A COMPARISON OF SOME METHODS COMMONLY USED FOR THE COLLECTION OF SUBLITTORAL SEDIMENTS AND THEIR ASSOCIATED FAUNA

Five replicate samples were collected using each of four different col lection methods, at two sites with contrasting sediment types, namely soft mud and hard muddy sand. The four methods were: 1. direct collect ion by divers; 2. collection using a Craib corer; 3. subsampling from an USNEL pattern box corer; and 4. subsampling from a van Veen grab. A s far as possible, biasing factors, other than the actual sampling ope ration at the sediment surface, were controlled. Nematodes in the samp les were identified and the resulting data analysed using a range of t echniques. Univariate measures of community structure, such as diversi ty indices, did not reveal any significant difference between sampling methods, neither was any significant difference detected in diversity profiles (k-dominance curves). The conclusion which might be drawn fr om these experiments is that it makes little difference which method o f sampling is chosen for a particular study. Such a conclusion would b e erroneous, as multivariate analyses did reveal significant differenc es in community structure between samples retrieved by different metho ds. The patterns of differences at the two sites were dependent, in de tail, on the transformations used in the analyses, but in general ther e was a smooth shift in community samples, to subsamples from grabs. T he detected differences, although significant, were small, and did not suggest a strong, or ecologically meaningful, explanation. A modified index of multivariate dispersion is described, and this is used to ex amine differences between sampling methods inn intra-sample variabilit y. This showed that differences in variability contribute to, but do n ot fully explain, the observed differences. It should also be borne in mind that factors which might not be expected to affect different sam plers in the same way, such as water depth, weather conditions and sam ple size, were controlled. It is likely that these have biasing effect s as important as the actual operation of sampling gear on the seabed, where explanations of the differences between sample quality from dif ferent methods, such as bow-wave effects, have been focused. Copyright (C) 1996 Elsevier Science Ltd