Early presence of Homo sapiens in Southeast Asia by 86–68 kyr at Tam Pà Ling, Northern Laos

The timing of the first arrival of Homo sapiens in East Asia from Africa and the degree to which they interbred with or replaced local archaic populations is controversial. Previous discoveries from Tam Pà Ling cave (Laos) identified H. sapiens in Southeast Asia by at least 46 kyr. We report on a recently discovered frontal bone (TPL 6) and tibial fragment (TPL 7) found in the deepest layers of TPL. Bayesian modeling of luminescence dating of sediments and U-series and combined U-series-ESR dating of mammalian teeth reveals a depositional sequence spanning ~86 kyr. TPL 6 confirms the presence of H. sapiens by 70 ± 3 kyr, and TPL 7 extends this range to 77 ± 9 kyr, supporting an early dispersal of H. sapiens into Southeast Asia. Geometric morphometric analyses of TPL 6 suggest descent from a gracile immigrant population rather than evolution from or admixture with local archaic populations.

). The cave is situated in a limestone karstic tower known as Pà Hang, located on the south-eastern side of P'ou Loi Mountain. The latter belongs to the septentrional Annamitic Chain, oriented broadly NNW-SSE. The nearest major city is Vieng Thong ( Supplementary Fig. 1), also known as Ban That Hium on Google Earth mapping. From there, the site can be reached traveling north by road for ~20 km along the Nam Kam River to the small Hmong village of Long Nguapha. Surrounding Pà Hang, there are multiple cavities explored by us in the past and located in a radius of 1 km from the center of the village.

Geology
The geology of TPL consists of Palaeozoic granite and diorite 1 covered with a widespread folded poorly metamorphic sedimentary formation characterized by gray to yellow pelites, siltites and arkosic sandstones attributed to the Silurian, and perhaps partly to the Devonian 1 . This poorly metamorphic unit is capped by a thick limestone unit, which is badly dated at present, but which has been attributed by Saurin 1 to an interval between the Late Carboniferous (Moscovian) and the Permian. The limestone unit consists of a pluri-decametric massive sparitic dark-gray carbonate without marl intercalation. Coral fossils are quite common according to place. On the way to the fossiliferous sites starting from Vieng Thong, the road successively incises through the three main formations of the region, namely the granite massif, the weakly metamorphic folded series, and the carbonate unit. In the studied region, the landscape consists of tower karsts formed through the dissolution of Moscovian limestone beds, containing a dense network of caves and galleries. Most tower karsts in the region display widespread karstic networks and at present at least half a dozen sites rich in fossils have been discovered and are in the process of being analyzed.  Fig. 2). From the main chamber, close to the roof, exists a second network ~5 to 10 m wide, 2 to 4 m high and 100 m long. At the end of the gallery, a very narrow gallery ~20 m long crosses the mountain to another outside exit ( Supplementary Fig. 2).
In the main entrance and down the main talus slope, the presence of large limestone blocks and slabs testify to collapse of the roof of the cave through slab failure, allowing the formation of the current large opening of the cave. The collapse of a part of the roof at the entrance is argued by a cutting of the roof according to the directions of diaclases as well as by the presence of stalactites much smaller (younger) than inside the cave ( Supplementary Fig. 3). We observe layers in the stratigraphy that clearly lap up against and drape over the limestone blocks in the lower part of the sequence, consistent with the collapse of the cave preceding the deposition of the fine-grained sandy clays and clay silts. In this sense, the original cave floor strewn with limestone slabs formed the topographic template that governed subsequent sedimentation.
Further, we do not record signs of deformation of the fine-grained units beneath or adjacent to the large limestone slabs, which would be expected if they had fallen on to this rather plastic sediment.
In 2008, three test pits were opened on the cave floor to examine stratigraphy and sedimentology: Trenches 1 and 2 on the west side and Trench 3 on the east side. Trench 3, which can be reached by a steep slope ~60 m in length, is where all the fossils so far reported from TPL were found. In 2019, Trench 3 was extended east towards the wall of the cave, and by 2020, the excavators reached the limestone base of the TPL cavity, exposing a 7 m deep profile from which the stratigraphy and sedimentology is examined.
The TPL sediment consists primarily of brown silty clays that alternate with coarser, sandier layers.
The sandy layers are themselves intermixed with clay and exhibit a very slight color change (Stratigraphic section Fig. 3). The main cavity of the cave does not have any other obvious connections to the wider karst system which leads us to believe that the primary or only source of sediment can be via the current main entrance. During heavy rainfall we can clearly see water running down the slope, and occasionally small ephemeral pools of water temporarily form at the base of the cave. It is evident that the sediment originated near the cave entrance and was introduced to the cave interior via relatively slow, low-energy slopewash transport ( Supplementary Fig. 4), although changes in sediment characteristics indicate that higher-energy processes did sometimes occur.
Clay-rich layers contain small limestone clasts and iron-oxide pisolites. The former is derived from the host limestone bedrock (the autogenic component), whereas the quartz, clays, and secondary iron oxides are primarily derived from the platform in front of the cave entrance where these elements are present in abundance. The clay composition (vermiculite and kaolinite associated with iron and aluminum oxides) indicates hydrolysing conditions that are characteristic of "ferralsols-nitisols," which are strongly weathered soil types that suggest that they were deposited in a humid, subtropical climate.
A first profile was documented during initial excavations of Trench 3 (right part of Stratigraphic section Fig. 3). Subsequent excavations of Trench 3 near the East wall of the cave led to the documentation of a second profile, which is ~5 m from the first (left part of Stratigraphic section Fig. 3). By correlating the two profiles we find that the stratigraphy and sedimentation are consistent between the two trenches and the individual strata and laminae are contiguous and can be followed laterally across the excavated area of the chamber; alternations of sandy and silty clay layers are punctuated by ~12 calcareous layers-similar to moon milk-which are sometimes cemented; and that there is no evidence in the excavated strata of bioturbations, post depositional modifications, soft deformations, reworking or mixing. Ongoing microstratigraphic analyses by MWM and VPCH is likely to support many of these observations. The ongoing microstratigraphic analyses by VPCH and MWM have identified some combustion residues in the sediments sampled from the main excavation areas 2 that may suggest human activity in close vicinity of the cave or even from areas of the cave now covered in limestone slabs. Small traces of calcareous ash which are not likely to survive longer distance transport (e.g., downslope from the entrance area) as well as intact charcoal fragments that retain morphologies consistent with very short distance travel have been identified in thin section using optical microscopy and further work will quantify the amount and context of this material. However, despite ongoing microstratigraphic analyses, and the presence of the H. sapiens skeletal material found in the sediments, Tam Pà Ling was not likely an occupation site. No artifacts have been found in the cave and at this stage we must presume that while they were clearly out in the landscape, the cave was either not habitable or accessible at the time that these early humans were present in this area.

Magnetic Susceptibility
To assess inputs of pedogenic material from the cave mouth area, we have used magnetic susceptibility (MS) as a proxy for intensity of pedogenesis and the reworking of this material into the cave. Previously, MS has been used as a palaeoclimatic proxy in archaeological cave settings (e.g., refs. 3,4 ), but these studies serve to highlight the issues with disentangling elevated MS values being related either to pedogenesis or burning occurring at the site. Warmer and wetter climatic conditions can certainly accelerate pedogenesis, forming magnetic minerals such as maghemite and magnetite and resulting in elevated MS values.
Conversely, when cooler and drier conditions prevail, a reduction in pedogenesis occurs and the MS signature is correspondingly lower 5,6 . As the East Asian Monsoon (EAM) controls much of Southeast Asian rainfall, MS was performed on bulk sediment samples from TPL to determine if-and to what extent-MS serves as a proxy for wetter or drier conditions and the flux of soil into the cave interior.
A total of 39 bulk sediment samples (18 from  MS values support observations that periods of increased precipitation (resulting in increased pedogenesis) correlate with greater soil delivery to the interior of the cave from the cave mouth area during periods of greater rainfall and slopewash activity. This is supported by the rise in MS values between 5.32-3.72 m (P2), with sediment at these depths comprising predominantly silty clays (Stratigraphic section Fig.   3). The same pattern is observed between 3.4 -1.8 m (P1) and 3.52 -1.92 m (P2), where higher MS values occur when sediment is predominantly composed of silty clays. We note that higher MS values are less pronounced in P2 than P1, likely due to the fact these sediment columns are laterally 5 m apart (Stratigraphic section Fig. 3) and therefore other spatial aspects are controlling sediment delivery.
The MS data suggest a decline in overall moisture availability during much of MIS 4, potentially linked to higher seasonal variability related to a weakening of the EAM. A similar decline in rainfall in MIS 4 is also observed by Bacon et al. 8 at Duoi U'Oi, northern Vietnam. Wetter conditions return towards the MIS 4-3 transition related to a strengthening of the EAM, which aligns well with similar interpretations by Milano et al. 9 . A steady rise in MS values during MIS 3 (peaking around 2.8 m, ~46 ka), declining towards the end of MIS 3 and into MIS 2 also corresponds with the general drying trend observed by Milano et al. 9 at Tam Pà Ling and Yamaoh et al. 10

Bulk organic and carbonate content (Loss on ignition)
To further interrogate the formation of the TPL stratigraphy, the loss-on-ignition (LOI) technique was applied to TPL bulk sediment samples to provide a broad estimate of the amounts of organic matter (OM) and carbonate content (CC) in the sediment [12][13][14] . When combined with other sedimentological parameters, such as magnetic susceptibility (MS), LOI data can assist in determining the source and character of sediments deposited in caves.
LOI was undertaken on the same 39 bulk samples analysed for MS following standard protocols 14 using a time-programmed Across International CF1200 series muffle furnace with a 36-L chamber capacity.
Results were calculated using the below equations from the same protocol 14  The difference between amounts of OM and CC at similar depths in each profile are due to their positions in the cave (Stratigraphic section Fig. 3). The higher CC in Profile 2 is attributed to its location near the cave wall, with more calcium carbonate being transferred into the sediment in this area of the site.
Conversely, the higher OM in Profile 1 is most likely related to its location towards the base of the slope beneath the cave entrance, which regularly receives allogenic clay-rich colluvial slopewash sediments ( Supplementary Fig. 4). Fig. 6) recorded between ~6.92 -5.52 m suggest relatively dry colluvial input, probably originating from in and around the cave entrance ( Supplementary Fig. 4). This The elevated amounts of CC we record in the sediment (Supplementary Fig. 6) coincide with drier conditions as observed from the MS data. This is characterized in the stratigraphy by increased frequency of limestone clasts which may relate to periods of increased slab failure (autogenic limestone generation) (Stratigraphic section Fig. 3). These are often immediately followed or accompanied by an increase in the amounts of OM such as ~6.52, 5.92, and 5.52 m in Profile 2, and ~3.4 and 2.2 m in Profile 1 suggesting that these slab failures in the sequence relate to episodes of cave opening which then respectively allows greater deposition of organic-rich clays further into the cave. The observations on these aspects of the sedimentological record of TPL support the idea that it was a collapse of the cave ceiling that formed its large cave opening. Although these also suggest that it took several of these events to gradually open the cave to its current size beginning from at least ~86 ± 8 kyr.
unicornis, Dicerorhinus sumatrensis, have been recovered in the ~7 meters high stratigraphy. Among these twenty-eight teeth, only a lower third molar of Capricornis sp. was still fixed in a fragmentary jawbone, the others being isolated teeth.
Regarding the two teeth used for dating, TPL 73 is a germ of a lower molar (M1 or M2) with an incomplete crown, whereas TPL 74 is a lower molar (M1 or M2) displaying incomplete roots partly gnawed by rodents. Both molars show the typical morphological pattern of Caprinae, with smooth enamel (not ridulated as in cervids), and stylids more prominent than main cusps (metaconid and entoconid). On TPL 74, the ectostylid is absent.
There is no evidence that these Caprinae could have been prey hunted by the Homo sapiens individuals recovered at the same level in the TPL stratigraphy. There are no archaeological artifacts, no animal bones with cut marks, no traces of human activities, and the analysis of deposits suggests that the cave was not an occupation site. The presence of teeth of large mammals results from the same depositional process as that of human remains in the cave. Moreover, the roots of most isolated teeth of large mammals show gnawing traces of rodents, most likely porcupines, which indicates that they were the last accumulator agents. Such traces of activities of porcupines are a common phenomenon 16 , as observed in the mammalian assemblages in the same area, Nam Lot (86-72 ka) and Tam Hang South (94-60 Ka) 17 , from breccias deposits of the Pà Hang mountain 18 . The animal teeth, being gnawed, and the human remains, showing no trace of porcupine activity, document different taphonomic histories, which remain difficult to explain and could lead to the assumption that the human remains have been buried outside of the cave [19][20][21] and later washed in, which would only be putative at this stage of the analysis.
However, their presence along the 7meter high section of Tam Pà Ling results most likely from the same depositional process with clay-rich slopewash deposits 15 .  Table 9 for landmark and curve and surface semilandmark definitions.

Supplementary Table 1. Results of the cross-validated linear discriminant analysis (LDA) and
posterior probabilities using the H. sapiens sample and three groups (early, Late Pleistocene, and Holocene). The highest classification rates occurred in the mandible, followed by the maxilla, frontal, mandible, and anterior corpus. Highest probabilities in bold.

Dataset
Correct classification