78 



1S9S). In the Marattiaceae, however, the epibasal half of the 

 embiyo is turned away from the archegoniuni. 



Fig. BO (PI. VIII) shows a nearly medium longitudinal Section 

 of a young embiyo. This is very much elongated transversely, 

 •and to judge from the position of the cells the original basal 

 wall is followed by medium walls thus forming fairly regular 

 quadrants. A large cell (st) in one of the epibasal quadrants, 

 l)oth in form and position, agrees with the apical cell of the 

 stem described by Brebner for DanacK. I have found that a 

 very similar stem initial is developed very early also in Aun'i- 

 .opteris. The embryos, however, were too young to make clear 

 the relation of the cotyledon and primary root of the stem. 

 Three nearly transverse sections of an embryo of about the 

 same age are shown in Fig. 31 (PL VIII). 



As in the other Marattiaceae the cotyledon and primary 

 root grow in opposite directions. The epibasal quadrant is 

 turned away from the archegonium and the cotyledon thus 

 grows directly upward, piercing the overlying tissue of the 

 prothallium and emerging upon the upper surface. The root 

 grows downward so as to be almost in a straight line with 

 the cotyledon. It has been assumed heretofore that the root 

 originates from one of the hypobasal quadi-ants, but in Angi- 

 opteris this does not seem to be usually the case, and it 

 remains to be seen whether this is so in Kaulfiissia. The growth 

 of the root through the tissue of the foot explains the al)- 

 sence of a recognizable foot in the older embryo. The young 

 sporophyte is thus bipolar and its median portion is completely 

 surrounded by the prothallium. On a re-examination of my 

 preparation of the embryo of ][f(U'attia DoiKjIasii I am inclined 

 to l)elieve that in this species also the foot is developed from 

 the whole of the hypobasal portion of the embryo and all of 

 the other oi'gans are of epil^asal origin. It will not be strange 

 if this ju-oves to l)e the rule therefore troughout the Marat- 

 tiaceae, as well as in tlu^ Ophioglossaceae and Jsoetes. 



The cotyledon is l)roadly spatulate in outline and shows no 

 trace of the dichotomous branching that is usually characteristic 



