206 
JOURNAL OF ECONOMIC ENTOMOLOGY 
[Vol. 17 
purpose might not be defeated or perhaps served just as well if the nature 
of their dependability was recognized, admitted and furnished along with 
the figures. 
This brings up an old subject about which many books have been 
written and in which renewed interest has been taken in various fields of 
research within the last few years. It is desired to call attention to the 
disregard, especially by Entomologists, of the Theory of Probability, 
which in all available texts examined has been treated as indispensable to 
the application of Statistical Methods. 
Theory of Probability 
While, as has been stated, the earliest workers on Hessian fly em¬ 
ployed some sort of Statistical Methods in estimating its abundance 
and damage, there does not seem to be a single reference of an attempt to 
apply the Theory of Probability to determine the reliability of such 
methods. As a matter of history, the Theory of Probability is older 
than that of Statistics, and has now been developed to the place where 
its application lends a powerful tool to investigations in many lines of 
research. By the application of one of the principles of this theory the 
measure of unreliability of the determination of any value is given by 
the probable error of the determination. This is usually expressed in 
the form of a pair of values, one above and the other below the value 
determined, the chances being even that the true value lies between 
these limits. 
By the application of the Theory of Probability to methods of de¬ 
termination of the abundance of and damage by Hessian fly, it seems 
quite likely that data from which heretofore definite conclusions have 
been drawn really should have been taken merely as indications more or 
less reliable as the particular ease may have justified. Often per¬ 
centages of infestation and parasitism have been recorded without 
giving the probable error or any description of the method of deter¬ 
mination. All too frequently such percentages are expressed in frac¬ 
tions, sometimes even hundredths of one percent, entirely disregarding 
any error of method which doubtless exists to the extent of several 
percent. Such practices are often not only the basis of unfounded 
conclusions, conflicting recommendations and hair splitting estimates, 
but are in themselves misleading in that they assume a non-existent re¬ 
finement of method. 
