QUATEENAET FAUNA OF GIBEALTAR. 67 



In U. arctos the tooth is also triangular in form, but usually more angular behind, in 

 accordance with the more pointed shape of the corresponding upper molar. It has 

 usually no constriction on the outer border ; and the anterior and inner cusp is generally 

 but little developed. The grinding-surface presents only a few coarse folds or rugae, 

 and is never, so far as I have seen, tuberculated or granulated in the slightest degi'ee. 



The fourth lower premolar (pm. 4) is in some respects the most characteristic of all 

 the teeth; and were its characters rather less liable to vary by defect in U.fossilis and 

 U.ferox, it might almost by itself be considered sufficient to give assurance of the species. 

 Its distinctive characters in U. spelceus are well known ; and when fully developed and 

 unworn it seems to me to afford excellent distinctive characters also between U.ferox 

 and U. aixtos. 



In Ursus spelceus, besides the principal cone, there are usually on the inner side two. 

 and always one, smaller cusps, of which one is anterior in position to the principal cusp. 

 In all other Bears this tooth has either a single conical cusp, or at most a single small 

 internal tubercle posterior in position to the principal cusp, and corresponding to the 

 hinder of the two iuternal cusps in U. spelceus. 



In all the large carnivorous Bears this tooth pi-esents several common characters, the 

 differences exhibited in various species depending solely, as it would seem, upon the 

 degree of development or suppression of minute parts. 



In all cases the tooth presents a large conical cusp, which is placed nearer the 

 anterior than the posterior border of the crown. An acute ridge or keel, in perfectly 

 unworn teeth, descends from the point of the cusp in front to the anterior end of the 

 crown, where it terminates after making a slight curve inwards, in a more or less distinct 

 though always very minute tubercle. Two similar but more strongly marked ridges 

 descend from the apex of the cone towards the hinder border of the tooth. 



In fully developed and perfectly unworn teeth in U. fossilis, ferox, and maritimus 

 (and, I have no doubt, also in U. arctos, though I have seen no tooth of that species young 

 enough to show it), these hinder ridges are more or less distinctly serrated, especially the 

 outer one, which is always continued to the hinder border of the tooth, whilst the 

 internal ridge rarely reaches more than halfway between the base of the cone and the 

 hinder border. Now the differencesin the fourth lower premolar, as between U.fossilis, 

 ferox, and arctos, consist solely in the varying degree of development of these minute 

 parts. In well-marked teeth of U. fossilis taken in the genn-condition, the anterior 

 carina and the tubercle at its base are strongly developed. The outer of the two 

 hinder carince, which is deeply serrated, terminates at the posterior border of the crown 

 in a small tubercle, on the inner side of which is placed a second tubercle of equal size ; 

 so that the hinder extremity of the crown (or the talon, as it may be termed) might in 



appears to be deeply einuated on the outer border ; but this appears to have arisen from the circumstance that 

 the figure has been taken from Goldfuss's specimen in the British Museum, in which the tooth has a 

 piece chipped off at that spot. 



