The matrix (< 40 μ) of turbidites forms a possible clue to the density of turbidity currents and the origin of the graywacke matrix. Experiments in a circular flume provide a mechanism to study the relation between composition of suspensions at various speeds and their deposits. There is a close analogy to the lower part of turbidity currents. The lutum content of samples with median diameters greater than 400 or 500 μ is found to correspond to the suspended load of the pore water. The higher value for finer deposits can be recalculated to suspension concentration by use of the “sedimentation factor”. Hence, each turbidite carries, as it were, a sample of its depositing current. The lutum content depends not on the ratio of sand to lutum in the current, as tacitly assumed by many authors, but mainly on the ratio lutum to water, although also influenced by velocity.

The average lutum density of coarser recent deep-sea sands is 1-2%. This indicates turbidity currents with 5-10% lutum by weight (density 1.03–1.07). The sand must be added to ascertain the current density. In first approximation turbidity currents tend to have densities at their nose of 1.1–1.2, but higher and much lower values also occur.

The maximum original lutum percentage of coarse turbidites is below 10%. Higher values are very scarce and are due to post-depositional mixing, or we are dealing with slides. However, in fine-grained turbidites there is more matrix up to 20% for a median of 100 p. Hence, coarse graded marine graywackes with 20 or more per cent matrix are presumably weakly metamorphic turbidites, that originally held the same modest amount of lutum as recent turbidites of the same grain size. The Trask sorting of the experimental deposits is very good, like the average of natural turbidites. Most cumulative curves of turbidite grain-size analyses on arithmetic probability paper show a characteristic bend in fine sand or silt sizes.