APPENDIX. 43 



families^ a very manifest alliance^ exhibiting tlic prominent cha- 

 racteristics of monopetalous flowers, with a pistillum consisting 

 of a superior ovarium, a simple style, and a stigma generally 

 entire or 2-lobcd; the ovarium by the confluence of its carpels 

 heing normally 2-locular, with the cells placed always anteriorly 

 and posteriorly in regard to the axis ; and whenever the few 

 known exceptions occur of more than two cells, these will be 

 found to arise generally from an unusual extension of the pla- 

 centae, which always proceeding from the centre of the dissepi- 

 ment, produce abnormally other spurious cells. Among the 

 Nolanace^y the only genus that approaches this definition is 

 Grahowskytty all others differing essentially in structure, but 

 even in that genus the resemblance is more apparent than real. 

 In the Nolanace(By the carpels constituting the pistillum are more 

 numerous, and, excepting the case just mentioned, are always 

 free, springing from a fleshy receptacle surrounded by distinct 

 glands, and all receiving their fertilizing influence through the 

 medium of one common style, that has no direct communication 

 with the ovaria, but always indirectly, through the intermedium 

 of a supporting gynobase, in the same manner as the Borragi- 

 nace<2 and the Labiatce. In Grabowskya, although the two com- 

 ponent carpels are connate, the style apparently issuing from it 

 does not spring from the summit, as in an ordinary pistillum, 

 but may be traced down the axile line of their union to the base, 

 and may be separated from the adherent carpels : the style there- 

 fore, though concealed below by the confluent carpels, is truly 

 of gynobasic insertion. This essential difference in the struc- 

 tm^e of the pistil renders the union of the Solanacea and Nola- 

 nace<e quite indefensible. Another peculiarity is constant in the 

 Nolanace(By in which respect we find no parallel or analogy 

 among the Solajiacece : this consists in the fact, that whether 

 the ripened nuts be unilocular or many-celled (the cells in all 

 cases being 1-seeded), there is always seen at their basal point 

 of attachment, one or more scars, each closing the entrance into 

 a corresponding cell, w^hich scar, in every case, represents the 

 end of a kind of plug, evidently analogous to the strophiole (or 

 Calomphala of Schrader), so conspicuous in the nuts of the Bor- 

 raginacece. Another distinction will also be found to exist which 

 has not been noticed by M. Dunal : in the Solanacete the extre- 

 mity of the radicle never points immediately to the hilum, but is 

 directed to a spot removed from it, and even where the embryo 

 is straight, as in Mettemichia^ Cestntm, Fahiaria, &c., although 

 the end of the radicle points to the bottom of the seed, the 

 hilum is always lateral or marginal, at some little distance from 

 the base : in Nolanaceaj on the contrary, the extremity of the 

 radicle always points to the strophiolar cavity in the base of the 



G 2 



