8 INHERITANCE IN CANARIES. 



form and position, of frizzling, and certain others are not yet ready for 

 a report. The types of color that I have studied most thoroughly are 

 the original green and the yellow. These have been bred pure i7iter se, 

 and with other species. The cinnamon and lizard-color types have been 

 merely touched. As for the crest, this consists of a group of feathers 

 on top of the head arranged so as to radiate from a center between 

 the eyes and forming a small cap covering over the eyes and beak. 

 The questions are: How is the crest inherited, and how is the plumage- 

 color inherited ? 



B. MATERIAL AND METHODS. 



My original stock (1904) consisted of 4 yellow hen canaries (one 

 crested) of the short or German (Harz Mountain) type and 2 green 

 birds of the same type. Also 3 yellow cocks (one crested) and 2 

 greens (one crested). To these each year sundry other canaries were 

 added. These were purchased from a dealer in New York and had 

 been imported from Germany or England. 



The canaries were bred in a room at the Station for Experimental 

 Evolution, each pair being kept in a separate cage. The usual methods 

 of feeding were adopted. All birds bear numbered leg-bands by which 

 their identity is established. 



C. RESULTS. 



Series I.— THE INHERITANCE OF CREST. 

 STATEMENT OF PROBLEM. 



It is my experience with poultry (1906, 69) and pigeons that crest 

 is alternative in heredity and dominates absence of crest. I wished to 

 test heredity of crest in canaries also to see whether the conclusion that 

 I have elsewhere (1906, 87) reached holds, viz., "dominance and reces- 

 siveness depend upon a relation of the characteristics per se and not 

 upon any relation of the races into which they have been introduced." 



MATERIAL. 



Crested and plain Harz canaries were chiefly used, but also a few of 

 the "Norwich" type, which had a flatter and darker crest (plate l.fig. 2). 



RESULTS. 



It quickly appeared that crest is alternative in inheritance, for when 

 crested and non-crested birds were paired, the offspring were either 

 well crested or plain-headed and there were no intergrades. This leads 

 to the hypothesis that crest is dominant as in poultry and pigeons. To 

 test this hypothesis I made a number of matings, of which the detail is 

 given in Section E. 



