October 17, 1913] 



SCIENCE 



537 



the "Pigmentation of Bacon Fat" and on 

 the spajring of sows for fattening, have 

 found the great value of consultation with 

 the staffs of several large bacon factories. 

 There seems to be in this a general lesson 

 that before taking up any problem one 

 should get into touch not only with the 

 producers, but with the buyers, from whom 

 much useful information can be obtained. 



I feel that Biffen's work has borne fruit 

 in still another direction, for which perhaps 

 he is not alone responsible. Twenty years 

 ago agricultural botany took a very sub- 

 sidiary position in such agricultural exami- 

 nations as then existed. In some of the 

 agricultural teaching institutions there was 

 no botanist, in others the botanist was only 

 a junior assistant. It is largely due to the 

 work of Biffen and the botanists at other 

 agricultural centers that botany is now re- 

 garded as perhaps the most important 

 science allied to agriculture. 



I must here repeat that I am not attempt- 

 ing to make a complete survey of all the re- 

 sults obtained in the last twenty years. My 

 object is only to pick out some of the typical 

 successes and failures and to endeavor to 

 draw from their consideration useful lessons 

 for the future. So far I have not referred 

 to the work which has been done in the 

 nutrition of animals, and I now propose to 

 conclude with a short discussion of that 

 subject. The work on that subject which 

 has been carried out in Great Britain dur- 

 ing the last twenty years has been almost 

 entirely confined to practical feeding trials 

 of various foods or mixtures of foods, trials 

 which have been for the most part incon- 

 clusive. 



It has been shown recently that if a num- 

 ber of animals in store condition are put 

 on a fattening diet, at the end of a feeding 

 period of twelve to twenty weeks about half 

 of them will show live-weight increases 

 differing by about fourteen per cent, from 



the average live-weight increase of the 

 whole lot. In other words, the probable 

 error of the live-weight increase of a single 

 fattening ox or sheep is fourteen per cent, 

 of the live-weight increase. This being so, 

 it is obvious that very large numbers of 

 animals must be employed in any feeding 

 experiment which is designed to compare 

 the feeding value of two rations with rea- 

 sonable accuracy. For instance, to measure 

 a difference of ten per cent, it is necessary 

 to reduce the probable error to three per 

 cent, in order that the ten per cent, differ- 

 ence may have a certainty of thirty to one. 

 To achieve this, twenty-five animals must be 

 fed on each ration. Those conversant with 

 the numerous reports of feeding trials 

 which have been published in the last 

 twenty years will agree that in very few 

 cases have such numbers been used. We 

 must admit then that many of the feeding 

 trials which have been carried out can lay 

 no claim to accuracy. Nevertheless, they 

 have served a very useful purpose. From 

 time to time new articles of food come on 

 the market, and are viewed with suspicion 

 by the farmers. These have been included 

 in feeding trials and found to be safe or 

 otherwise, a piece of most useful informa- 

 tion. Thus, for instance, Bombay cotton 

 cake, when first put on the market, was 

 thought to be dangerous on account of its 

 woolly appearance. It was tried, however, 

 by several of the agricultural colleges and 

 found to be quite harmless to cattle. Its 

 composition is practically the same as that 

 of Egyptian cotton cake, and it now makes 

 on the market practically the same price. 



Soya-bean cake is another instance of a 

 new food which has been similarly tested, 

 and found to be safe for cattle if used in 

 rather small quantities and mixed with 

 cotton cake. The price is now rapidly ris- 

 ing to that indicated by its analysis. Work 

 of this kind is, and always will be, most 



