JEK 4A/64 



Whenever possible, a 

 positive Identification was 

 made of the fish making up 

 the schools. Most of the 

 Identifications were made by 

 skilled fishermen on the 

 wheel watch, and such identi- 

 fications are usually based 

 on the type of birds working 

 over the schools and the 

 general behavior of the birds 

 and fish. Generally, when a 

 positive Identification fol- 

 lowed, the fishermen's judge- 

 ment was confirmed. Positive 

 identifications were possible 

 only when the fish broke 

 water near the boat, or when 

 one or more were caught on 

 the trolling lines. This 

 happened so seldom that most 

 of the records lack positive 

 verification. However, we 

 believe that the errors are 

 few. 



Scope and Organization 

 of the Sightings 



The fish and bird sight- 

 ings were gathered from an 

 area of approximately 8 mil- 

 lion squai^ miles (fig. 2). 

 Over this area there were 

 546 days of observation 

 during the period October 17, 

 1950, to April 4, 1953. The 

 Intensity of sampling Is ade- 

 quate to give a general pic- 

 ture of the distrlDutlon of 

 birds and fish but In many 

 Instances is not adequate for 

 defining seasonal and geo- 

 graphical variations in abun- 

 dance for subareas. 



Figure 1. — The form used to record surface observa 

 tlons on POFI vessels. This log. Introduced in 

 January 1953, calls for essentially the same in- 

 formation as earlier editions. 



Svimmarlzatlon of the 

 sightings was facilitated by classifying the observations according to rectangles 

 1 degree of latitude by 1 degree of longitude, based on the ship's noon position. 

 This could result in assigning an observation to a locality as much as 60 miles 

 from its true occurrence, because of the distance travelled during the day. This is 

 of little consequence in areas far from land, where the environment changes slowly, 

 but may cause distortion close to land. All of the observations were entered on 

 pxmchcards and further coded by month and year. 



After coding, the observations were divided Into three groups for study and 

 analysis. The "Island" observations consist of those made on days during which the 

 noon position of the ship was within a 1° square a portion of which is within 60 

 miles of land (fig. 2). A second group, termed the "semioceani c" observations. Is 

 made up of sightings made on days when the noon position of the ship was v;ithln a 

 1° square a portion of which is between 60 and 180 miles from land. The balance 

 comprises the "oceanic" observations. It will be realized upon reference to figure 

 3 that these limits are not perfectly realized because the configuration of the 

 island shorelines necessitated a number of compromises. 



