52 PROCEEDINGS OF THE AMERICAN ACADEMY 



genus. B«t the fifth stamen is wanting in M. {Pentacoeliiim) bonti- 

 oides of Japan, and in the allied M. Chinense. The Japanese species 

 sometimes has the cells of the ovary reduced to four (unless there is a 

 slip or misprint in Zuccarini's detailed description) ; while, on the other 

 hand, one of the original species, M. Icetum, Forst.,* has a 3-celled and 

 3-Seeded putamen ; and this not by the abortion of a half-carpel, for 

 the ovary in the flower I examined was found to be trilocular with^a 

 single ovule in each cell. All this militates against De Candolle's 

 primary division of the genera, and against the validity of his genus 

 Polycceliurn (Pentacoelium, Zucc). The alternative evidently now is 

 the establishment of two genera, Polycceliurn and Pentaccelium, or their 

 reduction to mere sections of Myoporum. Tlie latter is, I confidently 

 suppose, the better view, and the species in question may be disposed 

 as follows : — 



§ 1. Pentaccelium (Pentaccelnun, Zucc). Ovarium 4-6- ssepius 

 5-loculare. Stamina 4, didynama. 



Myoporum Chinense. Polycceliurn Chinense, A. DC. Prodr. 11, 

 p. 706. — China. 



Mtoporum bontioides. Pentaccelium hontioides, Zucc. Fam. Nat. 

 Fl. Jap. 2, p. 27, t. 3. Polycceliurn bontioides, A. DC. 1. c. — Japan. 



§ 2. PoLYCCELiUM. {Polycceliurn, A. DC. 1. c. pro parte.) Ovarium 

 5 - 8-loculare. Stamina 5 vel 6, petalis isomera. 



Myoporum Sandwicense {M. tenuifolium, Hook. & Am. Bot. 

 Beech, p. 93, vix Forst. & R. Br. Polycceliurn Sandwiceiise, A. DC. 

 1. c.) : glabrum ; foliis oblongo-lanceolatis acutissime vel tenuiter acu- 

 minatis integerrimis, inferioribusve nunc parce serrulatis ; fasciculis 

 3 - 8-floris ; pedicellis petioluA subsequantibus ; corolla late campa- 

 nulata ad medium usque o-fida; staminibus 5 ; drupa 4- 8-loculari. 

 Ludit floribus 6-meris 6-andris, foliis aut 2 - 3-pollicaribus angustis 

 aut 3 - 5-pollicaribus multo latioribus. — Sandwich Islands, leg. Men- 

 zies, Beechey, Gaudichaud, Douglas, Nuttall {Prinastrum caidijlorum, 

 Nutt. in Herb. Hook.), Remy (nos. 461, 462, 463), &c., on almost all 

 the islands. Variable in the foliage, &c. One form is recorded as " a 

 decumbent shrub " ; another, as " a tree, forty feet high." But no 

 mention is made of its wood, which, according to Hooker and Arnott, 



* " Stamens 5," in the generic character of Hook. f. Fl. N. Zeal., is an evident 

 lapsus or misprint. 



