In the weathering experiments with the distillate fuels it was 

 fortunately possible to recover most of the evaporated oil by 

 simply syphoning it off the top of the salt water. In the 

 collection of the remainder of the oils by extraction with 

 hexane and subsequent removal of the hexane (to determine the total 

 recovered oil) there were slight additional losses, but experiments 

 indicated that these losses were less than 1% of the extracted 

 portion of the weathered oils. 



A method similar to that used in the present experiments had 

 been used by Smith and Maclntyre [2], These authors had used one 

 tenth as much oil, twice as much salt water, and one third of the 

 airflow. They collected the volatile constituents that were carried off 

 in the airstream by condensing them in a cooled trap. The collected 

 volatile material was dissolved in pentane and subjected to gas 

 chromatography; and, from the area of the oil peaks of the gas 

 chromatogram, the weight of volatile material was calculated. Smith 

 and Maclntyre considered the volatiles to be those components boiling 

 at temperatures not higher than 271°C, the boiling point of pentadecane 

 (C-15). They also sampled oils from controlled spills and from the peak 

 heights determined the losses of individual hydrocarbons. They found 

 that the losses in volatiles were much more rapid in oils spilled 

 at sea, especially when there was wind, and that the volatiles 

 lost at sea included much higher proportions of C-ja and C-j^ hydrocarbons, 



In the present investigation, the weathering was increased by 

 adding the effect of wind produced with a paddle. The airflow was 

 increased also, but increased airflow alone would have much less 

 effect than the use of the paddle. Airflow introduced at the bottom 

 of the flask pushes aside the oil, and increased airflow produces 

 relatively little increased contact. 



Gas chromatography is a commonly used technique in the study of 

 petroleum products, because it gives considerable information about the 

 composition of the relatively volatile components. The components come 

 off the column in order of volatility or boiling points (as shown 

 in Figures 1 to 4) , and the areas under the peaks are proportional to 

 the amounts of the various components. To a lesser degree, the peak 

 heights are also proportional to the amounts present. Both of these 

 relationships have been used for quantitative calculations. Thus, 

 Smith and Maclntyre have calculated the losses of specific hydrocarbons 

 from fuel oils [2] , and Sivadier and Mikolaj have calculated the loss 

 of volatiles from crude oil slicks [7]. 



The data from the gas chromatography of the original and weathered 

 Navy distillate fuels was used to calculate the portion of oil remaining 

 after weathering. For these calculations, the components boiling below 

 298°C were considered to be volatile , and the components boiling above 

 298°C were considered to be nonvolatile. The calculated results for 

 1 -week simulated weathering agree within 2% with the amounts of oil 

 that were recovered, as shown in Table 4. For the longer weathering of 

 Oil C, the deviation was greater. 



