EXTINCT  RODENTS Among extinct rodents, only a few of the more important types may be noticed . As to the origin of the

order, we are still to a great extent in the dark; and even the relations of the Duplicidentata to the Simplicidentata are not yet fully understood . With regard to the latter point, it is, however, considered probable that both are branches of a common stock, which diverged from each other before ali the typical rodent characters were acquired . As to the ancestral stock of the order, it has been suggested that this is re-presented by certain Lower Eocene European and North American mammals, at one time regarded as primitive Primates . In Europe these include Plesiadapis and Protoadapis, and in North America Mixodectes, Microsyops and Cynodontomys; the last three constituting the family Mixodectidae . Possibly the European forms, in which the dental formula has been given as i . , c . 8, pi, m . I, and there is a gap between the incisors and the cheek-teeth, are more nearly related to modern rodents than the American types, and may indeed belong to the same order . On the other hand, the American forms, which have one pair of large chisel-like incisors in the lower jaw, also possess a lower canine, and show no marked gap in front of the cheek-teeth, nor any indication of the characteristic rodent backwards movement of the lower jaw . On these grounds, while admitting that they are allied to the rodents, it has been pointed out that they can scarcely be included in the Rodentia, and the order Proglires has in consequence been proposed for their reception . Whatever may be the true affinity of these problematical mammals, undoubted rodents are known from the Lower Eocene of both Europe and North America .

In Europe these

form the genus Ischyromys and the family Ischyromyidae, and have premolars 't, and all the cheek-teeth low-crowned, with simple cusps or ridges . Possibly they are akin to the Sciuridae . In America, Paramys, with transversely ridged Molars, is allied ; and the European Sciuromys should perhaps find a place in the same neighbourhood . A more advanced phase is represented in the European Lower Oligocene by the Pseudosciuridae, with the genera Pseudosciurus, Sciuroides, Trechomys, Theridomys, &c., in which part of the masseter passes through the broad infra-orbital canal, and the premolars are i ; the molars being low-crowned, many-rooted and either cusped or ridged . These rodents are thought to be allied to the Anomaluridae ; and it is partly on their evidence that the family Pedetidae is placed next the latter . Here it may be mentioned that Leithia, from the Pleistocene of Malta, originally regarded as a giant dormouse, seems near akin to Anomalurus . In the highly specialized mastoid region of the skull, the North American Oligocene Protoptychus approaches to Dipopodomys, while the contemporary Gymnoptychus and Entoptychus likewise appear referable to the Geomyidae . The Upper Oligocene Cricetodon in Europe and Eumys in America are the earliest known forerunners of the cricetine Muridae; while at the same time primitive beavers appear in the form of Steneofiber, to be succeeded in the European Pleistocene by the gigantic Trogontherium . The still larger North American Pleistocene Castoroides, known by one species of the size of a bear, and the allied West Indian Amblyrhiza, appear to be specialized beavers, although they have been referred to a family by themselves . Near akin is the North American Miocene family Mylagaulidae, typified by Mylagaulus, but including Mesogaulus and Protogaulus . Although showing some dental characters approximating to the porcupines, these rodents are regarded as allied to the Castoridae, although forming an isolated type . The prominent feature, writes Mr E .

S . Riggs, is the unusual development of the premolar to the exclusion of the posterior teeth . Associated with this is the strength and sharpness of the lower jaw, the prominence and anterior pcsition of the masseteric

ridge, and the depth of the ramus from the alveolar line to the angle . These indicate unusual capacity for crushing or grinding; while the last premolar is a crushing implement, which has reached the highest degree of specialization known in Rodentia . It is suggested that these teeth may have been employed for cracking nuts or hard seeds, although also used for grinding . The remarkable North American Ceratogaulus, with a large bony nasal horn, belongs to the same family . To discuss the remaining Miocene and later fossil Simplicidentata would be doing little more than adding to the generic names referable to the various existing families . It may be mentioned, however, that the distribution of these later Tertiary types accords very closely with that of their existing relatives; the families of South American hystricoids being represented by a number of extinct genera in the formations of Argentina and .Brazil . Special mention may be made of Megamys, from the caves of Brazil, which, while apparently allied to the living viscacha, attained dimensions approximating to those of a hippopotamus . As regards the Duplicidentata, it appears that the families Ochotonidae and Leporidae had become differentiated as early as the Lower Miocene . Titanomys is the earliest form, from the Middle Miocene, succeeded by Lagopsis, and then by the modern Ochotona . In this line there is a tendency to lose the last upper molar, but in Prolagus, which ranges in the Pliocene from Sardinia and Corsica to Spain, and forms a side-branch, the corresponding lower toothhas likewise disappeared .

In contradistinction to Titanomys, in which the cheek-teeth are rooted, is the North American Upper Oligocene Palaeolagus, where they are rootless . In

general dental characters, especially the retention of three pairs of molars, this genus approximates to the Leporidae, although in the absence of post-orbital processes and the pattern of the molars it departs less widely from the modern Ochotonidae than does Prolagus .