THE USE OF MOLLUSC SHELLS AS TOOLS BY COASTAL HUMAN GROUPS: The Contribution of Ethnographical Studies to Research on Mesolithic and Early Neolithic Technologies in Northern Spain

In European archaeology, the malacological remains recovered in archaeological contexts have traditionally been considered almost exclusively as food waste. In other cases, this view has been broadened in order to study these remains as an expression of aspects of the social organization of the human groups, based on the use of perforated shells as objects of personal ornamentation. However, the study of these natural resources as raw materials for the manufacture of tools aimed at satisfying the production needs of the human groups has been very limited. This little-developed aspect of research is at variance with the abundant ethnographic information from many different periods and geographical settings showing that malacological resources were used in many complex and varied ways. This paper is an attempt at compiling a small part of this ethnographic information — a contribution which, through its critical application to the archaeological record, is of interest in establishing a methodology for studying this type of evidence. In the specific case of northern Spain, information from ethnographic studies has been used to develop an appropriate methodology with which to approach the analysis of this kind of archaeological evidence, as recently documented for the first time at the classic site of Santimamine (Basque Country). At the same time, the documentation of shell tools could provide an explanation for the scarcity of (i traditional technologies "that characterizes many Mesolithic and early Neolithic sites in northern Spain.


JOURNAL OF ANTHROPOLOGICAL RESEARCH
Analysis of the corpus of information on the technological use of shells indicates that the traditional view that molluscs were exclusively food resources is biased, or at least incomplete, as some scholars have suggested for European sites (Cade 1991;Cristiani et al. 2005;Dupont 2006;Gruet 1993;Vigie 1987Vigie , 1992Vigie , 1995Vigie andCourtin 1986, 1987). This is especially true if we take into account the many, varied archaeological contexts, in different latitudes of the Pacific, Mediterranean, and Atlantic (e.g., Cordova Medina 1991;Dacal Moure and Rivero de la Calle 1984;Guarch 1973;Izquierdo Diaz 1991;Jerardino 1998;Jones O'Day and Keegan 2001;Mansur and Clemente 2009;Marquardt and Payne 1992;Mery et al. 2008;Moore 1921;Smith and Allen 1999;Szabo 2008;Szabo et al. 2007), which suggest molluscs were used not only as a food resource but as the raw material to produce tools.

ETHNOGRAPHICALLY DOCUMENTED USE OF MOLLUSC SHELLS
Many types of tools made from mollusc shells have been used historically by numerous human groups exploiting coastal environments. Some researchers have pointed out a differential use of gastropods and bivalves. Gastropods have been used as spoons, money, anvils, plates, glasses, weights for fishing-nets, mills, cups, gouges, hammers, spindles, and knives. Bivalves have been employed as knives, borers, and scrapers, among other uses (Claassen 1998:201).

Knives
Among the tools used in productive activities, we find shells used expediently as knives or as the raw material to make handled knives of different kinds (formal tools). In this respect, the use of Unio shells (freshwater mussels) as knives has been documented at Hokkaido (Leroi-Gourhan 1945), while the Naujamiut of Greenland used Ostrea edulis shells to cut and scrape seal skins. In a similar way, Eskimos from Alaska used shells to cut the entrails of marine mammals (Dupont 2003).
This kind of use has been well documented for the Yamana (Yaghan) from Tierra del Fuego. The knife-scrapers made by the Yamana had four parts: (a) the mollusc valve sharpened by abrasion, ( b ) an elongated beach pebble used as a handle, (c) a strap used to attach the handle to the shell, and (d) caulk or moss used to cushion the handle to soften the blows, and to retard breakage. This artifact was employed in many different activities: as a hatchet; to scrape; to carve wood; to cut up meat, wood, bone, or hide; or even to cut the umbilical cord of newborn babies (Gusinde 1986;Mansur and Clemente 2009). To carry out more delicate work in wood, bone, or hide, the Yamana also used Tagelus shells (Gusinde 1986).
Other aboriginal groups in Tierra del Fuego, such as the Selk'nam, are also known to have used tools made from mollusc valves prehistorically. In this case, Mytilus valves were used to make knives, which were later turned into scrapers (Mansur and Clemente 2009).
Similarly, in Brazil, we know that several bivalve species were used to make knives, including Lucina/Phacoides , Mytilus , Mactra , Macrocallista , Ostrea , and Diplodon. These knives were used to cut materials such as meat or wood. The knives made from Lucinidae shells were prepared by making continuous denticulations between 2 and 3 cm long along the shell rim. In the same area of Brazil, the prehistoric culture known as the Sambaqui of Guaragua^u used Ostrea shells; the Kamayura of Bura^ao used them to open manioc, and the Humutima (Umatina) and the Bororo cut hair with them (Prous 1992). Emperaire (1958) documented several different uses of molluscs among the Alacaluf in Chile. They used Mytilus shells as knives in butchery tasks, employing the sharp edges. In the area of the Jeronimo Channel, Mytilus shells were sharpened with stones in order to cut hard firewood and bone (Emperaire 1958). In this area, the "choro" or mussel is quite common and is gathered by groups living on the coast. Nomadic groups on the coast of Chiapas (Mexico) also made various tools, including knives, with the shells of various molluscs (Linares Villanueva 2005).
The Patwin in the southern Sacramento Valley of California also used mussel shells and knives to cut and descale fish, and to cut meat (Johnson 1978:357).

Scrapers
Many of knives described above, as well as being used in cutting activities, were used for scraping; an example is the multifunctional knife-scraper of the Yamana.
Some of the most common scraping activities carried out with shell tools are tasks involved in treating hides and, as is logical for groups exploiting coastal areas, in processing fish. For example, the Chugach Eskimos in Alaska used molluscs to remove the scales from fish (Dupont 2003). In an analogous way, the use of mollusc shells to scrape bark to make fiber has been documented in Polynesia 82 JOURNAL OF ANTHROPOLOGICAL RESEARCH (Leroi-Gourhan 1945:215), and Mytilus scrapers were used in Tierra del Fuego in hide-processing activities, with the mussel strapped to a pebble as a kind of handle (Leroi-Gourhan 1945:219). In Brazil, several types of shell scrapers have been documented. Side-scrapers were made from species such as Tinela ventricosa and Macrocallista , retouched to form a straight or slightly convex active edge, one that is more abrupt and robust than the shell's natural edge. These types of scrapers were often used by the Itaipu, who lived on the shore near Rio de Janeiro. They were also used by the Waura to smooth their pottery before it was fired (also see below). Other kinds of scrapers are concave or notched, with the active part retouched obliquely. These are normally used to shape cylindrical wooden or bone staffs (Prous 1992).
The Alacaluf also used shells to scrape and thin seal and otter skins (Emperaire 1 958). Similarly, around the Gulf of Mexico, prehispanic groups made scrapers with different kinds of mollusc shells (Linares Villanueva 2005). In Cuba, bivalve shells were used to scrape manioc root (Dacal Moure and Rivero de la Calle 1984).
Finally, mollusc and gastropod shells have also been documented as sidescrapers. The most common types were made from hard Strombus and Spondylus shells, although occasionally Pecten shells were employed as well, using their natural toothed edge (Suarez 1974). This type of use has been documented among various groups in Venezuela.

Hooks
Numerous groups have used mollusc shells to make fishing hooks. They have been documented in such areas as Oceania, Polynesia, and Melanesia (Leroi-Gourhan 1945:168).
In Brazil the genus Strophocheilidae is generally used, separating the body or the reinforced lip, which is then polished and used for fishing (Prous 1992).
The Obispeno and Purisimeno of the Chumash cultural group, an indigenous people of California, fished with net weights and hooks made from shells (Greenwood 1978:521). The net weights were generally made from large species, such as Strombus and Busycon; this use is also known for indigenous groups in what are now New York and Ohio, as well as elsewhere in California (Suarez 1974). Some groups, such as the Alacaluf of Chile, used some molluscs as bait (Emperaire 1958); this use has also been documented among the Yamana on Tierra del Fuego, but in that case they fished without hooks (Gusinde 1986).
Similarly, on the islands of Polynesia, hooks made from small Trochus and Turbo shells (between 13 and 30 mm) were used by the aborigines from the fourteenth century (Allen 1996). In general, these artifacts made from shell are very common around the Pacific, and in northeastern North America, whereas in Central America the use of nets and arrows appears to have been predominant (Suarez 1974).

Polishers and Smoothers
Mollusc shells were also used to polish various materials, For example, in South America, the Guayaquis (Ache) Indians use freshwater bivalve shells in the final stage of polishing axe handles, and in several parts of Africa they are used to smooth ceramic vessels (Dupont 2003). This use, for smoothing pottery, has also been documented ethnographically in the Canary Islands (Rodriguez and Navarro 1999).
In Brazil, unmodified Strophocheilidae shells were used by the Bororo to polish their wooden bows (Prous 1992). The Alacaluf worked bone with shells, holding the broken convex face and using the edges of thick shells to polish the heads of harpoons (Emperaire 1958:233).

Gouges , Adzes , and Other Woodworking Tools
In Chile, the Alacaluf used shells as woodworking gouges to build their boats (Emperaire 1958). They hafted shells on handles and used the cutting edge to prepare tree bark for canoes and also to shape the oars. Various species were also used to plane the wood and for the final effect. They made wooden hooks, using sharp shells; with the arrival of Europeans, these were replaced by metal blades (Emperaire 1958).
Another similar implement, although smaller, has been documented among aboriginal groups in Brazil. These are micro-gouges, generally made from small bivalve shells, with the anterior or posterior edges retouched to form a right angle, or slightly vertical concave or concave in transverse section. Some of these artifacts are made from Diplodon shells, between 6 and 8 cm long, by direct retouching of the dorsal face, causing flaking on the ventral face (Prous 1992). In Brazil, saws were commonly made with mollusc shells. Robust shells were used, of species such as Lucina , Mactra , Macoma , and Macrocallista , modifying the edges by retouching at regular intervals. These tools could be used as saws or for scraping (Prous 1992).
In Oceania, mollusc shells were widely used as adzes or axes. They were employed for various functions, including woodworking (Leroi-Gourhan 1945).

Punches and Borers
In Brazil, shells were also used to make punchestools with a straight, sharp point. They are normally used by applying pressure, often with rotation, so their use may leave oblique use-wear marks. They are usually made from such species as Macrocallista , Maculata , and Strophoenelidae. Naturally pointed Ostrea fragments may also be regarded as punches (Prous 1992). Spines of sea urchins ( Echinoidea ), for example, were used as drills by the Ainu of Hokkaido, in the northern Pacific (Leroi-Gourhan 1945:153).
Borers made from shells are documented for the proto-agricultural stage in Cuba. To manufacture these tools it was necessary to shape the shell by light percussion (Dacal Moure and Rivero de la Calle 1984).

Weapons
Two types of weapons made from shells have been documented ethnographically: projectile points and clubs. These artifacts are mainly found in places where lithic raw materials are not abundant or of poor quality (Suarez 1 974). For example, in Brazil, arrowheads are known to have been made with Ostrea and used primarily in the area of the Bay of Paranagua and Cananeia. They are pointed objects, some of which possess a small tang (Prous 1 992). Similar projectile points were also used by the Brownsville culture on the coast of Tamaulipas, near the modern border between Texas and Mexico (Suarez 1974:22).
Clubs were made from thick shells, such as Busycon perversum and Strombus gigas , which were hafted to a transverse handle. These artifacts have been documented in Florida and on the Atlantic coast of North America (Suarez 1974:22).

Other Tools
The Guayaquis (Ache) in the tropical rainforest of Paraguay used freshwater bivalve shells to spread the wax that protected the sides of the baskets they used for gathering (Dupont 2003). In Brazil, the seringueiros (rubber-tappers) in the Amazon forest used Ampullaria gigas shells to collect rubber latex.
Both shell axes and hammers were generally made from the mantle of thick gastropod shells; in some cases they were hafted but in others they were used expediently. They have been documented in Vancouver, New England, Florida, and also in Honduras (Suarez 1974). In Chiapas on the Gulf of Mexico, hammers were made from the shells of large gastropods (Linares Villanueva 2005).
In the same way, we know that some groups used large gastropods as ards or plowshares, with little modification of the shell. On occasions, with a wooden haft, the lower part of the labrum was used as an ard. Another type was made from clam valves tied with a cord. These objects have been documented only for societies in the initial stages of agriculture (Suarez 1974).
Finally, some groups of aborigines depilate practically their entire body and even cut their hair with shells. For example, among the Yamana of Tierra del Fuego, Mytilus shells were used as pincers for depilation (Mansur and Clemente 2009). This use is also known among the Alacaluf in Chile (Emperaire 1958). In many cases, the valves were tied together with a cord through the area of the umbo; this type of artifact has also been documented over much of North America.

MESOLITHIC AND EARLY NEOLITHIC TECHNOLOGIES IN SHELL MIDDEN SITES IN NORTHERN SPAIN
One of the debates that has arisen as regards the Mesolithic in northern Spain focuses on the scarcity of artifacts found in certain contexts, especially in the western part of the region (Arias 1992a;Clark 1976;Gonzalez Morales 1982;Gonzalez Morales et al. 2004;Vega del Sella 1930). However, the relative scarcity of lithic and bone technologies at these Mesolithic sites is usually based on comparison of the quantity of knapping waste or retouched implements with the numbers of artifacts found at Upper Paleolithic occupations. Despite the lower density of tools, Mesolithic shell middens still provide evidence for a similar variety of productive activities as is found at Paleolithic settlements. Activities apart from the exploitation of molluscs were carried out at the sites, such as the consumption of mammals and fish. In addition, it is common to find human remains within the middens, as at Balmori, Cuartamentero, El Molino de Gasparin, Poza l'Egua, Colomba (Arias and Fano 2003), or El Toral III. In the same way, some of these occupations display a markedly stable character, as at Mazaculos II (Gonzalez Morales 1995a, 1995b). This well-known reduction in the quantity of lithic and bone technologies at Mesolithic settlements in the western (Asturian) part of the region is also seen in general terms at sites in the eastern (Basque) sector. The assemblages of the latter area, however, are more heterogeneous in this respect. Some sites display low tool indices; in other cases the technological range might be regarded as more substantial, although still smaller than at Paleolithic sites. Note, however, that few empirical data are available since many of these sites are still being studied, and different proportions of the total surface areas have been excavated at each site. The availability of flint in larger quantities and of higher quality in the eastern sector enabled a greater use of blade technology. In contrast, at the western sites, the scarcity of outcrops of high-quality flint was counterbalanced by the use of other materials, such as radiolarite and quartzite (e.g., Straus 1996). The latter was employed to make, among other things, the Asturian picks that traditionally have been used in the chronocultural definition of the Mesolithic settlements in the western sector (generally known as the Asturian Culture).
In the early Neolithic, the period for which the technological use of shells has been documented in Cantabrian Spain, coastal sites such as Santimamine seem to display a higher density of lithic artifacts in comparison with the Mesolithic occupations, generally with larger numbers of geometric pieces. Thus, while coastal sites in the east, such as Kobaederra (also Urdaibai, in Vizcaya; Zapata et al. 1997) and Los Gitanos (Samano, Cantabria; Ontanon 2005), seem to exhibit a relative abundance of these kinds of artifacts in their inventories, at an early Neolithic site in the western area, Mazaculos II (level A2) (Gonzalez Morales 1995b), the presence of geometries was limited to a single microlithic trapeze. In the same way, in the western sector, only Stratigraphic Unit 7 at Los Canes Perez 1992, 1995) seems to have geometric microliths in numbers similar to or greater than are found in the eastern part of the region. However, Los Canes is an inland site, where the presence of marine resources was negligible. Despite this, in the Neolithic of the western part of Cantabrian Spain, a balance can be seen between the use of quartzite (the predominant raw material in this area in the Mesolithic) and flint. This is reflected in a larger proportion of blade technologies in contrast with the predominance of macrolithic implements at Mesolithic sites (Gonzalez Morales 1995b).
This lack of "traditional tools"that is, those made from bone or lithic materialsat many shell middens in northern Spain has led to continuing discussions about the functionality and use of these Mesolithic and early Neolithic sites. Rarely is the evidence found at the shell middens clear enough to enable them to be defined as habitation sites, which has given rise to a dual debate about the technology used by these hunter-gatherers to carry out their productive activities and also about the places chosen by the human groups to perform these economic activities (e.g., Straus 1979). As regards the technology, in response to the small quantity of bone and lithic industries at certain of these sites, some researchers have suggested the use of industries made from perishable material, such as wood (Arias 1992b;Clark 1976;Gonzalez Morales 1995a). Unfortunately, it is difficult to document this kind of material directly in the Cantabrian archaeological record, and the only way to trace its use is based on indirect evidence, such as use-wear analysis of the nonperishable tools that have been recovered.
Similarly, in other parts of the world, different functions have been proposed for the shell middens. Some authors have interpreted certain sites as places used specifically for the processing of malacological resources, which were then stored (Henshilwood et al. 1994). However, compared with this type of site, the Mesolithic and Neolithic middens in northern Spain occur in different locations, have different morphologies, and, above all, evidence a greater variety of productive activities, extending beyond the mere gathering of coastal resources.
In this respect, in order to suggest an explanation for the scarcity of technology with which to carry out the wide range of productive processes seen in these Mesolithic and early Neolithic sites, with the analogic support of the ethnographic information given above, the hypothesis is proposed that shell implements were used to perform some of the activities carried out by these groups of huntergatherers in northern Spain.  Barandiaran, and E. Eguren (1931;Aranzadi and Barandiaran 1935). These researchers described a major deposit whose stratigraphic sequence covered much of the region's prehistory, as well as a magnificent ensemble of Magdalenian cave paintings. In the Upper Paleolithic levels they found evidence of the exploitation of coastal environments, which increased in importance in the Mesolithic, with a large midden made up of Ostrea edulis, Scrobicularia plana , Ruditapes decussatus , and Patella sp. as the main species. They divided this midden into two levels, based on the presence or absence of ceramics. Between 1960 and 1962, Barandiaran undertook more limited work (Barandiaran 1961(Barandiaran , 1962(Barandiaran , 1963 to straighten the wall of a stratigraphic section, and additional molluscs were found. The recent stratigraphic reappraisal of the deposit, carried out between , has revealed nine levels in which molluscs appear in varying quantities, from few remains in the Late Magdalenian and Azilian levels to considerable accumulations in the Mesolithic and Neolithic layers. The excavation, done with modern methods, has enabled a detailed study of the material and, consequently, has revealed the evolution of mollusc exploitation throughout the Santimamine sequence (Gutierrez Zugasti 2008).
The lithic material obtained by this recent archaeological work at Santimamine shows a clear quantitative decline in the Neolithic (levels Slm-Lsm) and Mesolithic (levels H-Sln) in comparison with the Azilian (level Arcp) and above all with the occupations during the Magdalenian (levels Sine, Almp, and Csn-Camr; Lopez Quintana et al. in press). However, as we have pointed out above, this striking reduction in quantity is common at sites in the region during the transition between the Pleistocene and the Holocene, such as at La Riera (Asturias; Straus and Clark 1986) and El Perro (Cantabria; Gonzalez Morales and Diaz Casado 1992).

Material and Methods
The archaeological material that has been analyzed consists of nine mollusc shell fragments from the Neolithic levels Sim and Lsm. To be precise, from level Lsm, there are two hinge fragments of right valves, a fragment with the impression of the adductor muscle of a right valve, and an edge fragment, all of Ostrea edulis (Linne 1758). From the same level an edge fragment of Patella sp. and another of Ruditapes decussatus (Linne 1758) have also been analyzed. From level Sim, the Of these nine shell fragments, seven were defined by microwear analysis as working implements (Cuenca 2009;Gutierrez Zugasti et al. 2010).
In order to test the hypothesis proposed through the functional analysis of these remains, an experimental program was undertaken to characterize the usewear exhibited by shell tools when they are used to carry out specific productive activities (Cuenca et al. 2010). This type of experiment has the aim of describing the use-wear marks that are generated on the shells, together with the definition of the processes that form them. This information is achieved by controlling certain variables, which are either unmodifiable (such as the substance being worked) or modifiable (the working action and its duration, or the working angle). By controlling the variables that are regarded as significant, the experiment becomes an analytical tool (Gonzalez Urquijo and Ibanez 1994:17). Based on the results of the functional analysis carried out on the malacological objects at Santimamine, the activities selected for the experimental program were the processing of wood, animal skins, plant fibers, and hides. The mollusc species chosen for the experiment were the same as the ones examined in the use-wear analysis, and which are found in different proportions in the assemblages at sites in northern Spain in the Mesolithic and early Neolithic (Gutierrez Zugasti 2008). Therefore, their selection for the experiment was both rational and opportune, enabling the results of the experiment to be extrapolated to the study of other sites in the same area.
The working implements in the experimental program were first observed macroscopically, with a magnification of between 5 and 72.5 x, in order to analyze and photograph larger marks, mainly rounding and chipping. Subsequently the pieces were viewed and photographed at between 100 and 200 x, with a metallographic microscope, in order to analyze the micro-wear. The polished surfaces and the bottoms of the grooves were recorded with a duplicator inserted in the microscope, which allowed the observation and photography of these features at up to 400 x.
The shells used in the experiment were cleaned in water before being analyzed and photographed, and if necessary, the active zones were wiped with a cotton puff impregnated with alcohol or naphtha to remove adhering particles and residue.
Before the experimental program was carried out, the taphonomic traits and natural alterations most commonly seen on the shells of each species were documented by recording and photographing shells that had been gathered but not used. This was aimed at differentiating any natural alterations seen on the shells from those caused by their use as implements.
The substances processed with the shell implements were fresh hazel wands {Corylus avellana Linne 1753), stems of fresh reeds {Juncus sp.), fresh lamb skin ( Ovis aries Linne 1758), dry roe-deer skin ( Capreolus capreolus Linne 1758), and strips of red-deer hide ( Cervus elaphus Linne 1758). The following actions were performed: bidirectional transverse scraping action with the shells of Ostrea edulis , Mytilus galloprovincialis , and Patella sp., to remove fatty matter and remains of meat, and to thin fresh Ovis aries skin and dry Capreolus capreolus skin. Bidirectional transverse action was used to thin and tenderize strips of Cervus elaphus hide. A unidirectional transverse scraping action was used with the natural edge of Ruditapes decussatus to extract plant fibers from Juncus sp. The same shell species was used with a unidirectional transverse scraping action to remove the bark and to shape Corylus avellana wands. The only longitudinal kinematics used in the experiment were to cut the Juncus sp. plant fibers, using the edge produced by a percussion fracture of a Ruditapes decussatus shell. The application of force in all cases was through the pressure of the implement on the substance being worked.
The experiments were carried out in all cases with three time-spans of 5, 1 0, and 15 minutes. During the experiment, the contact angles approximated right angles (about 90°) in both the transverse and the longitudinal actions. In some cases they varied between 75° and 90°, but in the working of hide with the Ostrea edulis hinge, the surface of which is flat, the friction took place at a nearly flat angle (180°).

92
JOURNAL OF ANTHROPOLOGICAL RESEARCH Figure 6. Use-wear marks in the experimental program: ( 1 ) edge of Mytilus gallopr ovine ialis used to process dry skin for five minutes, at 100x; (2) hinge of Ostrea edulis used to process strips of hide for 10 minutes, at 200x; (3) edge of Patella sp. used to process dry skin for 10 minutes, at 200x; (4) edge of Ruditapes deeussatus used to shape wooden sticks for 15 minutes, at 400*.

Results
The analysis of the shell fragments from Santimamine determined that seven of them had been used in the processing of different kinds of soft or medium-hard animal matter and in some cases of plant material (Gutierrez Zugasti et al. in press). All the shell tools had been used with a transverse scraping action, in some cases alternating with longitudinal kinematics.
Three Ostrea edulis fragments had marks on one of the natural edges, and another had use-wear marks on the vertex of the hinge, which was noticeably rounded. This hinge displayed polishing characterized by heavy shading, a dull gloss, and a greasy appearance.
In the same way, two fragments of Patella sp. and Mytilus galloprovincialis shells showed signs of transverse action on one or two of their edges, respectively, caused by working an abrasive animal matter. Finally, a fragment of Ruditapes deeussatus , with a highly rounded edge, displayed polishing with compact shading, and striations with a dark bottom, which are oblique and perpendicular to the edge. This fragment was identified as an implement used to process a plant rich in silica.
The results of the experimental program confirm the hypotheses that were proposed about the substances that were processed with the shell implements found at Santimamine. The findings suggest that the Ruditapes deeussatus fragment was Figure 7. Ostrea edulis fragment with the impression of the adductor muscle, displaying use-wear marks: it is noticeably rounded and its surface shows micro-polish, a dull gloss and "greasy" appearance. Circular depressions and numerous fine, shallow striations, in different directions can also be seen, giving the polishing a rough appearance.
used to process a non-woody plant and was not used to work wood. However, despite the similarity in the use-wear marks described for the archaeological and the experimental materials, they are more developed in the archaeological items that have been analyzed. Therefore, the Ruditapes decussatus fragment from level Lsm at Santimamine was probably used to process a non-woody plant harder than Juncus sp., or it was used for a longer period than during the experiments. This latter possibility is very likely since the maximum duration of the tests in the experiment was only 15 minutes, and the implement was still perfectly suitable for continued processing. With regard to the Ostrea edulis hinge with a rounded vertex, the experimental items used to tenderize small hide strips displayed very similar traits to those in the archaeological assemblage, although the use-wear was somewhat more developed in the experimental items. The degree of dryness (and therefore abrasiveness) of the hide may determine the depth of the marks, or the speed with which the marks develop. shell used in the experimental program to process dry skin for 5 minutes, at 200x.
The other shell fragments from Santimaminethree Ostrea edulis fragments with marks on one of the natural edges, and fragments of Patella sp. and Mytilus galloprovincialiswere interpreted as implements used to scrape soft or mediumhard animal matter. However, the experiments seem to suggest that they were probably used to process dry hide rather than fresh skin. Alternatively, some type of abrasive might have been used to tan the skin (such as ocher). This procedure has often been documented ethnographically; it does cause more developed marks, and therefore the implement surfaces display more striations.
This hypothesis is based on the development of diagnostic marks in both collections: the amount of polishing and the concentration of striations on the archaeological material more closely resembles the experimental material used to work dry hide. Dry hide is rather more abrasive than fresh skin, and causes more developed marks and, as we have seen, more abundant striations (Cuenca et al. 2010).

DISCUSSION
The use of information from ethnography and anthropology in archaeological studies has provoked discussion since the 1970s. In America, various studies appear to have resolved this debate definitively (Gould 1978;Kramer 1979), whereas in Europe the contribution that this kind of information can make has been a more recent topic of discussion (Estevez and Vila 1995;Gandara 2006;Mansur 2006;Manzi and Spikins 2008).
Our findings described here constitute a contribution to the available information about the use of shells as tools in productive processes, and this is the first time this kind of use has been identified in northern Spain. This use, as we shall see, is on occasion clearly comparable with the kinds of activities for which shells have been used in many ethnographical accounts, processing similar types of materials. In addition, this evidence could partly explain the scarcity of lithic and bone tools at many Mesolithic and early Neolithic sites in northern Spain. It may complete the technological range employed by some of these hunter-gatherer groups to carry out activities related to the exploitation of coastal resources.
The methodological shaping of an experimental program, the basis of the development of the use-wear analysis of productive tools, should be founded first on the productive activities as shown implicitly or explicitly by the archaeological contexts in the geographical area of the study. In parallel with this, the design of the experimental program should be coherent with the natural resources (malacological resources, in this case) used as raw materials in the geographical context defined by the research program. Certainly, this factor prevents at the outset the erroneous application, as a direct analogy, of information from archaeological studies carried out in contexts with very different biotopes. Despite this, we believe that a wide knowledge of the technological use of shells, as recorded ethnographical ly, could provide decisive background data to encourage and stimulate the formulation of appropriate methodological hypotheses for the questions being asked. In the case of northern Spain, the contributions made by ethnographic information about the technological use of shells have assisted in the development of analytical mechanisms with which to tackle concrete problems in a restricted geographical area, but probably (and in parallel) the use of this information allows these same mechanisms to be validated in other geographical and chronological contexts, thus laying down a more solid foundation on which to construct a methodology of analysis. The documentation of ethnographic information has been used to formulate a working hypothesis during the planning of an experimental program (Cuenca 2009) to contextualize the results of use-wear analysis carried out on several shell tools (Gutierrez Zugasti et al. in press). In fact, the eabundant ethnographic information in the literature shows the suitability of shells for use as tools in numerous productive processes carried out by different human groups. In this respect, some of the activities performed by these groups, and to a lesser degree some of the mollusc species used as tools, are similar to those recorded in Mesolithic and early Neolithic contexts in northern Spain.
For instance, we can highlight the suitability of species belonging to the Mytilus genus (mussel species found at many archaeological sites in the north of Spain) for scraping tasks, as well as the numerous examples that have been recorded of the use of shells in activities such as processing hides and wood, or preparing meat and fishtasks that are represented in Mesolithic and early Neolithic contexts in northern Spain and at many different coastal archaeological sites all over the world. The large amount of information available about the many uses given to mollusc shells, among a varied and wide range of human groups, in different geographical and chronological settings, should make us reflect on 96 JOURNAL OF ANTHROPOLOGICAL RESEARCH the almost exclusive consideration as food resources and ornamentation that has traditionally been given to malacological remains in most coastal archaeological contexts, particularly in the case of research undertaken in Europe. To the abundant information about the use of these kinds of tools among different human groups, we can add the functional potential of some shell tools shown by experimental programs (e.g., Cuenca 2009;Cuenca et al. 2010;Lammers-Keijsers 2008;Mansur and Clemente 2009;Toth and Woods 1989).
The steady increase in reporting this kind of evidence at European archaeological sites (Cade 1991;Cristiani et al. 2005;Gutierrez Zugasti et al. in press;Vigie 1987Vigie , 1992Vigie , 1995Vigie andCourtin 1986, 1987) will very probably expand further with the development of a specific methodology for the analysis of these types of working tools, as well as through the spread of this type of analysis to other coastal contexts.
However, it is equally true that until now no systematic functional research has been undertaken on shells. The lack of a well-established, specific analytical methodology for this type of material, together with the great importance given to lithic tools in archaeological studies, has meant that other raw materials, including molluscs, have been undervalued as to their potential use in productive processes (Mansur and Clemente 2009). Presently we can neither extend these results to the general interpretation of coastal sites in the north of Spain nor rule out the likelihood that these types of tools might be found at other sites, since microwear analysis has yet to be applied to many of the assemblages. In any case, we should not isolate this technology from the other means of production used by huntergatherer groups for the same aim. Therefore, once it has been shown that this kind of technological use of shells has also been employed at other sites, it would be interesting to relate this use with aspects such as the availability of goodquality lithic raw materials, or the differential use of each kind of material, such as stone, bone, antler and shell, in specific productive processes. In this respect, differentiated use of each tool material has been documented in certain contexts (Charpentier et al. 2004). However, in other cases in which use-wear analysis has been carried out on the whole tool assemblage, both lithic and shell, the results seem to suggest both technologies were used for very similar activities (Lammers-Keijsers 2008). Finally, in still other contexts, the use of mollusc shells has been interpreted as a solution in situations where good-quality lithic raw material is unavailable (Schmidt et al. 2001 ).

CONCLUSIONS
To date, the documentation of this type of evidence in northern Spain is limited to the seven tools recovered from the deposit at the cave of Santimamine (Kortezubi, Vizcaya). Thus, taking into account that the evidence documented so far corresponds to an area (the eastern part of northern Spain) exhibiting greater availability of high-quality lithic raw materials for knapping than the western sector, the hypothesis about the role that shell implements might have played in carrying out productive activities at coastal sites in the western area, in both the Mesolithic and the early Neolithic, is strengthened even further (Cuenca 2009).

USE OF MOLLUSC SHELLS AS TOOLS
. 97 Although we are unable to reach definitive conclusions about these questions in our area of study, it seems likely that such an abundant raw material, perhaps already brought to the settlement as a food resource, could greatly reduce the work involved in the provision of other raw materials, as well as the subsequent process of manufacture or shaping, when shells are used without further modification (Cuenca 2009). At the same time, the use of an implement that does not require any kind of repair (Castro et al. 1999:51) to maintain its effective functionality would have been an advantage in comparison with lithic tools, because as soon as the functionality of the shell implement began to decline, it could rapidly and cheaply be replaced with another one.
In this respect, we believe that the technological use of shells could have modified certain behavioral patterns of the human groups in connection with the processes of manufacturing their tools, and by extension with other productive processes. These would probably have been influenced by such aspects as the availability of good-quality raw materials in the local area, or the need to move long distances or develop a network of social relationships for the provision of the lithic resources needed to make their tools (Cuenca 2009).
In the future, as other archaeological sites and assemblages are studied, any identification of the use of shells as tools may help to explain the puzzling scarcity of lithic and bone assemblages at some Mesolithic and early Neolithic sites in the north of Spain. It will then be necessary to reappraise the role of shells in these societies, taking into account that the malacological resources may not only have been used as food and in ornamentation, but also as tools with which certain productive processes were carried out. It will be necessary to develop specific methodologies enabling the study of this type of evidence by applying use-wear analysis (Semenov 1964). The combination of both factors could contribute to greatly increasing our knowledge about these coastal groups of hunter-gatherers. This type of technological evidence could be extended to other areas outside northern Spain by developing specific analytical methodologies based on traceology, and also by enlarging our knowledge about the functional potential shells may have as working tools through the possibilities offered by ethnographical documentation.

NOTE
The second auther is now at the Department of Archaeology, University of York, England (UK), and can be reached at igorgutierrez.zug@gmail.com. The third author can be reached at ignacio@imf.csic.es. The authors would like to thank the University of Cantabria for financial support, J. C. Lopez Quintana, M. R. Gonzalez Morales, and A. Garcia Moreno for their help; and L. G. Straus, C. Claassen, D. Lubell, and two anonymous referees for their comments. The article has been extensively edited by JAR.