No. 490] ORTHOGENETIC VARIATION 



631 



Fusus, and always associated with the Fasus y>rotoconch, while 

 "Fusus serratus" with a distinct protoconch has whorls more 

 specialized than those of any other true Fusus, even those of the 

 present day. Why do the fusoid shells in which the normal post- 

 embryonic developmental stages of Fusus occur, always have a 

 Fusus protoconch, while those fusoid shells which do not show the 

 true Fusus protoconch do not show the normal ontogenetic stages 

 of true Fusus, if this is not an expression of inheritance, and of 

 more fundamental significance than Cossmann, Boettger, or Giard 

 are welling to admit ? It is high time that we cease making gen- 

 eralizations and tracing relationship by a superficial study of shell 

 characters. Such superficial study has deservedly brought the 

 whole subject of conchology into disrepute, so that morphologists 

 have come to look upon shell characters as the least reliable indices 

 of genetic relationships, whereas they are really the most reliable 

 and delicate of such indices, if subjected to a critical study. 



Elimination of Early Conch Characters by Acceleration. 



It not infrequently happens that the protoconch stops abruptly, 

 and the conch begins as abru])tlv. In fact, it may he sai<l that this 

 is normal for such specialized tvjx-s as the tii.soid shells .^tudie^ I, 

 figs. 1 and 5). In normal pnimtixr types the pn-tocmich char- 

 acters may be expected to nicruv bv slow dtuTccs into iho.se of the 

 conch Sn(h is the case e\en m spaiali/td along one direc- 



tion or another. In highly specialized types, however, we often 

 find an abrupt transition from the characters of the protoconch 

 to those of the conch, these latter beginning suddenly with a num- 

 ber of special characters. Thus in Fusus, the protoconch ends 

 abruptly with a strong varix, and the conch begins with round 

 whorls, ribs and spirals. Here protoconch and conch h-.wv l-eeii 

 telescoped, so that tlu tian-itional (haia(t(i^ nndoiil.K db po^- 



ditions of this kind exist \n a nunihei ot t^pes wludi nia , oi nia\ 



w^herever it occurs, for it is cleatl\ a siaue ni di \( lopint iit. and 

 therefore a condition, which may ap|)eai- m the specialized rei iuiiial 



