15 



5 0-5 -10 -15 -20 -25 -30 -35 

 AWT/WT SEA X lo"^ 



A IN AIR, WITH CONTAMINATED EXTERIOR 

 □ COVERED, WITH CLEAN EXTERIOR 

 O IMMERSED IN TAP WATER 



Figure 38. — Ratio of weight change to initial weight against salinity change. 



points are shown for groups 1 and 3. No line is shown for those in 

 group 2, since only two points are considered. It would be expected 

 that the points in group 2 should fall near those in group 1, which is 

 the case. The slope of these lines is approximately 1/100, which is 

 construed to mean that the change of salinit}^ is about two orders of 

 magnitude less than the change in weight. 



The initial test of the wash water by the flame photometer for 

 sodium content on the exterior of the polyethylene bottles showed a 

 blank contamination on the order of 0.05 to 0.10 part per million. 

 On all the remaining tests there appeared to be no greater concentra- 

 tion than found in the blank. If the weight loss were a result of the 

 salt solution rather than pure water passing through the walls of the 

 bottle, the concentration of sodium, considering the volume of wash 

 water, should have been about 4 to 6 parts per million. 



The environmental conditions within the storeroom varied to some 

 extent. The experiment began in the winter and ended in the summer. 

 There was a general increase in relative humidity and temperature 

 between ^=153 days in early May and ^=232 days in late July; this 

 most probably varied the osmotic effect. 



On the assumption that only pure water passes through the walls of 

 the bottles, the resultant sahnities based on the change of weight were 

 computed and are shown in table XIII. The average difference 

 between the calculated and the observed salinities for those samples 



76 



