Skip Navigation

Cultural Resources - Grassy Lake Site in Ballard County, Kentucky



The Grassy Lake Site (15Ba144) is located in Ballard County, in far western Kentucky.  The site covers 3 hectares (ha) or 7.4 acres of land located on floodplain known as the Barlow Bottoms.



Limited investigation of the southeast corner of this site generated a good sample of materials from sealed deposits.  This paper will describe the sites location, field investigations, materials recovered and conclude with a discussion contextualizing Grassy Lake relative to nearby terminal Late Woodland and Early Mississippi sites. 

Grassy Lake is one of the many lakes in old river channels on the Barlow Bottoms, a dynamic landscape that has been constantly forming since people first entered the US.  The Barlow Bottoms is more than 8 km wide, stretching from the dissected upland loess hills to the Ohio River and more than 20 km north and south.

The lake is 3 km north of the confluence of the Mississippi and Ohio Rivers.  The entire bottom has been timbered and farmed over the past 200 years.  Prior to the 1937 flood there were numerous small farmsteads and crossroad communities scattered throughout the bottoms but after the depression, the flood, and WW II the bottoms were largely abandoned and the land used for wildlife refuges, timber production, and intensive large scale farming operations.   In the fall of 2010 the Kentucky Archaeological Survey and Kentucky Natural Resources Conservation Service as part of a Southern Appalachian Cooperative Ecosystem project undertook limited excavations at the Grassy Lake site.



The area selected for excavation was based on surface artifacts and manual soil probing.  A total of 12 sq m were excavated.  Ten square meters in an area southeast of the ridge and two square meters on the south end of the ridge.  During excavation materials were collected in 1 X 1 m units from 10-cm arbitrary levels and screened through 6.35 mm mesh.  Flotation and radiocarbon samples were collected from units and levels.   



Five of the units were excavated as a small block on the southeast edge of the ridge. The remaining unit was placed on the end of the ridge at a slightly higher elevation.   Within the block, the upper strata consists of very compact light brown silt that overlays a dark gray brown sandy loam midden. 




Below the midden is culturally sterile buff or yellow colored fine sand.  Soils profiles within this block are interpreted as sheet midden deposited on an abandoned river levee and then covered with silt following flood events.  The midden contained a dense amount of charcoal, ceramic and lithic material.




The remaining unit was placed on the end of the ridge at a slightly higher elevation.  Its profile consisted of a plow zone that overlaid a brown sandy loam midden. Both contained a low density of sherds and lithic artifacts.  At the base of the midden a dark grayish brown oval shallow basin feature was identified that contained few artifacts except for calcined bone fragments suggesting a hearth.  Based on the low artifact density and the presence of the surface hearth, this area is interpreted as an activity locale situated near a structure as opposed to the block which is interpreted as a refuse disposal area.

The three Grassy Lake radiocarbon dates obtained on nut charcoal from the excavation block are suggestive of a site occupation that dates from A.D. 1025-1150. 

Floral Assemblage

The plant collection exhibits a classic varied Mississippian plant use system within which 12 row closed cupule corn and native cultigens like chenopod, maygrass, and erect knotweed were significant contributors to the diet.  Nuts like hickory and acorn are well-represented.  Gourd rind is present in two samples, and wild plant seeds, including grape, sumac, smartweed, wild bean, and grass seed also were recovered. 

The collection is unusual in terms of the small amount of wood charcoal that was recovered. The dominant species of wood charcoal recovered are American chestnut, sycamore, elm, hickory, ash, oak from the white oak group, and cane.  The small amount of wood charcoal and the high fragmentation of nutshell, suggest secondary discard of food and cooking remains removed from primary habitation context.

Faunal Assemblage

Overall faunal preservation was poor due to acidic soil conditions in the midden and burning in the feature.     

Chipped Stone Assemblage

The chipped stone assemblage consisted of formal tools, informal tools, and debitage.   Of the 1384 lithic artifacts recovered, 95 percent were debitage; the remainder is either formal or informal tools.  The four complete and two fragmentary projectile points are all triangular Madison points made of Mounds Gravel, except for one made of Dover chert.  Hoe fragments recovered from the site consist of two bit ends of the hoes.  Twenty polished hoe flakes were also recovered.  All of the hoe fragments and re-sharperning flakes are Mill Creek chert.  Two drill fragments were made from Mounds Gravel Chert.   Informal tools consisted of edge modified or retouched flakes, utilized flakes, biface fragments, and core or core fragments.

Mounds Gravel accounts for almost two thirds of the raw materials used in chipped stone tool production.  Other raw materials utilized were glacial cobble, Purchase Gravel, Dover, and Mill Creek chert, and Quartz.   For the most part lithic raw materials were collected from nearby sources and were probably procured as unmodified gravels and cobbles from local streams.   Mill Creek and Dover chert would have been procured through exchange relationships, or by individuals traveling to quarries in Southern Illinois or Tennessee.

Ceramic Assemblage

The Grassy Lake ceramic assemblage is characterized by a high percentage of cordmarked vessels, with most of the remaining sherds having plain surfaces of varying degree of smoothing; with only a few 1 having fabric impressed exterior surfaces.   Of the plain sherds, 25 have a red, orange, brown, or black slip on their exterior or interior surface.  Five cordmarked sherds have a slip on their interior surface.  Though shell is present in 41.2 percent of the sherds, it is the dominant temper in only 13.3 percent of the assemblage.  



Jars are the most common vessel form, followed by bowls, pans, and funnels.  Slightly more than 50 percent of rims exhibit no modification.  Of the remaining specimens, rimstrips, exterior protrusions, and rolled lips are associated with about one third.  In general, plain jars are smaller than cordmarked jars.  This suggests that plain jars may have been used for different purposes than cordmarked vessels.  Perhaps they were used in food preparation and serving, with cordmarked jars being used for cooking and storage. 


Appendages consist of castellations, lugs, handles, and rimriders.  Notching of the lip is the primary decorative treatment, being associated with about one quarter of the rims.  Several sherds have incised or punctuated decorative motifs on jar necks.

A distinctive ceramic type recovered from Grassy Lake is stumpware.  As with later Wickliffe thick funnels these vessels may have been used as pot or pan supports.




Discussion and Conclusions

Radiocarbon dates from Grassy Lake, suggests that the site was occupied ca. A.D. 1050-1100 or towards the end of the early Mississippian James Bayou and Jonathan Creek phases.  On the other hand if one relied solely on the material culture and in particular the ceramics, one would be tempted to argue that the site was occupied about ca. A.D. 950-1000 or about 100 years earlier.

Based on comparisons of the materials recovered from nearby sites, we have come to believe that perhaps the time has come to rethink some of the temporal trends in material culture that have been proposed for the Ohio-Mississippi confluence region.

In general, the ceramic assemblages from the excavation at Grassy Lake compare favorably with assemblages recovered from the earlier terminal Late Woodland Pettit Site in Illinois and the early Mississippian Dedmon site in Kentucky.  The Pettit Site (11Ax253) is located along the Mississippi River a little more than 33 km northwest of the Grassy Lake Site, while the Dedmon site is located a little more than 65 km east of Grassy Lake along the Tennessee River. 

Not surprisingly, relative to the earlier Petitt site, the Grassy Lake Site exhibits an increase in the use of shell to temper ceramic vessels.  There is also an increase in vessels with plain exterior surfaces relative to cordmarked surfaces. 

At both sites rimstrips are associated with about one-third of the bowls and pans, but they are much more common on jars at Petitt than at Grassy Lake.  At the former they are associated with more than fifty-five percent of the jars, compared to about fifteen percent of these types of vessels at Grassy Lake.   

Though the Grassy Lake ceramic assemblage has more in common with Dedmon than Pettit, there are aspects of both assemblages that are suggestive of interregional differences.  At both Dedmon and Grassy Lake, about seven percent of the assemblage is tempered with just shell.  Surprisingly there is little mention of mixed shell and grog tempered ceramics at Dedmon.  When mixed tempered sherds are considered, Grassy Lake shows a greater use of shell as a tempering agent.  In addition, at sites, sherds with cordmarked exterior surfaces account for about 75 percent of each site’s ceramic assemblage and plain sherds about one quarter, but fabric impressed and slipped sherds are slightly more common at Grassy Lake than at Dedmon.  The high percentage of cordmarked sherds at both sites, suggests that plain surface sherds do not dominate local Mississippian collections until after A.D. 1100.   

At Dedmon rimfolds, lugs, castellations and handles are not as prevalent as at Grassy Lake.  On the other hand, lip notching and punctuation as a decorative lip treatment is more common at Dedmon than Grassy Lake.

The overall similarity of the Dedmon and Grassy lake assemblages, however, suggests that use of Mulberry Creek and Baytown Plain continued well into the Mississippi period.  Sussenbach and Lewis made a similar observation at the Marshall site in nearby Carlisle County. 



Marshall, however, is distinguished from Dedmon and Grassy Lake by a much higher percentage of plain vessels, and greater use of shell temper.  Though Sussenbach and Lewis suggest that the initial use of Marshall dates from A.D. 900-1100, based on our work at Grassy Lake we cannot help but wonder if the Mississippian occupation of Marshall post-dates A.D. 1100. 

The data from Grassy Lake, Dedmon, and Marshall, suggest that with respect to ceramics the Late Woodland to Mississippian transition was not very abrupt.  Grog and grog mixed with shell continued to be a major tempering agent well into the 1300s in the confluence region, and ceramics similar to Baytown Plain and Mulberry Creek Cordmarked continued to be manufactured well into the 1100s.  That ceramics were more resistant to change than other aspects of subsistence, material culture, and socio-political organization should not come as much of a surprise.  Ceramic manufacture is a learned behavior that is passed from generation to generation, and during periods of transition we should not be surprised to find a great deal of variation.   Though limited, the data from Grassy Lake raises questions concerning the transition from Late Woodland to Mississippian periods in the confluence region, and forces us to reassess some of our temporal assumptions.  As more work is undertaken in this region, we should be able to arrive at a better understanding of the Late Woodland/Mississippian transition in western Kentucky.



Jacob Kuhn
Assistant State Conservationist

William "Bill" Sharp

771 Corporate Drive Suite 210
Lexington KY 40503