ANIMAL PRODUCTION. 369 



the author, was a specially constructed differential calorimeter provided with 

 sensitive electrical resistance thermometers. 



Dried and powdereil preparations of muscles and other tissues from cold- 

 blooded (fresh- water mussel) and warm-blooded (rabbits and pigeons) animals 

 were brought into contact with water in the calorimeter and the heat given off 

 during the process of imbibition (real molecular swelling) was measured. It 

 was found that all the materials tested were liable to imbibition, the process 

 being attended by a considerable development of heat. Generally speaking, 

 muscle tissue developed greater heat than kiudney and liver. 



Differences between tissues from different types of animals were less marked, 

 though noticeable. The sensibility of the method is said to be very great, one 

 division mark on the measuring wire representing an average of 0.001 or 0.002 

 gm.-calorie. 



The work is of technical interest as suggesting now methods of research into 

 the nature of some of the more obscure changes occurring in body tissue, and 

 possibly into the action of digestive juices, etc. 



Direct calorimetry of infants, with a comparison of the results obtained 

 by this and other methods, J. Rowland (Trans. 15. Intemat. Cong. Hyg. and 

 Dcmogr. Washington, 2 {1912), Sect. 2, pp. 438-451, fig. 1). — A description of* 

 the methods used and the results obtained in respiration calorimeter exiieri- 

 meuts with infants under various conditions of nutrition. 



ANIMAL PRODUCTION. 



The element of uncertainty in the interpretation of feeding experiments, 

 H. H. Mitchell and H. S. Geindley {Illinois Sta. Bui. 165, pp. 463-579, pi. 1).— 

 In this bulletin the authors point out that many simple feeding experiments are 

 of more or less ambiguous significance because of the dissimilarity which exists 

 among the gains of individual animals, due to variable, uncontrolled, and 

 largely unknown experimental conditions. Methods are proposed for "dealing 

 with the question in a systematic and rational manner, so that the sphere of 

 uncertainty surrounding the conclusions based on experimental results will be 

 reduced to a minimum and be defined as clearly as possible." 



It is first shown that all attempts to predict the result of repeating an 

 experiment must be based on the "tendency of comparable experimental data 

 to assume a definite frequency distribution, expressible by a frequency curve 

 capable of mathematical definition." It is further explained that " an average 

 is at best only an imperfect description of a series of experimental data, and 

 when used for comparative purposes is often extremely misleading." The use 

 of a factor, known as the standard deviation, is advised in comparing the gains 

 exhibited by 2 different lots of animals and may be defined as *' the square root 

 of the average squared deviation of all individual gains from the average gain 

 for the lot." On the basis of this standard deviation, the probable error to be 

 encountered in predicting the result of repeating a feeding experiment may be 

 determined. By the use of such a probability method it is possible to interpret 

 the results of feeding experiments in a fairly satisfactory manner. 



For extensive comparisons of variation " the coefficient of variation is used, 

 this coefficient being simply the standard deviation calculated at a i)€rcentage of 

 the average. The coefficient of variation of gains within lots is a good measure 

 of the experimental error." From an extensive review of experiment station 

 literature in this country it is found that the average coefficient of variation 

 gains for similarly treated lots of sheep is about 21, for steers and swine about 

 17, and for poultiy about 16 per cent. 



Calculations indicate that experimental lots should contain at least 10 to 14 

 animals, or even 25 to 30 animals, when the rations or other conditions under 



