FINCH: THE MECCA PROJECT 



would be 26.3 ppm, less than 90% of the present 

 ADI. 



CONCLUSIONS AND DISCUSSION 



These results provide evidence which would 

 permit FDA to raise the present interim guide- 

 line for mercurj' in fish to 1.5 ppm without 

 compromising public safety. 



At the higher level, it would still be necessary 

 to monitor some fisheries such as swordfish, 

 and perhaps big eye tuna, but the expensive 

 controls on tuna, halibut, and other species 

 would be unnecessary. The benefits of such 

 an action would be: 



1. To assist in price stabilization or possible 

 reduction of the affected products. 



2. To remove a heavy cost and administrative 

 burden from the processors. While no 

 precise estimate is available, this is be- 

 lieved to run into several millions of 

 dollars of unproductive costs annually. 



3. To increase the fish supply by avoidance 

 of rejection, with the corresponding nutri- 

 tional benefits which this unique food 

 supplies to the U.S. consumer. 



4. To offer a probability of restoring the 

 swordfish industry, since an FDA's market 

 survey on swordfish made in 1971, showed 

 that nearly 90% of swordfish samples taken 

 then did not exceed 1.5 ppm. Swordfish 

 at a higher mercury level might be ex- 

 ported to countries which, because of their 

 differing consumption patterns, may not 

 have limiting mercury requirements. 



In order to justify the application of the 

 results of the MECCA program to modify 

 present regulatory guidelines, several limita- 

 tions of the program must be considered. These 

 include: 



1. There are limitations on the accuracy of 

 returns received for any survey involving 

 over 40,000 mailed returns. Since families 

 at all levels of education participate, errors 

 in entries will be numerous. When entries 

 are missing, the error is apparent and 

 may be treated, but otherwise there is 



not usually any way to check. This is 

 especially true in this case where the 

 survey was performed for a different 

 purpose than the present application and 

 was completed 2 yr previously. In addi- 

 tion, there may be key punch errors in 

 transcribing the written report forms re- 

 ceived to cards. 



2. It is assumed that all the fish purchased 

 was consumed. Any errors in this respect 

 would mean real levels lower than those 

 shown by the program, and so provide a 

 safety factor. 



3. The transposition of common names under- 

 stood by the consumer to particular species 

 with identified mercury levels can give 

 rise to error. 



4. In some cases available mercury data are 

 limited. 



However, even with these potential short- 

 comings, we believe that the results of this 

 program present a much more realistic concept 

 than can be obtained from the present simple 

 assumption of a high level daily intake of 

 60 g/day fish. Particularly it may be noted 

 that, while the absolute values of intakes may 

 be in some error, the relative figures showing 

 the effects of guidelines are likely to be con- 

 siderably more precise since the same errors 

 in each run will tend to cancel in comparisons. 

 Thus it may be noted in the data base com- 

 parison that, while the average intake level 

 estimated for each type of run shown in Table 

 2 varies from 1.79 to 2.66 /Ug/day, the dif- 

 ference between the figures produced from the 

 two bases for the same runs varies only from 

 0.15 to 0.19 iug/day with Data Base 2 being 

 always higher. It may further be noted that 

 the daily consumption of fish in the United 

 States averages 14 g/day. Assuming a weighted 

 average mercury content of around 0.15 ppm 

 for U.S. species, this would provide 2.10jLzg/day, 

 which is less than the unconstrained average 

 of 2.48 jug/day predicted by the model. The 

 model figure includes consumption of sport 

 catch but, although no figures exist, this is 

 considered to be relatively small compared 

 with consumption of commercial catch. The 

 model results then are consistent with the 



623 



