640 PROCEEDINGS OF THE NATIONAL MUSEUM. vol.38. 



material, which is very poor, is obviously not related to this genus, 

 while the record of Sapindopsis elliptica from Fredericksburg by the 

 same author is based upon the remains of Rogersia longifolia Fon- 

 taine. Likewise the record of Sapindopsis obtusifolia from Deep 

 Bottom, Virginia, is based upon fragments of uncertain generic affinity, 

 but doubtfully related to Sapindopsis. The genus is notably absent 

 from the f ossiferous Patapsco beds at Federal Hill, Maryland, but 

 when present at a locality it usually occurs in the greatest abund- 

 ance, as at Fort Foote, Maryland, or in the vicinity of Brooke and 

 Aquia Creek, Virginia, where hundreds of specimens often of great 

 perfection have been collected. 



In modification of the diagnosis quoted above it may be said that 

 the vast majority of the leaves are abruptly pinnate and the terminal 

 leaflets, while usually confluent and decurrcnt, are sometimes petio- 

 late, Sapindopsis magnifolia in particular furnishing many individuals 

 lacking the decurrently winged rachis. 



The most closely related plants to Sapindopsis in the modern flora 

 are the various genera of American Tropical Sapindacese. Among 

 these the genus Matayba Aublet approaches very near to the Creta- 

 ceous form. Matayba embraces species with both opposite and alter- 

 nate leaflets having either entire or dentate margins. The rachis 

 lacks definite alee, but it is somewhat flattened with a vestigal wing on 

 each side in the form of a raised line which is wider at the point of 

 origin of the leaflets and decurrent to the next lower leaflets. So many 

 other genera of the Sapiiidaeeae have markedly alate rachises that the 

 presumption is strong that this genus or its ancestors were at some 

 time similarly provided. The most similar species seems to be 

 Matayba apetala (Macfarland ) Radlkofer, in which the leaflets are usually 

 more numerous than in Sapindopsis; although some specimens show 

 but three pairs, those with four pairs are common; they are some- 

 times subopposite, and as many as seven pairs are met with. The 

 venation is exactly like that of the fossils, as is the texture of the 

 leaves and their limits of variation. Another closely related species 

 is Matayba domingensis Radlkofer, also a native of the West Indies. 



It is exceedingly satisfactory to be able to establish upon a some- 

 what firmer basis Professor Fontaine's choice of the term Sapindopsis 

 for these Potomac plants. The}^ are so abundant in their occurrence, 

 so striking in appearance, their strict habit and glossy texture giving 

 them every appearance of some fern-like plant, as, for example, the 

 common Acrostichum aureum of the Tropics, that their original 

 describer deserves great credit for having correctly determined their 

 modern affinities. 



