CAHN: SENSORY FACTORS 





Figure 4. — Scale diagrams of tunas swimming with Plex- 

 iglas partition in rectangular channel. The 3 usual 

 positions assumed relative to the partition are shown. 



the amount that crossed to the other side was 

 barely perceptible. 



The data on the side-to-side spacing are sum- 

 marized in Table 3, where the standard errors 

 were calculated from the variance of the means 

 of the runs in Table 2. It will be noted that in 

 the presence of partitions the variability is much 

 greater. A three-way analysis of variance on 

 the group means (Table 4) showed that the 

 increased spacing in the presence of partitions 

 was highly significant. Although there was a 

 tendency for the fish to remain slightly more 

 compact in the larger groups, these differences 

 were not statistically significant; the eflfects of 

 shape of the test area was very close to but not 

 quite significant. 



Table 3. — Summary of data on side-to-side spacing of 

 Eiithynnus in schooling orientation. 



Group means and standard errors 

 Urn) 



^ There was only one mean in this cose,- also only 4 fish in the school. 

 ^ The two means in this group were the same. 



Table 4. — Three-way analysis of variance' of group 

 means for side-to-side spacing of Euthynnus. 



Source <;,,.„„{ Degrees 



of ^"'^°f of ""'"" /-Ratio 



variance =^^°'^= freedom =^"°'^= 



Mean 



Significance 

 level 



Swimming speeds in the absence of partitions 

 ranged from 50 to 64 cm/sec in both rectangular 

 and octagonal areas. With partitions, in the rec- 

 tangular channels the two orienting fish swam 

 at about the same speed within the range of 50 

 to 64 cm/sec. During turns, usually made at the 

 ends of the rectangles, they almost always ma- 

 neuvered to begin the next lap together. Orien- 

 tation persisted for periods as long as nine days 

 for one pair. In the octagonal area the duration 

 of fish-to-fish orientation was never longer than 

 two days. The size differences between the oc- 

 tagonal channels necessitated differences in 

 swimming speed to maintain fish-to-fish orien- 

 tation. For example, the fish in the outermost 

 octagon had to swim excessively fast (over 65 

 cm/sec) to maintain contact with the fish in the 

 innermost section, swimming abnormally slowly 

 (34 cm/sec). Therefore most of the data was 

 obtained in the two larger octagonal channels, 

 where the fish swam at close to normal speeds, 

 at almost 50 to 64 cm/sec. 



201 



