A paper presented at the 2002 Annual Meeting of the Society for American Archaeology in the Symposium -- Late Paleoindian Foraging Variability During the Early Holocene of the Great Plains and Rocky Mountains. The Symposium was organized by Jason LaBelle and John Seebach.

Variation in Paleoindian Lithic Assemblages Through Time

Tony Baker
March, 2002

Paleoindian lithic assemblages are often considered the "best in breed". The individuals, who have studied them, believe they can recognize a Paleoindian assemblage even if it is void of projectiles. On the other hand most people will suggest that Paleoindian assemblages, which are void of diagnostics, are indistinguishable across the Paleoindian Period. To test this hypothesis of similarity, a simple study of five assemblages spanning 1600 years was conducted. This paper reports on this study and its findings.

The Sites

Figure 1
The landscape in the Central Rio Grande Valley of New Mexico during the late 1950s and early 1960s suffered from a drought and high winds. By the mid 1960s it was ideal for surface collecting and my father, Ele Baker, and I were fortunate enough to be there at that time. During just a few years we found numerous Paleoindian sites and many of these were first reported by Judge (1973).

Five of these sites, two Folsom, two Belen and unfortunately only one Cody are amenable for testing this hypothesis of similarity of Paleoindian lithic assemblages. The sites are sufficiently large for statistical testing and considered single component campsites since only one type of projectile was recovered. They are located in the same environmental setting and within a maximum of 53 miles of each other (Figure 1). Except for South Belen, all are located on the Albuquerque West Mesa west of the Rio Grande. South Belen is on the east side of the River.

These five sites were obviously not dated. However, 14C dates for Folsom and Cody sites are well documented in the literature and in this paper we will assume dates of 10, 500 and 8900 years B.P., respectively. Belen has an unknown chronology. My father named the point in the 1960s after concluding it represented a new point type (Baker 1968). Today, I believe the Belen point belongs to the poorly defined Plainview type that also suffers from the lack of a good date. For this paper, Belen will be assumed to have an age somewhere between Folsom and Cody.

The Analysis

I have observed that scientists are often guilty of over-analyzing their data. In a sense, they measure with a micrometer data that were originally marked with chalk and cut with an axe. The data listed in Table 1 were cut with an axe. The artifacts were collected from deflated surfaces and there is evidence to suggest that the surfaces had been exposed a number of times prior to the 1960s event. As a result, these assemblages are probably missing the smaller flakes that were removed by the wind. On the positive side, I believe my father and I were the first to collect these sites because the vast majority of artifacts were found on the day each site was discovered. These sites have been watched for the subsequent years and during that time they have yielded perhaps 10%, at most, of the artifacts listed in Table 1.

Table 1 – Site Statistics

 North FolsomSouth FolsomNorth BelenSouth BelenCody
Lithic Mass Collected0.9 kg2.3 kg1.2 kg1.8 kg1.6 kg
# of Points & Fragments1624141720
# of Scrapers & Fragments19473912360
# of Other Artifacts
(incl. Debitage)

Table 2 presents the lithic mass from the five sites separated into the three broad categories of bifacial tools, unifacial tools and debitage. Projectiles and preforms are included in the bifacial group along with channel flakes from the Folsom sites. Scrapers are unifacial tools unless constructed on a biface. Debitage are those artifacts that exhibit no visual indication of rework or use to the unaided eye. The P value is the Student’s "t" probability that the two Folsom assemblages are the same as the three Belen and Cody assemblages.1

Table 2 – Mass Distribution

 North FolsomSouth FolsomNorth BelenSouth BelenCodyP
Bifacial Tools14.4% 16.3%14.5%6.0% 16.4%<0.515
Unifacial Tools27.8%34.3%61.9% 53.6%53.3%<0.010

Figure 2 is the same data as in Table 2 except the sites have been combined by Paleoindian group. The reader will note there are obvious differences in the mass percentages of unifacial tools and debitage. On the other hand the mass percentage of the bifaces is statistically the same across the three groups and less than 17% in any of the sites. Of that 17%, only about half of the mass is projectiles and projectile fragments. The other half of the bifacial mass is more interesting. The non-projectile half is largely small fragments of bifaces and very few whole artifacts. This suggests that bifaces with any significant mass were valuable as a source of lithic material and were removed from the site when the people left.
Figure 2 -- Mass Distribution

The reader should note that the goal of this study was to test the hypothesis that the lithic assemblages without their diagnostics were the same across the Paleoindian Period. This means that the projectile mass must be removed from the data to perform the analysis. Since the entire bifacial mass was not statistically different between groups, it was removed for the remaining analyses.2

Figure 3 depicts the mass distribution after the bifacial mass was removed. Not surprising, the obvious difference noted in Figure 2 is still apparent. So, at this early juncture, this study has rejected the null hypothesis that all Paleoindian assemblages are the same if the diagnostics are removed. The remainder of the paper will discuss how and why they are different.
Figure 3 -- Mass Distribution

To investigate this difference, I chose to focus on the unifacial mass and only occasionally include information from the debitage. The unifacial mass consists of exhausted tools that vary from the most expedient to the most formal. However, tools suggest function and function varies from site to site within the same Paleoindian-group and between Paleoindian-groups. Site function was not the objective of this study, so tool type was not analyzed.

The parameters chosen for the analysis are those usually associated with debitage. They are material type (local vs. exotic), cortex (presence vs. absence), platform type (flat vs. faceted) and size (grams per artifact). Only the simplest measurements were made to avoid over-analyzing the data. The size analysis was also performed on the debitage. Figures 4 presents the finding for the first three parameters and Figure 5 presents the size findings for both the unifacial tools and the debitage.

In Figure 4, the data are plotted so they all trend upper as Paleoindian groups become more recent. The percent of exotic material is 100% minus the percent of the local material. This same procedure can be used for the other two parameters. Unlike the material and cortical parameters, platforms are not present on all artifacts. In fact, only about 35% of the mass from each site had platforms that could be classified as either flat or faceted. The remaining mass had either no platforms or platforms so badly damaged that the distinction between flat and faceted was not possible. The percentage of mass with platforms from each site varied little and was not statistically significant with a P<0.593.
Figure 4 -- Unifacial Mass Distribution

The three parameters in Figure 4 are dependent and so they should have similar trends. The more the local material is utilized, the greater the chance of having cortical tools. Similarly, more flat platforms occur when flakes are removed directly from cores of local material as opposed to bifaces. It would have been totally unexpected if one of the three parameters had demonstrated a reverse trend to what they exhibit in Figure 4. In essence, Figure 4 is only saying that Belen and Cody were utilizing more local material than Folsom.

Figure 5 indicates there is also a size difference between the assemblages. The later Paleoindians created the larger artifacts. This may seem novel initially, but after consideration this trend is also what is expected with the increasing use of local material. Average artifact size can be considered a measurement of duration of life cycle. The larger the discarded artifact, the shorter its life cycle. Since, the later Paleoindians are discarding local material one would expect it to be larger. Therefore, artifact size is also related to the use of local material.

I purposely have not discussed types of exotic material recovered from each site. I believe this information is not pertinent to this paper and would only be a red herring. However, all five assemblages had exotics and I have included the exotic data collected during this study in Note 3.

Figure 5 -- Artifact Size

The Discussion

In sum, the two Belen and one Cody (Later Paleoindian) assemblages are statistically different from the two Folsom (Early Paleoindian) assemblages. Specifically:

  1. The Later Paleoindian unifacial tools contain more local material, cortical material, and flat platforms than the Early Paleoindians assemblages.
  2. The Later Paleoindian unifacial tools and debitage are respectively larger.
  3. Based on mass percentage the Later Paleoindian assemblages contain more unifacial tools and less debitage.

The findings in 1 and 2 can be condensed into a single finding which is there was an increasing use of local material by the Later Paleoindians.

The findings in 3 are not as transparent as 1 and 2, but they are also related to material availability. Hypothetically, consider a knapper who is located in a lithic rich region. He reduces a 100-gram rock of local material into eight 8-gram tools and 36-grams of debitage. When the eight tools are exhausted, he discards them and begins with a new 100-gram rock. This knapper is creating an archaeological record that has a tool to debitage mass ratio of 64:36. Now consider a knapper who is located in a lithic void region. This knapper will create the same eight 8-gram tools and 36-grams of debitage, except it will be from exotic material. When these eight tools are exhausted, the knapper will not discard them because his lithic stores are very valuable to him. This knapper will refurbish or modify them and in so doing create more debitage. Lets assume he converts his eight exhausted tools into 13 3-gram tools and an additional 25-grams of debitage. When these are exhausted, he discards them and archaeological record has a tool to debitage ratio of 39:61. Additionally, the tools discarded in the lithic void region will be 3-grams in size compared to the discarded 8-gram tools in the lithic rich region.

At the inception of this study, the sites that produced the assemblages were selected near each other to minimize variability due to site location (Figure 1). Therefore, the Early Paleoindians in the area of these sites had the same access to the local material as the Later Paleoindians. So, why did the Later Paleoindians choose to use more local material?

The archaeological literature has numerous examples with similar trends as this study and a common explanation of the addition of the local material is "more" sedentism (Andrefsky 1998:221-229). This is an unsatisfactory explanation for me. In the Central Rio Grande Valley, there is no evidence such as structures, grinding stones, or even fire rocks, to suggest that any Paleoindians (Early or Later) were even slightly sedentary. The only evidence for more sedentism is the finding of this study that the Later Paleoindians used more local material. There is no other evidence.

The Hypothesis

In this section I will offer a different explanation for the varying percent of local material used in the Central Rio Grande Valley during the Paleoindian period. To do this I must add another piece of information to the record. Local material does not mean abundant and easy to gather. It means the material can be found within a 25-mile radius or so. In the Central Rio Grande Valley good lithic materials are available on the gravel terraces. However, even on the terraces, the material is not abundant or easy to locate. Additionally, the terraces are not located in the best areas for bison grazing. The terraces tend to be located along the drainages. Therefore, acquiring a five-gallon bucket of local material could be a several day effort for people on foot.

The data developed in the study indicates the Paleoindians were making a choice on how to deal with their constantly depleting lithic stores, a choice between curating or adding energy (Binford 1973) to their existing stores, or searching out and collecting additional local material.4 The Early Paleoindians chose to add the additional energy and Later Paleoindians chose to acquire more local material. These two different behaviors are apparent in the archaeological record in the percent of local material utilized and the size of the individual tools and debitage.

I propose this difference in Paleoindian behavior was the result of climatic changes. During the Early Paleoindian Period (Folsom) the climate was wet and cool. Ponded water and abundant grass existed virtually everywhere on the landscape. However, in the Later Paleoindian Period (Belen and Cody) the climate had become warmer and drier (Holliday 2000). Bison located on a wet landscape can wander in any direction to find water and food. They become dispersed and don’t have strong herding behavior. In a sense, each animal performs its own random walk across the landscape. Conversely, in the arid climate the animals are tethered to the few available resources. They were forced to routinely return to the few streams or live springs that remain on the landscape. The animals move in herds and become predictable as to their location during the day and the season. Animals in wetter conditions are much less predictable as to location (Elaine Anderson, personal communication 2002).

This difference in animal behavior precipitated the difference in Paleoindian behavior. The Early Paleoindians were tethered to the unpredictable animals. If they were to leave the small group of bison they were following for any length of time they probably could not relocate it. If a scout party was to leave the band for a few days to seek local lithic material, it was difficult to relocate the band. As a result, Early Paleoindians developed a lithic technology that permitted them to remain with the animals for the maximum length of time. It was a parsimonious behavior designed to conserve their lithic stores. When they ultimately exhausted they lithic material, they were forced to leave the bison and travel to a quarry. When this became necessary, they probably were unable to return to their prior group of animals.

During the Later Paleoindian period it was the animals that were tethered to the landscape and their movements were predictable. This predictability permitted the Later Paleoindians to leave the herd for several days to hunt for lithic supplies. Subsequently, they could return to the same herd because they knew where they would be located. Additionally, the heads were tethered to the streams and this is the location of the greatest number of gravel terraces. So the Later Paleoindians were often preying on animals that were located on the gravel terraces and they didn’t even need to leave the animals to acquire the local material.5

The Summary

This study began with the goal of testing the hypothesis that lithic assemblages were similar across the Paleoindian Period if the diagnostics were removed. Early in the analysis it was demonstrated that the assemblages of Early and Later Paleoindians from the Central Rio Grande Valley were statistically different. The Later Paleoindians used more local material, created larger tools and made less debitage. An explanation for the differential behavior is different climates. The animals of the Early Paleoindian Period grazed on a subpluvial landscape and were therefore unpredictable and moved in small or no herds. The Early Paleoindians chose to tether themselves to these animals and adopt a parsimonious lithic technology. The animals of the arid, Later Paleoindians Period were predictable and existed in larger herds around the limited water resources. These predictable herds allowed the Later Paleoindians to exploit the local material and not become disengaged with them.


Return to Paleoindian & Other... Home Page.


1 The Student’s "t" comparison of two sample means with unknown population variance was the only statistical analysis performed for this study. The two Folsom assemblages were averaged to create a mean and variance. The two Belen and one Cody assemblages were the bases of the second mean and variance. This combination was chosen based on the similarity of assemblages in Table 2.

2 The vast majority of Paleoindian lithic research revolves around projectiles because projectiles represent the vast amount of data. However, in the five assemblages investigated in this study, projectiles represent only 8% of the lithic mass. Obviously, Paleoindians broke rocks for reasons other than projectiles. A study of projectiles will illuminate information about projectiles. However, it does not say anything about the other 92% of the lithic mass. For example, prior to conducting this study, I believed only Folsom imported the exotic material Chuska into the Central Rio Grande Valley. This belief was founded on the knowledge that only Folsom projectiles in the area were made from Chuska. Investigation of the non-projectile mass for this study indicates Belen and Cody also imported Chuska. See Note 3.

3 Table of exotic material found in the unifacial tools. These are percents of the exotic mass and not percents of the unifacial tool mass.

 North Folsom
South Folsom
North Belen
South Belen
Alibates   4.0  
Chuska47. 8.5
Edwards   18.2  
Lake Valley 8.711.251.9 6.8
Obsidan16.717.8 2.21.1
Pedernal 3.20.411.6  
(Opaque Obsidian)
3.05.0   1.7 
White Opal22.511.4   7.97.4
Total100.0100.0 100.0 100.0 100.0

4 An experiment was conducted to measure pressure flake mass. Bob Patten, an accomplished knapper, attempted to make average to large pressure flakes. His efforts always produced flakes in fragments. Selecting and weighing the larger fragments yielded an average weight of 0.3 grams per fragment. The mass of the Folsom debitage in this Study has an average weight of 0.7 grams per artifact. This large size difference strongly suggests the Folsom people were using percussion to create their smallest tools. Percussion is an easy technique on larger masses, but it becomes increasing difficult as the mass becomes smaller. Increasing difficulty requires more effort be added to achieve small tools with percussion.

5 Bob Patten suggested the hypothesis of predictable/unpredictable associated with wet and dry times. Jan Cummings offered the fact that the gravel terraces are associated with the drainages.