XIV ISOETACEJE 545 



not. In case all of these primary archegonia prove abortive, a 

 small number, apparently not more than five or six, may be 

 formed subsequently ; but so far as my observations go, the pro- 

 duction of archegonia is limited, as is the growth of the pro- 

 thallium itself.^ 



The development of the prothallium goes on without any 

 increase in size, until the first archegonium is nearly complete, 

 about which time the spore opens along the line of the three 

 ventral ridges, and the upper part of the enclosed prothallium 

 is exposed, but projects but little beyond the opening. In case 

 all the archegonia prove abortive, the prothallium continues 

 to grow until the reserve food material is used up, but then dies, 

 as no chlorophyll is developed in its cells, and only in very rare 

 instances are rhizoids formed. 



Miss Lyon (3) figures a longitudinal division of the neck 

 canal cell in /. laaistris, and Arnoldi ( i ) states that a similar 

 division may occur in /. Malinvcrniana. 



The Embryo 



Besides the earlier account of Hofmeister, Kienitz-Gerloff 

 (6) and Farmer (2) have made some investigations upon the 

 embryogeny of /. laciistris, which correspond closely, so far as 

 they go, with my own on /. echinospora. 



The youngest embryos seen by me had the first division w^all 

 complete (Fig. 313, D). This is transverse, but more or less 

 inclined to the axis of the archegonium. The nuclei of the tw^o 

 cells are large and contain several chromatin masses. The sec- 

 ond division in the epibasal and hypobasal cells does not always 

 occur simultaneously, the low^er half sometimes dividing before 

 the upper one, and at times the second walls are at right angles 

 instead of in the same plane. Of the quadrants thus formed, 

 the two lower form the foot, and the two upper ones the cotyle- 

 don and primary root. The stem apex arises secondarily at a 

 later period, and probably belongs to the same quadrant as the 

 root ; but as it does not project at all, and is not certainly recog- 

 nisable until after the boundaries between the quadrants are no 

 longer evident, this cannot be positively asserted. 



Sometimes the quadrants divide into nearly equal octants, 



* In old pfothallia of /. lacustris according to Kienitz-Gerloff (6), there 

 may be 20 to 30 archegonia. 



35 



