630 STUDIES IN FOSSIL BOTANY 



being present at each bifurcation. This variation occurs, 

 in varying degrees, in the middle of the reproductive 

 region of the most vigorous plants, and is extraordinarily 

 frequent, sometimes extending to a majority of the 

 sporophylls on the shoot. Professor Thomas compares 

 the arrangement in these compound sporophylls with 

 that in Cheirostrobus (p. 1 16) ; a closer parallel may be 

 found in the repeatedly forked sporophylls of Spheno- 

 phyllum majus (p. 113), with which the agreement 

 appears to be almost exact, except for an unimportant 

 difference as regards the number of sporangia in each 

 group. In other instances observed by Thomas the 

 synangium acquires an elongated stalk, with somewhat 

 pendulous sporangia, as in Bowmanites Romeri. Some 

 similar variations were also observed in Psilotum. 

 The repeatedly forked form of sporophyll, which appears 

 to be so common as to fall under the head of normal 

 variations, is unfavourable to the idea of any near 

 affinity between Psilotaceae and the Lycopods, while 

 it affords strong evidence for a relationship to the 

 Sphenophyllales. Such a relationship also explains 

 the forked form of the normal sporophyll characteristic 

 of the family. Taking the anatomical characters also 

 into account, there is thus a strong case for an affinity 

 between the Palaeozoic Sphenophyllales and the recent 

 family. I do not, however, go so far as Professor Thomas 

 and Professor Bower, who include the Psilotaceae in the 

 class Sphenophyllales. The differences are, in my 

 opinion, too great to allow of so close a union. The 

 most obvious distinction is in the phyllotaxis, which is 

 spiral, or at least alternate, in the Psilotaceae, while it 

 is always verticillate in the Palaeozoic Sphenophyllales. 



