By Amanda Castañeda
Most of the previous ASWT blog posts have focused on our ongoing excavations in Eagle Nest Canyon, with a few thrown in about earth ovens and our undying love for burned rock. So I thought it was time for a little change of pace! This post highlights another very common archaeological feature found in many of the sites within Eagle Nest Canyon and elsewhere across the Lower Pecos Canyonlands —ground stone bedrock features. Bedrock features, as I will often refer to them, include “slicked” areas, shallow grinding basins, deep mortar holes, and everything else in between.
I first became interested in bedrock features during my tenure at Shumla Archaeological Research and Education Center. We went to rockshelters across the region to record rock art, and bedrock features were a common occurrence in many sites. Most intriguing were bedrock features that were over 50 cm deep. I thought to myself, what on earth were the Lower Pecos inhabitants doing with these features when you can barely touch the bottom with your fingertips? Further, despite the ubiquitous nature of bedrock features in the Lower Pecos, they represent a largely understudied part of the archaeological record. In part, this is why I chose to explore this prehistoric technology for my recently completed Master’s thesis at Texas State University.
Bedrock Features 101
Archaeologists typically categorize these features by morphology and the perceived type of activity (e.g., pounding, reciprocal grinding, circular grinding, etc.). For example, grinding facets are shallow basins likely used to grind foods with a back and forth or circular grinding motion. Mortars are deep holes that were utilized for crushing or pounding, probably using straight up and down motions or possibly rotary or circular motions in some instances. Lastly, “slicked” areas are flat surfaces that have a shiny, smooth appearance and their function is unknown. The highly polished surface could be the result of multiple activities such as polishing hides or another activity that might include oily substances. Ethnographically, bedrock features of all shapes and depths were used for a variety of activities, mostly related to food-processing.
While generic terms, such as the ones listed above, have been used in previous bedrock feature research around the world, prior to my study there had not been any bedrock feature research completed in the Lower Pecos.
Therefore, I was most interested in creating a baseline data set of the morphological variation of these features. In other words, are we able to pick out any unique “types” of features and how are these morphologies distributed across the landscape? Further, using other lines of evidence such as use-wear patterns, can we determine what kinds of foods were being processed in these features? Essentially, I wanted to take a broad approach to my research and try to gain a better understanding of how these features were utilized by the hunting and gathering peoples of the Lower Pecos.
Recording Bedrock Features in the Lower Pecos Canyonlands
For my research I recorded morphological attributes (e.g., shape/size), use-wear characteristics (e.g., wear patterns left behind on the rock from different processing activities), and metric data (e.g., measurements) for 824 individual bedrock features at 10 sites using a combination of Structure from Motion (SfM) Photogrammetry (see SfM Revolution) and traditional field documentation methods. I recorded morphological attributes and macroscopic use-wear patterns using the form below.
I used SfM as the primary mapping and documentation method for each of the bedrock features I recorded. From the 3D data, I was able to create high resolution feature maps, and gather measurements for each feature in the mapping software ArcGIS.
Statistical Data Analysis
I focused the majority of my analyses on three different measurements: maximum depth of each feature and two axes across the opening of each feature. To summarize my results, I completed a cluster analysis using the metric data, which resulted in four distinct groups of bedrock features. The majority of the features (97%) fell into one very large, closely related group. I should note that Cluster 1 was comprised of four smaller subgroups, but there is only so much you can do for a M.A. thesis and still finish in decent time (as my committee wisely advised me!). The remainder of the bedrock features formed three smaller groups.
Looking at the graphic below, it’s easy to see how some of these groups were formed. Cluster 4 features are all extremely deep, between 40 and 60 cm. Cluster 2 features are moderately deep, between 20 and 30 cm. However, there are pieces of the puzzle the graphic doesn’t show. For example, the majority of Cluster 4 features have completely straight/vertical walls, while Cluster 2 features have sloping walls that form an overall conical shape. These types of observations were exciting because they are diagnostic attributes of these clusters that are independent of the cluster analysis. Said differently, they support the distinction of two different cluster which also had two different activities going on.
As for Clusters 1 and 3, they seem to blur together in the above chart. The reason for this is that Cluster 3 features are defined by at least one considerably long axis at the opening of the feature. Looking strictly at the length of the axis measurements, Cluster 3 separates itself from the rest of the sample. Cluster 1, in both graphics, is all over the map. This cluster contains a very wide range of variation in all aspects of the feature, depth, length, and width.
So essentially what the cluster analysis had shown me is that there are distinct morphological groupings of features, and each had a diagnostic characteristic that defined the group (except Cluster 1). The next thing I wanted to determine was if there were any other attributes of these clusters that further support distinguishing them from one another?
Bedrock Features on the Lower Pecos Landscape
First I looked at location. How are these clusters distributed across the landscape? Unsurprisingly, all ten sites have features that are included in the Cluster 1 group. In fact, four of ten sites have features that only fall into Cluster 1. The other three clusters are more restricted in their distributions. Cluster 2 occurs at three sites, Cluster 3 is present at five sites, and Cluster 4 only occurs at two sites. These data suggest that across the region, the majority of the food-processing that occurred could be completed in a non-specialized, Cluster 1-type feature. This could be due to the relatively small amounts of food being processed in most features or to the predominance of certain foods that did not need a specialized surface.
Use-wear Patterns of Bedrock Feature Clusters
The next characteristic I wanted to examine were the use-wear patterns. Did each cluster have distinctive or diagnostic wear patterns that might help me interpret the types of food that were processed in those features? Before I start throwing terms at you, let me first explain a little bit about use-wear studies on ground stone surfaces. In regards to ground stone bedrock features, differential use-wear across the surface of a feature can show what type of activity happened most recently. Is the surface pecked and rugged, or is it completely smooth to the touch? These conditions tell different stories about what happened last with that particular feature. When making use-wear observations, the objective is to observe traits about the macrotopography, or the high and low points.
Now going back to my data, the use-wear patterns further distinguished the Clusters from one another. Clusters 1 and 3 had very similar use-wear- they had rugged or pecked surfaces with either leveled or rounded high points. Cluster 1 feature walls was a rugged surface with either levelled high points or rounded high points. This suggests the area was first pecked to roughen the surface, and then different activities occurred to produce the modification on the high points of the feature. Levelled high points could have resulted from significant amounts of stone on stone contact (e.g., during fiber extraction), or if the processed material was hard in nature (e.g., seeds). In the instances with rounded high points, the surfaces were initially pecked, and then some sort of “soft” material was processed that smoothed the highs and lows of the peck marks. As the substance moved across the surface and around the high points, the surfaces became rounded. Softer materials potentially include a variety of plants (e.g., baked agave or sotol, nut meats, fruits) and animal tissue.
Cluster 2 features, the somewhat deep ones, had mostly leveled surfaces on the walls with some gradual, smooth rounded high points. This suggests that materials being processed in these relatively deep features were somewhat abrasive nature, and that the individuals using these features did not feel the need to re-peck the sides of the shaft to roughen the surface. The intensive levelling of the walls also supports the probability of a rotary motion being used, as to increase the contact between walls and the material being processed.
The most common use-wear pattern on the walls of Cluster 4 features are rugged upper walls and mostly leveled lower walls. This pattern suggests the upper walls did not come into contact with either the processing implement or the material being processed. Similar to the walls in Cluster 2 features, the lower half of these features must have been relatively full of semi-abrasive materials. This also suggests a pounding motion was utilized rather than a rotary or gyratory motion since the upper walls showed little signs of wear. However, two of the features in Cluster 4 are leveled on all portions of the walls throughout the shaft, suggesting a rotary motion may have caused the leveling.
Now for the Hole Story
So to sum it all up, there are definitely different morphological types of bedrock features in the Lower Pecos. One of my original goals was to put forth a regional typology of bedrock features. Although the cluster analysis resulted in four highly different morphological groups, Cluster 1 includes an incredibly large range of feature sizes and makes up the majority of the data set. Until Cluster 1 is examined more thoroughly for intra-cluster pattering, I think it is premature to create a formal typology. Clusters 1 and 3 are both highly variable and elude a classification that can encompass all of the morphological and metric variation. Other groups (Cluster 2 and 4) are less variable and likely represent a true morphological and functional type. At this time, I will tentatively classify features in Cluster 1 and 3 as general grinding surfaces, features in Cluster 2 as conical mortars, and features in Cluster 4 as cylindrical mortars.
Behaviorally, the overwhelming presence of generalized features such as the ones in Clusters 1 and 3 makes sense for a mobile, foraging group. These features take very little time to create, and they were likely used to process many different foods, whatever the group could find on any given day. Other feature types (e.g., Cluster 2, conical mortars; and Cluster 4, cylindrical mortars) were highly specialized and only occurred at certain sites. This pattern could have implications about general lifeways for Lower Pecos hunter-gatherers. Perhaps these foraging peoples were using the many sites with general purpose features for a majority of the year, but sites with specialty features could signal use during certain times, such as a harvest or large social gathering. These theoretical ideas along with experimental work can help archaeologists push our interpretations of ground stone bedrock feature technology past just food processing and into theories regarding site reuse and optimal technological adaptations.
Before I began my research, I expected the results to show many more distinguishable groups or morphological types. Undoubtedly, further analysis of the data will yield a greater insight into what morphological groups may be hiding in Cluster 1, which might change how I’ve interpreted the data thus far! There is so much more to learn, these types or groups are not set in stone…well they are, but you know what I mean. There are also other avenues to explore such as residue studies, which could give us an even better understanding of exactly what was processed in a feature. There are experiments to be done to figure out how long it takes to create a feature 50 cm deep in limestone! We’ve only just begun to peck the surface. But isn’t that the most exciting thing about archaeology? The more we learn, the more questions we have, and the whole process begins again.
This blog post is meant to be a simplified summary, there is much more to this story! If you want to learn more about bedrock features in the Lower Pecos, you can download my full thesis from the Texas State Library webpage here.
Dubreuil, Laure 2004 Long-term Trends in Natufian Subsistence: a Use-Wear Analysis of Ground Stone Tools. Journal of Archaeological Science 31:1613-1629.