There are three sets of samples in this study. One is a set of gabbros sampled on board the JOIDES Resolution during the course of Leg 176. Another is a set from Leg 118 (R. Emmermann, pers. comm., 1998), and finally, a set of three muds from the core splitting table. With the exception of the muds, all of the samples were chosen to be representative of major rock types and units and an effort was made to include all the major gabbroic lithologies. No attempt was made to sample felsic veins, as other investigators have adequately sampled these.
However careful a point-sampling program is, it is impossible to generate a perfectly representative sampling of a section. This is particularly true in the case of cumulate rocks, where small lithologic units with very high concentrations of some elements will dominate the average and may or may not be included proportionately in the mean. For this reason, it was decided to subsample the mud from the onboard saw table as a means of estimating the bulk composition of the hole. Details of the philosophy and practice behind this sampling are published elsewhere (Snow et al., 1999, unpubl. data).