343 



DELISSEA* ( iaiulicliaud 



Calyx-tube turbinate or oblou^', its lubes deutifonu or sulnilate; corolla more 

 or less curved, tubiilar, widening from a narrow base, shortly bilabiate, the dorsal 

 slit not extending beyond the middle, the termination of the dorsal and often 

 also of the lateral slits indicated in the bud by a knob or.gibbus; staminal 

 column free from tlie corolla, white, glabrous, only the two lower anthers tufted 

 at the apex ; stigmatic lobes shortly pubescent outside ; berry ovoid, orange or 

 blaek, with narrow epigynous disk, two-celled; seeds dull white, the thin testa 

 finely wrinkled in transverse wavy lines. — Unarmed, glabrous, often subher- 

 baceous shrubs, the stems branching or simple, more or less fleshy ; leaves entire 

 or laciniate, succulent when fresh, flaccid and transparent when dry, bright 

 green, giossj' ; flowers in axillary racemes, with small deciduous bracts and naked 

 pedicels, always white or greenish white with purplish tinge. 



The genus Ddissca* was first erected by Gaudichaud on three lobelioideous 

 plants found in the Hawaiian Islands, D. sithcurdala (type of the genus), D. 

 uiiihilaia and Dclissea acuminata, the last now Cyama acuminata. The genus 

 was founded on quite an artificial character; it was mainly based on the 5-dentate 

 calyx lobes and the two lower bearded anthers. Gaudichaud 's genus Cijaiica. 

 of which only a single species (('. Grimcsiana) was then known, was based on 

 the long foliaeeous calyx lobes of the above-mentioned species, which hajipened 

 to possess anthers which were all bearded. The latter character is, however, 

 not constant, as it appears in a number of species of ('ijaiua which have not 

 foliaeeous calj'x lobes. 



Hillebrand, who had much more material at hand and who was familiar with 

 the plants, having seen them growing in the field, certainly was justified in 

 rearranging the genera; the writer can only uphold Hillebrand 's arrangement, 

 which is an excellent one, as it brings together species which were previously 

 separated in different genera by mere artificial and untenable characters. 



The seven species of Dtlisst a are well defined and can never be mistaken for 

 Cyaneas. Tliey all have the deeply wrinkled, dull white seeds in common, and 

 the dorsal gil)bus of the corolla is always jiresent, characters absent in L'yanca. 



All species of Delis.ica seem to have a tendency to at least sinuate leaves, even- 

 D. suhcordata, whose leaves show signs of becoming sinuate to laciniate near the 

 base of the leaf. 



The fact, liowever, that the genus [)(tiss(a occiu's on nearly all the islands of 

 this group, fi-oni distant Kauai, the oldest island. In Hawaii, the iiresuniably 

 youngest, shows that they are not all one variable species, which may be classed 

 with ('ijnuca. but are distinct enough specifically, possessing a common charac- 

 teristic which separates them from Ciiama, forming thus the genus Ddissca. 

 While some of the specific differences may not be so readily recognized in dried 

 material, no one who has seen them growing in their natural habitat can for a 

 moment doubt their specific distinction. For example, the species Delissen 

 uudiitata and D. suhcordata; Gray unites these two species, not even recogniz- 

 ing the latter as a form of the former. Had he seen the plants growing he would 



* Dedicateil to A. M. Delisse, iiharniacist, naturalist of tlie Freiu-h e.xiiedition, 1800 to 

 1804. 



