Papers from the Institute of Archaeology Late Holocene lithic points from a Southern Brazilian mound: The Pororó site

. Late Holocene lithic


Introduction and background
The diversity of lithic points associated with hunter-gatherer industries in Southern Brazil has only started to be understood in the last decade, after researchers start questioning the currently outdated notion of "Umbu Tradition" (see, for example, Okumura & Araujo 2016; Moreno de Sousa & Okumura 2018; Garcia 2019a; Moreno de Sousa 2020, for a review of the subject). Although much work remains to be done, technological studies on Early Holocene sites revealed five types of points, of which three are associated with the same archaeological culture. The Garivaldinense Culture is characterised by the recurrent presence of Garivaldinense points, Montenegro points and Brochier points at sites dating from the Early to Middle Holocene in Central and Eastern Rio Grande do Sul State, Brazil. The fourth type of point, known as Star type points, are known from findings in Paraná state and are associated with the Tunas Culture. The last type is known as the Rioclarense type, and is associated with the Rioclarense culture (Moreno de Sousa 2020; Moreno de Sousa & Okumura 2020). These studies have corroborated the hypothesis of a greater diversity of lithic industries in southern Brazil than previously recognised.
The Garivaldinense Culture was firstly described in the Garivaldino Rockshelter Site (a.k.a. RS-Tq-58 site), located in Rio Grande do Sul state (Mentz-Ribeiro & Ribeiro 1999;Cheliz et al. 2020;Moreno de Sousa & Okumura 2020). Among the lithic types found, the Garivaldinense type points are the most common. Their length varies from 25-45mm, their width from 15-25mm, and their thickness from 5-9mm. Their width-thickness proportion varies from 2.1-3.1:1. In morphological terms, they usually present triangular bodies with irregular or straight edges, straight shoulders, and straight or bifurcated or convex stems. In technological terms, they are produced using the bifacial reduction method and the percussion technique, finished by pressure retouching. Reduction negatives are usually organised as selectivetrespassed or convergent untrespassed. However, some points are not reduced at all, in cases where thin flat flakes are just retouched. Here we refer to trespassed negatives as those which take more than half of the artefact width, which is different from overshot negatives. The latter refers only to negatives that take the entire width of the artefact, even removing a little bit of the opposite edge. Overshot flaking is not observed in these points.
Brochier points are also common, but in lower frequency in relation to the Garivaldinense ones. Their length varies from 22-34mm, their width from 9-15mm, and their thickness from 3-7mm. Their width-thickness proportion varies from 1.7-2.9:1. In morphological terms, they are usually lanceolate, tapered or irregularly shaped, and unstemmed. In technological terms, they are produced by retouching of agate thin flakes.
The Montenegro points are the least common type of point found associated with the Garivaldinense culture. Their length varies from 20-36mm, their width from 10-16mm, and their thickness from 5-7mm. Their width-thickness proportion varies from 2.1-2.5:1. In morphological terms, they usually present a triangular bladeshaped body with straight but serrated edges and a bifurcated stem. In technological terms, they are produced using the bifacial method and the pressure flaking technique, such that negative scars form a median ridge.
According to Moreno de Sousa (2020), many other sites from this region would also probably become affiliates as part of the Garivaldinense industry once technological studies on the lithic points begin to be conducted. In this study we present new technological data from the Pororó site, located in the proposed Garivaldinense coverage area, and presenting at least two of the three types of points associated to this Paleoamerican culture.

Pororó site characterisation
The Pororó site is an artificial mound located in central Rio Grande do Sul State (Brazil), in the municipality of Pinhal Grande, 10km from the right margin of the Jacuí river (Figure 1), atop of the Serra Geral formation between the valleys of Jacuí tributaries, but still with the Lajeado Pororó valley. The site was first recorded by the owner of the farm where the site is located, Mr. Lucas Somavilla. The owner had previously found some archaeological materials on surface, such as bolas (a.k.a. boleadeiras) and lithic points, in the region. According to him, small concentrations of materials were also found during the building of houses and artificial lakes (weirs) in the farm. However, it was only between 1998 and 2002 that he found a substantial concentration of archaeological materials, during the expansion of his house ( Figure  2). Somavilla, interested in the study of his findings, contacted the team of archaeologists of the former Laboratory of Archaeological Studies and Research of the Federal University of Santa Maria (LEPA-UFSM).
Somavilla generously donated his findings to the LEPA-UFSM. The material has been recently moved to the Museum of Pinhal Grande. The site excavations were carried out in 2010, directed by Professor Saul Milder (Figure 3). The excavations revealed an anthropic construction of the mound. The most relevant features in the identification of the mound construction were the presence of boulders and other rock fragments used as building materials and, especially, the artificial accumulation of the typical archaeological "dark earth" -dark humic clay sediment formed by human activities and high presence of organic material (Figure 4). The excavation reached a maximum of 42cm in depth and a homogeneous stratigraphy. A fixed anvil was also identified in the basalt outcrop next to the mound (Garcia 2016;.

Materials and methods
A total of 10,026 of lithic vestiges were originally identified, of which 52.9% was described as construction materials for the mound. The remainder were identified as flakes or fragments from tool production (Garcia 2020). The artefact assemblage is composed by five pyramidal cores, two scrapers, 27 points, one point recycled as a scraper, three points fragments, and 11 bifacial preforms. Our study focused on the technology of the 27 points and the classification of them into types. Of the 28 points, 20 were found by Mr. Somavilla (including the one which had been recycled as scraper). According to him, they were found when digging to flattening the terrain for the construction of the second house (west on the map). The other eight points were found during the excavations conducted by Professor Saul Milder, all which were discovered in southern zone of the mound (close to the second house).
Previously, Garcia (2019b) presented a diachronic description -also referred to by some authors as diacritic analysis -of each one of these points. In this study, we followed procedures for the technological study outlined by Moreno de Sousa & Okumura (2020). Protocols include the observation and measurement of many metric, morphological and technological attributes, bivariate and multivariate statistics, as well as the classification of artefacts into types according to their features' patterns. We do not aim to present a complete description of this method here, since it has already been described in Moreno de Sousa & Okumura (2020) work.
In this study, the statistical analysis (descriptive, frequency and bivariate or multivariate analysis) was not performed due the small size of the samples. We considered that the number of specimens of each type were insufficient for an accurate regional comparison of the technology by bivariate analysis to the dataset presented by Moreno de Sousa & Okumura (2020). Other types of analysis (e.g., principal component analysis) that consider only the metrical attributes could be performed, but they do not take into account technological attributes. Since our objective in this paper is to discuss the technology of the artefacts, these other types of quantitative analysis were not carried out. The data presented in this analysis is available to any researcher to use for such purposes, and we welcome researchers to use these data.
In order to classify these points into previously known types, we considered the individual cultural patterns (metric, morphological and technological attributes) of each specimen. The features observed in each artefact are presented in tables.

Data results
We identified three distinct types among the points discovered at the Pororó site: Garivaldinense type ( Figure 5; n = 11), Montenegro type ( Figure 6; n = 8), and a third, previously undefined, type (Figure 7; n = 5) which we propose to call the "Pororó type". We also identified three points that could not be classified into any of these three types ( Figure 07). In fact, one of these points (Por-27) seems to present the standard features of Pay Paso points, usually found in northern Uruguay. And at least one point was reshaped as a scraper (Figure 07).     Table 2 presents the features of the Garivaldinense type points from the Pororó site. Table 3 presents the features for the Montenegro type, Table 4 presents the features for the Pororó type, and Table 5 presents the features for the remaining points.

Interpretations of the data
Although we could classify most of the points into the Garivaldinense and the Montenegro types, some differences in relation to the original definition of Moreno de Sousa & Okumura (2020) could be noted (Figure 9).
In general, the Garivaldinense points from the Pororó site are relatively thinner, with proportions that are usually larger than 3.1:1. This could be related to the quality of raw material, since the points larger than 3.1:1 are made of flint or were made on flakes that were thin enough and were not reduced.
The Montenegro point are the ones which present the most differences, particularly with regard to size and shape. Points that present the usual Montenegro body shape are bigger in comparison with the ones used for the definition of the type (the Montenegro "holo-types"). This could be simply a regional or chronological variation, or suggest that points of this type became longer through time. Some points present bodies that are less "blade-like" and wider. This could also be a chronological variation of this type.     The points we proposed to call as "Pororó type" could be a Late Holocene innovation within the Garivaldinense lithic industry, but a more accurate chronological distribution is still lacking. These points have dimensions that variate between 32 and 42mm in length, 13 to 18m in width and 4 to 6m in thickness, with a width thickness proportion that varies between 2.8:1 and 3.6:1. In morphological terms, these points tend to present a triangular body with irregular edges, straight shoulders and concave or straight stems. In technological terms, these points were made by bifacial reduction by pressure, with parallel trespassed negatives, followed by bifacial retouch by pressure flaking. The Pororó points present metric and morphological features that are very similar to the Garivaldinense points. Technology, however, is quite distinct. In this sense, the Pororó points could be both an innovation within the Garivaldinense lithic industry, as well as a late technological variation of the Garivaldinense points, in which parallel pressure flaking is more commonly used than selective percussion.
Finally, there are three points that do not fit into any of the Garivaldinense, Montenegro or Pororó types. In fact, artefact Por-27 presents features that classifies it within the Pay Paso type definition proposed by Suárez et al. (2018), particularly in morphological terms. In technological terms, the Pay Paso point from the Pororó site is made by bifacial reduction by percussion, with parallel trespassed negatives in the body, convergent trespassed negatives in the stem, and finished by bifacial retouch by pressure flaking. Pay Paso points, however, have never been recorded at Brazilian sites before. According to Suárez et al. (2018), Pay Paso points are common in sites from northwestern Uruguay (along the Brazilian and Argentinian borders). In this sense, the presence of this type in southern Brazilian sites should not be particularly surprising. However, the Pay Paso points from Uruguay are dated between 12,000 and 11,000 BP. The presence of a single Pay Paso point at the Pororó site could suggest the persistence of this type from the Early to the Late Holocene. However, more studies on hunter-gatherer associated sites from Rio Grande do Sul state, especially in the central, western and southern regions, are necessary to understand the possible dispersion and persistence of Pay Paso points throughout Holocene in southern Brazil. The presence of a Pay Paso point in an industry more related to the Garivaldinense culture implies contact of the Pororó people with other groups. Garcia (2013) carried out a survey of lithic raw materials in the region. Sources of all types of raw materials identified in the lithic points from the Pororó site were found within a radius of 7km from the mound. However, sources of flint and silicified sandstone are scarce, with most of the pebbles and cobbles presenting internal fractures and/or dimensions that are inadequate for lithic point manufacture. According to Garcia (2013), the number of flakes and the presence of hammerstones and cores at the mound suggests that the points were produced there. Although no preserved evidence of faunal remains has been found, it is known that antlers were used for the production of lithic points, by both percussion and pressure flaking, including at sites associated to the Garivaldinense industry (Mingatos & Okumura 2020). For all the points where the blank was identified, they were made by the reduction of flakes, probably taken from pyramidal cores, considering this was the most common type of core identified in the site. However, the reduction of cobbles cannot be discarded, particularly on the points made from flint.
The presence of bolas (or boleadeiras) suggests these groups were highly mobile, probably moving to Pampean landscapes to capture animals, since bolas are not efficient on hunting of preys in the local forest. This mobility could result in finding other sources of raw material and probable contact with other cultural groups.
The production of lithic points of different types, although they can be used for the same purposes and are equally efficient as arrow or dart points, is probably related to different uses. Unfortunately, it is impossible to identify their use, since the shafts are not preserved. But it is logical to say, for example, that the biggest points were probably not used for small animals. The diversity of points could also imply their use in different landscapes.

Conclusions
Although the lithic material evidence, beyond the presence of the points, allow us to describe the Pororó site as a lithic workshop, the production of lithic artifacts was not the only local activity. All evidence suggests the mound as a habitation area (Garcia 2016), as was the case with other similar mounds in Uruguay and southern Brazil that have been called "Cerritos" or "earth mounds" (see Garcia 2019aGarcia , 2020. The hunter-gatherers from the Pororó site are probably some of the first groups to build mounds away from the coast in southern Brazil.
In the specific case of the Pororó site, according to the types of lithic artifacts identified by us, the hunter-gatherer inhabitants of the site were associated with the Garivaldinense culture. Compared to the Early Holocene Garivaldinense sites, differences probably indicate innovation across time and space, as well as adaptions and possible cultural exchange with other cultural groups. The diversity of types at the site is probably related to different functions or, alternatively, each type could be related to different social divisions of hunters or flintknappers within the group. However, only functional analysis has the power to elucidate these questions.
The region still lacks more complete study. According to Garcia (2016), similar artefacts were found in another four locations within the Lajeado Pororó valley. These locations may also contain mound formations, but only by interventions will it be possible to confirm this hypothesis, since the region is mostly characterized by hills. Even the Pororó site was only identified as a mound during the excavations on the top of a hill. In this sense, more studies are essential to further our knowledge of the region.