74 ROBINSON. 



324 (1882); Robinson, Proc. Am. Acad. liv. 358 (1918).— Variable 

 in leaf-form as follows: 



Var. a. crenatum (Hieron.), comb. nov. Leaves elliptical, obtuse, 

 rounded at the base, shallowly crenate, 1.7-4.5 cm. long, 1.2-2.4 cm. 

 wide. — E. jjseudofastigiatum, var. crenata Hieron. in Engl. Bot. Jahrb. 

 xxxvi. 468 (1905). — Cajamarca: near Cutervo, von Jclski, no. 789 

 (Berl., fragm. Gr.). Peru without locality: Mathews (fragm. 

 Gr.). 



Var. /3. pseudofastigiatum (Hieron.) Robinson. Leaves ovate- 

 lanceolate, 3-5 cm. long, 1.6-2.6 cm. wide, acute or acutish at the 

 apex, acute at base, finely serrate-dentate, the teeth subacute. — Proc. 

 Am. Acad. liv. 359 (1918). EJ fastigiatum Benth. PI. Hartw. 135 

 (1844), not HBK. E. loxense Hieron. in Engl. Bot. Jahrb. xxi. 331 

 (1895), not Klatt. E. jjseudofastigiaium Hieron. 1. c. xxxvi. 467 

 (1905). — Peru without locality: Mathews (N. Y.). [Southern 

 Ecuador.] 



Var. 7. crenato-dentatum (Hieron.), comb. nov. Leaves elliptic- 

 lanceolate, acute or acutish at the apex, rounded at base, 5-8 cm. 

 long, crenate-dentate or crenate-serrate, the teeth slightly coarser 

 than in the preceding. — E. pseudofastigiatum, var. crenato-dentata 

 Hieron. 1. c. 468. — Cajamarca: near Cutervo, von Jclski, no. 730 

 (Berl., fragm. Gr.). 



Var. b. lanceolatum (Hieron.), comb. nov. Leaves lance-oblong, 

 9 cm. long, 3.2 cm. wide, acutish at the apex, cuneate at the base, 

 crenate-serrate, the teeth again sparingly mucronulate-serrulate. — 

 E. pseudofastigiatum, var. lanceolatum Hieron. 1. c— Cajamarca: 

 near Tambillo, von Jclski, no. 729 (Berl., fragm. Gr.). 



It is possible that this species, highly variable in its leaf-form, will 

 ultimately be found to merge with E. fastigiatum HBK., but that is 

 said to have heads about 12-flowered and its leaves, although possess- 

 ing closely the pinnate venation and reticulation of E. exserto-venosum, 

 differ in being oblanceolate, bluntly mucronate at the apex and nar- 

 rowly cuneate at base. Much more ample material of these forms is 

 essential before they can be classified with confidence. With the 

 present fragmentary representation they can neither be reduced with 

 certainty nor maintained with much satisfaction. But, as they differ, 

 it seems best to keep them up until evidence of intergradation becomes 

 available. 



59. E. CHiLCA HBK. Glabrous somewhat viscid shrub; branches 

 opposite, erect, or curved-ascending, leafy when young, at maturity 

 apt to be denudated; leaves opposite, rather narrowly lanceolate, 



