PHANEROGAMS— PROFESSOR BAYLEY BALFOUR. 95 



Wight (in an elaborate essay Illustr. ii. 2), Payer (loc. tit.), and Boissier (Flor. 

 Orient, ii. 736). 



Bentham and Hooker (Gen. PI. i. 784) place it as an anomalous form in 

 Lythrarieae, and, so far as I know, they are the first authors who have done so. 



In this account of the Botany of Socotra I have followed Bentham and 

 Hooker, as a careful study of the whole question, in the light of our new 

 Socotran plant, convinces me that their judgment is the correct one, and that 

 Punica has its most natural position in Lythrarieae. Those who would have 

 the genus as a distinct family find but little support in morphology. The 

 fruit-character, upon which stress was laid, is now proved by this Socotran plant 

 to be what Lindley long ago pointed out was the case, and Berg, Eichler, and 

 others more recently have maintained, merely a special development of a con- 

 dition comparable with that in the several families with which the genus has 

 undoubted alliance. 



With Myrtaceae there is undoubtedly a very near affinity. The general facies 

 of the plant encouraged its union, and Lindley placed it in the vicinity of 

 Sonneratia, which he also included in the order. Of technical characters the 

 inferior ovary is a strong myrtaceous feature ; but this character is disposed of by 

 the discovery of our plant, and the differences separating Punica from Myrtaceae 

 are several and important, viz. : — the valvate calyx, plicate petals, nonstamini- 

 ferous disk, ovary not always inferior, pulpy seeds, and convolute cotyledons. 



The character of the ovules is one in which the genus approaches 

 Melastomaceae in the tribe Memecylece, but the calyx and stamens are quite 

 diagnostic. 



With Lythrarieae, in which it is here placed, it has a vast preponderance of 

 features in common. The sepaline, petaline, and staminal characters which 

 exclude it from Myrtaceae and Melastomaceae are just those of Lythrarieae, 

 the superior or half superior ovary of our Socotran plant is a thoroughly 

 Lythrarioid character, and breaks down the chief objection urged by Eichler 

 to Bentham and Hooker's allocation. In the pulpy seeds and convolute 

 cotyledons it is still an exceptional type in Lythrarieae, but Sonneratia, which 

 is included by Bentham and Hooker in this order, shares with it the 

 cotyledonary characters. So that the only aberrant condition of this genus, 

 when placed in Lythrarieae, is the pulpy seed coat, and I therefore consider 

 Bentham and Hooker's recognition of its affinity as the correct one. 



The discovery of our plant necessitates a recasting to some extent of the 

 generic character, and this I now give :— 



Calycis persistentis crasse coriacei tubus turbinatus, angulatus, ampliatus, liberus v. ovario 

 adnatus ; lobi 5-7. Petala 5-7, calycis fauci inserta, lanceolata v. obovata, corrugata. 

 Stamiua perplurima calycis fauci multiseriatim inserta, filamentis filiformibus incurvis ; 

 antherae versatiles, ovatae v. elliptic*. Ovarium liberuin sessile v. semi-inferum, v. inferum, 

 multiloculare, loculis 5-7-seriatis fuudo ovulifero v. plurmis 2- rarissime 3-seriatim 



