- Berkeley Series
- Connestee Simple Stamped
- Deptford Overview
- Deptford Brushed
- Deptford Incised
- Deptford Cord Marked
- Deptford Linear Check Stamped / Fabric Impressed
- Deptford Linear Check Stamped
- Deptford Linear Check Stamped/Cord Marked
- Deptford Linear Check Stamped/Simple Stamped
- Deptford Simple Stamped
- Deptford Zoned-Incised Punctate
- Deptford Check Stamped
- Oak Leaf
- Swift Creek Complicated Stamped
Textile Marked Wares
- Woodland Plain
- Dan River Series
- Etowah Complicated Stamped
- Irene Complicated Stamped
- Irene Incised
- Lamar Complicated Stamped
- Napier Complicated Stamped
- Oldtown Series
- Pee Dee Complicated Stamped
- Santee Simple Stamped
- Savannah Series
- St. Catherines Series
- Uwharrie Series
- Woodstock Complicated Stamped
- Historic Period
As practiced in South Carolina (and elsewhere) ceramic analysis varies widely. Some researchers apply type names they learned at the feet of the legendary Dr. ____ of the University of ____ without really giving specific definitions. A sherd of pottery will be referred to as "Deptford" for instance, with a reference given to Dr. A's groundbreaking research and we are left to assume it really matches his description. This is not meant to diminish the contributions or impact of these elders in any way. However, just like potters, archaeologists create "communities of practice" where we learn from the elders and pass our knowledge along. This has led to many of the problems we face in trying to understand pottery, and thus culture, as old Dr. A and his students might disagree with Dr. B and his students, and end up calling the same pottery by different names.
Others, not wanting to apply a possibly incorrect type name, thus attaching specious interpretive baggage, use strictly descriptive terms instead of trying to make a sherd fit into a type (Steen and Chapman 2000 for instance). While not inaccurate, this leads to referential problems. That is, without empirical dates one can only wonder if the, for instance, coarse sand tempered cord marked pottery under discussion is in any way related to pottery of the same basic description found elsewhere. Since sand tempered cord marked pottery was made for 3,000 or more years not taking a stab at assigning a date limits the ability to make a meaningful interpretation.
Still others look at a number of variables and apply statistics to identify attribute clusters and try to define types empirically (Anderson 1975; Cable and Cantley 1998). As discussed in the previous research sections, approaches to ceramic studies in the southeast were shaped by the great early archaeologists like A.R. Kelly, James Griffin, Joffrey Coe, and Joseph Caldwell, among others. But none of those people worked full time in the state. While the importance of their pioneering efforts are indisputable, for our purposes we will focus on the modern era. While there are many bright and talented archaeologists working in South Carolina today, a half dozen or so people's past efforts have been central influences: Stanley South, David Anderson, John Cable, Chester DePratter, Chris Espenshade and Mike Trinkley.
Stanley South's work, conducted in the late 1950s and revised in 1976, only covered part of Horry County, and was limited mostly to the coast, but he set up a broad framework for classification that is still valuable.
In 1975 David Anderson put together the SCIAA type collection and provided type definitions for the ceramics there. He looked at collections from all across the state, and mapped the distribution of the major pottery types. Twenty years later he and others gathered for a symposium that resulted in the printing of the draft version of "Indian Pottery in the Carolinas," which remains the most comprehensive source on the subject. Although the pottery type descriptions are useful, the transcript of the proceedings is valuable for the insights it gives. In addition there are also good articles by Chris Espenshade, Jane Eastman and others in the volume.
At about the same time in the later 1970s Mike Trinkley was working on a PhD that focused on shell rings and the people using Thoms Creek pottery. He was instrumental in showing possible continuities between the wares seen in North and South Carolina and built a comprehensive Thoms Creek typology.
John Cable has led the way in doing intensive detailed analysis that is well described and well illustrated, and more important, he has been working on this continuously for his entire career, allowing him to refine his approach and apply what he has learned.
Chris Espenshade is influential in that he developed an analysis approach that was adopted by Brockington and Associates. They are the largest of the state's CRM firms, and have excavated more sites than anyone else, so Chris's approach has been applied to more collections.
As the old aphorism goes, there are many paths to enlightenment. Intensive consideration of pottery may not be on a given researcher's path. But, as discussed elsewhere on this site, there is only limited comparability in the different approaches taken in South Carolina. A major problem is the descriptive terminology used.
In this section the terms used by major researchers and their methods for analyzing ceramics will be introduced, and to some degree, hopefully, standardized. It will be organized like a lab manual, not to suggest that this is the one procedure that should be followed, but to try to cover as many details as possible. The basic concepts underlying the consideration of pottery assemblages in South Carolina are presented.
Chris Espenshade devised a multi stage approach to analysis that is based on the level of work: survey, testing, and data recovery/mitigation. At the survey level sherds are identified by:
- type name; (discussed below)
- decoration (or surface treatment)
- vessel fragment (rim, body, base, etc.)
- temper type
- weight (in grams)
Pottery at sites which require more extensive excavations receive a more detailed analysis. This is used to introduce basic concepts and definitions. Analysis approaches and terminology used by other researchers will be introduced as well.
Analyzing Phase III. (Data Recoveries – project specific)
|1. Type Name||16. Use Wear|
|2. Decoration||17. Use Wear Location|
|3. Interior Surface Treatment||18. Coil Height|
|4.Vessel Form||19. Coil Thickness|
|5. Temper Type||20. Coil Width|
|6. Temper Size||21. Coil Profile Shape|
|7. Temper Shape||22. Rim Shape|
|8. Temper Density||23. Rim Thickness|
|9. Minor Temper Type||24. Rim Diameter|
|10. Minor Temper Size||25. Oxidation State|
|11. Minor Temper Shape||26. Appendages|
|12. Minor Temper Density||27. Paste Type|
|13. Dark Core Retension||28. Count|
|14. Sooting||29. Weight|
|15. Fire Clouding||(from Severts and Brady 2006)|
As detailed as this is, there are others who look at even more attributes and, further, not everyone looks at the same variables. Coil width, thickness, height and profile are not widely observed attributes, for instance, but to Chris Espenshade, who experiments with pottery making, they are very important to his interpretation. While some attributes should always be recorded, there are some that might serve to inform a particular research interest that might not be relevant to others.
It can also be argued that it is not necessary to microscopically examine every last attribute, because a given attribute can vary widely within a given type. At some level of generalization it is acceptable to use broad groupings, like Joe Herbert's “New River Series” (Herbert 2003) which can then be cut into more specific types that have temporal and cultural meaning. But, as stated previously, we have suffered from a lack of consistency in terminology and the application of some very basic concepts.
At minimum we should all record the first six attributes introduced above. These and other attributes are discussed below.
1) Type Name. The concept of “types” can be nebulous, and is applied with a lot of variability. For some all Thoms Creek pottery is the type, while others consider Thoms Creek Plain a type, and Thoms Creek Punctate, and so on, separate types. Others would consider all Thoms Creek a series, or a ware group. I tend to generalize, and consider plain, punctate, etc under the broad category of Thoms Creek wares. After all, sherds from different parts of, for instance, a punctate decorated pot might be considered separate types if one happens to not be decorated. Basically a pottery type consists of sherds that share enough attributes to make them stand out as a group.
In the course of doing background research and compiling the database some 65 different type names were found which have been (or could be) applied to pottery found on sites in the state. These are listed below, with further discussion and formal type descriptions (where available) and further discussion in the Series and Types section. Very few of these series, types, varieties or phases is clearly defined, and in many cases the terms used by the original researchers, myself included, to describe pottery attributes are vague, undefined or incorrect. So it is often unclear from reading a type description if the sherd in your hand can be assigned to the “type” the author was describing.
The descriptor “medium grit temper” is a good example, because this seems to describe most of the pottery temper found in Georgia, apparently. The latter illustrates one of the problems we face when a traditional approach to pottery analysis is taken, because two important terms- “medium” and “grit” -are not defined, and, for that matter, “temper” is a little ambiguous too (see discussion below, and Rice 1987, and Shepard 1954). Some use nominal terms like “large, medium, and small” while for others “medium” refers to a given standard, such as the Wentworth Scale. Ceramicists Prudence Rice, Anna Shepard and Orton et al don't even use the term grit. Others use it to refer strictly to purposely crushed rock (Wagner 2009: personal communication). Chester DePratter uses it to refer to macroscopically visible grains of sand (Personal Communication 2009). Some North Carolina researchers (Herbert 2003) use it to refer to granular size sand (2-4mm).
So, in short, categorizing sherds first by “type” may not be a good idea. Below are the type names encountered in the literature review. Most of the links below lead to the "Series and Types" introduction. Other links lead to more detailed type descriptions from David Anderson and others where possible. In some cases, Dave Moore's Cowans Ford and Burke types, for instance, the types have not been identified in the state by anyone, but could or should be found here.
2) Surface Treatment / Decoration
To a ceramicist the overall appearance of a vessel's surface that is a result of its basic shaping is thought of as a surface treatment, while additional adornment is thought of as a decoration. This is a fine point, and from an artist's perspective, a debatable point. One might argue that if a potter did not want a vessel to have a complicated stamped surface she or he would not have used an elaborately carved paddle to shape it. But we will defer to the experts here and refer to surface treatment and decoration as separate production stages.
Common surface treatments include several degrees of “plain” which range from barely smoothed to highly burnished; stamping with a carved paddle; and textile marking. The more commonly used terms will be presented below. Incidental marks such as fiber tracks are considered elsewhere.
Brushed This can look like simple stamping, scratching or shell scraping. It consists of shallow parallel lines that look as if they were made with a stiff brush or bundle of twigs while the clay was still wet.
Burnished - see Plain
Check Stamped These impressed marks are made by applying a carved paddle to the wet clay. Checks vary in size and range from squares to diamond shapes. They are often found in a continuous, regular coverage, but variants can include single or multiple lines of checks separated by a wide space (linear) or checks combined with more complex designs.
Cob Marked This is made by rolling a corn cob across the wet clay. It results in marks that usually only vaguely look like a corn cob, however. In some cases it appears that a dried ear of corn was used, as the marks look like cupule impressions.
Combed Shallow, parallel more or less equidistant lines. Not commonly seen, and can be mistaken for brushing or scraping.
Complicated Stamped These are made with paddles carved with complex motifs. There are many motifs and variants, and they can be very sensitive markers of time and ethnicity. At a basic level one might simply differentiate between rectilinear and curvilinear designs, but in a detailed analysis careful attention should be paid to the particulars. Often additional decorations such as punctate, incising, and appliques are seen, and these combinations are temporally sensitive.
Cord Marked Cordage is wrapped around a flat paddle and used to malleate the coils and even out the body. Though at a distance it can all look pretty much the same, researchers have learned to examine cordage carefully. Usually they look at a number of cordage variables including size; looseness or tightness of twist; and directionality of twist (ie, S- left to right, or Z- right to left). Application variables are often measured as well. Often cord marking is predominantly parallel to the rim, angled one way or the other, or perpendicular. Over stamping sometimes looks haphazard, but other times obvious patterns are formed. On other occasions just the paddle edge is used leaving a mark many have referred to as “cord wrapped dowel.” This consists of rows of cord marks separated by a ridge (or land). The spacing of the cords is meaningful as well, and ranges from nearly overlapping to several millimeters apart. Though not seen here cord marking is used elsewhere to make elaborate decorations.
Corrugated Coils are not smoothed at all. This is not common in South Carolina, but is seen occasionally.
Dentate Stamped These decorations consist of lines of rectangular to squared indentations which appear to be made with a multi pointed tool like a comb. A roulette tool can also make marks like this (see Holmes 1903). Sometimes several lines are found together, but in general the markings seem to be applied separately.
Fabric Impressed As with cord marking fabric marked pottery looks pretty similar at a distance. Yet it was used for thousands of years, and when you look closer there is actually a good amount of variability. Most obvious is the nature of the weave - the weft and warp. Some impressions are from loosely woven soft fabric, but in others a rigid, non-fabric warp element was used. The threads, in other words, were weaved around sticks or pieces of cane like a mat, and this was used as a paddle. In other cases fabric was wrapped around a paddle and the surface is thoroughly marked, with no apparent patterning. And, naturally, there are cases where it appears that great care was taken to replicate the original fabric.
Finger Pinched Finger pinched decorations are usually seen on Awendaw Phase Thoms Creek ware, but finger pinched rims are seen on some late wares as well. As the name implies the decoration was made by pinching the wet clay.
Finger Grooved This too is an Awendaw Phase Thoms Creek trait primarily. The otherwise plain surface appears to have been grooved with the end of the fingers, leaving broad, shallow marks.
Incised With incising lines are cut into the clay with a stylus. They are sometimes random, and in other cases form intricate designs. They have been referred to nominally as fine, medium, and bold by some. Measuring the width and depth of the marks may allow significant variation to be identified. Some styluses are thin and sharp, or squared, or rounded, while in other cases it looks as if a broken off twig was used, so stylus variables should be considered as well. As with complicated stamped designs, characteristic motifs are identifiable, and can have temporal and cultural significance.
Net Impressed Nets differ from fabric in that each intersection of the warp and weft is tied with a knot. Sometimes the net marks are about all that is visible leaving behind a roughened surface while in others the grid of the net is clearly visible. As with wide spaced cord marking, this is likely a conscious choice, as most net impressing is thought to have been done with a net wrapped paddle.
Plain Who would think that there could be so much variability in plain wares? Yet researchers have identified a spectrum of treatments that ranges from nearly no effort at smoothing to polished, burnished surfaces. These variations appear to be meaningful. Plain finishes can be made by smoothing with the hands, paddling with a flat paddle, smoothing with a tool and burnishing. In some cases surfaces are re-wet, or “floated” to smooth them, resulting in bits of temper material protruding.
Punctate As the name implies punctate decorations are made by punching an object into the wet clay. These marks are often random, but are also seen in areas defined by incised lines, and in more or less continuous arrangements, such as a line around a rim. Reeds, sticks, periwinkle and cockle shells, fingernails, beaver teeth (Sassaman 1993) and other tools are used to make the marks. Most people identify separate punctate (even those in a line) and drag and jab punctate, where the tool is dragged from punch to punch.
Scratched This term is usually used when the marks appear too deep to be “brushed” and too random to be incised. Nominal though it may be, some sherds do appear to have been scratched.
Simple Stamped Simple Stamped designs are made with carved paddles and possibly, split wood. Parallel linear marks look like cord marking, but no cord marks are visible. These are usually U or V shaped, but a squared off variant is also see, As with cord marking there appears to be meaning to the orientation of the marks and the degree of over stamping. In some cases over stamping deliberately mimics check stamping. In others alternating blocks at 45 degree angles look like “line block” complicated stamping.
Slip Decorated A slip is fine clay suspended in water. Pigments are sometimes added to produce a particular color. This tends to be a different color from the body. It can be used as an engobe to cover the body entirely or it can be used to paint designs. Red seems to be the only color used in South Carolina, and it primarily occurs in the 17th and 18th century on wares such as Altamaha, Mission Red Filmed, and Kasita. The Catawba added black to their color palette in the 18th century, but this may have been done with sealing wax, rather than a clay slip (Riggs and Davis 2007).
Smoothed In some cases a surface treatment is partially of completely obliterated by subsequent smoothing.
3) Interior Surface Treatment
Usually interiors are smoothed to some degree, but occasionally they are marked using either the same technique as the exterior, or in a contrasting manner. It is rare for interiors to be extensively stamped or marked with fabric, but it not unusual for the top few centimeters to be cord marked or simple stamped. Check stamping, and incising are sometimes seen, but brushing and scraping are more common.
Floated The surface is re-wet and smoothed with the fingers. Temper is not pressed into the body and may protrude, while still being coated with a film of clay.
Plain As with exteriors plain wares range from roughly smoothed to highly burnished. As a generalization one can say that there are two variants, finger or hand smoothed, and tool finished. With the former surfaces can tend to be lumpy, while with the latter the tool presses temper bits into the clay. This is a nominal category. Researchers usually use terms like poorly smoothed, well smoothed and the like. Burnished surfaces should be sorted out, as this tends to be an important trait on late wares.
Shell Scraped Parallel striations that look like the edge of a clam shell are often seen on vessel interiors. These are less random than “brushed” marks and not as deep as “scratched” marks.
Slipped Slip is defined above. It is sometimes used to coat interiors or accent an edge. Less often it is used to paint designs. This is common to the west, but seldom seen in South Carolina.
4) Vessel Part
In preliminary analysis it is usually sufficient to categorize sherds as rim, body and basal. Very few other types of sherds are seen in South Carolina, but one might encounter handles, tetrapods, rosettes and other applique elements, shoulders from carinated bowls, necks from restricted neck jars, collars and other oddities that should be noted at this stage. Sherds less than a given size (depending on the researcher) are usually considered “residual sherds” or “sherdlets,” counted and weighed, and set aside.
5) “Temper” or “Inclusions” or “Aplastics”
Temper is a word, or concept, that is thrown around loosely, as Prudence Rice (1987) pointed out long ago. For a number of reasons she prefers the term “inclusions” (Rice 1987: 13). When the word "temper" is used here it is the equivalent of Rice's "inclusions." Essentially temper is a material added to the clay to modify its “workability, drying characteristics, firing behavior and final fired characteristics” (Rice 1987: 408). A variety of materials are used worldwide, but here in South Carolina we see primarily fiber (organic temper), quartz sand, grog, and crushed rock (quartz, and feldspar primarily, but also limestone). Shell is a common tempering agent in the neighboring states, but is seldom seen here. Shell and limestone tempering are often discernible only by the holes they left behind when the calcium dissolved.
One problem with temper is deciding which inclusions are intentionally added, and which are constituents of the clay. Absolutely “pure” clay is hard to find. Nearly all sedimentary clays and even primary deposits have some amount of sand or unweathered rock. In reading about pottery I have found numerous exhortations for people to study local clay sources to begin to address this question, or at least to bear it in mind when discussing "temper" (Southerland 1973; Trinkley 1973; Saffir 1979, for example). Joe Herbert and Theresa McReynolds' "COMPOSITIONAL VARIABILITY IN PREHISTORIC NATIVE AMERICAN POTTERY FROM NORTH CAROLINA" is a good start at considering the nature of potting clays, but only barely reaches into SC. They studied clays and sherds from the area of the Johannes Kolb site (38DA75) on the Pee Dee.
The question of inclusion density is difficult to address. Inclusion density can be affected by local clay sources, resulting in clays that are naturally coarse and sandy or fine and silty. Two experiments have been conducted to illustrate the problem. In the first I ran natural clays through standard geological sieves. These did not match the Wentworth Scale precisely, but are close. The material included in each grade are what did not pass through the sieve (ie, size grade 1 is 4.75mm or larger).
- Grade 1 (#4 sieve) approximates pebble sized inclusions-4.75mm or larger.
- Grade 2 (#10) is Granular- 2-4.75mm.
- Grade 3 (#18) is Very Coarse sand- 1-2mm.
- Grade 4 (#35) is Coarse Sand- .5-1mm
- Grade 5 (#60) is Medium sand- .25- .5mm.
- Grade 6 (#120) is Very Fine sand- .125-.25mm
- Grade 7 is less than .125mm
Relatively “pure” white kaolin clay collected in Aiken County contained a fair amount of “background sand”. A 50 gram sample was run through the first four geologic sieves. Medium and smaller grains were not further separated. Of the 50 grams, 2.6g were coarse sand or larger (#1, 0g; #2, .1g; #3, .4g; #4, 2.1g). Though not found in the 50g sample, pebbles up to 30mm across were found in the clay I collected as "pure" kaolin. Thus 5.2% of the “pure” white kaolin clay is actually coarse to granule size sand.
I ran a 50g clay sample from Horry County through the same sieves, and later expanded the range with another two sieves. I collected this clay in the early 1990s and it served as a good illustration of the temper/inclusion question for me at the time. The deposit was exposed in a ditch bank, where it could easily be seen in a lens sandwiched between sand lenses. The top and bottom of the lens graded from sand to sandy clay to "pure" clay that is still some of the best potting clay I have found.
Of the 50g sieved I was surprised to find that 1.4g did not pass through the largest sieve (#1), with pebbles up to 10mm across present. 2.9g were granule to pebble size (#2), 3.5g very coarse (#3), and 4.1g coarse (#4). Thus nearly a quarter (23.8%) of the raw clay was coarse sand or larger. A pot made with this clay could easily be thought of as coarse to very coarse sand tempered, with occasional granule and pebble inclusions. So with no conscious effort by the maker a pot decorated with textile impressions could be considered "New River," "Cape Fear," "Mount Pleasant," "Deptford" or "Deep Creek." This was an important lesson, because the pottery on the sites we were examining at the time could easily be called either type (Steen and Errante 1993). Just for the record, an additional 7.6g was between .25 and .5mm (medium); 23.1g was between .125 and .25mm (fine); and .7g was less than .125mm (very fine/silt). The remaining 6.67g was lost in handling, or stuck in the sieves.
Two other samples were previously fired, and could only be observed. At around the same time in the 1990s I collected some clay at Peachtree Rock, near Lexington in the Sandhills. This deposit is sedimentary, having formed in an ancient embayment where it capped a thick layer of silicified sandstone. It has a high kaolin content, so it was hard to work, and didn't mature completely in the firing. This example would appear to be tempered with everything from grog to limestone to granular sized sand, but it is the unprocessed raw material- nothing removed or added. This is the clay used to make what we called the "Oak Leaf" variant of Deptford. This is named in honor of the site's owner, the Nature Conservancy.
Clay samples from Fig Island were discussed previously. Unfortunately no samples were retained for sieving, but an examination of broken sherds shows that fine to coarse sand is present and accounts for at least 30% of the surface area. Larger inclusions up to 1.3mm across were noted in this sherd as well. The surface was smoothed by hand - not tooled- and has a distinctively "sandy feel." Incidental plant fibers and hematite inclusions were present as well.
So care should be taken when considering inclusions. Again, this is something people have known for years, but it is still worth reiterating. John Cable and Sean Taylor both account for "background sand" in their microscopic examination of sherds, for example. An anomalous inclusion value might not necessarily communicate the intent of the maker. It might be the local clay source.
The second experiment involves making “temper bars”. In this pure commercial potting clay and inclusions of the first four sizes listed above (medium and smaller were not considered) were combined by volume in 10, 20, 30, 40 and 50% densities. This shows clearly the effect of inclusion density.
The first step was to dry the clay and reduce it to powder. To obtain the inclusions I first went to a local construction site and gathered a bucket of sand. This turned into a lesson on its own, as I soon discovered that Midlands Sandhills sand is almost 100% medium or finer. I went to Fort Jackson and collected more. Same thing. Finally I dug out a soil sample from Peachtree Rock and found enough coarse material to make 100ml bars in all four size grades.
This was done by first pouring the powdered clay into a graduated cylinder, shaking it to settle the material, and then pouring in the proper volume of inclusions. For example, a 10% density includes 10ml of inclusions and 90ml of clay. This was shaken thoroughly to mix the material evenly, and then wet and formed into patties. These were scored across one side to facilitate even breakage to expose a fresh surface. The size grade and percentage of inclusions was incised on the upper surface of both halves. These were documented photographically, and a control sample of pure clay was also made.
The patties were then dried, and fired in a wood fire that would approximate traditional firing practices. The firing process showed some interesting points on its own. First, there was breakage. With the granule to pebble size grade (#1) the bars at 10, 20, and 30% cracked in the firing, as did the 20 and 40% bars in the #4 grade (Very Fine to Medium sand). In other cases where this commercial clay was used firing breakage was also seen, so the clay may not be well suited for direct firing. The lack of breakage in the Coarse to Granular bars suggests that the addition of larger inclusions helps to alleviate this problem.
To allow these bars to be used for estimating inclusion density in the lab I have photographed them as a group, and in detail. These images can be printed out and used for direct comparisons. This differs from the Matthews Scale in that the 40 and 50% ranges are included, and the inclusion sizes are regularized. In the Matthews Scale, which probably approximates natural paste composition better than my method, sand of all sizes from Medium up to a given value (1, 2, or 3mm) are included.
For me and for others the issue of whether grog tempering was crushed sherds or purposely fired clay was avoided by grouping them as “clay tempered” wares. Sorting sherd and grog tempering seems to make sense. St. Catherines ware, for instance, has a distinctly finer temper than its supposed predecessor, Wilmington. But with most grog tempered wares it is only occasionally possible to definitively identify sherd surfaces, forcing one to assume a correct identification has been made.
Prudence Rice points out that, by her definition, the mixing of clays from different sources to improve workability is tempering too, so the “clay tempered” term as I have used it, is incorrect. She gives an example from Pakistan, but we can point to the historic Catawba right here in SC who mixed their "pan" and "pipe" clays carefully to get just the right recipe (see discussion in Holmes 1903. As a result "grog" is considered the better term, but sherd tempering should be identified when possible and noted. As with sand and crushed rock size should be considered and size grades defined.
Summary of Common Inclusion / Temper Types
Sand Clear and light colored quartz sand is most common, but rose and other colored quartz is seen at times. This may be a conscious choice. It is strongly suggested that the Wentworth Scale be applied, though most researchers group the very fine and fine fractions with “medium”:
Wentworth Scale Values
- Very Fine- .0625 to .125mm
- Fine- .125 to .25mm
- Medium- .25 to .5mm
- Coarse- .5 to 1mm
- Very Coarse- 1 to 2mm
- Granule- 2 to 4mm
- Pebble- 4mm+
Anguarlity should also be observed, usually, with categories of rounded, subrounded, sub angular and angular documented.
Crushed Rock. Again, quartz is most common, but limestone or marl, felspar, soapstone, chert and other minerals are also seen. This is distinguished from angular and subangular sand by the acuteness of the breaks, and lack of any edge rounding from water action. Some refer to this as grit, but given the otherwise nebulous use of that term, crushed rock is preferred.
Grog This is all previously fired clay used as filler or temper. Identifiable crushed sherds are sometimes seen though more often the pieces look more like crushed rock and are angular. Potters often prepare grog specially, rather than grinding sherds, so if a distinction can be made, it should be noted. There seems to be purposely made grog in St Catherines ware, for instance, while its ancestor Wilmington ware frequently has identifiable sherd surfaces.
Clay This is mentioned separately because of its previous usage to describe all grog/sherd tempered pottery. Unfired clays mixed together could be considered tempering, according to Rice (1987), though unless the mixing was purposely incomplete, as seen in historic agateware for instance, it would be difficult to identify.
Fiber Early Stallings and related wares were shown to have been tempered with Spanish Moss. Although it seems logical that pine needles or other fiber could be used, only moss has been positively identified.
Shell Both freshwater and salt water shell were used for tempering in the Mid-Atlantic and elsewhere in the Southeast, but the practice was uncommon in South Carolina. The shell can be present, but it is often leached out, leaving flat or laminar voids. These are distinct from the blocky voids left by leached limestone.
Sponge Spicules These are natural inclusions. They are seen in clays from Florida and are seen rarely in South Carolina on the St. Johns ware brought in with the Spanish at Santa Elena.
Counts and weights work together on many levels, and it is suggested that both be used at the preliminary analysis level.
Why weigh sherds? First, one sherd weighing, say, 100 grams is obviously different from a sherd weighing .5 grams. Larger sherds can indicate less post depositional transformations (see Schiffer 1978) like trampling and other destructive processes. So if context A produces 100 sherds with an average weight of 100 grams, while other contexts produce 100 sherds with an average weight of .5 grams, further research is obviously called for. This puts us at what Stanley South has termed “The Why Threshold” (South 1977). Once a tendency such as this is documented archaeologically, we must move on to the explanatory phase.
The specifics of intensive analysis may vary to suit particular sites. For instance, at a Qualla site there is far more variation in rim treatment than at a Stallings site, so special attention is required. Thusfar the outline developed by Chris Espenshade and used by Brockington and Associates has been used, but intensive, attribute based analysis has also been used by David Anderson (1975, 1982, et al; John Cable (2002, 1998) and Sean Taylor (in Steen ed. 2002). Using both macroscopic and microscopic observation of a combination of nominal and ordinal variables, statistics are applied to identify meaningful clustering.
While it can be argued that going into such detail is overkill, especially given the natural composition of local clays, this approach is based in the scientific method, and should be superior to anecdotal / personal experience based analysis. At the very least anyone attempting to decide whether what John Cable calls “Refuge” is the same as what Larry Lepionka called “Refuge” should be able to do so without guessing. Of course, the results still have to be interpreted, which can lead to less than scientific conclusions.
Sherds vs Vessels
Intensive analysis is sometimes done at the sherd level and in others at the vessel level. At the sherd level the characteristics of each individual sherd are recorded. For picking up subtle, demonstrable differences this is the superior method, but the second step, identifying vessels, should not be omitted, as the number of individual vessels present gives an important clue to the density and duration of occupation.
Vessel based analysis groups all of the sherds that appear to come from a single vessel and proceeds from there with summary observations of many of the same attributes considered in a sherd level analysis. While the first approach gives more scientifically valid results, the latter is quicker and thus more cost effective.
In cases where the subtlest of variations might indicate a succession of occupations in a single village, or the introduction of new techniques into an existing community of practice analysis beginning at the sherd level might be preferable. But if a site has multiple components covering hundreds or thousands of years potentially, grouping macroscopically similar sherds into vessels and analyzing patterning might be a quicker way of analyzing site structure and occupation areas.
|Cable Variables||Anderson Equivalent||Brockington Equivalent|
|Vessel Form Variables|
|Ceramic Form||Vessel Form|
|Rim Type||Rim Form||Rim Shape|
|Lip Form||Lip Finish|
|Rim Orientation||Rim Orientation|
|Wall Thickness||Thickness||NE - Rim Thickness?|
|Exterior Surface Treatement Type||Exterior Surface Finish||Decoration|
|Exterior Surface Treatment Features|
|Interior Finish||Interior Surface Finish||Interior Surface Treatment|
|Dominant Sand Size Range||Primary Inclusions||Temper Size|
|Other Clastic Inclusions||Minority Inclusions||Minor Temper Size|
|Sand Lithology||Temper Type|
|Grog/Sherd (Clay) Temper Present/Absent|
|Clay Temper Size|
|Clay Temper Density|
|Culture History Variables|
Until the table above was composed I never noticed how different the approaches were. As one who experiments with pottery making Chris Espenshade is interested in different areas that have meaning to him. Though not fully reflected here, in more recent studies he has emphasized rim production steps, core color and firing atmosphere, for instance. David Anderson considered several additional variables in his discussion of the results, but not in all cases. Cable does not address temper / inclusion density in lithic tempered tempered wares here, though in other cases he does (Cable and Cantley 1998, for instance).
Vessel Form Variables from Cable (ed.) 2002
- Ceramic Form. This refers to the vessel segment- rim, body, base or “other.”
- Rim Type. This refers to the construction of the rim. Cable identified plain, notched, scalloped, thickened, and augmented rims in the 2002 collection. Further treatments are discussed below. Espenshade might consider “rim production stage” in this step of analysis.
- Lip Form. This is the shape of the lip, or the brink of the rim. Most rims are rounded, though in some case they are flattened, beveled one way or the other, tapered, or augmented by thickening.
- Rim Orientation. This is also referred to as neck orientation, as the rim is essentially the ultimate outcome of the neck. This indicates the overall vessel form. Straight rims suggests jars or large bowls. Excurvate or flaring rims are usually seen on vessels with constricted necks, while incurvate rims suggest globular jars.
- Rim decoration. This is a secondary treatment. One might include notching, scalloping, castellation,and thickening/augmentation under this category, though Cable only includes stamping, incising, punctation and “ticking” under this category. Rim strips, and other appliques like rosettes were not considered in the 2002 assemblage, though this would be the logical place to put them.
- Wall Thickness. This was measured 10cm below the rim for the sake of comparison.
- Sherd Size. John Cable uses size grades, with nested squares. The first is 2cm on a side, the second 4cm, and so on. Some researchers measure length and width and express size in that way.
- Exterior surface treatment type. Cable identified 48 surface treatment categories in this analysis, and this is by no means an exhaustive list of every possible combination. Broadly, surface treatments are variations on smoothing; stamping with carved paddles; and textile impressing. Cable includes what others might call decoration in this step of analysis, so many of his treatments are a combination: check stamped under incised, for instance.
- Exterior surface treatment features. These are the details of the treatment- incising depth, width and shape and so on.
- Motif Type. Broad categories- cord, fabric, etc.
- Motif Dimensions. Metric data on motifs- cord width, check size
- Interior Finish. In this analysis smoothing with the hands, tooling and floating were examined. Scraping and wiping are also noted, and the degree to which inclusions extrude from the surface is noted as well.
- Interior decoration. This is secondary treatment- textile impressing, stamping, incising, punctates would be included here.
- Interior contour. Cable says that only two states were observed- regular and irregular. In combination with the finish this is thought to give a general idea of vessel form and function.
- Dominant Sand Size Range. This does not distinguish between temper and natural sand, and may include combinations of different sized inclusions. It refers to the dominant size present, using the Wentworth Scale.
- Other Inclusions. This includes other things present in the paste, such as fiber and incidental pebbles or granules. Grog could be included here, he notes, but it is considered separately.
- Sand Lithology. The sand found in the sherds was mostly white/clear quartz, but also includes rose quartz, ferrous magnesium and an unidentified black material. He includes mica flecks under this category.
- Presence/Absence of Sherd/Grog Temper. The distinction between sherd and grog temper was not always made, but the presence of one or the other sets the sherds apart from predominantly lithic tempered wares.
- Clay Temper Size. These are small (>1mm), medium (1-2mm) and large (<2mm).
- Clay Temper Density. Low density is less than 5% of the paste, Medium, 5-15%, and high, 15% or greater.
- Paste Hardness. Hardness is a difficult variable to control because of the generally soft surfaces and variation present even on a single sherd. He does not feel that the Mohs Scale and scratch tests that many use are valid, and set up four general categories. These are soft, compact, hard, and vitrified. With the first two the edge of a sherd can be broken off and reduced to a powder, while hard pastes were difficult to break and could not be reduced. Vitrification is caused by very high firing temperatures and results in a glassy surface. This is uncommon on earthenwares and is usually a characteristic of stonewares and porcelain.
Culture History Variables
- Series. Series are overall groupings, like ware groups.
- Type. Types are internal variants- like Thoms Creek Plain et al
Suggested Attributes to be Used in Intensive Sherd Based Analysis
These were assembled by Sean Taylor and the author, primarily for use at the Kolb site (38DA75), but also to attempt to exhaustively list attributes observed by others as well for this web site. As stated before, not all are universally useful, and many have limitations.
- Excavation contexts are proveniences. There is no single way of splitting contexts. For some a one foot level in a ten foot square is sufficient control. For others the same square would be cut into a hundred proveniences. Each provenience has a number or designator, but not all produce artifacts. Generally artifacts from each individual provenience are placed in bags in the field, which are numbered to maintain inventory control. At the analysis level the bag from each provenience should be the base unit. Sherds within each bag should receive a catalog or specimen number. A Catalog # = bag# plus entry # ie, 100.1, 100.2, etc.
- Number of sherds - residuals and sherds that are EXACTLY alike can be grouped (mends, for instance), otherwise each sherd gets a catalog number.
- Weight in grams
- Sherd size in grades -
This is adopted from Ken Sassaman and refers to square centimeters. Increase grades for larger sherds. Use nested squares but downsize to account for irregular sherds. Use the size grade that the sherd would most nearly fill up if it was perfectly square.
- Sherd thickness, taken 3cm below the rim preferably. Note if this measurement is made. If a body sherd is regular (no more than 1mm variation), use one measurement. If irregular, but non-basal, take three and average. If basal take maximum thickness and describe in notes.
- Color, exterior/interior. Color is highly variable even on a single vessel because of firing conditions, use, and post depositional weathering. In many cases sherds of distinctly different colors will mend. A vessel may have areas of gray firing clouds on an otherwise oxidized surface. Although some pottery does indeed as a group tend to be lighter or darker, do not waste time with fine distinctions at the sherd level. Prudence Rice suggests using Munsell colors, which may be possible in some cases, but in general basic color catgories are acceptable.
- Dark, Reddish
- Dark, Brown
- Dark, Gray
- Medium, Reddish
- Medium, Brown
- Medium, Gray
- Light, Reddish
- Light, Brown
- Light, Gray
- Color, Profile
- homogenous “brown"
- homogenous “red"
- homogenous "gray/black"
- light exterior, dark core, dark interior
- light exterior, light core, dark interior
- light exterior, dark core, light interior
- dark exterior, light core, dark interior
- dark exterior, dark core, light interior
- Paste / Temper Inclusions (inclusions that appear to be purposely added)
1= "0.0-0.5mm fine sand (combines very fine, fine, medium using the Wentworth Scale). This might be viewed as “temperless.” Given the discussion of temper experiments above this is probably natural sand.
2= "0.5-1.0mm coarse sand"
3= "1.0-2.0mm very coarse sand"
5 "2.0-4.0mm granule. Sand this size and larger is called “grit” or “gravel” by some.
6 “4.0+ pebble. Note maximum size
7= "0.0-0.5mm fine crushed quartz. Crushed rock is distinguished by fresh, angular breaks.
8= "0.5-1.0mm coarse crushed quartz.
9= "1.0-2.0mm very coarse crushed quartz.
10= "1.0-2.0mm coarse crushed quartz.
11 "2.0-4.0mm granule size" crushed quartz .
12 “4.0+ pebble size crushed quartz. Note maximum size
13 "0.0-0.5mm fine grog" (This does not differentiate between sherds and intentionally fired grog. If sherds are clearly identifiable, note them)
14 "0.5-1.0mm coarse grog"
15 "1.0-2.0mm very coarse grog"
16 2.0-4.0mm granule size grog
17 4+ mm pebble size grog (note maximum)
18 limestone (note size). Limestone, marl and shell may be identified by the angular holes left behind when the calcium carbonate leached out.
19 "marl"(note size)
20 "shell"(note size)
21 “hole” (indet. leached out calcareous temper-note size and shell vs. lime/marl)
22 "granite"(note size)
23 “soapstone”(note size)
24 felspar (note size)
26. hornblende (note size)
27 “Other”- describe in notes, alter database structure to accommodate. There are many constituent minerals in a sherd which are not readily identifiable to the non-specialist.
10. “Background Sand” This is a judgement call. Inclusions that appear to be a natural clay element, often in association with another inclusion type (fine sand with grog, for instance) might be included. This may be a decision that is based on the pottery assemblage as a whole. If all of the pottery has a similar density of a given inclusion then it might very well be a natural constituent of the clay.
1= "0.0-0.5mm fine sand (combines very fine, fine, medium using the Wentworth Scale). See note above.
2= "0.5-1.0mm coarse sand"
3= "1.0-2.0mm very coarse sand"
4= "1.0-2.0mm coarse sand"
5 "2.0-4.0mm granule"
6 “4.0+ pebble Note maximum size
11. Inclusion (Temper) Grain Morphology
1 "angular" (but with rounded edges)
2 "crushed" (sharp breaks)
4 "sub angular"
5 "sub rounded"
12. Sand Grain Morphology (background)
2 "sub angular"
4 "sub rounded"
13. Inclusion Density (using Matthews charts)
1 "low" up to 5%
2 "medium" 5-10%
3 "high" 10-20%
4 “very high” 20-30%
5 “very, very high” 30% -40%
6 “extremely high” 40% +
14. Paste “Feel”. Though un-measurable, it is sometimes easy to sort sherds by feel. Joffrey Coe's description of Pee Dee ware as having a "sugary feel" will strike a chord with anyone who has worked at Town Creek and handled the pottery. Joe Herbert and John Cable's Hanover I and II types can usually be sorted by feel as well, with Hanover I having a sandy paste, and Hanover II a more silty, smooth feeling paste.
1 Fine - "chalky," "silty," or "waxy"
2 Moderate- a little "friable" or "sandy"
3 Coarse- very sandy and abrasive
15. Paste "Look" This is the appearance in cross section. Another hard to quantify attribute that should probably be defined by the analyst when used.
16. Paste Hardness. This is a variable that will be hard to quantify as well. John Cable notes that standard tools such as the Mohs Scale are ill suited for pottery because of its general softness to begin with, and exacerbated by firing differences and the uneven presence of aplastic inclusions in the body fabric. This variable generally results from firing conditions.
1 "friable" (soft and crumbly)
2 "compact" (moderately hard, difficult to break)
3 "vitrified" (very hard, with partially fused body fabric)
16. Sherd Breakage Appearance. This can be useful in assessing post depositional disturbance and site formational processes.
3 "extremely eroded"
17. Sherds- Surfaces
Exterior Surface Treatment- Primary
The surface treatment is the finish given while forming the vessel. It includes paddle marks, even if they are patterned and symbolically representative. Secondary, and tertiary treatments such as incising and punctating are often seen. These are considered "decorations." Expand as necessary.
17.1. “Plain” Plain surfaces do not have any visible markings. Variability in finishing, patterned burnishing, scraping, and brushing should be considered secondary decoration.
17.1.1 “plain - corrugated
1.2 “plain - roughened/rusticated
1.3 "plain"- lightly smoothed
1.4 “plain” - moderately smoothed
1.5 “plain” - well smoothed
1.6 "plain - burnished
1.7 “plain - patterned burnishing
1.8 "shell scraped"
17. 2. Incised (incising is a secondary treatment or “decoration”). This is sometimes split into "bold" and "fine," or some other qualifier to sort line weights.
2.1 straight lines
2.2 curvilinear lines
2.3 random lines
2.4 identifiable motif (describe in notes)
Qualifier: design motif
- lines with repeating semicircles
- lines with repeating squares
- ladder type separations (horizontal lines punctuated with vertical bars or lines
- triangles, usually filled with lines
- framing line (often used to contain geometric punctate or incising)
- representational images
17.3. Punctate (punctation is a secondary treatment or “decoration”)
- punctate, hollow round (reed), random
- punctate, hollow round (reed), patterned (describe in notes)
- punctate, hollow round drag-and-jab
- punctate, solid round, random (does not include “bundled sticks” ala Allendale Punctate)
- punctate, bundled sticks (“Allendale Punctate”)
- punctate, solid round, patterned (describe in notes)
- punctate, solid round, random, drag and jab
- punctate, hollow squared, random
- punctate, hollow squared, patterned (describe in notes)
- punctate, hollow squared, drag and jab
- punctate, solid squared, random
- punctate, solid squared, patterned (describe in notes)
- punctate, solid squared, drag and jab
- punctate, shell, periwinkle
- punctate, shell, cockle
- punctate, shell, angelwing
- punctate, other (describe)
17.4. Finger Marks
4.1 "finger impressed" (wide, shallow grooves)
4.2 "finger pinched"
4.3 "fingernail punctate/impression
17.5. Painted (not common in SC, but seen occasionally)
5.1 monochrome (describe, make categories if needed)
rim strip, solid
rim strip, segmented
rim strip with secondary decoration (note incised, ticked, punctate, pinched, etc., separately under secondary/tertiary decoration)
Paddle Stamped Decorations
17.7.1 Simple Stamped
7.1.1 U shaped groove
7.1.2 V shaped groove
7.1.3 squared groove (“thong wrapped paddle")
(orientation, spacing, overstamping are considered separate variables)
17.7.2 Check Stamp
Sometimes seen in combination with other treatments.
7.3 Complicated Stamp
These are the most common types, but it is recommended that careful attention be paid to designs that might be particular to the site or area. These are among the most temporally sensitive attributes.
Qualifier: add 1 for fine, 2 for medium, 3, for bold, 4, for “sloppy”, 5 for smoothed, etc.
7.3.1 rectilinear (general- fine rectilinear check)
7.3.2 curvilinear (general)
7.3.3 concentric (nested) circles
7.3.4 split circle
7.3.5 quartered circle
7.3.6 owl eye
7.3.7 filfot cross
7.3.8 figure 9 scroll
7.3.9 figure p scroll
7.3.12 arc angle
7.3.13 line block
7.3.15 split (barred) diamond
7.3.16 concentric (nested) diamonds
7.3.17 concentric (nested) triangles
7.3.19 wavy lines
7.3.20 zig-zag lines
7.3.21 Nested stars
Textile Marked Wares
17.8 Cord Marked
8.1 cord marked, fine (0-1mm)
8.2 cord marked, medium (1-2mm)
8.3 cord marked, heavy (gt 2mm)
8.4 cord marked, paddle edge (“cord and dowel”) marked, fine (0-1mm)
8.5 cord marked, paddle edge (“cord and dowel”) marked, medium (1-2mm)
8.6 cord marked, paddle edge (“cord and dowel”) marked, heavy (gt 2mm)
(orientation, overstamping, spacing, smoothing, secondary decorations, and metrics are separate variables)
Qualifier: Cordage Type
- "S twist"
- "Z twist"
- tight twist
- loose twist
17.9 Fabric Impressed
9.1 fine, regular fabric
9.2 fine, bunched fabric
9.3 coarse, regular fabric
9.4 coarse, bunched fabric
Qualifier: Fabric Type
- Flexible weave
- Rigid weave
17.10 Net Impressed
- clear net (ie, net squares visible)
- bunched net (often only random knotted segments or knot impressions can be seen)
17.11 Cob marked
This appears to have been made with a dried corn cob often, though it has been noted that very similar marks can be made with a pine cone (Herbert 2003). In some cases corn kernel impressions are visible.
17.12 Dentate Stamp
Usually small rectangular and shallow impressions made with a carved stick or roulette. Often only a single row is present, but occasionally two or more are seen.
18 Secondary Decoration
1. use appropriate designation from above, expand as necessary.
19 Tertiary Decoration
1. use appropriate designation from above, expand as necessary.
20. Incidental Marks
.1 fiber tracks on surface
.2 mat/vegetation impressions (usually on bases)
.3 tool marks
.4 shoulder smoothing
.5 other (describe)
21 Orientation of Decoration to Rim/Body
Measured from the rim, in compass degrees with 90 straight down. These describe the central tendency. That is, marks that are more or less straight up and down (80-90 degrees) can be considered perpendicular. For angles 0-45 degree and 45-90 degree increments to the right and left are used.
.1 parallel (>10 degrees)
.2. perpendicular (90)
.3 angle right 1 (0-45)
.4 angle right 2 (45-90)
.3 angle left 1 (90-135)
.4 angle left 2 (135-180)
.6 Combinations (describe)
Surface Treatment Metrics
Explain in notes. Common measures are:
for check stamping, size of squares and width of lands;
for simple stamping, width and depth of grooves and distance apart;
for cord, width and spacing;
for fabric, thread width, distance between rigid warp elements, size of same;
for incising, width, depth
for punctate, size and depth. This is partially described in the motif description, but should be described in the notes (ie, solid rectangle, 1mm x 2mm}
- overstamped, random
- overstamped, patterned
- paddle junctions visible
- smudged decoration
- smoothed over decoration
- well smoothed, nearly obliterated decoration
- sooting (for dating)
- organic buildup
- cut marks
- drilled hole
- coil break (no entry= no)
- (note interior or exterior)
Interior Surface Treatment
The main differences are hand vs tool finishing. Degrees of smoothing are nominal.
- moderately well smoothed
- well smoothed
- shell scraped
- scraped- indeterminate
- “floated”- erupting inclusions often coated with clay.
Decorated- incising, simple stamping, cord and fabric marking and even check stamping are occasionally seen.
The rim is the top of the vessel, but is sorted from the lip because other associated attributes can be observed, such as folding or augmenting, and decoration that make more sense considered as a suite.
Lip Form The lip is the peak of the vessel wall.
- tapered flat, to exterior
- tapered flat, to interior
- tapered rounded, to exterior
- tapered rounded, to interior
- beveled flat
- "folded exterior"
- "folded interior"
- "thickened exterior"
- "thickened interior"
- "thickened exterior and interior"
- "thinned exterior"
- "thinned interior"
- "thinned exterior and interior"
- "lipped exterior"
- "lipped interior"
- "strongly everted"
Use codes from above (ie, rim dec= 3.1- hollow reed punctate). Common decorations including stamping, notching, castellation, scalloping, pinching, incising, punctate, and appliques. On occasion the interior will be stamped or marked for a few centimeters from the lip.
- cazuela, (angled)
- rounded, incurved bowl type
- rounded, flaring, restricted neck jar type
- handle (describe in notes- loop, lug and strap handles are most common)
- feet / pods
Ceramic Vessel Forms
(For most sherds and vessels found archaeologically vessel forms are difficult to determine. As with some surface treatments one can often only identify the basic type: straight sided, unrestricted jars, for instance. if discernible. details and measurements should be recorded on a separate vessel form. No entry= unidentifiable)
- bowl, restricted rim
- bowl, open
- bowl, everted rim (brimmed)
- bowl, cazuela (carinated)
- Jar, straight sides, unrestricted rim
- Jar, restricted rim (incurvate)
- jar, restricted neck
- jar, restricted neck, everted rim
- jar, collared
Use wear / Alteration
- Scraping (interior)
- “Mend” Holes
Other Ceramic Items
- "pipe" - whole clay pipes
- "pipe bowl" - bowl or fragment thereof
- "pipe stem" - the non-bowl element
- "baked clay object" - irregular lumps of clay, sometimes with holes through the center, and sometimes decorated.
- ear spool