528 



HERSEY AND BACKUS 



[chap. 13 



frequency migration are marked, with the exception of the low-frequency layer. 

 An examination of the frequency-depth relationship shows frequency varying 

 approximately as the 5/6 power of the depth in the high-frequency layer, and 

 as the 1/2 power in the mid-frequency layer. 



Observations made in two other areas reveal that the picture is not always 

 as simple as in the example of Fig. 18. In area B of Fig. 14 we found clearly 

 defined layers exhibiting the same general characteristics as those just des- 

 cribed. But in area C of Fig. 14 the relation between frequency and depth is 

 more subtle. The low-frequency layer is not significantly different from those 

 of other areas. We find layers in the mid-frequency range, however, which do 

 not appear to migrate in depth or frequency ; comparisons of the acoustic 

 pressure of the reflected signal at the beginning and end of the usual migration 

 period show a drop in pressure indicating that considerably more than half of 



100 200 



800 1000 



DEPTH (m) 



Fig. 16. Typical Sonagraph record of sound -scattering in the low-frequency range. The 

 vertical line at about 420 m and the extended straight horizontal lines are artifacts. 

 The low -frequency layer is represented by the pronounced blackening between 3 and 

 4 kc/s which starts at about 200 m (cf. Fig. 17). 



the population of the layer has migrated out of it, but in such a manner that 

 current analysis equipment is not sensitive to scattering from this migratory 

 portion. Other layers observed in this area show marked depth migration but 

 appear not to show the kind of frequency variation defined above as frequency 

 migration. Unfortunately, the frequency analysis for these layers may not be 

 valid. 



The behavior of the low-frequency layers has been difficult to define ever 

 since our earliest observations. In places the depth migration appears to be 

 distinct, while in others it is unrecognizable. After migration in resonant 

 frequency had been discovered, it was evident that there was little if any 

 frequency migration of these layers. Nevertheless they always appeared to be 

 strongly peaked. For example, Figs. 16 and 17 show the low-frequency layer 

 to be resonant, but Fig. 18 shows, for the same layer, considerable depth 

 migration accompanied by no apparent frequency migration. Granting the gas 



