of 

 swimmini;, lying at the bottom of pools in the daytime, and creeping 

 about at night on banks by means of their strong spines, and feeding 

 on soft substances more or less putrified (Weyenbergh). 



We are justified by the principles of geographical distribution 

 in attributing to Loricarina an antiquity like that of Lepidosiren. 



