184 CHARLES R. STOCKARD 



viduals of another type, his confidence is considerably shaken in 

 the vaUdity of the selection. I have had similar experiences in 

 attempting to isolate from large numbers of eggs those showing 

 the earliest indication of the cyclopean defects. There are no 

 doubt processes of regulation which may tend to correct and 

 obliterate an early unusual arrangement, yet in spite of the rec- 

 ognized probabilities for mistake, I nevertheless feel that the 

 foregoing indication of suppression of early buds has some real 

 value, since two actually double specimens were certainly selected, 

 as later development showed. 



Kaestner ('98-'07) has figured very early double primitive 

 streaks in the chick and Assheton ('08) double embronic shields 

 in the sheep. Kopsch ('95) has described in Lacerta agilis, the 

 European lizard, one blastoderm with two blastopores, and thus 

 showed that a double gastrulation had taken place. From this 

 observation he agreed with O. Hertwig that all twin formation 

 as well as all anterior duplication arose from a double gastrula- 

 infolding or proliferation. This position leaves, as Kaestner ('99) 

 has stated, the question of doubleness or twins merely moved 

 back to an earlier stage before the origin of the two blastopores, 

 it remained to be answered why the double infolding takes place, 

 and w^hy it is so rare? In the present study it is felt that both 

 questions are answered. A developmental arrest does away with 

 the normal advantage of the usual growing point and permits a 

 double gastrulation; the condition is rare for the same reason 

 that the apical or dominant bud rarely fails to grow. 



Returning to the consideration of the actual case in the trout, 

 we may judge indirectly by the degree of separation of the two 

 components in the several individuals as to the probable distance 

 apart of the original embiyonic shields or embryonic axes on the 

 germ-rings. Gemmill ('01), has found a rather high propor- 

 tion of complete twins among something more than seventy 

 double trout specimens that he examined, while Windel ('95), 

 had found only nine complete twins among 117 eggs containing 

 double trout, or a proportion of one to thirteen. Among my 

 double trout specimens there is one case of complete twins for 

 every eight. From these observations it may be concluded that 



