io8 MARATTIACEAE [CH. 



accompanies their larger size : but they also show that the number of cycles is not directly 

 proportional to the dimensions. The result of the whole comparison is to demonstrate in 

 a general way that increasing vascular complexity tends to follow increase in size, but as in 

 Leptosporangiate Ferns a direct numerical relation does not exist. These two series illustrate 

 a homoplastic advance, and their similarity in principle though not in detail suggests that 

 similar causal factors underlie them both. 



The rather irregular cycles of vascular tissue in the Psaronieae no doubt correspond to 

 the rings of isolated strands of the Marattiaceae : so also their frequently continuous leaf- 

 trace corresponds to the series of separate strands of the Marattiaceous leaf-trace (Fig. 

 400, B). It appears that in these points the modern Marattiaceae are more advanced in 

 vascular disintegration than the Psaronieae. But Mettenius {Abhufidl. K. Sachs. Ges. 

 Bd. vi, 1863, Taf. I, Fig. 2) has shown how in a plant of Aitgiopteris that died after a 

 long struggle under cultivation the vascular rings are almost continuous in the upper 

 region (Fig. 397, B). Presumably it had retrograded structurally and pathologically towards 

 the earlier type. The general result of these comparisons is that the vascular construction 

 of the Psaronieae and of the Marattiaceae is according to a common type, both being 

 highly polycyclic : that the fossils show a less and the living Ferns a greater sub-division of 

 the vascular tracts : and that both conform to the methods of construction already recog- 

 nised in Leptosporangiate Ferns in relation to increasing size. 



The Spore-producing Members 



The sorus of the modern Marattiaceae is strictly circumscribed, and has 

 no definitely organised indusium. It is true that hairs may be scattered 

 round its periphery in Angiopteris (Fig. 402, B, D), and that the sporangia 

 of Danaea may be partially enveloped by upgrowths of the surface-tissue 

 (Fig. 402, K) : but these hardly deserve the name of indusium. The sori are all 

 constructed on the radiate uniseriate plan, where a single series of sporangia 

 is disposed in a radiate fashion round a central receptacle. When the sorus 

 is circular, as in Christensenia, the receptacle is a central point: when it is 

 elongated, as in Danaea, the attachment is linear. All the sporangia of a 

 sorus originate simultaneously, a character general for the Simplices. They 

 may be separate or united synangially : they are massive with a broad base, 

 and each produces a large spore-output. The dehiscence is in all of them by 

 a slit or pore, which though varying in form lies always in the median plane. 



The structure of the mature sorus in the five older genera is illustrated in 

 Fig. 402, all of them being superficial in position on the leaf, though varying 

 in their spread from the margin inwards. They vary also according as the 

 sporangia are separate as in Angiopteris (Fig. 402, A), or fused synangially 

 as in Marattia, Danaea and Christensenia {E, G, f). The sporangia are all 

 of the same massive type, and are arranged so as to face centrally, each 

 opening by a slit on the oblique inner face. Of those genera with separate 

 sporangia, the sori o{ Angiopteris are near to the margin (Fig. 392, A), and 

 each consists of about 20 sporangia. In Macroglossum also they are marginal, 

 but they extend farther inwards, while in Archangiopteris (Figs. 392, B\ 

 402, C) the)' are about half-way between margin and midrib, and extend far 



