532 



Fishery Bulletin 105(4) 



biomass and in numbers. For both net flux and spill- 

 over, a positive value indicates net movement out of the 

 MPA; a negative value indicates net movement into the 

 MPA. For the combination of all species, overall spill- 

 over was lowest at the Tumon MPA, where only 1.7% 

 of all tagged biomass was exported (two-way ANOVA, 

 F=17.12, P<0.01, Tukey's HSD, P<0.05). Spillover did 

 not differ between the other two MPAs (20.3% at Piti 

 and 16.7% at Achang, Tukey's HSD, P<0.05). 



A significant interaction between species and location 

 (two-way ANOVA, F=7.73, P<0.0001) warranted a closer 

 inspection of fish movements at each location. Oranges- 

 pine unicornfish was the only species exported from all 

 three MPAs. The overall mean spillover for orangespine 

 unicornfish was 29.4% of biomass (Table 2) and 21.5% 

 of individuals (Table 2). This was significantly higher 

 than the other four species (2-way ANOVA, F=6.27, 

 P<0.0001). There was no significant difference in spill- 

 over of orangespine unicornfish between MPAs (2-way 

 ANOVA, F=2.71, P=0.34). Although yellowstripe goat- 

 fish showed low overall mean spillover (1.1% of individu- 

 als imported to MPAs but 16.4% of biomass exported), 

 it was actually very mobile; 34% of its tagged biomass 

 moved into the Achang MPA, and 31% of its tagged 

 biomass moved out of the Piti MPA. These net inward 

 and outward movements at different locations cancelled 

 each other in the calculation of overall mean spillover 

 for this species. No movement of this species occurred 

 across the Tumon MPA boundaries. In contrast, net 

 export of convict surgeonfish occurred at Achang and 

 Piti MPAs, but there was a net import of surgeonfish 

 biomass to the Tumon MPA (Table 2). Spillover of con- 

 vict surgeonfish did not differ between the Achang and 

 Piti MPAs (Tukey's HSD, P>0.05). Honeycomb grouper 

 showed an overall mean outward movement of 3.3% of 

 tagged biomass (Table 2). However, there was no net 

 movement of individuals across MPA boundaries at 

 all sites (Table 2); this result indicated that primarily 

 larger individuals moved out of the MPAs. 



In general, for most species, the percentage of bio- 

 mass exported from the MPAs was greater than the 

 percentage of individuals exported. The exception was 



that of yellowstripe goatfish at the Achang MPA, which 

 showed a net outward movement of 15.6% of individuals 

 (Table 2) but a net inward movement that was 34.2% 

 of biomass (Table 2), indicating that the fish moving 

 into Achang MPA were much larger than the fish mov- 

 ing out. In contrast, at Piti there was a net outward 

 movement of 15.6% of individuals and 30.8% of the 

 biomass for yellowstripe goatfish, indicating that larger 

 fish were primarily exported. The effect of resident 

 biomass on spillover of each species was examined by 

 plotting spillover (in biomass) against the biomass ratio 

 of each species at each MPA, measured as biomass in- 

 side the MPA divided by the biomass at the fished site. 

 The variation in density between censuses rendered it 

 impossible to create separate regressions for each spe- 

 cies or MPA. Thus, the overall mean biomass ratio and 

 mean spillover were pooled for all species and MPAs 

 into a single regression. No relationship was found 

 between the biomass ratio and mean spillover (r-=0.1, 

 P=0.34; Fig. 3). 



Discussion 



One of the primary effects of protection from fishing is 

 an increase in size and abundance of fishes, and there- 

 fore in biomass, in MPAs (Russ, 2002; Halpern, 2003). 

 Biomass of reef fish was greater in marine reserve areas 

 than in fished areas after about six years of effective no- 

 take protection at Sumilon and Apo Islands in the Phil- 

 ippines (Alcala et al., 2005). Moreover, these differences 

 became larger with increased duration of protection up 

 to 19 years. Russ and Alcala (2004) concluded that the 

 time required for full recovery of reef fish populations 

 at the Sumilon and Apo Island MPAs would be 15 and 

 40 years, respectively. Biomass of leopard coralgrouper 

 iPlectropomus leopardus) and spotted coralgrouper (P. 

 maculatus), two large roving grouper species, was five 

 times higher in no-fishing zones than in fished zones 

 of the Great Barrier Reef Marine Park, after 14 years 

 of protection (Evans and Russ, 2004). Numerous other 

 studies had found no difference in biomass of Plectro- 



