673 



side of tlie protoxylem. If we assume tliat the protoxjlem was 

 originally wliolly immersed in the metaxylem, but that afterwards 

 tlie development of traoheidal elements has been arrested at the 

 inner side, except in the very lowest part, we can explain the 

 existence of the interruption above the insertion of the pinna-bar. 

 For when the pinna-bar approaches the petiolar bundle and fuses 

 with it, the parenchymatous tissue at its adaxial side is enclosed. 

 The parenchyma associated with the protoxylems of the next pinna- 

 bar approaches in its downwards course the pei'ipheral loop formed 

 by the pinna-bar next above, and as the development of the procambial 

 cells into tracheids has been arrested, a break is formed in the loop. 

 Through this interruption the pai'enchyma at the inner side of the 

 pinna-bar is connected with that enclosed by the fusion of the 

 pinna-bar next above with the petiolar bundle. The parenchyma 

 which is enclosed and that which lies in tlie sinus is formed by the 

 fusion of the strands of parenchyma lying adaxially to the proto- 

 xylems of successive pinna-traces. These intei'ruptions in the peii- 

 pheral loop show some resemblance to the leaf-gaps in the stele of 

 many Ferns. Here, too, parenchyma situated adaxially to the proto- 

 xylems of the leaf-trace penetrates into the xylem of Ihe stem, either 

 connecting the softer tissue in the interior of the stele with that 

 without, or hollowing the xylem of the stem by the fusion of these 

 parenchymatous formations of successive internodes. In the first case 

 a little strand of parenchyma, ending below blindly, can be found 

 some distance below the insertion of the leaf-trace; in the other 

 case this funnel in the xylem is absent. The parenchyma enclosed 

 inside the peripheral loop may be compared with the pith, formed 

 after the second method, but the connection of the successive paren- 

 chyma-strands of successive pinna-traces is not caused by reduction 

 in tissue which was present before (in phylogenetical sense). This 

 structure, caused by the peculiar symraetrj- of the bundle, is present 

 on both sides. 



This species agrees in the form of the antennae with E. Scotti 

 Bertr.,') but differs from it by the simpler structure of the pinnae- 

 bundles, its smaller dimensions, and the more scattered position of 

 the idioblasts in the inner cortex. It differs from E. shorensis Bertr. ') 

 by having another form of the apolar. In this species the continuity 

 of the pinnabar is maintained for a rather long distance, but Ihe 

 presence of a peripheral loop has not yet been noted. A continuous 



1) P. Bertrand, 1909, p. 14.0—147, 209, pi. XVI, fig. Ill, 112. 

 ') P. Bertrand, 1911, p. 30—38, pi. II, fig. 23—31, 34, 35. 



