254 I'.KRTKAM G. SMITH. 



region of the early gastrula. Fig. 21, sketched a day later, shows 

 unmistakably that, during the process of overgrowth, material 

 lying immediately dorsal to the blastopore is being carried from 

 both sides toward the median line a phenomenon which I have 

 called "confluence." Fig. 22, sketched two days later than the 

 preceding, shows the same process carried further, with the 

 complication that the upper ends of the most laterally situated 

 marks are being carried downward and even slightly outward 

 through being involved in the process of convergence. The out- 

 ward movement is more readily comprehensible if we compare the 

 homologous regions in the embryos of an elasmobranch and a 

 teleost. 



Figs. 23 to 26 represent the history of another egg stained in 

 the same manner as the preceding. The latest stage figured 

 shows incidentally that overgrowth is accompanied by extensive 

 in-turning of the dorsal lip of the blastopore. The results as 

 regards confluence are even more striking; but the figures give 

 some evidence of a similar tendency in the material immediately 

 ventral to the early blastopore that is, in the future yolk plug. 

 This latter fact must be considered in our interpretation of the 

 movement of material in the dorsal lip. 



Figs. 27 and 28 represent an egg stained in the equatorial 

 region of the early gastrula in such a manner that the stain 

 extends in a meridional direction across the blastopore, con- 

 siderably further than in the preceding figures. The result in 

 later stages (Figs. 29 and 30) shows conclusively that the shifting 

 of tissues toward the median line just below the blastopore is 

 almost the mirrored image of that above the blastopore; the 

 differences in the two levels at the ends of the blastopore are to 

 be explained as due mainly to the process of convergence. 



These results suggest a mechanical explanation of confluence. 

 Does the material immediately above and below the blastopore 

 merely follow the law of liquids flowing through an orifice, ccfn- 

 verging toward the point of least resistance ? To test the matter 

 I studied the currents produced in shallow water which was 

 allowed to escape from a pan through a slit cut in the bottom. 

 When the slit was made slightly wider at the middle than at the 

 ends, the currents showed a tendency to converge, as shown 



