PRELIMINARY RESULTS OF A REPLICATNE STUDY: METATE RE-ROUGHENING, PECKING, OR POUNDING?

Dawn M. Reid and Mari A. Pritchard-Parker Department of Anthropology University of California Riverside, CA 92521

ABSTRACT

This paper discusses the results of a replicative study on the roughening of milling equipment, specifically metates. The intent of the replication was to ascertain ifdiffering results were achieved when various percussor types were used to re-roughen a metate-like surface. Hammerstones prepared by flaking and unmodified cobbles were utilized. The results obtained in the replication were then compared with archaeological specimens in an attempt to determine more information as to the manufacturing and use behaviors associated with milling equipment.

INTRODUCTION ment would be most efficiently accom­ plished with tools of different shapes and Stone metates or grinding stones have sizes, depending upon how much mass one long been a primary food processing tool for is trying to remove and where that mass is many agricultural as well as hunting and located. Where manufacture seeks to re­ gathering societies. Use of these imple­ move a relatively large amount ofmass, re­ ments to reduce both plant and animal ma­ roughening seeks to merely mar the grind­ terial has been well documented. When a ing surface to such a degree that it becomes metate is used in conjunction with a mano useful again. Identifying potential differ­ to grind foodstuffs, sharp edges are sheared ences in the types of tools used in the manu­ off ofthe metate surface and residues from facture and maintenance ofmilling imple­ the material being ground are deposited. ments will contribute to our overall under­ This process results in the metate surface standing ofartifact distributions and their becoming increasingly smooth, ineffectual, associated behaviors. For example, explana­ and in need ofmaintenance. Rejuvenation tions may be advanced for hammerstones ofthe metate can be achieved through re­ found in behaviorally significant numbers roughening ofthe grinding surface thus ex­ near completed metates or milling activity tending its usefulness. areas. Additionally, use wear patterns on hammerstones will vary depending upon the Replicative studies involving the manu­ way in which they were used. A knapping facture ofmilling equipment (Pritchard­ hammer will reflect glancing rather than di­ Parker 1991) sparked discussions concern­ rect action fracture whereas a tool used to ing possible differences in the type ofham­ roughen a grinding surface should reflect a mer tool used to manufacture metates ver­ definite fracture pattern, delineating it from sus those used to maintain the milling sur­ other hammer tools. The resultant debitage face through re-roughening. Both angular­ from such an activity, which may often ap­ edged hammers and unmodified cobbles pear anomalous, may also be diagnostic of have been used in the replicative manufac­ the metate re-roughening process. ture ofmetates (Pritchard-Parker 1991; Schneider et al. 1992). However, it may be Ethnographic literature contains vague possible that the differing tasks ofmanufac­ references to "pecking" stones used in the turing and re-roughening ofmilling equip­ re-roughening process (Mohr 1954; Lister

Proceedings of the Society for California Archaeology, 1993, Vol. 6, pp. 199-206. Copyright@1993 by the Society for California Archaeology. 1965; Huckell1986; and others). Pecking, as width from 25.5 em to 30.5 cm. defmed by Turner (1963), is performed with a "sharpened hammerstone". The few de­ Three hammerstones were used in the scriptions of these stones allude to a modi­ Phase 1 experiment (hammerstones 1, 2, fied cobble with sharp edges or points, a and 3 in Table 1). All 3 hammerstones were sharp broken stone, or a prepared core. No quartzite cobbles that had one end flaked to references describing the use of blunt ham­ form an angular-edged hammer, and the merstones for re-roughening were found. other end left unmodified (rounded). Using However, one account of excavations at the 2 ends of a single cobble as different Mesa Verde included pictures ofrecovered tools helped to control for the consistency of "pecking" stones (O'Bryan 1950) that appear the quartzite among the hammerstones. to be unmodified cobbles with the cortex The weights of the hammerstones ranged and naturally-occurring rounded edges in­ from 566.9 g to 577.4 g. This weight range tact. For this reason, it seemed appropriate was intended to approximate the average to suggest that the term "pounding" be weight ofarchaeological specimens as calcu­ adopted (Jeanne Binning, personal commu­ lated from the mean weight of a collection of nication, 1991) to describe the use ofblunt 216 hammerstones recovered from a site, or unmodified hammerstones. While it has ORA-1070, in Orange County (Mike Macko, been suggested that the rounded or blunt personal communication, 1992). ends of manos could have been used to roughen metate grinding surfaces by Riddell (1960) and others, no replicative work that Table 1 we know of has addressed this issue. Hammer Tool Attribute Summary The goal of this experimental project was to evaluate the results obtained with original Final Used on both angular-edged and naturally rounded # Material Weight Weight Shape* Slab or blunt hammerstones when each was used to roughen a simulated metate surface. We 1 Quartzite 566.9 9 482.0 9 combo Granite focused on the following question: Would 2 Quartzite 573.5 9 403.2 9 combo Granite rounded-edged hammers be as effective as 3 Quartzite 577.4 9 354.3 9 combo Basalt angular-edged hammers in the metate re­ 4 Rhyolite 763.6 9 481.8 9 angular Basalt & roughening process? We observed the mor­ Granite phology of the dents (Turner 1963) or indi­ 5 Quartzite 2365.6 9 1543.6 9 unmod Basalt vidual pits resulting from the blows admin­ 6 Quartzite 1071.4 9 1071.4 9 angular Weathered istered with the different tool shapes. In Granite building upon the results ofthe initial phase 7 Quartzite 1094.1 9 1072.2 9 unmod Weathered ofthis study, additional variables were con­ Granite sidered, including weight of the hammer­ stone and material of both the hammer and *combo=coMbination; unmod=unmodified the grinding surface.

PHASED EXPERIMENTS In conducting the experiment, each of the slabs was divided into halves. Hammer­ Phase 1 stones 1 and 2 were used on the granite Methods. In order to control the initial sur­ slabs, with half of each slab being "pecked" face characteristics of the simulated met­ with the angular-edged ends of the ham­ ates, 2 slabs of granite and one slab of basalt, merstones, and the other half of each gran­ obtained from a local monument works, ite slab being "pounded" with the rounded, were used. All of the slabs had smooth sur­ unmodified, ends ofthe hammerstones. faces cut with a rock saw. The smooth, uni­ Hammerstone 3 was used in the same man­ form surfaces allowed the effects of each ner on the basalt slab. The ends of each hammer type to be isolated. The slabs hammerstone were to be used to strike the ranged in length from 37 cm to 51 em and in simulated metate surface 150 times. In or­

200 der to evaluate use*attrition of the hammer* Phase 2 stones under conditions ofpecking and Due to the rates of tool weight loss pounding, each hammerstone of known caused by use-attrition during Phase 1 (Fig­ weight was fIrst used for pecking, then the ure 1, hammerstones 1, 2 & 3), it was con­ resulting debitage was weighed. Each cluded that the original weights of the ex­ hammerstone was then used for pounding, perimental hammerstones may have been and the newly produced debitage was too low to replicate actual tools used prehis­ weighed. A fmal weight of each hammer­ torically. Considering the possibility that a stone was also recorded. majority ofthe hammerstones recovered ar­ chaeologically may have been expended Results. The dents produced in the surfaces tools, the average weight ofhammerstones of the basalt and granite slabs by the angu­ during their use-life may have been signifI­ lar-edged hammerstones were irregularly cantly greater than that recorded for the ar­ shaped and had some depth, whereas the chaeological specimens. Itwas also rec­ blows administered with the unmodifIed ognized that the material of the hammer­ (rounded) tools resulted in elongated marks stones used in Phase 1 (quartzite) may ac­ on the slabs that infrequently broke the sur­ count for some degree of the use-attrition face sufficiently to create a dent. The only experienced. It was also desirable to know if difference in dent morphology between the hammerstone weight and material account­ basalt slab and the granite slabs was the ed for some degree ofthe use-attrition ex­ somewhat greater depth of dents produced perienced and if these factors had any effect in the granite slabs. on the morphology of dents produced in the struck surface. The damage to the angular-edged ham­ mers themselves was reflected in crushed Methods. To test the effect of hammerstone edges and step fractures. The rounded­ weight and material on hammers tone use­ edged hammerstones tended to develop bat­ attrition and dent morphology, 2 hammer­ tered bifacial edges and took on a chopper­ stones of greater weights were used, one of like appearance. which was made of rhyolite. The rhyolite cobble was flaked to form an angular-edged An extreme amount of tool loss and early hammerstone weighing 763.6 g (#4 in Table tool failure was experienced with the ham­ 1), and a quartzite cobble weighing 2365.6 g merstones, particularly with the initially was left unmodifIed (rounded; #5 in Table unmodifIed (rounded) tool edges (see Figure 1). The rhyolite tool was used to strike both 1, hammerstones 1, 2, and 3). 150 blows the granite and basalt slabs, and the quartz­ were administered with the angular ends of ite tool was used to strike the basalt slab. all 3 hammerstones, but due to loss and One hundred fIfty blows were administered breakage, the unmodifIed ends of the 3 on the basalt slab by each tool, and 50 blows hammerstones could only be used for 22, 70, were administered by the rhyolite tool on and 98 strokes, respectively. the granite slab. The debitage that was produced by the Results. A similar level of weight loss use-attrition ofthe angular-edged hammer­ through use-attrition was experienced with stones consisted of flakes with crushed and both of the large hammerstones, with the shattered platforms and angular shatter. rhyolite specimen (#4) losing 37% of pre­ Debitage from the initially unmodifIed use weight, and the quartzite specimen (#5) (rounded) hammerstones was similar to that loosing 35% of pre-use weight (Table 1 and produced by the angular-edged tools and Figure 2, hammerstones 4 & 5). The larger also included round and oval flakes, often tools enabled the use of both hands in strik­ without platforms, that were created by the ing and were therefore more comfortable to initial blows. The size of the flakes from the use. As in Phase 1, a difference in dent mor­ rounded hammerstones tended to be slight­ phology was produced with the angular­ ly larger than those from the angular-edged edged verses the unmodifIed (rounded) tools. edged tool. The only distinction in dent morphology noted between the Phase 1 and

201 Hammer No.1 Hammer No.2 Original Weight • 566.9 g. Original Weight • 573.5 g.

final walghl 88'110

"Pecking' lou 2'110 'Poundlng' Ioas 9'110

"Pounding' Ion 28'110 Unrelalnad Ioas 2'110 Quartzite Quartzite

Hammer No.3 Original Weight • 577.4 g.

"Pecklng'loas 10'lI0

"Pounding" 10•• 23'110 Quartzite

Figure 1. Hammerstone Experiment Phase 1.

202 Hammer No.4 Hammer No.5 Original Weight • 763.6 g. Original Weight • 2365.6 g.

Unretall18d loaa 8"" Rhyolite Quartzite

Figure 2. Hammerstone Experiment Phase 2.

203 Phase 2 experiments was a slight increase in weighing 1094.1 g (hammerstone #7 in the depth of dents produced when using the Table 1). In conducting the experiment, the heavier hammerstones. weathered slab was first divided in half, with halfofthe slab being "pecked" with the ang­ The damage to the angular-edged rhyol­ ular-edged hammerstone, and the other half ite hammerstone consisted of step fractures being "pounded" with the rounded, unmodi­ and edge-crushing, resulting in an object fied hammerstone. Each hammerstone was that might typically be classified as a "bat­ used to strike the weathered slab 200 times. tered core". Damage to the initially unmodi­ fied (rounded) hammerstone resulted in bi­ Results. The general shape ofthe dents facial feathered as well as stepped flake produced in the weathered slab were similar terminations and heavy edge battering. The to those observed on the cut slabs. Howev­ resulting object closely resembled archaeo­ er, there was a notable difference between logical "chopper" specimens. As both Phase the cut and weathered slabs in the depth of 1 and Phase 2 experiments produced edge the dents produced. The dents produced in configurations on initially unmodified the weathered slab by the angular-edged hammers that are distinctly "chopper" like, hammerstone were deeper than those pro­ the correct identification ofsuch "choppers" duced in the cut slabs by the angular-edged as a functional type might sometimes be hammerstones ofboth the Phase 1 and questioned. The "chopper" tool form may be Phase 2 experiments. In contrast to the oval a by-product ofuse rather than the result of marks produced by the unmodified (round­ intentional shaping. ed) hammerstone in Phase 1, and the very shallow oval dents produced by the heavier The debitage produced during use ofthe unmodified tool in Phase 2, "pounding" of quartzite hammerstone in Phase 2 was es­ the weathered slab produced oval shaped sentially similar to that observed during dents that were as deep as those produced by Phase 1. However, the debitage from the the angular-edged tool on the same slab. Phase 2 rhyolite hammerstone is much more angular and irregularly shaped than Damage to the initially unmodified that produced from any ofthe quartzite (rounded) hammerstone was restricted to bammerstones. battering with no measurable weight-loss due to use-attrition (Figure 3, hammerstone Phase 3 #6). The angular-edged hammerstone lost The fmal phase ofthe experimental stu­ only about 1% ofits pre-use weight (Figure dy considered weathering ofa grinding­ 3, hammerstone #7), with damage restricted stone surface as a variable that may effect to a minimal amount ofsmall step fractures, the re-sharpening process. While basin or again (as in Phase 2) resulting in an object trough metates would have a fresh surface that might easily be classified as a battered exposed during the manufacturing stage, "chopper". The slight amount ofdebitage initial use ofa slab metate often involves the produced from the angular-edged hammer­ use ofa natural slab or flat boulder that has stone was essentially similar to the smaller been subjected to deep weathering. Would debitage produced from the quartzite ham­ re-sharpening a weathered stone surface re­ merstones in Phases 1 and 2. sult in the same patterns of dent morphology and hammerstone use-attrition as was ob­ served for the cut slabs used in Phases 1 and CONCLUSIONS 21 The results ofthe Phase 1 and 2 experi­ Methods. An unmodified, weathered gran­ ments suggest that an angular-edged ham­ ite slab was used as the simulated metate merstone would have been the most effec­ surface for the Phase 3 experiment. Two tive tool for re-roughening the grinding sur­ quartzite hammerstones were used: one was faces ofmetates and that "pecking", as re­ an unmodified cobble weighing 1071.4 g ferred to in the literature, would have (hammerstone #6 in Table 1), and the other indeed been the method of choice. Howev­ was flaked to form an angular-edged tool er, in light ofthe results ofthe Phase 3 ex­

204 Hammer No.6 Hammer No.7 Original Weight • 1071.4 g. Original Weight • 1094.1 g.

Final _Ight 100'11. Final walght ".cklng" loll. 1'11.

Unnttalned 10M 1'11.

Quartzite Quartzite

Figure 3. Hammerstone Experiment Phase 3.

205 periment, these conclusions should be quali­ Pritchard-Parker, Mari A fied in the case of slab metates made from 1991 Results ofa Mano Replication Study. unquarried, weathered material. In such Paper presented at the 25th Annual cases, the use ofrounded hammerstones for Meeting ofthe Society for California the roughening process can be considered a Archaeology, Sacramento. viable possibility. Riddell, Francis A Ofparticular interest are the effects of 1960 The Archaeology ofthe }{arlo Site the metate re-roughening process on the (LAS-7), California. University ofCali­ hammerstones used. We should be aware of fornia Archaeological Survey Reports the damage patterns produced on hammer­ No. 53. Berkeley. stones that have been used in metate re­ juvenation and take this into account in our Schneider, Joan S., Philip J. Wilke, and Les­ analyses oflithic tools commonly classified lie Quintero as "choppers" and "battered cores". Seeming­ 1992 Making Metates-Producing Pestles: lyanomalous debitage associated with mill­ Experimental Replication Using Materi­ ing activity areas should also be carefully als from Aboriginal Quarries. Paper pre­ considered as possibly being the result of sented at 26th Annual Meeting of the the metate rejuvenation process. The re­ Society for California Archaeology, Pasa­ sults ofthis replicative study have suggested dena. some interesting preJiminary interpreta­ tions which are being pursued in additional Turner, Christy G., II experimental research. 1963 Petroglyphs ofthe Glen Canyon Re­ gion. Museum of Northern Arizona Bul­ letin No. 38, Glen Canyon Series No 4. REFERENCES CITED Flagstaff.

Huckell, Bruce B. 1986 A Ground Stone Implement Quarry on the Lower Colorado River, North­ western Arizona. Cultural Resource Ser­ ies Monograph No.3. Arizona State Mu­ seum, Tucson.

Lister, Robert H. 1965 Contributions to Mesa Verde Ar­ chaeology: II Site 875, Mesa Verde Na­ tional Park, Colorado. University ofCol­ orado Studies, Series in Anthropology No. 11. Boulder.

Mohr, Albert 1954 The Deep-Basined Metate ofthe Southern California Coast. American Antiquity 4:394-395.

O'Bryan, Deric 1950 Excavations in Mesa Verde National Park 1947-1948. Medallion Papers No. 39. Privately printed, Globe, AZ.

206