Fig 1.
Overview and position of Mertenhof shelter.
(a): Elevation model of southern Africa showing location of Mertenhof Rock Shelter (star) and other sites mentioned in the text (circles); BBC = Blombos Cave, DRS = Diepkloof Rock Shelter, KKH = Klein Kliphuis, KRM = Klasies River, PP13b = Pinnacle Point 13b. (b): Inset showing location of Mertenhof (star) relative to the Doring River (red line) and the boundary between the Cape System and Karoo System geological zones (white line). Other sites shown are 1 = Diepkloof, 2 = Klein Kliphuis, 3 = Hollow Rock Shelter, 4 = Klipfonteinrand, 5 = Putslaagte 8, 6 = Uitspankraal 7, 7 = Tweefontein. Digital elevation data from [21] (c): Lower panel shows a panorama of Mertenhof Rock Shelter (black circle) in the context of the kloof. The Ceberberg mountains are visible in the background on the right hand side of the image, and the shale-dominant Tra-traberge on the left.
Fig 2.
Chert and heat-treated silcrete artefacts in Mertenhof.
(a-g): Different types of chert in the analysed Mertenhof assemblages. (h): Surface of unaltered and experimentally heat-treated silcrete showing the colour change upon heating. (j): Untransformed Mertenhof silcrete of the yellow type to which no heat treatment was applied. (i, k): Artefacts made from this same silcrete type that were heat-treated. (l): Mertenhof artefact made from grey silcrete. (m): the contrast between a pre- and a post-heating scar on the same artefact showing the difference of roughness. (a-g) and (i-l) all have the same scale. Note that the roughness of the surfaces in artefacts (i-l) cannot be estimated in these photos that were taken with frontal light.
Table 1.
Raw material prevalence and implement types in the lithic sequence at MRS.
Table 2.
Frequency of heat treatment proxies on the silcrete artefacts for all contexts.
Table 3.
Frequency of heat treatment proxies on the silcrete artefacts by industrial (sub)grouping relative to raw material proportions for main rock types.
Fig 3.
Correlation plots showing the relation of the percentage of chert artefacts in relation to the percentage of heat-treated silcrete.
(a) By contexts (CT) and (b) by (sub)groupings. Chert percentages refer to the total number of artefacts in the contexts, heat-treated silcrete percentages refer to the number of silcrete artefacts in the contexts. The coefficient of determination of the correlation is r2 = 0.973, p<0.001 for (a) and r2 = 0.999, p<0.001 for (b).
Fig 4.
Correlation plots showing the relation of the raw material spectrum’s shift from fine-grained non-local materials towards coarse-grained local quartzite to the percentage of heat-treated silcrete.
(a) By contexts (CT) and (b) by (sub)groupings. The non-local fine-grained / local coarse-grained materials ratio is calculated as the addition of the percentages of chert and DWS/hornfels divided by the percentage of quartzite in each context. Heat-treated silcrete percentages refer to the number of silcrete artefacts in the contexts. The coefficient of determination of the correlation is r2 = 0.823, p = 0.005 for (a) and r2 = 0.921, p = 0.040 for (b).
Fig 5.
Correlation plots showing the relation between the number of cores/flakes and heat-treated silcrete in (sub)groupings.
Cores/100 flakes are calculated from the total number of artefacts in the (sub)groupings and heat-treated silcrete percentages refer to the number of silcrete artefacts in the contexts. The coefficient of determination of the correlation is r2 = 0.859, p = 0.073.