lithic analysis

by Tara L. Potts and Philip J. Carr

Jump to: Lithic Analysis and Organization of Technology - Culture History
Lithics and Reconstructing Cultural Lifeways -
Flakes, Tool Use, and Other Lithics - Conclusions
- References

Flakes, Tool Use, and Other Lithics

Analysis of flake debris (n=59,186) followed general mass analysis procedures (Ahler 1989a; 1989b; Shott 1994). All flakes for each provenience were passed through a series of nested screens to determine size grades: 1-inch, ½-inch, ¼-inch, and ⅛-inch. For each size grade, flakes were counted and weighed. The majority of debitage analyzed using mass analysis was caught in the ⅛-inch screen (n=44,235, 74.7%). Flakes caught in the ¼-inch screen are the second most abundant (n=13,562, 22.9%), followed by ½-inch flakes (n=1,363, 2.3%), and the least amount in the 1-inch screen (n=26, 0.1%). Average flake weight in the ¼-inch size grade provides evidence to assess core reduction and tool production. Generally, a low average weight in this size grade is associated with tool production, and a relatively high average weight with core reduction. General comparative data are available from the Wickliffe Mounds (15BA4), a Mississippian town and mound complex near the juncture of the Mississippi and Ohio rivers. Carr and Koldehoff (1994) found that locally available Mounds Gravel—used in the production of utilized flakes from small, freehand cores and small triangular bifaces—had an average weight in the ¼-inch size grade between 0.52 and 0.65 g for different components (Carr and Koldehoff 1994: Table 4). In contrast, Mill Creek and Kaolin cherts are thought to have arrived at the site as large bifaces or finished tools, and the average weight for the appropriate size grade ranges from 0.41 to 0.58 g. For the Madison Park site assemblage, discussion of mass analysis results considers raw material types.

The majority of the material examined by mass analysis is quartz (n=50,021; Table 6-4). For quartz, the average weight in the ¼-inch size grade is high (0.69 g), a figure that supports a focus on core reduction. These data, in conjunction with the number of cores recovered, indicates quartz core reduction as a primary activity at Madison Park.

Not surprisingly, the average flake weight for Knox chert is much lower (0.41 g; Table 6-5), and indicates a somewhat different strategy in its reduction. This interpretation is supported by the relatively high percentage of Knox chert bifacial cores and low percentage of bipolar cores. The focus of the use of Knox chert appears to have been the production of bifaces, albeit often with minimal design and manufacturing effort.

A minimal number of nonlocal ¼-inch flakes were recovered and some caution should be exercised in drawing inferences from such limited data. However, the average weights of ¼-inch flakes for the Tallahatta chert (Table 6-6), Coastal Plain chert (Table 6-7), and Hollis quartzite (Table 6-8) are all lower than quartz and close to the average of Knox chert. Focus on tool production with these materials seems apparent; they arrived at the site largely as finished tools that were resharpened on the site. Again, this observation may be spurious given the small sample of non-local raw material.

Table 6-4. Quartz size grade data.

Size Grade

Count

% Count

Weight

% Weight

Mean Flake Weight

⅛-inch

37,299

74.57

5,175.45

17.57

.13

¼-inch

11,419

22.83

7,930.04

26.92

.69

½-inch

1,279

2.55

15,494.14

52.60

12.11

1-inch

24

.05

855.94

2.91

35.66

                SG1:SG2-4=2.93

Table 6-5. Knox chert size grade data.

Size Grade

Count

% Count

Weight

% Weight

Mean Flake Weight

⅛-inch

6,701

75.95

485.23

28.25

.07

¼-inch

2,047

23.20

849.01

49.43

.41

½-inch

73

.83

374.22

21.78

5.12

1-inch

1

.01

9.09

5.29

9.09

                SG1:SG2-4=3.15

Table 6-6. Tallahatta chert size grade data.

Size Grade

Count

% Count

Weight

% Weight

Mean Flake Weight

⅛-inch

28

50.00

2.35

5.98

.08

¼-inch

23

41.07

8.42

21.43

.36

½-inch

5

8.93

28.51

72.58

5.70

1-inch

-

-

-

-

-

                SG1:SG2-4=1.00

 

Table 6-7. Coastal Plain chert size grade data.

Size Grade

Count

% Count

Weight

% Weight

Mean Flake Weight

⅛-inch

1,623

77.51

12.71

15.56

.07

¼-inch

42

20.09

19.00

23.26

.45

½-inch

4

1.91

19.62

24.02

4.91

1-inch

1

.48

30.35

37.16

30.35

                SG1:SG2-4=3.44

 

Table 6-8. Hollis quartzite size grade data.

Size Grade

Count

% Count

Weight

% Weight

Mean Flake
Weight

⅛-inch

38

57.60

3.50

11.89

.09

¼-inch

27

40.90

13.74

46.67

.51

½-inch

1

1.50

12.20

4.44

12.20

1-inch

-

-

-

-

-

                SG1:SG2-4=1.35

Experimental studies (Ahler 1989) demonstrate that the ratio of ⅛-inch size grade flakes to the sum of the ¼-inch, ½-inch, and 1-inch size grade flakes can indicate the type of reduction that produced them. Low ratios indicate a focus on early stages of reduction (hard hammer percussion, bipolar reduction) and higher ratios indicate late stages (soft hammer reduction, pressure flaking). The sample sizes for Tallahatta chert and Hollis quartzite are small and, therefore, suspect for this analysis (Tallahatta chert SG1:SG2-4=1.00; Hollis quartzite SG1:SG2-4=1.35). For materials with adequate sample sizes, the lowest ratio is quartz (SG1:SG2-4=2.93), but the ratios for Knox chert (SG1:SG2-4=3.15) and Coastal Plain chert (SG1:SG2-4=3.44) are not dramatically higher. The high ratio for Coastal Plain chert is not surprising, since the majority of this material probably arrived on site as finished bifaces and hafted tools. The average weights and ratios for Knox chert and quartz indicate that these materials were used in all stages of reduction and a mixture of reduction strategies, including bipolar and bifacial.

Bipolar Analysis

A sample of 5,132 quartz flakes (1% of total quartz flakes) was size graded, classified as bipolar or non-bipolar, and the presence of cortex was noted. Of this sample, 2,457 (47.9%) are bipolar. Of this bipolar subset, approximately 77 percent exhibit cortex. This distribution is vastly different from the Knox chert flakes, which yielded approximately 4.7 percent (n=33) bipolar flakes. In this assemblage, only quartz and Knox chert were bipolarly reduced and both are local raw materials available only as relatively small water-rolled pebbles.

The majority of quartz in this sample fell in the ¼-inch size grade (n=4,462, 86.9%). The next most frequent size grade is ½-inch (n=607, 11.8%), followed by ⅛-inch (n=45, .9%), with the least frequent size grade being 1-inch (n=18, .4%). Interestingly, over 88 percent of the 1-inch size grade flakes are bipolar (n=16). Proceeding through the size grades, smaller flakes have lower percentages of recognizable bipolar flakes: ½-inch size grade yielding 470 bipolar flakes (77.43%), the ¼-inch size grade is 44.17 percent bipolar flakes, and none in the ⅛-inch size grade. This is due to the nature of bipolar reduction and the difficulty in identifying attributes indicative of this type on very small flake.

The bipolar reduction technique can be utilized for multiple reasons. It is an expedient way to produce a usable edge from a relatively small cobble. Bipolar reduction may also be utilized when low quality raw material does not permit other types of reduction (Andrefsky 1998:120). While bipolar percussion is useful for creating usable edges from small nodules, obviously there is a lower limit to the size a cobble can be worked. The larger the quartz nodule (again, within limits), the more likely bipolar percussion will be used. As mentioned above, our raw material study indicates that ½-inch pebbles are more likely to produce usable material than are the larger cobbles from the local river banks.

 

Non-Quartz Sample Individual Flake Analysis

A sample of chert flakes (n=702) was analyzed individually for the following attributes: raw material, size grade, weight, portion, platform, platform facet count, dorsal cortex, and dorsal scar count (see Appendix H for a list of attribute states). Comments on heat damage or possible retouch were noted on the analysis forms for appropriate specimens. Based on recorded attributes, each flake is assigned to a reduction stage (early, middle, late) following (Magne 1985) or classified as shatter (sensu Bradbury and Carr 1995). Magne (1985) found, through flintknapping experiments, facet count and dorsal scar count to be the variables most effective in assigning individual flakes to a manufacturing stage. His analytical methods are followed here. Flakes with an intact platform are assigned to a reduction stage based on the number of platform facets (0-1 facets = early stage; 2 facets = middle stage; 3 or more facets = late stage). Flake debris without an intact platform, but with a distinguishable dorsal surface, is assigned a reduction stage based on the number of dorsal scars (0-1 scars = early stage; 2 scars = middle stage; 3 or more scars = late stage). Flakes lacking both an intact platform and a distinguishable dorsal surface (shatter) cannot be assigned to a reduction stage by this method.

While stage classification (utilized here) provides demonstrably more accurate inferences than the traditional PST typology or the interpretation-free method, its applicability to materials common in the project area, such as gravel cherts, have not been explored through experimentation. Experiments with nodular chert have shown that platform facet count tends to be a conservative measure of reduction stage, and dorsal scar counts are more liberal. That is, classification of flakes by facets usually overestimates early-stage flakes, and counts of dorsal scars overestimate late-stage flakes. Therefore, additional data are gathered to examine the accuracy of stage determinations. For example, if an assemblage is dominated by core reduction, there should also be a relatively high frequency of cortical flakes, a high average weight for ¼-inch flakes, and a high percentage of shatter. Through the use of multiple lines of evidence, inferences are strengthened or ambiguities revealed.

A majority of the non-quartz ¼-inch sample is Knox chert (n=677). The remaining material in this sample includes Coastal Plain chert (n=19), Hollis quartzite (n=4), Tallahatta Chert (n=1), and Fossiliferous Bangor (n=1). For this discussion, due to small sample size, all non-local materials are discussed in aggregate.

Insight into reduction stage classification can be gained through an examination of flake completeness. Of this sample, a majority is complete (n=393, 56.0%) and there is a relatively low percentage of shatter (n=36, 5.1%). All of the shatter from this assemblage is Knox chert, except for one piece of Tallahatta chert. The ratio of complete flakes to shatter can be used to interpret reduction activities. Baumler and Downum’s (1989) experiments with ¼-inch debitage indicate that a significantly larger number of complete flakes over shatter are created during late stage tool production and resharpening. Conversely, a larger amount of shatter compared to complete flakes indicates core reduction. In this sample, the number of complete flakes (n=393) and shatter (n=36) suggests that the non-quartz materials were primarily used for late stage tool manufacturing.

Two platform types are particularly instructive for examining lithic reduction. Flakes with a lipped platform are usually created during biface (especially late stage) production, and flakes with a crushed platform are most likely from core (especially bipolar) reduction. Only 54 flakes (7.7%) with a lipped platform were identified. This low number indicates little focus on late stage tool manufacture. This is illuminated in part by the technological analysis of the hafted bifaces, which demonstrate the minimal effort used in stone tool manufacture, particularly in the form of retouch. The number of flakes with a crushed platform is also low (n=27, 3.80%) and all are Knox chert. Somewhat surprisingly, only 83 flakes (11.8%) have cortex on the platform, which points to some ambiguity in the reduction stage analysis. But once again, cortical platforms only appear on Knox chert flakes. Additionally, two flakes exhibit cortex on the platform and are lipped. The low number of crushed platforms in the non-quartz analysis supports this as well.

Of the flakes with discernable facet or scar counts (n=633), 45 percent (n=285) are early stage. The remainder is more evenly distributed between middle (n=183) and late stage (n=165). Of flakes with a dorsal surface (n=601), a majority (n=285) exhibit three or more scars, which suggests late stage reduction activities. Of the chert flakes, middle stage reduction (n=156) and early stage reduction (n=160) are similarly present, but in much smaller numbers than late stage reduction. Of the complete flakes in this sample (n=393), a majority (n=179) exhibit three or more scars. In this assemblage, 102 complete flakes exhibit two dorsal scars and 93 exhibit one or no scars.

Interestingly, if the stage designations are determined by platform characteristics for this sample, stage determinations are different. Based on platform characteristics, a majority (n=188) should be classified as early stage, followed by middle stage (n=122), and late stage (n=83). These discrepancies indicate that the small nodule size of Knox chert makes a strict comparison to Magne’s experiments problematic. The conservative nature of platform facets compared to dorsal scars is evident. Flintknapping experiments with small chert nodules are necessary to better understand how dorsal scar and platform facet counts change through the reduction process. Overall, the results indicate that all stages of reduction are present in the assemblage.

Amount of dorsal cortex has long been used to gain insights into flake debris assemblages. The largest proportion of flakes have no dorsal cortex (n=340, 48.36%), followed by those with 1 to 49 percent dorsal cortex (n=206, 29.30%). However, considered together, more flakes do exhibit cortex than do not. An examination of cortex amounts supports a focus on core reduction at 1MT318 and also reflects the nature of the raw material.

For quartz and chert materials, all stages of lithic reduction occurred at this site, from core reduction to biface completion and resharpening. Ambiguity revolves around whether the majority of recovered flakes result from core reduction or whether core reduction, initial tool production, and tool finishing were equal contributors to the assemblage. Available nodule size makes comparison with extant flintknapping experiments problematic.

Chipped Stone Tool Use

In addition to an examination of stone tool manufacture and technological lithic studies, an understanding of how stone tools were used can provide additional insights into the lifeways of prehistoric peoples. Archaeologists have developed use-wear analysis to move beyond the oversimplified equation of form with function. Both low- and high-magnification techniques require significant skill and experience. A key element to the success of such studies is having a comparative collection of tools of known uses, and it is beneficial if this collection is made from raw materials found in the archaeological assemblage of interest. Here, a simplified examination of use is based on tool failure types in some cases and classifying diagnostic bifaces as arrow or dart points using statistical formulae.

Use Failure Types
The most extensive study of biface failure types was undertaken by Jay Johnson (1978). While a range of failure types is discussed, only two are unequivocally associated with use: haft snap and impact fracture. Impact fractures occur when a biface is subjected to enough force during direct impact with an object to cause the distal end to fail. Haft snaps occur when the blade breaks at the haft, usually during use but also during retouch (Johnson 1987:52). Other failure types that can occur during use, but are also common in production, are not discussed here.

Three bifacial tools exhibit impact fractures—one Madison point of Knox chert and two quartz bifaces. The quartz bifaces exhibit retouch, but can not be assigned to a culture-historical type. These are probably finished tools and may have suffered haft snaps as well. Given the low quality of the material, it is difficult to make this determination. Two Knox chert Hamilton points, one Knox chert Madison, and one Coastal Plain chert Madison exhibit haft snaps indicative of use as projectile points, but may also result from failure during resharpening. One Knox biface has in impact fracture and a haft snap, suggesting that this point was doubly damaged during use. Impact at the proximal end of the point may have caused the failure at the haft as well.

Arrows or Darts?
Madison and Hamilton diagnostic bifaces are considered arrow points. But determining how the other diagnostic bifaces at Madison Park were used is a more difficult task. Researchers have suggested metric measurement can give insight as to the use of a biface when direct information, such as microwear or failure types, is lacking (Thomas 1978; Shott 1993; Bradbury 1997). Since Madison Park has been dated to the Late Woodland, we expect that a majority of the hafted bifaces are arrow points. Researchers have conducted analysis of prehistoric tools of known function as well as experiments to determine biface measurements that most accurately indicate whether artifact forms were used as arrows or darts. These formulae are based on the observation that size limits functionality of arrow points. Specifically, some bifaces are too heavy to serve as arrow points. While there is a slight degree of misclassification in each study, we decided to apply Shott and Bradbury’s recently derived functions to data from Madison Park. Shott argues that misclassification of dart tips as arrow points could be caused by resharpening. A biface originally created for use as a dart would be discarded after resharpening in the shaft, resulting in a spent tool. Archaeologists then recover spent darts and interpret them as arrows due to their small size. Statistical characterization of these tools based on metric attributes is perhaps the most accurate way available to sort arrow from dart points.

Shott (1993) modified the original formula by Thomas (1978) to deal with broken and triangular bifaces. Bradbury (1997) later tested Shott and Thomas’s functions and further modified the formulae. Using discriminant analysis of large data sets, these formulae were created. The original formula created by Thomas remains relevant because it contains the variables used in subsequent derivations:

Darts: C=.188ML + 1.205MW + .392MT-.223NW – 17.552

Arrows: C=.108ML +  .470MW + .864MT + .214NW –  7.922

where C= calculation; ML= maximum length; MW=maximum width; MT=maximum thickness; NW=neck width (Thomas 1978:470).

Measurements for each biface were inserted into both formulae. The formula that produces the largest number as an outcome is used to classify the projectile point. Shott suggests for triangular bifaces that base width measurement should replace neck width. Bradbury tested these calculations and determined that the number of misclassifications is reduced by using only maximum width and neck width. Because most of the points from Madison Park do not have necks, few (n=14) could be used with Bradbury’s formula, so the remaining points were classified using Shott’s formula (n=95).

Twenty-three hafted bifaces were classified as darts, while the remaining 86 are arrow points. A preponderance of arrow points is not unexpected from a Late Woodland assemblage. But the questions now are: Why do we find a significant number of darts at Madison Park? Is this outcome due to misclassifications? If not, were Late Woodland people in central Alabama still using darts, or are these points not projectiles at all?

Of the points classified as darts, types represented are Madison (n=7), Hamilton (n=4), Camp Creek (n=3), Gary (n=3), Bakers Creek (n=2), and indeterminate (n=3). As mentioned above, Madison, Hamilton, and Camp Creek are all part of what Justice calls the Late Woodland/Early Mississippian Cluster. Additionally, Madison and Hamilton are widely accepted as arrow points. These three types of points were classified using Shott’s formula, which uses length and thickness. Bradbury (1993:212) argues that these measurements, length specifically, can lead to misclassification.

Therefore, it is possible that all the specimens of these types are exclusively arrow points. As for the remaining darts—Gary, Bakers Creek and the Indeterminate—all are large stemmed bifaces that we think are more accurately classified as darts, or at least would not be used on arrows. Vertical distributions of darts and arrow points do not suggest temporal differences in the use of these two tool types. Level 1 contained the most darts (n=7) and arrow points (n=32). Arrow points clearly dominate Level 2 (n=15) and Level 3 (n=14), compared to the darts in those levels (n=5 and n=3, respectively). To summarize the pattern seen here, although arrow points were produced at a higher frequency, darts were also manufactured contemporaneous with the arrow points.
But were the specimens classified as “darts” actually used as darts? There are two plausible explanations for the appearance of bifaces classified as darts at Madison Park: (1) darts were still being utilized in the Late Woodland period, or (2) these bifaces are not projectile points, but rather knives. The adoption of bow-and-arrow technology does not mean that dart manufacturing and use ceased. Is it possible that darts were still utilized, just to a lesser degree than before? Despite the small package size of the available raw materials, and the suitability of those pebbles for arrow points, we know the inhabitants of Madison Park obtained larger pieces of raw material from which they made large bifaces. Could these large bifaces be knives rather than projectile points? Further study, particularly use-wear analysis, could answer these questions. Based on this limited analysis, bifaces were used as arrow points, probably as darts, and possibly as knives.

Chipped Stone Discard
Archaeologists generally assume that the majority of artifacts recovered archaeologically, particularly those that comprise the chipped stone assemblage, are the result of discard by prehistoric peoples. While Schiffer (1987) has laid the foundation for the study of artifact discard, lithic analysts have given too little consideration to this final aspect of an artifact’s use life. Here, we simply raise the question: why are so many presumably usable arrow points were recovered from the Madison Park site? We suggest that the answer to this question is likely complex, but that an organizational approach to the chipped stone tool technology has the potential to explicate this question.

Organization of Technology
Prehistoric people at Madison Park exercised a variety of technological choices in terms of raw material use, manufacturing techniques, activity distribution, and discard. These choices were carried out in a particular environment and in response to social and economic strategies they pursued. Their residences were located near a source of low quality lithic material, but at times they made the choice to acquire higher quality material from a slightly more distant source in some abundance. The minimal amount of non-local lithics from great distances and our technological analysis of flake debris suggest that non-local materials arrived on site in toolkits, mainly as finished tools. Whether these toolkits were brought by visitors to the Madison Park site or by inhabitants of the site conducting logistical forays to distant places where these materials occur is unknown at this time.

The Madison Park inhabitants employed both curating and expedient technological strategies. Bifacial arrow points, darts, and knives were manufactured in advance of use and employed at places other than their place of manufacture, both key components of curation (Odell 1996). However, little design effort is evident in the manufacture of hafted bifaces. This minimal technological effort suggests little concern over the risk of tool failure or with extending the use life of the tool. The employment of quartz, a low quality material, in the manufacture of stone tools indicates that neither tool reliability nor maintainability (sensu Bleed 1986) were much emphasized. This conclusion fits well with broader observations by Parry and Kelly (1987) concerning a lack of formal tools and the expedient nature of Late Woodland toolkits.

In terms of activities, all stages of stone tool manufacture are present at the site, as well as the full-range of formal tools in the Late Woodland toolkit. The large number of tested pebbles is indicative of stockpiling. Taken together, these are indicative of a sedentary lifestyle. Based on the lithic assemblage, we hypothesize that some segment of the population occupied the Madison Park site year-round for some years. A use wear study of formal stone tools and potential expedient flake tools would aid in testing this hypothesis, as would consideration of other lines of evidence.

Other Lithics

Non-chipped stone lithics were recovered at Madison Park. All were analyzed using the technological approach developed by Adams (2002). For each artifact, raw material, portion, number of modified surfaces, use wear, wear type, and kinetics were recorded. In this assemblage, raw materials include greenstone (n=36), quartz (n=7), and sandstone (n=4). Greenstone formations appear throughout east-central Alabama, especially in the middle Coosa valley. Although no greenstone was recovered during raw material sampling from the Coosa and Tallapoosa rivers, some of this raw material could have been transported along these waterways. Acquisition of greenstone might not have occurred at the location of outcrop, but rather closer to Madison Park (though not close enough to be considered local). The only local raw material in this assemblage is quartz.

Methods of Analysis

A technological approach to the study of non-shipped stone artifacts began by recording the portion of the artifact: complete, proximal, distal, etc. Use wear, wear type, and kinetics were recorded for each surface that exhibited evidence of use or modification. Use-wear categories include light wear, moderate wear, heavy wear, worn out, and unused, following Adams (2002). Light wear is defined as little evidence of use, such as slight pecking or microflake removal. Moderate wear is obvious damage, but the original tool shape has not been altered during use. Heavy wear occurs when the tool shape has changed or been broken during use. Worn out tools are no longer usable for the intended activity. And finally, unused tools may have been created for use, but there is no sign of use wear (Adams 2002:25).

Another technological attribute recorded is wear type. Wear types include adhesive wear, abrasive wear, fatigue wear, and tribochemical wear (Adams 2002:27). Adhesive wear occurs when there is little movement of two surfaces against each other and may not leave evidence of use. If this movement continues and pressure is applied, fatigue wear can appear on the surfaces. Examples of fatigue wear are pecking and flaking from use. Abrasive wear occurs when a harder object is rubbed against the surface of a softer object, such as using a stone tool to cut wood (Adams 2002:30). Finally, tribochemical wear refers to the build up of films on the surfaces of lithics. A common form of tribochemical wear is polish (Adams 2002:31). The different types of wear are not mutually exclusive. Determining the type of wear is difficult if a tool is used for more than one activity; evidence of its initial use may be destroyed.

Finally, kinetics was recorded for all surfaces. Kinetics refers to the “motions and forces related to tool operation” (Adams 2002:41). Kinetic classifications are: circular stroke, reciprocal (back-and-forth motion) stroke, flat stoke, rocking stroke, crushing, pecking, chopping, and battering. Once again, each surface is classified according to the type of stroke, if possible.

Results

As noted above, each modified surface was analyzed and the artifact was classified according to tool type, if possible. The purpose of approaching this analysis in terms of technology is to reduce misclassification and to identify tool types based on use and morphological characteristics. Tools are discussed by raw material type.

Quartz
Seven quartz tools compose the non-chipped lithic assemblage recovered during unit and feature excavations. Three cobbles have light to moderate pecking on an end, therefore identifying them as hammerstones. These three hammerstones do not have any other evidence of use. Two exhibit moderate to heavy wear on the end and face.

Other quartz nodules (n=7) exhibit pecking on one or more surfaces, pecking that likely occurred during use as an anvil for bipolar percussion. Given the abundance of bipolar percussion at Madison Park, we are confident of this interpretation. Additionally, stone against stone produces more wear and damage to surfaces than use against softer objects, such as nuts or wood (Adams 2002:34). Another hammerstone exhibits pecking on one end and crushing on the other, indicating that both ends of this hammerstone were used.

Interestingly, the largest of the quartz nodules (Figure 6-15) exhibits evidence of multiple surface use. Fatigue wear in the form of crushing and pecking appear on two sections of the lateral edge and on one face of the nodule. This quartz nodule seems to have been used a percussor and an anvil for bipolar reduction. Polish is evident on one end of the stone and on another section of the lateral side. Although there is no residue to indicate on what material this tool was used, we can be sure it was softer than quartz. Otherwise there would be fatigue wear present on these polished surfaces.

Figure 6-15. Largest quartz nodule recovered from the Madison Park site, with evidence of multiple surface use.

 Greenstone
A total of 36 greenstone fragments were recovered during unit and feature excavations. Thirty of these pieces are fragmented and less than ½ inch across. The remaining six are not complete, but are much larger with modifications that suggest they are fragments of hafted percussion tools. Four (Figure 6-16) are the beveled bit edges of axes/adzes. Along the bit edges tiny flake scars from use can be seen with the naked eye. Proximal or medial sections of these tools were not recovered from this site, so we can not determine overall tool size and shape. Fatigue wear along the edges indicates that the tools were used, and the fragmentary nature of the bits indicates tool damage during use.



Figure 6-16. The beveled bit edge of an axe/adze of greenstone, FS 381.

Sandstone
Two pieces of modified sandstone were recovered. Sandstone is a locally available raw material that appears throughout central and southern Alabama. Both of these artifacts exhibit heavy wear in the form of flat strokes. One in particular (Figure 6-17) exhibits six surfaces that have been ground flat. This artifact is classified as a pigment grinder for creating slips during pottery manufacture.



Figure 6-17. Sandstone pebble with six surfaces ground flat, interpreted as a pigment grinder, FS 543.


Copyright © 2013 by The University of South Alabama
Last Updated:
Wednesday, August 14, 2013 3:02 PM