INBRED AND HYBRID ANIMALS AND THEIR VALUE 335 



We can now turn to examples. 



Decreased variation following inbreeding. — ^This is the more commonly 

 observed result. At the moment we need mention only a few examples, 

 such as reduction of variation in: intensity of coat color, amount of white 

 spotting, tissue speciticity, and reaction to bacterial inoculation. The 

 degree of reduction differs widely and is dependent on the relative impor- 

 tance of heredity and environment in determining the variation in the 

 foundation stock. Thus, variation in tissue speciticity seems to be deter- 

 mined mainly by genetic factors, for it is greatly reduced by inbreeding. 

 In fact, if this character is measured by percentage of "takes" in transplants 

 between individuals, there is commonly no variation left at all in an inbred 

 line, all the transplants being successful. On the other hand, variation in a 

 character like white spotting may be determined largely by the environ- 

 ment and, therefore, not greatly reduced by inbreeding. Wright and Chase 

 (22) measuring white spotting in the guinea pig on an appropriate scale 



to a point at which it is more sensitive to the environment than are the bulk of the 

 individuals in the foundation stock. On a scale on which environmental effects are 

 equal at all points, variation cannot increase as a result of inbreeding. A natural 

 scale of this type, with its simple logical relation to the causes of variation, is to be 

 preferred; and when a character does not fall easily into one it is sometimes possible 

 to devise such a scale and transform the data to it (16). If all characters could be 

 expressed in these terms the question of increased variation following inbreeding 

 would not arise. For many characters, however, no such scale has been found, and 

 in some of these there is reason to e.xpect that it would be too complicated for practical 

 purposes. In these cases we can only use the descriptive scales available. It must 

 be remembered that a measure of the variation on these scales, though it may be of 

 descriptive value, does not have the analytical value of statistics derived from data 

 recorded on natural scales. 



One of these scales is necessarily quite common in biology because of the frequent 

 occurrence of physiological thresholds in development. On one side of the threshold 

 the character is recorded as "normal," with no variation, while on the other side the 

 character falls into a graded series of ''abnormalities." Most of the examples of 

 increased variation following inbreeding given later probably involve physiological 

 thresholds, the random bred stock falling mostly, or entirely, on the "normal" side 

 of the threshold, and the inbred strain falling largely, or completely, on the " abnormal " 

 side. 



When a character is recorded only in two categories (e.g., 3-toed and 4-toed, 

 tumorous and non-tumorous, infected and not infected) an inbred strain is to be 

 regarded as more variable than its foundation stock if it falls closer to a 50:50 dis- 

 tribution in the two categories. Here, however, no particular value is obtained by 

 speaking of the "variation," for the distribution of data as recorded can be described 

 completely simply by stating the percentage in either category. 



