248 FOOT-NOTES TO EVOLUTION. 
going through a TZirolites stage in its early youth. 
Neither does the first discovery of a genus always cor- 
respond with its first appearance. For example, tere- 
bratuloid brachiopods are not known below the Devo- 
nian, but their existence in the Silurian is necessitated by 
the occurrence of a terebratuloid stage in the ontogeny 
of spire-bearing brachiopods (Azcylobranchia) of that age. 
One great drawback to this work is that the am- 
monite faunas of the various ages have been classified 
by different specialists and on different principles, but 
all artificial. Thus the Triassic ammonites are divided 
into Leiostraca (smooth-shelled) and Trachyostraca 
(rough-shelled), a classification that can not be extended 
even to Jurassic groups. The Trachyostraca are further 
divided into 7ropitide, with long body chamber, and Cera#i- 
tide, with short chamber. But neither of these groups is 
monophyletic, for it is quite probable, judging from their 
ontogeny, that members of both groups are derived 
from the Goniatitide, and others from the Prolecanitida. 
Further, the authorities agree in deriving the Zropitide 
from the Glyphioceratide, but the larval stages of some 
of the Zropitide show the undivided ventral lobe and 
an unmistakable resemblance to certain Prolecanitide ; 
other so-called Zvopitide show the divided ventral lobe 
at an early age, and a decided resemblance to the stock 
of Glyphioceratide. 
In the same way most authorities agree that the Tra- 
chyostraca were all extinguished at the end of the 
Trias, and that all the Jurassic and Cretaceous ammon- 
ites, with the exception of Lytoceratide and Phyllocera- 
tide, were derived from the radicle Psi/oceras, and this, 
too, in spite of the fact that many of the genera are 
rough shelled, and in their larval stages show marked 
likeness to trachyostracan genera. Any naturalist can 
convince himself of this by looking at the young stages 
