332 



LOWER EOCENE FLORAS OF SOUTHEASTERN NORTH AMERICA. 



base, acuininato distud. Texture lij^neous. 

 Surface oruanientpd with close-set, narrow, 

 longitudinal ridges. Lengt ii ranges from 1.6 to 

 2:2r> centimeters, and widtli froju 4.5 to 10.5 

 millimeters. 



Xiueteen so-called species based on Nyssa 

 stones have Ix'en descril)ed from the small 

 pocket of lignite at Brandon, Vt. Among 

 these the present species may be compared 

 with Ni/ssa myJticostata Perkins.' 



Occurrence. — Lagrange formation (in beds of 

 Wilcox age"*, Puryear, Henry County, Tenn. 

 (collected by E. W. Berry). Wilcox group, 41 

 miles southeast of Naborton, De Soto Parish, 

 La. (collected by O. B. Hopkins). 



Collections. — L'. S. National ^Museum. 



Xyssa eolignitica Berry, n. sp. 

 Plate XCIX, figure 8. 



Description. — Stones rather large, elliptical 

 m outline, rounded at both (Uids, slightly 

 compressed, with numerous narrow, close- 

 set, longitudinal ridges. Lengtli about 2.6 

 centimeters. Maximum width about 11 mil- 

 hmeters. 



This species is much larger and more elUpti- 

 cal in outline than Nyssa ivilcoxiana Berry. It 

 resembles the larger forms from Brandon which 

 Perkins - identified as iV;/s.sa Icsciiril Hitchcock, 

 although it is not identical with Hitchcock's 

 original form. It is also much like an unde- 

 scribed form that is abundant in the clays of 

 the Yegua formation and in the Catahoula 

 sandstone of eastern Texas. 



The genus Nyssa is represented in fossil floras 

 from the late Upper Cretaceous onward, the 

 water-side habit of many of the species and the 

 resistant nature of the fruit stones being very 

 favorable to fossilization. Tim genus com- 

 prises five or six species in the existing flora, all 

 of which are confined to southeastern North 

 America except one Asiatic form, wliicli ranges 

 from the eastern Himalayas to the island of 

 Java. 



Occurrence. — Lagrange formation (in beds of 

 Wilcox age), Puryear, Henry County, Tenn. 

 (collected by E. W. Berry). 



Collection. — U. S. National Museum. 



1 Perkins, G. H., Vermont State Geologist Kept. 1903^, p. 197, pi. 79, 

 fig. 103, 1904. 



2 Perkins, G. H., V'-rmont State Geologist Kept. I'.ios-ii, pi. 52, fig. 9, 

 1906. 



Superorder GAMOPETALai. 



Order PKIMTJXALES. 



Family MYRSINACEffi. 



Genus ICACOREA Aublet. 



Icacorea prepanicitlata Berry, n. sp. 

 Plate evil, figure .5. 



Description'. — Leaves lanceolate and many 

 .slightly inetiuilateral in general outline, widest 

 at or above the middle and tapermg to both 

 ends: the base more narrowly extended than 

 the tip, which is commonly lilunt. Length 

 between 8 and 9 centimeters. Maxunum 

 width about 2.75 centuneters. Margins entu'e, 

 slightly undulate by reason of a tendency to 

 be unevenly revolute. Texture coriaceous. 

 Petiole short, stout, and curved, much ex- 

 panded proxiinad, about 1 centimeter in length. 

 Midril) stout, curved, prominent, subparaUel, 

 eight to ten opposite to alternate pairs; they 

 diverge from tlie midrib at angles of 50° to 65°; 

 as a rule the more open angles are m the tip 

 or base of the leaf and the more ascending and 

 less open angled secondaries in the median part of 

 the leaf; all become subparallel with the lateral 

 margms, along which they continue for a con- 

 sidera'ole distance, fonning a series of regularly 

 dimuiishing arches until they finally merge with 

 the Tertiary areolation. Tertiaries pcrcurrent, 

 partly obsolete by inunersion in the leaf sub- 

 stance. 



The species here discussed is exceedingly like 

 the leaves of the existmg Icacorea paniculata 

 Sudworth, a shore shrub or small tree of the 

 Florida Keys, southern Mexico, the Bahamas, 

 and Antilles. 



The genus Icacorea is sometimes extended to 

 include the 200 existing species of the Tropics 

 and subtropics of both hemispheres which by 

 other authors are referred to the genus Ardisia 

 Swartz. Pax, one of the most experienced 

 students of the Myrsinaceie, makes Icacorea a 

 subgenus of Ardisia and restricts it to the 

 American species. Icacorea is iinquestionably 

 entitled to generic rank and should be sepa- 

 rated from Ardisia. Whether the other sub- 

 genera of Ai'disia are entitled to rank as genera 

 I am not prepared to say, although all have 

 received such rank at one time or another. 

 About a dozen fossil species have been referred 

 to Ai'disia, and several of these should probably 



