1 882. 139 Trans. N. V. Ac. Sci. 



vapors is obtained, may be again divided into, first, those where the oil- 

 surface is exposed, and called " open testers," and, secondly, those where 

 the oil-surface is more or less covered, and called " closed testers." 



Of these two divisions, the open testers are the first that were made 

 and used. About twenty years ago Tagliabue devised his open petro- 

 leum-tester. It is a very smiple affair, and consists essentially of a glass 

 cup to hold the oil, which fits into a vessel of water that can be heated 

 slowly. Having filled the glass cup with oil, a thermometer is hung m it 

 and the whole is placed in the vessel of water ; the latter is heated, and 

 a small flame is constantly brought near the oil until a blue flash passes 

 over the whole surface ; the thermometer is now noted, and the tem- 

 perature at which the blue flash appears is called the " flashing 

 point " of the oil. 



There are many details that affect the " flashing point " of an oil on 

 this apparatus. Among these, I would mention the fact that rapid 

 heating of the oil will lower the " flashing point," and also that the 

 quantity of oil in the cup and the distance of the ignition-flame from 

 the oil-surface cause the same result. 



While experimenting with this apparatus, I have noted a phenomenon 

 which has a bearing upon the safety of kerosene oil. If the apparatus 

 is used in a very quiet room with no air currents, on approaching the 

 oil-surface with a flame from time to time, it will be noler* that, as the 

 oil becomes heated, a temperature is attained at which the flame used 

 for ignition becomes perceptibly larger and has a bluish outer envelope, 

 which is very distinctly seen. This bluish exterior of the ignition- 

 flame is very characteristic, and is noted many degrees below the 

 flashing point of the oil in this apparatus. A very little thought will 

 show what this phenomenon means. It will be observed that this blue 

 extension of the flame becomes more and more marked until the oil 

 " flashes." It thus appears that, as soon as the flame begms to enlarge, 

 the oil is giving off inflammable vapors, but not in sufficient quantity to 

 produce a " flash " on the surface. The vapors given off before the oil 

 " flashes " are so light and easily diffusible that they are lost before they 

 can form an explosive mixture with the air, for the flash is nothing 

 more or less than a miniature explosion upon the oil-surface. Not until 

 the vapors become so dense that they cannot diffuse rapidly, will the 

 oil flash. The importance of this fact will be apparent later in this 

 paper. 



In testing oil with open testers, efforts have been made to produce 

 uniform results by filling the oil-cup always to the same level, and alsa 

 by various devices to obtain an ignition-flame of a given size and at a 

 given distance from the oil surface. The ignition-flame has been re- 

 placed by the electric spark, as in the apparatus of Saybolt ; but 



