The greatest single problem in the pelagic studies was the accurate 

 prediction of the natural variability to be encountered and a subsequent 

 statistical design that would adequately remove this mask. For example, 

 the FDC measurements provide a good example of when significant differ- 

 ences may indeed have existed between polluted and reference stations, 

 but sample frequency was too low to detect it. It is true that very 

 often field logistics may preclude a truely adequate statistical design; 

 in such cases it would be advisable to evaluate whether or not a lesser 

 effort will produce any usable results. Conducting fewer experiments 

 with better statistical designs may be more advantageous. These com- 

 ments also apply to areas other than the pelagic zone. Statistical 

 design for field experiments is thoroughly discussed in a recent work by 

 Moore and McLaughlin (1978). 



In most oil spills, direct impact on the pelagic system will be of 

 short duration (less than a month). The most important measurements in 

 the pelagic zone, therefore, will be those helping to define its impor- 

 tance as a medium that transports oil to other less resilient systems, 

 such as the benthos or the littoral zone. However, this is not meant to 

 understate the importance of effects on pelagic larvae and other plankton, 

 which might be of long term or trophic significance depending on the 

 circumstances of the spill. 



2. The littoral system : The net-bag method for sampling in the 

 Fucus belt has proven to be an efficient method of surveying the effects 

 on the littoral macrofauna. Studies of the effect of oil on littoral 

 and supra-littoral communities, particularly on plants, are warranted as 

 a means of evaluating the damage caused by the spill. These zones are 

 especially sensitive to various methods of cleanup, and so studies 

 should be included which assess the effectiveness or additional damage 

 of cleanup methods. 



In situ (e.g., suspended) bio-assays were not attempted during the 

 Tsesis study, although they have been recommended at times (MITRE, 

 1978). It is believed that in cases where direct studies of organisms 

 in place are possible, such as Tsesis , fewer doubts are inherent than 

 with bio-assays; thus studies of the existing system should be con- 

 sidered first. Furthermore, when samples and transects can be done 



52 



