In discussions with estuarine hauling crews, it is clear that they avoid landing their nets in areas of dense seagrass. This potentially confounds the results obtained in the survey of the 9 estuaries. Given that we must sample sites where hauling occurs and hence where the nets are landed (logically), there can be at least two ways of interpreting consistent differences in shoot density and/or leaf-lengths between the control and hauled sites. First, the results provide some evidence that hauling has an impact on seagrass beds. Second, that the results simply reflect industry's choice to land their nets in areas of reduced amounts of seagrass and hence hauling has little or no impact. Clearly, if this project is to make some recommendations regarding the impacts of hauling and hence its management, we should, if at all possible, avoid concluding the research with these 2 alternatives at the centre of debate.

This seemingly intractable situation is not as hopeless as it would first appear. Previous research (e.g. Larkum et al., 1984) has shown that Zostera capricorni undergoes a cycle of growth in the spring and summer months followed by a dieback in the autumn and winter months. Given that the sampling for the survey has already been completed i.e. well prior to the seagrass attaining maximal shoot densities and leaf-lengths, it would be possible to take advantage of this period of enhanced growth to overcome the shortfalls discussed above. To do so would require that the the same sites be sampled again in late summer because if hauling was having an effect on shoot density and/or leaf-length it would be more clearly evident at this time of year. Furthermore, as we would have a baseline dataset we would predict that the change in shoot density and/or leaf-length from time 1 to time 2 would differ at the hauled compared to the control sites This would occur because the growth of the seagrass would be unaffected at the control sites where it would reach maximal shoot densities and leaf-lengths compared to the hauled site where it would not.

Clearly, the additional sampling described above would require an extension of the project beyond its current completion date in December, 1996. It is envisaged that a further 6 months would be required to complete the additional work necessary. It is important to note that this increased cost of identifying the potential impacts of hauling on seagrass will better serve the fishing industry as the greater scientific rigor gained will reduce the likelihood that the industry is blamed for changes that are not of its own making. Moreover, this approach would likely be the most cost-effective solution to overcome the potentially confounded result that will stem from the existing sampling. A far more costly solution would be to carry out a large-scale field experiment over a number of years to provide 'Before' and 'After" data for hauled and control sites.