262 B'ulletin American Museum of Natural History. [Vol. XXVII, 



meatus and connecting the post-glenoid and post-tympanic processes. Thus 

 as seen from above the tympanic fossa of Irfops montanus is very broad and 

 square externally, whereas in Erinaccus, through the coalescence of the post- 

 glenoid and post-tympanic process, the tympanic fossa is triangular. The 

 mastoid exposure in Ictops faces backward, in Erinaceus backward and out- 

 ward. In this one point Erinaceus agrees better with the Zalambdodonts 

 than does Ictops. The large round })it in the under surface of the basi- 

 sphcnoid of Erinaceus is not represented in Ictops. Occipital condyles are 

 represented in one specimen (Douglass, /. c, pi. xxii, fig. 1) as extending 

 entirely across the basioccipital. The paroccipital processes appear to have 

 lieen absent {cf. Microgale). The occiput was broad and nearly flat. The 

 angle of the mandible was gently inturned, as in certain species of Erinaceus. 



In the skeleton Douglass's figures sho\\' that the cvu'ved shaft of the ulna 

 was stout, the olecranon short, the femur had a well developed third tro- 

 chanter, the fibula was already fused with the tibia; the astragalus had a 

 well-grooved narrow trochlea and the anterointernal end of the astragalus 

 was produced into a broad prominence, which is not however prolonged 

 into a distinct process (contrast Mircogale); the tuber of the calcaneum was 

 stout and not long, betokening a ])lantigrade foot. All these are characters 

 which might be expected in an ancestor of Erinaceus. 



In conclusion, Ictops was apparently much more primitive than Erinaceus: 

 in the very small size of some of its species, in the long slender muzzle, small 

 subcaniniform upper canines, procumbent lower incisors, very primitive 

 narrow tritubercidar upper molars, slender mandible, horizontal well 

 developed malar, unfenestrated palate, unexcavated basisphenoid, large 

 post-glenoid ridge, separate post-glenoid and post-tympanic processes. In 

 practicalh' all these characters Ictops makes some approach backward to- 

 ward the stem of the Zalambdodonta. 



The foregoing review of the chief characters of Ictops, the most primitive 

 of the I^eptictida\ confirms the lo\\ly position of that family and its ancestral 

 relation to the Erinaceidse. Ictops also })ossessed a number of the characters 

 assumed below (p. 272) for the ancestors of the Tupaiidse and other JNIeno- 

 typhlous Insectivora, but how far these indications are valid is at present 

 impossible to say. 



The Leptictid skeleton and skull, so far as known, do not show any fos- 

 sorial adaptations (unless the fusion of the tibia and fibula be so interpreted) 

 and some of the smaller species may have been arboreal. 



Of the recent Erinaceidse Gymnura and the nearly allied Ilylomys ^ 

 are more primitive than Erinaceus in the following characters : 



1 Excellent figures of the skeleton of Hylnmys peguensis are given by Anderson (1874, pi. 

 Ixiv). 



