724 MR W. T. GORDON ON 



the latter joins the stem, and among other ferns, e.g. the Osmundacese, this is exactly 

 where numerous roots occur. This trace, therefore, may quite well be a root. 



One point which must be noticed here is that the two ends of the petiole-trace never 

 show the same stages simultaneously. The difference in development of the departing 

 traces at opposite sides of the petiole increases as we ascend. At the level of PI. II. 

 fig. 23, the difference in phase of the traces at the two ends is not great. The ends 

 of tins trace are shown more highly magnified in PL III. figs. 33 and 41. From 

 a comparison of fig. 41, showing one end of the trace, with fig. 36, showing the 

 other end at a higher level, it will be seen that they are in the same phase. Between 

 these two sections two others were prepared, so that the difference in phase between 

 the two ends of fig. 24 corresponds to about four sections, or ^ in. to £ in. Near 

 the union of the petiole with the stem the difference in phase is much less. In PL I. 

 fig. 10, pet. tr., the only dissimilarity is that one end of the trace has become free 

 from the stem before the other. This difference is always exhibited until the xylem 

 of the petiole has become completely merged in that of the stem, as is shown in PL II. 

 figs. 19, 18, and 17, where the top part of the trace is distinctly more merged in the stem 

 than the lower end. In fig. 18 the petiole-trace as a whole is becoming completely 

 incorporated into the stem-xylem, and in fig. 17 the fusion is still more perfect. The 

 petiole-trace has now lost its identity except for the protoxylem groups which are still 

 very distinct (prx^ and prx. 2 ). 



The stage before fig. 17 is seen in PL II. fig. 16, where again the protoxylem groups 

 are distinct {prx. x and prx. 2 ). The elongated shape of the whole stem-xylem, the trun- 

 cated outline at one end, and the two distinct protoxylem groups are the only indica- 

 tions of petiolar departure that can be seen. PL I. figs. 15 and 14, represent still earlier 

 stages in the departure of the petiole-trace, and the protoxylem groups are coming to lie 

 almost on the periphery of the internal xylem. At the earliest stages observed, the stem- 

 xylem is slightly elliptical, with two adjacent protoxylem groups lying at the border 

 of the inner xylem. This fact that the petiole is decurrent into the stem for such a long 

 distance makes it impossible to say whether the stem had protoxylem groups peculiar 

 to itself or whether the only protoxylem groups present belong to the petiole-traces. 



Among the Botryopteridese we meet with this same difficulty. Another interesting 

 fact is that, as in the Botryopteridese, the tracheides of the petiole of Diplolabis 

 romeri are larger than those of the stem. The average diameter of a stem tracheide is 

 160 /x for the outer xylem and 100 m to 120 m for the inner xylem. The average for 

 the petiole-trace is 200 /n. This difference in size of the tracheides of stem and petiole 

 is very marked in Botryopteris antiqua Kidston. 



As mentioned above, the stages of the petiolar departure were examined in more 

 than one series, but no one series gave a complete sequence of these stages. There is, 

 however, a great amount of overlapping (see text-fig. 1), and it has been possible to 

 give an estimate of the distance between the departure of the petiole-trace from the 

 stem, and its assumption of the typical Diplolabis form. This distance is between 



