600 Annals of the South African Museum. 



a natural result, the range of stages of development seen in one colony 

 is very much smaller than in F. Rousseletti. 



Secondly, there are certain outstanding differences in development, 

 the significance of which is difficult to understand : (1) The mother- 

 cell passes through the period of enlargement with loss of cilia and 

 subsequent down-sinking into the parent colony just as in F. Rousse- 

 letii, but no change in the orientation of the cell constituents follows ; 

 as a consequence the polar axis of the cell remains radial, the first 

 two divisions take place in two radial planes at right angles to one 

 another, and the phialopore is directed outwards (fig. 7, E), lying 

 immediately below the parental pore formed by the sinking of the 

 initial cell, and surrounded by the " perigonidial crown " of somatic 

 plastids (Klein, 1889, p. 20). 



Here and there a similar case is observed in F. Rousseletii, while, 

 conversely, a shifting resulting in the pore being placed laterally is 

 occasionally seen in F. capensis ; but in general the above rule holds 

 for the two species respectively. 



At first it was thought that this curious divergence in development 

 might be correlated with the direction of escape of the mature male 

 globoid, that the shifting might result in a globoid which escaped 

 outwards, while its absence resulted in a globoid which would eventu- 

 ally escape inwards into the hollow of the parent, and that it was 

 consequently to be correlated with the distribution of the sexual 

 cells, but there seems to be no evidence to support this view. In 

 both species the rule is for the globoid to escape outwards into the 

 surrounding medium and only exceptionally inwards into the parent. 

 Further, since the parent vesicle which encloses the mature male 

 globoid is amply large enough for it to rotate freely within it, there 

 seems no reason for the orientation of the globoid during develop- 

 ment to determine the direction of escape. 



(2) Possibly the difference of position may be connected with the 

 second divergence in behaviour this is in the rate at which inversion 

 proceeds. As yet inversion has only been watched in comparatively 

 few cases in F. capensis, but in every case it was found that the rate 

 of inversion was far more rapid than in F. Rousseletii ; whereas 

 in that species the stages from the opening of the pore to completion 

 of inversion took from forty-five minutes to two hours or more, in 

 F. capensis they are usually completed in under half an hour, and 

 although the process follows similar lines the successive " hat " 

 stages were not nearly so clearly defined (fig. 7, E-J). It is possible 

 that the position of the connecting strands relative to the phialopore 



