644 EIGHTH PACIFIC SCIENCE CONGRESS 



through groups of points. It was usual to examine the deep zone first, 

 since it was the easiest to interpret. Then the upper zone was located. 

 The remaining points fell in the boundary zone, which was interpreted 

 in one or more segments as indicated by the data. Although this proce- 

 dure is extremely simple, there are opportunities for differences of opin- 

 ion, as in any method of interpretation. Therefore, it is necessary to 

 adopt certain conventions to insure uniformity. 



The surface observation was plotted at unit depth (0.1, 1, or 10). 

 This is justified because the surface sample in a serial observation is 

 invariably taken at some depth ranging from one-tenth to ten meters. 

 Simple Structure 



Simple structure occurs when there is a regular alteration of water 

 masses and transition layers. These appear as nearly homogeneous 

 zones and boundary zones. There is a considerable variety of these sim- 

 ple structures including the types shown in Figure 2. Although most 

 of the data indicated that there were three zones present, there were 

 cases where either the upper or deep zones were absent or had become 

 homogeneous. However, all of these examples indicate a simple struc- 

 ture involving a single water mass, or one well-defined mass overlying 

 another. This is the basic structure, and all others are modifications 

 or derivatives of it. 



In the case of the temperature no distinction is made in the classi- 

 fication between a negative gradient (temperature decreasing with 

 depth) and a positive gradient (temperature increasing with depth). 

 Complex Structure 



Certain complex structures such as those shown in Figure 3 cannot 

 be reduced by approximation to simple structures. In these there are 

 two or more segments in the boundary zone in which the successive 

 gradients of properties continue to increase, or decrease, rather than 

 alternate in slope. In the limit, where there are many such segments, 

 the log-plot appears as a curve, and the straight line rule cannot be 

 readily applied. However, in the majority of cases, there are no more 

 than two or three such segments, and the interpretation is obvious. 

 Conventions 



There were instances, particularly in the boundary zones, where the 

 data could be interpreted in several ways, as shown in Figure 4. Where 

 three or more points fall on a straight line, as shown in the boundary 

 zone in structure A, there is no doubt about the interpretation. Simi- 

 larly the two points in the boundary zone of structure B define the 

 position and slope of the boundary but are no guarantee that it is a 

 logarithmic gradient. The boundary zone in structure C may be in- 

 terpreted as having three points, as shown by the solid line, or as having 

 only one point, as indicated by the dashed lines. Evidently the true 



