Hickford and Schiel: Gillnetting in southern New Zealand 



675 



the net (Hickford and Schiel, 1995), a motion that 

 often results in a fish freeing itself from the net. 



The deep-bodied Latridopsis ciliaris were mostly 

 tangled and gilled. The small number of this species 

 that were found wedged in net mesh is probably due 

 to none of the mesh sizes being large enough to al- 

 low larger blue moki to enter the nets any further 

 than their gills. The significantly greater number of 

 fish tangled in the 2.5" mesh is a result of larger fish 

 becoming tangled by their large fins and protruding 

 fin rays. Large, laterally compressed fish, such as L. 

 ciliaris and Nemadactylus macropterus, are not 

 strong swimmers (Doak, 1991). They rely on muscu- 

 lar undulations from head to tail in order to swim, 

 and they brake with their pectoral fins. This weak 

 swimming ability, coupled with their large spiny fins, 

 resulted in L. ciliaris often becoming entangled by a 

 single fin rather than being truly enmeshed in the 

 net. 



Winters and Wheeler ( 1990) stated that the differ- 

 ence in fishing power between nets of various mesh 

 sizes may be a result of differences in the proportion 

 of fish caught by each entanglement mode in each 

 mesh size. They stated that the three modes of cap- 

 ture have different fishing powers that may vary with 

 mesh size, but in general, wedging is more effective 

 than gilling, and both these modes are much more 

 effective than tangling. However, the results of our 

 study show that for total fish numbers caught in all 

 mesh sizes combined, most fish were gilled (60%), 

 whereas wedged (17%) and tangled (23%) fish made 

 up significantly lower proportions of the catch. This 

 result suggests that, in the case of our study, gill nets 

 do in fact "gill" fish rather than capture them by tan- 

 gling or wedging. 



Mesh-size selectivity was evident from the mean 

 length offish captured by each method in each mesh 

 size. Although the fork length of gilled and wedged 

 fish increased with increasing mesh size, the fork 

 length of tangled fish was less uniform in its rela- 

 tionship with mesh size. 



The results of our study show that tangling is not 

 the result solely of size selection and is not consis- 

 tent across mesh sizes. The proportions of Odax 

 pullus caught by each method, when plotted against 

 fork length, show a clear transition as fork length 

 increases from most fish being wedged to the major- 

 ity being tangled. This transition would not occur if 

 the size of tangled fish were independent of mesh 

 size, because fish of all sizes would then become 

 tangled in any given mesh size. 



Although the concept of gill net "saturation," or 

 diminishing returns with increasing effort, is gener- 

 ally recognized as a limiting factor in catch per unit 

 of time (Minns and Hurley, 1988), there have been 



relatively few studies directed at exploring the 

 mechanisms that limit the catch. Our study shows 

 evidence of a set- time saturation effect with all three 

 mesh sizes. Neither the number of fish caught nor 

 the number of species caught were significantly dif- 

 ferent between the six and fifteen hour sets. How- 

 ever, these different set times had only a small over- 

 lap diurnally and saturation may have been influ- 

 enced by differing periods offish activity. There was 

 no evidence in either the gill nets or in both under- 

 water observations and videos (Hickford and Schiel, 

 unpubl. data) that predators affected catches. 



Space limitation in the gill net itself is regarded 

 as a major component of the saturation effect. Once 

 a fish has been captured, the particular cell that it 

 occupies and the cells immediately surrounding it 

 are not capable of catching other fish. Koike and 

 Takeuchi (1982) examined this feature and found 

 that fish were repulsed around a captured individual 

 for some but not all mesh sizes. Kennedy (1951) cited 

 additional ways in which the efficiency of a gill net 

 decreases with time. These included the presence of 

 captured, struggling fish (which makes the net more 

 obvious and could frighten other fish away) and the 

 presence of dead fish (which may cause other fish to 

 avoid the area). Kennedy speculated that the greater 

 the catch during the initial time period, the greater 

 the difference between the initial (observed) and fi- 

 nal (expected) catches. 



The effect of set time on total and species catches 

 in gill nets has a direct bearing on the use of this 

 gear in assessing the abundance and species diver- 

 sity of fish populations. Some studies have focused 

 on comparing multi- to one-night catches (Richards 

 and Schnute, 1986; Minns and Hurley, 1988).The 

 evidence presented here, however, suggests that net 

 saturation can occur during a single night, although 

 this may be confounded by the varying behaviors of 

 the fish species present. 



The apparent similarity in catch rates of Odax 

 pullus, Arripis trutta, Aplodactylus arctidens, and 

 Latridopsis ciliaris during day and night sets was 

 unexpected. These species are more active during the 

 day and would be expected to be caught in signifi- 

 cantly greater numbers in the day sets. Greater av- 

 erage numbers of these species were caught during 

 the daytime, but the catches were so variable that 

 any patterns may have been masked. 



The condition offish in the landed catch is closely 

 related to the length of time a net is in the water. 

 The catch of nets set for longer than six hours will 

 contain a high proportion of damaged fish. The rela- 

 tion between set time and condition is confounded 

 by the fact that nets set for periods longer than six 

 hours were usually left in the water overnight. Dur- 



