312 Journal of Agricultural Research voi. xvii, no. 6 



4. It is of some interest to examine the weighted mean value of the 

 skewness for all the curves, the weighting being in proportion to the 

 number of individuals involved, in comparison with the skewness ex- 

 hibited in the variation curves of other characters. AVe have for the 

 weighted mean value of the skewness for mean weekly yield, the 11- 

 year curve being omitted, a value of +0.1047. ^o'' the variation curves 

 for fat content the weighted mean value of the skewness is +0.1338. 

 It was shown by Pearl and Surface (75) that in variation in annual 

 egg production in Barred Plymouth Rocks the skewness is always 

 negative and usually significa.nt. This difference in skewness between 

 the two characters milk production and egg production is striking. 

 Curv'es of variation in egg production tail off more on the side toward 

 low egg production, whereas the curves of variation in milk production 

 tail off more on the side toward high production. The weighted mean 

 values of the skewness for annual egg production in three successive 

 years were found to be —0.280, —0.122, and —0.108. In other words, 

 the values in general were of an order of magnitude not far from that 

 here found for the skewness of curves of variation in milk production. 



5. It might at first thought be supposed that the direction of the 

 skewness in milk productive curves was due to selection — that is, to the 

 continued culling out of the poor producers. Since, however, the same 

 factor of selection in the direction of the high producers was operative 

 to as great or even a greater extent in the making up of the flocks from 

 which the annual egg production variation curves were obtained, it seems 

 perfectly clear that selection can have had very little to do with bringing 

 about the difference in direction of skewness exhibited by egg and 

 milk production curves respectively. The inference would then seem 

 strongly justified that selection had nothing to do with the production 

 of the asymmetry of the variation curves in either case considered by 

 itself. 



6. Additional interest is given the matter when an examination is 

 made of the facts regarding the direction of the skewness in the varia- 

 tion of the hen's egg in size characteristics. vSuch data have been fur- 

 nished by Pearl and Surface (/6). They show (/?. 184) that in the 

 variation of egg length, egg bulk, egg weight, and egg breadth the skew- 

 ness is positive and significant in all cases except that of breadth. We 

 see here again emphasized a point which comes out frequently in bio- 

 metrical work — namely, that there is frequently between characters 

 a parallelism in variation corresponding to a parallelism in the 

 underlying physiological bases of the characters. This relation 

 is clearly apparent in the present instance. The size of the egg is 

 primarily determined by the secretory activities of the oviduct. It is 

 a character which is physiologically much more directly comparable 

 to milk production than is total annual production. Primarily the 

 latter depends physiologically upon quite another thing — ^namely, the 



