THE STUDY OF THE EVOLUTION" OE THE FLO WEE. 275 



of an outer whorl of five leaves and an inner of three single and one branched stamens 

 {st), so that it simulated a pentamerous jflower. One of the stamens {st) was situated 

 opposite the bract. But the lateral flowers (fig. 25) of both specimens resembled those of 

 G. elliptlca — the stamens were alternate with the perianth segments. Bentham seems to 

 have fallen into error through examining the terminal flowers of the inflorescence. 



The male inflorescence of G. Thuretii, a hybrid between G. elliptica (male) and 

 G. Fadyenii (female), called in error G. Macfay diana by Carriere, discovered by 

 M. Gustavo Thuret at Antibes about 1862, combines the characters of both parents 

 This point was overlooked by Carriere * in his descriptions published in the ' E-evue 

 Horticole,' of 1869 and 1879, and subsequently by Baillonf, who states that the male 

 flowers resemble those of G. elliptica. The material upon which the following observa- 

 tions are founded w^as gathered at Kew from plants obtained from France about 

 the year 1896. The "flowers" occur singly in the axil of each bract, agreeing thereby 

 with the female parent, but each stalk bears in reality three, more or less congenitally- 

 united flow^er-heads, so that there is a marked correspondence with the male parent. 

 There were from four to six perianth leaves in the specimens examined, and in some 

 cases two or three were congenitally united to form a lobed structure. The stamens 

 varied in number from four to twelve : the outer were usually of normal size, the 

 inner smaller and often rudimentary. 



The female flower consists of a bicarpellary ovary terminated by two, almost free, 



■ 



styles and surmounted in G. elliptica by a pair of minute folioles, which are situated 

 almost at right angles to the styles. Lindley J regarded the folioles as a calyx and they 

 are described as such by Harms § in 'Die Pflanzenfamilien.' Baillon H, on the other 

 hand, points out that the terminal flowers of the inflorescence often bear on the sides of 

 the ovary near mid-height two bracts which alternate with the two preceding normally 

 orientated ones ; for this reason he considers that the supposed sepals are nothing more 

 than displaced bracts. Wangerin^ investigated the matter in greater detail, and his 

 original figures show a complete series in the displacement of bracts, thus: — The folioles 

 are absent in G. gracilis (Wang., fig. 5, D) ; they are situated at about one-third the height 

 of the ovary in G. Wrightii (Wang., fig. 4, D) ; they are depicted higher up still in 

 G. launfoUa (Wang., fig. 2, D), and terminal in G. Fremontii (Wang., fig. 2, C). Having 



r 



examined herbarium specimens of several species of Garry a, notably G. VeilcUi and 

 G. laurifolia^ the writer is able to confirm Wangerin's observations. 



The development of the female flower of the hybrid, G. Thuretii, was described by 

 Baillon ** as follows: — " Sa fleur femelle est-elle representee au debut par un mamelon 

 plein qui occupe I'aiselle d'une bracte'e de Tin florescence. Bienl6t sur ce mamelon se 

 montre, a droite et a gauche, une feuille carpellaire. Toutes deux s'elevent en devenant 

 connees et constituent la cavite de I'ovaire, que leurs sommets couronnent ensuite en 



* E. A. Carriere in Eevue Hort. 1869, 17, and 1879, 154. 



t H. Baillon in Eull. Soc. Linn. Paris, i. (1877) 139. 



t i. Lindley, Veg. Xing. 2ud ed. 295 (1847). 



§ Harms in Engl. u. Prantl, Nat. Pfl., Teil iii. Aht. 8, 256. I! H. Baillon, I.e. 



f VV. \Yangerin in Das Pflauzeur. H(ift 41 (1910), Garryaceae, 1-17. ** H. Baillon, I c. 140. 



