Evaluating Tooth Surface Mould Techniques

The lower right jaw (dentary) of an adult specimen of Archosargus probatocephalus (Perciformes: Sparidae) was dissected and mounted on a temporary base to facilitate manipulation. Two worn teeth with obvious variation in surface texture were selected from among the molariform teeth of the jaw: one exhibiting little wear, with a relatively smooth, enameloid surface; the other, more worn, with a relatively rough surface of exposed dentine (the enameloid having been worn away). A needle was used to scratch two intersecting perpendicular lines across the centre of each tooth surface, dividing it into quadrants. Within each quadrant a relocatable 100 × 145 μm area was identified, based on recognisable surface features, so that data could be collected from the same location on the replicated surfaces (Supplementary Fig. S3; areas designated NE, SE, SW, NW). Before the moulds used in this study were collected, tooth surfaces were cleaned by applying a random light body impression medium to the surfaces, which was then discarded.
Seven impression media were selected, representing different viscosity levels (Table 1). Four are polyvinylsiloxane compounds, two room temperature vulcanising (RTV) rubber compounds, and one heat accelerated RTV compound. Moulds were taken using each of the different media in a random order. Some media allow use of an applicator gun, which standardizes the mixing of two-components by extruding them through a helical nozzle; others required the body and activator components to be mixed and applied manually.
For each medium we tested accuracy and precision of replication, and for three media we also tested the effect of how they were applied (manual versus applicator gun, and application by different operators). The latter test was based on moulds taken using three different impression media, representing the compounds currently used in dietary microwear analysis: two moulds of manually mixed Speedex, each made by a different operator, to test for effects of variability between operators; two moulds of President Jet Light Body, one applied to the surface using the applicator gun, the other applied manually; two moulds of President Jet Regular Body, one applied to the surface using the applicator gun, the other applied manually. Manual versus applicator comparison was not possible with Speedex, because an applicator version is not available.
Epoxy casts were produced from each mould using EpoTek 320LV. In many studies, particularly of tooth microwear, transparent/translucent epoxy casting material is used, but in order to optimise data acquisition (using focus variation microscopy; see below) we used the black pigmented EpoTek 320LV, which in other respects has similar properties to the commonly used transparent EpoTek 301. After all moulds were taken, data were acquired from the original tooth surfaces (gold coated, using an Emitech K500X sputter coater, for three minutes to optimise data acquisition). Throughout the text, each cast is referred to by the name of the impression media from which it was created.