THE SALMON FAMILY. 219 



fish by the heads will cause limited exudation of milt and 

 ova. The pressure, or weight of the roe above on the roe 

 beneath, is sufficient to produce this partial exudation. 



" The facts stated in the preceding paragraph are sufficient 

 not only to account for impregnation taking place at inter- 

 vals, for the impossibility of its being done tout d'un coup, 

 but also for the improbability of impregnation by intermis- 

 sion or coitus either before or at the time of the deposition of 

 ova. Mr. Young dissipated all doubts upon this point by 

 the following experiments : He took a female Salmon, exu- 

 ded by manipulation a portion of her ova, and having simply 

 done so, he buried it beneath the gravel of one part of an 

 artificial spawning-pond. From the same Salmon he exuded 

 another portion of ova, but before he covered it over with 

 the gravel of another portion of his spawning-bed, he impreg- 

 nated it by pressing milt from the male Salmon, and causing 

 it to ^come in contact with the last ova deposited. He then 

 covered them in beneath the gravel, and in due time they 

 produced fish. The ova he had covered in without impreg- 

 nation produced nothing. He repeated the experiment fre- 

 quently, and always with similar results. He has even taken 

 two female Salmon in the act of spawning. The ova of one 

 he impregnated with milt from a male ; the ova of the other 

 he did not impregnate. He covered in each under equal 

 conditions, apart in the same spawning-bed. The ova that he 

 had caused to be impregnated were productive ; the other 

 proved perfectly barren. This experiment was repeated, and 

 the result was ever the same. 



' It may be asked, how is it that ova and milt are not 

 swept away by the action of the rapid water in which they 

 are deposited that they are not swept away in the act of 

 deposition, and before they can be covered in beneath the 

 gravel? It would appear that at the critical moment of 



