the antiquities of this district have been shown by Squier to be chiefly due to the Indian tribes occupying that section at the time of its discovery by the Europeans, it is unnecessary to note the distinguishing characteristics. The works are chiefly enclosing walls, remains of palisades, and burial mounds.

(5) The Appalachian district, including western North Carolina, eastern Tennessee, southwestern Virginia, and part of southeastern Kentucky.

The characteristics which appear to warrant us in concluding that the works of this region pertain to a different people from those in the other districts, at the same time seem to show some relation to those of the Ohio district. Such are the numerous stone pipes, the altar-like structures found in some of the mounds, and the presence of mica plates with the skeletons. But the peculiar features are the mode of burial, the absence of pottery, and the numerous polished celts and engraved shells found in the mounds.

Although it is probable that there are at least three districts in the southern portion of the United States, they appear to pass from one into the other by such slight changes in the character of the works as to render it exceedingly difficult to fix the boundaries between them. I therefore mention the following, provisionally, as being those indicated by the data so far obtained.

(6) The Middle Mississippi area or Tennessee district, including southeast Missouri, northern Arkansas, middle and western Tennessee, southern and western Kentucky, and southern Illinois. The works of the Wabash valley possibly belong also to this district, but the data obtained in regard to them are not sufficient to decide this point satisfactorily. This district, like the others of the south, is distinguished from the northern section by its larger mounds, many of which are pyramidal and truncated and often terraced, and which were, beyond question, used as domiciliary mounds. Here we also meet with repeated examples of enclosures though essentially different from those of Ohio; also ditches and canals. From the Lower Mississippi and Gulf districts, with which, as we have said, it is closely allied, it is distinguished chiefly by the presence of the box-shaped stone cists or coffins, by the small circular house-sites or hut-rings, and by the character of the pottery. This is pre-eminently the pottery region, the typical forms being the long-necked, gourd-shaped vase and the image-vessels. In this district the carved stone pipes are much less common than in the Illinois, Ohio, and Appalachian districts.

(7) The Lower Mississippi district, including the southern half of Arkansas, Louisiana, and Mississippi. There are no marked characteristics by which to distinguish it from the Middle district; in fact as we move southward along the Mississippi from the mouth of the Illinois river, the works and their contents indicate a succession of tribes differing but slightly in habits, customs, and modes of life, the river generally forming one natural boundary between them, but the other boundaries being arbitrary. For example, the Cahokia region appears to have been the home of a tribe from which at one time a colony pushed northward and settled for a while in Brown and Pike Counties, Illinois. The extreme southeastern counties of Missouri were probably the seat of another populous tribe which extended its borders into the western part of southern Illinois and slightly into northeast Arkansas, and closely resembled in customs and art the ancient people who occupied that part of the Cumberland valley in middle Tennessee. This subsection is principally distinguished by the presence of the small circular house-sites, which are slightly basin-shaped, with a low ring of earth around them. As we move farther southward into Arkansas the house-sites change into low circular mounds, usually from 1 to 3 feet in height, and in nearly every instance containing a layer of clay (often burned) and ashes.

These small mounds, which are clearly shown to have been house-sites, were also burial places. It appears to have been a very common custom in this section to bury the dead in the floor, burn the dwelling over them, and cover the whole with dirt, the last operation often taking place while the embers were yet smouldering. Burial in graves was also practiced to a considerable extent. As we approach the Arkansas River, moving southward and from thence into Louisiana, the pottery shows a decided improvement in character and ornamentation.

(8) The Gulf district, including the Gulf States east of the Mississippi. The works of this section appear to be closely allied to those of the Lower Mississippi district, as here we also find the large flat-topped pyramidal mounds, enclosing walls, and surrounding ditches and canals.

The chief differences are to be found in the forms and ornamentation of the pottery and modes of burial.

As we approach the Mississippi River the distinguishing features gradually disappear, although there appears to be a distinct subdistrict in the northern part of Mississippi, and as we enter the Florida peninsula a change is observed which appears to indicate a different people, but the data so far obtained are not sufficient to enable us to outline the subdistricts.

This districting is to be regarded as a working hypothesis rather than as a settled conclusion which will stand the test of future investigations. It is more than likely that other subdivisions will be found necessary, and that the boundaries of some of the districts given will have to be more or less modified; still, I believe the arrangement will be found substantially correct.

As a very general and almost universal rule, mounds of the class under consideration are more or less conical in form, and are common to all sections where earthworks are known to exist, in fact they form almost the only ancient remains of some localities. Often they are isolated, with no other monuments near them, but more frequently they occur in groups or are associated with other works. Squier and Davis say "they are generally of considerable size, varying from 6 to 80 feet in height, but having an average of from 15 to 25 feet."[1]

This is probably true in regard to the mounds explored by these archaeologists in Ohio, but is erroneous if applied generally; as very many, evidently used and intended as burying places only, are but two or three feet high, and so far as the more recent examinations made in other sections—especially the explorations carried on under the Bureau of Ethnology—have shown, tumuli of this character are usually from 3 to 10 feet high, though some, it is true, are of much larger dimensions; but these are the exceptions and not the rule.[2]

As the authors just alluded to are so frequently referred to by writers, and their statements in reference to the works explored by them are taken as of general application, I will venture to correct another statement made by them in regard to mounds of this character. They assert that "these mounds invariably cover a single skeleton (in very rare instances more than one, as in the case of the Grave Creek mound), which, at the time of its interment, was enveloped in bark or coarse matting or enclosed in a rude sarcophagus of timber, the traces, in some instances the very casts, of which remain. Occasionally the chamber of the dead is built of stone rudely laid up, without cement of any kind."