95 
clusion is hardly more legitimate. The hypothetical "reptilian cone” can 
not always be obliged to appear first, if one of the other cusps in 
the process of development grows larger; it may even be allowed. to 
disappear if it has become functionless. It is most probable that 
the cusp which on-the adult tooth is meant to be the most vigor- 
0us may also appear as the first, in order to get the necessary 
start; Woodward's demonstration of the succession of tooth-cusps 
of various Insectivora (le, 1896) seems to: me to confirm this view. 
We thus adopt none of the Concrescence Theories; but what 
about; the Differentiation Theories? Let us start with the latest 
one. Gidley's statement of the construction of the Dryolestes molar 
seems very good; but his assertion that the hypocone should be of 
different origin in the various tooth-types surely needs further 
affirmation. With the exception of the Multituberculata the origin 
of the teeth of all mammals is undoubtedly monophyletic; the teeth 
of the most primitive tertiary and recent forms are all of the same 
fundamental pattern, though it may be somewhat veiled (Edentata, 
Cetacea). At any rate the rodents do not differ from the remaining 
mammals. Consequently there is no reason for searching after 
derivatives of the Triconodon, Dryolestes and Dicrocynodon molars 
in different recent orders. 
Tims's Cingulum-Cusp Theory is essentially based on the 
Conditions present in the various species of Canis; but for one thing 
this material is much too scarce, and again the teeth of Canidæ 
åre too specialised to indicate the evolution of the mammalian 
molar. Moreover Tims. has not been aware of the difference be- 
tween the constant cusps and the variable cusplets which may 
individually appear or disappear; for which reason he indicates 
sometimes a. cingulum-cusp, at others an accessory cusplet, and at 
still other times the real protocone by the term. protocone. His 
remarking that the more cusps a. tooth has, the less primitive it 
18, ig essentially correct; but in concluding from this that the 
3-cuspid upper molars must be more primitive than the 5-cuspid 
lower molars. he overlooks. the. fact. that a tooth can also lose its 
