286 BERRY: MESOZOIC FLORA OF ATLANTIC COASTAL PLAIN 
in the Coffee sand and in the McNairy sand is a sure indication 
of conifers, and the petrified wood found at Coffee Bluff is also 
coniferous and represents an undetermined species of Cupres- 
sinoxylon. Araucaria bladenensis, which is so common in the 
Black Creek and Lower Eutaw and ranges upward into the 
Cusseta sand, has not been discovered in Tennessee. 
If the two Tennessee horizons be considered separately it will 
be noted that the Coffee sand flora as at present known consists 
of fourteen named species and an unidentified Cupressinoxylon. 
None of these are new. It is contrasted with the McNairy sand 
flora by having but one common species (Manthotites georgiana) 
and by a considerable number of forms that come up from older 
horizons. Thus it has four species that originate in the Raritan, 
although none of these are characteristic of the Raritan but 
of slightly younger horizons. It has three species common to the 
Woodbine sand, five to the Dakota sandstone, seven to the 
Magothy formation, and eight to the Tuscaloosa formation. 
Nine of the Coffee sand forms are common to the basal Eutaw 
and twelve of the fourteen species are found in the Black Creek 
formation of the Carolinas. The Coffee sand has not yielded an 
invertebrate fauna, although on stratigraphic grounds Stephenson 
refers it to the Exogyra ponderosa zone. The present collections 
unquestionably confirm this reference. Its exact position in 
this zone is not directly determinable since the possible equivalents 
of the Coffee sand in Mississippi and Alabama are strictly marine 
formations without fossil plants, so that we have the familiar but 
much involved problem of comparing a section in one area with 
plants and without invertebrates, with a section in an adjoining 
area containing invertebrates but no plants. While the intef- 
pretations resulting from the two classes of evidence are in sub- 
stantial agreement the paleobotanical evidence would seem to 
indicate that the Coffee sand may be slightly older than Stephen- 
son (0p. cit.) postulates, unless we are to assume that 86 per cent. 
of the Coffee sand flora comes up from older horizons and then 
becomes extinct in the relatively short interval between the 
Coffee and the McNairy sand. This is of course possible and 
future work may demonstrate its truth, but in the present state 
of our knowledge it seems improbable. Two of the Coffee sand 
