MENI8PEBMACEJE. 27 



When Menispermacece have like Schizandrece indefinite stamens and 

 albuminous seeds, the only absolute difference left between the two 

 groups lies in the form of the embryo ; which is short and relegated 

 to the extremity of the albumen in the latter, elongated with 

 flattened or leafy cotyledons, nearly equal in length to the albumen 

 itself in the former. Anonads, which so nearly approach Magnoliads 

 in almost their whole organization, must of necessity be very difficult 

 to distinguish absolutely from Menispermads ; and the more so as 

 the latter may occasionally possess a ruminated albumen. Anonads, 

 it is true, have often a polyandrous androceum, while this is excep- 

 tional in Menispermads ; but as the latter have not constantly six 

 stamens, and several Anonads may present a definite number, we 

 must look to the gynseceuni for a clear distinction : either the ovules 

 are indefinite in Anonacece, which is the case in no known 

 Menispermad, or there are but one or two ovules ; these however 

 are then ascending with the carpel looking downwards and outwards, 

 while in Menispermacece they are descending with the micropyle 

 upwards and outwards. The Nutmegs have, as we have seen, strong 

 analogies with those Menispermads whose albumen is ruminated ; 

 but they are easily distinguished by their simple perianth, their 

 completely diclinous flowers, their solitary carpel, and the direction 

 of the ovule, the same as in the uniovulate Anonads. Berberidacece 

 are closely allied to Menispermacece by the series Lardixabalece. 

 These have always compound leaves : which has been found in but 

 one genus of Menispermacece. 1 The floral symmetry, like diclinism, 

 is the same in both types, but no Menispermad is known to possess 

 a multiovular ovary, which is constant in Lardizabalece? Moreover 

 a relation of uncertain degree has been indicated, not without good 

 grounds, between EupUorbiacece and Menispermacece. A. Saint- 

 Hilaire noticed that the male flowers may be quite the same in 

 both groups, especially when the stamens are monadelphous. 3 But 

 the gynseceum and fruit are totally different. Finally, Menispermacece 

 may be considered at about the same distance from Lauracece 



1 Burasaia, which as we have seen, with all especially in the sarmentose species and those 

 the fundamental organization of 3Ienispermacece, whose male floral glands are sufficiently developed 

 was long relegated to Lardizabalece. to simulate the small petals of the Menispermads; 



2 Benthah & Hooker ((?e«., 30) also dis- accordingly several species of the group Phyllan- 

 tinguish Berberidacece by the valvicidal anthers ; thece have been referred to this order as new 

 but this does not apply to Lardizabalece, &c. generic types. 



s This is peculiarly the case with Phyllanthus, 



