manner as previous years. There may be some question on the oper- 

 ation during the summer of 1966 as the transducer face became 

 covered with ice at that time, which weakened its sensitivity. How- 

 ever, in the fall of 1966 the transducer was reestablished well below 

 the ice and there can be no question that the layer was not present 

 in 1967' After this long hiatus, the layer appeared again in late 

 March 1968 and remained imtil October 27 with a long gap in late svm- 

 mer. Recent records just returned from T-5 show that the layer again 

 reappeared on April 15, 19^9 • 



For four seasons the layer has thus made a summer appearance and 

 apparently is making a fifth appearance this year. It has appeared 

 as early as March and disappeared as late as November. It has never 

 been observed in December, Jantiary, or February. The lack of its ap- 

 pearance in 1966 may possibly be due to instnamental problems but the 

 lack of appearance in 1967 is believed to be real. Since T-5 is 

 drifting continually it is possible that the area of drift in 1966 

 and 1967 lacks the scattering layer and that this gap is explicable 

 in terms of geographic distribution. This will be discussed further 

 in the next section. 



A diurnal variation in the scattering layer characteristics is 

 also notable at certain times of the year. The best examples occur 

 close to the auttminal equinox. Diurnal variations in solar radia- 

 tion are most pronounced at the equinoxes, and the diurnal scatter- 

 ing behavior is very likely related to the light variations. Examples 

 of this behavior are shovm in Figure 5 for 1964 and in Figure k for 

 1965. The records are aligned so that recurring daily characteris- 

 tics are evident. The scale is in terms of local sun time. In Figure 

 5 a distinctive lajrer between the depths of kO and 100 m appears at 

 about 1900 and disappears at about 0500. This dense layer is svir- 

 mounted by a shallow layer of discrete echoes which may represent 

 isolated specimens from the scattering layer or which may be pre- 

 dators feeding on it. This nighttime layer is centered generally 

 about local midni^t. In 1964, its duration was observed to in- 

 crease fairly regularly each day from five hours on September 12 

 to 11 hours on September 29. During this same period the hours of 

 darkness increased from approximately 8 to 13. There seems to be 

 a direct correlation between the duration of the nighttime layer 

 and the hours of darkness, with the layer appearing about one hotir 

 after sunset and disappearing about one hour before sunrise. 



Variability in Space 



T-3 is drifting in a circular current pattern (Figure 5). The 

 ice moves clockwise about the gyre and completes one orbit in 

 about five years. Between 1963 and 1969, T-3 covered slightly 

 more than one revolution in an orbit with a diameter of about 800 

 km. The ice station was drifting in water depths of more than two 

 km throughout the period. The presence of the scattering layer on 



439 



