July. 1921.] OQ UBA :—OAPS OF STELE IN POL YFODIACEAE. 1 23 



stele above described should not be regarded as the true foliar gaps ; 

 and if it be so, would it not be a contradiction to put Polypodiaceae 

 among Pteridopsida ? The typical foliar gaps following his definition are 

 present in some species of Adiantuw, and also according to GwynnE- 

 Vaughan^^ in Loxsoma. There are some cases where a trace divides as 

 soon as it departs from the gap (as in Atbyrium), and if this tendency 

 becomes more prominent two traces will leave the stele separately 

 from both sides of a gap (as in Diplazium, etc.) ; and the cases when 

 three or more traces depart separately from a gap may result. Such 

 a case was observed in sporelings of Marattiaceae.^^ The primordial 

 leaves in this family have only a single foliar trace, while the follow- 

 ing ones have a trace divided into two at the upper part, though it 

 remains single at the base ; and in some subsequent leaves two traces 

 leave separately from a gap, and at last the particular type of 

 traces develops gradually. Thus, it will be natural to consider that 

 the gaps above described as belonging to the true foliar gaps, and 

 also that the Polypodiaceous steles having such gaps are the typical 

 dictyosteles. 



In some younger part of the rhizome with dictyostele, we may 

 find the solenostelic condition (such as in Athyrium nippotiicum, Dipla- 

 zium japonicum). Considering these facts and the occurrence of the 

 transitional types between two stelar forms, it seems natural to con- 

 clude, as many authors have done, that the dictyostely may be 

 derived from a solenostely. But, two conditions or factors mentioned 

 before on the formation or appearance of the dictyostely, do not 

 hold true in this respect, because even in the rhizomes with long 

 internodes dictyostelic condition is generally present irrespective of 

 the length of internodes. The dictyostely of Polypodiaceae, there- 

 fore, at least in the species I have investigated, can be ascribed to 

 the constant relation of foliar gaps and commissure between them, 

 and not to the length of internodes. 



Summary. 



1. In some creeping rhizomes of Polypodiaceae, leaves are ar- 

 ranged in two rows on the dorsal side of the rhizome. This is, 



1) Gwynne-Vaughan, 1. c. 



2) Farmer J. B. & Kill, T. G. (1902) On the arrangement and structure of 

 vascular strands in Angiopieris evecta, and some other Marattiaceae.— Ann. Bot., Vol. 

 16 : West, C. (1917) A consideration to Uie study of the Marattiaceae.— Ibid., Vol. 31. 



