Rising Star cave system map. The relative locations of current entrances and underground fossil localities are indicated, including the Dinaledi Subsystem and Lesedi Chamber. The inset is an expanded map showing the Dinaledi Subsystem with the Hill Antechamber, Dinaledi Chamber, and associated fissure passages including the U.W. 110 locality.

Dinaledi Subsystem map and detail showing floor space of the Dinaledi Chamber and Hill Antechamber. The floor visualization is derived from LIDAR and photogrammetric survey. Differences in color in this visualization reflect the lighting used in different parts of the subsystem during data capture, and do not accurately depict contrasts in sediment color. Excavation units are indicated.

Dinaledi Subsystem plan view and cross-sectional view. In the plan view (bottom), floor elevation acquired from FARO scanning is indicated with the color ramp. The location of the cross section is indicated with the red line running from the Hill Antechamber via one of the interconnecting passages into the Dinaledi Chamber. The cross-section, indicated with a red line, includes both the Hill Antechamber Feature and Puzzle Box excavation areas. The cross-section view (top) indicates the relative elevation of these areas and the intervening floor surfaces.

Excavation plan in the Hill Antechamber. The location of the three excavation units described in this study are shown, with the skeletal and dental material of the Hill Antechamber Feature in place. The other two excavation areas each produced little skeletal material, with only unidentifiable fragments in N100W50 and 10 teeth with few long bone fragments in S150W150, discussed in text.

Hill Antechamber Feature. (A) Photogrammetric model of the feature during the course of excavation with surface partially exposed. North arrow in this image depicts grid north, which is offset by 20 degrees from magnetic north. (B) Plan view of skeletal material identified from CT segmentation. (C) Transect of photogrammetric model showing the floor surface of the profile flanking the feature excavation unit. A model of the segmented skeletal material within the feature is superimposed on the floor profile in the same orientation to illustrate the position of remains relative to the floor surface. (D) Skeletal material segmented within the feature as viewed from the west of the excavation. The material occupies volume across a substantial thickness of sediment.

Hill Antechamber Feature associated elements. Groups associated by spatial position, articulation, or clustering of contiguous anatomical elements.

Dental elements identified within the Hill Antechamber Feature, viewed from below in their relative positions in the feature. Identification of elements from the CT data is uncertain in a number of cases, represented here by “?”. These provisional attributions are provided to evaluate the minimum number of individuals and likely individuation of elements.

Stratigraphic situation of the Hill Antechamber Feature. At left, a photogrammetric visualization of this excavation areas shows the locations of four cross-sections labeled A, B, C, and D, which are pictured at right. These are arranged from north to south across the designated feature, with the final one being the excavation profile exposed as sediments to the south were cleared during the pillaring of the feature. All sections are shown at the same scale. (A) Section 20 cm from south end of feature. Four digits of the articulated foot are visible at lower left of this section. A concave rubbly layer along the bottom of the excavated unit slopes upward at both sides of the section, while an area of large clasts occupies the center of the section above and to the right of the foot. (B) Section approximately 10 cm from south end. Tarsals of the articulated foot at center-left. Here, the rubbly layer is a concave region below the tarsals and bone cross- sections at far left. (C) Section approximately 5 cm from south end of feature, at same approximate scale. The bottom half of this section is dominated by the same rubbly clast-rich area. Cross-sections of tibia and fibula are visible at center-left of section. (D) Sediment profile at immediate south end of feature. Orange layers are rich in LORM content, with many visible clasts, and these alternate with darker-colored layers. The layering is roughly parallel with the east-west slope of the chamber floor in this profile, somewhat increasing in slope in layers below 10 cm. This pattern to the outside of any skeletal remains is not parallel to the situation where skeletal remains are present, despite being only a few centimeters from the section shown in panel C, and an additional 5 cm from the section shown in panel B.

Vertical separation and angulation of articulated and semi-articulated material in the Hill Antechamber Feature. Top: Plan view of articulated hand (next to stone), articulated foot with lower limb elements, and disarticulated hand and foot elements. Bottom: View of same elements from the west. The angle of the floor in a north-south profile is indicated, which contrasts with the angulation of the articulated foot, the disarticulated hand and foot elements, and the orientation of the stone.

Microtomographic sections within the Hill Antechamber Feature. (A) Cross-section through the U.W. 101-2074 jacketed mass. The excavation surface is top; the right side of this section were in contact with the larger U.W. 101-2076 mass prior to excavation. In (A) small, partially fragmented bone elements are visible at the top of the section, embedded within a sediment mass with low radiodensity (darker on this image). Below these bones and low radiodensity sediment is higher radiodensity sediment characteristic of the Hill Antechamber Feature as a whole. The highest radiodensity (lightest in color) objects in the section are LORM clasts, with a distribution characteristic of the feature as a whole. Square shows the region illustrated in panel B. (B) Detail of section shown in panel A showing sediment interface below bones, with invertebrate burrows visible. (C) Section near that shown in panel A, with similar features but gradient of sediment radiodensity. The interface shown in these panels is interpreted as invertebrate activity and diffusion or percolation of material associated with decomposition of soft tissue or diagenesis of bone. (D) Section of U.W. 101-2074 with same orientation as in panels A and C, but without bone evident. Texture and composition is typical of Unit 3 sediment throughout the subsystem. (E) Section of U.W. 101-2074 with an isolated right talus visible at right. No sediment contrast or halo is evident near this bone, which was likely isolated at the time of burial in this position. (F) Section of U.W. 101-2075, with orientation as in other panels but different scale. The top of this section includes tibia and fibula shafts. As in panel D, the Unit 3 sediment here underlying skeletal material shows no evident layering and chaotic orientation of LORM clasts.

Excavation plan in the Dinaledi Chamber. The location of the three excavation areas described in this study are shown, together with a plan view of the skeletal material in the Puzzle Box area and the Dinaledi Feature 1 area.

Dinaledi Feature 1. (A) Photogrammetric model of excavated unit including Dinaledi Feature 1 and part of Dinaledi Feature 2. These data derive from a time before the cutting of the profile on the south edge of the feature as shown in Figure 15. (B) Schematic showing excavated area, neighboring areas, and skeletal material. All skeletal remains are shown including those identifiable within the unexcavated feature and those excavated from above the current level. While there is additional material visible within the unexcavated portion of the feature, the boundaries and remaining depth of such material is not yet known. (C) Positions of excavated elements as viewed from east excavation face. Elements remaining in the feature are not included in this view.

Skeletal element localization in the Dinaledi Feature 1.

Skeletal part representation in Dinaledi Feature 1. Excavated fragments are shown in dark gray, in addition to the portions of mandible that are unambiguously identifiable within the site. Other material, both within the deposit and within the excavated collection is identifiable to region but not necessarily to element. For example, a partial cranial vault is in place, many ribs are represented, some partial phalanges and fragments of metacarpals, and some vertebral fragments. These are not indicated in the diagram; none of them duplicate any identified element. Excavated elements are listed in the Supplementary Information. At least two postcranial elements attributable to a juvenile individual are represented; these are indicated at right. Recording forms after Roksandic et al. (2003).

Sedimentology and stratigraphy of unlithified mud clast breccia and laminated orange-red mud clasts surrounding Feature 1. (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 (UMCB) 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. Note that (D) and (E) have been intentionally rotated 90° right from their in situ position in (C) for easier viewing of microstratigraphy of LORM.

Comparative analysis of sediment particle size distributions and sorting characteristics. (A) Particle size distribution curves for sediment around Feature 1, showing volume percentage as a function of the mean grain size. (B) Violin plots representing the mean grain size in µm for each sample group. C) Violin plots illustrating the sorting of sediments in each sample group. In violin plots, internal box plots show the interquartile range and white dots denote the median sorting value.

Harker variation plots showing the relationship between CaO content (wt.%) and other major oxides (wt.%) and loss on ignition (LOI, wt.%) for the five sediment groups analyzed within and around Feature 1. The Pearson correlation coefficient (r) for each sediment group is provided, indicating the strength and direction of the linear relationship between CaO content and the respective major oxide or LOI.

Harker variation plots illustrating the relationship between CaO and selected trace elements for the five sediment groups analyzed within and around Feature 1.

Harker variation plots illustrating the relationship between Zn and selected trace elements for the five sediment groups analyzed within and around Feature 1. SB plots distinctly apart from other sediment groups, which tend to overlap or are mingled for most trace elements.

Principal component analysis (PCA) of the geochemistry of the five sediment groups analyzed within and around Feature 1. FS2280 is a sample of unlithified mud clast breccia (UMCB) from the Puzzle Box excavation area. (A) Scatter plot of PCA showing the distribution of major oxides and loss on ignition (LOI) over the first two principal components (PC1 and PC2). (B) Scatter plot of PCA showing the distribution of trace elements and rare earth elements (REEs) over PC1 and PC2. (C) Biplot of PCA loadings for major oxides, selected trace elements and REEs, LOI and mean grain size.

Matrix-supported elements within Dinaledi Feature 1. (A) Photogrammetric model of unexcavated elements. (B) Detail showing ribs including those matrix-supported in subvertical orientation.

Skeletal material from the Puzzle Box area in position. Elements included in these views are greater than 2 cm in their maximum dimension; isolated teeth and smaller fragments are not included. The rectangular box at left represents the location of a partial infant skeleton, discussed in text. Top: plan view. Bottom: View from east side of excavation area.

Spatial data on element positions within Puzzle Box area. (A) Photo taken during excavation. (B) The combination of white-light surface scans and high-resolution laser scans representing the same excavation stage as photo in A. (C) Elevation data for elements and surrounding context. (D) Three-dimensional coordinates of identified elements in plane with X and Z dimensions. (E) Three-dimensional coordinates of identified elements in plane with Y and Z coordinates. Scale in millimeters.

PLACEHOLDER Spatial orientation and clustering of elements from the Puzzle Box area. (A) Spatial locations of elements excavated from the Puzzle Box area in 2013 and 2014 in oblique 2.5D view. For larger specimens 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 of fossil material collected on the surface of the Dinaledi Chamber together with excavated remains. A single high density cluster comprises the majority of excavated fossils (green) with two smaller peripheral clusters (purple, gold) and outliers (red points).

Spatial positions of elements attributed to the DH7 skeleton. Most recognized elements are localized within the western part of the Puzzle Box area, particularly ischium and lower limb elements, with a mandibular fragment and humerus fragments all within 10 cm. The more distant elements are hand bones.

Spatial positions of elements attributed to the DH1, DH2, DH3, and DH4 crania within the Puzzle Box area. An outline of the DH2 vault outcropped upon the surface at the time of excavation and the portions preserved from the feature were enclosed in sediment. This partial skull was underlain by the left half of the DH3 cranium, with the left hemimandible in contact with it and with other fragments located within 20 cm at a similar depth. Fragments of the DH4 cranium were recovered at greater depth, localized toward the east side of the area, with pieces of the left frontal approximately 15 cm from a concentration with most other fragments. The DH1 cranium is the most complete but was highly fragmented at the time of discovery, with many refitting parts displaced across 35 cm and a range of depths, including the occipital and temporal fragments in the deepest part of the deposit.

Postcranial elements for which conjoining fragments are spatially displaced from each other in the Puzzle Box area. The processes leading to displacement of these fragments did not result in significant loss of detail of fracture surfaces, enabling direct refitting to match them. The processes causing movement of these fragments after fracturing were capable of changing depth, compass bearing, and plunge angle of fragments, including those not supported by contact with other underlying fragments. However, those processes did not move any of these fragments more than 15 cm apart.

Context of DH1 fragments with U.W. 101-1475 femur. The DH1 occipital and parietal bones are fragmented into pieces that vary from ∼1 cm to > 6 cm in diameter. (A) Diagram showing all DH1 fragments with U.W. 101-1475 from south direction. U.W. 101-1475 is a proximal femur emplaced vertically in the deposit, with its distal end in contact with multiple fragments of DH1 occipital and parietal bones. (B) Photograph at time of excavation of the deepest DH1 fragments showing contact and relative positioning of these fragments with U.W. 101-1475.

Postcranial shaft fragments in vicinity of DH1 cranial fragments and articulated Foot 1. The range of plunge angles of these fragments could not have been maintained without sediment support at the time they attained these positions.

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 were collected.