PARRY CEANOTHUS. 171 



After extended observation and careful study of C. papillosus, where 

 it exhibits its fullest development, in the Santa Cruz Mountains, I am 

 forced to the conclusion that it properly includes the two forms above 

 noted, heretofore regarded as distinct species. When seen in its most 

 luxuriant state on the lower slopes of the Santa Cruz range, it is a 

 densely-branched shrub, 4-6 feet in height, with large oblong leaves 

 (50 x 10 mm.), corrugated on the upper surface, and abundantly gar- 

 nished with projecting papillae, tomentose below, and only slightly rev- 

 olute on the margin. Farther up, the leaves become reduced fully one- 

 eighth the size of the lowland form, and are strongly revolute, almost 

 to the midrib; in this state, the clustered and long pedunculate inflor- 

 escence of a deep blue is attractively conspicuous, suggesting at once 

 the published figure of C. floribundus. Hook., Bot. Mag., pi. 4,806, but 

 still farther, it also forces the conclusion that C. dentatus, Torr. & Gray, 

 is only a form of this species, destitute of papillae. Before coming to 

 this conclusion, I was led to investigate the true nature of these papillae, 

 and find that they are only the usual form of glands so commonly met 

 with on the edges of the leaf, where they can most readily break 

 through, while on the upper surface the thicker epidermis usually covers 

 them as protuberances, though not unfrequently revealing their true 

 nature by pushing through, as regular glands. It thus appears that, 

 though a conspicuous character, it is not necessarily constant, or of 

 specific vah e. As in my view, it is only the absence of the papillae 

 that separates C. dentatus from the reduced forms of C. papillosus. I 

 have re-examined an authentic C. dentatus, corresponding precisely to 

 the figure in Bot. Mex. Bdy., pi. 10, and find that while surface papillae 

 are not to be seen, yet the projecting glands at the margins of the leaf 

 are not directly on the edge, but on the revolute folds of the upper sur- 

 face, so that the essential point of difference is narrowed down to a 

 very small margin. Besides, it is not unlikely that the older herbarium 

 specimens in which C. dentatus is generally represented, may have lost 

 the traces of surface glands by age and pressure, a fact which I have 

 been able partly to verify in an examination of old herbarium specimens 

 of C. papillosus. On these grounds I feel justified in relieving botani- 

 cal observers from the care of a vain search for these doubtful forms. 



19. C. tmpressus, Trelease, 1. c. Coast range, north of Santa Bar- 

 bara. Known only from imperfect material. 



Group V. Integerrimus. 



Branches flexible, or spinose in No. 22; leaves oblong or broadly 

 lanceolate, usually smooth, and inconspicuously veined, entire; inflor- 

 escence thyrsoid, on leafy peduncles. 



