This work was done within my MA thesis project entitled Techno-typology and chronometry in the Late Mesolithic of the Sado valley: the case of Amoreiras shell-midden (Alcácer do Sal), directed by Prof. Mariana Diniz (Lisbon University, Faculty of Letters) and concluded in 2012.

1 - The Amoreiras shell midden: an overview of the data and problems

1The known shell middens of the Sado valley are located along a roughly 15 km long stretch of the river, at 40-50 km from the present estuary, in the municipality of Alcácer do Sal. The shell midden of Amoreiras (Cabeço das Amoreiras or S. Romão) – is located on a hill at the left margin of the Sado River (figure 1). It was excavated during the 1950’s and 1960’s by Manuel Heleno, director of the National Museum of Archaeology at the time (MNA). Thousands of archaeological materials were recovered, but the results were not published and the artifacts were only preliminarily studied in the 1980’s under the research directed by J. Arnaud (1989, 2000).

2In spite of the scarcity of systematic studies, Amoreiras was assumed as one of the most recent Sado shell middens and a place of interactions between Mesolithic and Neolithic groups. The presence of ceramic fragments with Cardial decoration in the lowest levels of the area excavated by

3J. Arnaud, and the radiocarbon dating results, led this researcher to defend an occupation during the transition from the 6^th to 5^th millennium BC, when the neolithization process was occurring (Arnaud, 2000). G. Marchand, focusing on the numerical dominance of segments among the geometrics referred by J. Arnaud (1989), inserted this site in the last phase of the Late Mesolithic chrono-typological model proposed for the Central-Southern Portugal (Marchand, 2001, 2005).

4More recently, the dating of the “skeleton 5” from Amoreiras – Beta-125110: 7 230 ± 40 ^[14]C BP (Cunha, Umbelino, 2001) – led to retreat the site occupation to the final of the 7^th – beginning of the 6^th millennium cal BC (Diniz, 2010). The ceramic analysis, which suggests different cultural occupations of the site during the Neolithic (Diniz, 2010), increased the controversy.

5Given the problems of the Amoreiras site, a systematic analysis of the lithic remains from the excavations carried out by Heleno appeared to be imperative.

2 - The lithic industry of Amoreiras: excavations by Manuel Heleno

6A sample of materials recovered from the central area excavated by Heleno (“extension of the survey A”) was analyzed, following the propositions of the “chaîne opératoire” concept. The lack of contextual data was restrictive to this study, since the excavations were done through artificial levels, leading us to consider the materials as a whole.

7The analyzed 1 592 lithic remains document all the technological categories of the “chaîne opératoire”, representing a microlithic industry, with scarce variability and quantity of retouched tools (table 1; figure 2).

Table 1- Lithic technological categories from Amoreiras (ext. of the survey A), by level (total assemblage, including fragments).

Category	Level 1	Level 2	Level 3	Total
				%	N
Cores	34	21	11	415	66
Core preparation/maintenance	9	11	1	13	21
Debitage products	285	178	46	3 197	509
Blades	5	4	0	057	9
Bladelets	212	120	35	2 305	367
Flakes	68	24	11	835	133
Retouched tools	133	100	36	1 690	269
Geometrics	71	53	18	892	142
Retouched bladelets	26	20	10	352	56
“Common fund” tools	14	16	2	201	32
Debris	365	335	27	45,67	727
Fragments	331	281	24	3 995	636
Chips	12	33	2	295	47
Microburins	22	21	1	276	44
Total	826	645	121	10 000	1 592

Materials deposited in the National Museum of Archaeology – MNA

8The main knapping purpose would be the bladelet debitage from prismatic cores. Part of the bladelets was transformed, mainly into highly normalized geometric microliths. Segments are dominant (66.9% of the geometrics), mostly symmetrical forms (figure 3). The quantity of unretouched bladelets (23.1% of the whole lithic set) suggests that they would be used without transformation. The debitage of small and medium‑sized flakes is also attested, although less expressively and mostly remaining unretouched.

9The application of the prismatic method appears prevalent with the major exploitation of prismatic cores (72.7% of the complete ones), often until their exhaustion. Less expressive, the use of the ‘random’ method (Carvalho, 1998) to extract non-standardized flakes and scars is visible in some inform / polyhedral cores (16.4%) and flaked cobbles (2 exemplars). The production of tools (endscrapers and stone drills) from thick flakes, or directly using the cobbles as supports, is also significant.

10The data does not show the existence of operative schemes depending on the raw material type. In spite of the diversity of the available rocks, there is a preferential se of siliceous slates and cherts for all the technological categories (table 2). Those rocks would be locally accessible as cobbles, in secondary deposits. A spared use of the best siliceous rocks from the core maintenance material exclusively obtained on chert is also visible.

Table 2 - Representation of the raw-material types in Amoreiras (%), by technological category

Category	Chert	Siliceous slate	Jasper	Porphyr	Quartz- diorite	Quartz	Hyaline Quartz	Quart- zite	Other	Undeter- mined
Cores	3 788	4 242	152	303	303	152	303	455	000	303
Core preparation maintenance	5 238	4 762	000	000	000	000	000	000	000	000
Debitage products	3 143	4 774	570	118	098	354	806	039	020	079
Retouched tools	4 387	4 275	892	000	074	000	186	037	000	149
Debris	2 077	5 791	922	055	055	289	646	014	000	151
Total % (relative to the total assemblage)	2 927	5 132	760	075	082	251	597	044	006	132

11Some signs of pressure bladelet debitage are present (table 3), but not in a dominant way. The employment of other flaking techniques, individually or in conjunction (Pelegrin, 2012: 467), is probable, namely the indirect percussion (table 4). Some flakes with big butts, pronounced bulbs and ripples (Carvalho, 1998: 79) also suggest the use of direct percussion. Although scarce, macroscopic signs of heat treatment were detected (thermic luster and homogenous texture, sometimes associated to more aggressive thermal alterations, following Boix Calbet, 2012) in 8.3% of the assemblage, mostly on chert. This procedure is attested in all the technological categories, namely on bladelets and microliths (table 5). Although often detected in several Neolithic contexts of the Southern Iberia (Manen et al., 2007), the cultural origin of the heat treatment in Amoreiras – Mesolithic or Neolithic – is not clear, due to the scarcity of stratigraphic data.

12The analysis shows the occurrence of knapping operations at this site, by using local resources and trying to spare both the provision and the manufacture efforts. The uniformity of the lithic set is visible, although there are some dissonant elements, like the heat treatment and a blade fragment with a pecked butt.

Table 3 - Characters of the bladelets from Amoreiras suggesting the use of pressure technique

Character	%	Reference
Paralel edges	35.37 (complete bladelets)	Inizian et al., 1999
Narrower butt than the max. width	5.38 (MNI)	Inizian et al., 1999
Absent ripples	74.39 (total)	Pelegrin, 1984; Inizian et al., 1999
Concave or plunging distal profile	59.76 (complete bladelets)	Binder, 1987; Brunet, 2012
Small and pronounced bulbs	46.24 (MNI)	Tixier, 1984; Binder, 2012
Facetted butts	17.20 (MNI)	Binder, 1987
Heat treatment	11.44 (total)	Inizian et al., 1999; Boix Calbet, 2012
Character	mm	Reference
Mesial thickness (constancy)	2.36 ± 1.01 (MNI)	Inizian et al., 1999

Table 4 - Characters of the bladelets from Amoreiras suggesting the use of indirect percussion technique

Character	%	Reference
Flat butts	43.55 (MNI)	Binder, 1987; Carvalho, 1998
Fairly pronounced bulbs	41.94 (MNI)	Tixier, 1984

Table 5 - Heat treatment incidence (%) on the lithic industry of Amoreiras (total assemblage)

Technological category	%
Core preparation and maintenance	2 381
Cores	2 000
Bladelets	1 144
Blades	000
Flakes	602
Retouched tools	1 375
Debris	399

3 - What cultural value for the lithic industry from Amoreiras?

13Amoreiras has a long occupational diachrony - a burial dated to the final of the 7^th millennium cal. BC, two radiocarbon dates fitted between the second half of the 6^th and the beginning of 5^th millennium cal BC, some Cardial ceramic fragments, vessels from different phases of the Neolithic, and finally, a microlithic industry with a clear predominance of segments.

14The absence of domestic fauna (Albizuri Canadell, 2010) suggests that the post-Mesolithic occupations were incipient, but they cannot be excluded from the interpretation.

15Thus, the inclusion of Amoreiras exclusively in the final phase of the Late Mesolithic, or its generic assumption as one of the recent shell middens of the Sado valley appears to be a malad-justed hypothesis. The late character of segments from the Mesolithic lithic assemblages also needs to be questioned, given the lack of stratigraphic data from the Sado shell middens.