235 



The presence of seabrosity on the corolla lobes in Puna specimens of this 

 species, make C. platyphylla little remote from Cyanea scabra. 



Cyanca Fcrnaldii, C. plaiyphyUa, C. noli-me-tangpre, Cyaiiea profiisa, C. 

 scabra and C. rollandioidrs are so closely related to each otlier th;it tht-y may be 

 looked upon as one very polymorphic species. Their haliit of growth is, how- 

 ever, quite ditt'ereut, in some instances at least, whicli permit us to recognize 

 them as distinct species One could trace a species througli a nuiuber of others 

 even to species belonging to other sections. What Cyanca plalyphfiUa is on 

 Hawaii or Cyanea scabra, for instance, on Maui, Cyanea angustifulia is on 

 Oahu, with its many relatives on JIaui, Lanai, and its main relatives or, better 

 said, ancestors in Cyanea coriacca, C. spatlndafa, C. Faurici and C. Hardyi on 

 Kauai. They all form a continuous chain, but leave isolated the older species 

 of the section palmaeformes as Cyanea leptostegia, C. arborca, C. Giffarclii, etc., 

 pi'obably the oldest Lobelioifleac of the Hawaiian Islands. 



The Jol)ed forms of C. scabra unite that species with C. hoJophyUa. C. sola- 

 nacea, C. lobata to C. fcro.r, and again from C. fero.r to C. asplenifolia to 

 C. Grimesiana, the true pinnate type' of Cyanca. The varieties of the latter 

 species, especially C. Grimesiana cylindrocalyx, make this circle perfectly com- 

 plete. 



Cyanea Fernaldii Rock in Torrey Bot. CI. Bull. 44:231, pi. 11. l'»17. 



'(P)ate 1:^8.) 



Leaves lanceolate-(ibli)ng, thin in te.xture when dry (chartaceous), dark gi-een 

 above, paler underneath, with prominent midrib, subentii'c with somewhat wavy 

 margin, minutely and closely denticulate, bluntly acuminate at the apex, 34-40 

 cm long, puberulous on both sides, the apparently fleshy petiole 12-19 cm or 

 more in length, near the blade of the leaf showing signs of murication ; peduncle 

 fleshy (apparently), nearly as long as the petioles, 8-10 cm long or a little more, 

 naked its entire length, pubescent with round sears near the ape.x ; bracts folia- 

 ceous, 3.5 cm long, dentate, only present at the apex of peduncle: pedicels short, 

 filiform, 12-15 mm long, pubescent ; calycine lobes ti'iangular, 2.5 mm long, the 

 ovarian portion of calyx 7 mm long, oblong, nearly cylindrical, pubescent; corolla 

 evidently white, pubescent outside, slightly curved, 4 cm long, 6 mm wide, 

 glabroiK inside, the lobes not muricate: staminal column glabrous, as well as 

 anthers, the lower onl.y penicillate; fruit unknown. 



HAWAII: Without definite locality, 1851-1855, J. Remy no. 301, type in 

 the Gray Herbarium, and co-type in Herbarium IMuseum Paris; — Waipio Val- 

 ley, flowering July 1909, Rock no. 4663 in the herbarium of the College of 

 Hawaii. 



The specimen in the Paris Museinn, a photograph of which is here repro- 

 duced, is labeled Eullandia and Dclissea ; the duplicate in the Gray Herbarium 

 is marked RoUandia Hiimboldtiana Gaud. That the species is not a IxoUandia 

 is evident by the free staminal column, while the absence of dorsal knobs would 

 separate it from Dclissea. The species belongs, in fact, to the section Cyaneae 

 ejenumae and comes exceedingly close to Cyanea platyphylla Hillel)r., differing 

 in the long, narrowly oblong leaves and in the exceedingly long peduncles; other- 

 vdse the two species correspond closely. 



The plant is named for Professor Fernald of the Gray Herbarium, to whom 

 the writer is indebted for many favors received and especially for the loan of this 

 particular species. The type is in the Gray Herbarium, Remy no. 301. The 



