APPENDIX , 357 



tijmbler pigeon and a white of pure blood would produce 

 young, from the offspring of which pure dominants and 

 recessives, closely resembling the original parents, could be 

 segregated, because they differ only in color. On the other 

 hand, if a bulldog and a spaniel were crossed, the chances 

 of producing a specimen resembling either parent would 

 be remote. The various paired characters would be trans- 

 mitted intact and could be segregated, but the exact num- 

 ber of dominants and recessives required to make a typical 

 representative of either breed probably never would occur 

 in a single individual. 



The values and limitations of Mendelism already are ap- 

 parent to the reader. In the making of new breeds, by 

 crossing those already in existence, it is invaluable; the 

 knowledge of. its workings is of tremendous advantage to 

 the horticulturist. But to the animal breeder who must im- 

 prove his stock by working with what he has, rather than 

 by crossing, its value has not yet been demonstrated. It 

 may be that soon the many investigators will show us how 

 Mendelism inheritance can be recognized in the tiny char- 

 acters which matter to the breeder of fine stock. At present 

 our knowledge is not sufficient to be of use in this 

 field.i 



How, then, are we to proceed in order to achieve the de- 

 sired end ? Since Mendelism in its present state cannot give 

 us what we need, we must find other means. Selection is 

 the force with which we can do most to control heredity. 

 It is by this method that our wonderfully diverse domestic 

 breeds have been produced. 



Variation is of two sorts, continuous and discontinuous. 

 Differences of the first group are almost unnoticeable and 

 are those by which most of the slow but sure progress in 

 the improvement of our breeds is made. The sudden breaks 

 or jumps, known as mutations, or " sports," are the means 



