The Laws of Heredity. 177 



We remark, in the first place, that the empiric formulas given 

 above are capable of being simplified and reduced under two chief 

 heads : immediate and mediate heredity. When we find a child 

 resembling its father or mother, the fact appears perfectly simple, 

 either because it is so common, or because we judge it to be quite 

 natural that like should produce like. But when we see the great- 

 grandchild resembling the great-grandfather, or the great nephew 

 resembling the great uncle, and this without any intermediate 

 stage to explain the resemblance, this case of heredity appears 

 to us so strange that many have rejected it. It would then be a 

 great point if we could show that this mediate heredity resolves 

 itself into the other form. To do this we must make a brief 

 digression. 



All naturalists are agreed that no studies are of more advantage 

 for them than those of comparative anatomy and comparative 

 physiology ; that the knowledge of rudimentary organisms has 

 given them a better understanding of organs and functions, and 

 that these results have been specially remarkable as regards 

 generation. The study of the lower forms of this function has 

 greatly enlarged their views, and even .entirely modified the ideas 

 of scientific men on that subject Among these discoveries, that 

 of alternate generations appears to us, of all others, the best fitted 

 to throw light on the subject which now engages our attention. 



In 1818 Chamisso's studies on certain molluscs called biphorae, 

 or salpae, led him to the discovery that these animals are alter- 

 nately free and aggregated. In the first generation strings of 

 biphorae are found, the product of gemmation ; in the second, 

 solitary biphorae produced from spores ; in the third, the strings 

 reappear: so that the young never resemble the parent, but always 

 the grand-parent 



ist generation Aggregated salpae Grandfather 

 2nd Free Father 



3rd Aggregated Son 



The researches of Saars, Steenstrup, Owen, and Van Beneden, 

 show that in some animals the cycle is not limited to three genera- 

 tions, but that often it is more extended, and that the resemblance, 

 instead of passing from the grandfather to the grandson, passes 

 from the great-grandfather to the great-grandson. In those species 



