714 
a aes 
of all gauge ratios by an amount greater than three times the standard de- 
viation from the mean, that ratio was considered suspect and gauge readings 
for the two shots were examined individually for possible errors, Usually 
suspect values found in this manner were also suspect on the basis of the 
analysis of percentage difference between the two gauges of a given pair.) 
The mean of the ratios so obtained gave the average difference between the 
two explosives for an equal volume of charge measured at equal distances 
from the charge. This particular type of comparison may or may not have 
been the one desired for a particular application but other types of ratios 
could be calculated as discussed below provided the weight and distance ex-— 
ponents and the densities of the various explosives were known, 
When secondary variables were being studied it was necessary to Zo 
through similar analysis comparing the results from the same explosive be-— 
tween the different strings, When it was not desired to study the effect 
of secondary variables it was the practice to change the charge-to-gauge 
distances betwee pee Tes so as to give a good logarithmic distribution of 
the values of WL 3h, where Wis the charge weight and R is the charge-to~ 
gauge distance. This allowed a well-distributed series of points when the 
results were plotted on log-log paper. 
The standard deviations calculated in the above analyses gave an esti- 
mate of the significance of the results obtained, The 5=—percent significance 
level was usually employed at UERL (a significant difference between two 
values must exceed twice the standard deviation of the difference between 
these values). This level of significance was obtained by using from )0 to 
50 mechanical gauges and 8 piezoelectric gauges on two or three strings, and 
it was usuatly possible to distinguish between two explosives which differed 
by as little as 2 percent, By increasing the number of gauges on each shot, 
or by increasing the number of strings it would be possible to obtain a 
higher level of signiftcance or make closer distinctions. However the large 
increase in the number of gauges or number of shots necessary to appreciably 
raise the significance level was not deemed warranted for this work. 
(b) Conversion from one type of explosive ratio to anaes -- The 
directly determined experimental ratios from the RELIANCE work were usually 
the ratios of the sauge readings for cqual volumes of explosives at equal 
charge-to-gauge distances (Ova). (For convenience, ratios are designated 
in the form vyg, where the létter on the line indicates the ratio and the 
subscripts indicate the quantities which remain constant. In particular, 
Dya means the ratio of the sauge readings D for equal volumes of explosives 
V at equal charge-to-gauge distances de) For certain purposes it was de- 
sirable to compare explosives on a different basis. For example, in air=— 
craft where the pay load is the limiting factor it would be desirable to 
compare the explosive powers from equal weights of explosives at equal 
distances (Dyq)e Also it is often desired to determine the weight of a 
particular standard explosive that is necessary to do the same damage at 
the same distance from the charge as a unit weight of some new explosive. 
The ratio of the weights (Wpq) is referred to as "equivalent weight." 
