118 



occur in the pith as in I. tomentella or whether as in 1. ovalis 

 the medullary tissue is quite normal. The latter proves to be 

 the case and furthermore if a node be examined in cross- 

 section , it is to be seen that the opening left by the outgoing 

 leaf-trace-bundle is closed by lignified primary tissue exactly 

 as in /. ovalis. 



Pyr en a cant ha scandens Harv. 



a native of South Africa, belongs to a small group of plants, 

 of which the systematic affinities are not altogether certain. 

 While Baillon classes the genus among the 'Phytocreneac. Ben- 

 tham and Hooker merely append it to this order among the 

 „genera afjlnia aut exchtsa, vel duhici'\ The investigation of the 

 anatomy of the genus derives therefore additional interest 

 from the light it may throw upon the systematic relations of 

 the plants in question. Although in general the application of 

 anatomical characters in systematic botany has not met with the 

 hoped for success, still in groups of such strikingly characte- 

 ristic structure as the Bignoniaceae , Sapindaceae , and Phgiocre- 

 neae anatomical features have certainly considerable systematic 

 significance. It argues therefore without doul)t for the correct- 

 ness of classifying the genus Pyrenacantha with the Phytocreneae 

 that one of its species, Pyr. scandens, possesses an anatomical 

 structure remarkably like that of Phytocrene. In PI. XIX Fig. 6 

 the cross-section of a small Pyrenacantha stem has been repre- 

 sented from dried material. The wood, as may be seen, even 

 at this early stage of its development clearly shows five pro- 

 minences (/). Alternating with them stand five bast-plates {b'p) 

 closely reseml)ling those of Phytocrene in being richly supplied 

 with hard elements and comparatively large sieve-tubes. Not 

 only does the histological character of these plates but also 

 their number '), arrangement, and relative size recall the si- 



1) Although in general the number of bast-plates in stems and branches of P/ij/- 

 tocrene species is from 8 — 13 never-the-less at the base of the Ph.macrophylla, and 

 in a small piece of a stem of Ph. bracteata Griff. I have found but 5 plates present. 



