EMBRYOLOGY OF THALASSEMA MELLITA 175 



formation of the fourth quartet, mitotic figures appear in the 

 four cells at the upper pole (i^i. i, etc.) and the apical rosette 

 is budded off leiotropically (Text-Fig. 2, E). These cells lie 

 entirely on the surface at first and are much larger than in 

 Podarke. Although in later stages they occupy a smaller space 

 at the surface, they never sink in to the extent that they do in 

 Podarke and Lepidonotus. At this stage two of the polar bodies 

 are very commonly seen imbedded in the stem-cells of the 

 cross (PL I, Fig. 2) or in the rosette, but they interfere in no 

 way with the division of these cells. This sinking in of the polar 

 bodies is of very general occurrence in annelids and molluscs 

 (Treadwell, '01). 



The next divisions take place in the third quartet. They are 

 leiotropic and somewhat unequal (PL I, Fig. 8). From the 

 lower cells in the a, c and d quadrants will later arise part of 

 the mesenchyme. The prototrochal cells divide equally and 

 leiotropically and at the same time the intergirdle cells divide in 

 the same direction and unequally (PL I, Fig. i). The smaller 

 moieties lie in the spaces between the four groups of cells which 

 constitute the primary prototroch and later form a part of the 

 secondary prototroch as in Podarke (PL I, Fig. 6). The embryo 

 now consists of fifty-six cells and is perfectly radially symmetri- 

 cal (PL I, Fig. 2). 



This stage marks the end of the purely radial cleavages, as in 

 Podarke, and from now on bilateral or morphogenic cleavages 

 take place in increasing numbers. The first of these initiates 

 the formation of the cross. The posterior stem cells (i<^].i.2 

 and i^Tj J 2) divide bilaterally, while the anterior divide spirally 

 and leiotropically (PI. I, Figs. 2 and 3). This cleavage, however, 

 is subject to a number of variations as will be indicated later 

 (see p. 187). From now on in the praetrochal region tlie posterior 

 quadrants (c and d^ always divide before the anterior, whereas in 

 Podarke the reverse seems to be the case. As Child ('00) has 

 indicated there are in annelids two types in regard to the forma- 

 tion of the cross. In Nereis, Polyinnia, Spio, Aricia, AmpJutrite, 

 Clymenella, Sternaspis and Arenicola the spindles lie meridion- 

 ally, whereas in Podarke, Lepidonotus and CJieetopterus spirally 



