THE MANUFACTURE OF A PETROGLYPH: A REPLICATIVE EXPERIMENT James C. Bard and Colin I. Busby From a review of the literature (Grant 1967, Heizer and Baumhoff 1962, Mallery 1893, Rudner 1971, Steward 1929, Turner 1963, Willcox and Pager 1967) it was found that many authors had advanced the methods of abrasion, incision or pecking as the most likely means of executing petroglyphs. For instance, Grant (1967:12) stated: "Most petroglyphs are produced by rock pecking. This can be done in two ways--by striking the surface of the rock with a sharp piece of harder stone, or, for more precise control, by chiseling the rock, using a hammerstone to pound on the stone chisel. The design is usually started with a series of dots joined into lines by continued pecking. Flat tones are indicated by close all-over pecking or by abrading the surface (rubbing or scraping with a harder stone)." Turner (1963:11) adds these further details: "Pecking was done by two methods: (1) hammerstone and chisel, which resulted in very accurate removal of the surface stone and equidistant placement of each pecked dint and, (2) sharpened hammerstone, which gives a sloppy appearance imposed by varieties of muscular coordination. Abrading or incising the surface rock, with another stone, stick, or bone, will produce deep lines and is graphically effective. In general this was not done where the stone was highly patinated and where even a lightly pecked line would stand out strikingly. In any case, the incising of elements takes considerably more time than does peck- ing, to judge by personal experience." However, there exists little experimental data on how petroglyphs were made and the length of time it might have taken to produce them. What data that does exist on the subject is either only briefly mentioned or is quite often subjective and inadequate, with no real base (see Bock and Bock 1972). Occasionally the data is of a more scientific and quantitative nature. Sierts' (1968) article is an excellent example of the scientific approach and method. 84 The present replicative experiment was devised to help fill some of the gaps that exist in the knowledge about petroglyph manufacture with special reference to the different methods, materials, and with particular emphasis on the time and effort involved in their manufacture. Raw Material Collection and Geological Identification For the sake of replicative accuracy, care was taken to insure that only material that would have been available to the original petroglyph makers was used in the experiment. Since the authors were familiar with the Grimes Point petroglyph site (NV-Ch-3) in western Nevada, the raw materials for the experi- ment were collected in the vicinity of this site (see Heizer and Baumhoff 1962 for a description of this site). The basalt material collected was similar to that used at the Grimes petroglyph site and for comparative study, a slab of basalt from the Edwards Creek drainage area in the Desatoya Mountains, approxi- mately 125 miles southeast of Grimes, was also obtained. The pecking tool raw material sample (except for some Texas chert nodules) was obtained near the vicinity of the site. The rocks with suitable surfaces were selected primarily on the basis of ease of transport, e.g., the heaviest and largest basalt boulder or slab that could be lifted by one man. Prime consideration was given to surfaces which were relatively smooth, even, and heavily patinated. Many possible surfaces were rejected because they were either rough and pitted or lacked a weathered "patina" (the so called desert varnish). Others were rejected because they had large areas rendered unsuitable for use by cracking or exfoliation. The bulk of the collected materials was small tabular slabs of basalt with the appropriate surface characteristics. These slabs were later described by R. Leitz (personal communication) as hornblende andesites. The pecking tools were collected from a number of localities in the vicinity of the site. Careful attention was paid to size, shape and to the condition of the proposed working end. Sub-rounded, water-worn, hand-sized cobbles with a fairly small surface area on the proposed working end were considered to be the ideal pecking tools. Tools fitting the criteria were easily grasped, comfortable to work with and could be manipulated with accuracy when directing the blow on the rock surface. The raw materials that were selected for the pecking tools are commonly referred to in the archaeolo- gical literature as rhyolites and basalts. These raw materials were subse- quently identified by Leitz as hornblende andesites (commonly called brown basalt), silicified tuffs (referred to as rhyolite) and vesicular olivine basalt (commonly known as grey basalt). 85 Terminology The two methods of fabrication used in the experiment were direct pecking and a process called pecking/grinding. In direct pecking, the pecking tool is brought sharply into contact with the rock surface with the total distance of travel being approximately eight centimeters. From observation, it was found that an average of 128 blows per minute could be delivered to the rock surface by this method. Treganza (1955:21) refers to this technique as "crumbling". and describes it as a " ...Technique whereby a hammerstone is struck against a specimen, instead of the removal of a large or small flake as in the percussion or pecking method, a granular powder results from the blow.," In the pecking/grinding method, direct pecking is used but, instead of the pecking tool being lifted immediately away from the surface of the rock after contact, the tool is pulled or pushed along the surface of the rock, either towards or away from the person doing the pecking, for approximately two to five centimeters. This results in an abrasion/grinding process taking place on the petroglyph surface. After this abrasion/grinding, the pecking tool is lifted and the process is repeated. It was found that an average of 112 operations per minute could be completed with this process. Since no definition of what constituted a finished petroglyph was found in the literature and as one could not be formulated by observation from weathered petroglyphs present at the Grimes site, it was necessary that an arbitrary definition would have to be decided upon. Therefore, for the purposes of the experiment,, a petroglyph was judged to have reached completion when the desert varnish had been broken through to reveal the unweathered, dark grey basalt surface beneath. Methodology Using a template, a uniform design pattern was put on each slab of ba- salt with a light coat of spray paint. This design consisted of a "rake" element that had been previously described at the Grimes Site (Heizer and Baumhoff 1962:Fig. 40, Design L). The main element was found to be 6.5 inches long uy 1.0 inches wide (Fig. 6). The surface area was calculated to be 18.5 inch or approximately 47.0 centimeters . The process of manufac- ture, using one of the two methods and one of the four raw materials as a pecking tool, was timed throughout, from start to finish. The number of blows per minute, using each of the two methods, was obtained by the counting of the number of blows in a five minute time span. A mean for each method was calculated from the results and was later used in computing the total number of blows per surface for each of the two methods and for the various raw materials used. This rate was determined by observing the technique of one person and it was decided that-in order to minimize the differences that 86 would be caused by two different people working on the same petroglyph, one person would do all of the pecking and pecking/grinding for the experiment. The pecking tools were measured and photographed both before and after their use in an experiment (Table 2). Generally, only one pecking tool was needed to complete the design. However, occasionally the raw material in use was flawed in some way and a second tool was substituted in order to complete the petroglyph. The basalt surfaces were occasionally brushed when the dust produced by the various operations obscured the design features, but apart from this operation, no other actions were carried out. Upon completion, the petroglyph was washed thoroughly to remove any rock dust still adhering to the surface and was left to dry. When dry, the design was chalked in to accentuate its features and was photographed with a medium red filter to further heighten the contrast. As a further control on the experiment, a large boulder was used as a uniform (control) surface. By the method of direct pecking five strips 20.3 cm by 2.5 cm (8 inches by 1 inch) (Fig. 6) were pecked into the surface using the four different raw materials available. The results obtained from this uniform surface, allowed us to roughly calculate how many square centimeters could be pecked per minute under ideal conditions by each of the tool materials. By applying these results to the measurements of various prehistoric petro- glyphs done on similar surfaces, it would be possible to estimate the number of man-hours that were employed in their execution. Results Direct pecking was employed on Surfaces A, B, C, D, H and I. As a group, the silicified tuffs were quite effective in pecking out the design. Surface C was accomplished in 35.0 minutes; Surface I required 36.0 minutes and Sur- face B, from the Desatoya Mountains, took 82.5 minutes from start to finish. A Texas chert nodule was found to be the most effective pecking tool, requir- ing only 30.0 minutes to execute the rake element design on Surface H. Less effective than either of the silicified tuffs or the Texas chert was the pecking tool of vesicular olivine basalt. This tool required 50 minutes to complete the petroglyph on Surface D. The least effective tools for direct pecking were found to be the hornblende andesites which required a total of 124.5 minutes in the completion of the design on Surface A (Fig. I and Table 1). The method of pecking/grinding was employed on the remaining surfaces E, F,G, J, and K. Again it was found that the Texas chert was the most 87 effective pecking tool requiring only 40.0 minutes to complete the design. Next in order of effectiveness was the vesicular olivine basalt, which took 65.0 minutes in the completion of the design. The hornblende andesite completed the design in 82.5 minutes. The silicified tuffs were found to be the least effective in the pecking/grinding method as they took 99.0 and 125.0 minutes respectively in completing the design (Fig. 2 and Table 1). In executing the designs by the pecking/grinding method, the Texas chert, silicified tuffs and vesicular olivine basalt all required more time than in the direct pecking process. With the pecking/grinding method, only hornblende andesite tools showed a decrease in the time required to complete a petroglyph. On the'Uniform Surface direct pecking experiment, the Texas chert again proved to be the most effective, completing the line in 12.0 minutes. The silicified tuffs were the next most effective, taking 18.0 minutes and 22.0 minutes respectively. The vesicular olivine basalt required 44.0 minutes to peck out the line and the hornblende andesite required the maximum time of 55.0 minutes (Fig. 3 and Table l). As a general statement, it should be noted that these results were ob- tained under ideal laboratory conditions and not under the field conditions encountered by the aboriginal petroglyph makers. This does not mean that the data should be rejected entirely, but that careful consideration and thought should be employed when analyzing or applying this data to any situation outside of the laboratory. Inferences From the experimental data and from observaton, it was possible to draw several inferences and conclusions that appear to be consistent with the results obtained. First and most important, this experiment has shown that direct pecking is the most efficient means, both in terms of time and labor, of manufacturing a petroglyph when compared with the pecking/grinding method. The time needed to produce a design of some exactness if surprisingly short when a sufficiently hard material is employed as the pecking tool. Even with a "softl" tool material, the time required is still acceptably short. It can be inferred, based on this experimental data, that the aboriginal petroglyph producers, with the raw materials available in the vicinity of the Grimes Point site, used direct pecking, as the means of producing most if not all, of the petroglyphs present at the various areas in the site. From observation, several of the petroglyphs manufactured in the experi- ment seemed to be easier to produce than others. This appeared to be due to material irregularities that were present on their surfaces. These irregu- 88 larities include hairline fracture holes, and weaknesses in the rock due to stress. It is quite possible that the aboriginal manufacturers deliberately sought out surfaces with these irregularities in order to reduce the time and labor involved in making a petroglyph. Again from observation, it was noticed that the basalt slabs had three distinct layers or "tskins" present on their surfaces. The first layer con- sisted of the dark brown "idesert varnish". Beneath this layer, a light brown layer was found, and underneath that layer was the pristine grey heart rock. A quite satisfactory petroglyph, in terms of contrast, can be made by just pecking through the "desert varnish" to the light brown or red layer beneath. It is entirely possible that many petroglyphs were made in this manner, thus time and labor involved would be minimized. To account for the differences in the results for the silicificed tuffs, Leitz (personal communication) suggests that the submersion of the basalt slabs and boulders under the waters of Lake Lahontan may have induced chemical changes to occur in the basalt (see Morrison 1964 for a discussion of the geology of the area). That is, the rate of weathering for the basalt may have been increased due to the effects of the submersion. This effect of differen- tial weathering, coupled with minor compositional differences among the basalt surfaces, and viewed from the perspective of technique variability from surface to surface, may help account for the differences between the similar raw materials. However, differences in technique from surface to surface do not seem to account for the results obtained by direct pecking between the surface of a Grimes Point boulder and a surface on a slab obtained from the Edwards Creek drainage area of the Desatoya Mountains (Fig. 4). Both are hornblende andesites and both were pecked using the direct pecking method with a sili- cified tuff as a pecking tool. The time required to peck a design on the Desatoya surface was roughly 2.3 times longer than the time required to peck the same design on a surface from the Grimes Point area. It would appear that the differences in the amount of surface weathering present on each of the two slabs could account for this time difference. While the Grimes Point slab had a dark brown patina of "desert varnish", the Desatoya basalt slab was relatively unweathered. That is to say, only a very light beige patina was present on its surface. From these results, one could infer that weathered rock material would be preferred to unweathered material, at least in the case of a basalt surface, for manufacturing a petroglyph. According to Sierts (1968:282), wetting the surface with water, saliva or some other liquid tended to soften the surface, increase the visibility of the design outline and helped in the removal of loose particles.' After the com- pletion of the replicative experiment, a fresh slab of basalt was alternately wetted and pecked. Qualitatively speaking, the results of this exercise agree with Sierts. The work appeared to move along faster thus implying a softening of the surface; the dust was kept down, visibility of the design increased 89 and the noise from the pecking operation was somewhat muffled. Hence, it is quite possible, that during the manufacture of a petroglyph, the surface was frequently wetted in order to increase efficiency and to decrease the time involved in its execution. From a personal inspection of the various petroglyph sites in the Grimes Point area (see Heizer and Baumhoff 1962 for a full discussion), the authors found numerous rounded and end-battered cobbles of hornblende andesite present in close association with many of the petroglyphs. This would seem to imply that the petroglyph makers utilized these cobbles as pecking tools. This would appear to be supported by the fact that the surface rocks of hand size in the vicinity of the sites are sharp, angular, and irregularly shaped. Rounded cobbles would seem to suggest a stream or water environment which is not pre- sent or evident around the sites. However, the experimental data indicates that the best possible results would be obtained by using the silicified tuffs, "cherts" and "rhyolites"tfound in the area, rather than hornblende andesites or vesicular olivine basalt. Since only one or two "hard" pecking tools would be needed for a moderate-sized design, there would be no great hardship involved in procuring the pecking tools and transporting them to the work site. The authors propose that the silicified tuffs, "cherts" and "rhyolites" available in the area were used along with the various basalts present as pecking tools. At this site, the basalts were discarded and the cherts and rhyolites may have been reused as raw material in the manufacture of other chipped stone artifacts. This tool reuse would also account for the paucity of silicified tuff, "chert" and "rhyolite" tools at the sites, while the large number of hornblende ande- site tools at the sites can be accounted for by the fact that they are totally unsuitable as a raw material in the manufacture of chipped stone artifacts. Scattered chert flakes have been found at NV-Ch-71 near the Grimes Point petroglyph site (Karen Nissen, personal communication) in association with the petroglyphs. These flakes could represent damage done to chert tools through use as pecking implements. While this evidence is scanty at best, it does seem to offer some, albeit meagre, support in favor of the hypothesis. With- out question, further research is needed to disprove or validate this idea of raw material reuse. Although this experiment was confined to one specific area and was limited in scope, its general outline and content could be expanded and applied to the study of other petroglyph areas. It would seem that each area would have to be tested individually due to the varying factors present at any given site. Undoubtedly, future research whether at the Grimes Point site or at other petroglyph sites, will shed more light on the mechanics and inferences that can be derived from the manufacture of petroglyphs through replicative means. 90 Acknowledgements The authors gratefully acknowledge the advice, assistance and encourage- ment provided by Dr. Thomas R. Hester (University of Texas at San Antonio) and Ms. Karen Nissen (Department of Anthropology, University of California, Berkeley). To Mr. Robert Leitz, Department of Geology, University of California, Berkeley, we express our special thanks for analysis and identification of the raw materials. Note The authors applied the data to a petroglyph locality near Walker Lake, Nevada. Two petroglyphs, LMA 2-30804 and 2-30800, from this area now de o- sited in the Lowie Museum of Anthropology, Berkeley, were measured in cm . For example, if LMA 2-30800 was produced by direct pecking, using as pecking tools, either hornblende andesite, red silicified tuff, white silicified tuff, or vesicular olivine basalt, the following time requirements for design com- pletion were extrapolated, 69.3 minutes, 27.8, 22.8 and 55.7. Extrapolations of this sort are useful in a rough way, if factors such as varying stroke rate, fatigue, environmental conditions, and available raw material can be evaluated. 91 to~ ~ ~ -t 4.J ob (J-- .,4' U- 4r) $-U4 d d-4 ). u o -Ha) u 4G .) C4 91 C) CD F S 4 -4O 0 00 0 0 0 0 co 0 0 0 %O .I C4 'D0 cu c) 00 0 0 - -. 0 00 00 00 '.t '- 0 cP en ,fl a't Lt - o 00 ( 0 0 ( tl 0 00 c) '.0 rO ,D44 L L( a 0N - -4 r-% ON ON 4 -4 zo -4 _ -4 CO ~4 ooo x 0 ~ ~ 00 Q0) d 4 OO O > 0 0 (N E -4U m - 00 ,4 cn'o C~) .(fl t: 4 ,n co A ca 0,0W o o 0 F . H ' a) ,% .,, 44 cn $4 C N 1- uz :1 _ a 44 a d0 -H O ). v-4~~~~~~~~~~~~~~~ 4-4 0 0~~~~~~~~~~~~~~~~~~~~~~~~~- ca~~~~~~~~~~~~~~~~~~~~~~~~~~~c 0~~~~~~~'0 _ A a) - a 4- e r.4 < H - a)% ( 14 - .o o o 'H " w ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~~~~~~s U 'H 0H , C. C) P H 4 ) 4 H A (D u --4 ?) Ln 44n : X X cd n i 3 3 3 3 ac "04-i a)~~~~~~~~~~~~~~~~~~~~~~~~~~~i 4 ~~~~ 'r4 0 C)~~~~~~~~~~~~~~~~~~~~ 0~~~~~~~~~~~~~~~~~~ 4-4 p- a) co 4-1 92 0 It4 OW PW PW PW 04 N P LnP Xann Q m * - - - C m N PW N C m - N - Cm 0 @ cl 0 0 0000 0 0 CD P4 P P4 0 on o 0 C: CD o4 o4 ot o o o0 o 0o o q r- ooo on _ O4 a) cn~~~~~~~~~~ -I C c:~ ~ w O l 4 O CD Ln CD CD CD CD e- t 00 > _-1 C 00 0 r-r .r 4 e U e 0 0 0 e a A 1 X? OOOOOOQOOO4OOOOOOOOO 0 o C-44 C: w Q) < C _ r C CD C Dn - s _4 0 O 0 - 0 0 ON c pq e4 w I' I' Ln Ln Ln --I Ln Ln _n r- rl 00 Ln > 1 r- r- ItO 'n a) 4. 0) a) ? ?0 433 Ca 0) ~4.4 0)a) L4 4QLRa)uQa) L44 Q $4~( bO ?-4 *H 4 r* 4 H a r4A a .0: C --0 CC U- - l =)c : U g4 C) _ a -4 O p --4 -4 $4 H -4 X -4 C .r4 X 'H -4 0 0 r4 U 0 cu JJ -r *-I4 C - II , O 0,1 *H '-40 *H0)-H'-1a *1E -4 0 4 "0 a) 0 r Y Q t A r cu S aC c 4 QC')~.0r'%oOO~Q r 4-'%oO0'4 a ) a4 ;_~~~~~- -4 -4 -4 -4- -4 -4 < n -4- c.' tn aX Z 93 S 4 N u o - 0 0 tQ ~ ~ ~ c CC 0 CQ 00 0 , - *HI ~~0 -4 I 0 94 v H V 0d G) 0 C P %= w4 'H .C .u ".4 O H - h4 _X E z \l + \' 4-4 (Q .,-4+' , 0( 3 cv uzl~~~~ Ho 4 4 Xs V Eq n ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~U) C C-, 0 S Q) 0~~~~~~~~~~~~~~~~~~~~~~~ "-4 o~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~b '--4 O ~ ~ ~ ~ E II u 5:1 1ZIZIZZIIZI XI~ 0 0 0000 n ~ ~ c C' . 9C ~ 0 %\4 y-~~ S-. q- y S fLU ~ U. . JTQ 94 0 0 0 0 0 a CIQ U" 0\ E-iH 0\ C H 4-i 0 9 C H 'd H CH CH 14 0 o C) C) 0 C) H e Ut U) -p Ut 0s 0 Cu H ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~co 0 z C) :q C-,~~~~~~C 4.4 0 zi C) co 4-i t ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~94 p4 U)~~~~~~~~~~~~~~~~~~~~~~~~~~~4 CW14~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ CD -r. u. O ~ 0O Z~ ~ ~ ~ ~~~~~~~9~.UWU U~ 95 o o 0 0 0 * * 0 0 0 4.) 6,-I N 4) .r 0 ~~~0 enj 4) ho 0 C)4 0, g t H + QkCo 1 0 H 0 0 "4 .9 H 4) ~~~~~~~~~~~~~~~~~~~~~~44 rz4 2 u 0~~~ U' C"I co O '-4 ~~~~~~~~~~~ClP4. i-I 4~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~$ 0~~~~ o = r3- 3s Oa 0 coH 0 o Q) c 0~~~~~~~~~~~~~~~~~~~~~ 0 HI I