255 
SALVINIA ELLIPTICA Newb. 
Pil 206, £. 14,'t4a, 13: 
Leaves small, 1, in. long, 3-3 in. wide, elliptical to al- 
Most orbicular in outline, cordate at base, more or less emargin- 
ate at apex; midrib well defined; surface marked with relatively 
large but few pits, arranged in rows, which radiate from the mid- 
rib at an acute angle upward. 
Formation and locality: upper cretaceous (?) Carbonado, 
Washington. 
This beautiful little Sa/vinia is clearly distinct from any spe- 
cies hitherto described. In outline the larger specimens do not 
differ much from S. Mildeana Goepp.*? (f. 5), and the smaller ones 
‘pproach the forms of S. cordata Ettingsh.*4 (f. 7), S. formosa 
Heer *3 (f. 6) and S. Ekrhardti Probst. *9 (f. 2), but from all of 
these it may be readily distinguished by its few large pits, as com- 
Pared with the numerous smiall ones of the species mentioned. 
It is, perhaps, more nearly like the latter, all things considered. 
S. Spinulosa Probst.* = (f. 4) is different both in size and shape, 
and the others offer no pretence for comparison whatsoever; in 
fact, we may well doubt whether some of these should be re- 
tained in the genus. The peculiar, reticulate nervation and ab- 
sence of pit markings in S. Ad/eni Lesq.*¢ (f. 12), S. reticulata (Et- 
tingsh, in part), Heer *« (f. 13) and S. cyclophylla Lesq.*7 (f. 11), 
at once serve to separate them from all the others, and the latter 
'S €specially open to suspicion on account of its well defined mid- 
rib with irregular, pinnately arranged branches—a structure en- 
rely foreign to the genus. It is significant that S. re tcwlata was- 
°riginally described as a pod of Dalbergia, and S. Alleni as a leaf 
- Ophiog lossum, while S. cyclophylla would not be described to-day 
“$8 Salvinia. All but the latter one of these three are, however, 
SO Very close in form, nervation and apex, to Zimesipteris Bernh., 
“Specially in the case of S. Alleni, that I wonder at their never 
fore having been recorded in this connection. | . 
_, S: oligocenica Staub *™ (f 1) and S. aguensis ** (f. 9) are too 
fragmentary and imperfect for exact comparison, but they oe. 
be More or less closely allied. S. attenuata Lesq.** (f. 10) is 
"nique, on account of its lack of any midrib, and the nervation 
Which extends from base to summit. Were it not for the ap- a 
