178 
JOURNAL OF ECONOMIC ENTOMOLOGY 
[Vo], 17 
sense, fortunate and, in another sense, rather unfortunate. By a broad 
use of the word “estimating” we have drawn the fire of the statisticians. 
In the Survey’s work it has become increasingly more evident that the 
estimates we are receiving from our collaborators lose much of their 
potential value in that; (a) they are not based upon sufficient nor de¬ 
tailed enough observations and (b) that each reporter is forced to adopt 
original methods of estimating insect abundance which are usually not 
comparable with the estimates of other collaborators. I do not wish 
to be construed as criticizing our collaborators as we have gained much 
in the past three years of the Survey’s work in locating the areas of intense 
damage of many of our more important insect pests and wish our collab¬ 
orators to continue to send in miscellaneous observations of insect 
abundance as there is still much to be done along this most profitable 
« 
line, and it is probably the only line which we can pursue to advantage 
with our limited funds and organization, but the ultimate object of 
the Survey is to use this mass of data in statistical entomology and, in 
order that it may be used for this purpose, it must meet with the re¬ 
quirements of the statistician, that is, the requirements of mathematical 
common sense. 
I wish to digress here for a moment to impress upon you the ridiculous¬ 
ness in carrying out figures in reports which are arrived at by statistical 
methods to such a point that they indicate the most accurate mathe¬ 
matical precision. An amusing report was recently brought to my 
attention which eminated from a fish hatchery and gave, after a wordy 
report, figures similar to the following: indicating that during the past 
year 13,280,642 fish and fish eggs had been distributed by the hatchery. 
We do not wish to perpetuate the old inductive method whereby an 
entomologist, by virtue of his experience and familiarity with the sub¬ 
ject, would go into the field and estimate (in the sense of guess) at the 
percent of damage being done or the relative number of insects present. 
These guesses were more or less accurate, usually less, but, even if we 
disregard the personal equation, they are absolutely useless for statistical 
work. Even a so-called statistical survey on insect abundance, arrived 
at by actual count of presumed sample units, is often based on counting 
and calculating processes and samples which introduce not only a 
possible error but a probable error, which is greater than the results 
obtained. 
One of the most important things that we will have to pre-determine in 
this work is a definite standard of accuracy and every record should be 
brought up to this standard. The data which we use must be based 
