INTRODUCTION. 
XXI 
(p. 204) at some Icngtli. Eosiren, tliongh primitive, is already in all respects a typical 
Sirenian, almost the only important characters distinguishing it from the later members 
of the order being: (1) the ])resence of traces of the second and third pairs ofdncisors 
and of the canines ; and (2) the rather less degree ot reduction in the ])elvis (see fig. G8 B, 
p. 214), which still ])osscsscs a well-defined acetabnlum. In fhe case of the incisors 
and canines reduction is already far advanced and they have also been thrust out on 
to the side of the snout, possibly by the development of the horny plate, which 
most likely already replaced them functionally. In Eotheriiim, from the lower beds 
of the Mokattam Hills, the incisors and canines are larger and occupy their normal 
position on the edge of the jaw, and the pelvis has a completely closed ohUirator 
foramen and a large and well-defined acetabulum, showing that probably the femur 
was still large and perhaps to some degree functional. In these points Eotherium 
approaches a normal land-mammal, but in other respects, so far as known, is 
essentially a Sirenian and its actual terrestrial ancestor must be sought in earlier 
deposits. One of the most striking Sirenian characters of the skull in both this genus 
and Eosiren is the deflection of the snout, a peculiarity evidently of great value to a 
short-necked, long-bodied, aquatic animal feeding on plants growing at the bottom 
of the water in which it lived ; in the most primitive type of Sirenian, the imperfectly 
known Prorastomus described by Owen (Quart. Journ. Geol. Soc. vol. xi. 1855, p. 541) 
from the Eocene beds of Jamaica, this character has not yet been acquired. 
The question of the origin of theSirenia is of great interest, and there seems to be a 
considerable amount of evidence in favour of the view first put forward by de Blainville, 
that they are intimately related to the Proboscidea. In the first place, the occurrence 
of the most primitive Sirenians with Avhich we are acquainted in the same region as the 
most generalised Proboscidean Moeritherium is in favour of such a view, and this is 
further supported by the similarity of the brain-structure and, to some extent, of the 
pelvis in the earliest-known members of the two groups (see pp. 202 & 214). Moreover, 
in the anatomy of the soft parts of the recent forms there are a number of remarkable 
points of resemblance. Among these common characters may be noted the possession of ; 
(1) pectoral mammae, (2) abdominal testes, (3) a bifid apex of the heart, (4) bilophodont 
molars with a tendency to the formation of an additional lobe from the posterior part 
of the cingulum. The peculiar mode of displacement of the teeth from behind 
forwards in some members of both groups may perhaps indicate a relationship, 
although in the case of the Sirenia the replacement takes place by means of a 
succession of similar molars, while in the Proboscidea the molars remain the same 
d 
