vSPOROPHYTE. 29 



the shoots arise often after the section lines are obliterated. They lie 

 above (ventral to) the stdmatic line of the petiole. At maturity the shoot 

 is from 3 to 8 mm. (average of 10 ==5. 4 mm.) from the rhizome. It is in 

 all respects a stem. Its tubular stele is attached nearly at right angles to 

 one margin of the trough-like vascular bundle of the petiole (fig-. 82). 

 The stele usually appears truncate at its point of origin, but it may be slit 

 on its upper side, giving a trough-like shape (figs. 93-96). I have never 

 found any commissural strand connecting the shoot directly with the stem. 

 as noted by Gwynne-Vaughan (1903) in other solenostelic ferns. The 

 inner parenchyma of the petiole is continuous with the medulla of the shoot. 

 Occasionally two shoots arise on opposite sides of a single petiole; they are 

 then either opposite or as much as 2 mm. apart. The shoot may remain 

 dormant as a mere papilla for two or three years, or its growth may be 

 extremely slow, or it may from the first nearly equal in size the leaf from 

 which it springs, or the leaf may fail to develop (owing to injury by fungi, 

 etc.) and the shoots alone continue. Ultimately the leaf-shoots produce 

 normal rhizomes, as shown by one which I measured in March, 1905. It 

 was 12 cm. long, 1.5 mm. in diameter at its origin and 2.6 mm. at 4 cm. 

 from its origin. During previous seasons it had borne three leaves, and 

 two more were ready to develop in 1905. The apex had just forked. 



The descriptions of leaf-development given by Sadebeck (1873, 1874, 

 1878, 1898), Kny (1875), and Campbell (1887, 1895, 1905) contain some 

 points which I can not interpret, and the leaf of Dcnnsttcdtia differs in its 

 development from that of the Hymenophyllacese as given clearly by Prantl 

 (1875, a, />). The leaf takes its origin from a deep four-sided prismatic 

 cell in the apex of the stem (figs. 103, 109, 118, 121, 128). The cell is 

 sometimes recognizable by its large size in the fourth segment from the 

 stem initial (fig. 109). It seems reasonably certain that not every stem 

 segment gives rise to a leaf, nor even two out of every three segments. 

 The location of the leaf -initial in the segment also varies much, according 

 to the size and rapidity of growth of the stem. It is usually near one mar- 

 gin of the segment (figs. 103, 109), and may lie very near the stem-initial 

 until many divisions have occurred in it (figs. 128, 129). The leaf-initial 

 in its earliest stage extends into the stem as far as the future boundary of 

 medulla and plerome. Its divisions are different from those of the neigh- 

 boring cells a fact which is related to the formation of the leaf -gap. No 

 definite order could be discovered in the early divisions of the leaf-initial. 

 It maintains its four-sided shape for some time (figs. 118, 121), then is cut 

 obliquely (figs. 121, 122-127) and becomes tetrahedral. After forming a 

 few segments on three sides, the initial ceases to form segments on one of the 

 three sides and it becomes ' 'two-sided. ' ' Meanwhile, all of the cells derived 

 from the primitive leaf -initial are elongating and forming" a papilla on the 

 stem apex, at the tip of which the definite apical cell of the leaf is situated 

 (fig. 70). This cell is wedge-shaped (figs. 123, 130) and very broad across 



