I go I.] Matthew, Additional Observations on the Creodonta. 1 9 



us a sure key ; for no case of exact parallelism in both teeth and 

 feet between two unrelated types has yet been found or is likely 

 to be." {Osborfi, 1893, p. 10.) If Clcenodo?i be totally unrelated 

 to the Ursidae it is an exception to this statement, and, as far as 

 I recall, the only one. And yet the evidence is very strong for 

 deriving the Ursidae from primitive Dogs. 



IV.— Family H Y^ENODONTID^ Cope. 

 In this family are included, as Wortman has already indicated 

 ('99, p. 139), most of the genera grouped by Schlosser under the 

 name Proviverrid^. Deltatheriuvi belongs with the Oxyclaenidae ; 

 the position of Didelphodus and Palceosinopa is uncertain, as they 

 fail to show the progressive characters of either group. Sinopa 

 and Cynohycenodon clearly belong in it, Proviverra somewhat less 

 clearly, while Quercytherium appears to be an aberrant member. 

 The progressive characters of the two families may be contrasted 

 thus : 



HytEnodontida:. OxyclcenidcB. 



Upper molars becoming quadrate. 

 Fourth lower premolar becoming mo- 

 lariform. A considerable diastema 

 developing behind canines with disap- 

 pearance of p^. Dentition function- 

 ally insectivorous (cp. Lemurs, 

 Opossums, and some Insectivores). 



Upper molars with connate exter- 

 nal cusps, developing a shear, espe- 

 cially between m^ and mg. Last 

 upper molar becoming transverse and 

 disappearing. Heels of lower molars 

 disappearing, me- disappearing, pa^ 

 advancing to form a cutting blade. 

 Dentition functionally carnivorous (<:/". 

 Felidae). 



All the above Hyaenodont characters are exhibited to a greater 

 or less extent by various species of Sinopa, though always in an 

 early stage of progress. Most of them are exhibited by Cynohy- 

 CBfiodon and Proviverra, which besides are extremely close to 

 Sinopa. The same is true of Quercyt lerium except for the very 

 peculiar premolars (which are approximated in Cynohycenodon 

 minor if the reference to this species be correct of a fine speci- 

 men in the Paris Museum). Palceosinopa and Didelphodus can 

 be placed here only on account of general resemblance to the Pro- 

 viverrines and lack of the progressive characters of any other 

 group. The generic definitions will be : 



Hyccnodon. M^ absent ; m^~^ without protocone, para- and metacone united ; 

 metastyle forming a shearing blade larger than the united pa. and me.; 

 mi„3 without metaconid or heel. Jaw long, slender. 



