320 PROCEEDINGS OF THE AMERICAN ACADEMY 



one lobe wholly exterior in the bud, while an unpublished Tahitian 

 species exhibits, in the few flower-buds we have for examination, a 

 regular quincuncial imbrication. That third section may, however, 

 be confirmed by admitting the nature of the placentae to a leading 

 place in the character. 



The ovules of Geniostoma are amphltropous, as suggested by End- 

 licher, rather than anati'opous, as stated by Blume ; but the seeds 

 become almost anatropous. The minute embryo assigned by Blume, 

 and figured in his G. iasiostemon, is not borne out by G. ligustrifoliiim 

 and G. ruprestre, both of which have the cylindrical embryo nearly 

 as long as the albumen, as .remarked by Alphonse De Caudolle in the 

 Mauritian G. ovation.^ 



The corolla, as remarked by Bentham, is not at all funnel-form, but 

 almost rotate, — in fact, between rotate and campnnulate. The phrase 

 " corolla mhinfundibidiformis " found its way into the generic char- 

 acter from Forster's figure of the flower of G. rupestre, which repre- 

 sents the tube of the corolla much too long. It is really so short that 

 Sprengel, describing an original specimen, calls the corolla " suhrosacea 

 pentapetaloidea" The fine pubescence on the lobes is variable in the 

 same species, or even wanting. The villosity of the throat also vai'ies 

 in amount in different forms of G. rupestre, in some cases being re- 

 duced to a small tuft at the insertion of the short filaments, or merely 

 upon them ; in one of our species it is wholly wanting. 



The stigma, I believe, is capitate throughout the genus, or at first 

 depressed-globose, after anthesis sometimes becoming obovate. Here, 

 again, Forster's " stigma cylindricum, " — utterly at variance with his 

 figure, and that, in turn, doubtless incorrect as to the strong lobulation 

 — has given rise, on the one hand, to De CandoUe's " stigma crassius- 

 culum" on the other, to Endlicher's " stigma sidcatuyn." Sprengel's 

 " stigma capitatum, pid)escens, sid)lamellosum" is better, and would be 

 better still without the last word. "While occupied Avith the pistil of 

 G. rupestre, I may remark that the difference between Forster's char- 

 acter '^^ stylus jiliformis tuho longior" and his figure (d), which exhib- 

 its a very short style (such as Sprengel terms "brevissimus"), is in 

 fact exemplified in what I regard as different forms of this species ; — 

 the style being often twice or thrice the length of the ovary and the 

 stigma, and sometimes reduced in length even to a minimum. As to 

 Sprengel's " stylus hasi villosus" that was probably suggested by Fors- 



