340 PROCEEDINGS OF THE AMERICAN ACADEMY 



The anomaly of five nucules to the fruit, which suggested Hooker's 

 specific as well as De Candolle's generic name, is so far from constant, 

 that I could not detect a single instance in a long suite of specimens, 

 although Dr. Pickering appears from his notes to have been more 

 lucky. I have found six nucules ; but a supernumerary carpel is not so 

 extraordinary ; and this suggests that the five nucules, when this num- 

 ber occurs, result from the abortion of one cell or half-carpel. In all 

 essential respects this species is a Heliotrope, in which genus even the 

 inequality of the sepals is not altogether unexampled. 



Without hesitation, we may reduce to the genus Heliotropium Nut- 

 tail's Euploca (as I had already indicated), Endlicher's Schleidenia 

 (Preslea, Mart.), and De Candolle's Pentacari/a, and (with Fresenius) 

 associate Tournefortia with it rather than with Ehretia. But the plant 

 which (in Mem. Amer. Acad. n. ser. 6. p. 403) I had inadvertently 

 referred to Heliotropium, viz. H. Japonicum, is only a variety (with 

 broader leaves and longer style) of Ammam's Arguzia ( Tournefortia 

 Arguzia, DC), a connecting link between Heliotropium § Heliophytum 

 and Tournefortia. 



Considerations analogous to those which forbid the dismemberment 

 of Heliotropium, point, though perhaps less directly, to the reunion 

 under Coldenia of several plants which have been distinguished as 

 genera.* 



* If we retain under Coldenia both Stegnocarpus, DC. and Tiquilia, Pers. (which 

 is Galapagoa, Hook. f". ), and add Eddya, Torr. (and even Ptdocalyx, Torr.), a well- 

 marked genus, of uniform floral characters and not incongruous in habit, will be the 

 result. Otherwise we shall have four or five genera for barely twice as many spe- 

 cies. The genus may be thus disposed in sections : — 



COLDENIA, Linn. 



I. Fructus e nuculis 4 trigonis dorso convexis inms faciebus planis arete con- 

 jnnctis, crassis, crustaceis. 



1. EccoLDENiA, DC. Styli 2, breves. Fructus globoso-quadrilobus : nucnlaa 



geminatim subconnatae, demum partibiles. C. procumbens, Linn. 



2. Stegnocarpcs, DC, Torr. Stylus bifidus. Fructus globosus in nuculas 4 



secedens. C canescens, DC. Stegnocarpus canescens, Torr. in Pacif. R. R. 

 Rep. 2, p. 169, t. 7. 



II. Fructus alte quadrilobus, e nuculis 4 (vel abortu paucioribus) ovatis parvis 

 angulo interno basi styli mediante tantum connexis, pericarpio tenui. 



3. Eddta, Torr. Stylus superne bifidus. Nuculee tenuiter crustaceae, papilloso- 



scabrae. C. hispidissima, Torr. 1. c, t. 9. 



