70 ME. BENTHAM ON LOGANIACE^I. 



Anthocleista ; valvate in Stryclinos, a genus spread over the whole 

 of the tropics, Brehmia, peculiar to Africa, and possibly also in 

 Labordea from the Sandwich Islands, and imbricate in the Masca- 

 rene JVicodemia. 



15. Deseontaikea, Buiz et Pav. 

 This plant, for the genus consists but of a single species, al- 

 though repeatedly described and figured, has given rise to much 

 difference of opinion as to its affinities, and even to considerable 

 uncertainty as to the real structure of its ovary. Referred by 

 some from its foliage to Ilicinece or Theophrastece, by others from 

 its fruit to Solanece, from its bitter principle to Gentianece, from 

 some supposed affinities quite unintelligible to me, to the vicinity 

 of Diapensia and Galax, it is only recently that its real place 

 among Loganiacece has been pointed out in the ' Gardener's Chro- 

 nicle,' although even there an affinity is also suggested with 

 Legnotidece, whose connexion with Loganiacece I have already said 

 I am unable to comprehend. I have now been enabled to clear 

 up all doubts as to the structure of the flower • the rich materials 

 at Kew, where the Hookerian herbarium alone contains specimens 

 from seventeen different collections, have given me the means of 

 examining several ovaries taken from the most different-looking 

 forms, and all confirm the association of Desfontainea with Lo- 

 ganiacece, showing the closest affinity with Fagrcea, and like 

 Fagrcea forming a connecting link between Loganiacece and Gen- 

 tianece. Indeed, besides the dentate leaves and the colour of the 

 flowers, almost the sole generic distinction between Desfontainea 

 and Fagrcea consists in the number of cells of the ovary, which in 

 the former are usually five, although sometimes reduced to four 

 or even three, whilst in Fagrcea they are always two only. The 

 raised line connecting the petioles, the five-leaved calyx, the aesti- 

 vation of the corolla, the form and position of the stamens as well 

 as the structure of the gynoecium (always excepting its number 

 of parts), are the same in both. In both genera, although the 

 ovary is completely divided into cells at its base, the dissepiments 

 show an occasional tendency to separate from each other in the 

 upper part. In some species of Fagrcea, as I have already ob- 

 served, they do not even meet in the centre ; in Desfontainea I 

 have always found them to meet, but they are often very easily 

 separable, and if the ovary under examination has not been 

 thoroughly soaked, they appear on a transverse section to leave a 

 vacuity in the centre. This has probably been the cause of the 



