EMBRYOLOGY OF THALASSEMA MELLITA 191 



cells in every quadrant (Text-Fig. 3, D^ instead of only in the 

 anterior and lateral, as in the normal type (Text-Fig. 3, C). In 

 the next division of the cross the stem cells normally divide 

 equally in the anterior, but extremely unequally in the posterior 

 arms. Yet, here again, in the radial type we have an equal and 

 strongly marked leiotropic division in every arm (Text- Fig. 3, D). 

 In this tj'pe, too, we have a retention of radial division of the 

 post-trochal region. No X^ o cell is formed and the second 

 division of the third quartet is exactly alike in every quadrant. 

 Accordingly, at 72 cells we find a complete reversal to radial 

 cleavage. I have been able to follow this radial type to over 

 130 cells and, although certain bilateral features, in varying 

 degree, creep in, the great majority of the divisions are purely 

 radial. There is, also, strong reason for believing that this type 

 never develops into an adult ; for I found a few gastrula stages 

 in which no X group, no ^/ cells, no larval mesenchyme cells 

 from the third quartet and no large anterior oesophagoblast 

 (2^2 2-) could be distinguished. It is significant to note, how- 

 ever, that certain cells (mesenchymatous) were sinking in from 

 the first quartet. 



These two types, the purely radial (Text- Fig. 3, D) and the 

 normal bilateral (Text-Fig. 3, C), are connected by numerous 

 transitions. The X, o cell may be formed in embryos where the 

 praetrochal region continues radial and the second division of 

 the third quartet is also radial, or bilaterality may be alone ex- 

 pressed by the division of the posterior intergirdle cell. 



There is another distinct, but very rare, type which seems to 

 be a step toward a purely bilateral cleavage (Text-Fjg. 3, ^). 

 In this case the cross is formed by equal and bilateral divisions 

 in every quadrant, the rudimentary cells are formed as usual at 

 the second cleavage in the posterior arms of the cross and the 

 cells next to the rosette all divide toward the median line. The 

 second division of the anterior and posterior intergirdle cells is 

 nearly equal, but the rudimentary cells are budded off as usual 

 in the lateral. 



Treadwell has studied a few (only three cases) similar varia- 

 tions in Podarke. He calls attention to the fact that here " the 



