n o. 3
PREVIOUSLY PUBLISHED IN CERAMIC NOTES:
CERAMIC NOTES no. 1.
Annotated Bibliography of Ceramic Studies, Part I. Analysis:
Technical and Ethnographic Approaches to Pottery Production and Use. Compiled by Prudence M. Rice and Marian E. Saffer
OUT OF PRINT
CERAMIC NOTES no. 2.
Ceramic Technology at a Weeden Island Period Archaeological
Site in North Florida. Ann S. Cordell t1984).
The James A. Ford
Library of Anthropology
Florida Museum of Natural History,
Gift of.:" m_ ti C 71,
CERAMIC NOTES: OCCASIONAL PUBLICATIONS OF THE CERAMIC TECHNOLOGY LABORATORY, FLORIDA STATE MUSEUM, is a publication series for articles and monographs dealing with anthropological and archaeological studies of pottery and related materials. The series is edited by Prudence M. Rice, with assistance for production of issue number 3 provided by Jenette Savell, Cindy Cart, and Lorena Cantu. Address all correspondence to P.M. Rice, Florida State Museum, Gainesville, Florida 32611.
Checks for CERAMIC NOTES no. 3 should be made out in the amount of $8.00 US payable to UNIVERSITY OF FLORIDA FOUNDATION. Postage is included.
Copyright ( 1986 by the Florida State Museum of the University of Florida.
CERAMIC NOTES NO. 3
PAPERS IN CERAMIC ANALYSIS
EDITED BY PRUDENCE M. RICE
TABLE OF CONTENTS
The Technology of Ceramic Production of Wanka and Inka Wares from the Yanamarca Valley, Peru Melissa B. Hagstrum 1
A Formal and Functional Analysis of St. Johns Series Pottery from Two Sites in St. Augustine, Florida Mary K. Herron 31
The Southeastern Fiber-Tempered Ceramic Tradition Reconsidered
George Ward Shannon, Jr. 47
Analysis of Ceramics from the South Prong I Site (8-HI-418), Hillsborough County, Florida Jeffrey M. Mitchem 81
Extraction and Thermal Alteration of Pollen Embedded in Clay
Donna L. Ruhi 111
A Preliminary Report on Investigations of Sponge Spicules in Florida "Chalky" Paste Pottery Nina Thanz Borremans and Graig D. Shaak 125 GOWER: A Program to Calculate Similarity from Mixed-Level Data
Marian E. Saffer, Roger K. Biashfield, and Prudence M. Rice 133 Pottery Manufacture and Design Symbolism of Late Swift Creek Phase Ceramics at Kings Bay, Georgia Rebecca Saunders 145
The Technology of Ceramic Production
of Wanka and Inka Wares
from the Yanamarca Valley, Peru
Melissa B. Hagstrum
Studies of prehistoric ceramics focusing on technological aspects of ceramic manufacture contribute to an understanding of productive specialization in prehistoric society (cf. Arnold 1975; Rice 1981; Shepard 1942; Van der Leeuw 1976). This study endeavors to show how a technical understanding of ceramic wares can aid in determining the economic organization of craft production. In particular, this study aims to discern differences in the manufacturing technique of wares produced for general household consumption and for state uses. To furnish evidence for determining the form of craft specialization, technological studies require an understanding of the ceramic production sequence: vessel forming, finishing, and firing.
In this study, variables related to the production sequence were measured to assess both standardization and labor investment in the manufacture of ceramic wares. Differences in standardization and labor investment in manufacture point to differences in the organization of production. Specifically, measures of standardization and labor investment provide information about the form of ceramic production, whether competitively based, geared to extra-community demand, or non-competitively based, geared to intra-community demand. These measures also contribute to an understanding of the relationship of producer to consumer. This research focuses on two forms of specialist production: general craft specialists who are independent self-supporting producers; and attached craft specialists who are supported in a contractual relationship by an elite or a state organization (Earle 1981:231).
Three wares from the Upper Mantaro region of Peru (see Fig. 1), dating to the late prehispanic period (A.D. 1000-1533), were studied. They include two local Wanka wares, Micaceous Self-Slip and Base Clara, which were produced throughout this period, and the state Inka ware, which was produced from about A.D. 1460 until the Spanish conquest. The measures of standardization and labor investment suggest that Wanka Micaceous Self-Slip was produced by general specialists for extra-community
Fig. 1. The Upper Mantaro region of Peru. The elevation of this region ranges from 3500 m. to 400. m.; the dashes topographic line at 3500 m. designates the Yanamarca Valley. Dots represent archaeological sites. (Redrawn from Earle et al. 1980:6).
consumption where production was competitive; that Wanka Base Clara was produced by general specialists for intra-community consumption were production was non-competitive, marked by a close contractual relationship between producer and consumer; and that the state Inka ware was produced by attached specialists for the elite and state apparatus in a close contractual relationship, where production was non-competitive.
PREHISTORY OF THE YANAMARCA VALLEY
The prehistoric chronology of the Yanamarca Valley for the 500 years prior to the Spanish conquest provides a unique setting for studying developments in the organization of ceramic specialist production. This period marked a transition from simple, socially undifferentiated communities (Wanka I phase) to increasingly complex, hierarchically organized polities (Wanka II phase), and finally to rapid incorporation into the Inka state organization (Wanka III phase) (Earle et al. 1980:3-5). These events are significant for economic institutions generally and for ceramic craft specialization specifically.
Wanka I (A.D. 1000-1250). During the Wanka I phase the population was dispersed among several small communities averaging approximately 2.5 ha. in size. These communities did not vary substantially in size, and there is little evidence for special civic and ceremonial architecture indicating social differentiation and regional organization (Earle et al. 1980:20-21).
Wanka II (A.D. 1250-1460). During the Wank4 II phase there was a
dramatic local increase in population which became concentrated in fewer, larger communities. These communities ranged from five to nearly 100 ha. in size. As the average community was 28.7 ha., this represents more tnan a tenfold increase in community size from Wanka I. There was a marked concern with security, demonstrated by the shift to large communities situated in defensible locations (Earle et al. 1980:21-23). The increase in the Wanka population during the Wanka I and II phases may have led to increased competition for resources and incipient social differentiation because effective leadership in warfare and alliance formation would have been prompted by competitive conditions (LeVine 1979).
Wanka III (A.D. 1460-1533). The Wanka III phase began when the Inka army conquered the Mantaro region. This phase was characterized by a dual settlement system in the form of (1) an imposed state organization controlling the Mantaro region and linking the area to the Inka empire, and (2) a local settlement pattern affected by the state's presence yet engaged primarily in community activities (Earle et al. 1980:42).
Standardization and Labor Investment in Pottery Manufacture
Assessment of the level of standardization and of labor investment in pottery manufacture can be used to reconstruct the organization of craft production in a prehistoric society. The level of standardization of
ceramic wares is one indication of the level of specialization in ceramic production. The level of labor investment in ceramic wares indicates whether the specialist production is competitive or non-competitive (see Table 1).
Table 1. Standardization and Labor Investment in Pottery Manufacture.
4JAttached specialist General specialist
S (High) non-competitive non-competitive
(Low) General specialist
If a ware shows standardization and cost-efficiency in production,
then it was produced by general specialists under competitive conditions for extra-community demand. In this case large-scale production dictates standardization in manufacture and production costs dictate labor efficiency. The relationship between producer and consumer lacks common ties of either kin or community.
If a ware does not show standardization but does show labor intensity in production, then it was produced by general specialists under non-competitive conditions for intra-community demand. On this domestic level, the general specialist produces ceramics for family members and perhaps for a small outside group, likely to be kin, as well. The relationship of producer to consumer is a close one.
If a ware shows both standardization and labor intensity, then it was produced by attached specialists under non-competitive conditions for single group, a state and/or an elite. The producer-consumer relationsnip is a close contractual one.
The Production Sequence
The production sequence represents the various operations the potter performs in transforming raw clay into completed vessels. Studying these operations provides an objective means for determining levels of standardization, labor investment, and skill in manufacture (Feinman et al. 1981:871-884). The production sequence includes vessel forming, vessel surface finishing, and vessel firing (Rye 1980:3-4).
Vessel forming refers to the many techniques used for deforming the plastic clay to render the pot's basic shape. These techniques leave characteristic markings in the clay, and if not obliterated by subsequent surface finishing procedures, they are diagnostic attributes containing information about the level of sophistication and the scale of production. For example, wheel-thrown and mold-made pottery have distinctive markings, and these methods of manufacture signal potential specialist production because of their labor efficiency and capability for standardizing vessel form.
Surface finishing treatment is the most variable and the most
diagnostic procedure in the production sequence, and surface treatment is consequently the major attribute used by archaeologists to define ceramic types. There are two basic categories of finishing techniques, modifications of the vessel surface (e.g., scraping and polishing) and applications to the vessel surface (e.g., slipping and painting). Because a single pot may receive single or multiple surface treatments, the amount of labor invested in this stage of production is highly variable. For example, a vessel which receives multiple finishing treatments--slipping, burnishing, and painting--is more labor intensive than a vessel w4hicn receives a single finishing treatment--scraping or smoothing. N'ot all surface finishing techniques may be observable on the completed vessel; later steps may obliterate earlier ones. For example, slipping or polishing may obliterate earlier scraping and wiping.
Vessel firing, like surface finishing, is also a variable procedure. Of the many variables related to the firing procedure, the most critical are the physical and chemical properties of the clay and the firing temperature and atmosphere achieved. The properties of a clay body are assessed by the potter in order to fire the clay optimally within (1 ) the proper temperature range to avoid under- or over-firing and (2) tile proper firing atmosphere to realize the desired coloration. The firing atmosphere, whether fully oxidizing, fully reducing, or incompletely oxidizing, may be inferred from the colors on the surface of a vessel (for example, reds signal an oxidizing atmosphere in iron-rich clays) and in the paste of a vessel (for example, the presence or absence of a black or gray core) (Shepard 1980:106-107; Rye 1980:114-118). Ordinarily, attaining a purely oxidizing atmosphere requires more control than does attaining either a reducing or an incompletely oxidizing atmosphere (see vessel firing section.) Thus, the atmosphere achieved in firing ceramic wares contains some information about the producer's knowledge of the clays used and skill in controlling the firing event.
THE PRODUCTION SEQUENCE OF WANCA AND INKA WARES
The Ceramic Sample
Wanka and Inka wares are particularly well suited to the aims of this research. From the Wanka I to II phases the increasing scale of political units, marked by larger settlements and their organization into larger
polities, favored the development of independent specialist production in pottery manufacture. During the Inka period (Wanka III to Contact), the centralized political bureaucracy favored development of attached specialist production in pottery manufacture. Of the several ceramic types found in the Yanamarca Valley, jar rim sherds of Wanka Micaceous Self-Slip, Wanka Base Clara, and Inka Wares define the study sample. Micaceous Self-Slip (97 sherds) and Base Clara (64 sherds) come from Tunanmarca, a Wanka II phase settlement, and from Marca, a Wanka III phase settlement. Inka ceramics (38 sherds) come from Marca only (see Fig. 1 for site locations). The sherd sample for this study was collected during the initial stage of the Upper Mantaro Archaeological Research Project excavations in 1982.
Wanka I and II Phase Ceramics. Micaceous Self-Slip is a plainware
type produced during the Wanka I and II phases. In the early phase, the vessel form was a tall-neck jar which was replaced in the later phases by a low-neck jar (Earle et al. 1980:18). The Micaceous Self-Slip jars examined in this study were formed by coiling. Subsequent scraping with a gourd or similar tool refined the pot's shape by obliterating the joints between the individual coils. Wiping the surface of the vessel, wnicil was still wet enough to be malleable, with a damp piece of leather or cloth (cf. Donnan 1965:124) was likely to have been the final step in shaping the pot. By exerting subtle pressure with the fingers over the leather or cloth, the pot's lip was given its characteristic form because the clay was still wet. Wiping also smoothed the surface of the vessel by removing the drag marks caused by scraping, giving the surface a uniform finish. The paste of these Micaceous Self-Slip vessels was generally reddish yellow, indicating that they were probably fired in an oxidizing atmosphere.
Base Clara is a painted ware, a local development of the earlier
Huacrapuquio ceramic type. In the Wanka I phase, the jar and bowl forms exhibited a blue-purple paint applied to a background color which varied from light orange to beige or cream; in the Wanka II phase, the paint color changed to a dark-gray or gray-brown color (Earle et al. 1980:15), resulting from a change in paint color rather than in firing conditions since the paste color remained the same (Earle 1983, personal communication). The three varieties of Base Clara, "unslipped," "slipped," and "cream slip," exhibit a production sequence similar to that of Micaceous Self-Slip, differing only with the final addition of paint and/or slip. Unslipped Base Clara jars were ordinarily painted with a thin wash-like paint, applied with a broad brush; slipped jars were slipped and painted; and cream slipped jars were slipped but not painted.1 In all cases, the slip was merely brushed on with no additional burnishing or polishing. The paste of these Base Clara vessels was also generally reddish yellow with evidence of grey firing cores indicating that the ware was probably fired in an atmosphere which did not achieve comDlete oxidation.
Wanka III Phase Ceramics. The Micaceous Self-Slip and Base Clara
wares of the Wanka II phase continued to be produced during the Wanka III
phase without modification. State Inka ceramics were, however, added to the assemblages of the local settlements. The Inkaic vessel forms were more numerous and more varied than were those of the Wanka wares: aryballoid jars, open-mouthed jars, pedestal-based cooking allas, and plates were produced (Earle et al. 1980:24). The Inka aryballoid jars examined in this study were also formed by coiling and subsequent scraping and wiping. The vessels were slipped and highly burnished or polished. The final step was the addition of painted decoration; on the sherds examined it was a line ringing the lip edge. The paste of these Inka vessels was generally reddish yellow, similar to Micaceous Self-Slip although more clear and red, indicating that the ware was probably fired in an oxidizing atmosphere.
The production sequence of the Wanka and Inka wares is the focus of the remaining sections of this paper. Vessel forming, finishing, and firing will be discussed for each ware in terms of standardization and labor investments in manufacture.
Vessel formation refers to the process of molding clay into a desired shape. This process can be studied by vessel surface attributes and by determining the sequence of their execution. Examples of attributes are surface marks, fractures, or indentations. Because clay proceeds from plastic to non-plastic during the forming stage, attributes made at the plastic stage were executed earlier than those made at the leather-hard or dry stages. Furthermore, attributes produced early in the forming procedure may be obliterated by those produced at a later stage. Surface
evidence of the early stages of vessel production varies with vessel shape. For example, the interiors of narrow-necked jars and bottles are inaccessible when the shape is completed, and nence they often preserve evidence of earlier forming stages. By contrast, because both the interior and exterior of bowls and plates are refined for aesthetic and functional reasons, they do not reveal evidence of earlier forming stages. In this study, jars were selected for analysis to see as much as possible of the forming processes.
Micaceous Self-Slip. The Micaceous Self-Slip vessels appear to have
been carefully and con-sistently formed. Although coil ridges were felt on some sherds, the coils were generally obliterated by subsequent scraping and wiping. For a given sherd the vessel wall was maintained at a consistent thickness over its surface. Even the interiors of jars appear to have been carefully smoothed. The vessel rims were consistently formed, as will be discussed below. Details of the vessel rim form, constructed by finger pressure during the wiping procedure, show dexterity in execution.
The forming process determines vessel shape, and standardization in forming procedures results in standardization of vessel shape. The variables defined for assessing standardization in Micaceous Self-Slip jar formation consisted of attributes related to vessel form: rim and collar diameter, neck height, wall thickness, and neck form (Table 2).
Table 2. Attributes of Wanka Vessel Form.ab
Cm. MSS % BC
Small 10 15 28 13
Large 16 36 72 87
Small 3 12 38 46
Large 13 29 62 54
Short .6 1.8 23 36
Tall 1.9 10.3 77 64
Thin .3 .6 30 10
Thick .8 1.1 70 90
B 96 28
A 4 12
aFor the categories of rim diameter, wall thickness, and lip detail, the Micaceous Self-Slip sample = 89 sherds, and the Base Clara sample = 40 sherds; for the categoreis of collar diameter and neck height, the Base
Clara sample = 28 sherds.
bThe ranges given in this table were determined from scatter plots. See Figures 2 and 3.
cMeasured 1.5 cm below lip edge. dSee Figure 5.
The histograms drawn for rim and collar diameter and neck height show that there is a bimodal distribution emerging in these data; that is, small and large vessels had been manufactured (see Fig. 2).
Rim diameter was plotted against collar diameter in an effort to
assess standardization in vessel form of the Micaceous Self-Slip sherds. Figure 3 (top) shows a correlation between the sizes of rim and collar diameter for Micaceous Self-Slip jars, with a coefficient of .78. In terms of standardization of vessel form for Micaceous Self-Slip then, the sizes of rim and collar diamter are in general closely correlated. Assessing standardization in vessel form is a complicated problem, particularly when the investigator must rely solely on the information derived from potsherds.
10 20 30cm
10 20 cm
Fig. 2. Micaceous Self-Slip form attributes.
1 2 3 4cm
* .n .
10 15 20 25 cm
Fig. 3. Rim diameter plotted as a function of collar diameter. For Micaceous Self-Slip the linear regression is: Collar + 1.794 + .639 (Rim), with the correlation coefficient r = .78. For Base Clara the linear regresion is: Collar = 3.58 + .570 (Rim), with the correlation coefficient r = .56.
10 15 20 25 cm
The labor invested in the forming procedure can also be discerned from attributes of vessel shape. For this study, labor investment in the forming stage was measured by sherd neck form. The individual Wanka neck forms, defined by LeBlanc (n.d.), have been ordered hierarchically in terms of the amount of labor required for their construction (Fig. 4). LeBlanc grouped the rim of Wanka vessels into four neck form categories:2 B, A, N, and C. In general, B is less labor intensive than A, A is less intensive than N, and N is less labor intensive than C. Ordered in relation to labor investment, B ;; A > N > C.
Fig. 4. Base Clara form attributes.
Form C requires care in construction to achieve the continuously curving line of the neck, whereas form N's directly outflaring neck requires less special handling. To form a continuously curving shape, the clay must be added in gradual stages so that the additional weight of the wet coil is supported by the drier and stiffer coils beneath. To form a directly outflaring shape, the clay need not be added in such small increments, since it is not required to "hold" a curve and is therefore more rapidly constructed. For the same reasons, form A requires more care in construction than does form B.
Micaceous Self-Slip jars were constructed with neck forms requiring minimum labor investment. The B forms are represented in 95% of the
sample, the A form is represented in only 4% of the sample, and the N and C forms are not represented. In summary, from the rim sherds examined, it appears that the Micaceous Self-Slip ware had a standardized vessel form that was constructed with little labor investment.
Base Clara. The Base Clara jars were formed by coiling as could De seen on the interior of those sherds with collar constriction and intact neck-to-shoulder joint. Very often the coils were left unsmoothed in this transition zone, because they did not show on the completed vessel and since scraping and wiping would have been difficult in this region of the vessel. Nevertheless, carefully constructed jars of any type do show smoothing of the transition zone from neck to shoulder. The lack of attention to coil smoothing, together with other characteristics of Base Clara construction, suggest that these vessels were produced with little attention to detail. The overall impression is that the ware had been constructed with haste and at a relatively low level of technical skill.
The variables defined for assessing standardization in Base Clara jar
formation are attributes of form: rim and collar diameter, neck height, wall thickness, and neck form (Table 2 shows the attribute frequencies, and Fig. 2 shows the sample distribution for the attributes measured). The histogras for rim and collar diameter and neck height show an emerging bimodal distribution in the data (Fig. 5). Thus, like Micaceous Self-Slip, small and large vessels were constructed.
I M! --
10 20 cm
-5 Collar diameter
10 20 30 Cm
5 Neck height
I n P 0 q R n F1 5 10 cm
Fig. 5. Wanka neck forms. Arranged hierarchically, from low to high labor
investment, left to right.
To assess standardization in vessel forming, rim diameter was plotted against collar diameter. Figure 3 (bottom) shows that there is some correlation between the sizes of rim and collar diameter for Base Clara jars, with a coefficient of .56. Thus, Base Clara is a less standardized ware than is Micaceous Self-Slip. First, the correlation coefficient for rim to collar diameter is lower (.56 versus .78). Second, 60% of the sample sherds are accounted for by two neck forms, N and C (versus 100% in two categories. )
In terms of labor investment, note that forms B and A (representing 40% of the sample) require lower labor investment and are more standardized than forms N and C (representing 60%), which require higher labor investment (see Table 2). In summary, this study suggests that Base Clara was less standardized in form but more labor intensive in construction than was Micaceous Self-Slip.
Inka. That Inka jars were formed by coiling is evident by remnant
coilF-T7hges felt and seen on some sherds. The formation of this ware was observed to have been consistently highly skilled with great attention to detail. The possibility that the Inka jars were formed by using molds was discounted because no traces of mold seams running either perpendicular or horizontal to the vessel rim were encountered on the sample sherds (cf. Donnan 1965:118-119; Rye 1980:81).
The Inka sherds were sufficiently small in size so that assessing
standardization of vessel form was precluded for this study: because the sherds were fragments which contained no information on collar diameter or neck height, a quantitative assessment of standardization in vessel form could not be made. The impression gained from examining the rim sherds, however, was that the Inka ware was highly standardized in form. The scatter plot of Inka rim diameter (Fig. 6) shows a wide range of vessel opening sizes, from 12 cm. to 44 cm. Examination of these rim sherds suggests that Inka vessel form is probably standardized for all values of opening rim diameter, that is, the same vessel shape was produced in various sizes. The lip detail frequencies (Table 3) show that d majority of the sherds examined (63%) fall into two categories, lip details 1-3 and J-1.
E 20 30 40 CM
Fig. 6. Inka rim diameter.
Table 3. Attributes of Inka Vessel Form.ab
In cm %
Small 12 20 37
Medium 21 28 35
Large 29 44 24
Thin .7 .9 63
Thick 1.0 1.6 37
Rim Form/Lip Detaild
aThe Inka sample = 38 sherds.
bThe ranges given in this table were determined from scatter plots. See Figure 6.
CMeasured 1.5 cm below lip edge.
dSee Figure 7.
For this analysis, the individual Inka rim forms and lip details have been classified3 and ordered hierarchically in terms of labor investment in construction (see Fig. 7). The rims were grouped by rim form (designated by the letters "I" and "J") and lip detail (designated numerically by 1-3). The Inka rim form J is more labor intensive than is rim form I for the same reasons that Wanka neck form C is more labor intensive than is neck form N: form J required care in construction to create the continuously curving line of the rim, whereas form I's directly outflaring rim required less special handling. The hierarchical ordering of Inka lip details shows that in general the higher the numerical designation of the lip detail, the more labor intensive was its construction technique. Although the fine Inka surface finishing technique largely obliterated evidence of Inka vessel formation, the Inka lip details were probably formed by a finger pressure technique. Therefore, the rounded lip detail of form J-1 was created by exerting a simple even pressure on the interior and exterior surfaces of the rim while turning the vessel; the blunt lip detail of forms 1-2 and J-2 was also created by exerting a simple even pressure on both rim surfaces while simultaneously running a finger atop the lip; and the angular lip detail of forms 1-3 and J-3 was created by exerting more subtle opposing pressures on the rim surfaces, causing one edge to flare out and the other to tuck under. For the Inka sherds, the rim form/lip detail frequencies show that 74% of the vessels exhibited the more labor intensive rim form J., and 55% exhibited the more labor intensive lip details 2 and 3 (see Table 3). In summary, a high level of standardization in vessel form and a high level of labor investment in construction is suspected for the
manufacture of Inka jars.
Fig. 7. Inka rim forms and lip details. Arranged hierarchically, from low to high labor investment, left to right.
Vessel finishing changes the texture and enhances the aesthetic
character of the vessel. The two categories of finishing techniques are modifications of and applications to the vessel surface. Modifications include scraping, smoothing, polishing, burnishing, incising, impressing,
and carving; applications include the employment of different materials such as pigments, slips, glazes, and organic coatings. These decorative techniques are closely related to the forming procedure because they are linked directly to the varying properties of clay bodies as they dry. In general, modifications of the surface must be executed when the clay is wet or leather-hard, and applications may be added when the clay is leather-hard or dry. Finishing techniques are readily observed on ceramic artifacts unless preservation is poor. To focus on standardization and labor investment in Wanka and Inka finishing technique, the production step measure described by Feinman et al. (1981:871-884) is employed here to quantify vessel finishing procedure for the study wares.
The production step measure, used to determine labor investment in
ceramic manufacture, is an ordinal scale index of production costs which focuses on the number of steps required to produce a given ware and which enables comparisons of the relative labor costs to produce different wares. The measure indicates approximate production complexity.
The basis of this measure is a series of ethnographic studies which described the requisite tasks for manufacturing various nonwheel-made ceramic wares (see Feinman et al. 1981:872). According to the method, one
point is tabulated for each step in the manufacturing process with the exception of actual vessel shaping, because all vessels receive this production step. The method focuses primarily on vessel finishing techniques (applique, surface finish, and decoration) because these processes are readily observed on archaeological sherd samples (Feinman et al. 1981:873-874.)
The frequencies tabulated for Wanka and Inka surface finishing technique show surface modifications to include scraping, wiping, burnishing, and polishing, and surface applications to include slipping and painting. Following Feinman et al. (1981:873), each of these steps is accorded one point (see Table 4).
Table 4. Score Assignments for Vessel Finishing Techniques.
Scraping and/or wiping
However, two issues relating to score assignments are noted. First, the production step of scraping is given one point, as are burnishing, slipping, or painting, for example. Scraping may be understood to o~e Jot1 a forming and a finishing technique; it is the step that thins tne walls
and smooths the coils. Thus scraping is a very labor intensive technique. The other techniques (e.g., burnishing, slipping, etc.) are less labor intensive in terms of actual time and attention involved, but they are labor intensive per pot because they are additional treatments, and because they cover up evidence of scraping (Rye 1980:62,86). Scraping is accorded a single point in this study, however, because on some vessels scraping was the only surface treatment, and no subsequent techniques were employed to obliterate the scraping procedure.
Second, points are not accrued for materials procurement (Feinman et al. 1981 :873) because it is rarely known where the inhabitants of prehistoric settlements obtained the raw materials (clay, temper, pigments) necessary for ceramic manufacture. This limitation in the method is one that could be eliminated with ethnographic study so that, for example, the additional costs of slips or pigments could be factored into the measure. The production step measure, though conservative in its score assignments for particular tasks, is a valuable comparative tool.
In this analysis, levels of standardization and labor investment are
determined for each ware (see Table 5).
Table 5. Vessel Finishing Treatment Scores.
Treatment Score % N
5.1 Micaceous Self-Slipa
1. Scraping/Wiping 1 94 91
2. Scraping/Burnishing 2 1 1
3. Wiping/Burnishing 2 3 3
4. Scraping/Wiping/Burnishing 2 2 2
5.2 Base Clarab
1. Slipping only 1 6 4
2. Scraping/Wiping/Slipping 2 47 30
3. Wiping/Painting 2 16 10
4. Scraping/Wiping/Painting 2 3 2
5. Wiping/Slipping/Painting 3 19 12
6. Scraping/Wiping/Slipping/Painting3 9 6
1. Wiping only 1 3 1
2. Burnishing only 1 8 3
3. Wiping/Slipping 2 3 1
4. Slipping/Burnishing 2 12 5
5. Slipping/Polishing 2 2 1
6. Burnishing/Painting 2 13 5
7. Slipping/Burnishing/Painting 3 37 14
8. Slipping/Polishing/Painting 3 21 8
aMicaceous Self-Slip sample = 97 sherds
bBase Clara sample = 64 sherds
Clnka sample = 38 sherds
Standardization in vessel finishing is assessed by the frequencies tabulated for each finishing sequence; labor investment is assessed by the median and mode of the assigned scores. A discussion of the finishing procedures for each ware is presented below, followed by assessments of their standardization and labor investment.
Micaceous Self-Slip. The surface finishing procedures exhibited on the Micaceous Self-Slip jar rim sherds include a combination of scraping,
wiping, and burnishing. Because no applications, slip or paint, were observed on the study sample, this ware is considered to be an "undecorated" ware; however, fugitive paint has been observed on a few examples.
Table 5.1 shows the score assignment and the frequencies tabulated for each finishing sequence. These frequencies suggest standardization in Micaceous Self-Slip finishing treatment. Note that 94% of the sherds fall into a single finishing treatment category, that of scraping and/or wiping
(treatment #1 ).
Table 5.1 also shows the median and mode of the Micaceous Self-Slip finishing treatment scores to be one point, suggesting that this ware is not labor intensive in finishing treatment. It is important to bear in mind that an undecorated ware employs, by definition, fewer operations in finishing than does a decorated ware. Therefore, manufacture is more efficient because the forming and finishing procedures are executed virtually at the same stage in the production sequence; in the case of Micaceous Self-Slip this operation is observed as surface scraping. F )r a decorated ware, by contrast, vessel finishing is a distinct stage ii toe production sequence, because other materials are applied and modifications are made to the vessel surface at various stages of drying.
Base Clara. The surface finishing procedures exhibited on the Base Clara jar rim sherds include a combination of scraping, wiping, slipping, and painting. Because surface applications, slipping and painting, were employed, this is a decorated ware.
Table 5.2 shows the score assignments and the frequencies tabulated for each Base Clara finishing sequence. The various surface treatments 00-6)
correspond to the defined subtypes, "cream slip" (#1,2), "unslipped" (#3,4), and "slipped" (#5,6). The frequencies suggest that by comparison to Micaceous Self-Slip, Base Clara was less standardized. Note that each of the finishing treatment categories has more equal representation for Base Clara than for Micaceous Self-Slip: 91% of the Base Clara sherds
fall into four of six categories (#2,3,4, and 6) whereas 94% of the Micaceous Self-Slip sherds fall into one of four categories. Note that this comparison is made between an undecorated and a decorated ware.
Table 5.2 also shows the median and mode of Base Clara finishing treatment scores to be two points, suggesting that by comparison to Micaceous Self-Slip (median, mode = one point), this ware is more labor intensive in finishing treatment, and the various surface treatments for Base Clara employ a greater variety of production steps, including slipping and painting.
Despite this increase in labor investment with respect to Micaceous Self-Slip, the Base Clara finishing treatment gives the impression of having been rapidly and carelessly executed, without attention to detail. For example, the Base Clara slips exhibit technically inferior characteristics of adherence to the vessel, a phenomenon which may have
been remedied by care in slip application. In fact, the Base Clara slips generally have worn off and are poorly preserved archaeologically.
Slip is a fluid suspension of clay in water, and therefore it snares many properties with clay. Thus, good adherence of slip to the clay body of the parent vessel is achieved by (1) use of a slip clay similar in shrinkage and thermal properties to the clay body used in constructing the vessel (cf. Shepard 1980:67; Rhodes 1973:250); (2) care in treatment of the vessel surface prior to slip application as well as post-application, in terms of burnishing (Shepard 1980:67); and (3); attention to the dryness of the vessel at the time of slip application, because slip applied to a dry vessel will separate from the surface, but applied to a sufficiently moist surface will minimize differential shrinkage (cf. Rhodes 1973:250; Rye 1980:141).
The poor adherence of the Base Clara slips may have resulted from properties of the vessel surface prior to slip application. Three possible explanations of Base Clara slips are advanced. First, for the Base Clara jars, it appears that the surface treatment was not suited for slipping. The vessel surface was ordinarily cloth wiped after having been scraped. Because both the cloth and the vessel surface had been wet, the slipping operation may have left inclusions in the paste exposed at the surface. The resulting roughness may have caused the slip to form pinholes, an effect caused by tiny air bubbles which formed as the slip was applied to the rough surface. Alternatively, these pinholes may have resulted from the expansion of inclusions in the clay body (probably limestone, Costin, personal communication) during or after firing. Finally, the lack of burnishing is an additional reason for poor slip adherence on Base Clara vessels because burnishing is what usually makes a slip adhere well. That the slip applied to Base Clara vessels was not
burnished is further evidence of the careless execution of this ware, and as noted earlier, burnishing is a labor intensive procedure per pot.
An additional point to be made about Base Clara finishing treatment is that in the painted varieties, the painted decoration is without exception carelessly executed. Paint was applied with a broad brush in bold designs betraying haste in application. As might be expected, much of the painting has worn off with the slip.
In summary, although the Base Clara vessels show increased labor
investment in finishing treatment over Micaceous Self-Slip, the procedures were carelessly and rapidly executed: specifically, the poor adherence of the slips supports this observation as does the careless painted decoration.
Inka. The surface finishing procedures exhibited on the Inka jar rim sherds include a combination of wiping, burnishing, polishing, slipping, and painting. Because this ware was slipped and painted in addition to being burnished and polished, Inka vessels were a highly refined and
carefully executed ware.
Table 5.3 shows the score assignment and the frequencies tabulated for each finishing sequence. The frequencies show that 92% of the sherds examined fall into five of eight categories (#3,4,6,7, and 8). By comparison to Base Clara (91% in four of six categories), the Inka finishing procedures are relatively more standardized.
Table 5.3 also shows the median and mode of Inka finishing treatment
scores to be three points, suggesting that by comparison to Base Clara (median, mode = two points), this ware is relatively more labor intensive in finishing treatment. The Inka finishing treatments consistently employ a greater number of production steps.
Besides the increased labor investment, the Inka finishing treatment was executed with skill. In order to achieve the highly polished surface
characteristic of Inka ceramics, the vessel surfaces were burnisned and/or polished both before and after slipping. An indication of burnishing prior to slipping is that the Inka paste exhibits coarse inclusions which, if merely wiped with a damp cloth, would have been left exposed it the surface; burnishing has the effect of lodging inclusions below a surface layer of fine paste in any clay body. An indication of ournishing subsequent to slipping is the surface luster. Whereas Inka potters may have been fortunate to have used a clay with a natural luster, this quality was doubtless enhanced by additional mechanical burnishing/polishing. Vessel painting was careful and dextrous.
The Inka ceramics are a remarkably well-executed ware. Observation of the surface finishing sequence in terms of standardization, labor investment, and overall skill suggest that this ware was produced by accomplished craftsmen. In contrast to the Base Clara ceramics, wnich display a vessel finishing sequence executed with naste and with iiiniinal attention to detail, the Inka ceramics display a labor intensive finishing sequence executed with skill and precision in detail.
The main purpose of firing is to subject vessels to sufficient heat for a sufficient period of time to ensure destruction of the clay-mineral
crystals (Rye 1980:25). Once these crystals have been destroyed, clays take on the characteristic properties of a ceramic: vitrification,
hardness, and stability under a wide range of chemical and physical conditions.
The principal variables related to firing which are controlled by the potter are the rate and duration of heating, the maximum temperature attained, and the atmosphere surrounding the objects. There are, however, difficulties and uncertainties involved in ascertaining prehistoric firing method from the properties of pottery. (Many of these difficulties can be resolved by experimental studies.) The main difficulty is that the properties of pottery are affected by a number of independent variables: the physical and chemical properties of the clay, and the temperature,
length, and atmosphere of firing.4 moreover, in non-kiln firings the temperature and atmosphere may vary greatly even within a single firing episode.
Without the aid of experimental studies to pin down the properties of clays and the firing methods used prehistorically, it is more difficult to determine how prehistoric vessels were fired than it is to decide how well they were fired. Fired pottery shows direct evidence of the effectiveness of firing but not necessarily of the method of firing, because identical firing conditions will produce different results with different clays. For example, a given time/temperature/atmosphere combination will fire one clay to a dense, hard pottery, and another to a porous, soft pottery; or this combination may bring out a strong clear color in one clay but leave another gray. It is important to determine how effectively potters were working within the limitations presented by the material they used and the needs they intended to serve in order to judge the level of technical expertise they had attained (cf. Rye 1980:3-5, Shepard 1980:213-215).
In this section, the effectiveness of firing of the Wanka and Inka wares is ascertained by analyses of paste color and hardness. Standardization in firing technique is assessed by the frequencies tabulated for each firing atmosphere (derived from paste color) and by the median, mode, and range of paste hardness. Estimates of labor investment in firing technique are determined by the tasks required, either prior to or during the firing episode, in order to achieve a particular atmosphere. That is, oxidizing and reducing atmospheres imply that the prehistoric potter maintained some control over the firing atmosphere, whereas an incompletely oxidizing atmosphere implies that standards required to achieve full oxidation were not met (Shepard 1980:86).
How well vessels were fired, whether fully reduced, fully oxidized, or incompletely oxidized, only can be approximated from prehistoric potsherds. Paste color is the primary evidence for determining firing effectiveness; paste hardness is corroborating evidence. The paste of each sherd in the study sample was assigned a Munsell color using the soil color chart. The individual Munsell color assignments were subsequently grouped into color families using Munseil's divisions as a guide. b Following Shepard (1980:106-107) and Rye (1980:114-118), the paste colors were interpreted in terms of firing atmosphere. The interpretations of pottery color drawn for this analysis are summnarized in Table 6.
Table 6. Interpretations of Pottery Colors for Determining Firing
Clear throughout cross-section of wall Fully oxidized
Clear on surface, gray in wall interior Incompletely oxidized
Light gray surface, dark gray wall interior Incompletely oxidized
Brown, light to dark, throughout cross-section of wallInopeeyoizd
Gray, light to dark, throughout cross-section Inopelyxizd
of wall Incompletely oxidized
Paste hardness corresponds only roughly to firing atmosphere. Clay
hardens when it is subjected to heat, and its hardness increases with increased firing temperature. The expectation is that wares fired in a reducing atmosphere should be harder than those fired in an incompletely oxidizing atmosphere, which in turn should be harder than those fired in an oxidizing atmosphere.6 Wares fired in a partially or completely reducing atmosphere are often harder, because iron and other fluxes begin to sinter at lower temperatures in conditions of reduction, thereby promoting incipient vitrification at lower temperatures than in oxidizing
atmospheres (Rice 1983, personal communication). A field scratch test (Table 7) was used to determine paste hardness for the sample sherds.
Table 7. Scratch Test for Paste Hardness
Field Method Mohs Scale
0 = Fingernail 2.2
1 = Copper penny 3.5
2 = Penknife 5.1
3 = Harder than penknife 5.1+
If a sherd was scratched by a fingernail it was assigned a hardness of 0; if scratched by a penny, 1; by a penknife, 2; and harder than a penknife,
3. The test is imprecise, but it provided a means for comparing the
Using the information on pottery color and firing atmosphere, each
sherd in the study sample was tabulated according to its firing atmosphere as determined from its paste color. In addition the results of the scratch test were tabulated, and the median, mode, and range of the sherd hardness were found for each ware (Taole 8).
Table 8. Firing Atmospheres and Paste Hardness of Wanka and Inka Wares
0% R% 10% Mn-H Md-H Rng-H (in %)
0 1 2 3
MSS 54 17 29 2 2 27 69 4
BC 25 25 50 1 1 1 62 30 7
INKA 87 8 5 1 1 90 10
Abbreviations: 0 = oxidation; R = reduction; IO= incomplete oxidation;
Mn-H = median (hardness); Md-H = mode (hardness); Rng-H = range
Micaceous Self-Slip. For Micaceous Self Slip, 54% of the sample was well oxidized in tiring, 17% was fully reduced, and 29% was incompletely oxidized. Because oxidizing and reducing atmospheres suggest some control
over the flow of oxygen, these percentages show that 71% of tne sample was fired with some consistency of the oxygen flow. In terms of standardization and labor investment, the majority of this ware was fired consistently in an atmosphere which required some control of the firing conditions.
The median and mode of paste hardness, 2, show that most of these sherds were scratched by a penknife, with hardness of 2 (69%), a few
resisted the penknife, hardness of 3 (4%), and some were scratched by a copper penny, hardness of 1 (27%). The mode of Micaceous Self-Slip paste hardness, represented by 69% of the sherd sample, and the range, spanning three categories, demonstrate that by comparison to Base Clara (mode represented by 62%; range spanning four categories), the firing conditions were more standardized, and to Inka (mode represented by 90%; range spanning two categories) firing conditions were less standardized.
Base Clara. For Base Clara, 25% of the sample was well oxidized in
firing, 25% wa .s fully reduced, and 50% was incompletely oxidized. Because incomplete oxidation implies that the standards to achieve full oxidation
were not inet, these Base Clara vessels were probably fired without attention to the conditions of fuel and vessel arrangement, fuel type, and/or duration of firing requisite for full oxidation. In terms of
standardization and labor investment, firing was not standardized (only a slight tendency toward incomplete oxidation), and correspondingly labor investment was not high.
The median and mode of paste hardness, 1, show that the majority of
these sherds were scratched by a penny (62%), and although 1% was softer, scratched by a fingernail (hardness of 0), and 37% were harder: 30% were scratched by a penknife and 7% were harder than the penknife. The mode (62% of the sample) and the range (four categories) show that relative to the other two wares, the firing conditions were the least standardiized.
Inka. For Inka, 87% of the sample was well oxidized, 8% was fully reduced, and 5% was incompletely oxidized. These percentages show that this ware was highly standardized and that it was fired consistently in
the same atmosphere.
The median and mode of paste hardness, 1, show that the majority, 90%
of these sherds were scratched by a penny and only 10% were harder, scratched by a penknife. The mode of Inka paste hardness (90% of the sherd sample) and the range (two categories) demonstrate that Inka firing conditions were highly consistent or standardized.
In summary, the information about firing atmosphere presented in Table
8 suggests how well the Wanka and Inka vessels were fired and how
effectively the Wanka and Inka potters were working within the limitations presented by the materials they used. According to two contemporary potters having experience with open-fire (i.e., non-kiln) techniques (Rhodes 1973:264; Riegger 1972:86-91), and according to Rye (1980:98), a potter/archaeologist, it is very difficult to maintain a true oxidizing
atmosphere throughout the firing.
Thus it is apparent that the Inka and Micaceous Self-Slip wares
required a labor investment to achieve oxidizing conditions (87% of the
Inka sherds and 54% of the Micaceous Self-Slip sherds examined were fully oxidized). The conclusion that these wares were fired well and that the potters had a working knowledge of the materials they used seems warranted. The Base Clara ware, however, does not present such a clear picture. Although 50% of the Base Clara sherds were incompletely oxidized, this may be the result of arranging the pots in a dense configuration so that oxygen may not be allowed to circulate around the vessels, preventing them from becoming completely oxidized.
The median of paste hardness does not order the three wares as
expected in terms of firing atmosphere: theoretically, Base Clara which has the lowest proportion of oxidation (25%) should have the highest median, followed by Micaceous Self-Slip (54% oxidation) and Inka (87% oxidation). That Base Clara does not have a higher median than Micaceous Self-Slip may be explained by the fact thdt the clay bodies are different. Of the two clays, the Micaceous Self-Slip oody is somewhat less coarsely textured. Nevertheless, the test for paste hardness is instructive, particularly in terms of assessing the degree of standardization in firing technique, because it shows the frequency within the mode of the paste hardness for each ware. The results of the scratch
test provide a basis for ordering the three wares examined by the level of standardization achieved in firing: Inka is the most standardized and consistently fired ware, followed by Micaceous Self-Slip and Base Clara.
This discussion should demonstrate that understanding prehistoric firing technique--from the point of view of effectiveness, standardization, and labor investment--is a complicated endeavor which merits careful experimentation with clays and firing events, enabling tne archaeologist to make definitive assessments of the technical achievements of prehistoric potters.
SUMMARY AND CONCLUSIONS
This paper has examined Wanka Micaceous Self-Slip, Wanka Base Clara, and Inka jar rim sherds in terms of standardization and labor investment in the ceramic production sequence. Understanding the processes of vessel forming, finishing, and firing provides insights about the economic organization of craft production. The wares examined exhibit characteristics of two forms of specialist production, general and attached, geared respectively to general household consumption and to state uses.
Table 9 summarizes the information on standardization and labor investment in the production sequence of the Wanka and Inka wares examined.
Table 9. Characteristics of Specialist Production in Wanka and Inka Wares
Standar- Labor Form of
dization Intensity Production
MSS + General specialist for extracommunity demand (competitive
conditions of production).
BC + General specialist for intracommunity demand
conditions of production).
INKA + + Attached specialist for single
group (elite, state)
(non-competitive conditions of
These measures show that Micaceous Self-Slip exhibits a high level of standardization and cost-efficient production. This plainware type was remarkably standardized, especially in vessel forming and finishing techniques, and labor investment was low in all aspects of its
production. Moreover, archaeological evidence from the Upper Mantaro Valley region suggests that Micaceous Self-Slip exhibits little intersite variability (Earle 1983, personal communication). Therefore, the characteristics of Micaceous Self-Slip production, standardization and cost efficiency, are those of a ware produced by general specialists for extra-community demand under competitive conditions. The relationship of producer and consumer probably lacked ties of either kin or community.
The measure of standardization and labor investment in the manufacture of Base Clara show this ware to be relatively unstandardized though moderately labor intensive in all aspects of the production sequence. The producers of Base Clara are likely to have produced this ware for their own communities. Archaeological evidence suggests that there is substantial intersite variability in the ware, but little intrasite variability in the ware (Earle 1983, personal communication). Thus, the ware was probably not exported since the additional transportation costs would have provided an incentive for cost-efficient production. The
characteristics of Base Clara production, low standardization and high labor investment, typify manufacture by general specialists for intra-community demand under relatively non-competitive conditions. In this setting, the relationship between producer and consumer was close, perhaps kin-based.
Standardization and labor investment in the manufacture of the state Inka ware were consistently high in vessel forming, finishing, and firing. This ware was produced for a state apparatus with well-defined
canons for ceramic production, and archaeological evidence suggests that Inka ceramics exhibited little intersite variability in the Upper Nantaro region (Earle 1983, personal communication). Therefore, this ware shows high levels of standardization and labor investment characteristic of attached specialist production for a single group, the Inka state. Inka ceramics were produced non-competitively, where producer and consumer were tied by a close contractual relationship.
Clearly, discerning levels of standardization and labor investment in pottery manufacture provides clues to both the organization of craft production and to the overarching economic organization of society.
I am grateful to Timothy K. Earle for providing me the opportunity to study a sample of pre-Columbian Peruvian ceramics and for his perceptive comments and suggestions on an earlier version of this paper; to Cathy L. Costin and Catherine Scott LeBlanc for teaching me the ceramic typology of the ?antaro region; and to John A. Hildebrand for critically reading several versions of the manuscript. This study was made possible by the field and laboratory assistance of those people from Ataura, Peru, wiwo worked with the Upper Mantaro Archaeological Research Project. This research was supported in part by NSF Grant #BNS82-03723.
I add my thanks to Prudence M. Rice, editor of Ceramic Notes, for her constructive and insightful comments.
(1) The three Base Clara subtypes are not considered individually in this
analysis because of (1) Sample size and (2) consistency; the sample
sizes of the subtypes were too small to be used separately.
Furthermore, neither the Micaceous Self-Slip nor the Inka wares were
divided into subtypes, although these may exist.
(2) Two additional forms, D and P, classified for undiagnostic snerds,
were not used in this analysis. Neck form D is used when not enough
of the neck exists to make a positive distinction between A or B, and
neck form P serves the same function for C and N (LeBlanc [n.d.]).
(3) The classification scheme presented here has relied largely on the
Inka rim form categories defined by D'Altroy (1981:373).
(4) The term "firing atmosphere" has two implications, which although
related, should be stated explicitly. First, "firing atmosphere"
refers to the amount of oxygen available to burn the amount of fuel
supplied. If insufficient oxygen is provided, reducing conditions are
produced. If excess oxygen is provided, oxidizing conditions are
produced. If the amount of oxygen provided is not controlled or is
poorly controlled, incomplete oxidation results (Rye 1980:96).
Second, "firing atmosphere" connotes firing effectiveness for a
particular clay body. Achieving full reduction or full oxidation
implies that the potter controlled air conditions and rate of burning,
and knew the clay body's firing requirements; incomplete or partial oxidation implies that the potter lacked this skill, expertise, and
technology (particularly in New World situations lacking kilns).
(5) Each chart, for example 7.5 YR, is divided graphically by horizontal
and vertical dark lines which group the color chips by chroma and
(6) Shepard (1980:113-114) cautions that the relation between hardness and
firing temperature is not uniform and direct because of the
variability of clays. Hardness is affected by impurities in the clay and by fineness of grain and density, all of which promote sintering
of the clay body. Hardness is also affected by amount and kind of nonplastic inclusions which affect the ease of sintering. Finally,
hardness is affected by firing atmosphere; to the extent that the atmosphere acts on constituents of the paste and converts them to stronger fluxes, this is true of reducing atmospheres and iron, in
Arnold, D. E.
1975 Ceramic ecology of the Ayacucho Basin, Peru: implications for
prehistory. Current Anthropology 16(2):183-194.
Costin, C. L.
1982 Prehistoric craft production in the Yanamarca Valley, Peru. Ph.D.
dissertation proposal, Department of Anthropology, University of
California, Los Angeles. Photocopy.
D'Altroy, T. N.
1981 Empire growth and consolidation: the Xauxa region of Peru under
the Incas. Ph.D. dissertation, Department of Anthropology,
University of California, Los Angeles. University Microfilms, Ann
Donnan, C. G.
1965 Moche ceramic technology. Institute of Andean Studies, University
of California, Berkeley. Nawpa Pacha (3):115-138.
Earle, T. K.
1981 Comment to P. M. Rice. Current Anthropology 22(3):231.
Earle, T. K., T. N. D'Altroy, C. J. LeBlanc, C. A. Hastorf, and T. Y. LeVine
1980 Changing settlement patterns in the Upper Mantaro Valley, Peru.
Journal of New World Archaeology 4(1 ):1-49.
Earle, T. K., C. A. Hastorf, C. J. LeBlanc, and T. N. D'Altroy 1978 Preliminary report of the 1978 field season of the Upper 1,antaro
Archaeological Research Project. Submitted to the Instituto
Nacional de Cultura, Lima. Photocopy.
Earle, T. K., T. N. D'Altroy, and C. J. LeBlanc 1977 Regional archaeology of the Late Prehispanic periods in the Upper
Mantaro: report of the 1977 field season of the Upper Mantaro
Archaeological Research Project. Submitted to the Instituto
Nacional de Cultura, Lima. Photocopy.
Feinman, G. M., S. Upham, and K. G. Lightfoot 1981 The production step measure: an ordinal index of labor input in
ceramic manufacture. American Antiquity 46:871:884.
LeBlanc, C. J.
n.d. Rim forms of Huanca pottery. Photocopy. 1981 Late prehispanic Huanca settlement patterns in the Hanamarca
Valley, Peru. Ph.D. dissertation, Department of Anthropology,
University of California, Los Angeles. University Microfilms, Ann
LeVine, T. Y.
1979 Prehistoric political and economic change in Highland Peru: an
ethnohistorical study of the Mantaro Valley. M.A. thesis, Archaeology Program, University of California, Los Angeles.
1973 Clay and Glazes for the Potter. Chilton, Pennsylvania.
Rice, P. M.
1981 Evolution of specialized pottery production: a trial model.
Current Anthropology 22(3):219-240.
1972 Primitive Pottery. Van Nostrand Reinhold, New York.
Rye, 0. S.
1980 Pottery Technology: Principles and Reconstruction. Manual on
Archeology No. 4. Taraxacum, Washington, D.C.
Shepard, A. 0.
1942 Rio Grande glaze paint ware: a study illustrating the place of
ceramic technological analysis in archaeological research.
Contributions to American Anthropology and History, 39.
1980 Ceramics for the Archaeologist. Publication No. 609. Carnegie
Institution of Washington, Washington, D.C. (Originally published
Van der Leeuw, S.
1976 Studies in the Technology of Ancient Pottery. Organization for the
Advancement of Pure Research, Amsterdam.
A Formal and Functional Analysis
of St. Johns Series Pottery
From Two Sites in St. Augustine, Florida
Mary K. Herron
This report has two primary goals. The most basic of these is to
define and analyze the forms and functions of St. Johns series pottery as it occurs in domestic contexts of the Eastern Timucua culture area. The second goal is to compare aboriginal ceramic use patterns at two temporally related Spanish colonial and Timucua sites, in order to study Spanish-Timucua acculturation processes.
"Timucub" is a linguistic abstraction which refers to the historic Indians of north and central Florida and southeast Georgia (Deagan 1978a:89). There were, however, significant differences in environmental adaptations, dialect, ceremonial and political practices, and geographical distribution among the Timucua-speaking groups. These differences resulted in the division of the Timucua peoples into two major sub-groups, the Eastern Timucua and the Western Timucua (Deagan 1978a; Milanich 1978; Goggin 1953). The Eastern Timucua have been further divided into seven tribes which were distinct in the sixteenth and seventeenth centuries (Deagan 1978a). One of these, the Saturiwa, occupied the area around St. Augustine during the early historic period, ca. A.D. 1560-1650 (Goggin 1952:54ff).
Until recently, little archaeological research had been done on
Timucuan village life. Most work focused on mound sites of the St. Johns lib period (A.D. 1300-1513); although village sites existed, there was little specific information on them in the early 1950s (Goggin 1953). Milanich (1972) and Merritt (1977) improved upon this situation when they investigated village occupations in the Eastern and Western Timucua areas. While these investigations have supplemented our knowledge of Timucua lifeways, no analysis has yet been done on the formal and functional characteristics of indigenous Timucua pottery from domestic contexts, and its role in food preparation technology.
The second goal of this study is to investigate Spanish-Timucua acculturation processes. Previous studies of acculturation of the Timucua, as reflected in the material conditions of domestic life, have
been limited to the studies cited above. The lack of archaeological data from Timucua village sites precluded definition of an indigenous Timucuan domestic material culture, and there were no sixteenth or seventeenth century Spanish colonial sites excavated in St. Augustine until 1975 and later to provide comparative data. Furthermore, because the Timucua were extinct by 1725 (Deagan 1978a:115), the lack of data from early colonial sites has necessarily limited the study of Hispanic influences on Timucua material culture. It is known, however, that Spanish-Timucua interaction included the incorporation of St. Johns series pottery into Spanish food preparation technology of the sixteenth and seventeenth centuries (Bostwick 1976; Deagan 1978b, 1983:233-234).
A comparative analysis of aboriginal ceramic occurrence was undertaken at two sites of differing cultural affiliation and temporal range. The objective was to investigate the nature of acculturation processes among both Spaniards and Indians as they occurred in domestic activity spheres in early St. Augustine. The two sites chosen were the Fountain of Youth Park site (8-SJ-31) and the Josef de Leon site (8-SA-26-I).
The Fountain of Youth Park site is an Eastern Timucua village site
inhabited by the Saturiwa during the St. Johns IIc period, A.D. 1513-1565 (Goggin 1952:54). The site is located in St. Augustine off Magnolia Boulevard (Figure 1), and was excavated by Florida State University in 1976 and 1977 (Merritt 1977). Earlier excavations at the site were conducted in 1951 by Goggin (Seaberg 1951 and in 1934 by Ray Dickson and Matthew Stirling of the Smithsonian Institution (see Merritt 1977).
The Josef de Leon site is located at the corner of Marine Street and Bravo Lane, in St. Augustine. This site was occupied slightly later in the sixteenth century than the Fountain of Youth site, and had an exclusively Spanish colonial (rather than Indian) occupation (Singleton 1977; Deagan 1978b).
ATTRIBUTES OF THE ST. JOHNS CERAMIC SERIES
The St. Johns ceramic series is a category of aboriginal pottery identified as being of Timucua manufacture in the protohistoric and historic periods of northeastern Florida. It encompasses a number of types (Griffin 1945:218-223; Goggin 1952:99-105), including St. Johns Plain (Griffin 1945:220) and St. Johns Check Stamped (Griffin 1945:220), which were predominant in the assemblage recovered from the Fountain of Youth Park site.
St. Johns Plain and St. Johns Check Stamped are manufactured through a process of segmental coiling. The paste generally contains sponge spicules (Borremans and Shaak, this volume), although sand also occurs, particularly in late St. Johns II sites, to such a degree that sherds may appear to be samples of a grit-tempered series (Goggin 1952:101). Paste texture is generally homogeneous and "chalky" to the touch, with the exception of those samples containing sand, which are rather coarse. Core
Fig. 1. St. Augustine (crosshatching) and its environs. Youth Park site; B, Castillo de San Marcos; C, Josef de
A, Fountain of Leon site.
color ranges from black to gray-black (Goggin 1951:101-104), while surface colors range from "buff" to light brown to gray (Griffin and Smith 1949: 346). In thickness, these types range commonly from 4 17 mm., with a hardness usually of 2.0 2.5 (Griffin and Smith 1949:346). It should be noted, however, that occasional examples of this ware are much harder, particularly in later St. Johns II coastal sites (Goggin 1952:101).
St. Johns Plain and St. Johns Check Stamped are* very similar in
technological characteristics, varying primarily in temporal range (Plain is earlier than Check Stamped; Goggin 1952:54) and in surface treatment. St. Johns Plain has a surface finish which is normally smoothed and occasionally scraped (Griffin and Smith 1949:346). This type is not decorated, though sculpturing and modeling often occur in ceremonial contexts (Goggin 1952:101). St. Johns Check Stamped is decorated with a stamped check pattern applied with a carved wooden paddle, usually covering the exterior of the vessel completely, and may be either neatly or randomly executed (Goggin 1951:169). The size of the check is variable and suggests temporal as well as geographic differences in a large-small sequence (Griffin 1948:52-53).
Vessel forms of St. Johns series are variable: a large shallow bowl is most common (Goggin 1952:101), while a large deep bowl is less common (Goggin 1951:170). Small plain bowls of unusual shape also occur, usually in ceremonial contexts, and include small-necked pear-shaped jars, jars with constricted mouths and flaring orifices, gourd-shaped jars, boatshaped vessels, and other eccentric forms (Goggin 1952:102). St. Johns rim forms are also variable but are most commonly straight and slightly incurved, though unusual forms do occur, such as flanged, pointed, and reed-punctated types (Goggin 1951:170; 1952:102). Lip forms are simple, being flat, beveled inward, and/or round (Griffin and Smith 1949:346). Basal portions of most vessels are rounded, with occasional instances of sub-conoidal forms (Griffin and Smith 1949:346), though basal appendages, ranging from mere pinching to well-modeled supports, were not uncommon in St. Johns Plain, tetrapods appearing with the greatest frequency (Goggin 1952:102).
The St. Johns series has a wide range of formal and decorative
variability not discussed here because sherds recovered from both the Fountain of Youth Park and Josef de Leon sites are not of the eccentric type. Rather, they fall generally within the common range of vessel forms and decorative motifs of the St. Johns tradition.
ABORIGINAL POTTERY FROM THE FOUNTAIN OF YOUTH PARK SITE
The village area of the Fountain of Youth Park site contained only a St. Johns II component, ca. A.D. 1000-1600 (Merrit 1977:95), which immediately underlay modern humus (soil zone 1). At about 4-45 cm b.s., a house pattern with postholes, pits, and hearths was located. This house, plus four features inside the house and two outside, yielded a total of 1165 aboriginal sherds of four major types: 479 St. Johns Check Stamped, 465 St. Johns Plain, 39 sherds of San Marcos (Smith 1948), and 182 unnamed
Within these types, rim sherds of sufficiently large dimensions were used for the analysis of forms. Functional analysis was pursued by looking at vessel forms and identifying evidence of charring or sooting on the vessel exteriors,1 which was taken as an indication of use in cooking.
St. Johns Series Vessels
With respect to vessel forms, the analysis of St. Johns Series sherds was disappointing, because their small size meant that only 9 rim diameters and 6 profiles could be obtained. St. Johns Checked Stamped had rims, ranging in diameter from 20 to 66 cm. All but one of these was uncharred. Three vessels were in the form of shallow globular bowls, while two were straight-sided bowls of indeterminate depth; all were
charred (Table 1). St. Johns Plain had one large rim sherd suggesting a deep-sided globular bowl with a diameter of 29 cm.; it was not charred.
Table 1. Rim Sherd Diameter Projections
Youth Park site.
and forms at the Fountain of
St. Johns Ck. Stamped
St. Johns Plain UID sherd tempered
20 cm 20 cm 31 cm 32 cm 33 cm
43 cm 66 cm
shallow globular shallow globular shallow globular straight sided
deep globular shallow globular
*vessels with charring.
A preliminary count of charred versus uncharred sherds in these types suggested some apparent trends in Timucua vessel use (Herron 1977). Charring appeared on a greater percentage of Check Stamped pottery than on Plain St. Johns ware. In the final tabulation of 944 St. Johns sherds, it was found that 300 of 479 St. Johns Check Stamped sherds (62.9%) were charred, while only 70 of 465 St. Johns Plain sherds (15%) showed evidence of charring (Table 2). This was taken to indicate deliberate selection of Check Stamped pottery as the- primary ware for cooking purposes at this site.
Table 2. Charring Patterns of Sherds at the Fountain of Youth
St. Johns Ck. Stamped 300/62.6% 179/37.3%
St. Johns Plain 70/15.0% 395/84.9%
San Marcos 2/ 5.0% 37/94.8%
Unnamed sherd tempered 139/76.3% 43/23.6%
Two hypotheses were advanced in an effort to explain this phenomenon. Check stamping may be a functional as well as a decorative technique in that this surface treatment
(1) may provide a rough surface which is more easily handled in a
cooking ware; and/or
(2) may have desirable thermal effects, such as greater heating
capacities in cooking vessels.
Both of these functional hypotheses, if true, suggested an element of increased food preparation efficiency.
Subsequent investigations focused on the heating and cooling
capacities of these two types of pottery. The studies were based upon the assumption that heat convection currents were involved: if two objects of the same size and composition are exposed to identical conditions of temperature change, the one having a greater surface area relative to its mass will heat more quickly and cool more quickly than will the other. A check stamped sherd, in other words, with the surface variations caused by the stamp impressions, has a greater surface area than a plain sherd and should therefore absorb heat more quickly than a plain sherd. The stamped sherd should likewise radiate heat (cool) faster than a plain sherd.
Ideally, to test this hypothesis on the two ceramic types of interest, sherds of exactly the same size and thickness should be used. The selection of perfect test samples was not possible, however, because of the differential thicknesses of these sherds resulting from the manufacturing methods of the aboriginal potters. It was decided that the problems could be effectively eliminated by selecting only flat check stamped sherds to test the heating and cooling capacities of both check stamped and plain types. The reason for this compromise was that the plain and unstamped interior surface of St. Johns Check Stamped sherds is virtually identical in treatment and finishing to the exterior of St.
Johns Plain sherds.
Ten sherds of St. Johns Check Stamped type, each 1.5 square cm. in size, were used as samples for analysis. A 1/16" diameter hole was drilled into the side of each, and an inron-constantan thermocouple inserted to measure temperature changes during heating and cooling of the sherd. The thermocouple in turn was attached to a chart recorder calibrated from 0*-100* C.
Each sherd was heated twice, with first the check stamped and then the plain surface exposed to heat and then allowed to cool at room temperature. To insure that only the intended surface was subjected to temperature changes, each sherd was placed into a small box containing fiberglass which effectively insulated all but one surface. The insulated sherds, with the thermocouple inserted through the side wall to the approximate center of the sherd, were heated in an oven set to 2500 C, and temperature changes were recorded in graph form for both heating and cooling from the 40*-100* C. range.
A standard change formula was used in the calculations:
where T = temperature
&T t = time
At= kT(OK) k = constant
v= OC +2730
Average heating and cooling rates for check stamped and plain surfaces calculated by these procedures are given in Table 3.
Table 3. Rates of heating and cooling for check-stamped versus plain surfaces of St. Johns wares.
Plain surfaces Check-stamped surfaces
Heating 1.023 */sec. 1.37 */sec.
Cooling 0.990 O/sec. 1.02 O/sec.
From these figures, it can be seen that the plain surfaces heat more slowly than the check stamped surfaces, or by inference, that St. Johns Plain vessels will heat more slowly than St. Johns Check Stamped. Thle cooling rates are approximately equal for the two surface treatments. A one-tailed Mann Whitney test of significance was performed on the results of the analysis, using an alpha level of .05. For heating, the null hypothesis that the sample means are equal, or that there is no difference in the rate of heating between the two types, could be rejected at the .05 level because P = .0016:S .05. For cooling, the null hypothesis
of no difference between samples could not be rejected, with P.5787 > .05.
The differential rates of heating are hypothesized to result from the restricted movement of hot, less dense air imposed by the resistance of surrounding cooler, more dense air moving from the outer macro-atmosphere into the inner micro-atmosphere of the grooves on check-stamped sherds. This convectional movement across the gradient from warmer to cooler atmospheres will lead to different rates of achieving equilibrium temperatures between the two types of sherds.
The primary consideration here is not that check stamped sherds heat more quickly than plain sherds, but rather that the Timucua apparently
found some qualities in these vessels to be desirable for cooking wares. Certainly when considered cumulatively, the average heating and cooling rates support the idea that St. Johns Check Stamped is a more efficient cooking ware, and this efficiency would translate not only into time but into costs of fuel used for cooking.
No effort was made to test the possibility of advantages offered by the roughened check stamped surfaces, and this hypothesis is unproven at present. Nonetheless, the high percentage of sooting and the results of the heating rate tests suggest that check stamped wares were preferred by the Timucua occupants of the Fountain of Youth Park site for cooking activities. The St. Johns Plain type is believed to have served functions other than cooking, either as storage, drinking, or eating containers. The dearth of vessel form data makes it impossible to offer more specific observations on usage of this type.
Other Type Classes
Two additional ceramic types besides St. Johns wares were found at the Fountain of Youth Park site. One of these is San Marcos ware, which was a type produced in protohistoric times by the Guale Indians of the Georgia coastal area (Smith 1948:313-319). Through Spanish interaction with that
area, San Marcos pottery appears in the St. Augustine area after about AD 1570. It is found in abundance at Spanish domestic sites dating from the sixteenth through the eighteenth centuries, functioning as the primary cooking ware for the colonists (see Deagan 1983:233-235). Bo0th plain and linearly stamped varieties are known, and formal and functional analyses of the ware (Otto and Lewis 1974) have indicated that the stamped variety was used predominantly for cooking. It was essentially unmodified in formn from its occurrence in aboriginal contexts. Plain varieties, however, were more subject to modification in traditional forms and probably were used for non-cooking activities.
The second type of pottery found at the Fountain of Youth Park site is an unnamed sherd-tempered ware (Goggin 1952:57, 112) associated with the St. Johns II horizon of Florida prehistory. This type contains inclusions of sand and crushed sherds that are red-brown to orange in color, varying from 1 4 mm. in size. The sherd-tempered sherds have plain or cob-marked surfaces, though Goggin noted fabric-marked, complicated stamped, and check stamped surfaces as well.
As in the St. Johns series types, there seems to be patterning in the occurrence of charring on sherds of these two types from the Fountain of Youth site as well (Table 2). Of the 39 sherds of San Marcos pottery, only 2 (5%) were charred and this type seems to have been exempt from use
in cooking. San Marcos pottery was apparently brought into the site through interaction with the Georgia Gua Ile Indians (Smith 1948), and perhaps because it was not of indigenous manufacture (nor was it present in sizeable quantities), it was not used for cooking purposes.
Of the 182 sherd-tempered sherds, 139 (76.3%) of them were charred, indicating that this type was commonly used as a cooking ware.2 The origins of the ware are not known, though the sherds from the Fountain of Youth Park site are said to be virtually identical to a sample recovered from a Tacatacuru Timucua mission site on Cumberland Island, Georgia (Milanich, personal communication to Merritt, 1976). These two areas sustained frequent political interaction in the early historic period, and by A.D. 1600 were under the same chief (Deagan 1978a:102).
A tentative hypothesis to account for the pattern of use of these two wares can be proposed: the Saturiwa Timucua used only wares manufactured by themselves and other immediately related groups, such as the Tacatacuru Timucua, for food preparation activities.
ABORIGINAL POTTERY FROM THE JOSEF DE LEON SITE
The Josef de Leon site is of Spanish cultural affiliation, and its
temporal occupation (ca. A.D. 1570-1980) overlaps that of the Fountain of Youth Park site in the sixteenth and seventeenth centuries. A comparison of the occurrence of St. Johns series and other aboriginal ceramics at these two sites, which have differing cultural affiliations and temporal spans, should be useful in understanding Hispanic/Indian acculturation processes and their archaeological manifestations.
Spaniards in the Southeastern U.S. during the sixteenth, seventeenth, and eighteenth centuries made extensive use of aboriginal earthenwares as cooking vessels. This probably resulted in part from an erratic supply of Hispanic utilitarian wares, but more important was marriage of Spanish colonists to Indian women (Deagan 1983:123-124). Thus, the issue is not whether the Spaniards were aware of the local preference for Check Stamped wares (St. Johns series) or even desired the same performance characteristics from utilitarian earthenwares. Rather, the predominantly male orientation of Hispanic culture in St. Augustine necessitated the adoption of indigenous material culture elements associated with female activities
such as food preparation technology. At the de Leon site, as at all other Spanish sites in St. Augustine, aboriginal ceramics predominated in the
material assemblage (Bostwick 1976:4-5; Singleton 1977).
It should also be noted that by the middle of the seventeenth century, the Guale Indians of the Carolina and Georgia coast were increasingly
moving southward toward St. Augustine in reponse to depredations of English settlers in the vicinity of Charleston. The Guale are responsible for the introduction into Florida of San Marcos pottery, which was more abundant at the Josef de Leon site than at the Fountain of Youth Park
On the basis of the analyses of pottery from the Fountain of Youth
Park site, plus the foregoing sketch of the cultural affiliation of the
Josef de Leon site, three hypotheses may be advanced to predict the patterns of ceramic use at the latter site:
(1) St. Johns series pottery will be used at the de Leon site in the
same way they were used at the Fountain of Youth site;
(2) Because the de Leon site is later and incorporates Guale elements,
a greater percentage of San Marcos ceramics will be used in food
(3) Other aboriginal pottery imported into the area, manufactured by
people other than the Timucua or Guale, will not be used as
cooking wares at the de Leon site.
St. Johns Series Ceramics
Analysis of vessel forms of St. Johns pottery from the de Leon site was more informative than that from the Fountain of Youth site (Table 3).
Table 3. St. Johns Rim sherd Diameter
Josef de Leon Site.
Projections and Forms at the
St. Johns Ck. St.
16 cm 1 7 cm
straight sided straight sided
12.5 cm deep globular 15 cm deep globular
shallow globular deep globular deep globular
straight sided shallow globular shallow globular
24 cm 25 cm 38 cm
20 cm 22 cm
deep globular straight sided
shallow globular shallow globular
shallow globular shallow globular
St. Johns Check Stamped vessels ranged in diameter from 16 to 38 cm. (14 vessels measured), while 12-sherds of St. Johns Plain exhibited a similar range, from 12.5 to 35 cm. A marked divergence in vessel forms was noted, however, not only in comparing the check stamped and plain types, but also between charred and uncharred vessels within these categories.
The predominant form in St. Johns Check Stamped was a deep straightsided bowl of indeterminate depth. This form was preferred for cooking, as indicated by the frequency of charring (Table 3). The more globularshaped bowls of the Check Stamped type seem to have served some function other than cooking, as they are not so frequently charred. In St. Johns Plain type, only one straight-sided bowl was noted, and it was charred. Other vessels within this type that exhibited charring were globular in shape, and nearly all appeared to be rather deep. All of the uncharred Plain bowls were shallow.
The occurrence of charring on the two types of St. Johns series
pottery at the de Leon site supported the patterns noted at the Fountain of Youth. Of a total of 500 St. Johns Check Stamped pottery at the Ci Leon site, 233 (46.6%) were charred; of 625 St. Johns Plain sherds, only 123 (19.6%) showed charring (Table 4). Thus the first hypothesis is supported.
Table 4. Charring Patterns of Sherds at the Josef de Leon Site.
St. Johns Ck. Stamped 233/46.6% 267/53.4%
St. Johns Plain 123/19.6% 502/80.3%
San Marcos Plain 68/18.3% 303/81.6%
San Marcos Stamped 86/28.3% 21 7/71.6%
Aborig. Stamped 23/33.8% 45/66.1%
Aborig. Plain 78/75.7% 25/24.2%
Other Pottery Types
San Marcos pottery was present at the de Leon site in considerably
greater quantities than at the Fountain of Youth Park site. With a total of 674 sherds, San Marcos comprised 34.2% of the total ceramic sample at de Leon, as compared to 12.6% at the earlier site.
For purposes of analysis, San Marcos ceramics were divided into two categories, stamped and plain. San Marcos Plain category comprised 371 sherds, and of these 303 (81.6%) were uncharred. Of the 303 San Marcos Stamped sherds, 217 (71.6%) were uncharred. Although, as at the Fountain
of Youth Park site, San Marcos pottery was not charred as frequently as St. Johns pottery (and therefore presumably was not so much used in cooking), there is a much higher frequency of charring at the de Leon site
than at the Fountain of Youth Park site: 154 sherds, or 22.8% of San Marcos sherds at the de Leon site, as compared to 2 or 5% at the Fountain of Youth site. A final observation is that the degree of differentiation between stamped and plain San Marcos sherds is not as pronounced as thadt between the St. Johns types at both sites. These data support the second hypothesis concerning the greater utilization of San Marcos ceramics in food preparation and general domestic contexts at the de Leon site.
An interesting feature of the de Leon site was the range of aborignal ceramic "tradewares" rarely found in the St. Augustine area during historic times (among these are Altamaha, Irene Lamar-like Incised wares, unidentified wares, Leon-Jefferson wares). The increased number of these sherds at the site (171 sherds, or 8% of the ceramic assemblage) may be attributed to the Spanish mission presence over a large portion of Florida during the time that the de Leon site was occupied. These types, like San Marcos, can be divided into stamped and plain categories for comparative purposes (Table 4). Of 68 aboriginal plain sherds, 45 (b6.1,,) were charred; of 103 stamped sherds, 78 (75.7%0) were uncharred. A higher
percentage of these trade wares exhibit charring than did the San iMarcos sherds: 34% of the tradewares, as opposed to 23%/0 of the San Marcos wares. This is in apparent contradiction to the third hypothesis. An alternative explanation may be that as part of the acculturation process, distinctions among and between categories of native pottery suitable for cooking, such as "local" versus "trade" began to break down.
One goal of this study was to define, explain, and compare patterns of aboriginal ceramic use at two sites, the Fountain of Youth Park site and the Josef de Leon site, in St. Augustine, Florida. This study has suggested that there was an aboriginal preference for use of vessels with stamped surfaces for cooking purposes, and that this preference was a reflection of the superior efficiency and economy of these surfaces through their heat transfer qualities. In addition, this preference appears to have carried through from prehistoric times well into the historic period, when Spanish males married Indian women who brought aboriginal food technology with them into their households. The hypotheses and conclusions generated by the initial study (Herron 1977) have since been tested and supported at numerous sixteenth century Spanish colonial sites in St. Augustine (Deagan 1978:110-124).
Two more general applications of this approach may be suggested for
future research. First, the Eastern Timucua groups were characterized by a conservatism reflected in the long continuity and homogeneity of the archaeological record throughout the 2000 years of the St. Johns tradition. This suggests that the analysis of St. Johns series ceramics from the Fountain of Youth Park site is relevant not only to the ceramic technology of the St. Johns Ic period, but also to earlier periods as well.
Second, decorative ceramic techniques constitute a multifunctional phenomenon which satisfies the artistic and practical requirements of potter and user alike. The results of the present study suggest that a similar approach might be helpful in investigations of indigenous ceramic industries of the Southeast when there are dichotomies between plain and decorated wares. A re-examination of Southeastern ceramics, with an emphasis on the dual functions of plastic ceramic surface treatments, is recommended for a better understanding of aboriginal food preparation technologies and domestic activities in general.
This paper was originally written in 1918, when the author was an undergraduate student at Florida State University, and has since been substantially reworked. The author wishes to acknowledge with gratitude Dr. Kathleen Deagan, then her advisor at FSU, and the students of the FSU archaeological field schools of 19/6 and 1977. In addition, Johin King and Dr. David C. White provided invaluable assistance for the technical
aspects of measuring and interpreting thermal conductivity of the pottery.
1. This study was completed before the investigations of Hally (1983) into the composition and patterns of deposition of soot on vessel surfaces as indicators of aboriginal usage.
2. Ed. note: The thermal advantages of crushed sherd temper principally the comparability of thermal expansion coefficients of clay matrix and temper may be a factor in the use of this ware in cooking.
Borremans, N. T. and G. D. Shaak 1985 A Preliminary Report on Investigations of Sponge Spicules in
Florida "Chalky" Paste Pottery. Ceramic Notes 3.
1976 Aboriginal Creamics in Pre-18th Century Colonial St. Augustine,
Florida: the de Leon Site. Paper presented at the conference
on Historic Sites Archaeology, Tuscaloosa, Alabama.
Degan, K. A.
1978a Cultures in Transition: Assimilation and Fusion Among the
Eastern Timucua. In Tachachale: Essays on the Indians of Florida
and Georgia During the Historic Period, edited by J.T. Milanich
and S. Proctor. Gainesville: University Presses of Florida.
1978b The Material Assemblage of Sixteenth Century Spanish Florida.
Historical Archaeology, 12:25-50.
1983 Spanish St. Augustine: The Archaeology of a Colonial Creole
Community. Academic Press, New York.
Goggin, J. M.
1951 Fort Pupo: A Spanish Frontier Outpost. Florida Historical
Quarterly 30: 139-192.
1952 Space and Time Perspective in Northern St. Johns Archaeology,
Florida. Yale University Publications in Anthropology #41,
1953 An Introductory Outline of Timucuan Archaeology. Southeastern
Archaeological Newsletter. 3(3): 4-17.
Griffin, J. B.
1945 The Significance of the Fibre-Tempered Pottery of the St. Johns
Area in Florida. Washington Academy of Sciences, Journal J5
1948 Towards Chronology in Coastal Volusia County. Florida
Anthropologist 1: 49-56.
Griffin, J. W. and H. G. Smith 1949 Nocoroco, A Timucua Village of 1605 Now in Tomoka State Park
Florida Historical Quarterly, 27 (4): 341-361.
Herron, M. K.
19/7 An Analysis of St. Johns Series Ceramics from the Fountain
of Youth Park Site with Notes on the Functional Aspects of Check
Stamping. Southeastern Archaeological Conference Paper,
Merritt, J. D.
19/17 Excavations of a Timucua Village in Northeast Florida. Paper
presented at the Southeastern Archaeological Conference,
Milanich, J. T.
1972 Excavations at the Richardson Site, Alachua County, Florida:
An Early 17th Century Potano Indian Village (With Notes on
Potano Culture Change). Bureau of Historic Sites and
Properties Bulletin #2, Florida Department of State, Tallahassee.
1978 Patterns of Change Among the Western Timucua. In Tacachale: Essays
on the Indians of Florida and Georgia During the Historic Period, edited by J.T. Milanich and S. Proctor. Gainesville: University
Presses of Florida.
Otto, J. S. and R. L. Lewis
1974 A Formal and Functional Analysis of San Parcos Pottery from
Site SA-16-23, St. Augustine, Florida. Bureau of Historic Sites
and Properties Bulletin #4, Florida Department of State,
1977 The Archaeology of a Pre-18th Century House Site in St. Augustine.
Unpublished Masters Thesis, on file with the University of Florida,
Smith, H. G.
1948 Two Historical Archaeological Periods in Florida. American
Antiquity 13: 313-319.
The Southeastern Fiber-tempered
Ceramic Tradition Reconsidered
George Ward Shannon, Jr.
The objective of this paper is to document the similarities and
associations among the morphological attributes of the four fiber-tempered plain pottery types of the Southeastern Fiber-tempered Ceramic Tradition (or SFCT). The SFCT is a significant part of early pottery manufactures in the tni-state area of Alabama-Florida-Georgia between approximately 2500 and 500 B.C. Although there has been nearly a century of active research into the nature and extent of the SFCT, the interrelationships of certain types of the tradition are still poorly understood.
Four classificatory units are at the core of the SFCT (Figure 1): the Stallings Island, Wheeler, Orange, and Norwood Plain pottery types. These
are accepted as representative of four distinct ceramic technologies diagnostic of four discrete ceramic "series" (a number of types bearing obvious relationship to each other; (Willey 1949:61). These series are the basis of the spatial and temporal boundaries underlying interpretation of archaeological cultures in the tni-state area: the Stallings Island culture (Stoltman 19/2), the Wheeler complex (Jenkins 1975a), the Orange culture (Bullen 1972), and the Norwood phase (Phelps 1966). Together, the four "series" have been incorporated into a single fiber-tempered ceramic "tradition," which is defined as a group of series sharing a stylistic integrity of ceramic traits that continue through time (Willey 1945:53).
This paper will attempt to explain the SFCT more clearly as a cultural problem by first outlining the currently accepted formal, spatial, and temporal dimensions of the tradition within the Southeastern United States, and then by discussing the implications of the findings of this analysis for understanding the SFCT and Southeastern prehistory. This reconsideration of the stylistic and technological attributes originally used to differentiate the plain types of this ceramic tradition challenges the usefulness of these types as criteria for establishing four discrete cul tures.
Stallings Island Distribution
Orange Distribution I Norwood Distribution
Wbaoeler Distribution & Formative Sites
0 Miles 100
Fig. 1. Spatial distribution of the Southeastern Fiber-Tempered Ceramic Tradition.
THE SOUTHEASTERN FIBER-TEMPERED POTTERY TRADITION
The Stallings Island series is composed of two formally described types: Stallings Plain (Griffin 1943:155-168) and Stallings Punctated (Griffin 1943:155-169). Stallings Incised and Stalling Simple Stamped, although never formally described, have long been recognized as integral constituents of some Stallings ceramic assemblages (Stoltman 1972:41).
Bilbo phase pottery seems to have been a direct extension of Stallings ceramic technology (Waring in Williams 1977). In addition, many archaeologists recognize St. SimoWs phase pottery as Stallings (Milanich 1971: 120). The geographical distribution of the people who made and used Stallings pottery extends from the banks of the Savannah River south of the Fall Line, onto the coastal plain of northeastern Georgia and southeastern South Carolina. The writer has recovered Stallings pottery as far north as the Tar-Pamplico drainage of northeastern North Carolina. The southern boundary of Stallings wares appears to coincide with the Georgia-Florida border. Stoltman (1966:872) has determined that the Stallings Plain wares from the Rabbit Mount site in South Carolina suggest an antiquity on the order of 3000 B.C. (Stoltman 1974:232). Stallings Plain pottery appears to be the earliest pottery in North America (Bullen 1961; Stoltman 1966; 1974). The Stallings ceramic series persisted into the early second millennium B.C. and ended about 1000 B.C. (Stoltman 1972:37).
Four ceramic types have been formally described for the Wheeler
series: Wheeler Plain (Haag 1939:2), Wheeler Simple Stamped (Webb and DeJarnette 1942; formerly referred to as Pickwick Simple Stamped by Haag 1939:3), Wheeler Punctated (Haag 1939:4; formerly Bluff Creek Punctated Griffin 1939), and Wheeler Dentate Stamped (Webb and DeJarnette 1942:514; formerly Alexander Dentate Stamped Haag 1939:5). The people who made and used Wheeler pottery resided within the Pickwick and Wheeler basins of the Tennessee River Valley of northwestern Alabama. Other concentrations of Wheeler pottery exist in the Yazoo Basin of Mississippi, around the mouths of the Mississippi and Pearl Rivers in southeastern Louisiana, and along the stretch of the Tombigbee River that crosses the Alabama and Mississippi state line. Most recently, fiber-tempered pottery has been identified in Late Archaic contexts at the Nebo Hill site located in Missouri's western plains region (Reid 1984), possibly expanding the known distribution of Wheeler pottery further inland and far above the Fall Line. No radiocarbon dates for Wheeler pottery have been obtained; however, this complex has been dated by relative means from 1200 to 500 B.C. (Jenkins 1975b).
Two ceramic types have been formally described for the Orange series: Orange Plain (Griffin 1945:219-221) and Orange Incised (Griffin 1945: 218-223). One apparent variant of the Orange Incised form is a type known as Tick Island Incised (Moore 1894:601). These two types, Orange Incised and Tick Island Incised, co-existed but by 1450 B.C. Tick Island Incised seems to have vanished (Milanich and Fairbanks 1980:156). In many Orange assemblages, a punctated variety has been recognized (Bullen 1972: Figures
1 and 9; Jahn and Bullen 1978: Figures 5e and 5j), but has never been formally described as a type. The people who made and used Orange pottery were located in the St. Johns River and Indian River drainages of the Atlantic coastal plain of Florida. Ripley Bullen (1972:14-24) suggested that the people who made and used Orange pottery first inhabited the St. Johns River Valley and from this area gradually expanded westward toward the Lower Mississippi River Valley, covering most of Florida by ca. 1200 B.C. Orange pottery has been found in the Glades area (Griffin 1971:329),
Manatee region (Willey 1949:178), central Gulf Coast (Griffin 1971:326), northwest Gulf Coast (Bullen 1958), and central Florida (Griffin 1971: 332). The Orange period started around or shortly after 2000 B.C. and ended before 1000 B.C. (Atkins and MacMahan 1967:140; Bullen 1961, 1972:9).
The Norwood series is composed of two ceramic types: Norwood Plain (Phelps 1965:66-67) and Norwood Simple Stamped (Phelps 1965:68). Also, two provisional types are associated with the series: Norwood Punctated (Phelps 1969:13) and Norwood Incised (Jahn and Bullen 1978: Figures j and m). Norwood pottery is found primarily in the Big Bend region of Florida, an area delimited by the far banks of the Apalachicola and Suwannee Rivers. It has also been reported, however, from sites such as the Palmer site located well to the south of this region (Bullen and Bullen 1976), from sites like Lower Peach Tree Ferry and Stallworth (Sears n.d.:51,55) located far to the west of this area, and from sites like Stafford North and Table Point (Milanich 1971:52) located far to the east of the Big Bend region. Two radiocarbon dates, Samples FSU-67 and M-394, have been determined for the Norwood series: 1012 B.C. + 120 (Stipp et al. 1966:52; Phelps 1966:19), and 1195 B.C. + 250 (Bullen 1961:104; Phelps 1965:67).
THE CERAMIC ANALYSIS
The objective of this analysis is to test the validity of our current classification of plain fiber-tempered sherds. To begin this test a sample of sherds (refer to Tables 1-4) was drawn from the collections of Stallings, Wheeler, and Orange plain sherds formally described by Sears and Griffin (1950), and from a collection of Norwood plain sherds formally described by Phelps (1965). For the purposes of this analysis, it was assumed that the sherds under consideration were derived from an independent random sampling of all possible fiber-tempered plain pottery and that the resultant distributions are normally, distributed. This sample, consisting of 178 sherds, representing all four fiber-tempered plain types, was subjected to a modal analysis. Both stylistic and technological ceramic attributes were investigated to determine if their occurrence among these four plain types could be attributed to random sampling fluctuations, or were due to non-chance factors--that is, to the falsity of proposed null hypotheses. All of the null hypotheses (Ho:) state that the ceramic attribute in question is distributed independent of type identification. These hypotheses hold that there are no significant differences between ceramic attributes as identified, and their distributions among types. The alternative hypotheses (Hl:) state that ceramic attributes are related to type identification.
The attributes measured were: rim, lip, body, and base form;
hardness; exterior, interior, and core color; exterior and interior surface finish; temper; and paste texture. These attributes were measured
because the outline for describing southeastern pottery standardized by Ford and Griffin (1960) recognizes these measures as "the guiding features for typing pottery" (Ford 1961:20). The measurement of rim diameter was accomplished with the aid of a graduated concentric centimeter scale.
Table 1. The Stallings Plain Sample (N=42).
Ca .-, 0. 4Jh a- ),
0} P} ,-0P .
.A 0 :
1 E -41 Ai4 0C W 0 0 0
A-4 0r 04 4-1 4 44
41) 44 aC 4J 4-4 01 W) 0 M 00 0 c
; a1) w -0 0 4) 44$4 444. 4 1 4 j
01 I-a -4 0 4 2 C a
Cd) -90-.Co C
1 ( 9612) 2 (31579) 3 (31579) 4 (31579) 5 (31579)
6 (31579) 7 (2715) 8 (7127) 9 (7772) 10 (2715) 11 (31579) 12 (31816) 13 (31579)
14 (31579) 15 ( 2715) 16 (31816) 17 ( 2715) 18 (31816) 19 (31816)
I OYR6/ 6 10oYR3/1
10YR4/2 I OYR4/ 2 IOYR613 1 OYR5/3 I OYR6/3 10OYRS3
10YR3/2 IOYR5/6 10YR5/3 I 0YR6/4 10YR6/3 10YR6/4 IOYR5/2 10YR6/3 7.5YRN5/ 1 OYR6/4 10OYR6/4 1 OYR6/4 IOYR6/4
10YR412 1 OYR6/3 10YR4/1
7.5YRN5/ S10YR5/3 7.5YRN5/ 1 OYR4/1 10 YR6/4 10YR4/1 10YR4/1
10YR3/1 7.5YRN5/ S0YR3/1 7.5YRN5/ 7.5YRN3/
sm C E sm A E smn A E sm C F sm A E sm C F sm C F rg C E sm A E sn C F sm C E
fl A E
smn A E sm A J
fl C F
sm A I fl C E rg A E rg A E
2.5 2.5 2.0 2.5 2.5 2.5 2.5 2.5 2.5 2.5 2.5 2.5
2.5 2.5 2.5 2.5
I OYR6/6 IOYR3/1 10YR4/2 I OYR6/3 7.5YR6/6 I OYR6/3 I 0YR3/2 1 OYR5/6 1 OYRS/3 1 OYR6/4 I 0YR6/3 1 OYR6/4 5YR4/3
1 0YR6/3 S10YR6/4 I OYR6/4 7.5YR5/2 7.5YRN5/ 7.5YRN5/
Table 1. Continued.
20 ( 7127) 21 (31579) 22 ( 7772) 23 (31579)
24 ( 7773) 25 ( 7127) 26 ( 7773) 27 (31579) 28 (31579)
29 (31568) 30 (31816) 31 (31579) 32 (31579)
33 (31579) 34 (31568)
35 (31568) 36 (31816) 37 (9612) 38 (7127)
39 (31579) 40 (31816) 41 (31579) 42 (31579)
2.0 2.0 2.5 2.5
2.0 2.5 2.5 2.5 2.5 2.5 2.5 2.5 2.5 2.5 2.5 2.5 2.5 2.5
2.0 2.5 2.5
1 OYRS/3 10YR4/1 1 OYR6/4
IOYR5/2 10OYR5/3 1 OYRS/3 S10YR6/4 1 OYR6/3 10OYR6/3 1 OYR5/4
1OYR6/4 1OYR6/3 10YR6/3
IOYR6/3 10OYR5/4 1 OYR5/4
IOYR6/4 LOYRS/2 10YR6/4 I OYR5/3
(7127) Irene site, Georgia; and (9612) Swift Creek site, Georgia.
IOYR5/3 1 OYR5/3 10YR6/4 1OYR4/1
10 OYR5/3 10 OYR6/4 10YR5/2
1OYR6/3 10OYR5/4 10 OYR6/4 10YR5/2
10YR5/2 I OYR6/3 I OYR5/4 1 OYR5/4
7.5YRN5/ 10 OYR6/4
1OYR4/1 10YR5/2 10OYR6/4 10YR5/3
7.5YRN5/ 1 OYR4/ 1
1OYRS/2 10YR5/2 10YRN4/ 7.5YRN5/
1OYR6/3 1 OYR4 /1 O10YR3/1 7.5YRN5/
IOYR4/1 7.5YRN5/ I OYR3/1
Note. University of Michigan, Museum of Anthropology, Ceramic Repository provenience data: (31579, 2715, 31816, 31568) Stallings Island site, Georgia; (7772, 7773) Bilbo site, Georgia;
Table 2. The Wheeler Plain Sample (N=42).
tU 4) :3
U) u-U (0 04 4-44P
A1- r- 2 ( 9
U ~ O ~ 4 r4 4 U) .0 0
. 0 A 0 1 .4 4-4 0 0
4j 44 ,:j 4j 4-. ) dt t ) Q
P- 0) ) 'TI 0 41J P~ 44 4 1 4 3 4-)
,a w U) k r-I 0 4 V9 0 t
4 4 -- -,0. 0 ( Co 0uto C
1 (7044) 2 (7044) 3 (7097) 4 (7097) 5 (7044) 6 (7044) 7 (7097) 8 (7044) 9 (7042) 10 (7097) 11 (7097) 12 (7097)
13 (7044) 14 (6883) 15 (6883) 16 (6883)
17 (7097) 18 (7097) 19 (6883)
2.0 2.5 2.5 3.0 3.0 3.0
IOYR5/3 IOYR5/3 I OYR5/4 1 OYR5/4 7.5YR5/6 1 OYR6/3 10OYR6/4 2.5Y 4/2 10YR5/3 10 OYR6/3 1OYR5/3 10OYR6/6 2.5Y 4/2 1 OYR3/1 10YR4/1
1OYRS/2 O10YR6/3 10YR4/2 IOYR4/3
1OYR6/4 1OYR5/3 2.5Y 4/2 1 OYR5/4 7.5YR5/6 1 OYR5/3 O10YR3/2 IOYR5/4 I OYR6/4 1OYR4/I IOYR5/6 10OYR4/3 10 YR3/l IOYR3/1 IOYR4/1 IOYR6/3 IOYR5/2 10YR5/4 IOYR3/1l
7.5YRNS/ 1 OYR5/3 2.5Y 4/2 1 OYR5/4 7.5YR5/6 1 OYR4/1I 1 OYR3/1I
2.5Y 4/2 O10YR3/1 I OYR4/1I IOYR4/1 IOYR4/1I 10YR3/1 10OYR3/1 IOYR4/1 1OYR3/1 1OYR4/1
B E A E A E A E A E A E A E B I B E B E B E A E B E B E B E A G B E B E B E
Table 2. Continued.
20 (7097) 21 (7097)
22 (6883) 23 (6883) 24 (6883) 25 (7044) 26 (7044) 27 (6883) 28 (6883) 29 (7097) 30 (6883) 31 (6883)
32 (7042) 33 (7097) 34 (6883) 35 (7044) 36 (7044) 37 (7044) 38 (7097) 39 (7045) 40 (7097) 41 (7045) 42 (6883)
3.0 2.5 2.5 2.5
3.0 2.0 3.0 2.0 3.0
3.0 3.0 3.0 3.0
3.0 3.0 3.0 2.5 3.0 2.5
1 OYR5/3 1OYR6/4
10YR6/4 IOYR5/8 I OYR7/4 7.5YRN3 / 1 0YR5/6 IOYR6/3 10YR6/3 10YR4/1
10 OYR6/3 10 OYR6/4 7.5YR5/2 1 OYR5/4 10OYR6/4 7.5YR5/2 1 OYR6/4 7.5YR5/2 10YR5/3
I OYRS/2 10OYR6/4
10YR3/2 1 0YR4/2 10YR6/4
IOYR5/4 IOYR5/4 7.5YRN3/ 10YR5/6 1 OYR6/3 O10YR4/2 7.5YR6/6 1 OYR5/3 1 0YR4/1
7.5YRN5/ 1 0YR5/3
7.5YRN5/ 7.5YRN5/ IOYR4/2 IOYR4/1 S10YR6/4 I OYR5/2 I OYR6/6
Hobbs Island, Alabama.
O10YR4/1 10YR4/1 7.SYRN5/ 10YR4/1 10YR4/1
7.5YRN3/ 10YR5/6 IOYR4/1
1OYR4/2 10OYR4/1 10YR4/1
10YR4/1 7.5YRN3/ 1OYR3/1 10YR3/1
7.5YR5/2 7.5YRN5/ 7.5YRN5/ O10YR3/1 1 OYR3/ 1 IOYR4/1 10YR3/1 1 OYR3 /1
8 ? 8 cf 10 ?
20 cf 21 ?
Note. University of Michigan, Museum of Anthropology, Ceramic Repository provenience data: (7044, 7045, 7042, 7097) Lauderdale County, Pickwick Basin, Alabama; (6883) Colbert County,
Table 3. The Orange Plain Sample (N=47).
to (a Cl) 0 4)i
u 14 4 44k
U- H to4*d( 00
A~" 11 ., 0 W0 -r4, 4 404 4j
0 A, 0 ~ 44 04)4 4C)1
4- -'j ) 44 a 4 0 0) 0) ( 40 01
4 ) 10 0 41i 4-1 P 44 43 43 4-) 4,a W U 46 4 (
0 (3 0 0) (dr4
C4) t4 14 p Q L 4 PL4
I (v67-267) 2 (v67-180) 3 (v67-205) 4 (v67-139) 5 (v67-179) 6 (v67-180) 7 ( 95098) 8 (v67-138) 9 (v67-170) 10 ( 95098a)
11 (v67-152) 12 (v67-168)
13 (v67-162) 14 (v67-212) 15 (v67-212) 16 ( 95098a) 17 (v67-168) 18 (v67-156) 19 (v67-167)
1 50 a 40
1 38 s 40 s 42 1 42 a 32 s 50 s 50 1 50 s 46 i ? i 50 1 38 1 32 1 46 s 50
a 48 a 48
2.5 7.5YRN3/ 2.5 1 OYR6/3 2.0 7.5YR7/4 2.5 7.5YR4/2
2.5 IOYR4/1 2.5 7.5YRN6/
2.5 5YR3/3 3.0 7.5YR4/2
3.0 5YR5/1 2.5 5YR5/3 2.5 I OYR5/2 2.0 7.5YR5/2
2.5 5YR7/6 2.0 7.5YR4/2 2.5 O10YR6/3 2.5 7.5YR6/4 2.5 7.5YR6/4
7.5YR7/4 7.5YRN3 / IOYR5/1 7.5YRN4/ 7.5YR4/4 7.5YR3/4 7.5YRN5/
5YR4/4 7.5YR4/2 IOYRS/2 5YR5/3
10YRS/2 10 OYR6/3 5YR4/1I 10YR6/6 1 OYR6/3 I O0YR5/1 7.5YR6/2 7.5YR6/6
7.5YRN4/ 7.5YRN3/ 7.5YR5/4 7.5YRN4/ 7.5YRN4/ 7.5YRN3/ 7.5YRNS/
5YR3/1 10OYR2/1 1OYR4/1I 5YR4/1 10YR3/1I 10YR6/3 10YR5/1 7.5YR3 / 7.5YRN4/
fl A sm D fl A sm A fl A smn A
fl A fl A fl A fl A fl A fl D fl A fl A
sm A fl A fl D fl A fl A
Table 3. Continued.
20 (v67-154) 21 (v67-169) 22 (v67-158) 23 (v67-136)
24 (v67-215) 25 (v67-184) 26 (v67-172)
27 (v67-167) 28 (v67-168) 29 ,(v67-162) 30 (v67-136) 31 (v67-151) 32 (v67-175) 33 (v67-162) 34 (v67-151) 35 (v67-180) 36 (v67-161) 37 (v67-174) 38 (v67-217) 39 (v67-180) 40 (v67-120) 41 (v67-205) 42 (v67-156) 43 (v67-162) 44 (v67-152) 45 (v67-205) 46 ( 95098a)
12 i 40 12 r 13 i 50 14 r
2.5 10YR5/2 2.5 7.5YR4/2 3.0 O10YR7/4
2.5 5YR6/3 2.5 7.5YR6/4
3.0 5YR5/3 2.0 5YR5/4 2.0 7.5YR8/4
2.0 7.5YR4/2 2.5 7.5YR5/2
3.0 2.5YR5/4 2.0 7.5YRS/2 2.5 7.5YR4/2
2.0 7.5YR7/4 2.0 7.5YR5/2 2.0 10OYR4/1 2.5 7.5YR6/4
2.5 10YR5/2 2.0 7.5YR6/2 2.0 7.5YR6/4 2.0 1 0YRS/3 2.0 1 OYR6/4 2.0 10YR5/2
2.5 7.5YR5/4 3.0 1 OYR6/2 2.0 7.5YR5/2
2.0 7.5YR6/2 3.0 1 OYR6/3
Volusia County, Bluffton Midden (8Vo22), Florida.
I O0YR5/2 7.5YR5/6 10OYR4/1 5YR3/1 7.5YR6/2
5YR6/2 5YR4/1 7.5YR8/4 7.5YR4/2 7.5YRN5/
7.5YR4/2 7.5YR7/4 7.5YRN5/ 10YR4/1 7.5YR5/4
5YR4/4 7.5YR6/4 7.5YR5/2 10YR4/1 I 0YR6/4 10YR6/3 7.5YR5/4 I O0YR4/1 7.5YR7/4 7.5YR6/2 1 OYR5/2
10YR4/1 7.5YRN4/ 10YR4/1
5YR4/1 7.5YR8/4 7.5YRN3/ 7.5YRN5/
5YR5/2 7.5YR5/2 7.5YRN4 / 7.5YRN5/ 7.5YRN4/ IOYR5/3 7.5YR5/4
SYR3/1 7.5YRN4/ 7.5YRN4 / 10OYR4/1I I O0YR6/4 10OYR4/1 7.5YRN3/ IOYR4/1 7.5YR5/2 7.5YRN4/
A E A E A E A E A E A E A E A E A E A E A E A E D E A E A E A E A E A E A E A E A E A E A E A E A E A E A E A E
Note. University of Florida, Florida State Museum provenience data: (v67, 95098, 95098a)
Table 4. The Norwood Plain Sample (N=47)
A W0 oo
*d 0 *r 44 ,M, r r r r
44 a) 4. 4. a t t 0 o
* *r *r O P 4 4J r* 44
CL( J r Mr 10 En 04 $ 41t
44i 0 *O.d 0.d 4 4)(
01 41 6J-3 b44 P3 M4 (4 D 0 En~ A4 00
1 (Fr 4) 2 (Fr 4) 3 (Fr 23) 4 (Fr 23) 5 (FR 4) 6 (no pro)
7 (Fr 4) 8 (Fk 22) 9 (no pro) 10 (Fr 4)
11 (Ta 35) 12 (Fr 4) 13 (Fr 4) 14 (no pro)
15 (Fr 4) 16 (Fr 4) 17 (Fr 4) 18 (Le 103)
19 (Le 103)
2.0 3.0 3.0 2.5 3.0
2.0 2.5 2.5
2.0 2.5 2.5 2.5 3.0 3.0 3.0 3.0
S10YR5/4 10YR4/1 10YRS/2
10YR4/2 10YR2/1 7.5YR5/4 10YR4/2 10 OYR6/3 7.5YR4/2 10YR5/2 10 OYR6/4 7.5YR5/4 10YR3/2
10YR3/1 1 OYR3/ O10YR3/1 10YR4/2
10YR3/2 IOYR4/1 1OYR5/2 1 OYR6/4 1 0YR5/2 7.5YR3/2 1OYR4/1
O10YRS/1 7.5YRN3/ 10 OYR5/ 2 1 OYR6/4 10YR4/1 10OYR3/1 IOYR4/2 1OYR4/2 O10YR3/1
1OYR5/2 O10YR6/4 I OYR6/4
10YR3/2 1OYR4/1 10YR3/1
10YR4/1 lOYR5/2 7.5YRN3/ 10 OYR3 /1 I 0YRS/1 7.5YR4/2 IOYR4/1 I OYR6/4 10YR4/1
10OYR3/1 1OYR3/1 IOYR6/1 1 OYR3/1 I0YR4/1
C H C E C E C E C E C E C J A E C F A E C J C E C F C H C H
A E A E C F C F
Table 4. Continued.
20 (Fr 4) 21 (no pro) 22 (Fr 4) 23 (Fr 4) 24 (Fr 4) 25 (Fr 4) 26 (no pro) 27 (Fr 4) 28 (Fr 4) 29 (no pro) 30 (Ta 35) 31 (Fr 4) 32 (Fr 23) 33 (Fr 4) 34 (Fr 4) 35 (Le 15) 36 (Ta 35) 37 (no pro) 38 (Fr 4) 39 (Fr 4) 40 (Fr 4) 41 (Fr 4) 42 (Fr 4) 43 (Fr 4) 44 (Fr 23) 45 (Le 24) 46 (Le 24) 47 (no pro)
2.5 2.5 3.0 3.0
2.0 3.0 2.5 2.5 3.0 3.0 2.5 3.0
3.0 2.5 2.5
2.5 2.5 2.5
2.0 3.0 3.0 2.5 2.5
2.5 2.5 3.0 3.0
IOYR3/1 O10YR3/1 5YR4/3
IOYRS/3 IOYR3/2 7.5YR4/4 IOYR3/1 5YR3/4 10YR3/1 10OYR3/2 10YR6/3 IOYR8/1 7.5YR6/2 10 OYR3/2 10 OYR4/2 10YR5/3 10YR3/2 1OYR3/2 10YR4/2 10YR4/1 10YR3/1 1 0YR5/3 1OYR6/2 10YR4/2 1 0YR6/4 1 OYR6/4 1 OYR6/4
IOYR4/4 IOYR4/1I 7.5YR4/2 10YR3/1
10OYR4/6 1 OYR6/1
IOYRS/2 IOYR4/3 1OYR5/2 IOYR4/1 1 OYR5/2 IOYR7/1 7.5YR6/2
1OYR4/3 IOYR4/2 10OYR5/3
10YR6/4 1 OYR6/1 I OYR4/2 IOYR5/I 1 OYR5/2 10OYR5/3
I OYR5/1 10YR6/2 10OYR6/4 10OYR6/3 1 OYR6/3 1 OYR6/4
11 f 11 f 13 ?
Note. Florida State University, Tallahassee, Department of Anthropology and Archaeology provenience data housed in the Research Laboratory of Sociology and Anthropology, East Carolina University, Greenville.
IOYR3 /1 1OYR4/1 5YR3/1 10YR3/1 10YR3/1 10YR4/1
IOYR3/2 1 OYR3/I 1OYR5/2 10YR3/1
7.5YRN4/ 10YR3/1 I OYR4/2 10YR3/1
O10YR4/1 10YR4/1 1 OYR4/1 IOYR3/1 IOYR4/1 O10YR3/2 10YRS/2
10YR4/1 IOYRS/1 10 OYR3/1
Key to Tables 1-4:
Rim forms may take one of three variations: strdight (s), inward
slanting (M), and outward slanting (o). Lip forms may be either round (r) or flat (f). Base forms are round (r), unidentifiable (?), flat with a concave underside (cf), or flat (f). Exterior and interior surface finish is denoted as either smoothed (sm), roughened (rg), or filmed (fl). Paste temper is described as either vegetal fiber with imprints that appear to be cylindrical when viewed in cross section (A), vegetal fibers with imprints that appear to be flat or angular in cross section (B), clastic inclusions and vegetal fibers that appear to be cylindrical when viewed in cross section (C), or clastic inclusions and vegetal fibers that appear flat in cross section (D). Paste texture was recorded as smoothed and vesicular (E), smoothed and grainy (F), contorted and vesicular (G), contorted and grainy (H), laminated and vesicular (1), or laminated and grainy (J).
Thickness measurements were determined to the millimeter with a sliding calipers. Measurements of paste hardness were derived from the application of a simple scratch test using Mohs standard. Color determinations were standardized via the application of the Munsell (1975) soil color chart. The use of a hand lens and a steroscopic binocular microscope facilitated the examination of temper and paste texture.
Interval scale measurements obtained from an examination of rim, lip, body and base thickness, hardness, and vessel diameter were subjected to the T-test (Tables 5-11; Thomas 1976). T-test null hypotheses are presented in terms of a two-tailed test with an alpha level of 0.10. Variations in measurements of these attributes tend to cluster around a mean (Table 5). It is important to recognize that the results of these T-tests do not prove similarity among ceramic attributes; rather, they indicate to what degree similarity may be apparent. Observations on core, interior -and exterior color, interior and exterior surface finish, paste temper, paste texture, and rim, lip, and base form were coded as nominal level variables and then subjected to the chi square analysis (Tables 12-18; Doran and Hodson 1975). Chi square test null hypotheses are presented in terms of a two-tailed test with an alpha level of .05. The chi square test results provide an indication of whether or riot the presence of a given attribute is independent of type. Both of these tests provide a good indication of whether or not a given null hypothesis is sufficiently unlikely, given a decision rule. And, both provide an objective means of evaluating just what are the critical attributes that allow analysts to label each of these four plain types as a discrete ceramic entity.
Test results indicate that among the nineteen attributes examined,
only five (i.e., lip and body thickness, surface finishes, and temper) are critical in differentiating the series. The attribute of lip thickness presents an interesting dichotomy. The rejection of the fundamental hypothesis (H0:) in the Orange/Stallings comparison indicates that there is a statistically significant difference in mean lip thickness found between these two plain types. This difference of nearly 3.00 mm. may be interpreted as a reflection of a preference for flat-lipped vessels by the people who made and used Stallings pottery, a preference not indicated by the Wheeler, Orange and Norwood sample.
Differences in body thickness (Table 8) may be attributed to the
uneven, somewhat lumpy, and often contorted nature of the paste that is usually apparent regardless of type. This test makes it apparent that several variations in mean body thickness are statistically significant, but these differences are not interpreted as indicative of culturally shared ideas about the proper thickness of a vessel. A consideration of the overlapping range of body thickness apparent within this sample makes this point obvious. Stallings sherds range in thickness from 8.00 to 16.00 mm., Wheeler from 6.00 to 17.00 mm., Orange from 6.00 to 11.00 mm., and Norwood from 8.00 to 21.00 mm. The test results suggest that Stallings and Norwood Plain are both rather thick and therefore similar in
respect to this attribute.
Table 5. Mean Values of Interval Scale Data.
Attribute Stallings Orange Norwood W'heeler
Rim Thickness 10.68 mm 9.30 mm 10.25 mm 9.50 mm
Lip Thickness 9.68 mm 6.80 mm 8.50 mm 8.37 mm
Body Thickness 12.14 mm 8.40 mm 12.58 mm 9.86 mm
Base Thickness 20.00 mm 11.50 mm 14.00 mm 13.40 mm
Hardness 2.42 2.40 2.57 2.77
Rim Diameter 39.47 cm 44.10 cm 49.50 cm 42.75 cm
Table 6. T-Test of Rim Thickness.
Type (Ho:) (Hi:) alpha s df T obs. T crit. decision
Nor/Whe 7= 9.50 X/ 9.50 .10 3.31 3 .5 2.35 fail to reject (Ho: Nor/Sta 7=10.68 X10.68 .10 3.31 3 .26 2.35 fail to reject (Ho:) Nor/Ora X= 9.30 7/ 9.30 .10 3.31 3 .57 2.35 fail to reject (Ho:) Whe/Sta X=10.68 X/10.68 .10 1.31 7 2.56. 1.89 reject (Ho:)
Ora/Sta 7=10.68 /10.68 .10 1.68 20 3-38 1.72 reject (Ho:)
Whe/Ora X= 9.30 X/ 9.30 .10 1.31 7 .43 1.89 fail to reject (Ho:)
Table 7. T-Test of Lip Thickness.
Type (Ho:) (Hi:) alpha s df T obs. T crit. decision
Nor/Whe = 8.37 1 8.37 .10 3.10 3 .08 2.35 fail to reject (Ho:)
Nor/Sta = 9.68 9.68 .10 3.10 3 .76 2.35 fail to reject (Ho:)
Nor/Ora = 6.80 # 6.80 .10 3.10 3 1.13 2.35 fail to reject (Ho:)
Whe/Sta = 9.68 I/ 9.68 .10 2.87 7 1.29 2.87 fail to reject (Ho:)
Ora/sta X= 9.68 X/ 9.68 .10 3.87 20 3.42 1.72 reject (Ho:)
Whe/Ora X= 6.80 / 6.80 .10 2.87 7 1.55 1.89 fail to reject (Ho:)
Table 8. T-Test of Body Thickness.
Type (Ho:) (Hi:) alpha s df T obs. T crit. decision
Nor/Whe = 9.86 / 9.86 .10 2.99 38 4.94 1.60 reject (Ho:)
Nor/Sta X=12.14 V12.14 .10 2.99 38 .93 1.60 fail to reject (Ho:)
Nor/Ora = 8.40 3V 8.40 .10 2.99 38 8.89 1.60 reject (Ho:)
Whe/Ora =12.14 V/12.14 .10 2.60 28 4.75 1.70 reject (Ho:)
Ora/Sta =12.14 412.14 .10 1.79 21 9.84 1.70 reject (Ho:)
Whe/Ora X= 8. 40 8.40 .10 2.60 28 3.04 1.70 reject (Ho:)
Table 9. T-Test of Base Thickness.
Type (Ho:) (Hi:) alpha s df T obs. T crit. decision
Nor/Whe 1=14.00 -,14.00 .10 5.74 2 .09 2.92 fail to reject (Ho:)
Nor/Sta =20.00 X/20.00 .10 5.74 2 1.72 2.92 fail to reject (Ho:) Nor/Ora X=11.50 111.50 .10 5.74 2 .84 2.92 fail to reject (Ho:)
Whe/Sta X=20.00 _20.00 .10 8.48 1 .99 6.31 fail to reject (Ho:)
Ora/Sta =20.00 p20.00 .10 2.30 3 7.39 2.35 reject (Ho:)
Whe/Ora X=11.50 X11.50 .10 8.48 1 .41 6.31 fail to reject (Ho:)
Table 10. T-Test of Degree of Hardness.
Type (Ho:) (Hi:) alpha s df T obs. T crit. decision
Nor/Whe = 2.77 2.77 .10 2.55 46 .43 1.60 fail to reject (Ho:)
Nor/Sta = 2.42 2.42 .10 2.55 46 .51 1.60 fail to reject (Ho:) Nor/Ora = 2.40 2.40 .10 2.55 46 .56 1.60 fail to reject (Ho:)
Whe/Sta = 2.42 2.42 .10 .87 41 2.69 1.60 reject (Ho:)
Ora/Sta X= 2.42 / 2.42 .10 .20 46 .68 1.60 fail to reject (Ho:) Whe/Ora X= 2.40 34 2.40 .10 .87 41 2.84 1.60 reject (Ho:)
Table 11. T-Test of Rim Diameter.
Type (Ho:) (Hi:) alpha s df T obs. T crit. decision
Nor/Whe X=42.75 142.75 .10 1.00 3 13.00 2.35 reject (Ho:)
Nor/Sta R=39-47 Z39.47 .10 1.00 3 20.06 2.35 reject (Ho:)
Nor/Ora X=44.10 X/44.10 .10 1.00 3 10.80 2.35 reject (Ho:)
Whe/Sta =39.47 V39*47 .10 6.59 7 1.40 1.89 fail to reject (Ho:)
Ora/Sta X=39.40 Y39.40 .10 6.08 19 3.45 1.72 reject (Ho:)
Whe/Ora X=44.10 X/44.10 .10 6.59 7 .57 1.89 fail to reject (Ho:)
Table 12. Chi Square Test of Core Color.
Gray Brown fo fo
Norwood 35 11 46
Stallings 31 12. 42 .058
Total 66 22 88
Norwood 35 11 46
Wheeler 32 10 42 .000
Total 67 21 88
Norwood 35 11 46
Orange 35 9 44 .130
Total 70 20 90
Stallings 31 11 42
Wheeler 32 10 42 .040
Total 63 21 84
Stallings 31 11 42
Orange 35 9 44 .400
Total 66 20 86
Wheeler 32 10 42
Orange 35 9 44 .130
Total 67 19 86
Note. (Ho:) states that the attribute of core color is independent of type. Chi square critical value is 3.84, df (1), alpha level of .05.
Table 13. Chi Square Test of Interior Color.
Gray Brown 2
Type f'o fo Total X abs.
Norwood 17 30 47
Stallings 6 32 38 4.35
Total 23 62 85
Norwood 17 30 47
Wheeler 10 31 41 1.39
Total 27 61 88
Nor'wood 17 30 47
Orange 19 22 41 1.11
Total 36 52 88
Stallings 6 32 38
'Wheeler 10 31 41 5.62
Total 16 63 79
Stallings 6 32 38
Orange 19 22 41 8.59
Total 25 54 79
Wheeler 10 31 41
Orange 19 22 41 4.30
Total 29 53 82
Note. (Ho:) states that the attribute of interior
color is independent of type. Chi square critical
value is 3.849 df (1), alpha level of .05.
Table 14. Chi Square Test of Exterior Color.
Gray Brown 2
Type fo fo Total X obs.
Norwood 11 34 45
Stallings 4 36 40 3.02
Total 15 70 85
Norwood 11 34 45
Wheeler 4 36 40 3.02
Total 15 70 85
Norwood 1L 34 45
Orange 7 35 42 76
Total 18 69 87
Stallings 4 36 40
Wheeler 4 36 40 .00
Total 8 72 80
Stallings 4 36 40
Orange 7 35 42 .77
Total 11 71 82
Wheeler 4 36 40
Orange 7 35 42 .77
Total 11 71 82
Note. (Ho:) states that the attribute of exterior color is independent of type. Chi square
critical value is 3.84, df (1), alpha level of .05.
Table 15. Chi Square Test of Interior Surface Finish.
Smoothed Filmed Roughened
Type fo fo fo Total X2 obs.
Norwood 28 2 17 47
Stallings 28 3 11 42 1.24
Total 56 5 28 89
Norwood 28 2 17 47
Wheeler 33 7 2 42 14.84
Total 61 9 19 89
Norwood 28 2 17 47
Orange 12 35 0 47 52.82
Total 40 37 17 94
Stallings 28 3 11 42
Wheeler 33 7 2 42 8.20
Total 61 10 13 84
Stallings 28 3 11 42
Orange 12 35 0 47 44.68
Total 40 38 11 89
Wheeler 33 7 2 42
Orange 12 35 0 47 30.31
Total 45 42 2 89
Note. (Ho:) states that the attribute of interior surface
finish is independent of type. Chi square critical value
is 5.99, df (2), alpha level of .05.
Table 16. Chi Square Test of Exterior Surface Finish.
Smoothed Filmed Roughened
Type fo fo fo Total X2 obs.
Norwood 27 7 13 47
Stallings 25 5 12 42 .14
Total 52 12 25 89
Norwood 27 7 13 47
Wheeler 34 6 2 42 12.49
Total 61 13 15 89
Norwood 27 7 13 47
Orange 12 34 1 47 33.80
Total 39 41 14 94
Stallings Wheeler Total
Stallings 25 5 12 42
Orange 12 34 1 47 41.06
Total 37 39 13 89
Wheeler 34 6 2 42
Orange 12 34 1 47 30.38
Total 46 40 3 89
Note. (Ho:) states that the attribute of exterior surface finish is independent of type. Chi square critical value is 5.99, df (2), alpha level of .05.
Table 17. Chi Square Test of' Paste Texture.
Vesiculated Grainy 2
Type fo o Total X obs.
Norwood 30 17 47
Stallings 29 13 42 .25
Total 59 30 89
Norwood 30 17 47
Wheeler 40 2 42 8.86
Total 70 19 89
Norwood 30 17 47
Orange 47 0 47 20.74
Total 77 -17 94
Stallings 29 13 42
Wheeler 40 2 42 9.80
Total 69 15 84
Stallings 29 13 42
Orange 47 0 47 17.07
Total 76 13 89
Wheeler 40 2 42
Orange 47 0 47 2.23
Total 87 2 89
Note. (Ho:) states that the attribute of paste texture is
independent of' type. Chi square critical value is 3.84,
df (1), alpha level of' .05.
Table 18. Chi Square Test of Paste Temper
Fiber Fiber/Clastic 2
Type fo fo Total X obs.
Norwood 12 35 47
Stallings 25 17 42 10.49
Total 37 52 89
Norwood 12 35 47
Wheeler 39 3 42 41.27
Total 51 38 89
Norwood 12 35 47
Orange 43 4 47 42.08
Stallings 25 17 42
Wheeler 39 3 42 12.86
Total 64 20 84
Stallings 25 17 42
Orange 43 4 47 12.33
Total 68 21 89
Wheeler 39 3 42
Orange 43 4 47 o6
Total 82 7 89
Note. (Ho:) states that the attribute of paste temper
is independent of type. Chi square critical value is
3.84, df (1), alpha level of .05.
Finishes on interior and exterior surfaces were recorded separately to determine whether or not surface finish is independent of type. Test results indicate that surface finish is dependent upon type (see Tables 15 and 16). This is true of both interior and exterior surfaces. Orange specimens are predominantly filmed and burnished, while Wheeler is smoothed, and Stallings and Norwood are roughened.
Determinations of paste temper were derived from the visual inspection of the interior, exterior, and core surfaces of each sherd. Recognition of two broad categories of temper is possible: (1) pure vegetal fiber-tempering, and (2) vegetal fiber-tempering with associated clastic inclusions. Fiber tracks observed in cross section were recorded as either flat or cylindrical. Table 18 presents the association of temper among the four plain types in question.
For many years it has been believed that the method of tempering
pottery has cultural importance (Griffin 1939:1160). Therefore, temper is an attribute that may be related to type identification. Traditionally, the consensus among southeastern archaeologists has been that Stallings pottery is tempered with Spanish moss (Sears and Griffin 1950), Wheeler pottery with some form of grass (Sears and Griffin 1950), Orange pottery with shredded palmetto fibers, and Norwood pottery with sand and Spanish moss (Phelps 1965). Unfortunately, the usefulness of temper as a discriminating agent is limited because the paleobotanical identification of these fibers remains enigmatic (Weaver 1963; Brain and Peterson 1971; and Simpkins and Scoville 1981).
From this sample it is apparent that the agent used to temper Wheeler pottery leaves a characteristic wide and flat imprint, which may be distinguished from the more cylindrical fibers used to temper the remaining three types examined. Not all fibers used to temper Wheeler vessels conform to this wide and flat shape, however. Approximately 30 percent of the sample possessed fiber tracks that were absolutely indistinguishable from those used in the manufacture of Stallings, Orange and Norwood Plain types.
It is also apparent from the results of this analysis that too much
attention has been given to the sandy nature of the Stallings and Norwood pots. The latter have often been set apart and labeled as "semi-fibertempered wae" (Bullen 1953:88), and as a result, many analysts aware of the succeeding temper horizon (sand) have been tempted to interpret Norwood pottery as late or transitional. It must be recognized that the semi-fiber-tempered designation is not exclusive to any one type, and, therefore should be reconsidered.
A quick perusal of the original plain type sherds formally described for all four series makes it apparent that the total sample contains a tremendous amount of variation in the range of clastic inclusions deemed acceptable within the paste of each plain type. The results of the temper
identifications make it very clear that Orange and Wheeler are associated in terms of their demonstrated preference for pure fiber-tempering, that
is fiber-tempering with very few if any appreciable amounts of clastic inclusions within the paste. However. the converse of this observation -the association of Stallings and Norwood because both have high percentages of sherds characterized by clastic inclusions--may not be accurate. The presence of particles of quartz within a sherd's paste may not reflect some culturally determined criterion of proper vessel construction but rather reflect the use of a local clay source that may contain these inclusions naturally. Thus, the real meaning of temper preference may be primarily a geological or environmental one rather than a cultural one, and therefore its not "temper" at all.
Fourteen of the nineteen attributes examined during this analysis
could not be interpreted as critical and therefore useful for discriminating one plain type from another. Such non-discriminatory attributes include: rim thickness (Table 6), rim diameter (Table 11), base thickness (Table 9), hardness (Table 10), core color (Table 12), interior color (Table 13), exterior color (Table 14), and paste texture (Table 17). In addition, the attributes of method of manufacture, rim form, lip form, base form, tooling marks, and total vessel form were not found to be instructive for distinguishing one plain type from another.
The objective of this analysis has been to identify critical ceramic attributes for formally discriminating between the Stallings, Wheeler, Orange, and Norwood Plain types of the sample. If the types as originally described are valid, then their attributes should separate into four discrete groups. This, however, did not occur.
Nineteen ceramic attributes were investigated and compared between the four plain fiber-tempered types of the sample. The morphological attribute similarities as noted on Table 19 far outweigh the differences. Table 19 demonstrates that a statistically significant degree of similarity and association exists among the morphological attributes shared by Norwood, Wheeler, and Orange Plain, but that the same high degree of similarity and strength of association does not extend to the Stallings Plain type. To a lesser degree five ceramic attributes do vary between types; these include: lip and body thickness, interior and exterior surface finish, and temper.
This preliminary study demonstrates that these attributes were not
significantly different in three out of four taxa for plain fiber-tempered pottery. It is suggested that this is so because the critical attributes historically used in delineating the variability among types and for
sorting these four series are imprecisely described, and thus they are seldom instructive for making comparisons between types. Because these critical attributes are not reproducible, both the writer and Stoltman (1974:19) have found it impossible to apply the currently accepted classification objectively. Thus, these test results suggest that the type distinctions that have been drawn historically (e.g., Sears and Griffin 1950; Phelps 1965) are open to question.
Table 19. A Comparison of Attribute Similarities
and Associations Between Types.
Method of Manufacture Rim Thickness Rim Form Rim Diameter Lip Thickness Lip Form Body Thickness Base Thickness Base Form
Hardness Exterior Color Interior Color Core Color Exterior Surface Finish Interior Surface Finish Tooling Marks Paste Textures Paste Temper Vessel Form
[Stallings (Stallings [Stallings [Stallings Stallings Stallings [S tailings Stallings S tailings
[Stallings [S tailings
[Stallings [Stallings [Stallings [Stallings [Stallings
Norwood Norwood Norwood] Norwood (Norwood (Norwood Norwood :Norwood Norwood Norwood Norwood Norwood Norwood Norwood Norwood Norwood
Orange Wheeler] Orange Wheeler] Orange [Wheeler] Orange Wheeler] Orange Wheeler] Orange Wheeler] Orange Wheeler] Orange Wheeler] Orange Wheeler] Orange Wheeler] Orange Wheeler] Orange Wheeler] Orange Wheeler] Orange Wheeler] Orange Wheeler] Orange Wheeler) Orange Wheeler]
[Stallings Norwood][Orange Stallings [Norwood Orange
Note. Brackets indicate either a high degree of association or the presence of similarity between the types enclosed.
The tests also indicate that there may be less variability within the SECT than previously thought. The morphological attributes of fibertempered pottery have a clinal distribution over the Alabama-FloridaGeorgia area. Thus, their usefulness for differentiating archaeological cultures is dubious, and a new classification may be in order. The SECT, regardless of how it has been divided in the past, is in actuality characterized by a remarkably homogeneous foundation consisting of nearly identical plain types. This analysis makes explicit what has always been implicit and therefore overlooked within the formal type descriptions of
this pottery: most of the morphological ceramic attributes utilized to discriminate between plain types approach identity. Quite possibly, the differences that do exist between these wares may be interpreted as
reflections of geological or environmental, rather than cultural, factors that limit the potter's choice.
In the past, analysts have failed to recognize the homogeneous nature of these four plain types, partly because of typological inadequacies and
partly becaUse of the practice of pigeonholing most of the fiber-tempered specimens found since the late 1960s into these four estaolisnei formal types. Thus the difficulty in accepting the cultural interpretation of the fiber-tempered tradition is attributable to the fact that this ceramic tradition was and continues to be derived from an intuitively generated typology well-grounded in subjective inference (Spaulding 1976:59-60).
As presently conceived, the types Norwood and Wheeler Plain are not
very useful as refined tools for the elucidation of time-space problems because they share both stylistic and technological attributes that readily identify them as Orange Plain. Phelps concedes that:
Norwood plain includes some sherds which would be
identified as Orange, Stallings or Wheeler Plain if
viewed out of regional context, but which are generally
rougher in appearance, less well made, than the "type"
sherds in other series (Phelps 1965:65).
The distributions of the types Orange, Wheeler, and Norwood Plain are known to overlap in both time and space. It is becoming increasingly more apparent that at their boundaries, they influenced the development of one another strongly, if not directly. Thus, it is proposed that these three types constitute a ceramic "complex" that may be objectively set apart from the Stallings sphere of influence to the north and east. A complex is synonymous with culture period in that it represents the group of pottery types or the various series of types that occur together in the same general area at the same time (Willey 1949:6).
From an analytical point of view, it has proven beneficial to combine the St. Simons, Sapelo, Bilbo, and Stallings Island ceramic wares of the coastal plain of Georgia and South Carolina into one taxonomic group, the Stallings ceramic complex (Griffin 1943; Stoltman 1972:52). It may prove equally beneficial to recognize the existence of a proposed Norwood-
Orange-Wheeler amalgamation. One taxonomic label, the Orange ceramic complex (Bullen 1955, 1971, and 1972), could be expanded to incorporate these Norwood and Wheeler types.
Sullen (1972:24) proposed an hypothesis that, if correct, would Degin to explain the apparent techno-stylistic similarity common to Norwood, Orange, and Wheeler Plain pottery. Central to his line of reasoning is his (Sullen 1969:42-44, 1972) interpretation of the last movement of the people who made and used Orange pottery into western Florida. The absolute dates known for Norwood and the relative dating assigned to Wheeler indicate that neither of these series reached their apex until well after this proposed westward expansion began.
Sullen (1959, 1971) introduced a concept known as the "Florida Transitional Period" to explain this movement. According to Sullen, pockets of ceramic regionalism existed in the southeast ca. 1200-500 B.C., when the people who made and used Orange pottery were gradually expanding westward from their Florida, St. Johns River valley homeland, across the 3ig Bend region, and into the lower Mississippi valley (Bullen 1972:24). This period was marked by considerable population growth, clear regional adaptations, and the apparent interregional spread of Formative lifeways. Walthall (1980:83, 92) refers to this time as the Middle Gulf Formative Period. It is not understood whether the Late Orange-Early Woodland transition that took place during this period was initiated through trade and culture contact with resident ethnic groups in surrounding areas, or by actual population movements onto the Gulf coastal plain and then upstream into Alabama and Mississippi. Only future investigations will demonstrate which of these alternatives, if either, is valid.
Obviously, much more research is imperative if we are to determine the exact relationships between these two early ceramic complexes and begin to delineate the path of their development. It seems that the recognition of the Stallings enterprise as an entity apart from a Norwood-Orange-Wheeler amalgamation is a start. Certainly such a revised view of the nature and extent of southeastern fiber-tempered wares, if correct, has important implications for tracking down and tracing the origin(s), development, and movement of Formative peoples.
The results of this preliminary analysis lead to the suspicion that there may be a great deal less variability within the Southeastern Fiber-tempered Ceramic Tradition than previously thought. This conclusion makes apparent the need for an integrative synthesis of all forms of material culture within the SFCT. Not until such a synthesis has been performed can the underlying regularities inherent in the cultural patterning of the SECT be fully revealed and a realistic definition and understanding of this tradition be obtained. If this study serves no other purpose it might at least demonstrate the inadequacies and difficulties inherent in the long-sought practice of defining
archaeological cultures solely upon the spatial and temporal coordinates of ceramic "index fossils." Problem-oriented classifications must be created in the future if we are to achieve our goal of identifying the cultural boundaries within this apparent ceramic continuum.
I take this opportunity to express my gratitude to the following individuals for their guidance and support with this reasearch. Drs. James B. Griffin, Jerald T. Milanich, and David S. Phelps made access to the original type sherds examined in this analysis possible. I offer a special word of thanks to Dr. Griffin for making available collections and facilities of the Ceramic Repository of the Division of Archaeology of the Museum of Anthropology, University of Michigan, Ann Arbor, for the purpose of this research. Various drafts of this manuscript have been read by Drs. William H. Sears, Moreau S. Maxwell, Charles E. Cleland, and William A. Lovis. I express my gratitude and sincere appreciation to all these individuals for their constructive comments and creative insights.
Atkins, S. and J. MacMahan
1967 The Zabski site, Merritt Island, Florida. Florida Anthropologist
Brain, J. and D. Peterson
1971 Palmetto tempered pottery. Southeastern Archaeological
Conference, Bulletin 13:70-76.
Bullen, R. P.
1953 Culture changes during the Fiber-tempered period in Florida.
Southeastern Archaeological Conference Newsletter 4:45-48.
1955 Stratigraphic tests at Bluffton, Volusia County, Florida.
Florida Anthropologist 8(l).
1958 Six sites near the Chattahoochee River in the Jim Woodruff
Reservoir area, Florida. Bureau of American Ethnology, Bulletin
1959 The Transition period of Florida. Southeastern Archaeological
Conference Newsletter 6:43-53, 59-62.
1961 Radiocarbon dates for southeastern fiber-tempered pottery.
American Antiquity 27:104-106.
1969 Excavations at Sunday Bluff, Florida. Contributions of the
Florida State Museum, Social Sciences 15.
1971 The Transitional period of southern southeastern United States as
viewed from Florida, or the roots of the Gulf tradition.
Southeastern Archaeological Conference, Bulletin 13:63-70.
1972 The Orange period of peninsula Florida. In: Fiber-tempered
pottery in the southeastern United States and northern Columbia: its origins, context, and significance. Florida Anthropological
Society, Publications 6:9-33.
Bullen, R. P. and A. K. Bullen 1976 The Palmer site. Florida Anthropological Society, Publications
Doran, J. E. and F. R.Hodson
1975 Mathematics and computers in archaeology. Harvard University
Ford, J. A.
1961 A quantitive method for deriving cultural chronology. Pan
American Union, Technical Manual 1.
Ford, J. A. and J. B. Griffin 1960 A proposal for a conference on pottery nomenclature for the
Southeastern United States. Southeastern Archaeological
Conference, Newsletter 7(l):5-21.
Griffin, J. B.
1939 Report on the ceramics of Wheeler Basin. In: An archaeological
survey of Wheeler Basin on the Tennessee River in northern
Alabama, by W.S. Webb. Bureau of American Ethnology, Bulletin
1943 An analysis and interpretation of the ceramic remains from two
sites near Beaufort, South Carolina. Bureau of American
Ethnology, Bulletin 133(22).
1945 The significance of the fiber-tempered pottery of the St. Johns
area in Florida. Washington Academy of Sciences Journal
Griffin, J. W.
1971 Prehistoric Florida: a review In Archaeology of the eastern
United States, edited by J. B. Griffin, pp. 322-334. Chicago.
Haag, W. G. (editor)
1939 Southeastern Archaeological Conference, Newsletter 1(I).
Jahn, 0. L. and R. P. Bullen 1978 The Tick Island site, St. Johns River, Florida. Florida
Anthropologist 31(4):part 2.
Jenkins, N. J.
1975a The Wheeler series and southeastern prehistory. Florida
1975b Archaeological Investigations in the Gainesville Lock and Dam
Reservoir: 1974. Report on file at Mound State Monument.
Milanich, J. T.
1971 The Deptford Phase: An Archaeological Reconstruction. Doctoral
Dissertation, University of Florida, Gainesville.
Milanich, J. T. and C. H. Fairbanks 1980 Florida Archaeology. Academic Press, New York.
Moore, C. G.
1894 Certain sand mounds of the St. John's River, Florida. Journal of
the Academy of Natural Sciences of Philadelphia, second series,
Munsell Soil Color Charts
1975 Munsell color. Macbeth Division of Kollmorgen Corporation,
Phelps, D. S.
1965 The Norwood series of fiber-tempered ceramics. Proceedings of
the 20th Southeastern Archaeological Conference, Bulletin
1966 Early and late components of the Tucker site, 8 Fr 4. Florida
1969 The Norwood phase of North Florida and its extra-regional
relationships. Paper presented at the 68th Annual Meeting of the
American Anthropological Association, New Orleans, Louisiana.
Reid, K. C.
1984 Fire and Ice: New Evidence for the Production and Preservation
of Late Archaic Fiber-tempered Pottery in the Niddle-Latitude
Lowlands. American Antiquity 49(l):55-76.
Sears, W. H.
n.d. NSF G-5019. An Investigation of Prehistoric Processes on the
Gulf Coastal Plain. ms. on file, Florida State Museum,
Sears, W. H. and J. B. Griffin 1950 Fiber-tempered pottery of the Southeast. In Prehistoric Pottery
of the Eastern United States (loose leaf series), edited by J. B.
Griffin. University of Michigan, Ann Arbor.
Simpkins, D. L. and D. A. Scoville 1981 Isolation and identification of fiber inclusions from a sample of
Stallings and Orange Series Ceramics. Paper presented at 38th
Annual Meeting of Southeastern Archaeological Conference,
Spaulding, A. C.
1976 Multifactor analysis of association: an application to Owasco
ceramics. In Culture change and continuity: essays in honor of James Bennett Griffin, edited by Charles E. Cleland, pp. 59-68.
Academic Press, New York.
Stipp, J. J., G. A. Knauer, and H. G. Goodell 1966 Radiocarbon Dating Laboratory, Department of Geology, The Florida
State University, Tallahassee, Florida. Radiocarbon 8:46-53.
Stoltman, J. B.
1966 New radiocarbon dates for southeastern fiber-tempered pottery.
American Antiquity 31(6) :872-874.
1972 The Late Archaic in the Savannah River region. In:
Fiber-tempered pottery in southeastern United States and northern
Columbia: its origins, context, and significance, edited by R.
P. Bullen and J. B. Stoltman. Florida Anthropological Society,
1974 Groton Plantation: an archaeological study of a South Carolina
locality. Peabody Museum of Archaeology and Ethnology,
Monographs 1. Harvard University, Cambridge.
Thomas, D. H.
1976 Figuring anthropology: first principles of probability and
statistics. Holt, Rinehart, and Winston, New York.
Walthall, J. A.
1980 Prehistoric Indians of the Southeast: archaeology of Alabama and
the Middle South. The University of Alabama Press, University.
Weaver, E. C.
1963 Technological analysis of prehistoric Lower Mississippi ceramic
materials: a preliminary report. American Antiquity 29(l):49-56.
Webb, W. S. and D. L. DeJarnette 1942 An archaeological survey of the Pickwick Basin in the adjacent
portions of the states of Alabama, Mississippi, and Tennessee.
Bureau of American Ethnology, Bulletin 129.
Willey, G. R.
1945 Horizon styles and pottery traditions in Peruvian archaeology.
American Antiquity 2:49-56.
1949 Archeology of the Florida Gulf Coast. Smithsonian Miscellaneous
Collections 113. Washington, D.C.
Williams, S. (editor)
1977 The Waring papers: the collected works of Antonio J. Waring.
Papers of the Peabody Museum of Archaeology and Ethnology 58.
Harvard University, Cambridge.
Analysis of Ceramics From
The South Prong Site (8-1-1-4 18),
Hillsborough County, Florida
Jeffrey M. Mitchern
The southern and eastern regions of the Central Peninsula Gulf Coast culture area of Florida (Fig. 1) are poorly known archaeologically (Milanich and Fairbanks 1980:24-26). They appear to be geographically and culturally transitional with the adjacent Okeechobee and Central Florida culture areas, and are poorly known because most previous archaeological work in the area has been concentrated on coastal sites, which often
include impressive earthworks.
During the 1970s, extensive phosphate mining in the area prompted a large number of cultural resource surveys in order to satisfy state and federal environmental impact legislation. In some cases, mitigative
excavations were carried out to salvage archaeological information before sites were destroyed (Welch 1983:20-21). Such excavations were rare, however, and were generally not funded at a level to allow in-depth analyses of the various artifact classes recovered. One survey in southeastern Hillsborough County by Martin (1976) located the South Prong I site adjacent to the South Prong of the Alafia River. The site consisted of a scatter of ceramic and lithic artifacts and debitage. On the basis of artifact types recovered from the surface and from test pits, Martin identified Archaic (6000-1000 B.C.), Weeden Island (A.D. 700-1000), and Safety Harbor (A.D. 1000-1650) components at the site (1976:22).
Plans to construct a phosphate mine utility corridor over the site
necessitated its excavation in 1982. This excavation was carried out on the basis of a research design composed of seven goals, which included: determining the spatial boundaries of the site, ascertaining what cultural components were represented, determining the nature of specific episodes
of occupation, determining site function, locating discrete activity areas, investigating patterns of resource procurement, and determining the relationship of the site to other sites in the region and in adjoining areas (Welch 1983:1). As part of the investigation, a ceramic technological analysis was conducted on the 264 sherds from both Martin's (1976) and Welch's (1983) work at the site. Of these, 98 are included in the present report (sherds smaller than 2 cm. in one surface dimension
Fig. 1. Map of the Central Peninsula Gulf Coast Culture area showing the location of the South Prong I site. (After Welch 1983.)
were eliminated). The assemblage of sherds was diverse, representing pottery from the Archaic, Weeden Island, and Safety Harbor periods, a
possible time span of 3650 years (the first pottery in Florida was made in the late Archaic around 2000 B.C.). Also, it appeared that more than one
archaeological culture area was represented in the pottery, as types normally found in the Circum-Glades area were present along with types commonly found along the Central Peninsula Gulf Coast.
The ceramic materials appeared to be concentrated in particular areas of the site. Some small clusters were noted, but in general the overall sherd distribution was very sparse. The presence of sherds in some portions of the site and their absence elsewhere is most likely due to changing intrasite settlement patterns as the site was reoccupied in different time periods. The excavations did not reveal any house patterns or other direct evidence of domestic structures. Welch (1983:158-159)
feels that the patterning of sherds indicates repeated short-term use of the site rather than long-term occupations.
Inl the context of the goals of the excavation's research design, five specific areas of investigation were chosen for study in the ceramic analysis. These are listed and discussed below.
1) One objective was to determine whether the ceramic remains from the
site can provide any data on temporal, cultural, or geographical affiliations with adjacent culture areas, most specifically the Central Peninsula Gulf Coast area to the west. This objective was addressed by a typological study of the pottery to see if recognizable non-local categories could be identified on the basis of paste composition and other
attributes used to segregate pottery types in central and southern Florida.
2) Another goal was to seek evidence of trade or contact with groups outside of central and south Florida. In addition to the identification of exotic pottery types as mentioned above, evidence of trade wares would be provided by the presence of inclusions which do not occur naturally in the physiographic region in which the site is located. The presence of such aplastics as sponge spicules or mica could be viewed as evidence of trade or outside contact if such inclusions could be shown to be absent in the vicinity of the South Prong I site and in the Central Peninsula Gulf
Coast and South Florida culture areas.
3) Determination of the use or function of individual vessels was another goal. This aspect of the study is based primarily on study of vessel form reconstructed from rim sherds (Braun 1980; Shapiro 1983, 1984) and the determination of types of use alteration such as sooting or oxidation coloration (Hally 1983b). Such data allow inferences to be made regarding site function.
4) Data on the technology of pottery production at the site, including manufacturing and finishing techniques were sought. Microscopic analysis and observation of aspects reflecting technological proc Iessses were
undertaken, as described in Shepard (1956) and Rye (1981). Refiring experiments were also performed to gain insight into temperature and
circumstances of original firing of the vessels represented at the site.
5) Finally, it was deemed important to record basic qualitative and quantitative data on clay sources and sherds to begin the formation of a data base for future comparative ceramic studies in the area. In addition to recording technological data for the sherds, an attempt was made to
collect and analyze local clay sources.
Clay Sample Study
In May of 1982, the South Prong I site was visited for the purpose of locating and sampling clay deposits which might have been used as raw materials by the aboriginal inhabitants. The collection and study of clays from archaeological sites can provide muchl information to aid archaeologists in making inferences about past behavior and activities at sites where pottery is found. The identification and measuremnt of aplastic inclusions in the clay and in pottery can be useful in determining whether certain vessels were locally produced or imported. Firing of the clays to different temperatures and/or with various tempering materials added not only aids in determining vessel origin, but may allow inferences to be made concerning technological choices made by the potter and cultural traditions present in the village or region (Rye
According to SCS Soil Survey maps (Leighty et al. 1958), the South
Prong I site lies primarily on alluvial land, which varies from fine sand to fine sandy clay. None of the directly adjacent soils are described as containing clay. Open excavation units on the site and on the edges of a borrow pit next to the site were checked for clay deposits, but none was encountered. The banks of the South Prong of the Alafia River, which is adjacent to the site, were also searched.
Along the northern bank, two samples were collected from layered deposits which appeared to contain clay. The layers were exposed by
shovel in eroded banks, and it should be noted that both sources were considered very poor quality at the time of collection. Initial manipulation of the samples in the field indicated that the clays were somewhat lean or "short," meaning that they appeared to contain so much sand that they would not be adequate for producing pottery.
Firing experiments were conducted on the two samples, consisting of forming wetted samples into bars and firing briquets in an electric kiln after drying (Mitchem 1983a:44-45). The results of these experiments demonstrated that neither of the deposits sampled were likely to have been utilized by the aboriginal potters as sources for the pottery analyzed here. Both samples contained large quantities of quartz sand, making them
The results of the above experiments do not prove that the site
inhabitants were not producing pottery at the site, however. It is likely that suitable clay deposits exist near the site, but were not located during excavation or survey.
The 98 sherds from the South Prong I site were examined at the Ceramic Technology Laboratory of the Florida State Museum. Each specimen was first subjected to microscopic examination using a 70x binocular microscope equipped with a micrometer. These examinations were performed to identify aplastic inclusions, determine frequency and predominant shape of different inclusions (using Wentworth's size classificaiton), and the relative frequencies of different sizes of the various inclusions.
Categories of aplastic inclusio-ns identified were quartz (Q), sponge spicules (S), mica (M), whitish (W), ferruginous (F), gray (G), and brownish (Br) particles. Predominant shape categories were subjective ordinal categories: angular (A), subangular (SA), subrounded (SR), and rounded (R). Frequency measurements were also subjective ordinal categories. For both frequency of different types of inclusions and relative frequency of different sizes, the terms abundant, common, scarce and very scarce were used. These terms, as well as those used for predominant shape, are relative terms which were based on shapes or frequencies encountered in other specimens of the sherd population under study.
Color measurements were also recorded for each sherd, using Munsell Soil Color Charts. To provide as much standardization as possible, all specimens were measured under identical lighting conditions (fluorescent
light) by the author. Colors were recorded from the exterior and interior surfaces, and from a freshly broken section of the core. Ranges of colors and interpolations were recorded when appropriate.
The thickness of each sherd was also recorded. Sliding calipers were used to measure the range of thickness in centimeters for each specimen. The results of the microscopic, color, and thickness measurements are presented in Appendix 1.
Macroscopic examination was undertaken to determine whether forming and finishing techniques could be ascertained from the sherds. Evidence of coiling, smoothing, burnishing, slipping, and surface decoration was sought. The results of these studies are discussed later in this paper.
In order to group the sherds into categories for comparison, each was assigned to a typological category on the basis of inclusions and other paste attributes. Most of these categories are commonly used by
researchers in the Central Peninsula Gulf Coast culture area of Florida and are described below.
Sand-tempered plain refers to those sherds lacking surface decoration and having large amounts of Quartz sand inclusions in the paste. This is not a formal category, but a miscellaneous residual class of sherds which do not satisfy the requirements for inclusion in any of the other typological groups. Formerly, sand-tempered plain sherds from this region of Florida were classified as Glades Plain or Glades Gritty Ware (Goggin 1939:37-38, 1944:1) by most researchers. These types, however, have proven to be too restrictive in terms of the variation present in plain pottery in the area, and also suggest connections with the Glades area of
southern Florida which may not always be present.
Sand-tempered plain with spiculite paste refers to sherds which would be considered sand-tempered plain, but exhibit microscopic sponge spicules present in the paste. These sherds lack the "chalky" feel which is characteristic of St. Johns and Belle Glade ceramics (see below).
St. Johns Plain was first defined by Goggin (1940) as Biscayne Chalky Ware. Later, Biscayne Plain was used as its name (Goggin 1944:5; Willey 1949:444). Researchers now refer to it as St. Johns Plain because it is found most abundantly in the St. Johns River drainage of east Florida. The predominant characteristics of this pottery are its "chalky" feel and very soft paste. Microscopically, sherds of this pottery exhibit an abundance of sponge spicules in the paste (Borremans and Shaak, this volume). The spicules occur naturally in the clay. Some sand may also be present. Though at least one researcher has suggested that the ware was produced solely in the St. Johns River Basin and exported elsewhere (Crusoe 1971:41), this hypothesis has not been adequately tested. The abundance and distribution of the ware all over Florida would argue against its being manufactured only in the St. Johns Basin. No spiculite clays have been recorded along the Gulf Coast, but this is probably due to inadequate search and sampling for such clays. In the present work, the St. Johns wares are considered locally made types.
St. Johns Check Stamped was first mentioned by Goggin (1940) as Biscayne Check Stamped. This definition was expanded by Ferguson (1951:27-28). It is a ware with "chalky" St. Johns paste which has a waffle-like design on the exterior. St. Johns Incised has been described by Griffin (1945:220) and Ferguson (1951:26-27); it is an incised version of St. Johns Plain. St. Johns Red on Buff was described by Goggin (1948:7); it is St. Johns Plain painted with red lines.
Belle Glade Plain, first described by Goggin (1944:5), is generally considered intermediate between St. Johns Plain and sand-tempered plain. It has a slightly "chalky" feel, but includes common to abundant amounts of quartz sand in the paste. Rims are often flat, and the Belle Glade Plain of the Glades area of southern Florida typically has marks on the exterior indicating that the vessel was shaped when leather hard by cutting or shaving off portions with a sharp edged tool (J. T. Milanich, personal communication, 1981). Examples from north of the Glades area also show this trait, but not as often (Griffin and Smith 1948:19-20). In the analysis of the South Prong I ceramics, the main criteria for classifying sherds as Belle Glade Plain were a slightly "chalky" feel and
a paste harder than St. Johns (due to sand inclusions). Flattened rims were another characteristic. It was probably locally made, though Luer and Almy (1980:212) feel that it was made in the Lake Okeechobee Basin and traded out.
Pasco Plain was first described by Goggin (1948:8-9). It is generally characterized-as a limestone-tempered ware, with pockmarked surfaces due to leaching of some of the aplastic particles. Recent work by Deming (1975:24) has shown that some of the inclusions are apparently fuller's earth (a siliceous clay) rather than limestone. None of the inclusions in the Pasco Plain sherds from the South Prong I site reacted to hydrochloric acid (HCl), so the inclusions are not limestone. Some of the sherds have pockmarked surfaces, however.
Norwood Plain includes both fiber-tempered and sand-and-fiber-tempered pottery found alng the Florida Gulf Coast (Phelps 1965). It is considered to be the Gulf Coast variant of Orange Plain (Griffin 1945:219-220). Fiber (probably Spanish moss) was added to the clay before firing. When the vessel was fired, the fibers burned, leaving vesicles in the paste.
The typological category for each sherd of the sample of 98 is designated in Appendix 1.
Refiring of Sherds
Ten sherds were utilized in a refiring experiment. The sherds were chosen on the basis of paste attributes, in order to obtain data about different paste types in the sample.
The 10 sherds were placed on a tile in an electric kiln and fired, beginning at 2500 C., for 15 minutes at increments of 5O' to a maximum temperature of 6500 C., making a total of nine temperature levels. After cooling, a piece was broken off each sherd and carefully labeled to ensure that the correct sherd and temperature were identified. In order to obtain fragments large enough to analyze, the refired sherd pieces were generally around 3 x 3 cm. This allowed about one square centimeter for
The main purpose of the ref iring was to get some idea of the range of temperatures used by the aboriginal potters to fire their vessels. This, is accomplished by looking at changes in color of the sherds after being fired at different temperatures. Because a great many factors can affect surface color changes (Shepard 1956:222-224), colors of the core are recorded and used in analysis. The rationale behind the method is that the temperature level at which a drastic change in core color is noted is probably just slightly above the original firing temperature. It should be noted, however, that the firing atmosphere, minerals present in the clays, and the length of time a vessel is fired can affect core color. Therefore, all of the probable firing temperatures must be viewed with these other possible factors in mind (Shepard 1956:103-105).
In order to arrive at a range of firing temperatures for a particular sherd, all of the refired fragments were examined in relation to the original sherd. Core colors were checked, and the lowest temperature level at which a color change was noted was recorded. If it was a drastic change, it was considered to have a high probability of being the level above which firing occurred. If the changes were more gradual, a range of firing temperatures would be a safer proposal.
Brief descriptions presented in Table 1.
well as others used in Mitchem (1983a, 1983b,
of the sherds used in the refiring experiment are For more detailed descriptions of these sherds as the present study, the reader is directed to
Table 1. Description of Sherds Used in Refiring Experiment.
SHERD # TYPOLOGY
375-4 513-31 569-7 594-8 646-1
StJPl Pasco Pl STP
Belle Gl STP STP(spic) Belle Gl
Rim, sone very coarse quartz inclusions; coiled. Crumbly, laminated paste; medium and fine sand abundant. Chalky with dark core; medium and fine sand common. Table 1 (continued). Whitish inclusions (not limestone); black core; surface pitted Rim; gritty paste; whitish inclusions; interior well smoothed while wet; black core Somewhat chalky; abundant medium and fine sand; surfaces appear slipped; core dark. Well-smoothed surfaces; parts of core very dark. Slightly chalky; abundant coarse and medium sand; rough surfaces; dark core.
Chalky; exterior slipped; fine and very fine sand abundant; dark core.
Chalky; fibrous inclusions; interior smoothed; black core.
Range 7.5 YR 2/U to 10 YR 5/4 Range 5Y 2.5/1 to 7.5 YR 6/6 Range 2.5YR 2.5/0 to lOYR 5/1
Range 2.5Y 5.5/4 to 2.5YR 3/0
Range 5Y 5/1 to 5Y 3.5/1
Range 7.5YR 4/4 to 2.5Y 2.5/0
STP StJPl Pasco
- St. Johns Plain P1 Pasco Plain
Belle Gl Belle Glade Plain STP (spic) sand-tempered plain with
RESULTS OF SHERD ANALYSIS
Microscopic and Macroscopic Analysis of Sherds
Evidence for the use of coiling as the method of forming was present on 21 sherds. This is not surprising, as most aboriginal pottery in Florida shows evidence of this forming technique. The evidence consists of coil fractures (fractures along poorly joined coils) and cross section views of the alignment of the particles and layers.
Evidence of several finishing techniques (see Shepard 1956:187-191) is present. The great majority of the 'sherds exhibit smoothing on one or both surfaces. In some cases, characteristics of the surface indicate whether the surface was wet or leather hard (a stage of drying where the water content has dropped below the plastic stage LRye 1981 :20-21 j) when smoothing occurred. Fourteen sherds show evidence of smoothing while the paste was leather hard and eight show evidence of smoothing on wet paste. Burnishing (which is usually performed on leather hard paste) is evident to some extent on five sherds. At least two of the sherds (1-594-83 and 23M) were definitely slipped, and several others may have been. Slipping can at times be difficult to distinguish in this pottery.
Two sherds (#307-12 and 581-1) exhibited mica in the paste. The mica was present in the clays from which the vessels were made. The sherds are fiber (and sand) tempered. Mica is rarely found in pottery from the area, and such vessels are generally considered non-local. Other types of aplastics noted in the sherd sample include quartz sand, sponge spicules, whitish particles, ferruginous particles, gray particles, and brown particles. Vesicular fiber tracks were also observed in fiber-tempered she rds.
The predominant shape of aplastics was recorded for all of the
sherds. Shape may be a useful criterion for use in determining whether inclusions were intentionally added as temper, or occur naturally in the clay. If the clays were deposited by water action, one would expect to find rounded, fairly fine, generally uniformly sized particles occurring as natural inclusions (since water tends to sort as well as erode particles). On the other hand, predominantly coarse, angular inclusions would suggest that they were intentionally added.
There are problems with this line of reasoning, however. First, many Florida clay deposits were not deposited by water action, but developed in place from leaching and othieTprocesses (N. Comerford, personal communication, 1985). Second, even if the above scenario was correct, potters might have added aplastics collected from a riverbed, which would be sorted and rounded. The lack of usable clay sources from the site prevents checking to see what shapes and sizes of inclusions occur naturally in the local clays.
A count of sherds containing quartz inclusions reveals that 80 (82%) of the 98 sherds have predominantly angular or subangular quartz
inclusions. The other 18 sherds (18%) have quartz inclusions which are predominantly round or subrounded.
Three (18%) of the 17 sherds with whitish inclusions contain
predominantly angular or subangular inclusions, while the remaining 14 (82%) contain predominantly rounded or subrounded inclusions. These whitish inclusions are not limestone. Most appear to be very small phosphate rock pebbles, which are very abundant in the region (the site was being mitigated due to anticipated impacts from phsophate mining). The 10 sherds which contained ferruginous particles, four with gray particles, and one with brown particles all contained predominantly round or subrounded particles. Mica and sponge spicules were not used in shape classification because they are always flat and rodlike, respectively.
The shape data suggest that the potters may have been adding quartz
intentionally at least part of the time, while the data on white particles suggest these inclusions tended to occur naturally in the clays. Unfortunately, the lack of clay samples precludes any testing of the hypotheses suggested by these data.
Size frequency data for quartz inclusions show that in more than hialf of the sherds, medium find sand (as defined by the Wentworth size classification) was common. This could mean either that potters intentionally selected these sizes as temper, or that these sizes naturally occur in the local clays. Table 2 summarizes the various frequencies of quartz grain sizes.
Table 2. Frequencies of Quartz Grain Sizes in Sherds.
WENT WORTH ABUNDANT COMMON SCARCE VERY SCARCE
2 T'granule) 0 0 0 2 (2'1'0)
3 (very coarse) 0 0 1 (1%) 37 (38%)
4 (coarse) 0 26 (27%) 28 (29%) 35 (3611)
5 (medium) 0 72 (73%) 16 (16%) 8 (8%)
6 (fine) 1 (1%) 62 (63%) 27 (27%) 8 (8%)
7 (very fine) 1 (1%) 34 (35%) 41 (42%) 18 (18%)
8 (silt) 0 6 (6%) 16 (16%) 31 (32%)
An attempt was made to determine vessel form by studying the 14 rim sherds in the sample. Rim profiles were drawn, and are shown in Figure
2. In addition, a rim diameter template was used to estimate the rim diameter of five of the sherds. The remaining nine rims were either too small or too irregular to estimate. The five estimates are as follows: #158-1, 52 cm.; #569-7, approx. 26 cm.; #20M, 28 cm.; #38M, 50 cm.; and #39M, approx. 36 cm. The majority of the vessels represented by the rim sherds depicted in Figure 2 appear to be simple open bowls or oeakers. Sherd #158-1 appears to represent a flattened-globular bowl with a chamfered lip. Its thickness and the estimated rim diameter indicate a large, heavy vessel.
563-1 404-10 20M
Fig. 2. Profiles of Rim Sherds.
In addition to determining typological placement of the sherds,
examining them microscopically, and attempting to determine vessel form and size, the sherds were examined for evidence of surface decoration and signs of secondary use. In the entire collection of sherds from the site, only two exhibit surface decoration. These are #593-12, a St. Johns Check Stamped sherd, and #48-7, which is a St. Johns Incised sherd. Two of the sherds exhibit some evidence of possible reuse after breakage of the vessel. Sherd #389-1 has one or possibly two grooves on the edge, and neither is recent (such as trowel or shovel marks). The sherd could either have served as a hone (for bone pins or some similar implements) or as some sort of pendant. Sherd #22M also exhibited a groove on one edge. None of the grooves are recent, because if they were, the dark core would show through. Reuse of sherds as convenient ad hoc tools was probably a common occurrence (Hally 1983a:176-177, 1983bTT2TTResults of Refiring Experiment
The results of the ref iring experiment varied in terms of their ease of interpretation. In some fragments, the temperature level of drastic core color change was obvious. Other fragments revealed a gradual lightening of the core, making accurate estimation of the original firing
level more difficult. Color measurements were taken only on those fragments which showed a significant color change (a significant color change is defined here as one which was apparent to the investigator's eye while viewing the fragments in sequence of increasing temperature level.) Further changes in color (above the initial change) were also recorded, as well as the temperature where all traces of dark coring were oxidized, and the temperature where no further changes were observed. A brief summary of the changes is presented in Appendix 2.
Four of the sherds (#158-1, 368-5, 569-7, and 646-1) were
sand-tempered plain. Three of these would appear to have been fired to about 350*-450 ,C., but one (#569-7) exhibited a much higher temperature level of 500*-600*C. This sherd contains abundant amounts of quartz sand, where as the others contain common amounts. This difference may have affected the changes in core color, possibly the result of a more open pore structure.
Two of the sherds (#375-4 and 32M) were St. Johns Plain. Both
specimens had a probable firing temperature of about 450*C. Color change in #375-4 was gradual, possibly due to the common quartz sand, as opposed to scarce amounts in #32M.
Specimens #594-8 and 23M were classified as Belle Glade Plain. These sherds exhibited different original firing temperature estimates. A 450*C. level was proposed for #594-8, whereas a range of 500*-550*C. was suggested for #23M. An interesting result of the ref iring of these two sherds was the finding that both had been slipped. Both the exterior and interior surfaces of #594-8 had been slipped, apparently with different
clays. The exterior slip tended to fire white, while the interior had a reddish or pink characteristic at temperatures above 500'C. The exterior of #*23M was also slipped. After refiring, clear borders were evident in cross section between the core and the slips of both sherds.
Specimen #513-31 was a Pasco Plain sherd. Its original firing
temperature appears to have been about 500*C. Specimen #5M1 was classified as sand-tempered plain with spiculite paste. An original firing temperature of 450*C. was proposed for this sherd, within the range of the sand-tempered plain sherds in the sample.
The results of the firing experiments indicate that the estimated
original firing temperatures varied from 3500to 6000C., with seven of the 10 sherds suggesting 450'C. as a possible temperature. Though the sherds represent the predominant paste types present in the total collection, the small number of sherds refired is not necessarily representative of the entire collection in terms of firing temperature. However, the data are the first for the Tampa Bay or south central Florida areas, and will be useful as comparative data for future ceramic studies in the region.
The pottery from the South Prong I site provides some insight into the behavior of the prehistoric inhabitants of the region. The basic typological study of the sherds indicated that examples of the earliest pottery type known from the region (Norwood Plain) were present at the site, as well as very late types such as St. Johns Check Stamped. This wide range of time periods, representing late Archaic (ca. 1000 B.C.) to Safety Harbor (ca. A.D. 1000-1650) periods, is generally in agreement with the interpretations of Welch (1983:1 "84-187) and Hemmings (1975) about the South Prong I and similar sites in the area being repeatedly occupied over very long periods, perhaps seasonally.
In terms of geographical and cultural relationships, the ceramics show more affinity with the Gulf Coast area, but the inhabitants probably interacted with Ockeechobee peoples as well. The Belle Glade Plain sherds at the site may have come from the area around Lake Okeechobee, but it is impossible to determine without clays from the vicinity of the South Prong I site.
The presence of inclusions in pottery which are known not to occur in the area is a good indication that the vessels were imported. The inhabitants may have imported clay or materials for tempering, but this would seem to be unlikely. Arnold (1975:192, 1985:50) suggests that potters tend to obtain their primary ceramic resources (aplastics, clay, and firing fuel) from within a five to seven kilometer radius of their residence. Evidence of trade or contact with groups outside of central and south Florida might be demonstrated if local clay sources could be located and studied. The two aplastic inclusions which would probably be most useful in trying to identify exotic wares would be sponge spicules and mica. Thus far, no spiculite clays have been recorded along the Gulf
Appendix I. Microscopic and Typological Descriptions.
Inc. Size fre. Shape Exterior
48-7 StJInc Q 5
6 S 8 65-3 STP Q 4
7 80-4 STP Q 4
8 81-12 StJPl Q 4
85-8 STP Q 4
8 102-8 STP(spic.) Q 3
7 S 8
114-I STP(spic) Q 4
8 S 8 158-1 STP Q 3
8 202-1 Belle Gl Q 4
S 8 261-13 STP Q 4
7 W 4
6 F 4 287-7 STP Q 4
8 263-6 STP Q 4
8 W 3
Range 5Y 3 to SY 5/1.
C o 1 o r
IOYR 3/1 7.SYR 2/0
/1 2.5Y 3.5/0 5Y 5/1.5
R 7.5YR 5/6 SY 3/1 Range 7.5R 2/0 0.9-1.1 cm.
to 7.5YR 6/6
A IOYR 7/3 10YR 4.5/1 2.5YR 2.5/0 0.7-0.8 cm.
A 5Y 3/1 IOYR 4/1.5 O10YR 5/4 0.8 cm.
SA Range 5Y 2.75/1 5Y 4/1 7.5YR 2/0 0.5-0.6 cm.
to 2.5Y 5/2
SA 2.5Y 4.5/2 2.5Y 6/3 2.5YR 2.5/0 0.5-0.6 cm.
SA Range 2.5Y 5/3 SY 3/1
to 7.5YR 3/0
Range 7.5YR 2/0 1.0-1.3 cm. to 10YR 5/4
R O10YR 6.5/4 2.5Y 4.5/2 7.SYR 2/0
A 5Y 4/1
SA 7.SYR 6/6
2.5Y 5/1 2.5Y 2/0
IOYR 5/1 Range 2.5Y 7/4 0.7 cm.
to 2.5Y 6/2
SY 3.5/1 2.5YR 2.5/0 0.6 cm.