28 ANNALS NEW YORK ACADEMY OF SCIENCES 



posed. Why they disappeared in this region, while the rival groiip of 

 Manatees survived is an interesting question; but the evidence as to the 

 distribution and range of the Sirenia during the Tertiary is so scanty and 

 incomplete that any further speculations are scarcely worth while. 



MOLAR-PREMOLAE For.MULA IIS^ SiRENIANS 



The molar-premolar formula in the Sirenia is difficult to state cor- 

 rectly, 2)artly because of certain i^eculiarities in the premolar replacement, 

 I)artly the doubtful interpretation of alveoli where the teeth themselves 

 are not kno^vn. So far as the Manatees are concerned, I have accepted 

 the interpretation placed by Thomas and Lydekker upo]i the cheek teeth, 

 involving an actual increase in the number from the primitive formula 

 of four premolars and three molars which pretty certainly characterized 

 the ancestors of all placental mammals. This increase in number of 

 molars would appear to be attained by extension of the dental lamina 

 posteriorly and budding from the tooth germ of the third molar, thus 

 continuing the process by which the third is derived from the second and 

 the second from the first. That such an increase, whether by this or other 

 means, does occur normally in the number of true molars in certain other 

 placental ph3da, appears beyond question. Otocyon, Gentetes, Myrme- 

 cohius and various Cetaceans may serve as illustrations. The abnormal 

 occurrence of an extra molar or j^remolar in the series is not a rare occur- 

 rence among other placental mammals; this is usually ascribed to re- 

 duplication. 



It is by no means clear that there is any such increase in the number 

 of either premolars or true molars in any of the other Sirenians living or 

 extinct. Andrews ascribes four lower molars and four premolars to 



'-* 1 4 4 



Eosiren, and Lepsius gives the formulas as : i^r c-— p-— m-— in Pro- 

 X & 3 14 4 



.1 1 3 4 . ,. ,.,. . .201-24 



rastomus; i-—r c— - p— r m -.- ii^ rlalitliermm; i— - c—- Ptttt m~r 

 (3) I 4 4 3 1 2-3 4 



in Ilalicore. This would seem to indicate four true molars as the normal 

 number in this family. Abel, hoAvever, has shown (Abel, 1906) that the 

 fourth milk molar in the Halicoridse is retained exceptionally late in life, 

 and sometimes intercalated between the last successional tooth and the 

 first true molar. He accounts in this way for the apparent series of eight 

 postcanine teeth in the lower jaw of Eosiren, IlaUtherium and the latei 

 Ilalicorida^ without finding it necessary to suppose the addition of a 

 molar from behind to the usual placental series. Possibly the eight post- 

 canine teeth of Promstomus are to be explained in this way; but the inter- 



