356 



Journal of Agricultural Research 



Vol. XIII, No. 7 



permanence. The joints of these desiccators were covered with wax; 

 but even under the best conditions leakage could not be prevented 

 entirel}^ and it was found necessary to pump out the desiccators several 

 times a year. To obtain the effect of air (series C, F, G, and I) the glass 

 stopcocks of the desiccators were replaced by perforated rubber stoppers 

 and straight glass tubes which passed through the stoppers and dipped 

 into the sulphuric acid (series C and F) or into the water (series G and I). 

 To exclude light (series A, B, C, E, and G), the desiccators were placed 

 in a large oblong wooden box, lined with building paper, with an over- 

 hanging cover similarly lined. The cover was held by corner posts 0.5 

 inch above the top of the box, projected 0.5 inch beyond the sides of the 

 box, and overlapped 3.5 inches. In addition a strip 1.4 inches wide was 

 nailed to the outside of the box 0.5 inch below the edge of the cover. 



Fig. I— Apparatus used in the experiments to determine the stability of olive oil. 



This provided a continuous air passage 0.5 inch wide under all sides of the 

 cover and yet absolutely prevented the entrance of light, even by 

 reflection. To obtain the effect of light (series D, F, H, and I) the 

 desiccators were placed on the cover of the box and exposed to light 

 from a north window. All the samples were kept in the northeast room 

 of the Experiment Station dairy building, into which the direct rays of 

 the sun did not enter at any season of the year. The temperature was 

 not constant, but relative in all cases. Considerable time was consumed 

 in obtaining the necessary supplies and in preparing the samples, so that 

 the experiment did not actually begin until April 2, 1910. 

 EFFECT OF AIR, LIGHT, AND MOISTURE 

 ORGANOLEPTIC CHANGES 



Changes of an organoleptic character are difficult to measure and even 

 more difficult to express, particularly where the differences are slight. 

 The results are relative, however, if not strictly accurate, and are 

 recorded in Table I. 



