PHANEROGAMS — PROFESSOR BAYLEY BALFOUR. 93 



2. A. multiflora, Roxb. Flor. Ind. i. 426; DC. Prod. iii. 79; Boiss. Flor. 

 Orient, ii. 743 ; Clarke in Hook. Flor. Brit. Ind. ii. 570. 



A. senegalensis, Lamk., var. multiflora, Hiern in Oliv. Flor. Trop. Afr. ii. 477. 



Soeotra. Near Galonsir. B.C.S. n. 714. 



Distrib. India, Afghanistan, Persia, and tropical Africa ? 



I am not certain that the tropical African plants are this species. 



2. LYTHRUM. 

 Lythrum, Linn. Gen. n. 604 ; Bentb. et Hook. Gen. PL i. 779. 

 A cosmopolitan genus. 



L. hyssopifolium, Linn. Sp. 642 ; DC. Prod. iii. 81 ; Boiss. Flor. Orient, 

 ii. 739 ; Jacq. Flor. Austr. ii. 20, t. 133. 



Soeotra. On the plains. B.C.S. n. 699. 



Distrib. Widely spread over the globe. Absent from India. 



3. PUNICA. 



Punica, Linn. Gen. n. 618 ; Benth. et Hook. Gen. PL i. 784. 



The genus Punica, presenting as it does a most peculiar feature in fruit 

 structure, has been the subject of much investigation, and its position and 

 affinities have been frequently debated. The plant from Soeotra which we 

 refer to this genus differs in a very remarkable way from the type of the 

 genus in the fruit-character which has been its most distinguishing feature 

 hitherto, and thus necessitates a modification in the generic description 

 which I shall now notice. I shall, however, only say so much here regarding 

 the structure of the fruit as may be required to explain the difference in the 

 types, and as will suffice for a basis of a few remarks on the systematic 

 position of the genus. In the Appendix will be found a fuller morphological 

 account of the genus. 



The true structure of the fruit was first satisfactorily indicated by Lindley 

 (Introd. Syst. Nat. (1830, 64), and his account has been the basis of all 

 subsequent explanations, the subsequent detailed work of Payer and Berg 

 confirming the essentials of his description. The structure briefly is, — there 

 are two rows of carpels, an upper, comprising five or more, with parietal 

 placentas, and a lower of three or four, with central placentation. There 

 appear, therefore, in the fruit to be two tiers of carpels superposed, each 

 containing many seeds arranged on placentas of different position in the tiers. 

 Development shows that these two tiers of carpels are primarily concentric 

 whorls, the upper being at first outside the other ; but in the evolution of the 

 carpels the external whorl is carried upwards, and eventually lies above the 

 other. An interesting further modification is described by Payer (Organogenies 



