•ABT*21 A PLEISTOCENE FLOEA FROM TRINIDAD BERRY 7 



rounded lobes of vaiying sizes. The midrib is stout but not especially 

 prominent. The secondaries are thin, largely immersed in the coria- 

 ceous leaf substance ; they diverge from the midrib at wide angles at 

 irregular intervals, and are abruptly camptodrome at a considerable 

 distance inside the margins. The size varies from narrow elliptical 

 leaves 3.5 centimeters long by 1.7 centimeters in maximum width like 

 that shown in figure 3 on plate 2, to similarly shaped leaves 8 centi- 

 meters long and 5 centimeters in maximum width; from obcordate 

 leaves 2.1 centimeters long and 1.7 centimetei^ wide like that shown 

 in figure 4 on plate 4 to similar leaves 4.5 centimeters long and 3.4 

 centimeters wide like that shown in figure 2 on plate 4. Finally we 

 have the large irregular leaves like those shown on plate 2, figure 1, 

 and plate 4, figure 3, variously lobed and retuse, and without a paral- 

 lel outside the family Sapotaceae in so far as I know. The more 

 regular leaves of this species are distinguished with difficulty from 

 the associated leaves of Rhizophora but as I have remarked under the 

 discussion of the latter they are more coriaceous with more obsolete 

 venation and with different tips. The petiole is shorter and stouter, 

 and appears in the fossil material as more or less ribbed. 



I have been to the pains of examining all of the material of the 

 Sapotaceae preserved in the National Herbarium in my effort to ab- 

 solutely connect the fossil with an existing species. Outside the 

 genus Mhnusops the only species showing variations in form com- 

 parable with the fossil is Sideroxylon elegans DeCandolle of the 

 Guianas, in which the leaves are uniformly smaller and the venation 

 is somewhat more prominent. In the genus Mimusops the two most 

 similar species seen are Mimusops dujolicata Urban, a common Antil- 

 lean forest tree, and Mimusops halata schoniburghii Pierre from the 

 lower Orinoco region. In the latter the leaves average relatively 

 longer and narrower and have longer petioles. In the former ex- 

 actly the same variations iii outline are shown, but such variants are 

 usually smaller than the fossil, although specimens with regular 

 leaves may be larger. On the whole the fossil is closest to Mimusops 

 duplicata and I have described it as a possibly extinct form under 

 the name of preduplicata^ indicating a relationship, which may 

 really amount to specific identity were all the facts known. 



The genus Mimusops is a prolific and common tropical type in 

 both hemispheres, reaching northward to the Florida keys in this 

 hemisphere. In the fossil record it contains 3 lower Eocene species 

 in southeastern North America and a fourth in the Miocene of Haiti. 

 Several European species have been recorded, two coming from the 

 late Eocene of Hesse. 



Cotypes.—C?Lt. Nos. 37024-37030, U. S. N. M. 



