II. i CELL-DIVISION 47 



the bubbles the conditions observed in the four-celled stage of 

 other types Nereis, Arenicola, Unio, Aplysia, Discocelis may be 

 faithfully copied. 



It only remains to be added that the contact surfaces of the cells, 

 like those of the bubbles, make angles of 120 with one another. 



Robert has also imitated the eight-celled stage (the four 

 micromeres alternating with the four macromeres), the stage of 

 twelve cells (division of the micromeres), and that of sixteen 

 cells (second quartette formed). The bubbles of the second 

 quartette may be made to slide in between the macromeres and 

 so rotate the whole first quartette, as happens in the egg. The 

 division of the micromeres in the egg results in the arrangement 

 of four cells crosswise in the centre, four others occupying the 

 spaces between the arms of the cross. The bubbles behave in 

 the same manner. 



In the eight-celled stage the micromeres alternate with the 

 macromeres. In the case of the bubbles this is not necessarily 

 so ; the two sets of bubbles may be superposed if the ' polar 

 furrow ' in one tier is at right angles to that in the other, or if, 

 as pointed out above, the upper bubbles are small. Otherwise 

 superposition is a very unstable condition. 



It would appear then that many of the patterns exhibited by 

 eggs with a spiral cleavage are explicable by reference to the 

 laws of surface tension. The principle of least surfaces may 

 be extended to other cases. The first four blastomeres of 

 Ophiuroids and Asteroids form a perfect tetrahedron, though 

 this arrangement is subsequently discarded for one which could 

 not be imitated with soap-bubbles (we may notice in passing that 

 in the first case the egg is tightly invested by its membrane, 

 in the second it is perfectly free). In Ascaris megalocepkala 

 the four cells come to lie, as do four bubbles, in one plane, 

 and polar furrows have been seen in many eggs which belong 

 to another type of segmentation (in Coelenterates (Hgdractinia), 

 Sponges (Spongilla), Crustacea (Branc/iipus, Lucifer, Orchestia), 

 Vertebrates (Petromyzon, Rand), Ascidians, and Ampltioxus). 



The principle of least surfaces not more than three surfaces 

 meeting in a line, not more than four lines meeting in a point 

 is, however, not of itself sufficient to explain the whole of the 

 phenomena even in this most favourable tetrahedral type ; 



