SINCLAIR: MARSUPIALIA OF THE SANTA CRUZ BEDS. 
441 
Posterior premolar, transverse diameter . 
'My, antero-posterior diameter .... 
“ transverse “ .... 
“ height of middle of crown, above alveolar border 
My, antero-posterior diameter .... 
“ transverse “ .... 
M3, antero-posterior “ .... 
“ transverse “ .... 
My, antero-posterior “ .... 
“ transverse “ .... 
Depth of mandible below middle of sectorial . 
“ “ “My ... 
No. 15,425. No. 15,079. 
.0008 
.005 
.0023 
.OO45 
•0033 
.OO33 
.0022 
.002 5 
.003 
.003 
.002 
.002 
.002 
.0017 
.OO65 
.0065 
.006 
.006 
RELATIONSHIPS OF THE CRNOLESTIDzE. 
The most primitive Santa Cruz representative of the Csenolestidae is, 
undoubtedly, the genus Halmarhiphus , which represents, with little or no 
modification, a type which is not only ancestral to the Palaeothentinae, but 
agrees perfectly with the “minute insectivorous forms which, apart from 
the diprotodont modification of the antemolar teeth, possessed a full ante- 
molar formula,” indicated by Bensley’s studies as the ancestors of the 
Phalangerinae. In this interesting genus the dental formula and molar 
patterns are didelphid, affording striking evidences in favor of the theory 
of the didelphid origin of the Diprotodontia (see Bensley, 1903). It has 
already been shown (p. 417) that the lower molar patterns in the Palaeo- 
thentinae are readily derivable from the Halmarhiphus type by a reduc- 
tion in height of the cusps and the formation of cross crests. So far as 
our knowledge of Halmarhiphus warrants an inference, there can be little 
objection to deriving the Palaeothentinae from a similar ancestral form. 
The chief objection to regarding Halmarhiphus as directly ancestral to 
the Palaeothentinae arises from the fact that the latter are represented in 
formations older than the Santa Cruz ( P chufaitensis from the Pyro- 
therium beds, Amegh., 1897, P- 9^)- That Halmarhiphus is in the direct 
line of descent culminating in Ccenolestes will, it is believed, hardly be 
questioned after an examination of the accompanying plates. So far as 
the lower dentition is concerned, no argument can be advanced to the 
contrary. The recent genus shows a slight tendency toward the develop- 
ment of crescents, while, in its Santa Cruz predecessor, the cusps are high 
