722 MR W. T. GORDON ON 



Each arm in PI. II. fig. 28 is shorter in proportion to the length of the horizontal 

 bar than is the case in fig. 29. Further in towards the stem (PL II. fig. 27) the four 

 arms bend towards one another in pairs, and on the right-hand side at p 1 and p 2 , two 

 traces belonging to another pair of reduced pinnae can be distinguished. These two 

 traces are connected together by a thin line of crushed cells, and there is also a con- 

 nection between the two ends of the double trace and the tapering ends of the adjacent 

 arms. Thus we have in fig. 27 an island of parenchyma formed similar to that in 

 PL IV. fig. 42, between the bar b and the two arms. In this case, however, the 

 bar is poorly developed, and probably supplied a very reduced pair of pinnae ; the 

 island is not so broad as that in fig. 42. 



We have now seen two pairs of reduced pinna-traces entering the petiole, and the 

 only difference between them lies in this, that the lower pair is more rudimentary 

 than the upper. For, in the upper pair, the traces were followed in through the 

 cortex until they joined together, and the united trace became fused on to two adjacent 

 arms. In the lower pair, however, there was no actual fission of the two traces, 

 though the xylem was aggregated into two large groups which were connected to the 

 arms and to one another by a thin filament of xylem cells. Both pairs of reduced 

 traces failed to reach the exterior ; they disappeared in the outer cortex. Still 

 continuing downwards, the lower combined pinna-trace disappears gradually into the 

 arms, thus leaving them open. Below this level the arms are still closer than they 

 were before (PL II. fig. 26), and they are still shorter in proportion to the length of 

 the horizontal bar, which now constitutes the bulk of the trace. The trace has still, 

 however, a typical zygopterid form, having four arms with a protoxylem group lying 

 at the extremity of each. Still going lower down the petiole, we find the arms are 

 again bridged over by what is probably the xylem-bar of a still lower pair" of reduced 

 pinna-traces (PI. II. fig. 25, b). In this case, however, there is no evidence that the 

 connecting bridge of the xylem was ever aggregated into two groups, but it remains 

 as a thin strand one cell deep between the arms. Since the arms have become shorter, 

 as indicated above, the island of parenchyma between the two arms and the connecting 

 xylem bridge is very small. 



Before proceeding further, we had better recapitulate the results derived from 

 following the petiole-trace downwards. The arms of the petiole-trace have been 

 reduced in length compared with the horizontal bar, they have arched over towards 

 each other in two pairs, and they have lost the dilatations at their extremities. Three 

 pairs of reduced pinna-traces have been met with, and each pair is more rudimentary 

 than the pair immediately above. We have now reached the stage represented in 

 PL II. fig. 25. 



Continuing inwards, the bridge shown in fig. 25, b bends in at its centre towards the 

 end of the horizontal bar (which now constitutes almost the whole of the petiole-trace). 

 Thus the island formed by the bridge and the two small arms becomes reniform in shape 

 (Plate III. fig. 40). The two ends of the island c, c point outwards from the petiole- 



