ON INBREEDING IN JERSEY CATTLE. 



653 



Since a bull contributes no x chromoaome to his sons, all contribution of a sex 

 linked factor to the sons' progeny must be traced through their dams. Therefore 

 the paternal grandparent contributes no sex linked character to his granddaughters. 



It will be noticed that half of the sire's pedigree makes no contribution, however 

 good it may be. This is the side that is frequently most emphasised. Altogether, 

 in the fourth parental generation only half the ancestors need be considered in the 

 examination of a pedigree from this point of view. While, in a way, the sire's contribu- 

 tion to the heifer is the more important, it is only so because it is the more definite. 

 The dam has a greater accumulation of blood lines to draw upon, and if there are 

 several sex linked factors involved, may certainly make the bigger contribution on 

 the average. 



On making a close study of the work of Gowen and others at Maine in search of 

 facts which might disprove this hypothesis the writer was unable to find anything 

 definite in this direction. On the contrary Gowen, in his book (1924) dealing with 

 American Holstein Friesians from the Advanced Registry, gives figures showing the 

 correlations between daughter and parents and between daughter and grandparents. 

 Table II tabulates his results, and is drawn chiefly from pages 155, 188, 224, 252, 

 300, 309, 319 and 327. 



Table II. 

 Showikg Corbelation Coefficients of Heifees with their Parents and 



Geandpaeents (from Gowen). 



Thus, while the parents contribute about equally, there is considerable variation 

 in the grandparents, that of the parental grandsire being considerably and significantly 

 less than the other grandparents. That the correlations are small does not greatly 

 matter. The difference exists, and, as far as the writer is aware, Gowen has made 

 no attempt to explain it. This table may be represented diagrammatically in Figure V. 



FiGITEE V. 

 Showing correlation of milk yields with parents and grandparents from Table II. 



This squares fairly well with Figures III and IV, showing how sex linked characters 

 are inherited and how the paternal grandsire has no influence in the matter. 



Table III shows the correlation figures for cousins by the same grandparents 

 grouped according to the specific grandparent. This forms a useful confirmation of 

 the previous table. The effect of the paternal grandsire in this case is less than the 

 probable error, and may therefore be considered to be nil. The figures for this table 

 are also taken from Gowen's work. For interest the correlations of full sisters and 

 dam to daughter are also added. 



