Archeological Investigations

IV. ABORIGINAL HOUSE MOUNDSToC

The small, low, flattened mounds of the lower Mississippi Valley are a problem to archeologists. They have been studied principally near the Mississippi River, in Arkansas and Missouri, and for many years it was thought that in the latter State they are confined entirely to the southeastern portion. Recently they have been found much farther to the north and the west than they were supposed to exist.

A group, rather limited as to number and to the area covered, is at the head of a narrow valley trending northward from Granite Mountain in Iron County.

"Near Iron Mountain, in St. François County, more than 500 of these small mounds, arranged in parallel rows following the direction of the watercourses, were counted within a radius of 3 miles."[1]

The next group known north of this is on the right bank of Plattin Creek in Jefferson County, about 12 miles from the Mississippi.

"A group of some 50 similar mounds is situated on the right bank of the Meramec, about 6 miles above its mouth, in Jefferson County."[2]

The most northern group so far observed is near Ferguson in St. Louis County, Missouri, where 46 are located on a narrow ridge which has the same general elevation as the table-land. The ridge extends around the head of a ravine, and the mounds are placed along its crest or on the gentle slopes near the top. There are 10 or 12 at the southern edge of Ferguson, on an overflow bottom bordering a small creek.

Toward the west from the swamp region a small group is in a broad valley near Alton in Oregon County, which borders on Arkansas. They are scattered along a gentle slope which has a little stream at the foot.

In Dent County four groups are known. One is on the infirmary farm south of the town of Salem. Most of these are but slightly changed from their natural condition. Another group is 6 miles east of Salem. These also are largely intact. A third is on the road from Salem to Short Bend. The fourth is at the edge of Salem, on the Rolla road.

"On the high plateau of Dallas County, north of the Niangua ... within an area smaller than 10 square miles, 860 were counted."[3]

Three groups are well marked in Phelps County. A mile east of Rolla they begin at the line of the Frisco Railway and extend southward in a shallow valley or "draw." Some are on the overflow flat bordering the little stream, but most of them are on the slopes to either side.

South of Dillon they extend for a mile in a slight depression.

Beginning at the Soldier's Home in St. James, the largest number yet found out of the swamp region lie for 2½ miles on both sides of a small creek running eastward north of the Frisco Railway. These reach from low land subject to overflow to an elevation of fully 50 feet up the hillsides.

Several groups occur in Pulaski County. Four miles southwest of Big Piney post office, near the site of what is known as "The Ranch House," is a little wet-weather stream along both banks of which are probably a hundred of these structures. Farther up this stream are two other groups, the three including a distance of about 4 miles in length between their outer limits. West of these and south of Bloodland is a fourth group belonging with these.

In the level bottom between Big Piney River and the branch flowing from the Miller Spring 2 miles from Big Piney post office a number of these mounds formerly existed; and on the opposite side of the Big Piney, in an extensive bottom, were many of them. All these have now disappeared under cultivation.

On the outer bend of the Devil's Elbow, on Big Piney 3 miles above its mouth, some of these mounds stood. They are described as being from 2 to 3 feet high; the number was not stated, but there is not room for many in the narrow strip where they were located.

In the extreme western part of Morgan County, at Stover, is a group scattered over an area at least half a mile across in any direction. The distance between the mounds varies from 25 to 150 feet. They are mostly on gentle slopes, though some are on the crest of the ridges. Many of these are well preserved, some of them having never been under cultivation.

In Osage County there are more than a hundred at the eastern edge of Rich Fountain. They are in low flat ground which is muddy or even boggy in wet weather.

It will be noticed that all those from Alton westward and north-westward are in line with the route from southeastern Missouri to the plains of Kansas and Nebraska.

Practically, however, the northern limit of this type, in great numbers, is in St. François County, near Farmington. From here they extend almost continuously into Louisiana and Texas.

In nearly every part of southern Missouri east of the Iron Mountain Railway they occur in closely connected groups, reaching sometimes for miles except where the continuity is broken by a slough or other unfavorable condition. They are found everywhere—on high, well-drained levels; on sloping ground, sometimes so steep that it may well be called a hillside; in low "crawfish land"; in swamps where, in the driest weather, even after a prolonged drought, they can be reached only by wading through water or muck. The last, however, may have been more easily accessible when built, their present condition being due to the general subsidence of this region during the earthquake period of 1811. The existing sloughs and sluggish bayous are the widenings and extensions of streams which at the time these mounds were constructed were no doubt bordered by banks above ordinary overflow and readily reached by canoes. Manifestly the country was well populated, and therefore presumably practically timberless; consequently the flood water would rapidly pass away and the streams not be choked by drift and other débris as is the case at present.

Various theories, most of them advanced by persons who are but slightly, if at all, familiar with the country, have been propounded to account for mounds of this character. Their vast number has led some writers to believe that they can not be artificial but must be due to natural phenomena; as, for instance, that these, as indeed all mounds, were piled up by floods, Noachic, glacial, or local; or that they result from the industry and energy of burrowing animals, such as foxes, badgers, ground hogs, rabbits, prairie dogs, gophers, chipmunks, or even ants; the character of the assumed flood or the species of the supposed burrower depending to some extent upon locality, but principally upon the theorizer's insufficient knowledge of animal industry or of the action of torrential waters. Others are convinced they are formed by the piling up of earth around a bush, clump of grass, stone, or other object acting as a nucleus about which wind-borne material may accumulate—overlooking the fact that clay, gravel, or gumbo soil can not be carried by wind, and that lighter soil or sand will form elongated instead of circular masses. Another supposition is that they are due to stream erosion; flood waters washing away the soil between them and thus leaving the earth composing the mound in its original position. The same objection applies to this as to the wind-blown theory, namely, that we can not imagine water acting with such mathematical regularity and intelligent discrimination, especially upon slopes which lie at all sorts of angles with the trend of the current.

Persons who recognize their human origin have suggested that they were erected as stands for hunters, from which they could detect game at a greater distance, or could take better aim as the animal passed; or perhaps as camping places while waiting; but in many places more than half the area of the ground over several acres is occupied by such piles of earth, promiscuously distributed. This implies more hunters than animals.

For a long time it was supposed that they were burial mounds, like so many such structures found over the country; but this idea has been dispelled by the failure to discover in them any evidences of such purpose; no human bones nor any of the artificial objects commonly placed with the dead have ever been found in them unless under such conditions as to show their presence was accidental.

Two very plausible theories have found general acceptance: That they were the sites of dwellings, placed on them to be out of the mud in wet weather; and that they were in the nature of garden beds, thus elevated for growing any food products which needed a comparatively dry soil, or might be injured by temporary accumulation of water from excessive rainfall.

But they were not "residence mounds" or "house sites" in the sense that they furnished a base or foundation for structures which were used as dwellings; for there has never been found on their surface or in the earth immediately around them any of the débris invariably accompanying Indian huts or houses, such as fireplaces, ash beds, burned rocks, broken implements, or fragments of bones and pottery. These considerations also interfere with a full acceptance of the hypothesis that they are remains of houses built of wood and covered with earth. It is true that such evidence is very frequently found in other localities; but to establish the fact that they were residence sites, refuse of this kind should be found wherever the mounds occur.

J.B. Thoburn arrived at this conclusion from the resemblance of some of them in their outlines to the grass-covered houses of the Pawnees; and it is believed that this tribe in its migration from the south followed approximately the route along which these small elevations are found. When the Pawnees—assuming they were the builders—passed on westward they could not procure timbers of sufficient strength to hold up the earth, so they used light frames and covered them with grass.

Bushnell arrived earlier at the same conclusion. He says, concerning a few mounds of this character in Forest Park, St. Louis: "In the case of the seven mounds on the elevated grounds, the finding of potsherds, pieces of chipped chert, and the indication of fire, all on what appeared to have been the original surface, would point strongly to their having been the remains or ruins of earth-covered lodges." He gives citations from early explorers in support of this theory, and adds, "But in other mounds these indications did not occur."[4]

Such an explanation finds support in the vast number of these structures. In building, the aborigines naturally chose the sort of timber which was soft and light, consequently easy to cut and to handle, such as willow or cottonwood. This soon decays. But no matter what variety of wood was utilized, not many years would be required, under the conditions supposed, to weaken its fiber until it could no longer uphold the weight of earth on the roof, and a new house must be erected. Several such renewals would be needed in the course of a century; so that the ruins of an ordinary village might create the impression that a large settlement had existed on its site.

The explanation of "agricultural use" is probably correct in some instances, for frequently the mounds are made of earth gathered up around their base, and so not only would be of value in a wet season, but would afford a much greater depth of fertile soil for sustenance of plants. In some localities modern farmers find that on such mounds crops are much better than on the low spaces between them. On the other hand, a majority of the small mounds in the lower counties of southeastern Missouri are composed either of the hard, reddish, sandy clay which forms the subsoil of the land above overflow; or of the tough, waxy, black "gumbo" of the swampy or flat lowlands. In either case they are almost invariably sterile, so that in a cultivated field the position of a mound is easily determined even from a considerable distance by the feebler growth on its surface. Moreover, in many places, hundreds of them occurring within an area of a few square miles are built on clay lowlands where crawfish abound, within a few rods of sandy, well-drained ridges whose soil is never muddy more than a few hours after the hardest rain, and produces as fine corn and wheat as can be raised in any part of the State.

In short, no matter what suggestion has been offered as to their purpose or uses, objections to it can be brought and sustained. It is not improbable that, in the end, it will be found the difficulty lies in trying to place in a hard and fast category a variety of structures which are similar in appearance but which were intended for various uses. With more comprehensive study, it may be that a classification is possible which will interpret what is now obscure. Instead of uniformity, there was probably great diversity of motives, ideas, and beliefs which led to the building of these as well as of other mounds; and when the key is once obtained the explanation which will account for one may be very different from that which as clearly accounts for another.

A few of these mounds have been explored by the writer, but no discoveries were made upon which can be based a definite statement as to their probable purpose.

NEW MADRID COUNTY

On the farm of A.B. Hunter, 7 miles north of New Madrid, more than 60 of these mounds, irregularly placed, extend for half a mile along the west bank of St. John's Bayou, the extreme width of the group being about 200 yards. The largest mound, standing on the edge of the terrace, was 6 feet high and 75 feet across. On the original surface, over a small area at the central part, were decayed fragments of human bones; so this was probably erected as a tumulus. The others were much smaller; from a foot to 3 feet high, and 30 to 50 feet in diameter. Six of these, varying in size from the largest to the smallest, were thoroughly excavated within the original margin and down to the undisturbed earth beneath them. No artificial object was found in any of them except here and there a fragment of pottery or a small amount of ashes or a piece of charcoal, not intentionally deposited but gathered up and carried in with the earth in the course of construction. There were no distinct fire-beds or ash piles at the bottom, or in any part of the mound; nor were there any holes in which posts may have stood.

ST. FRANÇOIS COUNTY

Nearly 2 miles south of Farmington, on Quesnel's land, are about 30 very small, low mounds, none more than 18 inches high or 25 feet across. They are on the general level, some of them on a gentle slope, of the first upland above the St. François River and a mile from that stream at its nearest point.

Half a mile to the south of these is a group of similar mounds on the farm of Isaac Hopkins, on a gently sloping hillside, and from 30 to 40 feet above the level of the overflow bottom land. One of these has been gradually worn away by the encroachment of a gully until more than half of it has disappeared. While the curvature of its surface is very apparent, and the remnant of its margin sufficiently distinct to show its regularity of outline, careful inspection of the face formed by the erosion fails to reveal any trace of stratification, or line of demarcation between the bottom of the mound and the original surface. There is precisely the same uniformity of change from the grass roots to the underlying gravelly soil that exists in the exposed bank at any point to either side of the mound. Mr. Hopkins, desirous of knowing what might be in the mound, or why it was built, has noted the appearance of the earth from the time the gully reached its margin. At no time has its appearance differed in the least from what it presents now.

On the river bottom portion of Mr. Hopkins's farm, and on the adjoining Goings and Townshend farms to the southward, are many mounds lying along both sides of the Belmont division of the Iron Mountain Railway. Fully 100 were observed within a distance of a mile; and they are said to continue both up and down the river. They are all above flood stage, except in time of extreme high water. They range from a foot to 3 feet high, and from 20 to 40 feet across; but some of them have been lowered and broadened by cultivation. They are of the same earth as the ground around them. Mr. Hopkins says crops are much better on the mounds than on the area between them. This is no doubt due to the greater amount of productive soil in the one case, and to the excess of moisture in the other; the railway embankment impeding drainage in the lower part. Oak trees 4 feet in diameter grew on the mounds before they were cleared off.

Two of these mounds were completely removed, down into the subsoil. The first was 18 inches high and 35 by 40 feet across; the variation in breadth resulting from continual cultivation in one direction. It contained nothing whatever of artificial character, not even a scrap of pottery. There were no post holes, no indications of a fire bed, no trace of a distinction between the mound and the soil below it. In fact, except for the greater thickness of the superficial dark earth there was no difference between the appearance of the face of the excavation and that of a hole dug at random in the field.

The second mound was somewhat larger than the first, being 2 feet high and 40 feet across, and at a little higher level toward the edge of the field. It was the largest which could be excavated of this group. As in the first mound opened, there was no worked object, if a small flint flake be excepted; no ashes; no fire bed; no trace of demarcation between the mound and the original surface of the ground, though in each mound the excavation over the entire area was carried down into the gravelly, hard-packed subsoil. Its artificial origin is clearly proven, however, by four holes dug into the earth beneath it before its construction. Nine feet a little north of the center, which was assumed to be the highest point of the mound, was a hole (A) 12 by 14 inches and 14 inches deep, with a flat bottom, the sides as regular as could be expected in hard soil dug out in primitive manner. Nine feet west of the center was a hole (B) a foot across, 10 inches deep, with a solid though somewhat irregular bottom. Near the center was a conical hole (C) a foot deep and the same across the top. Four feet from it, west of north, was another (D) of about the same size and shape. The measures given are of course only approximate, as the sides of all the holes were somewhat uneven, but they are practically correct. The depth was measured from the top of the gravelly subsoil. Fourteen feet east of south from the center was an irregular hole (E) about 2 feet deep to the bottom of the loose dirt in it. This had not been dug, but was due to the decay of a tree which grew here before the mound was made. At the top of the dirt filling this hole was a piece of decayed bark, apparently oak, which had grown in the air; and farther down fragments of root bark. Eight feet east of the center was a hole (F), similar to the last, 10 inches deep and averaging 2 feet across. This, also, resulted from the decay of a stump.

A plan of the holes is given in figure 37. The dotted lines are merely to show direction and distance.

This mound offers confirmation of the belief that such structures, or some of them at least, mark the sites of dwellings. With the two trees, E and F, the posts, A and B, would form the corners of an irregular quadrangle; the two posts, C and D, would support the inner ends of roof timbers. While no trace of posts or roof timbers remained, it is difficult to imagine for what other purpose these holes would be dug; and in this heavy, wet earth all traces of wood must in time disappear. Conversely, the total absence of a fireplace, potsherds or other remains, and of any sign of a floor, would serve to dispel the assumption that this spot was ever inhabited even for a short time. The evidence is as strong one way as it is the other.

In short, the limited observations above recorded leave the question of origin and purpose just where it was.

Some years ago one of the mounds at Ferguson, St. Louis County, was opened. No remains of any sort were discovered, according to the report of the excavators; but on the original surface, at the center of the mound, was a fire bed in and about which were ashes, charcoal, and fragments of rude pottery.

No excavations have ever been made in the mounds near Granite Mountain; but a tortuous little stream has undercut several of them, thus making vertical sections as in the case of the mound at Hunter's, near Farmington. In some mounds only a small portion near the margin has been removed; in others the erosion has progressed to such an extent that observations were possible at varying distances, to and beyond the center. In every instance a monotonous uniformity of appearance prevails from the top of the mound into the underlying gravel. At no level is there a sign of a floor, fire bed, or other evidence of human work; and no difference can be detected between the earth upon which the mound rests and that on either side. Yet the mounds are indubitably artificial.

Exactly the same remarks apply to several mounds on the County Farm, near Salem. A little creek and a drainage ditch have cut away varying portions of them, and they merge insensibly into the soil and gravel on either side.

In further support of the theory that these mounds are the remains of earth-covered houses, a few extracts relating to the area under discussion will be given from Dr. Cyrus Thomas in the Twelfth Annual Report of the Bureau of Ethnology:

Near "Beckwith's Fort," in Mississippi County, Missouri, are (p. 189)—

In describing mound excavations in Crittenden County, Arkansas, the explorer states (p. 227):

The positions of the skeletons in this mound would indicate that while the inmates of the house were asleep the roof fell and killed them. It was customary among some southern Indians to bury the dead under the floors of the houses; but the text clearly shows that these skeletons were lying on the floor. It would be supposed from most reports, not only in the volume quoted, but from various other sources as well, that only the walls of these houses were plastered with mud, the roof being of thatch alone. It seems to be overlooked that the tops of the houses would have even more need of such protection than the sides. The marks indicating that the clay was "stamped apparently with an implement made of split reeds" are only the impressions of the reeds or saplings by which the clay was supported; the "brick like" or "clinker like" condition of the clay being due, of course, to the destruction of the house by fire.

Adair, in his History of the Southern Indians, says they daub their houses with tough mortar mixed with dry grass; that they build winter or hot houses after the manner of Dutch ovens, covered with clay. Again:

This seems to mean that the entire building was plastered with mud, and then covered with grass to shed the rainfall.

In a mound in Arkansas County, Arkansas (Twelfth Ann. Rept. Bur. Ethn., p. 231)—

The fact that no skeletons or utensils were discovered on the floor finds its most reasonable explanation in the supposition that the inmates, finding their abode to be unsafe, moved out and took their possessions with them. This would account, also, for the absence of such remains in similar mounds farther north. The abundance of pottery fragments found in this case, and in many others, may mean only that these were worked in as a part of the clay roofing. They would be of some service in holding the clay in place in wet weather.

It is quite probable that the continuous, though fragmentary, layer of burned clay on the floor so often noted is due in part at least to the material forming the roof. The walls would be more apt to fall outward than inward, and would be more liable to crumble than to fall as an intact mass. In fact, this is clearly shown by the statement (p. 229) that in certain house sites in St. Francis County, Arkansas,

Further, in describing explorations of certain "hut rings" at "Beckwith's Fort" in Mississippi County, Missouri (p. 187), the report states that they are

Note is made of

It appears, too, that certain sites were occupied for long periods, new houses being constructed when necessary. In describing mounds in Poinsett County, Arkansas, the same writer says (p. 205) that

These layers of "very dark" material undoubtedly are remains of mud from the adjacent swamps, which was mixed with or plastered over the grass roofs. It is difficult to understand how they could have become mixed after the burning.

As showing the extent to which this prolonged occupancy was carried, we are informed (p. 254) that in Coahoma County, Mississippi, a mound was—

In still other portions of the country besides those already mentioned are evidences of similar houses whose sites are now marked by mounds. In southern Ohio, especially, records of excavations contain numerous references to post holes under mounds both large and small. In the case of the former, so far as we may judge from the reports, the houses were destroyed before the mounds were built, and it does not appear that they were ever covered with earth. In the small, low, flat mounds, under which such holes existed, no thought was taken that these may mark the position of posts used to support a roof; all mounds were explored with the idea that they were for burial purposes, consequently no attention was paid to these features.

The Mandan houses, as described by Lewis and Clark, Catlin, and others, when fallen into ruins would leave exactly such mounds or hut rings as those found in Missouri and Arkansas.

It is now generally conceded that the wall or embankment at Aztalan, Wisconsin, concerning which so many wild theories have been promulgated, was simply a series of such house sites connected by a low ridge. The evidences of mysterious sacrificial altars seem to be due only to the destruction of such houses by fire.

In Wisconsin, also, and in Minnesota, are many small mounds apparently of this character which are due to an extinct tribe known to the Sioux and Chippewas as "The Ground House Indians."

In 1887 I became acquainted, at Munising, Michigan, with Mr. William Cameron. He was of the Scotch clan of Camerons, a nephew of a former Governor of Canada. Educated for a profession, he made a visit to relatives in Canada in early manhood, and the attractions of the wilderness proved so great that he never returned to his home. At the time I met him he was 84 years of age, in full possession of his mental faculties. For more than 60 years he had traversed the Lake region, his fur trading and trapping expeditions having carried him over all the country from Montreal to the mouth of the Mackenzie River. Much of his life had been spent among the Indians, especially the Sioux and Chippewas. He learned from them all they could tell him of their tribal history and former methods of living. The Chippewas told him that when they first came into the country they found the Sioux in possession, but finally, obtaining arms from the French, they drove the Sioux westward.

The "old men" of the Sioux corroborated this tradition and told Cameron that as they went westward they came to a race of people who lived in mounds which they piled up. These people were large and strong, but cowardly. "If they had been as brave as they were big," said the Sioux, "between them and the Chippewas we would have been destroyed; but they were great cowards and we easily drove them away."

Mr. B.G. Armstrong, of Ashland, Wisconsin, told me that he had taken great pains to investigate this tradition. From all that he could gather by much inquiry among the Indians and from his own observations, he was satisfied of its correctness. These people, whom the Sioux called Ground House Indians, built houses of logs and posts, over and around which they piled earth until it formed a conical mass several feet thick above the roof. Their territory extended from Lake Eau Claire, about 30 miles south of Lake Superior, to the Wisconsin River near Wausau or Stevens Point; down the Wisconsin a short distance; thence west into Minnesota, but how far he could not say; then around north of Yellow Lake back to the Eau Claire region. The Sioux exterminated the tribe, the last survivors being an old man and a woman who had married a Sioux. They were taken to the present site of Superior, near Duluth, and "died about 200 years ago"—that is, in the last quarter of the seventeenth century.

Gordon, an intelligent Indian living at the town of the same name, a short distance south of Superior, was familiar with this tradition, as were other Indians with whom I talked, and who accepted it as a well-known fact. Gordon related that he had heard "the old men" say these Indians erected their houses of wood and piled several feet of dirt over them; and they buried their dead in little mounds out in front of their houses and a few hundred feet away. He told of a mound that was opened near Yellow Lake in which the position and condition of the skeletons, two or three of children being among them, showed "as plainly as anything could" that they had been sitting or lounging around the fire, when the roof fell in and crushed them.

There is a "Ground House River" in eastern Minnesota, which probably derives its name from this people.

FOOTNOTES:

V. ARCHEOLOGICAL WORK IN HAWAIIToC

INTRODUCTION

The ethnologist or archeologist desiring to conduct explorations on the Hawaiian Islands will find it necessary to begin his labors at the Bishop Museum in Honolulu. This museum contains an extensive collection of articles, classified, arranged, and labeled, illustrating every phase of native life as it has existed since the islands have been known to white men, as well as many of the implements and objects pertaining to agriculture, fisheries, and domestic occupations of earlier times. Models or casts of houses, and of individuals engaged in various lines of industry, give the visitor a clear idea as to the routine of ordinary daily life. A careful study of all these things enlightens him in regard to what he may expect to find and to the meaning of such discoveries as he may make.

The extensive library which belongs to the museum contains every publication relating not alone to the islands but to all the archipelagoes of the southern Pacific that it is possible to procure; and among the most valuable of the volumes are the reports and memoirs of the museum itself, in which are set forth the observations and deductions of numerous investigators who, either in behalf of the museum or under its auspices, have endeavored to find a solution for the many problems involved.

Equally valuable to the student are the information, interpretations, and instruction freely placed at his disposal by those connected with the museum, especially by Dr. Brigham, the former director, whose long and busy life has been devoted almost entirely to a study of the Polynesian groups; by Professor Gregory, the present director, who with tireless energy is the impelling force behind various lines of scientific research; by Mr. Stokes, curator of the ethnological department, who for more than a score of years has been surveying, photographing, and collecting in every part of the islands; by Mr. Thomas G. Thrum, of Honolulu, who has completed, in manuscript, a volume containing a list and description of more than 500 heiaus on the islands; and by various other men who, in private life, have devoted much time and close attention to whatever may pertain to native life and customs.

MOLOKAI ISLAND

Following the advice of those whose knowledge gave them authority to speak decisively, the initial base of research was the island of Molokai, which presents the best conditions for study. It lies off the usual lines of travel, offers no inducement to tourists who wish to have the benefit of good roads and comfortable hotels, and consequently is seldom visited except by those who are called on business or who go as the guests of the few residents there.

Mr. George Cooke, one of the owners of a large cattle and sheep ranch on the island, and greatly interested in its aboriginal history, gave most generous aid in a reconnoissance of such parts as he had time to visit. He placed his beautiful summer residence at the disposal of Prof. Gregory and the writer, and conducted the explorers to nearly all the places of interest which could be approached by automobile. Mr. James Munro, manager of the ranch, also rendered valuable assistance. Owing to his long residence here he has become thoroughly familiar with every noteworthy feature, and pointed out many remains which, without his guidance, would have been missed altogether. Fully acquainted with the life of the Hawaiian people, he made clear the origin and purpose of many things that, lacking his intelligent explanation, would have been without significance.

Although there are now comparatively few Hawaiians on Molokai, it is evident that the island at one time supported a dense population. Along the southern, or leeward, coast are numerous fish ponds formed by building a stone wall across an inlet or, more frequently, by constructing it with the ends on shore and carrying it around a section of the open sea. The walls are strong enough to resist the waves, well above the level of high tide, and surround spaces of various areas up to 70 acres. These ponds were stocked with numerous kinds of fish which, thus protected from their natural enemies, increased rapidly and formed an unfailing food supply. The antiquity of these ponds is denoted by the amount of silt partially filling them, brought down from the mountains by erosion of the soil. They are still used to some extent by Hawaiians as well as by other residents.

Inland, low walls of stone or earth, or both, surround hundreds of old taro patches, one variety of these plants requiring an abundant supply of water during its growth. The poi made from taro was the principal vegetable food of the inhabitants. Sweet potatoes were also a leading article of diet. The fields in which they were grown may still be identified here and there by the little ridges heaped up. All these, with the addition of migratory birds and fowls which at certain seasons swarmed on the different islands, supplemented by various nuts and fruits growing spontaneously, provided a varied and ample food supply. Mammals, except the pig, dog, and rat (really a large mouse), which came in with the early natives, were unknown prior to the advent of the whites. There were no land reptiles and few indigenous noxious insects; although mosquitoes, not to mention certain domestic pests, abound in a few places, and there are some scorpions and centipedes; but these, like measles, smallpox, tuberculosis, and worse diseases, are adjuncts of an enforced civilization. The mongoose, brought in to destroy rats, and the myna bird, to devour insects, are themselves now beginning to be detrimental.

Along the coasts, on the headlands and lower hills, and to a less extent farther inland, are village sites, foundations of temples and houses, garden patches inclosed by stone walls, and long rows of stones, some of which are borders of roads or trails, others being for purposes which are undetermined. Among these, taro beds and sweet potato patches may still be traced.

The most remarkable among the remains are the great temple site on Senator Cooke's ranch, toward the east end of the island, and the "paved trail" 10 miles down the coast from Kaunakakai, the principal village and harbor. The former is rectangular in outline, built on irregular ground, of stones large and small, to form a level platform on which a thousand persons could assemble without being hampered for lack of room. The outer faces of the walls vary from 3 to 20 feet in height; and except at the lowest parts there are terraces or steps all around, about 5 feet in height and of differing width. Surrounding this platform, extending for half a mile up the little valley of which it marks the entrance, on the slopes to either side, and on the nearly level area reaching down to the sea in front, are all the indications of a populous settlement.

It is said that the ruins were formerly much more numerous and extensive, the larger part of them being swept out of existence by a great rush of water from the mountains "a long time ago."

The "paved trail" is a causeway of large stones. Some parts of it are obliterated by slides and encroaching ravines; other parts preserve the original condition and appearance. The width is not quite uniform, as the stones are of different sizes, but it departs very little either way from 6 feet. So far as can be judged in its present overgrown state, it extends in a straight line for about 2 miles, from the beach to a point on the hill at an altitude of fully 1,000 feet. To what it led, or why it was built, are questions awaiting an answer.

All of these places are now abandoned except a few villages along the coast. The people are not here to occupy them, and even if they were the conditions have become so changed that residence about them is no longer feasible. At the temple site, for example, the extent of the old taro beds predicates an abundance of water; at present, the one family living near by must carry it in a dry season from the well or spring of a neighbor. There is no steady water supply within miles of the "paved trail."

Clearly, extensive changes have taken place in recent times in climate and perhaps in topography. Fifty years ago forests of large trees grew over hundreds of square miles on the southern slopes of Molokai where at the present time there is only grass, or where algaroba trees, similar to the mesquite of the southwestern United States, are now spreading. This deforestation is still going on; dead or dying trees fringe the timber still standing. The cause of this progressive barrenness has not, so far, been, fully ascertained; there is undoubtedly a connection between it and the diminished water supply, though which is cause and which is effect, or whether both are due in common to some atmospheric phenomenon, is unknown. One result, however, is apparent. The roots of the forest trees do not extend deep into the earth, but spread out over the surface like those of pine trees. Thus much of the rainfall was prevented from escaping rapidly and such as was not absorbed by the roots made its way into the ground beneath the upper soil, whence it percolated downward to feed the springs. Now the greater part of the water runs off and is lost. For this reason large areas once well populated are no longer habitable.

Molokai, like other islands of the group, contains no stone except of volcanic or coral formation. There is no chert or similar material from which chipped implements can be made; nor, as would naturally be expected, is there any obsidian suitable for such manufacture. It may occasionally be seen on the sites of villages, but always in small angular fragments seldom more than half an inch in any dimension, always coarse-grained, even porous, and never of a quality which can be flaked into definite forms. No doubt its only use was as an abrasive, after being pounded fine. Rarely, quartz or chalcedony is found; it resembles the deposit around hot springs or in fissures, and, like the obsidian, is in fragments too small to be utilized except as a grinding or polishing material for smoothing wrought objects.

Manufactured stone specimens are confined principally to three general classes: Adzes, for working in wood; pestles, for pounding the taro root; and discoids, for games. The last are exactly similar to the chunkey stones so abundant in the States, except that none of them have concave or hollowed faces, and they are used in the same way. There were three forms of the game: To hurl or roll a disk farther than an opponent; to strike a pole or other mark set up; and to test the inherent magical powers of the stones by rolling them in such a way that they would collide, the object in this case being to see which one might prove victorious by breaking the other or forcing it out of its course. A suitable arena for the contest was prepared by carefully leveling and smoothing a straight, narrow strip of ground to any length desired, a slight wall being thrown up along each margin.

Pottery was unknown, there being no clay suitable for making it. Calabashes or gourds and wooden trays served as receptacles, though stone dishes or bowls are sometimes found. Along the coast occur sinkers, either plummet-shaped or half-ovoid like an egg divided lengthwise. This form has a groove around the longer diameter, crossing the flat face, and was tied to a white shell as a sinker in catching squids or cuttlefish, a hook being attached to the line. Coral was much used as files or rasps. There are a few objects whose purpose is problematical; and some highly polished black disks which, laid flat and covered with a film of water, make excellent mirrors; but aside from what is here mentioned, not much worked stone is found. Wood, bone, and shell served as the raw material for nearly all other needs.

Graves, or what are supposed to be graves, marked by cairns 3 or 4 feet high, or perhaps by only one or two layers of stones, are found, though rare. Many so-called caves—which are merely "tunnels," "bubbles," or "blow-holes" in the lava—were utilized as burial vaults. The natives vigorously protested against an attempt to excavate any of these, claiming that their ancestors or members of their families are buried in them and must not be disturbed. In the dunes human skeletons are frequently exposed by the shifting of the sands by the high wind. The natives seem to have little regard for these. Perhaps they are of the "common people," while cairns cover the chiefs or priests. There is a tradition that in "the old times" most of the dead were cast into the ocean as an offering to the Shark God.

There are no mounds or other structures of earth; everything was built of stone. All structures began at the surface of the ground. No evidence has been found of an occupation earlier than that of the present Hawaiian people. At no point examined in ravines or cliffs was there the slightest hint of human life at a period antedating that beginning with the race discovered by Captain Cook. Consequently no extended excavations were attempted. The results of some examinations made in three different places will be presented.

About 10 miles in an air line from Kaunakakai and the same distance from Mr. Cooke's home, on a mountain known as Mauna Loa, is a narrow, sharp ridge extending nearly south and terminating abruptly at the junction of two deep ravines. On the end of this are two house sites, or heiaus, which had never been disturbed. They are as nearly rectangular as the irregular stones of which they are built will permit. The larger (A) has its south wall at the edge of the low cliff, with its sides nearly on the cardinal lines. Omitting inches from the measurements, its outer dimensions are: North wall 38 feet, south wall 32 feet, east wall 33 feet, west wall 32 feet. The corresponding inside measurements are 21 feet, 19 feet, 21 feet, and 22 feet. Thirteen feet north from the north wall is a stone pile 13 feet north and south by 10 feet east and west, 18 inches high. Ten feet west of this is a single layer of stones covering an area 7 feet east and west by 4 feet north and south. At 9 feet out from the middle of the west wall is a platform 7 by 7 feet, its west edge on large stones in place. At the west end of the north wall are three large flat stones, one of them forming the corner, the two others west of this, the three being up-edged and in a continuous line.

Within the inclosure, at the southern end, is a closely laid pavement formed of a single stratum of loose stones, laid on the earth, and covering a space 20 feet east and west by 10 feet north and south. Along the inside of the wall, at the northeast corner, is a similar pavement 12 feet north and south by 4 feet 6 inches east and west, and a foot high. Both of these pavements were probably intended for seats and beds. On the larger pavement, 5 feet from the south wall, 9 feet from the east corner, was a boulder, its diameters 11, 12, and 15 inches, whose largest surface lay uppermost and was hollowed out to form, a deep saucer-shaped depression like a mortar; but as there was nothing to grind, it was probably to crack or pound nuts in. At the middle of the southeast quarter of the inclosure was a pile of stones 3½ feet across and 1 foot high; there was nothing under them. Seven feet from the north wall, 10 feet from the east wall, was a fireplace formed of two slabs on the east and west sides and a flattened boulder on the south side, all upedged, the north side being left open. Its bottom was undisturbed earth, a foot lower than the level of the platforms. It would seem, though this is uncertain, that the platforms or pavements were on the original surface level, the unpaved space being cleared out to the level of the bottom of the fireplace, and that this space had been filled with earth blown in by the winds after the spot was abandoned. From outside to outside the upedged stones measured 26 by 28 inches; the space inside 18 by 20 inches. On the west edge was a large grinding stone, the amount of wear on its surface indicating much use. A pavement 4 feet wide reached from the open side of the fireplace to the north wall.

In the cavity was about half a bushel of small stones, most of them burned. When meat was to be baked, a fire was made in the pit and as many of the stones as required were heated; they were placed in the body cavity, in the mouth, and in slits cut in the skin of the animal, which was then deposited in the pit, closely covered, and left until thoroughly cooked. Similar ovens or barbecue holes, and the same method of cooking, are still in use among the natives in their villages.

Views of this house site and of the fireplace, taken from various directions, are shown in plates 38-40.

Nearly north of the house site (A), at a distance of 91 feet, is the similar structure (B). The ground on which this is built is 6 feet lower than at (A). Its measurements are 23 by 24 feet outside, 13 by 18 feet inside, longest north and south. The entire interior is paved. For a space of 8 feet from the north end the pavement is a foot higher than in the south end. Beginning at the foot of the south wall, on the outer side, and extending for 29 feet toward (A), there is a closely laid stone pavement 10 feet wide at the wall and gradually diminishing to a width of 5 feet; its termination is nearly square, the slight curve being apparently not intentional. The west edge of this pavement is in a straight line, the east edge being curved.

Partial views are given in plate 41.

Neither (A) nor (B) has any opening for a doorway, nor is there any apparent method of easy entrance, though a slight platform on the north side of (A) may have supported steps of wood.

These walls, as in all other heavy structures observed, were made by carefully laying up two rows of large stones at a little distance apart and filling the space between them with stones of any convenient size, thrown in at random. Timbers set in them formed the skeleton structure of a house which was completed of poles and smaller growth, the sides and roof being thatched. The weight of the stones held the main timbers against the force of the wind even in severe storms.

The surface over hundreds of acres around these ruins is covered with house sites, long straight rows of stones, and garden lots surrounded by stone walls. Shop refuse, mostly chips and spalls from adz making, sea shells broken to extract the mollusks, coral for abrading, adzes in all stages of finish, and many "olimaikis" (chunkey stones) are found. A mile away is a chunkey yard or bowling alley about 600 feet long on the crest of a ridge which overlooks the ocean on both sides of the island.

THE RAIN HEIAU

A mile from the Cooke residence is a peculiar structure, said to be the only one of its kind in the entire Hawaiian group. Native tradition has it that "a long time ago" a rain wizard who was angered by the people of this district sent such rains that everything was on the point of being washed out to sea. Another wizard told the people to make a heiau (temple, or sacred building) with many small compartments which were to be left uncovered in order that the raindrops, each of which was as large as a man's head, could be caught and held in them, and burned. The rain would cease when the first wizard learned that he was being circumvented.