IOWA ACADEMY OF SCIENCE 
183 
inating oil without burning it and presents a most interesting relation 
between the illuminating power and the physical tests of the oils. In 
the 28 samples the average variation of the illuminating power from 
that shown by the curve is 1.3 per cent. When it is remembered that 
measurements of candle-power within one per cent requires very good 
facilities, the accuracy of the curve becomes more remarkable. 
Having given this very striking result, the question very naturally 
arises as to the correctness of the curve. The majority of the oils tested 
were from samples of oils sold in North Dakota, but three of them were 
obtained from independent refineries in Pennsylvania, Ohio and Kan- 
sas, and 13 of them we obtained on the market in the states of Minne- 
sota, Illinois, Iowa and South Dakota. Although the experiments do 
not demonstrate the limitations of the application of this curve, yet it 
is not expected that the results can be applied to oils coming from fields 
which are widely different. . In fact, the agreement obtained in these 
experiments is very surprising. It is fair to say that the writer does 
not lay much stress upon the numerical results obtained in this inves- 
tigation. In his opinion, the greatest value of these experiments is 
that they demonstrate beyond a doubt that there are possibilities in a 
study by physical methods which have not been appreciated hereto- 
.fore. 
