THE SPORANGIUM 641 



this hoop became simplified to a single cell-row ; and that it changed its 

 position in accordance with the soral arrangements, being oblique in some 

 Simplices and in all Gradatae, first with median but subsequently with lateral 

 dehiscence ; and that finally it swung again into the vertical position in the 

 Mixtae, but with the stomium still remaining lateral. 



Passing to the coritents of the sporangium, these are derived from the 

 archespovium, which in all Ferns, with occasional exceptions in abnormal 

 Eusporangiates, consists of a single cell. From this in the Leptosporan- 

 giatae the tapetum is first cut off peripherally, and the central cell which 

 remains divides up into cells which are commonly found to number some 

 power of two. This is a consequence of the successive divisions occurring 

 as a rule simultaneously within the sporogenous group, and the result is 

 that the spore-mother-cells number 2, 4, 8, 16, 32, 64, etc. Since each of 

 these undergoes tetrad divisions, the numbers of spores produced may as 

 a rule be 4, 8, 16, 32, 64, 128, 256, 512, etc. These may be styled the 

 typical numbers. 



There are two ways of computing the output of spores from a single 

 sporangium : either by examining preparations of sporangia with spore- 

 mother-cells, and estimating their number, or by actual counting of the 

 spores produced. Where the output of spores is small a reasonable 

 degree of precision is to be expected by either method ; but where the 

 numbers are large there are various sources of error, and the results must 

 be taken as mere approximations. The results of such computations will 

 now be given for various Ferns : the estimates for the Marattiaceae have 

 been made from study of sections traversing the sporogenous group before 

 tetrad-division, but in all the rest the results are those from direct 

 counting of the mature spores (see Table, p. 642). 



In estimating the value of these results it is first to be noted that the 

 largest numbers are associated with complete synangial fusion, a smaller 

 number being found in Angiopteris where the sporangia are separate than 

 in any of the synangial genera of living Marattiaceae. The potential 

 number estimated for Glcichema flabellata is nearly equivalent to that of 

 Atigiopteris, though the actual countings run considerably below ; still they 

 are in excess of those for any other Leptosporangiate Fern observed, and 

 as this is seen in a Fern in which the type of sorus is the same as in 

 Angiopteris, it strengthens the affinity between these two genera, notwith- 

 standing that the one is, as regards the mode of segmentation of the 

 young sporangium, a typical Kusporangiate, while the other shows 

 essentially the segmentation of the Leptosporangiates. 



From Gleichenia, as we pass through the table, successively lower 

 numbers are seen, and it is obvious that the larger numbers are 

 associated with those Ferns which on other grounds arc held to be 

 relatively primitive. Of the Simplices examined none have a lower typical 

 number than 128: and in this connection it may be noted that a rouuh 



estimate of the spore-output from a sporangium of 'Aygopteris, after 



2 s 



