414 MORPHOLOGY OF THE ANGIOSPERMS 



removed from the buttercup family and given family rank as the 

 Hydrastidiaceae (a family apparently not technically described). Evi- 

 dence from vascular tissue seems to show that the genus deserves 

 family rank. The vessel elements, like those of Faeonia, are mostly 

 scalariform; in all other genera of the Ranunculaceae, they are simple. 

 The vascular bundles of the stem approach the amphicribral type, the 

 phloem capping the xylem and like the xylem in amount, whereas the 

 vascular bundles of other genera in the family approach the amphivasal 

 type, where the phloem is almost surrounded by the arms of the xylem. 

 Another difference in the xylem is the scalariform pitting of the lateral 

 walls of the vessel elements, in contrast with the rounded pits of other 

 genera. These vascular differences are as great as those that set Paeonia 

 apart. 



If this family is established, the question of the position of the 

 Hydrastidiaceae must remain open until more information is available. 

 Whether the androecium is centripetal or centrifugal is uncertain; the 

 numerous stamens seem to mature simultaneously. Serological studies 

 also leave the position of Hydrastis in doubt. Serum reactions of the 

 genus show that it is apparently as close to PodophyUmn in the Ber- 

 beridaceae as it is to other genera in the Ranunculaceae. 



The Ranunculaceae stand out in the Ranales as a divergent herba- 

 ceous family in an otherwise dominantly arborescent order. Most of 

 the clearly primitive woody families in this order are characterized by 

 the presence of "ethereal oil cells"; none of the Ranunculaceae has these 

 cells. This histological character appears important as an indicator of 

 phylogenetic relationship in the group of highly primitive families of 

 this order. The absence of these cells from the Ranunculaceae widens 

 the gap between this family and the woody families of the order. 



The Ranunculaceae are sometimes cited as a good example of a 

 natural family, although, at one time, they were described as "one of 

 the most unnatural of natural families." The removal of Paeonia has 

 been a first step in a possible further breakup of this family; Hydrastis 

 will probably be removed, on the basis of the presence of scalariform 

 vessels in the wood and differences in pollen-grain type. 



The dominance of the herbaceous habit in the family strengthens the 

 position of the family as probably the most advanced in the order. 



Lardizabalaceae 



The chief morphological interest in the small family Lardizabalaceae 

 is in its primitive sporophylls: laminar stamens with deeply sunken 

 sporangia (in at least some genera), massive connective with elongate 

 apices; carpels with sessile stigma and laminar placentation. In con- 

 trast with these primitive floral characters, is the high specialization of 



