128 PHALANGEETD*. 



peared in most of the modern Phalangeridce *. To ttis conclusion I 

 am brought by the comparison of such premolars as those of Phalanger 

 and Pseudochirus, by which one sees that it is neither p.' nor p.* that 

 has disappeared, and, as to p/, by the position of the anterior premolar 

 (when present) in Trichoswms, and, above all, by the fact that 

 occasionally in Phalanger f two minute teeth are present between 

 the large anterior and posterior premolars, the numeration of which 

 would be impossible were the anterior functional premolar not 

 reckoned here, and therefore (presumably) elsewhere, as p.^ An 

 exception to this rule, however, is shown by Dromicia nana (see 

 below, footnote, p. 145), which rather appears to have lost p.' than 

 p.^ ; and therefore it is possible that as the loss of the fourth pre- 

 molar is certainly a very recent occurrence, different teeth may have 

 been lost in different genera. 



The homologies of the lower rudimentary teeth are far more 

 difficult to make out, and I am only able to make quite a tentative 

 interpretation of them. Taking, again, the largest numbers known, 

 we find that occasionally there are five, and, at least in one 

 instance J, six teeth between i.' and m.' ; of these six the last two 

 are obviously p.' and p.*, and the others should, I think, be reckoned 

 as i.^, i.', canine, and p.', the little anterior tooth at the base of i.' 

 being therefore i.^, and not the canine as it has been generally 

 considered. When there are, as is of course usually the case, much 

 fewer teeth in the jaw than this, they may apparently be generally 

 reckoned as having disappeared in the following order — canine, p.' or 

 i.", then p.", and lastly i.^ In one genus only, Distoechurus, p.* is lost, 

 but this loss is unaccompanied by that of p.^ or p.' Thus the four 

 intermediate teeth commonly present in Petaurus or Daetylopsila 

 (fig. 4) would be i.^, i.^, p.', and p.'', while the two so often only 

 present in Triehosurus would be i.^ and p.* This general rule 

 would, of course, be modified in individual cases by the variations in 

 the positions of the different teeth ; and again in some cases, as for 

 example in many specimens of Pseudochirus, it is impossible to 

 homologize every individual tooth, the variations in position being 

 apparently much too erratic §. 



As to the means of formulating the various dentitions found in 



* This was suggested, but not definitely asserted, in the paper in which the 

 homologies of the Polyprotodont premolai-s were worked out (see Phil. Trans. 

 1887, vol. clxxviii. p. 447, footnote). In the Mesozoio Plagiaulacidie, however, 

 also Diprotodont, it was clearly p.' that was first lost ; but they were evidently 

 even at that time so exceedingly specialized that they could not have been the 

 direct ancestors of any of the modem Diprotodonts, and therefore their evidence 

 on this point has but little importance. 



t E. g. the specimen (no. 104) of Fhalanger onentalis described by Jentink, 

 Notes Leyd. Mus. vii. p. 90, of which that author has kindly sent me a sketch. 



\ The same specimen referred to in the last footnote. 



§ An alternative arrangement would be by looking upon the tooth here 

 caUed i.' as the canine, on that called the canme as p.', and on that called p.' 

 as p.^ and therefore looking upon i.' instead of p.'' as the tooth always absent, 

 even in the many-toothed Phalanger above referred to. On the whole, how- 

 ever, until furtl\er evidence is brought forward, I prefer the homologies given 

 above. 



