Maps of the study locality. (A) Rising Star cave system. The locations of the Dinaledi subsystem (enlarged in B), Lesedi Chamber, and main cave entrance are indicated. (B) Location of Hill Antechamber within the Dinaledi Subsystem. (C) Floor profile of Hill Antechamber and Dinaledi Chamber showing height and angle of grade, and excavation areas.

Dinaledi Chamber burial features. (A) Photogrammetry model of the Dinaledi Chamber floor and excavation areas. Locations of 2013–2016 excavation area and two 2018 excavation units are labelled. The rectangle indicates the area of the other panels. (B) Photograph of excavation area including Feature 1 and Feature 2. (C) Three-dimensional reconstruction of excavation area including both the excavated skeletal material and the unexcavated material in spatial position. The oval area of Feature 1 corresponds to the sediment contrast and outline of skeletal material remaining in situ. Three excavated bones at left and one at right were stratigraphically higher and outside the feature. (D) Three-dimensional reconstruction from photogrammetry of Feature 1, indicating evidence of matrix-supported elements.

Sedimentology and stratigraphy of unlithified mud clast breccia and laminated orange-red mud clasts surrounding the burial Feature 1 on the Dinaledi floor. (A) North-facing overview of Feature 1 showing the relation of the sediments around the fossils and height of profile. (B) Profile view. Feature 1 occurs within unlithified mud clast breccia rich in orange-red clasts. A continuous laminated orange-red mud layer beneath unexcavated floor surface dips near the feature, where it becomes fragmented and muddled. (C)

Photomicrograph of sediment beneath the burial feature showing the in situ poorly sorted fabric of the unlithified mud clast breccia. (D) Close-up photomicrograph of a laminated orange-red mud clast. The clast contains up to 30% sand and has mm-scale laminations. (E) Close-up photomicrograph of laminated orange-red mud clasts coated and impregnated with secondary Mn- and Fe-oxyhydroxides in brown-grey silt and clay matrix of the unlithified mud clast breccia. (E) Principal component 2 (PC2) yields positive scores for fossil-bearing sediments and negative scores for sterile sediments (see Fig. 7). (G) PC2 loadings showing that elements with positive scores are phosphorus, sulfur, silicon and titanium.

Spatial data for Dinaledi Chamber surface collected specimens in relation to excavation area. (A) Position and skeletal element identifications of surface specimens including the 2013–2014 excavation unit. For fossil fragments large enough to be mapped as with two end points (n = 16), we calculated (B) plunge angle and (C) planar orientation following Brophy et al. (2021). Length of bars represent frequency of "two-shot" fossils with a given angle. (D) Density-based cluster analysis (7) identified six areas of higher-density surface specimen accumulations. These six areas are indicated with different colors here (arbitrarily chosen) with outliers in red.

Principal Component Analysis (PCA) of XRF major element chemistry of fossil-bearing and sterile sediments around Feature 1 on the Dinaledi floor, the Lesedi chamber and the Dinaledi Subsystem. (a) Three-dimensional (3D) biplot showing the three components that apply to all the variables (major elements). (b) Two-dimensional (2D) biplot showing component 1 vs component 2 for all the variables. Component 1 shows the dominance of Ti, Al, K and Ba vs Ca and Mg, which is consistent with the clay-rich mudstone sediments and their uncalcified nature. Component 2 shows the dominance of Si, P and S vs all other elements, but mainly Mn and Fe. It is useful for delineating fossil-bearing sediments from sterile sediments.

Principal Component 1 (PC1) scores and loadings. (a) Scores showing that almost all the samples from the Dinaledi floor have positive PC1 scores (uncalcified) compared to all the Lesedi Chamber samples and some samples from the Dinaledi Subsystem. The samples that have negative scores are calcified. Different colours represent different areas/chambers of the Rising Star cave system as annotated. (b) The positive and negative loadings of the elements on PC1 showing that Ca, Mg, S, P and Na have negative loadings and their composition and controls are different from all the other elements, which have positive loadings.

Principal Component 2 (PC2) scores and loadings. (a) Scores showing samples that come from sediments that contain fossils vs those that are sterile. The results are consistent with field observations and excavation results. For example, only one of the SA group samples from sterile areas around Feature 1 shows a positive score (presence of fossils) unlike all the SB group samples from inside Feature 1. Further, some samples of sediments (SC group) between features 1 and 2 and those from vertical profile on the side of Feature 1 (SE group) show a mixture of negative and positive scores. Different colours represent different areas/chambers of the Rising Star cave system as annotated. (b) The positive and negative loadings of the elements on PC2 showing that Si, P, S and Ti have positive loadings suggesting different controls of these elements compared to all the other elements, which have negative loadings.

Principal Component 3 (PC3) scores and loadings. (a) Positive scores show samples that are dominated by UMCB vs those with negative scores and are dominated by LORM. The samples from the sediments collected during excavation and opening of the features were sieved and not collected in situ, so their scores are probably biased towards LORM because of the abundance of LORM in the floor sediments. Different colours represent different areas/chambers of the Rising Star cave system as annotated. (b) The positive loadings of Mn,Ba and Fe on PC3 are consistent with LORM mud clasts that have been altered to UMCB vs the LORM mud clasts that are still unaltered and dominated by elements showing the negative loadings of mainly Al, Mg, K (see Figure S7).

Surface of Dinaledi Feature 1 remaining in situ with identifiable elements indicated.

Hill Antechamber feature. (A) Three-dimensional rendering of Hill Antechamber feature in situ prior to jacketing and extraction from the cave. Bone fragments and skeletal material are exposed on the excavation surface, and the feature is pillared with adjacent sediment profiles visible. (B) Overhead view of segmented skeletal material and teeth within the feature. The complete dentition is at top of frame, and the articulated foot is visible near the bottom of the frame. The artifact is at top left, colored in a tan color, with articulated hand elements above it. (C) Three-dimensional rendering of Hill Antechamber surface slope and profile, view from eastern side. At top, approximately 2 meters of slope and adjacent cave wall are visible, with the excavation unit at center. At bottom, the excavation unit is shown at larger scale with the skeletal material within the feature rendered. In the lowest portion of the feature, the articulated foot and leg material are emplaced at an angle opposite to the slope of the floor. Sedimentary layers here parallel the floor slope, and the leg, foot, and adjacent material cut across stratigraphy. The thoracic, upper limb, and cranial material along the top of the feature have been compressed into a plane at a shallower angle compared to surrounding stratigraphy. (D and E) Sections of CT data from Hill Antechamber feature. (D) Transverse (E-W) section, east at left. In this section the articulated foot is clearly visible near the bottom left with loose homogeneous sediment approximating the soft tissue. The dashed line delineates an internal fill, bounded by clasts and frequent voids that lie above a cut separating the fill from underlying sediment. The vertical line indicates position of section shown in E. (E) Sagittal (S-N) section, south at left. Here the longitudinal axis of the foot is visible at left, with additional skeletal material trending north. The dashed line follows the cut separating the foot and other skeletal material surrounded and supported by clasts, with some voids visible.

Hill Antechamber Artifact 1 (HAA1) showing surface from 8 different angles with 2 different lighting directions. The 3D model results from the segmentation of the synchrotron scan at 16.22um, the artifact being still in situ in the paster jacket.

Hill Antechamber Artifact 1 (HAA1) close-up from the previous figure with detail showing striations visible on both faces and intersection of these striations with sharp edge of artifact showing appearance of serrations.

(A) Dinaledi Chamber 2013-2014 excavation data with skeletal element identifications in oblique 2.5D view. For larger specimens able to be mapped with two end points (n = 79), we calculate (B) their plunge angle and (C) planar orientation frequencies (16). (D) 3D density-based cluster analysis (7) finds that a single high density cluster comprises the majority of excavated fossils (green) with two smaller peripheral clusters (purple, gold) and outliers (red points).

Bulk chemistry obtained from x-ray fluorescence (XRF) in weight percentage (wt.%).

Particle size distribution (PSD) of sediments from the Dinaledi Chamber floor and the Lesedi Chamber represented using the Folk and Ward Method.

In situ sampling localities around Feature 1 on the Dinaledi Chamber floor. (a) Top view of the Dinaledi floor showing the exposed Feature 1 and the group A (SA) samples from areas outside any feature, group B (SB) samples from inside Feature 1, and group C (SC) samples between Features 1 and 2. (b) Vertical wall where profile group E (SE) samples of Feature 1 were collected.

Excavation grid plan for Hill Antechamber

Locations of U.W. 108, U.W. 109, and U.W. 110 localities within the Dinaledi Subsystem.

SEM photomicrographs of sample 2280 from the Dinaledi floor S of the initial excavation pit. (a) Overview BSE image of the unlithified mud clast breccia (UMCB) with abundant secondary Mn- and Fe-oxyhydroxides occurring as polycrystalline infilling (b) and concretionary infilling (c). (d) EDS chemical maps showing the abundance and distribution of the major elements. Mn is more abundant and localized whereas Fe is more widespread. The mudstone contains appreciable Mg and K than Ca.

SEM photomicrographs of sample L01 from the UW108 passage. (a) BSE image of the UMCB similar to the Dinaledi Chamber floor. (b) EDS chemical maps showing the mud clasts with abundant Fe but little Mn, which is different from the Dinaledi floor UMCB.

SEM photomicrographs of sample L02 from the UW108 passage. (a) BSE image of the UMCB in this sample with abundant visible Mn-oxihydroxide compared to L01 in the form of discrete grains (b) and mud clast infilling (c). (d) EDS chemical maps showing the presence of fragmentary dolomite and discrete Mn-oxihydroxide grains in L02.

SEM photomicrographs of sample L03 from the UW108 passage. (a) BSE image of the highly calcified UMCB containing dolomite clasts that are cemented with calcite. (b) EDS chemical maps showing abundant presence of CaO and MgO and the sparse Mn-oxihydroxide.

SEM photomicrographs of sample L04 from the UW108 passage. (a) BSE image of the unaltered LORM mudstone with a very fine-grained texture. (b) EDS chemical maps showing the dominant clay composition of the LORM mudstone without CaO and MnO.

SEM photomicrographs of sample PF02 from UW110 above a ledge of broken chert horizon. (a) Overview BSE image of the UMCB sediments showing Mn-Fe-oxihydroxide grains in bright grey colour. (b) Mud clast with Mn-Fe-oxihydroxide concretions. (c) Mud clast with infilling of crystalline Mn-Fe-oxihydroxide. (d) EDS chemical maps showing that the mud clasts contain abundant very fine-grained Fe and Mn not occurring as Mn-Fe-oxihydroxide.

SEM photomicrographs of sample PFL01 from UW110 above the second chert ledge from the floor. (a) Overview BSE image of the muddy sandstone with little to no mud clasts and Mn-Fe-oxihydroxide. (b) EDS chemical maps showing the chemistry and textures of the quartz sand and gypsum grains.

SEM photomicrographs of sample HAP1 from the passage used to access UW111. (a) Overview BSE image of the LORM sediments. (b) EDS chemical maps showing the chemistry and textures of the LORM mud clasts, which contain abundant very fine-grained Fe. (c) Close up BSE image of a LORM mud clast containing very fine sand (d).

SEM photomicrographs of sample HD03 from the UW111 locality. (a) Overview BSE image of the UMCB sediments from beneath a thin flowstone. (b) EDS chemical maps showing that the UMCB sediments are not cemented by calcite but contain discrete, fine-grained dolomite.

Plot of samples vs their mean grain size. The mean grain sizes of half the number of samples is around 350 μm, whereas the other half are variable between 200 μm and 600 μm. The SA and SE groups of samples were sampled from in situ sediments at different depths, and their mean grain sizes show upward fining sequences shown by the black arrows (arrows point towards the top of sediments, see Figure S1 for relative depths of samples). The DF group of samples also include samples from different depths, and they show a mixture of upward and downward fining sequences consistent with their mixing during sieving.

Excavation unit in Dinaledi Chamber from original 2013–2014 excavation work and locations of 2018 excavation units plotted by photogrammetric reconstruction.

Hill Antechamber excavation unit S150W150 prior to opening excavation. This area had a collection of non-overlapping flat stones on the surface.

Top surface of the Hill Antechamber feature after full exposure. Photo (left) and 3D model based on photogrammetry (right). The bone material at the northmost extent of the feature is powdery and highly fragmented, with more complete skeletal elements visible toward the south.

Hill Antechamber feature after pedestaling and separation from surrounding sediment.

Sagittal (north-south) section of Hill Antechamber feature. North is at left of frame. This section is at approximately 55% of the east-west breadth of the feature. The articulated foot is visible in longitudinal section at right of frame, with cross-sections of other bones and teeth further to the left of frame. The layer that constitutes the top of the feature is packed with bone material including articulated, semi-articulated and loose material, flattened into less than 5 cm thickness.

Hill Antechamber feature after jacketing of largest block (U.W. 101-2076) in six layers of plaster bandages, prior to separation from sediment at its base.

Fossil mass within plaster jacket after packing into waterproof caving bag for exit from the cave system.

Detail of medical-resolution CT image of Hill Antechamber feature. This is a horizontal section with north at bottom of frame and west at right of frame. In this image, the bright object at lower left is a cross section of HAA1. At its left, cross sections of four rays of the articulated hand are visible; there is also a bone visible in the gap or space adjacent to the artifact that is a fragment of intermediate phalanx.

2013-2014 Dinaledi Chamber excavation area in (A) photo and the combination of white-light surface scans (B) and high-resolution laser scans (C) used to collect 3D data on the location of specimens excavated from the original Dinaledi assemblage. 3D data plotted as (D) X- and (E) Y-axis profiles of identified specimens including the articulated hand and foot.

Sediment profile showing east wall of N100W50 excavation unit in Hill Antechamber. The sediment is a dark brown unlithified breccia containing laminated orange-red mud (LORM) clasts. In this unit, the clasts make up a small fraction of the sedimentary deposit with little evidence of layering or stratigraphic differentiation.

Hill Antechamber excavation east wall profile. North is at left of frame. The ellipse is drawn around a 15 cm by 10 cm by 5 cm collapse of the wall that accompanied removal of the plaster jacketed block, resulting in some distortion to the profile in this localized area. The darker patch at the right side of the ellipse is a shadow from the collapse edge, not a dark-colored inclusion in the sediment. Layering of the unlithified mud clast breccia is visible with some layers having a higher content of LORM clasts and laminae, with color variation less evident here than in the section shown in Figure S22. The layering is approximately parallel to the slope of the chamber floor. LORM content and clasts are less toward the north edge of the excavation, at left of frame.

Stratigraphic profile of sediments directly adjacent to south boundary of Hill Antechamber feature. At left of image is the west side of the unit. This profile represents the sedimentary structure of the S50W100 and S50W50 units before the excavation reduced the feature to its rounded south edge. Horizontal layering of unlithified mud clast breccia (UMCB) with denser orange-red LORM-clast bearing laminae is evident for the top 10 cm of the profile. This layering becomes subhorizontal with horizontal depth with an east-west trending slope. The lowest layer visible trends into the horizontal floor of the excavation unit. This layering is not paralleled by the skeletal material, fill, or LORM clasts visible within the feature itself.

Rendering of overall CT segmentation of Hill Antechamber feature. (A) View from west side; (B) view from north side; (C) view from overhead with elements labeled.

CT section of Hill Antechamber feature. This east-west transverse section is at approximately 50% of north-south length of the feature. At the bottom of the section, many small LORM clasts are visible, with two notable voids taking the form of vertical cracks. The disordered array of LORM clasts continues to the right of image with frequent voids (west side of feature).

Identifiable dental elements within the Hill Antechamber Feature. This view is oriented from below the feature to maximize the occlusal details of teeth other than the associated maxillary dentition of Individual 1, at top right. Enamel and dentin exhibit high contrast within the CT data. The 3D surface models generated from segmentation are the result of a smoothing algorithm after voxel selection. While this smoothing assists in visualization and identification in many cases, it creates some distortion in other cases. This resolution does not make it possible to identify all elements without ambiguity. For the teeth attributed to Individual 1, the presence of occlusal ordering within the maxilla and mandibular fragments enables a clear assignment of nearly all elements. For the teeth inferred to belong to Individuals 2 and 3 the present data leave ambiguity about the identity of some of the elements.

CT section of Hill Antechamber feature. This is a transverse section on the east-west plane at approximately 65% of the north-south length. The five rays of the articulated foot are visible at lower left of the section. This section cuts across the metatarsals. The bones of the foot are immediately surrounded by a halo of sediment that approximates the shape of the foot’s soft tissue. This lower-density sediment separates the bones of the foot from surrounding, more radio-opaque LORM clasts and sediment. Above the foot some small voids in the sediment are visible; small voids are also visible toward the left of this section directly above a disordered arrangement of LORM clasts.

Bright orange LORM patch immediately beneath Hill Antechamber feature. (A) Bottom of Hill Antechamber feature after jacketed extraction and inversion. The bright orange patch is visible centrally slightly toward the right (west) side of the inverted feature (B) Excavation unit immediately after jacketed extraction of feature and cleaning of surface. The corresponding orange patch is visible at center of unit.