to seal them from further contamination. These bottle outsides 

 were later to be washed, and the wash water was to be analyzed for 

 salinity, hopefully to reveal if salts were being passed through the 

 walls and were collecting on the outside. 



Group 3. The remaining six polyethylene bottles were stored in a 

 second sterilizer, immersed in tap water to the same level as the liquid 

 in the bottle and covered with a tight-fitting lid. This is the extreme 

 case of water condensation simulation and was to indicate if water 

 crosses into the bottle from osmotic pressure. 



The bottles in group 1 were washed, inside and outside, with sea 

 water and the exteriors were allowed to dry by natural evaporation. 

 The bottles from groups 2 and 3 were washed thoroughly on their 

 outsides with tap water and then with distilled. One bottle from 

 group 2 was also washed on its inside and then filled with distilled 

 water. This was placed in an SOO-ml. beaker containing distilled 

 water to the level of the water on the inside of the bottle. The 

 bottle was agitated, allowed to soak 10 minutes, then agitated again. 

 The wash water was run on a Woods Hole Oceanographic Institution 

 flame photometer (Spencer and Woodcock, 1963) to determine if 

 there was any sodium and thus any residual salts on the outer surface 

 prior to filling the inside with sea water. 



The polyethylene bottles were weighed empty (with caps) after 

 washing and then weighed again after filling with sea water. All 

 weighings except the final ones at ^ = 232 days were done on a torsion 

 balance scale; the exceptions were weighed by a Mettler Type K-7 

 scale. An inter-comparison was made between the two types of 

 scales to determine the corrected value for the final weighings. The 

 accuracy of the weighing is considered to be ±10 mg. 



The polyethylene bottles exposed to the atmosphere, those in the 

 two sterilizers, and the citrate bottles were kept in a storeroom with 

 an average relative humidity of about 45 percent. Tap water was 

 added at intervals to the sterilizer containing group 3 to maintain 

 a semiconstant level. 



Salinity determinations were made by an oil-bath conductivity 

 bridge of the Woods Hole Oceanographic Institution, with a precision 

 of ±0.003°/oo. The day after fiUing all bottles the first glass citrate 

 bottle was run to determine the salinity for ^ = 0, which was 

 34.700°/oo. The time interval for examining the salinities and weight 

 changes was scheduled for approximately 1, 2, 3, 4, 6, and 8 months 

 after t=0. 



Results and Discussion. The bottles in groups 1 and 2, after an 

 apparent initial loss of salinity, showed an increase in salinity with 

 time and an associated weight loss (fig. 37 and tables XI and XII). 

 The one exception to this was the final bottle of group 2 (not plotted, 

 but see table XI), which showed a relatively large increase of weight. 



74 



