214 Bulletin 307 



The question, therefore, whether the acre or the orchard is to be used 

 as unit, resolves itself into a question of the purpose of an orchard survey. 

 If it is to obtain a census of the existing conditions of the industry, the 

 yield and income per acre should be used. The theoretical value of 

 different methods of orchard management, however, may be better ex- 

 pressed by using the orchard as unit. 



The amount of difference in the yield per acre and the mean yield per 

 acre per orchard depends on the uniformity in the size of orchards. In 

 cases in which the size of orchards is fairly uniform, the difference between 

 the two means is negligible. When some of the orchards are large and some 

 small, the influence of the size of the orchard is shown in the mean. 



Value of standard deviation and coefficient of variation in orchard surveys. — • 

 Further research in orchard surveys would be of value in establishing the 

 degree of risk and the probability of the mean values being reached under 

 various types of orchard management and for various kinds of fruits. 

 Such figiires would be of value from the standpoint of investment, both 

 for those who have already invested and for prospective orchardists. 

 For such purposes the standard deviation is of value. It is a measure 

 of the absolute variability of yields and incomes of the orchards of the 

 group for which it was determined, and shows, also, the concentration 

 of the values about the mean. In this way, the probability of the mean 

 value and deviations from the mean are indicated. A deviation of one 

 half the standard occurs with a frequency of 61.7 per cent of the cases; 

 a deviation greater than the standard in 31.7 per cent ; a deviation of twice 

 the standard occurs in comparatively few cases; and one of three times 

 the standard is very rarely found. 



When it comes to a question of the comparative value of the different 

 types of orchard management, or of different fruits, the coefficient of 

 variation is of value. Such a measure of relative variability would answer 

 the question concerning the comparative profit in raising peaches or apples. 



BIBLIOGRAPHY 



Biometrika. 



Davenport, C. B. Statistical Methods. 



Davenport, Eugene. Principles of Breeding. Ginn & Co. 1907. 



Galton, Sir Francis. Natural Inheritance. The Macmillan Company. 



1889. 

 Johnson, W. W. The Theory of Errors and Method of Least Squares. 



John Wiley &. Sons, New York. 1893. 



