114 DOMESTICATED ANIMALS AND PLANTS 



taken to, say, one quarter inch and the weights more accurately ? 

 The answer is that experience has shown that these are suffi- 

 ciently accurate, and if the measurements had been taken finer, 

 say to the quarter inch, the labor of calculating would have been 

 doubled, and all without altering or improving results in any 

 substantial way. What, now, is to decide the question as to 

 accuracy of measurements ? 



Speaking generally, the object is not so much to get accurate 

 measurements of all the individuals as such, as it is to make 

 them comparable one with others ; and it will be found by trial 

 that measurements taken to the half inch in length of ears of 

 corn, for example, will give practically the same results as those 

 taken at a closer measure, as one fourth or one eighth of an 

 inch. Not only is this true, but it is practically impossible to get 

 the length of an ear of corn correctly within an eighth of an 

 inch, as will be found by trial. 



On the other hand, if we should take the measurements only 

 to one inch, they would be too far apart for smooth distributions. 

 The best easy test of the measurement to be chosen for "class " 

 grouping, as it is called, is whether it gives a fairly " smooth " 

 distribution. A glance at the distribution of length or of weight 

 of ears will show that the figures slope off each way from the 

 middle at a fairly uniform rate without any sudden break and 

 without any number being greater than its neighbor nearer the 

 middle. This is the best test of sufificient accuracy. In order to 

 save labor the measurement will be taken as " coarse " as 

 possible, but not so coarse as to break up the smoothness of the 

 distribution or to make the groups too few. A little experience 

 soon develops a judgment at this point which is better than 

 anything that can be learned by instruction ; but with all the ex- 

 perience of experts some trials have to be made whenever a 

 new problem is taken, in order to determine the most desirable 

 " scheme of measurements." 



Suggestions as to grouping. After the scheme of measure- 

 ment has been decided — as inches, half inches, ounces, pounds, 



