19U3.J NATURAL SCIENCES OF PHILADELPHIA. 685 



mainly east of the pine barrens, but quite across the Cape IMay 

 peninsula. 



Habitat. — Low sandy ground or sandy woodland. 



Description. — Flowering plants, Tinicum, Delaware county, Pennsyl- 

 vania, April 25, 1903. No. 5,141, Herb., W. S. Flowers 15 mm. 

 wide, blue, close to the color of the double cultivated species (between 

 "campanula blue" and "mauve" of Ridgway), base white, lower 

 and laterals with dark lines, beard spreading from the basal spots on 

 to the blue area, lower petal and upper ones as well pubescent with 

 scattered hairs; petals all rather broad and rounded, lower one almost 

 cup-shaped and truncate, all distinctly emarginate; sepals lanceolate- 

 acute; scapes 60-80 mm. Leaves thick, fleshy and glabrous, nearly 

 triangular, crenate, with basal teeth slightly more prominent, decurrent, 

 15 X 25 mm. ; petiole glabrous, 40 mm. Later specimens, May 9, 

 have erect cleistogenes on peduncles 75 mm. high. 



Fruiting plants, Tinicum, June 21, 1903. No. 5,143, Herb., W. S. 

 Leaves 50 x 5 mm., strictly triangular, margin dentate-crenate, coarser 

 toward the base, petioles 200 mm., peduncles of cleistogenes 120 mm. 

 high. 



This seems to be the typical V. cmarginata. Other colonies in 

 Cape May county, New Jersey, have the summer leaves very broad 

 across the base, with the teeth enlarged into deep lobes (PI. XXXVI). 

 Still others in Chester county, Pennsylvania, in woodland, develop 

 leaves 90x90 mm., much expanded and lobed at base; petioles 300 

 mrn. long! I have not had the opportunity of studying either of these 

 last in flower, so cannot state whether their blossoms conform in 

 color to the Tinicum plants. The leaves of the Chester county 

 specimens, while still thick and fleshy, are slightly but minutely 

 pubescent. 



The above descriptions will give some idea of the immense varia- 

 bility of the sagittata group, comprising numbers 22-25. V. fimbri- 

 atula seems to be the nearest to the supposed parent or papilionacea 

 type. From it we pass to sagittata, of which emarginata seems to be an 

 extreme development, with apparently a marked deviation in floral 

 characters. The resemblance in leaf-form between e7narginata and 

 some aberrant hrittoniayia has already been commented upon. Whether 

 the apparent intergrades which serve to confuse the sagittata group 

 are really such, or the extremes of a somewhat variable species, is a 

 difficult matter to determine. 



