The density of the agitated water coliJimi in Tahle II is greater 

 than the density to at least 15 meters in Table III; therefore, work 

 was performed by the system in raising water through a vertical dis- 

 tance in excess of 15 meters. Comparison of salinity and density 

 data of Table II and the plotted curves of Figure 6 reveals that 

 water similar to the entire water column at Station 1 is found at 20 

 meters at Station 3, indicating that the water was raised at least 

 20 meters. 



The eventual cooling of the entire water column to the freezing 

 point indicates that vertical transport of sensible heat from depth 

 was not a factor in the maintenance of the artificially created 

 polynya, except during the initial stage of the ice formation. Con- 

 sequently, an understanding of the physical process involved must be 

 sought along other lines. 



Elementary ice particles are probably disk-shaped and devoid of 

 crystalline form. Ordinarily they flocculate and grow into true 

 crystals. The turbulent energy of the induced currents may destroy 

 the crystals before they enlarge or may effectively prevent crystal- 

 line growth about ice nuclei. Ice particles at the surface of the 

 bubbled area are rapidly swept from regions of divergence into 

 regions of convergence where, by means of descending currents, they 

 are transported beneath the sirrface to be eventually dispersed from 

 the polynya area. 



I^^drographic Office ice observers, stationed at Thule Air Base 

 throughout the winter of 1959-60, noted that the polynya gradually 

 narrowed; by the end of December width ranged from 12 feet at the 

 eastern end to 50 feet at the western end, where an auxiliary air 

 hose was used to augment the bubbling activity by inducing more vig- 

 orous currents. Dimensions of the ice-free area gradually increased 

 during spring as the air temperature rose to approximately 0° F, 



A plot of sea ice tensile strength versus temperature (Assur, 

 1958) shows a marked increase of strength as the temperature of the 

 ice drops below -9.2° F. At this temperature sodium chloride is 

 precipitated from the brine pockets in the ice. During periods of 

 extremely low air temperature in winter, the weakest point of the 

 ice should be at its undersurface where the temperature approaches 

 that of the water. 



Measurements made during April i960 show that ice thickness 

 directly above one of the polyethylene pipes averaged approximately 

 10 inches while thicknesses ranged between ij-l and kk inches at 

 locations 60, 200, and 375 yards north- northeast of the pier. Abra- 

 sive action of induced currents apparently inhibited ice growth in 

 zones of most vigorous flow. 



