August 26, 1921] 



SCIENCE 



173 



absolute surface tension values in terms of 

 dynes. Such an instrument with its proper 

 formula might well be of service in such meas- 

 urements because of its simplicity of manipu- 

 lation. 



Frederick S. Hammett 

 The Wistar Institute op 

 Anatomy and Biology 



variation of individual pigs in economy 



OF GAIN 



In a very interesting article by Ashby and 

 Malcomson, published in the Journal of Agri- 

 cultural Research, Volume XIX., pages 225- 

 234, the following statement is made on page 

 232: 



The resultant coeflScient of correlation 



J- = — 0.452 ± 0.068 



shows a distinct negative correlation, between rate 

 of gain and economy of gain, entirely disproving 

 the apparent relation shown by Tables IX. to XV. 



This conclusion is very interesting, espe- 

 cially since it is contrary to the usual belief 

 and usual experience. The writer has re- 

 calculated the coefficient of correlation from 

 Table XVI. on page 232 and found a different 

 result from that given by Ashby and Malcom- 

 son in that r = — 0.166 ± 0.083 which is not 

 a significant correlation. 



Thinking that a different treatment of the 

 data might throw further light on this point, 

 a new correlation table was made between the 

 rate of daily gain and the amount of feed re- 

 quired to produce 100 pounds of gain. This 

 correlation was found to be 



r = + 0.140 ±0.083. 



This is not a significant correlation, but it is 

 interesting to note that it is positive instead 

 of negative. 



Again Ashby and Malcomson combined 

 cases of animals fed on pasture and of those 

 fed in the dry lot. From a statistical point of 

 view, this is not advisable, since the food de- 

 rived from the pasture was not taken into 

 consideration. 



The average daily gain of 27 animals fed on 

 pasture was 1.14 pounds and the average 

 amount of feed required to produce 100 pounds 



of gain was 361.2 pounds. The average daily 

 gain of 36 animals fed in the dry lot was 1.41 

 pounds and 391.8 pounds of feed were required 

 to produce 100 pounds of gain. It is readily 

 seen from these figures that the rate of gain 

 among pasture fed animals was less than 

 among those of the dry lot, and at the same 

 time the amount of feed for 100 pounds of 

 gain was less, because the part of the feed 

 from the pasture was not taken into con- 

 sideration. 



Separate correlations were found for these 

 two groups. In the pasture-fed group 



r = — 0.181 ±0.126, 



while in the dry lot group 



r = — 0.036 =±: 0.112. 



Both of these are negative but not significant. 

 When treated together 



r= + 0.140 ±0.083, 



a positive correlation but still not significant. 

 One can not accept the conclusion of Ashby 

 and Malcomson that there is a negative cor- 

 relation between the rate of gain and economy 

 of gain for the following reasons: 



1. On the basis of their own data, there is 

 no significant correlation. 



2. From a statistical standpoint it is not 

 legitimate to pool cases of animals fed on 

 pasture and animals fed in the dry lot, for the 

 purpose of determining the correlation be- 

 tween rate of gain and economy of gain since 

 the two groups are dissimilar. 



Other factors which might influence the 

 results are initial weight, age, length of feed- 

 ing period and methods of feeding. In their 

 discussion, the possibility is suggested of 

 using these individual differences as a basis 

 for selecting strains which are more econom- 

 ical producers. But these variations which 

 they found can not be said to be genetic be- 

 cause of too many uncontrolled factors such 

 as cited above. It is interesting to note that 

 even when pigs are self-fed, selection is ex- 

 ercised by individual pigs in the kinds of 

 feed which they consume. The following 

 table taken from a preliminary report by Ashby 

 on page 201 of the 1916 Proceedings of the 



