810 DEPARTMENT OF THE INTERIOR 



2 GEORGE V., A. 1912 



to its relation to existing forms, there seems to be little or no room for denying 

 the relation of the present material to the existing American elm, of which it 

 is undoubtedly the ancestral form. The most prominent respect in which it 

 differs appears to be in the rather broad zone of vessels in the spring wood, 

 and the somewhat different form presented by the distribution of the wood 

 parenchyma in the summer wood. Both of these features are of a variable 

 character in the white elm and quite conformable to what is found in the fossil. 

 That both U. americana and U. racemosa should be represented in the same 

 formation by equivalent forms, is in no way surprising when we recall their 

 constant association in the Pleistocene and also in existing floras. There is 

 therefore no reason why the prototypes of these familiar species should not be 

 similarly associated in the early Tertiary. The diagnosis of this species is as 

 follows : — 



TJlmus protoamericana, n. sp. 



Transverse. — Growth rings variable, often very narrow, with no obvious distinc- 

 tion between spring and summer wood except through the location of 

 the large vessels. Wood cells at first rather large and rather thin- 

 walled, soon reduced and passing somewhat gradually into small, thick- 

 walled cells at the outer limits of the growth ring, very variable and 

 unequal throughout, rarely disposed in radial rows, the structure dense. 

 Vessels at first large and prominent, often with round or oval, trans- 

 versely or more generally radially 2-3 seriate: forming a zone \ to \ 

 the thickness of the growth ring and abruptly followed by smaller vessels 

 with wood parenchyma which form tracts of variable extent, radially or 

 transversely extended, or more or less coalescent so as to form diagonal 

 tracts or tangential zones of indefinite extent; the parenchyma elements 

 within such tracts often conspicuously resinous. Medullary rays pro- 

 minent, numerous, upwards of 4 cells wide, sparingly resinous. 



Radial. — Kay cells all of one kind, low and more or less contracted at the ends; 

 the upper and lower walls thin and not pitted; the terminal walls 

 sometimes thick and strongly pitted; the lateral walls without obvious 

 pits. Vessels of the spring wood broad and short, 1£ to 2 times longer 

 than broad, the radial walls with multiseriate, hexagonal pits with 

 large, transversely oblong pores; the smaller vessels fibrous, but with 

 similar construction, the pits often reduced to a single row; thyloses 

 of the large vessels often strongly developed, but more or less strictly 

 localized. 



Tangential. — Rays numerous, medium, upwards of 4 cells wide; the small, rounded- 

 hexagonal cells forming a dense structure. Vessels as in the radial 

 section. 



250 , 



— of 1903. Ulmus columbiana, n. sp. 



c 



Plate vrn. 



Among the woods represented in the collections of 1903, was a specimen 



believed to be a new species of Rhamnacinium, and provisionally referred to that 



250 



genus under the number . A more critical examination proves it to be an 



c 



elm of a type not readily assignable to any known species. Its diagnosis is as 



follows : — 



