XX] 



SORI 



109 



along the veins, each consisting of a much larger number of sporangia. A 

 fossil type, Danaeopsis Heer, has lately been shown by Halle {Archiv fur 

 Botanik, xvii, No. i. 1*921) to have sori extending as in Danaea the whole 

 distance from margin to midrib: but here as in Angiopteris all the sporangia 

 are separate. The genera with synangial sori illustrate like differences. 

 Marattia bears its synangia near to the margin (Fig. 392, Cy. in Proto- 

 inarattia (Hayata) they are submarginal, and rather longer: while in Danaea 

 (Fig. 392, E) they may extend along the whole length of the veins. Thus the 



C ^5 jl ^JZ jx 



Fig. 402. Sori and sporangia of Marattiaceae. A, />', Angiopteris crassipes Wall. A, sorus: B, two 

 sporangia, one in surface view from without, the other cut longitudinally. C, D, Arckaiigiopteris 

 Henryi Christ and Giesen. C, sorus ; D, two sporangia in section. E, F, Marattia fi-axinca Lin. 

 E, synangium; F, same in section. G, H, Christensenia {Kaidfussia) aesctdifolia (BL). C, part 

 of the lamina seen from below, with three synangia ; H, section through a synangium. J, K, 

 Danaea elliptica .Sm. J, two synangia; A', section through a synangium. [A, B, E, G, J, after 

 Bitter. C, D, after Christ and Giesenhagen. F, H, K, after Hooker-Baker. From Engler and 

 Prantl.) 



degrees of extent of the sorus and its relation to the margin appear to be 

 similar, whether the sori be synangial or with the sporangia all separate. 

 Finally, in Christensenia (Fig. 402, G), which is a shade-loving, broad-leaved, 

 reticulate type, the rosette-like sori are dotted over the broad under-surface 

 of the pinna. Comparison with the leaves of Danaea, and especially with 

 those which are only partially fertile, gives a clue to an explanation : for 

 there the elongated sori are found to show occasional fissions, and the partial 



