Texas Beyond History
TBH Home
Tejas Main

Late Caddo, A.D. 1400-1680

"Soule photograph"
Famous "Soule photograph" of family compound of Chief Long Hat near Binger, Oklahoma taken around 1870 by William Soule. Many of the construction details match those described in Spanish and French accounts dating to the late 1600s as well as archeological patterns dating hundreds of years earlier including those uncovered at Late Caddo sites in several areas of the homeland. Courtesy Smithsonian Institution.
Avery Engraved bowl
Late Caddo period Avery Engraved bowl from Foster site, Miller County, Arkansas. From C.B. Moore, 1912.

Click images to enlarge  

Belcher Engraved Bottle
Late Caddo period Belcher Engraved bottle from Foster site, Miller County, Arkansas. From C.B. Moore, 1912.
late phases
Approximate extent of some of the phases archeologists have defined for the early 1500s Late Caddo world. (The people responsible for the Fort Coffee phase in the Arkansas Basin may not have spoken Caddo.) Some of the Late Caddo sites mentioned in the text are also shown. Adapted from Perttula, 2000, and other sources.
stone tools
Stone tools from Late Caddo site in northwest Louisiana. Courtesy Tim Perttula.
Teran map detail
This detail from the famous Terán map shows a hedge-ringed compound or farmstead where an extended family probably lived. There are two houses, two raised granaries, and a ramada (shade arbor).
Hatchel Mound in 1938
Hatchel Mound as it appeared in 1938 when WPA crews began excavations. This is thought to be the "templo" mound depicted on the Terán map. TARL archives.
Eli Moores site
Low mound at the Eli Moore site on the Red River west of Texarkana that is thought to have been part of the Upper Nasoni village depicted on the Terán map. Texarkana phase, Late Caddo period. Photo courtesy Tim Perttula.
"Ash Mound"
The "Ash Mound" at the Frankston phase A.C. Saunders site in 1931 prior to initial excavations there by A.T. Jackson from the University of Texas. TARL archives.

About 500 years ago, Caddo population reached an all-time high. Caddo ritual centers, villages, hamlets, and farmsteads existed in all but a few areas of the Caddo Homeland, such as parts of the middle Sabine Valley and upper Sulphur River basin. Such areas had been settled earlier and may have been abandoned because of localized climatic conditions, declining soil fertility, or, more likely, because they fell between the territories of rival groups. Early Spanish accounts make it clear that the Caddo world was made up of small competing societies (polities) who were never politically united. Instead, various groups formed alliances and friendship pacts to protect their own territories, provide access to trade routes and scarce resources, and to join forces to fight enemies. Lists of enemies complied by the Spanish show that Caddo groups sometimes counted other Caddo groups among their enemies. This fluid world of small competing societies and shifting alliances seems to fit well the archeological evidence of the Late Caddo period (ca. A.D. 1400-1680).

The ending date for the Late Caddo period of A.D. 1680 is chosen because soon thereafter the European invasion began to cause major, irreparable changes to Caddo societies. Archeologists sometimes use the term protohistoric to refer to the period after the first contact with European culture and before the period of sustained interaction. In the case of the Caddo, the first contact was with DeSoto's army in 1542-1543, but it would be another 143 years (A.D. 1686) before Europeans returned to the area. In the intervening protohistoric period, European diseases, horses, Old World plants, and trade goods likely began to reach some Caddo groups. But the real impact of these European introductions did not take hold until the late 1600s.

It is during the period between about A.D. 1400 and A.D. 1680 that we see the greatest number of distinct geographical clusters (territories) across much of the Caddo Homeland. These are relatively small areas about 60-100 miles long and 25-50 miles wide within which we find similar Late Caddo pottery styles and other ways of doing things. Archeologists call these geographic clusters phases and they probably each represent the combined territory of several allied groups that intermarried and shared close ritual, economic, social, and military ties. The accompanying maps show some of the phases that have been defined, but you should realize that these are just approximations that are only as accurate as the field information. In areas that have seen lots of archeological work, such as much of northeast Texas, the phases are reasonably well known. For many other areas, they are not.

In the Great Bend area of the Red River, Caddo archeological sites dating after ca. A.D. 1400 are included in the contemporaneous Belcher and Texarkana phases. Texarkana phase sites occur on the Red River northwest of Texarkana to the Arkansas/Oklahoma state lines, as well as on the lower Sulphur River. Belcher phase sites (ca. A.D. 1500 to the late 1600s) are distributed from about Fulton, Arkansas, to below Shreveport, Louisiana. Upstream from the Texarkana phase, the McCurtain phase represents another distinct Late Caddo cluster of related sites in northeast Texas and southeast Oklahoma.

In the Texarkana and Belcher phase areas, sites include large, permanent settlements with mounds and cemeteries, hamlets, and farmsteads. The mound centers were marked by the construction of earthen mounds that were used as temples, burial mounds, or ceremonial fire mounds. These settlements were inhabited by sedentary Caddo agricultural communities with complex societies led by individuals with high status who lived at the mound centers. Mound centers such as Belcher, Battle, Hatchel, and Cabe represent the larger villages or towns. Hamlets or farmsteads have been investigated at the Sherwin and Atlanta State Park sites for the Texarkana phase and the Cedar Grove and Spirit Lake for the Belcher phase.

Both the larger communities and the smaller hamlets/farmstead had pole and grass structures (usually circular), outdoor ramadas or arbors (for shade), household cemeteries, and middens filled with household refuse. The 1691 Terán de los Ríos map of the Nasoni village on the Red River documented the likely character of individual farmstead compounds and the layout of a dispersed village. In northwestern Louisiana, Belcher phase sites are distributed along Red River channels and meanders that are no longer active today (the river's course shifted repeatedly in historic times). Cowhide Bayou and Red Chute Bayou in the Shreveport area are good examples of once-active channels along which Late Caddo sites are found.

Texarkana phase sites appear to date from ca. A.D. 1400-1700, and mound construction seems to coincide with the main period of settlement at the Hatchel site, from about A.D. 1400-1700. Both archeological and historical information suggests that the main platform mound (with many levels of structures and features) at the Hatchel site may be the templo or temple mound shown on the 1691 Terán map of the upper Nasoni village. The Horace Cabe mound may also have been occupied at this time, as well as the Eli Moores mound site, and the Moore/Higginbotham mound center and community a few miles downstream from the Horace Cabe site probably represents the upper Cadohadacho village that was eventually abandoned in 1788.

Like other Caddo groups on the Red River, the McCurtain phase settlement pattern includes numerous habitation sites (with household cemeteries and substantial midden deposits) and mound centers. The mounds appear to have mainly been constructed between ca. A.D. 1300-1500. In some instances, mound centers were not directly associated with permanent settlements or middens. The McCurtain phase mounds were generally constructed in one or two stages over important public structures, with the structure abandoned, dismantled and/or burned, then capped with a fill zone of soil scraped from nearby deposits. Simple and elaborate single and multiple burials were also placed in the mounds, as with the East Mound at the Sam Kaufman site on the Red River.

Caddo settlements along this stretch of the Red River seem to closely resemble those depicted in the Terán map and Soule photograph; Caddo villages were composed of individual compounds of houses and other structures associated with mounds and the residence of a caddi or chief. The density of McCurtain phase sites indicates that greater numbers of people are living in closer proximity than before. At the Sam Kaufman site the mound in McCurtain phase times was used as a place for the burial of the social elite. A shaft tomb containing 10 individuals and many grave goods was found near the center of the mound, which had been excavated through the remains of a burned temple structure. Special purpose salt-processing sites, such as the Salt Well Slough site (41RR204) are also common in the vicinity of the Sam Kaufman site.

The Frankston phase is comprised of farmsteads, hamlets, and small villages in the Neches and Angelina river basins in East Texas that date from ca. A.D. 1400-1680. The only Frankston phase site known to have mounds is the A. C. Saunders site. Excavations there in the early 1930s by crews from the University of Texas led by A. T. Jackson provides us with a unique glimpse of Late Caddo ritual life. The first mound (Feature 1) itself was not very high (2.2 meters or 7 feet) but it was broad covering an area about 26 by 33 meters (85 by 107 feet). Trenches into it showed that it was made up of thick deposits of ash capped by a clean layer of sandy loam; very few artifacts were recovered in the excavations there.

Belcher Engraved bowl
Late Caddo period Belcher Engraved bowl from Foster site, Miller County, Arkansas. From C.B. Moore, 1912.
Late and Protohistoric sites
Some of the better known and most important sites dating to Late Caddo and Protohistoric times, about A.D. 1400-1680. Graphic by Dee Ann Story.
Frankston effigy bowls
Frankston phase effigy bowls. TARL archives.
Belcher phase artifacts
Belcher phase artifact assemblage. From Webb, 1959.
Teran Map
"Terán Map" of Upper Nasoni settlement on Red River, produced by the Spanish expedition of 1691-1692, led by Terán del Rio. This village was part of the Cadohadacho alliance. The map shows that the community consisted of small farmsteads or extended family compounds, each depicted as being surrounded by hedges or rows of trees or bushes. The settlement shown is believed to have stretched along several miles of the Red. The "templo" mound is believed to be the archeological site known as the Hatchel mound.
Hatchel Mound today
Hatchel Mound as it appears today.
Cabe Mounds site
The Cabe Mounds site was one of the major ritual centers along the Red River during the Texarkana phase and probably represents part of the upper Cadohadacho village that was eventually abandoned in 1788.
animal bones
Animal bones (mostly deer) from 1930s excavation of a Frankston phase midden at the A.C. Saunders site in the Neches Valley of northeast Texas.
bone and shell tools
Tools made from animal bone and clam shells from the A.C. Saunders site. TARL archives.
A. C. Saunders site
Excavations in progress of the midden and big building at the A. C. Saunders site in 1935.
reassembled bowls and jars
Partially reassembled sections of large bowls and jars from the midden at the A. C. Saunders site.
Ballard Brushed jar
Tall Bullard Brushed utilitarian jar. Frankston phase, 41AN12.

Near the mound was Feature 2, a large midden mound that capped a large circular structure (13.5 meters or 44 feet across) with a large doorway (gap in the posts) facing east. Within the building was one hearth and there was a larger one at its entrance and several others outside the building, some of which were used after the building had been dismantled and capped with incredibly rich midden deposits containing fragments of an estimated 1291 pottery vessels (dominated by Poynor Engraved carinated bowls, Bullard Brushed and Maydelle Incised jars, and LaRue Neck Banded jars), many fragments of elbow-shaped ceramic pipes, Perdiz arrow points, stone drills, mussel shell digging tools, and an assortment of bone tools (awls, needles, and beamers) as well as some shell columnella beads. Large quantities of animal bones, apparently mainly deer, were found in the midden. Unfortunately, the archeologists did not save the animal bones, other than the bone tools, as was typical practice at the time.

The Feature 1 deposits may represent a fire temple, similar to those described a few centuries later by the Spanish not far to the south among Hasinai groups. The best evidence for this is the massive ash accumulation itself and the fact that the ash pile was purposefully sealed, as special buildings often were; however, the limited UT investigations did not identify any structures in Feature 1. The Hasinai fire temple was said to have a large central hearth, which was not the case here, but there is no reason at all to believe that each Caddo group built and arranged their fire temples the exact same way.

The rich midden mound deposits with so many smashed pots, smoking pipes, and animal bones is exactly the kind of deposit we might expect at places where communal feasts were held. The Spanish noted that successful fall harvests occasioned major festivals at the principal villages that drew kinfolk and allies near and far for several-day celebrations of Caddo life. Such events would have involved feasting, special ceremonies, tobacco smoking, dancing, performances of stories and songs, trading, negotiating, courtship, and many other activities. Of course, we cannot say for certain that this is what took place at the A.C. Saunders site, but such an interpretation is consistent with what was found there.

circular pattern of posts
This large circular pattern of posts defined a large building that may have been a fire temple.
smoking pipe fragments
Fragments of elbow-shaped, ceramic smoking pipes from the midden at the A.C. Saunders site. Tobacco smoking was part of many rituals. As with the sherds, many of the pipe fragments look pristine, as if they had been intentionally broken. TARL archives.
elbow pipes
Ceramic elbow pipes from the A.C. Saunders site. TARL archives.
Frankston phase pottery
Frankston phase pottery, Late Caddo period. TARL archives.
saline pool
Saline pool at the Hickman Salt site in the Saline River Valle. Salt coats the vegetation in the foreground. Arkansas Archeological Survey photo. Courtesy Pictures of Record.
Mound A
Mound A at the Ferguson site in Little Missouri River drainage of southwest Arkansas before excavations began in 1972. Arkansas Archeological Survey photo. Courtesy Pictures of Record.
Mound A
Close up view of the cross-section of a ritually dismantled and burned temple (one of several) found within the south part of Mound A at Ferguson. Courtesy Pictures of Record.

Mid-Ouachita and Social Hill phase sites in the Ouachita and Saline valleys of south-central Arkansas include specialized saltmaking communities such as the Hardman site. Excavations there under the direction of Ann Early uncovered evidence of a long-lived Late Caddo and protohistoric settlement including a fenced compound such as the one shown in the accompanying reconstruction. The Caddo people who lived at this site along Saline Bayou in the Ouachita River valley collected salt brine from natural seeps and boiled it down in heavy flat ceramic pans. The salt was probably traded mainly to non-Caddo corn farming peoples living in great numbers in the nearby valleys of the Mississippi and lower Arkansas rivers. People depending on maize (corn) for much of their diet would have required additional salt in their diet. Thus, it makes sense that all of the salt-making sites known in the Caddo Homeland date to relatively late times (Middle and Late Caddo periods) after corn became king.

The excavation of a large two-part platform mound dating to about A.D. 1300-1500 at the Ferguson site in the Little Missouri River drainage of southwest Arkansas exposed a series of temples and houses built in successive stages. Temples were erected at the higher south end of the 150 by 70-foot Mound A and houses (perhaps of priests or leaders) at the north end. Excavations were done at the Ferguson site in the early 1970s by the Arkansas Archeological Survey with the assistance of the Arkansas Archaeological Society under the direction of Frank Schambach. The accompanying photographs show some of the extremely well preserved remains of ritually dismantled and burned temples. The Ferguson site spans the latter Middle Caddo and earlier Late Caddo periods.

Late Angelina phase sites in the Lake Sam Rayburn area seem to date after ca. A.D. 1450. The Walter Bell site had small midden deposits, three circular structures ranging in diameter from 6.4-12 m in diameter. This site also had a small cemetery with extended and flexed burials of children and adults. Funerary objects placed with the deceased included one to six ceramic vessels per individual, Perdiz arrowpoints, conch shell beads, deer ulna tools and deer meat offerings, mussel shells, and engraved bird bone flutes.

Artist Ed Martin's depiction of saltmaking based on historic accounts and archeological evidence from the Hardman site in the Ouachita Valley of Arkansas.
reconstructed salt pan
Partially reconstructed salt pan from Caddo saltmaking site near Arkadelphia, Arkansas. Arkansas Archeological Survey photo. Courtesy Pictures of Record.
Mound A
Cross-section through the south end of Mound A at the Ferguson site. Courtesy Pictures of Record.
two circular houses
Outlines of two small circular houses with extended entranceways.
carbonized logs
Carbonized logs along the east wall of the burned temple shown above. Late Caddo Ferguson site, southwest Arkansas. Arkansas Archeological Survey photo. Courtesy Pictures of Record.
burned temple
Overview of another of the burned temples at the Ferguson site. This rectangular building was oriented on cardinal directions. Courtesy Pictures of Record.
House #2
House #2 (about 21 feet in diameter) at the Walter Bell site beneath what is today Sam Rayburn Reservoir in Sabine County, Texas. Angela phase, Late Caddo period. TARL archives.
post molds
Post molds from two central posts in the center of House #2 at the Walter Bell site. TARL archives.
Angelina phase pottery vessels
Angelina phase pottery vessels from burials at the Walter Bell site in Sabine County. TARL archives.

The Titus phase (ca. A.D. 1430-1680) represents the archeological remains of related Caddo groups who lived between the Sabine and Sulphur rivers in the East Texas Pineywoods and Post Oak Savannah [see next section]. These Caddo peoples lived in dispersed year-round settlements where they farmed and hunted, buried their dead in planned cemeteries, and manufactured culturally distinctive ceramics of considerable stylistic and functional diversity. The same may be said for the contemporaneous Frankston phase groups in the Upper Neches and Angelina River basins, and Caddo groups living in the Angelina and Attoyac River basins.

These Pineywoods and Post Oak Savannah Caddo groups were like the Cadohadacho groups on the Red River: relatively small-scale societies led by hereditary leaders and priests. There are considerable archeological similarities in patterns of mound construction and burial practices among the Titus phase groups and the Belcher, McCurtain, and Texarkana phase groups along the Red River. Unfortunately, the Caddo groups in the Sabine and Cypress drainages were only cursorily described in early historic times before they disappeared, probably being absorbed into the well known Cadohadacho and Hasinai groups to the north, east, and south.

Follow Caddo History

bird bone flutes
These two bird bone flutes were found in one of the burials excavated at the Walter Bell site. Triangular designs were incised into the delicate bone. On the right is a detailed photo and drawing of the design on the intact flute. TARL archives.