with Observations on the General Arrangement of the Articulata. 299 



Scolopendra the full number of teeth is acquired at an early period, usually 

 after the third or fourth change of tegument, when the individual is very 

 small, and before it has gained its full complement of legs. Hence the 

 number and form of the teeth are good characters of species in all that have 

 assumed the adult form. These are, perhaps, the best structural characters 

 in the adult Lithohius when taken in connexion with the number of ocelli. In 

 Cermatia they are uniform in their appearance, and consequently afford only 

 generic characters ; but in the great genus Scolopendra the dental plates and 

 the teeth are so developed as to afford both characters for the division of the 

 genus into groups and for the identification of species. Thus in the first divi- 

 sion, which seems to be nearest allied to Lithobius, the Parvidentatce, the plates 

 are very short, transverse, and almost quadrangular, and the teeth are very 

 small and more numerous than in the other divisions of the genus. In the 

 second division, the Latidentatce, the plates assume a more peculiar character : 

 they are large, distinctly divided from the lip by a deep suture, and often 

 have the posterior external angle elongated. The teeth on their anterior 

 margin are usually very large* and fewer in number than in the preceding 

 division, and are more adapted for cutting and tearing. The internal one on 

 each plate is usually broad and spatulate, and the external triangular and 

 acute. In Rhomhocephalus the plates are very much narrowed at their ante- 

 rior border and widened at their posterior, and are less distinct from the labium 

 than in the other section, and the teeth are fewer or smaller. In Heterostoma 

 the teeth are elongated, and are triangular, larger, and more acute than in 

 either of the others. In Cryptops the plates exist, but the teeth are absent ; 

 while in Geophilus the plates also are undeveloped. 



The spines on the posterior legs, and the form of the anal plates, as pointed 

 out by Brandt, afford the next best characters for the species and subdivisions 

 of Scolopendra. The spines are not developed on any part of the leg in Scolo- 

 pendra but the femur, and perhaps also the coxa. The posterior internal angle 

 6f the femur is always more or less elongated into a spine, even when spines on 

 other parts of the joint are entirely absent. But the characters derived from 

 this part of the body can only be depended on when the organ is one of the 

 original members, and is not a reproduced limb, since if it happen to be the 

 latter, — which may be known by comparison with the limb of the opposite 



VOL. XIX. 2 R 



