Archaeology by Experiment

John Coles

• Publisher: Routledge
• Year: 2014
• eBook ISBN: 9781317606086
• Topic: Social Sciences
• Subtopic: Archaeology
• Index: Social Sciences

Experimental archaeology is a new approach to the study of early man. By reconstructing and testing models of ancient equipment with the techniques available to early man, we learn how he lived, hunted, fought and built.

What did early man eat? How did he store and cook his food? How did he make his tools and weapons and pottery? Such everyday questions, besides the more dramatic mysteries associated with the monuments of Easter Island and Stonehenge and the colonization of Polynesia, can all be explored by experiment.

1

FOOD PRODUCTION

FOREST CLEARANCE AND CROPS

The indisputable fact that vast areas of Europe, North America and Asia were forest-covered when the first agriculturists penetrated meant that some procedures had to be immediately introduced to clear the land for cultivation. Although there is increasing evidence that hunter-gatherers had already initiated the process of deforestation through fire, pollen analyses have always shown that there were phases in prehistoric times when forest cover was rapidly depleted, and most of these occur at times when technological advances were particularly marked, or where other pressures must have built up through movement or increase of population. The use of timber in mining activities (for shoring and for smelting), in hill-fort building (for lacing the earthen ramparts and for walling), and in cremating the dead (for funeral pyres) must have depleted a forest cover already decimated in favoured areas by the need for larger and larger tracts of land for cultivation and grazing.

The clearance of forest may be said to represent the first major impact of man upon his environment, as it was the first step leading to a landscape controlled to a great extent by man. The clearance of such forest tracts has always been a subject of some concern to archaeologists and ethnographers who have tried to visualise the ease with which early man coped with his environment. Experiments in this particular subject have ranged from simple functional studies with reconstituted felling-axes, to sequences of work involving not only clearance but cultivation of experimental crops. Some of these have been closely related to pollen and other botanical investigations, in attempts to record the processes of vegetational change in altered areas.

Numbers of experiments concerned with the felling of trees have been conducted; perhaps the first was by the Vicomte Lepic who used a polished flint axe, hafted in an oak root, to chop down an oak tree (Evans 1897, 162). In Denmark, a trial clearance of 2000 square metres of oak forest was made in 1954 (Iversen 1956). Flint axes were used, hafted in accordance with surviving prehistoric hafts, and the trees were felled at knee height. Oaks of greater than 35 cm diameter were not felled but were girdled and left to die, and each of the smaller trees was chopped down in about 30 minutes. A small area of 500 square metres was cleared by 3 men in 4 hours, and from this it was concluded that one man could have cleared .2 hectare in a week.

A directly comparable experiment was recently carried out in Czechoslovakia in a variety of woods including pine (Pinus silvestris), spruce (Picea excelsa), alder (Alnus glutinosa), birch (Betula verrucosa), ash (Fraxinus excelsior) and willow (Salix petraca). It was found that a small tree, of 14–15 cm diameter, could be felled with a polished stone axe in only 7 minutes (Stelcl and Malina 1970). As in the Danish work, the method of chopping was with the axe held at an angle of about 50 degrees to the trunk of the tree, when a series of short blows would splinter off long chips; the stem of the tree when felled would have a sharpened end, like a pencil, and the stump at knee height would have a smashed top (Pl. 3).

Near Leningrad, an experiment with a polished axe of nephrite found on a Neolithic site showed that a pine of 25 cm diameter could be chopped down in 20 minutes (Semenov 1964, 130), and these figures are comparable to those recorded many years ago by Smith working on hafted flint blades and pinewood (Smith 1893).

Most of the comments made following these experiments have referred to the efficiency of stone axes, and there seems little doubt that such axes on soft woods are almost as effective as metal axes. A direct comparison between stone and steel has been made, however, that purports to show differently (Saraydar and Shimada 1971). A steel axe, weighing 2.3 kg, with a cutting blade 14 cm wide, was compared with a granite axe, 1.8 kg and blade 5 cm wide. The measurement made was of the kilocalories consumed by the manual effort in chopping at 24 strokes per minute for 5 minutes on a 15 cm diameter maple tree. The volume of air expelled was measured and its oxygen content was reckoned, leading to calculations of energy expended. The use of the stone axe consumed 5 times as many kilocalories as did the use of the steel axe, and the conclusion was that the stone axe took 5 times the energy, and 6 times as long, to accomplish the same work as the steel axe. The lack of detail given about precise methods of operation of the axes, and differences in blade thickness, as well as the differing widths and weights of axe, suggest that the work could be repeated with profit.

In the Yukon, recent observations on the felling of spruce trees by chiselling out splinters with wedges or stone axes in a band around the trunk, then pushing the tree over, show that the manner of use of stone axes may not have been that chopping action used by modern man with steel axes. In splintering spruce trees, a contest between stone and steel resulted in a win for the stone axe (Leechman 1950).

The few observations made about the rate of clearance of trees are not particularly useful. The Danish work suggested .2 hectare of clearance per man per week; in upper Canada in the eighteenth century .4 hectare could be cleared by one man with a steel axe in a week, the trees felled at knee height, but in a working year a man would rarely clear more than 2 hectares. He would have many other things to do, including burning the land, and it was as important there as in Europe to keep the cleared areas small, with trees around, to allow forest regeneration after land exhaustion. Work in central America has also emphasised these aspects.

If sufficient forethought was given to the process, and if time or energy was a problem, trees to be felled in a few years could be ringed and allowed to die before chopping down. This would have eased the woodchoppers’ task. Fire-setting would also have been a method of felling particularly large or hardwood trees. Shaw has recently pointed out that this is an effective procedure for truly enormous trees, and is still in use. In west Africa he observed that a large tree, 44 metres high, was impervious to all efforts with a steel axe that bounced off the hard wood (Shaw 1969). Fire-setting around the base of the previously ringed tree gradually burnt through, and after 60 hours the tree fell as planned by the locals. The effort was estimated as 6 hours of bark stripping, collecting wood and tending the fire. Windrow-felling would also be possible on a slope; the trees would be half chopped through so that when the uppermost trees were finally felled, their fall would knock down the neighbouring trees and so on. It would be a spectacular sight to see the whole hillside cleared at once, although in practice this never occurs completely.

When the efforts of prehistoric man at clearing forest are placed beside those of modern man, the impact of early man appears minute, but in some ways his approach to the problem has hardly been improved. One example must suffice. In the Mato Grosso of Brazil, vast areas of forest are in the process of being cleared so that cattle ranches can be established. The size of some of these ranches beggars the imagination (Fazenda Suia Missu 400000 hectares, Fazenda Dr Paulo 40 × 32 km), and the method of clearance has been described as a mixture of crude destruction and cunning. Using steel axes, men chop down most of the trees, leaving only a few; one man clears .4 hectare in 10 days. The wood is burnt in the dry month of August, burnt again the next year, and the next, and then coarse grass seed is strewn over the charred remains and tree stumps; the cattle, from zebu stock, are then driven in to graze in this inhospitable area. The scale of the operation is such that nearly 12000 hectares were cleared from one ranch in only 5 years, and 20000 head of cattle established. Beside this, the 1 or 2 hectares of Neolithic man seem puny indeed, but the method of clearance and working of the land is probably identical.

The Danish experiments continued with the burning of half of the cleared area (Steensberg 1955). The brushwood was spread evenly over the ground, and a 10-metre belt was fired, then the burning logs were rolled and pushed forward to the next strip of ground. Care was taken not to burn the ground too heavily, but it had to be heated thoroughly to clear it and to increase its fertility through incorporation of ashes and release of mineral constituents in readiness for the next stage; most clearances of forest involve burning of one sort or another, and controlled experiments have demonstrated the necessity for this in terms of crop yield. Observations on bush fires lasting several days indicate that the soil is heated, and weed roots destroyed, to depths rarely below 20 cm, and a flash fire racing through a lightly wooded area would hardly heat the soil below 7 cm. Clearly, the deliberate clearance of forest for cultivation of crops would necessitate carefully organised and controlled burning of felled timber to prolong the heat and destroy the weed roots.

In the Danish experiment, the area cleared was then hand-sown with wheat and barley. The seed was spread evenly onto the warm ground and raked in with a forked stick. The unburnt but cleared area was also sown in a similar manner. Other experiments in planting have been concerned with the seeding of ploughed ground, and uncertainties have been expressed over the depth of furrow (Reynolds 1967). It has been suggested that to broadcast seed into furrows perhaps 12 cm deep would be wasteful, yet to broadcast onto a ploughed and raked field would only feed a variety of animals, particularly birds. The Roman method of sowing was to broadcast by hand either above furrow or under furrow; seeds sowed onto a flat field would often be ploughed into furrows so that hoeing was possible. Raking would also cover the seeds.

In the Danish experiment, the seed was hand-sown and raked in. Both burnt and unburnt areas were hoed and weeded, and the resulting crops were then compared. The unburnt control area yielded hardly any crop at all, but the burnt area produced a strong crop. In the second year, however, the area’s yield was much reduced, and it seems from this that slash-and-burn cultivation of this type was only effective for a very short time before the land lost its fertility. Burning again, the introduction of cattle for manuring, manuring from a farmstead, and fallowing, would all have contributed to the field’s regeneration, and no doubt all of these methods were practised; no comparative experiments have been carried out on this vital aspect.

Work in Latin America has provided some confirmation of the results of clearance, burning and successive crops. In one experiment, 10 mecates (c. .5 hectare) of Yucatan forest was cleared without the use of metal tools (Hester 1953). The vegetation here is dense, with one estimate of between 100 and 200 woody plants per mecate. Wood up to 6 cm diameter was broken by hand, and untreated limestone lumps were used successfully to chop down larger timber except for some trees that were girdled. The stone tools showed no sign of wear after two weeks of work, and would probably have been unrecognisable by archaeologists as anything but natural lumps. The felled wood could have been burnt at almost any time in the year, preferably just after the rainy season, and this would have helped control the grasses, weeds and insects, and furnished ash with mineral salts for the crop.

Another experiment was concerned with the fertility of the soil which tends to be rather thin in much of the southern Maya lowlands (Steggerda 1941). A small experimental plot of one mecate was cleared and planted with corn for 8 successive years. For the first 4 years the field was irregularly weeded, in contemporary native fashion, and the yield by kilogram of shelled corn fell (32–28–16–7); as a control, the same field was weeded regularly for the next 3 years, and the yield was 34–15–24, The competition from weeds was clearly a factor, but more recent work has suggested that soil fertility depletion was probably the major agency in failing crops (Cowgill 1962). That this is, however, not a universal phenomenon has been shown by a lengthy experiment in Colorado (see below).

Observations among surviving groups in the area provide more reliable indications of cultivation procedures and population densities in earlier times. The cycle of shifting slash-and-burn cultivation in use suggests that where a cleared field is planted for one year only, it may be left to regenerate for 1–4 years; if 2 years of crop are obtained then a gap of 3–8 years is likely. Each farmer therefore needs large quantities of land, for both crops and fallow, and one estimate for this area is 1.2–1 .6 hectares per person; this leads on to densities of 150–200 per square mile (600–800 per 1000 hectares). Such observations, supported by experimental work, allow much firmer estimates of ancient behaviour patterns in these areas than we can hope to advance in areas where the ethnohistorical material is absent.

An early experiment in the growing of corn was carried out by Franke and Watson in Mesa Verde National Park, Colorado, over a period of years. The report of the experiments after 17 years drew some interesting conclusions (Franke and Watson 1936). In 1918, a small field, one hectare, of heavy red clay soil was chosen near a group of early Pueblo and Basket Maker sites, on the assumption that the same ground was likely to have been cultivated by these ancient settlers under conditions not unlike those of today; in these, rainfall was and is of extreme importance. The average annual fall from 1908–20 was 43 cm, varying from a drought of 25 to 56 cm; the fall occurs twice, February–March and July–December, and the driest season is June when the crops are at risk.

The field was cleared of brushwood, and fenced against animals and humans. Each year it was ploughed and planted in early May by Navajo Indians; the main crop was of calico corn or squaw corn, but beans and squash were also grown. The planting method was that used by the Navajos on their own farms. A hole about 38 cm wide and 13 cm deep is dug, and a smaller hole then dug in the base; the seed corn, 10–12 seeds, is dropped in and buried by soil filling only the small hole. When the young shoots emerge, they spread upwards and it is not until they reach the ground surface that the larger hole is filled in. The purpose of this method is to obtain deeper supplies of moisture for the seed. The small hills of earth that result from this process are 1–2 metres apart, and hoeing is done twice in the summer. Manuring of the field only occurred twice in 17 years.

In the early part of the experiment, planting sticks were used by the older Indians and ‘a prayer was tamped into every hill’. In later years, shovels were used by younger Indians, and the prayers were neglected. The absence of prayer had no noticeable effect on the yield, another interesting experiment.

Each year, germination tests of the crop were car...

Table of contents

• Cover
• Half Title
• Title Page
• Copyright Page
• Original Copyright Page
• Table of Contents
• Plates
• Acknowledgements
• Table of equivalents
• Introduction
• 1 FOOD PRODUCTION
• 2 HEAVY INDUSTRY
• 3 LIGHT INDUSTRY
• Conclusions
• Bibliographical references
• Index