378 



SUMMARY OF EXPERIMENTS. 



INTEOOrCTORY. 



In the preceding detailed account of the various crosses effected are 

 included combinations between forms of teleosts. ranging from closely re- 

 lated species within the same genus to species belonging to widely separated 

 orders. Their relationships are summarized below, the figures set opposite 

 each indicating the number of different combinations made in each group :^ 



Between different species of same genus 2 



" " genera of same family 4 



" " families of same order 1 



" " orders of same class 17 



A number of interesting facts appear from the above table and from a 

 closer inspection of the more detailed Table 9. In all the crosses attempted 

 with the exception of tlie cross in which Opsanus tan was used as the male, 

 impregnation was possible. The sperm of the single Opsanus tan specimen 

 used was not ripe in the three combinations attempted, so that it is impossi- 

 ble to say whether these crosses are possible. 



CHAEACTF.R OF IMrKEG?rATIO>'. 



In many of the crosses the impregnation was wholly normal. In some 

 there was in addition to the noruially impregnated eggs, a varying number 

 of dispermic and polyspermic impregnations. Among the abnormally im- 

 pregnated eggs the disi>ermic was very much more common than the poly- 

 spermic condition. In the dispermic eggs tiie protoplasmic disc, as is well 

 lvno\\'n, falls at once into four cells. Sections of these conditions show that 

 tAvo male pronuclei fuse with the egg pronucleus ; wliether additional sperma- 

 tozoa enter such eggs, but i-smain functionless so far as early cleavage is 

 concerned, I am not able to say. In the polyspermic eggs the protoplasm 

 falls at once into six or more cells. The cases coming under my observa- 

 tion in which many cells at once appeared, have been rather rare. ■ 



PERCENTAGE OF FERTILIZATION. 



A Striking fact is the large iiercentage of eggs impregnated. In fully 

 two-thirds of the crosses this ran above 50 per cent., and in many of the 

 combinations it ran above 75 per cent. A glanc« at Table 9 will show that 

 this is not in any way correlated with the nearness of relationship. 



1 This is represented in more detailed form in Table 9. 



