1906.] 



a Neiv Natural Order of Plants. 



235 



ment. Taking the morphological characters seriatim, it is evident that the 

 closest relationships are with the Anacardiaceae and Cupuliferse. The 

 absolute separation of the sexes and the very great diversity of the floral 

 structure of the sexes, associated with pinnate leaves, offers a combination 

 of characters probably without a parallel. 



Beginning with the foliage, the Julianiaceae have alternate, exstipulate, 

 imparipinnate leaves in common with at least eight different ligneous orders, 

 but here the affinity, or, rather, resemblance ends so far as six of them are 

 concerned, and the comparisons need be carried no further. There remain 

 the Anacardiaceae and Juglandaceae, both of which are also resiniferous, both 

 have unisexual flowers with reduced envelopes, at least as to some of their 

 members, and both have solitary, exalbuminous seeds. Other points of 

 resemblance or similarity in the Juglandaceae are the dissimilar male and 

 female flowers, the broad, stigmatic lobes of the style, and the single-coated 

 ovules. Juglans has also a funicle of unusual development. But the 

 combined characters in common of the Julianiaceae and the Juglandacese 

 cannot be regarded as constituting a close affinity. In some respects there 

 is a nearer relationship to the Anaeardiaceae. The anatomical characters of 

 the two orders are very much alike ; but as Dr. F. E. Fritsch will describe 

 and discuss the anatomy in a separate paper, it is unnecessary to enter 

 into particulars here. 



The nearest approach I have found to the singular funicular development 

 of the ovule is in the Anaeardiaceae, but the resemblance is remote and the 

 ovules of the latter are double coated. Coming to the seed and the embryo, 

 however, the resemblance is complete, and, apart from the slight obliquity 

 of the cotyledons of Juliania, the description of the seed and embryo of 

 Cotinus or Rhus would do for Juliania. With this the affinities to the 

 Anacardiaceae are exhausted, and they are not sufficiently strong to justify 

 the juxtaposition of the two orders. The next comparison is with the 

 Cupuliferse, taking the order as limited by Bentham and Hooker. There is 

 nothing in the secretions nor in the foliage to warrant an approximation of 

 the two orders, and in habit of growth the Julianiaceae are very different. 

 But divergences as great, or greater, exist between closely associated orders, 

 and even between genera referred to the same order ; and when we come to 

 the inflorescence and flowers, affinities are evident; that is if affinities are 

 deducible from similarities in structure. 



The male inflorescence, the male flowers, and the pollen of Juliania 

 adstringens are so near in texture, structure, and form to the same parts in 

 certain species of oak that, detached, they might be referred to the genus 

 Quercas. In fact, there is much greater dissimilarity in the male inflorescence 



