180 IXIIERITAXCE, FERTILITY. AND SEX IX PIGEOXS. 



interchanged. The germs of different species are specificallj- unlike, so that although 

 they set out in very similar ways, they graduallj- diverge and end in all the varieties 

 that inhabit the earth. Relatively slight initial differences are sufficient for greater 

 and greater and more numerous differences as development runs on. All this is to 

 be expected from a physical standpoint; special pangens to steer are superfluous; 

 their theoretical importation explains absolutely nothing, and only leaves us with 

 greater difficulties to account for than the phenomena themselves. 



Ontogenetic recapitulation in a given species is most wonderfully exact. It 

 is this perfection of reproduction that seems to require some extraordinary or 

 supernatural agencies. We easily forget that only physical processes can approach 

 such exactness. It is absurd to think such processes need or can have assistance in 

 reaching exact results. The more we reflect upon this, the clearer it becomes that 

 recapitulation must be physically directed, i.e. self-directing. 



Ontogenetic recapitulation is, then, a fact that admits of no explanation except 

 from a physical standpoint. In this fact lies the whole of heredity. 



Now, while ontogeny is so wonderfully exact that we never cease to be amazed 

 at its performances, we must not forget that germ-cells are subject to slow variation. 

 In fact, it is only germ-variation that has to be considered in phylogeny as in onto- 

 geny. Conseciucnth", when the germ-cell takes a step forward, ontogeny begins with 

 an initial difference that sets the whole series of ontogenetic stages on a diverging 

 line that digresses so little as to be undiscoverable until near or at the end of 

 development. 



Succeeding generations start with recapitulation at the new level or at the new 

 point in the same level. As variation in the germ moves, so the recapitulation 

 shifts; to be physically exact it must do so. Variation in results, then, is no evidence 

 of a fault in exactness, and the shift is so slow that recapitulation is closelj' ideal. 



ON MENDELIAN HEREDITY. 



In an address of last j'ear,' I had no occasion to do more than refer to Mendel's 

 brilliant work. Although ]\Iendel did not undertake to formulate any theorj' of 

 evolution, the problem of the origin of species was certainly in his mind, and he 

 hoped to get some decisive results by crossing varieties or species. A finer model of 

 experimental work and careful anah'sis has not been seen. The results were verj- 

 remarkable and are now accorded first rank by many prominent naturalists. 



I have devoted a good deal of time during the last 10 years to crossing various 

 wild and tame species of pigeons, and do not intend at this time to go deeply into 

 the subject of h3-bridization, but may refer briefiy to some experiments which lead 

 me to believe that ]Mendcl's experiments with different varieties of peas, important 

 as they are, do not reveal any fundamental and universal law of heredity. 



]\Iendel saw clearly that the problem of the origin of species must be attacked 

 from the analytical and experimental side. The first step, then, was to select 

 particular "specific characters" that were as sharply defined as possible. He 

 selected 7 pairs of contrasting characters. For example: tall and short stem; axial 

 and terminal flowers; smooth and icrinkled seeds; yellou- and green cotjdedons; gray 

 and brou-n seed-coats; inflated and constricted pods; green and yellow pods. 



' This part of the chapter is the latter part of a lecture delivered at the Wisconsin Natural History Society 

 Februar>- 27, 1908. (Z 5) 



