OF WESTERN SOUTH AMERICA. 153 



di&tiuguislied from AntirrTiinitm by the form of the corolla. 

 Several years later, with the ample additional materials supplied 

 by American explorers, Asa Gray, in one of those remarkable 

 papers which he has published ia the ' Proceedings of the 

 American Academy of Sciences/ reduced the scattered fragments 

 to order, and arranged the North-American species in sections of 

 the genus AntirrMnum^ including in the section Gamlelia the 

 Manrandia juncea^ Benth. In his final revision of the Order for 

 tlie ' Genera Plantarum,' Bentham adopted the views of Asa 

 Gray as to the limits of AntirrJiiiittm and the groups under 

 wiiich tlie American species should be classed ; but he maintained 

 the genus Galvesia for tlie Peruvian species. On a scrutiny of 

 the generic characters assigned to this and to AntirrJiinum,! can 

 find only three points in which they appear to differ. The style 

 in Galvesia is said to be minutely two-lobed. In several Antir- 

 rJiina the tip of the style is slightly thickened, as in Galvesia, 

 and the very minute cleft in the latter is all the difference that 

 can be pointed out. The cells of the anther in Galvesia are said 

 to be confluent at the summit, while in Antirrhinum they are 

 separate. Althougli I liave been unable to verify this character, 

 I place absolute reliance on tbe unfailing accuracy of Mr. Ben- 

 tham ; but I do not know whether the anthers of the sections 

 Maurandclla and Gamlelia of Ant ir7'hinu7n have heen examined, 

 or whether the assigned difference has been found to exist. 

 Finally, the cells of the capsule in Galvesia are said to open t>y 

 an irregular pore below the summit. I am not sure whether 

 Bentham considered the ^\ord. irregidar to indicate any important 

 diff^erence between Galvesia and the AmQric^n Antirrhina, which 

 have the capsule nearly or quite ovoid and the cells oi)eningby a 

 single pore to each cell. But in any case the diff^erence is much 

 less than that existing between the latter and the European 

 species, in which the capsule is very oblique, one cell opening by 

 a single pore, the other by two pores with toothed edges. I may 

 mention that in Galvesia the corolla is more gibbous at the base, 

 and therefore nearer to the typical Antirrhinum than it is in A. 

 jnnceiun, I must not omit to notice the fact that, from the figure 

 and description of a plant brought from Cerros Island and 

 described by Kellogg in the * Proceedings of the Californian 

 Academy,' vol. ii. p. 17, as Saccularia Veatchii, Bentham conjec- 

 turally added a second Californian species to the genus Galvesia^ 

 with which Saccularia appeared to him to be identical. 



