32 BULLETIN 103, UNITED STATES NATIONAL MUSEUM. 



radial view) ; toward the median region the rays are three or four 

 cells broad (the radially elongated cells in the radial view). 



Felix states that in the Antigua material the rays were usually 

 biseriate, while uniseriate and triseriate rays were rare. I do not 

 know the extent of his material, but in the case of that from Panama 

 I had but few radial sections cut. Ray cells frequently filled to a 

 greater or less degree with gum. 



Remarks. — Fragments of the wood of this species are very com- 

 mon in the collections from Panama, but a good deal was rather badly 

 decayed before petrification. That which has formed the chief basis 

 for the foregoing description and all of the photographs and draw- 

 ings is beautifully preserved. The species is clearly identical with 

 the type, as very insufficiently described and illustrated by Felix. 

 One highly ferruginized and fairly well preserved quadrant of a 

 trunk indicates a large tree, with a diameter of at least 25 cm. 



The genus Taenioxylon was established by Felix in 1882 with 

 T. varians from Antigua as the type. He has since described 7 addi- 

 tional species including 2 additional from Antigua, 1 from southern 

 Brazil, 1 from East Indies, 1 from Philippines, 1 from Caucasus, 

 and 1 from the Swabian Alps. All are of Tertiary age and show 

 resemblances to various members of the 3 Leguminous families, 

 Caesalpiniaceae, Miomosaceae, and Papilionaceae. Felix considers 

 the present species to be a member of the Papilionaceae, and it agrees 

 entirely with Solereders account of the anatoiny of this family. The 

 two kinds of ray cells described have, according to Saupe, been 

 shown to occur in the following tribes in this family, namely the 

 Podalyrieae, Genisteae, Galegeae, Hedysareae, and Sophoreae. With- 

 out much recent comparative material, which is unavailable, it is 

 impossible to allocate the present species more definitely within this 

 extensive family. 



Occurrences. — Bohio formation, middle Bohio Ridge (poorly pre- 

 served) quadrant of a large trunk indicating a tree with trunk at 

 least 25 cm. in diameter. Cucuracha formation, upper part. Green 

 clays of Gaillard Cut (locality 6845) Oligocene limestone. Orbi- 

 toidal limestone, 2 miles north of David (locality 6523) (all above 

 collected by D. F. MacDonald). Culebra formation, upper part. 

 Near top of big slide, just north of Culebra. Collected by M. I. 

 Goldman (figured material). 



Collections. — U. S. National Museum, Johns Hopkins University. 



Family MIMOSACEAE. 

 Genus INGA Willdenow. 



INGA OLIGOCAENICA, new species. 



Plate 16, fig. 2. 

 Deselection. — Leaflets rather above medium size, elliptical-ovate 

 and very inequilateral in general outline. Apex abruptly acute, not 



