and with binoculars about once each 5 minutes. During the first half of 1953 the nnen concentrated 

 their search on the forward quadrants of the horizon in accordance with their usual practice. 

 Following some study of the data and development of a theory that schools intermittently sound 

 and are not detectable for periods of time, the fishermen were instructed to search the after 

 quadrants of the horizon as carefully as they did the forward quadrants. After finding fair num- 

 bers of schools, some of them directly astern of the vessel, the fishermen were convinced of the 

 necessity of this, and we believe that subsequently we achieved uniform attention to all sectors of 

 the horizon. 



Whenever a school of fish accompanied by birds was sighted, the scientist on watch was 

 immediately notified and he maintained a minute -by-minute record of the behavior of the birds, 

 the bearing of the school relative to the course of the vessel, which was never altered until the 

 school was abeam, and the estimated distance of the school from the vessel. After the school 

 was abeam, we frequently approached and fished it, if bait was available. 



Estimates of distance over the open sea are expected to be subject to a great deal of error. 

 We are convinced, however, that our fishermen obtain consistent and reasonably accurate esti- 

 mates of the distances of schools from the vessel. The consistency is evident in data which will 

 be subsequently discussed. The distances were estimated to the nearest 1/4 mile when under 

 about 2 miles and usually to the nearest 1/2 mile from 2 to 4 miles. The smoothness of the 

 frequency curves of schools sighted against distance indicates that there is no important tendency 

 to favor a particular distance over adjoining distances. Various methods of checking the accuracy 

 were tried, including the use of a range finder, a stadimeter, and a sextant. None of the conven- 

 tional methods for measuring distance under such conditions proved practical, but we did find 

 that with a calm sea large flocks of birds could be detected and their distance accurately measured 

 by radar. On two occasions we obtained radar measurements independent of the fishermen's 

 estimates, and we found good agreement (table 1). Apparently the fishernnen have developed a 

 very considerable skill in estimating the distance to birds by relating their appearance to the 

 average time required to run them down. 



Table 1. --Comparison of estimated distance and radar distance of bird flocks 

 a. CHARLES H. GILBERT cruise 12 - May 11, 1953 (about 80 birds in flock) 



b. HUGH M, SMITH cruise 22 - September 8, 1953 (about 300 birds in flock) 



— The radar readings indicate the spread of the rather diffuse 

 first flock (May 11, 1953) and the middle of the more 

 compact second flock (September 8, 1953). 



