ON THK STJIUCTUKE AND CLASSIFICATION 
rapidly alon, diverse lines as eonditioned by the mode of addition of seeondary 
cusps to the crowns can now be to be the single 
cone b„ our present evidence does not support the hypothesis that there was but one 
of addition of cusps to this cone. On the other hand as descri ed in detail in 
the foregoing sections, we 4nd evidence of three or more modes of addit on. 
u In one series, to which the larger number of genera belong, lateral cusps 
were added upon tbe anterior and posterior faces of the protocone forming the tricouo- 
donl fvpc. This tvpe persisted in one line of genera (Tncarmlori), with a gradual 
increase of the lateral cusps or para- and metacones. In another line of genera the 
imru-and metacones were rotated inwards to form the <,*,4crc«?ur type (Syiuluco- 
rtcmmi, AMmclon). In a third line of genera, this process was probably completed 
liefore the upper Jurassic period, together with the addition of a heel, forming the 
tnlyercuhir sectorial type i^Stylacodoii). 
h. In a second series, only tlie anterior cusp or paracone was developed upon the 
protocone, tlie metacone being replaced by a heel which became continuous with the 
internal cingulum, and the latter gave rise to internal cusps {AmpMthenum) . 
c. In a third series, neither the paracone nor metacone were developed upon the 
protocone but the crown was reinforced by the development of cusps from the posterior 
heel and from the internal cingulum. Or, the postero-internal basal cusp represented 
the metacone and the antero-internal the paracone {Peralestes) . 
<1. In a fourth series, represented by the single genus Leptocladus, there was 
simply an clongjite heel behind the rnain cone, the molars having the same general 
|>atteni as the premolars. 
e. The mode of development of the prismatic columnar crown of the genus 
Kurlodon, is unknown. It may have sprung from the tritubercular type, in which a 
complete union of the internal cusps has left a record of the transverse valley in the 
line of enamel extending across the crown. 
The following is an hypothetical scheme of the mutual relations of the Mesozoic 
families and genera, founded upon the homologies and reduction of the teeth, as con- 
sidered most probable in the above analysis. It is iirtended, not to show the actual 
line of succession, for our paheontological record is far too imperfect for such an 
attempt, but as an outline of a possible line of succession in which the genera are 
taken merely as types representing certain stages of development of the molars. At 
the sjime time, the diagram does express the author’s present views of the degree of 
separation of the families from each other. There is, for example, little doubt that 
the Stylw'odonhdiv have diverged from the common stem at an earlv period, since they 
present the most modem type of molar known at this period, excepting perhaps the 
Kurtodontidw. * The central line is through the Tricemodontidee. The Spalacotherii- 
foe and Pliascolotheriince may have branched from this. The position of the Arnplii- 
