Baum et al : Bycatch of Hippocampus erectus in a Gulf of Mexico sfirimp trawl fisfiery 



727 



Table 3 



Contingency tables on effects of year, lunar phase, and area on population structure of seahorses in the bycatch of the Hernando 

 Beach bait shrimp trawl fishery. Yates corrections were applied to 2x2 contingency tables. 



Source 



Juveniles: Adults 



Sex ratio 



df 



df 



X~ 



Reproductive state 



df 



P 



Year 



Lunar phase 

 Area 



2.46 



11.00 



8.84 



>0.10 



<0.025 



<0.05 



0.15 

 3.69 

 13.3 



>0.50 

 >0.25 

 >0.05 



27.89 



7.80 

 21.64 



<0.001 



>0.05 



<0.001 



M 



(Table 3); almost half of the reproductively ac- 

 tive males were found in one area (S2),and 83% 

 of them were caught on one of the three nights 

 with very high CPUE (16 July 1998). 



Mortality and injury 



Fewer than 1% of seahorses died during tows or 

 sorting, but 4,7% (;i=28 of 588 seahorses) lost 

 tail rings. The mean loss was 22 of the usual 36 

 tail rings (Lourie et al., 1999), or 61% of the tail 

 (range=8-31 rings lost). Approximately 14% of 

 the losses (n=4 of 588 seahorses) appeared to 

 be the result of new wounds, probably caused 

 by the focal trawl. There was no effect of year 

 (P=0.25), sex (P=0.75), or reproductive status 

 (P=0.75) on incidence of seahorse injury. Postre- 

 lease mortality is unknown. 



Discussion 



Direct effects of the trawl fishery on seahorse 

 mortality 



We estimate that this fleet catches almost 72,000 sea- 

 horses incidentally per annum, based on the mean CPUE 

 and given that 31 boats fished approximately 240 nights 

 per year. Most seahorses were returned to the wild in the 

 Hernando Beach fishery, but this may not be the case in 

 other live-bait shrimp trawl fleets in Florida (Vincent, pers. 





_a 



jiMi 



i 



wkk 



41) Ml Sll M)ll i:0 1411 IM) IXd :iKi 



Standard length (mm) 



Figure 5 



Length-frequency histogram for seahorse bycatch in 1998 and 1999. 



obs.). We could not determine the potential fishing-induced 

 mortality for the Hernandez Beach H. erectus population, 

 even when all trawled seahorses were retained, because 

 seahorse catchability in roller beam trawl gear is unknown 

 and no studies have estimated H. erectus density in the 

 wild. Focal behavioral studies on congenerics similar in 



