OF HEARTHS AND HOUSES

One goal of the 1993 Field School was to explore the area around Feature 3 looking for evidence of a house . This was accomplished by opening a 6 meter by 6 meter unit referred to as Block I . Feature 3 had been excavated in 1992 and found to be a 1.2 meter in diameter, round, basin shaped pit containing a large amount of daub, bone , and Caddoan pottery sherds. Near the bottom of the pit was a zone of ash. Charcoal and mussel shell from Feature 3 yielded three calibrated radiocarbon dates of about 1425 AD (Middlebrook 1993).


INTRODUCTION
During the 1993 East Texas Archeological Field School conducted at the Tyson site (41SY92) in western Shelby County, the junior author had an opportunity to participate in the excavation of a Caddoan hearth. The work was directed by Linda Lindsay , a graduate student in Anthropology at Southern Methodist University. This paper describes our findings and a few features of hearths and houses.
One goal of the 1993 Field School was to explore the area around Feature 3 looking for evidence of a house . This was accomplished by opening a 6 meter by 6 meter unit referred to as Block I . Feature 3 had been excavated in 1992 and found to be a 1.2 meter in diameter, round, basin shaped pit containing a large amount of daub, bone , and Caddoan pottery sherds. Near the bottom of the pit was a zone of ash. Charcoal and mussel shell from Feature 3 yielded three calibrated radiocarbon dates of about 1425 AD (Middlebrook 1993).
When Block I was completely exposed, a number of other pits and postholes were seen in plan view ( Figure I). Our activity focused on Feature 9 on the western edge of Block l . This I.I 5 meter by O. 9 meter oval hearth was first revealed at 20 cm depth when ash was encoun-tered ( Figure 2). The feature contained large amounts of ash from in situ burning, nuggets of fired clay, a small amount of bone, and several burned sherds with ash adhering to their surfaces. The hearth was slightly basin-shaped and approximately I 5 cm thick. A discontinuous thin layer of bright orange clay near its bottom was observed. The hearth had been prepared for use by digging a very shallow pit but no intentional "clay liniJlg" was seen. Two large postholes were found in the area of Feature 9. Feature 17 was discovered beneath the eastern end of the hearth. It was 30 cm in diameter and bad a smoothly rounded bottom at 75 cm below ground surface. Feature 12 was a very distinct posthole of similar proportions just west of the hearth. The diameter of F 12 was 27 cm and the depth was 65 cm below surface.
How do we understand this feature? Specifically, does Feature 9 represent the central hearth of a Caddoan house? This question is currently difficult to answer because the outside wall of a putative house has not been identified. Possibly, Block I lies entirely inside a large house. The question may be easier to answer after reviewing accounts written by early Europeans visiting the area and reviewing the archeological findings at other East Texas Caddoan sites.

ETHNOHISTORIC RECORD
Early Spanish writers described for us the amazingly rapid construction of Caddoan houses; these buildings were often completed in less than a day by many members of the community su-11 pervised and directed by their governing elites (Bolton 1987;Swanton 1942;Wyckoff & Baugh 1980). The initial stage of construction involved the placement of tall poles in a circle around a  removable central post or tree and then pulling the poles together at the top. Espinoza (Hatcher 1927;quoted in Jackson 1936) discusses this: The laths are placed in a circle and in the middle they put up a very tall pole with knots on it for climbing. Two Indians are placed on top ... They continue to tie them until they have formed a figure like a half-orange. . . They work so dexterously that a little after midday they are finishing the hut. . . The building finished, they cut the middle post at the bottom and the building is thus left standing . . .

.F27
There are additional indications in early accounts that the location of the removed central post was subsequently used for a fireplace. While references are made to fires inside ordinary Caddoan houses, the most detailed descriptions involve the fireplaces inside the residences of important religious figures and in special temples. Fire had possibly both functional and sacred significance to the Caddos of East Texas. The homage paid to fire is reflected in these quotations from Morfi (Chabot 1932 in Jackson 1936) and Hidalgo (Hatcher I 927 in Jackson 1936 ): Both the former and the latter fear that the fire will get angry with them; and to < ..,... satisfy it, they make offerings to it of their first corn, tobacco, meat from the chase; and, in a word, all of their crops . . When in need they always appeal to the fire And:

13
The fire the Tejas Indians have in their houses was brought from the house of their high priest.
If the fire goes out they start immediately for the house of the priest to get new fire . It never goes out in the house of sacrifice .. .
In summary, then, the ethnohistoric record suggests that at least some Caddoan houses had a central hearth overlying a central posthole, and that fire had special significance binding each house to a perpetual fire at some religious center. Parenthetically, the practice of placing offerings in a central hearth may be illustrated at the Henry M Site ( 4 lNA60) along Bayou Loco in western Nacogdoches County (Middlebrook 1988). Henry M is a small Allen phase farmstead probably dating to the early part of the 1700s and contemporaneous with the nearby and better known Deshazo site (41NA27). Excavations to date have indicated the presence of at least one Caddoan house with a large, central posthole covered with the ash of a possible hearth. Interestingly, a well preserved Busycon sp. scoop, shaped by a groove and snap technology, was bw-ied within the ash directly over the posthole (Figure 3). A valuable implement obtained through trade would not have likely been randomly tossed aside.

ARCHEOLOGICAL FINDINGS
In the sandy soils of East Texas, hearths are often evidenced by the presence of ash (perhaps mixed with charcoal) and, frequently, burned clay. Ash is produced by the nearly complete burning of plant material and is composed of extremely alkaline potassium carbonates. Ash can reflect burning in place or the deposition of ashes from another location. Archeologically, ash may be found in several different contexts; a few examples include: (1) prepared clay hearths; (2) 14 the scatter of embers along a living surface; (3) large ash filled pits that may have functioned for cooking, pottery making, or refuse disposal; ( 4) smaller ash and smudge pits possibly used in smoking animal products, "repelling insects" , food preparation, or warming houses; (5) ash on floor of extended entranceway to ceremonial building; (6) central or peripheral fireplaces within ceremonial building; (7) fallen structural elements of a burned building; (8) an ash mound formed from the cumulative disposal of sacred ashes of a "fire temple" ; and (9) the site of cremation burials. Table I presents a summary of hearth, ash, and architectural structure data from an extensive sample of excavated Caddoan sites in Deep East Texas (DET; that portion of the Caddoan area south and east of the Sabine River). Table 2 displays for comparison a limited sample of sites in the nearby regions of Northeast Texas and Northwest Louisiana.
The DET data in Table I demonstrates the restricted nature of archeological information available concerning Caddoan house construction and site organization in the area. Initially , 121 structures excavated in the region appears to be an impressive number. However, 104 (86%) of these structures were located within only 3 (6%) of the 52 sites reviewed. Our perception of DET Caddoan architecture is dominated by, and therefore easily distorted by, findings at a single Early Caddo site (George C. Davis with 51 structures), a single recently excavated possible Middle Caddo site (Oak Hill Village with 42 structures), and a single Historic Caddo site (Deshazo with 9 structures). Only 11 ( or 21 % ) of the DET sites revealed hearths during excavation although other features containing ash or signs of burning were relatively common. While 24 (48%) of the 50 hearths were shown to be "central" hearths within houses, all the central hearths were reported from just four DET sites: George C . Davis, Oak Hill Village, Bryan Hardy, and HeruyM.
The central hearths found at the George C. Davis site in Cherokee County were under Mounds A and B and in the village areas. Newell and Krieger (1949:24) noted: Fireplaces were found in 14 of the 34 (house) outlines, always approximately in the center. . . A large central posthole was present in more than half the outlines, usually at about the center of 15 Volume 6, Number 4 the fireplace when the latter was present. Some extended through the fireplace, but others had definitely been cut off before the fireplace was made ... Spock (1977) carefully reanalyzed Davis site architectural features. She divided the 51 structures in her study into categories of "domiciles" and "special function structures" . Interestingly, only 2 of the 30 domiciles had central hearths while 14 of the 21 special function structures bad them. Twelve of the 16 central hearths at George C. Davis had associated center post (Spock 1977:30, Table I).
The findings of Spock ( 1977) might suggest that at least in some Caddo mound sites the use of a central hearth was less likely in nhabitations of the populace" than in more complex structures used in the religious or ceremonial life of the community. The observations at George C. Davis are not inconsistent with those of Webb ( 1959) who excavated eight houses in two Belcher site mounds. Seven of the eight houses had a central hearth or ash bed. Three were clay lined and one was found to have a central posthole beneath it. Large ash beds, possible the residue of ceremonial fires, were found around the interior periphery of Houses 6 and 7.
The construction of a large "ash mound" at the A.C. Saunders site is almost certainly related to the five nhearths" associated with a large building thought by A.T. Jackson to be a "fire temple" (Kleinschmidt 1982). The close relationship between central hearths and special use structures was also supported by White side's ( 1958) excaation at the Bryan Hardy site, where he found a house with an extended entranceway and a central hearth under a low mound. The recent excavations at the Oak Hill Village site in Rusk County (Cruse 1995;Perttula, personal communication) are somewhat more problematic because the function of the site bas not been fully established. Forty-two structures surround an apparent plaza. Four circular buildings (Structures I, 7, 9, and 17) contain central hearths over  central postholes. Two rectangular buildings (Structures 38 and 39), probably related to an earlier occupation of the site, had hearths not located centrally to the structures and not over posts. The only central hearth in the reviewed DET studies that seems to be completely unrelated to a possible ceremonial or special function site is the one excavated at the Henry M site (Middlebrook 1988).
Most of the remaining reports in this review failed to associate hearths with houses due to the small areal extent of excavation or other factors negatively influencing posthole detection. This is certainly not the case, however, at the extensively studied Deshazo site in Nacogdoches County, where three hearths were uncovered in the vicinity of overlapping structures in Unit 1 (Story 1982). Hearth 2 was likely a small utilitarian fireplace about midway between the center post and the exterior wall. Hearth 3 is a feature with a posthole underneath, but not clearly associated  with any structure. Its use is uncertain. Hearth 1, near several center posts, may not have been associated with any of the three houses. Carolyn Good (1982) speculated that this feature may well have represented the locale of an all-night bonfire in the "annual renewal" ceremonies described by early Spanish writers. Belcher site mounds. Seven of the eight houses had a central hearth or ash bed. Three were clay lined and one was found to have a central posthole beneath it.Large ash beds, possibly the residual of ceremonial fires, where found around the interior periphery of Houses 6 and 7.
The construction of a large "ash mound" at the A. C. Saunders site is almost certainly related to the five "hearths" associated with a large building thought by A. T. Jackson to be a "fire temple" (Kleinschmidt 1982). The close relationship between central hearths and special use structures also be supported by Whiteside's (1958) excava-tion at the Bryan Hardy site where he found a house with an extended entranceway and a central hearth under a low mound. The recent excavations at the Oak Hill Village site in Rusk County (Cruse 1995 ;Perttula, personal communication;ed. note: see pp. 23-25) are somewhat more problematic because the function of the site has not been fully established. Forty two structures surround an apparent plaza. Four circular buildings (Structures I,7,9,and 17) contain central hearths over central postholes. Two rectangular buildings (Structures 38 & 39) probably related to an earlier occupation of the site had hearths not located central to the struc-tures and not over posts. The only central hearth in the reviewed DET studies that seems to be completely unrelated to a possible ceremonial or special function site is the one excavated at the Henry M site (Middlebrook, 1988).

Volume 6, Number 4
Most of the remaining reports in this review failed to associate hearths with houses due to the small areal extent of excavation or other factors negatively influencing posthole detection. This is certainly not the case, however, at the extensively studied Deshazo site in Nacogdoches County where three hearths were uncovered in the vicinity of overlapping structures in Unit 1 (Story 1982). Hearth 2 was likely a small utilitarian fireplace about midway between the center post and the exterior wall. Hearth 3 is a feature with a posthole underneath, but not clearly associated with any structure. Its use is uncertain. Hearth I, near several center posts, may not have been associated with any of the three houses. Carolyn Good ( I 982) speculated that this feature may well have represented the locale of an all-night bonfire in the "annual renewal" ceremonies described by early Spanish writers.

CONCLUSIONS
In summary, the reviewed archaeological findings in Deep East Texas Caddoan sites, modestly but inconsistently, support the predictions regarding hearths based on the ethnohistoric record . Major gaps still exists in our understanding of fireplaces. Hearths are certainly not created equal; specifically , they vary in size, shape, location, preparation, function, duration of use, and contents. There may be limitations to applying infonnation about other Caddoan hearths to the Tyson site. Nevertheless, our working hypothesis that Feature 9 represents a central hearth is appealing because of its large size, associations with large postholes, and proximity to other likely interior house features. The structure here may have been the residence of an important local Caddoan political or religious figure given the very elaborate grave offerings associated with two juvenile burials adjacent to Feature 9 (Middlebrook 1994 ). Careful analysis of the fauna}, paleobotanical, and ceramic contents of the hearth is scheduled and may reveal more clues to Feature 9's function.