34 



THE CASUARINAL DIVISION. 



A SCLEPI ADEiE. 



There is such a well-marked coincidence in structure between Violacece and this family in the stamens, that although perhaps it has never been suggested that any near affinity exists between 

 them (unless between the coronet of Stapelia and Passifloracece), yet as they are more nearly related to Violacece than to any other polypetalous family, they would by inference be their monopetalous 

 form. This similarity is especially remarkable between the stamens of Hoya pulchella and lonidium commune and the spurs of the latter are distinctly the same in their mode of 

 formation, the difference being that in the former the filament is longer and the anther cells much shorter, while the connective continues elongated (which is frequent in Asclepiadece), which occasions 

 the spur in lonidium to come off from the filament much lower down. Although in lonidium only 2 of the stamens have appendages, yet in Hymenanthera all equally have them, and the anthers are 

 connate as in Asclepiadece. In Asclepiadece the appendages, when erect, are sometimes much elongated and even meet over the centre of the flower, and their being erect in Hymenanthera (which 

 are the same as those of lonidium become erect,) is together with the connate anthers a singular coincidence in structure, because in Viola they are directed downwards. 



Although they differ from Violacece in having an axial placentation, yet in the very nearly allied Apocynacece the placentation is in part of the genera parietal. They also agree remarkably 

 with Violacece (i. e., when the ovules are suspended,) in the position of the raphe, and by placing them here as a monopetalous form of that family they are brought into their natural position in 

 Table VII., which I venture to believe is an admissible argument, with reference to true station, where affinities are doubtful. Mr. Griffith, I believe, first noticed the position of the foramen of the 

 ovule, where they are numerous and horizontal, (Trans. Linn Soc. vol. XX. p. 391,) which is on its upper margin, and Viola tricolor corresponds very remarkably in having this character, 

 Jpocynacece being the only other family in which I have observed it. With other families of this Division they agree in the production of caoutchouc, of medicines called Ipecacuanha, and very 

 strong fibre, that of Marsdenia tenacissima being amongst the strongest known. 



J UGLANDEiE. 



This family is generally compared with Anacardiacece by systematic botanists, but in all the amentaceous families the ovule when pendulous has the raphe next the placenta or ventral, from 

 which it is at least very probable that it would be so in Juglandece if the ovule were pendulous, and this would be an anomaly among the allies of Anacardiacece. In their compound leaves and 

 bifid stigma they agree with Euphorbiacece, that family, although so often deleterious, being also remarkable for edible nuts as Conceveiba Guianensis. 



