University of Washington Archaeological Field School in Vietnam Summer 2015 Reader

University of Washington Archaeological Field School in Vietnam

Summer 2015

Reader

The Institute of Archaeology, Hanoi, Vietnam CULTURE SHOCK*

Kalervo Oberg

I would like today to make a few remarks about culture shock, a malady which I am sure has afflicted most of us here in varying degree. We might almost call culture shock an occupational disease of people who have been suddenly transplanted abroad. Like most ailments it has its own etiology, symptoms, and cure. Culture shock is precipitated by the anxiety that results from losing all our familiar signs and symbols of social intercourse. These signs or cues include the thousand and one ways in which we orient ourselves to the situations of daily life: when to shake hands and what to say when we meet people, when and how to give tips, how to give orders to servants, how to make purchases, when to accept and when to refuse invitations, when to take statements seriously and when not. Now these cues which may be words, gestures, facial expressions, customs, or norms are acquired by all of us in the course of growing up and are as much a part of our culture as the language we speak or the beliefs we accept. All of us depend for our peace of mind and our efficiency on hundreds of these cues, most of which we do not carry on the level of conscious awareness. Now when an individual enters a strange culture, all or most of these familiar cues are removed. He or she is like a fish out of water. No matter how broad-minded or full of good will you may be, a series of props have been knocked from under you, followed by a feeling of frustration and anxiety. People react to the frustration in much the same way. First they reject the environment which causes the discomfort: “the ways of the host country are

*Presented to the Women’s Club of Rio de Janeiro, Brazil, August 3, 1954.

Copyright, 1954, Kalervo Oberg All rights reserved bad because they make us feel bad.” When Americans or other foreigners in a strange land get together to grouse about the host country and its people – you can be sure they are suffering from culture shock. Another phase of culture shock is regression. The home environment suddenly assumes a tremendous importance. To an American everything American becomes irrationally glorified. All the difficulties and problems are forgotten and only the good things back home are remembered. It usually takes a trip home to bring one back to reality. Some of the symptoms of culture shock are: excessive washing of the hands; excessive concern over drinking water, food, dishes, and bedding; fear of physical contact with attendants or servants; the absentminded, far-away stare (sometimes called the tropical stare); a feeling of helplessness and a desire for dependence on long-term residents of one’s own nationality; fits of anger over delays and other minor frustrations; delay and outright refusal to learn the language of the host country; excessive fear of being cheated, robbed, or injured; great concern over minor pains and eruptions of the skin; and finally, that terrible longing to be back home, to be able to have a good cup of coffee and a piece of apple pie, to walk into that corner drugstore, to visit one’s relatives, and, in general, to talk to people who really make sense. Individuals differ greatly in the degree in which culture shock affects them. Although not common, there are individuals who cannot live in foreign countries. Those who have seen people go through culture shock and on to a satisfactory adjustment can discern steps in the process. During the first few weeks most individuals are fascinated by the new. They stay in hotels and associate with nationals who speak their language and are polite and gracious to foreigners. This honeymoon stage may last from a few days or weeks to six months, depending on circumstances. If one is a very important person he or she will be shown the show places, will be pampered and petted, and in a press interview will speak glowingly about progress, goodwill, and international amity, and if he returns home may well write a book about his pleasant if superficial experience abroad.

2 But this Cook’s tour type of mentality does not normally last if the foreign visitor remains abroad and has seriously to cope with real conditions of life. It is then that the second stage begins, characterized by a hostile and aggressive attitude towards the host country. This hostility evidently grows out of the genuine difficulty which the visitor experiences in the process of adjustment. There is maid trouble, school trouble, language trouble, house trouble, transportation trouble, shopping trouble, and the fact that people in the host country are largely indifferent to all these troubles. They help but they just don’t understand your great concern over these difficulties. Therefore, they must be insensible and unsympathetic to you and your worries. The result, “I just don’t like them.” You become aggressive, you band together with your fellow countrymen and criticize the host country, its ways, and its people. But this criticism is not an objective appraisal but a derogatory one. Instead of trying to account for conditions as they are through an honest analysis of the actual conditions and the historical circumstances which have created them, you talk as if the difficulties you experience are more or less created by the people of the host country for your special discomfort. You take refuge in the colony of your countrymen and its cocktail circuit which often becomes the fountainhead of emotionally charged labels known as stereotypes. This is a peculiar kind of invidious shorthand which caricatures the host country and its people in a negative manner. The “dollar grasping American” and the “indolent Latin American” are samples of mild forms of stereotypes. The use of stereotypes may salve the ego of someone with a severe case of culture shock but it certainly does not lead to any genuine understanding of the host country and its people. This second stage of culture shock is in a sense a crisis in the disease. If you overcome it you stay, if not, you leave before you reach the stage of a nervous breakdown. If the visitor succeeds in getting some knowledge of the language and begins to get around by himself, he is beginning to open the way into the new cultural environment. The visitor still has difficulties but he takes a “this is my cross and I have to bear it” attitude. Usually in this stage the visitor takes a

3 superior attitude to people of the host country. His sense of humor begins to exert itself. Instead of criticizing, he jokes about the people and even cracks jokes about his or her own difficulties. He or she is now on the way to recovery. And there is also the poor devil who is worse off than yourself whom you can help, which, in turn, gives you confidence in your ability to speak and get around. In the fourth stage your adjustment is about as complete as it can be. The visitor now accepts the customs of the country as just another way of living. You operate within the new milieu without a feeling of anxiety although there are moments of strain. Only with a complete grasp of all the cues of social intercourse will this strain disappear. For a long time the individual will understand what the national is saying but he is not always sure what the national means. With a complete adjustment you not only accept the foods, drinks, habits, and customs but actually begin to enjoy them. When you go on home leave you may even take things back with you and if you leave for good you generally miss the country and the people to whom you have become accustomed. Now before going on to consider the nature of culture shock, it might be well to point out that the difficulties which the newcomer experiences are real. If individuals come to a tropical area from a temperate one they quite often suffer from intestinal disturbances. Strange foods sometimes upset people. In Rio, for instance, water and power shortages are very real. When these physical difficulties are added to those arising from not knowing how to communicate and the uncertainties presented by strange customs the consequent frustrations and anxieties are understandable. In the course of time, however, an individual makes his adjustment, you do what is essential about water, food, and the other minutiae of daily life. You adapt yourself to water and power shortages and to traffic problems. In short, the environment does not change. What has changed is your attitude towards it. Somehow it no longer troubles you, you no longer project your discomforts onto the people of

4 the host country and their ways. In short, you get along under a new set of living conditions. Another important point worth considering is the attitude of others to a person suffering from culture shock. If you are frustrated and have an aggressive attitude to the people of the host country, they will sense this hostility and, in many cases, respond in either a hostile manner or try to avoid you. In other words, their response moves from a preliminary phase of ingratiation to aggressive ridicule and on to avoidance. To your own countrymen who are well adjusted you become somewhat of a problem. As you feel weak in the face of the host country people you tend to wish to increase your dependence on your fellow countrymen much more than is normal. Some will try to help you, others will try to avoid you. The better your fellow countryman understands your condition the better he is able to help you. But the difficulty is that culture shock has not been studied carefully enough for people to help in an organized manner and you continue to be considered a bit queer – until you adjust yourself to the new situation. In general, we might say that until an individual has achieved a satisfactory adjustment he is not able to fully play his part on the job or as a member of the community. In a sense he is a sick person with a mild or severe case of culture shock as the case may be. Although I am not certain, I think culture shock affects wives more than husbands. The husband has his professional duties to occupy him and his activities may not differ much from what he has been accustomed to. The wife, on the other hand, has to operate in an environment which differs much more from the milieu in which she grew up, consequently the strain on her is greater. In an effort to get over culture shock, I think there is some value in knowing something about the nature of culture and its relationship to the individual. In addition to living in a physical environment, an individual lives in a cultural environment consisting of man-made physical objects, social institutions, and ideas and beliefs. An individual is not born with culture but only with the capacity to learn it and use it. There is nothing in a new born

5 child which dictates that it should eventually speak Portuguese, English, or French nor that he eat with a fork in his left hand rather than in the right or use chop sticks. All these things the child has to learn. Nor are the parents responsible for the culture which they transmit to their young. The culture of any people is the product of history and is built up over time largely through processes which are, as far as the individual is concerned, beyond his awareness. It is by means of culture that the young learn to adapt themselves to the physical environment and to the people with whom they associate. And as we know, children and adolescents often experience difficulties in this process of learning and adjustment. But once learned, culture becomes a way of life, the sure, familiar, largely automatic way of getting what you want from your environment and as such it also becomes a value. People have a way of accepting their culture as both the best and the only way of doing things. This is perfectly normal and understandable. To this attitude we give the name ethnocentrism, a belief that not only the culture but the race and the nation form the center of the world. Individuals identify themselves with their own group and its ways to the extent that any critical comment is taken as an affront to the individual as well as to the group. If you criticize my country you are criticizing me; if you criticize me you are criticizing my country. Along with this attitude goes the tendency to attribute all individual peculiarities as national characteristics. For instance, if an American does something odd or anti-social in a foreign country which back home would be considered a purely individual act, this is now considered a national trait. He acts that way not because he is Joe Doaks but because he is an American. Instead of being censured as an individual, his country is censured. It is thus best to recognize that ethnocentrism is a permanent characteristic of national groups. Even if a national criticizes some aspect of his own culture, the foreigner should listen but not enter into the criticism. I mentioned a moment ago that specific cultures are the products of historical development and can be understood not by referring to the biological or psychological peculiarities of its human carriers but to an understanding of

6 the antecedent and concomitant elements of the cultures themselves. Brazil and the United States, for instance, have different cultural origins and different culture histories which account for present day differences. In this case, however, the differences are not great, both cultures being parts of Western civilization. It might be useful to recognize here that the study of culture per se is not the study of individuals. Psychology is the study of individual personality. Sociology is the study of groups and group behavior. The student of culture studies not human individuals but the inter-relationships of cultural forms like technologies, institutions, idea and belief systems. In this talk we are interested not so much in the study of culture as such, but its impact upon the individual under special conditions. Now any modern nation is a complex society with corresponding variations in culture. In composition it is made up of different ethnic groups, it is stratified into classes, it is differentiated into regions, it is separated into rural and urban settlements, each having its distinctive cultural characteristics. Yet superimposed upon these differences are the common elements of official language, institutions, and customs which knit it together to form a nation. These facts indicate that it is not a simple matter to acquaint oneself with the culture of a nation. Similarly the culture of one’s own nation is complex. It, too, differs by region and class. Americans, for instance, who go abroad in various governmental and business capacities, are usually members of the middle class and carry the values and aspirations of this class, some of which are an accent on the practical or utilitarian, work as a means to personal success, and suspicion of personal authority. Accustomed to working in large hierarchical institutions like business corporations, governmental agencies, or scientific foundations which have a life of their own and persist in time, Americans tend to become impersonal. Individuals, no matter how able, are replaceable parts in these large institutions. To Americans, personalism which emphasizes a special individual, like a political leader or a business leader or religious leader as solely responsible for the existence and success of an institution is somewhat strange. To the American it is the organization that

7 counts and individual beings judged according to their ability to fit into the mechanism. This difference in interpersonal relationships often comes at least as a minor shock. A new pattern has to be established which has to take into consideration class society, the symbols of individual status, the importance of family relationships, and the different importance given work, leisure, and the values people strive for. The rather sketchy remarks I have made here about culture and its elements is for the purpose of showing how important an objective treatment of your cultural background and that of your new environment is for understanding culture shock. There is a great difference in knowing what is the cause of your disturbance and not knowing. Once you realize that your trouble is due to your own lack of understanding of other people’s cultural background and your own lack of the means of communication rather than the hostility of an alien environment, you also realize that you yourself can gain this understanding and these means of communication. And the sooner you do this, the sooner culture shock will disappear. The question now arises, what can you do to get over culture shock as quickly as possible? The answer is get to know the people of the host country. But this you cannot do with any success without knowing the language, for language is the principal symbol system of communication. Now we all know that learning a new language is difficult, particularly to adults. This task alone is quite enough to cause frustration and anxiety, no matter how skillful language teachers are in making it easy for you. But once you begin to be able to carry on a friendly conversation with your maid, your neighbour, or to go on shopping trips alone you not only gain confidence and a feeling of power but a whole new world of cultural meanings opens up for you. You begin to find out not only what and how people do things but also what their interests are. These interests, people usually express by what they habitually talk about and how they allocate their time and money. Once you know this value or interest pattern it will be quite easy to get people to talk

8 and to be interested in you. When we say people have no interests we usually admit the fact that we have not bothered to find out. At times it is helpful to be a participant observer by joining the activities of the people, to try to share in their responses, whether this be a carnival, a religious rite, or some economic activity. Yet the visitor should never forget that he or she is an outsider and will be treated as such. He or she should view this participation as a role playing. Understanding the ways of a people is essential but this does not mean that you have to give up your own. What happens is that you have developed two patterns of behavior. Finally a word on what your fellow countrymen can do to help you get over culture shock. It is well to recognize that as the persons suffering from culture shock feel weak in the face of conditions which appear insuperable, it is natural for them to try to lean heavily on their compatriots. This may be irritating to the long-term resident but he should be patient, sympathetic, and understanding. Although talking does not remove pain I think a great deal is gained by having the source of pain explained, some of the steps towards a cure indicated, and the assurance given that time, the great healer, will soon set things right.

Reference: DuBois, Cora, Culture Shock. This talk was present as part of a panel discussion at the first Midwest regional meeting of the Institute of International Education in Chicago, November 28, 1951.

9 *Archaeologists Field Handbook 29/6/04 5:23 PM Page 32

CHAPTER TWO NAVIGATION AND

MAPPING W W

W W W

WHAT YOU WILL LEARN FROM THIS CHAPTER

W How to use maps to navigate to sites W How to understand scale W How to calculate a grid reference from a map W The uses of different kinds of maps W How to read bearings from a compass W How to use a GPS W Basic bush survival skills W Basic bush camping skills

Navigation and mapping are two of the most important skills in archaeological fieldwork. You will need these skills to get to and from sites, to avoid getting lost, and to be able to record the location of sites accurately. The two fundamental components of these skills are being able to understand a map and extrapolate from that map to the physical features you can see around you, and knowing how to read a compass. With these two skills mastered, you will be able to perform most of the basic surveying processes you will ever need in the field.

32 *Archaeologists Field Handbook 29/6/04 5:23 PM Page 33

NAVIGATION AND MAPPINGW33 USING MAPS

Maps are simply devices for transforming points on the curved surface of the Earth into a flat, two-dimensional plane. There are various methods for doing this, but the most common is the Universal Transverse Mercator (UTM) Grid system, which—as its name suggests—is a worldwide mapping system that divides the world into 60 equal zones from west to east. Australia covers twelve zones of this system, from 47 through to 58. These are the grid zone numbers depicted in the legend of every map. The area around Armidale, in New South Wales, for example, is designated ‘56J’ in the UTM system. Until very recently, most maps in Australia were based on the Australian Map Grid, a UTM projection system which was the best estimate of the Earth’s shape around the Australian continent (Geoscience Australia 2003). Recent updates on this system have produced the Map Grid of Australia 1994 (MGA94), which has been designed to supersede earlier AMG-derived maps. Because maps drawn at different times will have used different mathematical models of the Earth to flatten its curved surface, it is important to note the specific system used to generate each map (all maps will explain the system they use in the legend at the bottom or to one side). Because these systems do not match up, it is important to note which system you are using to calculate a grid reference, as the same grid reference under different systems will not refer to the same point in space. Likewise, it is important to note which system your GPS unit has been set to (see ‘Using a Global Positioning System’ below) to make the grid reference accurate. There are several different types of map that you will become familiar with in the course of archaeological fieldwork. The most common of these are topographic maps, geological maps and orthophoto maps. All of them use the same UTM grid system—they just plot different kinds of information. Topographic maps depict all the visible natural and built surface features, whereas geological maps depict rock formations, geological zones and soil types. Orthophoto maps combine aerial photography with topographic information. They can be very useful, but their coverage is usually restricted to major towns and their immediate environs. The maritime equivalent of maps is charts, which depict the depth of the ocean rather than the height of the land. Most maps are oriented with north to the top, south to the bottom, east to the right and west to the left. At the bottom of each map is the scale, a legend to explain the mapping conventions used to depict features and a small boxed section telling you how to read a grid reference. This box will also specify the grid zone designation for the map, which you will need to know when calculating a grid reference (see ‘Using a map to calculate a grid reference’ on page 36). Maps also come in a standard range of scales (although not all scales are available in all areas), from very small-scale maps which cover enormous areas (such as 1:1000 000 maps covering nearly 3000 square kilometres) to large-scale maps covering much smaller areas in greater detail (such as 1:25 000 scale *Archaeologists Field Handbook 29/6/04 5:23 PM Page 34

W34 THE ARCHAEOLOGIST’S FIELD HANDBOOK maps covering only 750 square kilometres). Scale is simply the ratio of the size of a feature as it is drawn on the map to its actual size. This ratio is given as a representative fraction, so that 1:100 means that one unit of measurement on the map (whether 1 millimetre, 1 centimetre or 1 metre) represents 100 of the same units when measured on the ground. Thus 1 millimetre on the map equals 100 millimetres on the ground, 1 centimetre equals 100 centimetres and so on (see Table 2.1). To make things even easier, the representative fraction is also converted into a linear bar scale at the bottom of the map to show you this relationship graphically. As you can see in Table 2.1, at scales of 1:250 or larger, something which is 50 centimetres long is far too small to plot accurately on to a map.

TABLE 2.1: Relationship between scale of map or plan and real distance Scale of plan Real measurement Scaled measurement 50 cm on the ground = 2 cm on the plan 1 m on the ground = 4 cm on the plan 1:25 2 m on the ground = 8 cm on the plan (1 cm on the plan = 3 m on the ground = 12 cm on the plan 25 cm on the ground) 5 m on the ground = 20 cm on the plan 10 m on the ground = 40 cm on the plan 50 m on the ground = 2 m on the plan 50 cm on the ground = 1 cm on the plan 1 m on the ground = 2 cm on the plan 1:50 2 m on the ground = 4 cm on the plan (1 cm on the plan = 3 m on the ground = 6 cm on the plan 50 cm on the ground) 5 m on the ground = 10 cm on the plan 10 m on the ground = 20 cm on the plan 50 m on the ground = 1 m on the plan 50 cm on the ground = 0.6 cm on the plan 1 m on the ground = 1.2 cm on the plan 1:75 2 m on the ground = 2.4 cm on the plan (1 cm on the plan = 3 m on the ground = 3.6 cm on the plan 75 cm on the ground) 5 m on the ground = 6 cm on the plan 10 m on the ground = 12 cm on the plan 50 m on the ground = 60 cm on the plan 50 cm on the ground = 0.5 cm on the plan 1 m on the ground = 1 cm on the plan 1:100 2 m on the ground = 2 cm on the plan (1 cm on the plan = 3 m on the ground = 3 cm on the plan 100 cm on the ground) 5 m on the ground = 5 cm on the plan 10 m on the ground = 10 cm on the plan 50 m on the ground = 50 cm on the plan 50 cm on the ground = 0.4 cm on the plan 1 m on the ground = 0.8 cm on the plan *Archaeologists Field Handbook 29/6/04 5:23 PM Page 35

NAVIGATION AND MAPPINGW35 Scale of plan Real measurement Scaled measurement 1:125 2 m on the ground = 1.6 cm on the plan (1 cm on the plan = 3 m on the ground = 2.4 cm on the plan 125 cm on the ground) 5 m on the ground = 4 cm on the plan 10 m on the ground = 8 cm on the plan 50 m on the ground = 40 cm on the plan 50 cm on the ground = 0.2 cm on the plan 1 m on the ground = 0.4 cm on the plan 1:250 2 m on the ground = 0.8 cm on the plan (1 cm on the plan = 3 m on the ground = 1.2 cm on the plan 250 cm on the ground) 5 m on the ground = 2 cm on the plan 10 m on the ground = 4 cm on the plan 50 m on the ground = 20 cm on the plan 1 m on the ground = 0.2 cm on the plan 1:500 2 m on the ground = 0.4 cm on the plan (1 cm on the plan = 3 m on the ground = 0.6 cm on the plan 500 cm on the ground) 5 m on the ground = 1 cm on the plan 10 m on the ground = 2 cm on the plan 50 m on the ground = 10 cm on the plan 1 m on the ground = 0.13 cm on the plan 1:750 2 m on the ground = 0.26 cm on the plan (1 cm on the plan = 3 m on the ground = 0.39 cm on the plan 750 cm on the ground) 5 m on the ground = 0.65 cm on the plan 10 m on the ground = 1.3 cm on the plan 50 m on the ground = 6.5 cm on the plan 1 m on the ground = 0.1 cm on the plan 1:1000 2 m on the ground = 0.2 cm on the plan 1 cm on the plan =) 3 m on the ground = 0.3 cm on the plan 1000 cm on the ground) 5 m on the ground = 0.5 cm on the plan 10 m on the ground = 1 cm on the plan 50 m on the ground = 5 cm on the plan

Scale will make a difference to how finely you can plot features onto a map or how accurate you need to be when measuring an area for a site plan. (See ‘Drawing horizontal surfaces (plans)’ on page 286). To illustrate the difference which scale can make, imagine you’re plotting the location of a site by placing a 1 millimetre dot on a map from the ballpoint pen in your pocket. On a 1:25 000 scale map, the diameter of your dot will be something like 25 metres, on a 1:100 000 scale map, its diameter will have increased to 100 metres, and on a 1:1000 000 scale map, the same dot will now be covering an area of some 20 kilometres. If you were then to assign that dot a grid reference, so that the location of the site was permanently recorded, obviously the grid reference determined from the 1:25 000 scale map would provide a much more accurate fix. Generally speaking, *Archaeologists Field Handbook 29/6/04 5:23 PM Page 36

W36 THE ARCHAEOLOGIST’S FIELD HANDBOOK a larger-scale map is better for survey purposes, as it will have a better resolution of surface features, making it easier and more accurate to plot sites. While most surveys can get away with plotting sites on a 1:100 000 scale map, 1:25 000 is by far the most accurate.

FIGURE 2.1: The relationship between the area of a map’s coverage and its scale

USING A MAP TO CALCULATE A GRID REFERENCE

When you look at any official published map, you will see solid black lines running vertically and horizontally across the map. These are the grid lines. The lines that are drawn vertically on the map and on which the figures increase in magnitude from west to east, are known as eastings; the lines that are drawn horizontally across the map and on which the figures increase in magnitude from south to north are called northings (Biddle et al. 1974: 39). Eastings represent the distance that each line is east of the start of the map zone. Northings represent the distance each grid line is below the equator. By using eastings and northings, it is possible to describe the location of any point on a map by *Archaeologists Field Handbook 29/6/04 5:23 PM Page 37

NAVIGATION AND MAPPINGW37 specifying the number of metres a point is east of the nearest vertical grid line and the number of metres it is north of the nearest horizontal grid line. This is called a grid reference and, depending on the degree of accuracy you want to achieve, it is possible to assign a grid reference which specifies a unique point on the surface of the Earth. To read or assign a grid reference, you must follow three simple rules:

1 Always read from the bottom left hand (southwestern) corner of the relevant grid square. 2 Always read the easting (the vertical line) before the northing (the horizontal line). 3 Always read eastings from left to right and northings from bottom to top.

This means that, when you are reading a grid reference from a map, you need to read the closest easting to the left of the site (because eastings are always read from left to right) and the closest northing below the site (because northings are always read from bottom to top). Because this is the convention, if you wrote down the easting incorrectly as being the closest line to the right of the site, for example, you would actually be indicating to the reader that the site was to the right of that line and giving a grid reference which placed the site too far to the east. Figure 2.2 illustrates this. The grid reference for a hypothetical heritage site on this map can be specified at various levels of precision. Because few maps are ever at a scale of 1:1 (that’s the point of maps: to reduce a large area so it fits on to a relatively small sheet of paper), the degree of accuracy you will achieve in plotting a site via a grid reference will depend both on the scale of the map you are using and the order of accuracy you give to the grid reference. It is perfectly possible to plot a site on a map using only a four-figure grid reference (see Table 2.2 on page 39). This is called an area reference but will only be accurate to a 1000 metre square around the site (i.e. all it will tell you is that your site occurs some- where within that 1000 metre square). In Figure 2.2, the area reference for the site is 2672. To define a more accurate position, you need to refine the scale of the map down by dividing that 1000 metre square into successively smaller squares (see Figure 2.2 on page 38). Each of the lines separating these smaller squares will be numbered from 1 through 9, beginning in the bottom left hand corner of the square. To calculate a six-figure grid reference, you initially follow the same process as for a four-figure grid reference, adding two extra steps (see below). In Figure 2.2, the six-figure grid reference for the site is 268723. While far more accurate than an area reference, this still only positions the site within a 100 metre square. To be even more accurate, you need to locate the site to within the nearest 10 metres. To do this, follow the same basic procedure, adding a few extra steps to give an eleven-figure grid reference. In Figure 2.2, the eleven-figure grid reference for the same site is 32687 807235. If you are working with a map of large enough scale (anything from 1:4 000 to 1:25 000), you will be able to be even more accurate and pinpoint the site *Archaeologists Field Handbook 29/6/04 5:23 PM Page 38

W38 THE ARCHAEOLOGIST’S FIELD HANDBOOK FIGURE 2.2: Calculating a grid reference

to the nearest metre. This will give a full grid coordinate for the site. By including the grid zone number and section letter from the map legend, you make the full grid coordinate unique, which means that no other point on the Earth will have this same reference (Biddle et al. 1974: 40). You can use the grid zone number and letter in conjunction with the 100 000 metre square designation and any area or grid reference to give such a unique reference. In Figure 2.2, the universal grid reference for the site is 55KCA 32687 807235. *Archaeologists FieldHandbook29/6/045:23PMPage39

TABLE 2.2: Steps in determining grid references To determine an area To determine a six-figure To determine an eleven-figure To determine a full grid reference (a four-figure grid reference grid reference coordinate (a universal grid grid reference) reference) • Locate the nearest vertical • Locate the nearest vertical • Locate the nearest vertical grid line • First identify the zone number and grid line (easting) to the grid line (easting) to the left (easting) to the left of the site section letter of the map (this is left of the site of the site • Read both the small black figures and given in the map legend as the grid • Read only the large black • Read only the large black the large black figures which label the zone designation) figures which label the line figures which label the line line (they will be on the top and the • Note the two letters which identify (they will be on the top (they will be on the top and bottom of the map, either beside or the 100 000 metre square in which and the bottom of the the bottom of the map, on the line itself) the site occurs (this will be given map, either beside or on either beside or on the line • Write these numbers down beneath the grid zone designation) the line itself) itself) • Divide the grid square which contains • Write these numbers and letters • Write these numbers down • Write these numbers down the site into ten equal vertical strips, down • Locate the nearest • Divide the grid square and identify the number of the • Locate the nearest vertical grid line horizontal grid line which contains the site into nearest tenth strip line to the left of (eatsing) to the left of the site (northing) below the site ten equal vertical strips, the site • Read both the small black figure • Read only the large black and identify the number of • Write this number down after the and the large black figures which figures which label the line the nearest tenth strip line small and large figures label the line (they will be on the • Write this number after to the left of the site • Subdivide this smaller square also top and the bottom of the map, the easting number • Write this number down into ten equal vertical strips and either beside or on the line itself) • The easting and northing after the large figures identify the number of the nearest • Write these numbers down after numbers together are • Locate the nearest tenth strip line to the left of the site the zone number and section letter your Area Reference horizontal grid line • Write this number down • Divide the grid square which (northing) below the site • Locate the nearest horizontal grid line contains the site into ten equal • Read only the large (northing) below the site vertical strips, and identify the black figures which label • Read the small black figures and number of the nearest tenth strip the line the large black figures which label the line to the left of the site • Write this number down line • Write this number down after the easting numbers • Write these numbers down after the • Subdivide this smaller square also • Divide the square into ten easting numbers into ten equal vertical strips and equal horizontal strips and • Divide the square into ten equal identify the number of the nearest identify the number of the horizontal strips, and identify the tenth strip line to the left of the site *Archaeologists FieldHandbook29/6/045:23PMPage40

TABLE 2.2: continued To determine an area To determine a six-figure To determine an eleven-figure To determine a full grid reference (a four-figure grid reference grid reference coordinate (a universal grid grid reference) reference) nearest tenth strip line number of the nearest tenth strip • Imagine that you are subdividing below the site below the site this square yet again and estimate • Write this number down • Write this number down how many of these tenths there are after the large numbers • Subdivide this smaller square also between the second strip line and • The easting and northing into ten equal horizontal strips and the site numbers together are your identify the number of the nearest • Locate the nearest horizontal grid grid reference tenth strip line below the site line (northing) below the site • Write this number down • Read the small black figures and • The easting and northing numbers the large black figures which label together are your grid reference the line • Write these numbers down after the easting • Divide the square into ten equal horizontal strips, and identify the number of the nearest strip line below the site • Subdivide this smaller square also into ten equal horizontal strips and identify the number of the nearest tenth strip line below the site • Write this number down • Imagine that you are subdividing this square yet again and estimate how many of these tenths there are between the second strip line and the site • Write this number down • All of the easting and northing numbers, together with the grid zone designation and the section letter, are your universal grid reference *Archaeologists Field Handbook 29/6/04 5:23 PM Page 41

NAVIGATION AND MAPPINGW41 The choice of which kind of grid reference to use will largely be a matter of how accurate you want your grid reference to be. Obviously universal grid references are far more accurate than area references. As a general rule, it is best to aim for the highest level of precision when recording any site location, even if you are not going to publish these precise grid references (for example, some Indigenous groups may request that you use an area reference rather than a universal reference in your published report so that the exact location of sites can be concealed). Before you decide on which kind of grid reference to use, it is also worthwhile checking with the relevant government heritage authority to find out its requirements. The New South Wales National Parks and Wildlife Service, for example, requires you to use an eleven-figure grid reference on all its standard site recording forms.

If you are calculating a grid reference for a large site, you can either specify the coordinates for the corners of a square drawn around the site, or simply give a single set of grid coordinates for the centre of the site.

Understanding contour lines Maps don’t just contain information about the type and location of features on the ground: they also give some indication of the rise and fall of the land, or its contours. The usual way to show this on a flat map surface is by using contour lines, or imaginary lines joining places of equal height. These are the thick and thin wavy lines you can see all over the map, some of which have small numbers attached to them at various intervals. On the map, each contour line represents a particular height above sea level, and the numbers will tell you how many metres above sea level each contour line is. The most important things to remember about contour lines are that:

• when they are close together on the map they indicate steep slopes; • when they are far apart on the map they indicate gentle slopes; • when they are uniformly spaced on the map they indicate uniform slopes; • when they decrease in spacing (when read from high to low), the slope is convex (outward sloping, like a hill); • when they increase in spacing (when read from high to low), the slope is concave (inward sloping, like a valley).

Understanding contour lines will help you to plan your surveys. In the initial decision- making stage, particularly if you have to estimate how much time and money the survey *Archaeologists Field Handbook 29/6/04 5:23 PM Page 42

W42 THE ARCHAEOLOGIST’S FIELD HANDBOOK FIGURE 2.3: Contour lines and slope

is likely to cost (see ‘How to prepare a tender’ on page 243), it will be very useful to know whether you are going to be walking up the side of a mountain or surveying around a gorge, as opposed to walking on relatively flat ground. Each of these terrains has specific features and challenges, as well as different potential for containing sites, and being aware of the rise and fall of the land is important to planning the amount of time it will take to conduct a survey.

MAKING MUD MAPS

One other kind of map which you may see referred to in reports and field notes is a ‘mud’ map. Strictly speaking, these are not really maps, but rough sketches of an area which are *Archaeologists Field Handbook 29/6/04 5:23 PM Page 43

NAVIGATION AND MAPPINGW43 not usually drawn to scale (i.e. they are not measured accurately). Mud maps are made in two main situations:

• to record the route you travelled to a site, so that you or someone else can find the site again; and • to give a general idea of the shape and context of a site if you don’t have enough time to record it properly but need to know the basic layout of the site.

Travelling mud maps are an important and relatively easy way of making sure you can relocate a site. They are made fairly quickly and only need to contain the information that is essential to finding your way back to a site. Usually, you will make these mud maps as you are driving to a site, and in this circumstance this is best done by a passenger. Travel- ling mud maps are something that may be used by other researchers at some stage in the future. These people will not have access to any local information that you may have in your head, so it is best if you are as explicit in your directions as possible. Some fundamental rules for making travelling mud maps are:

• Make sure that the core information relates to the decisions you make at turning points on the route. Every decision to turn should be recorded. • While mud maps are not made to scale, some indication of distance travelled between way points is essential. You can do this by simply noting the speedometer reading in the car, or by counting the number of paces you have walked. • Major features on the landscape should be recorded to indicate way points on the route. It is important that these features be permanent, like a house, rather than ephemeral, like ‘cows in a field’, or ‘beer bottle in fork of tree’. • If you are using buildings as way points, don’t just identify them in terms of their owners. The house that is currently owned by A. Gorman will in all likelihood be owned by somebody else in the future.

Mud maps of sites are made to give a general overview of what the site looks like. They are made when you don’t have time to record the site properly but still need some basic information. Mud maps turn abstract phrases such as ‘a large scatter of stone artefacts’ into a concrete image of the site itself. Again, these maps are made quickly and only contain the most significant information and will not be useful for in-depth research or for making critical management decisions (such as how close to the site a road can be built). For either of these purposes you will have to return to the site and make an accurate map, drawn to scale (see Chapter 4: Site surveying). If you are working by yourself, you will probably be establishing the dimensions of the site through pacing (see ‘Using the compass and pacing technique’ on page 90). Some fundamental rules for making mud maps of sites are: *Archaeologists Field Handbook 29/6/04 5:23 PM Page 44

W44 THE ARCHAEOLOGIST’S FIELD HANDBOOK FIGURE 2.4: A travelling mud map *Archaeologists Field Handbook 29/6/04 5:23 PM Page 45

NAVIGATION AND MAPPINGW45 • Include at least one permanent feature on the map, and all of the major features. • Show the approximate location of all archaeological materials, the extent of their area and any areas of damage. • Include basic information, such as a north arrow, your name and the date on which you made the mud map. • Dimensions can be included in approximate terms (e.g. approximately 20 ✕ 8 metres), but it is best if you pace this out. • Don’t forget to augment the mud map with a universal grid reference, photos and notes and never rely on a mud map as the only recording of a site’s location or its contents.

FIGURE 2.5: A mud map of a site *Archaeologists Field Handbook 29/6/04 5:23 PM Page 46

W46 THE ARCHAEOLOGIST’S FIELD HANDBOOK USING AERIAL PHOTOGRAPHS

Aerial photographs, as their name suggests, are literally photographs of a section of the Earth’s surface taken from the air. While aerial photographs were originally taken in order to map major environmental features, they also form a good basis for archaeological research and interpretation. An aerial perspective can be helpful in understanding past environments, as features such as recessional beach-lines and dunes are much easier to identify from a distant perspective than close up. This data can be integral to understanding relationships between environments and people in the past, especially if you are researching Indigenous settlement patterns. For the historic period, aerial photographs can provide valuable snapshots of changing demographic and settlement patterns. In the same way that a larger-scale map is better for identifying particular sites or features, the lower the level of the aerial photograph (i.e. the lower the plane was flying when the photograph was taken), the greater the level of detail it will provide. The first aerial photographs were taken in the 1930s and, if you can obtain a sequence of photographs for the same area taken at different times, you may well be able to literally see the major changes in your area over the last 50–60 years. Unfortunately, aerial photographs are limited in their ability to show the spatial layout of sites because only large archaeological features will be visible from the air, but they are very useful for identifying general areas which may contain sites and which therefore might be useful to survey. In Australia, aerial photographs can be obtained through the state departments that deal with topographic maps. Aerial photography can be used to:

• key into a surveying strategy for a region; • compare regional settlement patterns through time; • detect specific sites, such as the faint outlines of old settlements, buildings or roads; • relate your site to the surrounding environment; • assess relationships between sites within a region; • assist in regional environmental analysis; • assess a region’s potential for significant sites.

At a practical level, the best approach to aerial photography is to learn how to interpret photographs taken by others. The amount and kinds of information you can deduce from an aerial photograph will depend on your knowledge, training and experience. While you don’t need to be a specialist in geology or geomorphology, some experience in these fields can be helpful. It is best not to use aerial photographs for navigational purposes, as their scale can vary substantially. *Archaeologists Field Handbook 29/6/04 5:23 PM Page 47

NAVIGATION AND MAPPINGW47 USING A COMPASS

A compass uses a magnetised needle to indicate north. Following the principle that magnetic force is supplied by the molten iron at the Earth’s core, the coloured end of the needle (or the end with the letter ‘N’ on it) will point towards the north pole of the earth’s magnetic field. To align a compass to north, you simply turn the compass dial so that North (0°) on the dial matches the direction of the north end of the compass needle. Unfortunately, the Earth’s poles are constantly moving and the north pole indicated by the compass needle (‘magnetic north’) is not the same as the real north pole (‘true north’). This compass error is called declination and, depending where on the Earth you are, can result in a difference of anything from a few degrees to up to 20–30° away from true north. Fortunately, on the bottom of all topographic maps there is a simple chart which indicates the yearly degree of declination away from true north in any area (see Figure 2.6 on page 48). You should bear in mind, however, that because the poles are constantly moving, this declination is also constantly changing and may be inaccurate on older maps. True north (TN) is the direction towards Earth’s geographic north pole, magnetic north (MN) is the direction in which the compass needle points, and grid north (GN) is simply the direction of all the vertical grid lines on a map. Grid north is always used as the reference point for bearings and grid references.

Remember that a compass will point to anything metallic or magnetic that is made of iron. This means that it cannot be used accurately if you are standing beside wire fence- lines or scrap metal, inside cars or under or beside electrical power lines. Even metal objects on your person can affect your compass reading.

Taking bearings The direction in which you want to travel to get to your destination is called a bearing. Because there are 360 degrees in a circle, compass bearings are referred to by the number of degrees they are away from north, always counting in a clockwise direction. For example, east is one quarter of the circle away from north (which translates to 90°), south is directly opposite to north or halfway around the circle (at 180°), and west is three- quarters of the circle away from north, at 270°. The numbers on the moveable compass dial are these degrees. *Archaeologists Field Handbook 29/6/04 5:23 PM Page 48

W48 THE ARCHAEOLOGIST’S FIELD HANDBOOK FIGURE 2.6: Using a topographic map to calculate the difference between true and magnetic north

How to read a compass bearing • Point the compass towards your destination. If you are using the compass to make a plan of a site, this will be the point to which you want to measure (see ‘Using the compass and pacing technique’ on page 90), otherwise it will be the direction in which you want to travel. • Keeping the compass fixed on your destination, move the dial so that 0° is aligned to the magnetic compass needle (which will be pointing towards magnetic north). • The bearing of the point, or the distance it is away from north, can now be read off the compass dial (remembering always to read the numbers on the dial clockwise from 0°). This will be slightly different depending on the type of compass you are using (see Figure 2.7 on page 49). • For greater accuracy, use your arm to extend the directional arrow, or sight the compass on a ranging rod, or other highly visible feature, held at your destination. *Archaeologists Field Handbook 29/6/04 5:23 PM Page 49

NAVIGATION AND MAPPINGW49 • When you reach your destination, you should always take a back-bearing to your original location. This is a check on the accuracy of your original bearing. The difference between your original bearing and your back-bearing should always equal 180°, although an error of plus or minus 2° is often unavoidable. Taking and recording accurate bearings and back-bearings with a compass should allow you to plot your movements accurately and will be invaluable if you are navigating across country.

FIGURE 2.7: Taking a compass bearing. Using a prismatic compass (top) you must look through the compass to sight on your destination; using a protractor-type compass (bottom), you must sight along the compass using the arrow at the tip as a guide. *Archaeologists Field Handbook 29/6/04 5:23 PM Page 50

W50 THE ARCHAEOLOGIST’S FIELD HANDBOOK If you are using a compass to navigate your way across country, particularly if you are moving through dense vegetation, it is important to make continuous checks to ensure you have not moved away from your bearing. One way to do this is by continually checking your compass. However, by far the best way to maintain a bearing is to select a clearly visible object which lies along your path (such as a tree or a rock outcrop) and move towards it. If you can’t find any obvious natural features, then send someone on ahead to act as a marker. Continue to do this for each leg of your journey. This process is called point-to-point bearing. If you are using a map and a compass in conjunction to navigate towards a site, you will need to compensate for the declination. In other words, compass bearings taken on the ground must be converted into grid bearings before they can be plotted on to a map. Likewise, grid bearings taken from a map must be converted into compass bearings before they can be used with a compass on the ground. To convert between bearings, simply either add or subtract the degrees of magnetic variation shown on the declination chart at the bottom of the map. For example, using the chart in Figure 2.6, imagine that you have just taken a magnetic bearing from your compass towards a scarred tree. The bearing reads 50°. To convert this to a grid bearing which you could then plot on to your map, you would simply add 7 degrees (because, over time, magnetic north has moved so that it is currently 7° east of grid north). Conversely, if you had taken a grid bearing of 50° from the map and wished to use it to navigate towards the scarred tree, you would first have to convert the grid bearing to a magnetic bearing. To do this you would subtract 7 degrees, thus giving you a magnetic bearing of 43° to follow on your compass. If you are using a GPS you can set your GPS to give you directions in magnetic degrees, thus eliminating the need to worry about compensating for declination (see ‘Using a Global Positioning System’ on page 51). Once you have converted your compass reading to a grid bearing, you can use a protractor to measure this bearing directly on to a map. Like a compass, the outer rim of a protractor has degrees marked around its edge (either from 0°–180° if the compass is a half circle, or from 0°–360° if the compass is a full circle or a square). To measure a bearing, draw a light pencil-line from the point where you are standing to the point to which you are navigating. Place the protractor on the map so that the centre point or hole is directly over where you are standing and the north line (or 0°) is pointing to grid north on the map (i.e. the top), parallel to the eastings. Read the degrees off the edge of the protractor where it meets your pencil line.

How to orient a map Whenever you use a map in the field for navigation, make sure you have the map correctly oriented to start with. This means ensuring that the directions on the map *Archaeologists Field Handbook 29/6/04 5:23 PM Page 51

NAVIGATION AND MAPPINGW51 correspond to the directions of the same features on the ground. The easiest way to do this is simply to look for any readily identifiable features around you and turn the map so that it and you are facing in the same direction relative to those features. If you cannot identify any surrounding features, and are still unsure how to orient the map, you can use a compass. • Lie the map flat. Place your compass on it and adjust the vertical lines on (grid north) to match up with magnetic north on the compass (the swivelling arrow needle). • Check the declination in the legend at the bottom of the map and make an adjustment to take into account the variation between magnetic north and grid north. For example, if the declination on your map is specified as 5°, you would need to shift the compass slightly so that its north arrow is pointing to a spot which is 5° east of grid north. • The map will now be oriented correctly.

USING A GLOBAL POSITIONING SYSTEM (GPS)

Global Positioning Systems have greatly simplified the process of navigating to and from sites and of recording site locations. A GPS can be used either on land or on water. It calculates a grid reference for your position from orbiting satellites, and operates on the same map projection systems as official published maps. When using a GPS on land, make sure that you set the map system to match that of your maps. If your GPS readings are taken using a different map system to your printed maps, then your grid references will not plot accurately. Thus the Australian Geodetic 66 or 80 system will be most appropriate for older maps and the WGS84 (World Geodetic System) will suit more recent maps. If you are having persistent problems in plotting your GPS readings, this is one of the first things to check. When using your GPS on water, you will have to use latitude and longitude to plot location instead of eastings and northings, as this is the system employed on all charts.

Tips for using your GPS efficiently • Most satellites orbit the equator, so when using a GPS in Australia hold the unit facing north so that your body doesn’t interfere with the satellite signal. • When using a GPS to navigate to a known grid reference, walk along until you have correctly located yourself along either the easting or the northing, and then follow that until you reach the correct position. *Archaeologists Field Handbook 29/6/04 5:23 PM Page 52

W52 THE ARCHAEOLOGIST’S FIELD HANDBOOK • When you first turn on your GPS receiver, it will ‘lock on’ to all available satellites, provided the receiver has a clear and uninterrupted view. It is often worth leaving the GPS on for a few minutes before using it for serious work, as sometimes the initial readings can be misleading.

The best low-end models are the Garman and Magellan units, both of which are accurate to within 10–15 metres and are easily affordable. They are also water and shock resistant, as well as light and easy to carry. Because the degree of accuracy of your GPS will depend partly on the brand, make sure you always note down the make and model of your GPS in your field notes and report. GPS units are simple to use and can be a great asset in the field, provided you keep some basic factors in mind. Most can be set to either magnetic north (good if you are navigating using a compass) or true north (good if you are navigating using a map). Make sure you know which one you are using. Because GPS units can be used all over the world, they also have to be set to the relevant map zone to give an accurate grid reference. This means that, even if you only travel between the Northern Territory and New South Wales to use your GPS, you will have to readjust the map zone. Since it is possible for errors to creep into any GPS reading (as a result of factors such as the number of satellites which are accessible, their geometry, atmospheric conditions or signal reflection), it is always wise to check every reading against a topographic map to be certain of your location. And because GPS units are prone to running out of battery power at the most inopportune times, or may be unable to get a fix on a sufficient number of satellites for an accurate reading (particularly if you are surveying under dense tree cover), it is always wise to know how to calculate a grid reference manually (see ‘Using a map to calculate a grid reference’ on page 36). If you plan to use a GPS under heavy tree cover, you should investigate buying one with an external aerial.

Stephen Sutton’s tips for successful fieldwork • Always have a good pair of boots. I recall a report at a conference in which the presenter admitted that part of the study area hadn’t been surveyed ‘because there were bindies and we didn’t have any boots’. • Take good care of your field assistants. It’s essential to the success of fieldwork that your assistants are well fed and that their health and safety are assured. • Plan your fieldwork. You need to be aware that the only data you have is what you collect. By this I mean that when you get back to the lab, and are working up your notes or collection to find something missing—too bad. It’s almost always *Archaeologists Field Handbook 29/6/04 5:23 PM Page 53

NAVIGATION AND MAPPINGW53 impossible to reconstruct from memory (either yours or your colleagues’), or to find surrogates for the missing data. The only way to do it is to return to the scene and this is expensive and time-consuming. • There is no ‘one way’ to do anything. Whilst careful attention to data collection is fundamental, it should also be appropriate. Work should be structured to ensure that the final product—the reason you’re in the field in the first place—can be achieved, but after that there is a capacity to waste resources collecting data that will never be used. • Don’t rely on technology. There are digital machines for everything these days: GPS, laptop computers, EDMs, data-loggers, digital cameras and so on. They all have an amazing capacity to fail at the critical moment. There is no substitute for traditional skills like map-reading and surveying and the use of back-up materials like rag-paper survey notebooks.

FINDING NORTH USING YOUR WATCH

In case you are ever stuck in the field without either a compass or a GPS, you can use any analogue watch to find north. This is not accurate enough to be able to map a site (see ‘Using the compass and pacing technique’ on page 90) or to navigate to a site, but might be enough in a pinch to prevent you from being lost. To find north using your watch, point the number 12 on your watch towards the sun. North will always lie mid-way between the hour hand and the number 12 (see Figure 2.8 on page 54). If you have only a digital watch, make a sketch of an analogue watch and draw the correct time on to it, leaving off the minute hand. Use the sketch in the same manner as a real watch to locate north.

SURVIVAL IN THE BUSH

Archaeologists often travel in remote areas on tracks, dirt roads and off-road. This can be highly dangerous, particularly if you are not prepared or are not being sufficiently careful. Never go into the bush unprepared—archaeology is not worth dying for. Most advice for working in the bush is common sense. Make sure you take enough water, particularly if you are going to be surveying large areas and doing lots of physical activity (you should always aim to drink the equivalent of between eight and ten glasses of water a day). If you are going to a remote area, ensure that you have spare tyres for your vehicle, emergency food rations and some means of reliable communication (such as a UHF radio or satellite phone, not just a mobile phone). Make sure that you tell someone exactly where you are going and when you expect to return. *Archaeologists Field Handbook 29/6/04 5:24 PM Page 85

CHAPTER FOUR SITE

SURVEYING

W W

W W W

WHAT YOU WILL LEARN FROM THIS CHAPTER

W The basic principles of a chain survey W How to produce quick and accurate site plans W How to keep your errors to a minimum W How to set up and use an automatic (‘dumpy’) level W How to record survey information in your field notebook and on recording forms

The basic surveying toolkit At least two 25 metre, 30 metre or 50 metre tape measures (larger sizes are best) At least two 5 metre, 8 metre or 10 metre retractable tape measures (larger sizes are best) Compass Nylon builder’s line/string Plastic clothes pegs (for fixing a tape measure to a string baseline) Tent pegs, or wooden stakes (for fixing the ends of a baseline) Mallet Plumb-bob Ranging rod Drawing equipment (see Chapter 9)

85 *Archaeologists Field Handbook 29/6/04 5:24 PM Page 86

W86 THE ARCHAEOLOGIST’S FIELD HANDBOOK Optional Levelling instrument (automatic level, EDM, theodolite, total station) Tripod Stadia rod or prism staff Walkie-talkies

THE BASICS

A site plan is the simplest and most effective way to record spatial information about a site. Even if there is only enough time to draw a sketch plan, you can still use this to convey the basics of the extent, form and main features of the site. A clear and accurate site plan is an essential part of any site recording (see ‘What to do when you find a site’ on page 80).

Making your survey as accurate as possible Obviously you can’t just go and make a series of totally unconnected measurements at a site and then expect to be able to use them to draw up a coherent site plan. It is no use knowing that the barn is 5.8 metres from the fence if you have no idea where the fence is in relation to anything else. Or knowing that the fence and the barn are 5.8 metres apart if you didn’t measure how long the fence was or where, precisely along its length, the barn was located. Even if you made sure to measure the position and dimensions of every feature by moving systematically around the site (‘the house is 12 8 metres, the fence is 3 metres from the southeast corner of the house and is 25 metres long, the barn is 5.8 metres from the ninth fencepost from the northern end’), if you have made one mistake at any point in this sequence, then every other subsequent feature which you measure in relation to this will also be out of position. At the end of the exercise you will still not be able to create an accurate site plan and all your time will have been wasted. The most important element in an accurate survey is to ensure that the location of all features can be tied together in such a way that there are no ‘floating’ measurements and the sources of error can be kept to a minimum. This can only be done through the use of a fixed reference point from which each feature is measured, and therefore to which the location of each feature can be securely tied (this is called a site datum). The datum may be a wooden peg which you hammer into the ground, or some other fixed feature of the site. This datum is often related to an arbitrary line that you establish through your site for the purposes of survey. Such an arbitrary line is called a baseline, and if possible one end of this baseline should be fixed to the datum point. Any measurement from a baseline to a feature is called an offset. *Archaeologists Field Handbook 29/6/04 5:24 PM Page 87

SITE SURVEYINGW87 Your baseline may be a long tape measure (this is certainly simplest, because the distance along the baseline is easy to read off the tape), or a string-line of known length which you have fixed to pegs in the ground. Where you place it will depend on the size and shape of your site; because it is an arbitrary line which you are using for convenience, it doesn’t matter where in the site it is located, as long as you can plot most or all of the features from it. If you are recording a stone arrangement, for example, you might choose to fix the baseline through the centre of the arrangement so that you could conveniently plot all of the stones on either side of it. Similarly, if you are recording a collection of farm buildings, you may be able to use a fixed fenceline (provided you have measured it first) as your baseline. If your site is particularly large or spread out, you may need to establish more than one baseline to be able to reach all features. In this case, you could use the position of the first baseline to establish the position of the second and so on, but bear in mind here that if you have made any errors in establishing the position of the second baseline, then any subsequent baselines will also be out and there is a real likelihood that your errors will be compounded. To try to control this process of cumulative error, the most reliable means of surveying an area is to set up a series of baselines around the perimeter to form a framework which encases the site (see Figure 4.1 on page 88). This uses all the same principles of measurement as a single baseline, simply repeated around each side of the framework until all the details have been plotted in. When establishing a framework, you need to keep a few basic rules in mind:

• Use as few lines as possible to keep your errors to a minimum (any form of quadri- lateral is ideal). • Make sure that the framework is ‘rigid’—in other words, that your starting point is also your end point. • Make sure that you check the accuracy of your framework before you begin detailed measurement by not only measuring the length of the outside perimeter, but also the diagonals between points. This is why surveyors commonly divide an area to be surveyed into triangles, then measure all of the sides (Hobbs 1983: 44–45). These diagonals are essential checklines which will help to keep your framework rigid. • Try to keep the angles of the triangles in your framework between 30° and 120°. If your angles are greater than 120°, it is easy for errors to creep in.

By establishing a framework first, you give yourself a method for keeping a constant check on the accuracy of your plan without running the risk of getting bogged down in the details. A good framework will use the key elements of a site to establish the main spatial relationships between features. Once you have established this, you will easily be able to check how accurately you have positioned the details in relation to these features *Archaeologists Field Handbook 29/6/04 5:24 PM Page 88

W88 THE ARCHAEOLOGIST’S FIELD HANDBOOK FIGURE 4.1: A framework of baselines, the essential elements of any site plan

and in this way your errors should be kept to a minimum. This is what surveyors mean when they use the maxim: Always survey from the whole to the part. This simply means establishing your baseline or framework first and only then measuring in the detail of particular features (see Figure 4.1). When establishing any baseline, either alone or in a framework, there are several rules to bear in mind:

• All measurements along a baseline should begin from the same point (i.e. the same end of the baseline). This will be less problematic if you use a tape measure as a baseline, because all measurements will begin at ‘zero’. • Always measure along the baseline first and only then to the feature, and always record your measurements in the same order. • Do not move or disturb the baseline until all measurements from it are complete, including any last-minute checks you need for accuracy.

Keeping your errors to a minimum The biggest problem with all forms of survey is the potential for error to creep into even the most careful measurements. When you are making a series of measurements, it *Archaeologists Field Handbook 29/6/04 5:24 PM Page 89

SITE SURVEYINGW89 is very easy for one measurement to be slightly out, thus making the next slightly out, and so on. If these errors are allowed to accumulate, you could end up with serious problems. This is why the other basic premise of all surveying is to constantly be on the alert for potential errors, and to keep these to a minimum wherever possible. This may be something as simple as making sure that all sides of your framework can be placed in areas which are free of excessive vegetation or other obstacles, making sure that everyone is familiar with the tape measures you will be using so that no mistakes will be made in calling out measurements, or making sure that all measurements are recorded consistently in a log book. You will have to accept that some margin of error will be inevitable in all surveys (unless you are lucky enough to be using state-of-the-art electronic surveying equipment), but always aim to keep this within acceptable limits (see Table 4.1 on page 113). The only way to achieve this is to make independent checks on your measurements as often as possible. Remembering the tendency of surveyors to divide their survey areas into triangles, the simplest way to make an independent check on your framework is to not only measure the sides of your framework, but also to measure and plot the diagonal checklines between corner points (see Figure 4.1 on page 88). Obviously this will not always be possible, particularly if there is vegetation in the way or if you are plotting a site from a central baseline only (see ‘Using the baseline/offset technique’ on page 96). In this case, you can make independent checks on the position of key features by plotting them from more than one point along the baseline (see Figure 4.1 on page 88). You should do this for all major features of your site anyway, and for any features which are particularly important to your research goals and thus which need to be plotted accurately.

TECHNIQUES FOR CONSTRUCTING A SITE PLAN

The simpler forms of site plan use basic orienteering skills which anyone can master; the more complex require particular items of equipment which may not always be available. The two simplest and least technical methods for constructing a site plan are:

• compass and pacing; • baseline and offset.

The advantage of these techniques is that they require little in the way of equipment and can be used in any situation. If you have to carry your equipment a long way to reach a site, there are obvious advantages in using simple techniques requiring basic equipment. There are disadvantages, too, in that these methods will not give the same accuracy as more sophisticated survey equipment, such as an automatic level or a total station/EDM. *Archaeologists Field Handbook 29/6/04 5:24 PM Page 90

W90 THE ARCHAEOLOGIST’S FIELD HANDBOOK The choice of which method to use will once again come down to time and resources. Probably the best way to deal with this dilemma is to decide how detailed your plan needs to be. If you were mapping a site which was going to be destroyed, then you would want to record everything in minute detail, as this may well be the only recording which will ever be made of this site. In this case it would be best to use some form of electronic distance- measuring device, such as a total station, as this would give you a highly accurate and detailed result. If, on the other hand, you simply wanted to produce a plan showing the main physical features of a site as part of public record or for publication (bearing in mind that most plans are greatly reduced in size for publication and that, as the scale becomes smaller, the ability to depict finer units of measurement decreases—for more information see ‘Using maps’ in Chapter 2 on page 33 and ‘Drawing horizontal surfaces (plans)’ on page 286), then it is probably more effective to use a simpler and less time-consuming technique. Whichever method you choose, make sure that you include an accurate description of it when writing up the methods in your report (for more information, see Chapter 10: Getting your results out there). If at all possible, it is best to plot your results into a site plan while you are in the field (i.e. at the same time that you are taking measurements). This way you will be able to catch any mistakes as they arise and will be able to check on compass readings or particular measurements immediately. If you leave the site and then wait for a few days before you try to turn your measurements into a site plan, you run the risk of lumbering yourself with mistakes which you may not be able to correct. Of course, this may not always be practical (particularly if you are recording a site by yourself), but try to draw up your plan as soon as possible while it is still fresh in your memory.

USING THE COMPASS AND PACING TECHNIQUE

This technique uses the length of your own stride to measure distance and a compass to plot direction. Its great advantage is that it enables one person to create a reasonably accurate plan of a small area relatively quickly. Before you can employ this technique in the field, however, you need to calculate the length of your stride, or your pacing unit. This is the average length of a single step and varies from person to person. It also varies for the same individual, depending on whether you are walking uphill (when your pace will be shorter) or downhill (when your pace will be longer). It may even vary at the beginning and the end of each day according to how tired you are. The compass and pacing technique is really only accurate for small areas, although the basic principles (i.e. taking a bearing with a compass to give direction and then pacing the distance to give a measurement in metres) can also be useful when navigating (to keep track of where and how far you have walked) or to give you the rough dimensions and *Archaeologists Field Handbook 29/6/04 5:24 PM Page 91

SITE SURVEYINGW91 spatial relationships of a site when drawing a mud map. Your pacing unit will be useful any time you need to make quick approximate measurements.

How to calculate your pacing unit To do this, you will need an area of relatively flat ground. Measure out a length of 25, 50 or 100 metres (you are actually going to calculate your pacing unit as an average of 100 metres, but if you can’t find that much room lay out a smaller length and multiply it accordingly later) and walk this distance at least ten times. Because your unit will be an average, the more times you do this, the more accurate your final pacing unit will be. Note the number of paces it takes you to walk this distance each time. Add these up and divide by the number of times you walked the distance to average the number (if you have used a smaller measured length of 25 or 50 metres, you will need to then multiply the average by four or two to arrive at a figure per 100 metres). Divide 100 by this average to give you your pacing unit. For example, if you’ve walked a length of 100 metres twenty times and the sum of your paces came to 2500 then, divided by 20, this would give you an average of 125 paces per 100 metres. Divide 100 by this average to equal a pacing unit of 0.80, or 80 centimetres. To use your pace as a unit of measurement, simply pace out the dimensions of whatever you are measuring and multiply by your pacing unit to arrive at a figure in metres. If your pacing unit was 0.80, for example, then a feature which you measured to be 6.5 paces in length would be in reality 5.2 metres long. Remember to maintain a comfortable walking pace as you pace out your measured length and whenever you use your pacing unit to measure features at a site. If you exag- gerate your steps, you will never obtain an accurate or replicable measure of your pace. You should also perform this exercise on sloping ground (on around a 15–20 per cent slope) to give you a unit for measuring up or down hills.

How to plot an area using the compass and pacing technique • Work out a framework for your survey, such as the outer boundaries of the site, so that you can plot in all of the detailed features from this framework. Use as few lines as possible, but make sure that your framework is rigid—in other words, that the starting and end points are the same. If you leave your framework open, you will have no way to check on the accuracy of your measurements. Each side of this framework is called a traverse. • Beginning at the starting point (Point A), take a compass bearing (see ‘Taking bearings’ on page 47) to the second point (Point B) on the framework by aligning the compass either on a feature (if this is what you have used as a point), or on a ranging rod held *Archaeologists Field Handbook 29/6/04 5:24 PM Page 92

W92 THE ARCHAEOLOGIST’S FIELD HANDBOOK at the point. In surveying terms, such a bearing is often referred to as a foresight. Note down the bearing and count the number of paces it takes you to walk to that point. Make sure you walk in a straight line, noting the distance along the line of any features which cross it and the nature of these features. When you reach the second point you need to take a back-bearing (in surveying terms, a backsight) to the starting point as an independent check. Remember that the difference between your foresight and your backsight should equal 180° (see ‘Using a compass’ on page 47), although an error of plus or minus 2° is perfectly acceptable. If the difference is greater or less than 180° you will have to check your bearings and possibly take them again. It’s important that you get this part of the exercise right, otherwise your plan will not be accurate. If you are still having trouble, look around for any sources of magnetic interference which might be present, such as power lines or fences. • Continue around the framework taking bearings to each new point, noting the distance and nature of features along each bearing and taking back-bearings (see Figure 4.2 on page 93), until you arrive back at your starting point (Point A again). This is called closing the traverse. Once you have completed your traverse, you can take bearings and measurements to any additional features located inside or outside the framework from any of the points on the framework. • Plot your plan on to graph paper. Using a scale rule, work out a suitable scale so that you can fit all of the features on to the paper (you may need to roughly sketch in the dimensions and shape of your framework by hand to make sure that it fits on the page), aligning magnetic north to the vertical lines on the graph paper. • Beginning at your starting point (Point A), use a protractor to plot your first compass bearing on to the graph paper (see ‘Taking bearings’ on page 47). Convert the number of paces to metres by multiplying the number of paces by your pacing unit and measure this distance onto the page along the protractor bearing. Continue this process for all of the points around the framework. Because inaccuracies are inevitable in both the compass readings and the pacing measurements, it is unlikely that your traverse will close (i.e. that your starting and end points will plot on to the graph paper in the same location). Because of this, we will call the end point (which is really the same as the starting point) Point A1. The distance between your original starting point (Point A) and end point (Point A1) is called the error of closure (see Figure 4.3 on page 94). The best level of accuracy that can be achieved by a compass and pacing survey is approximately 1:300, or an error of 1 metre in every 300 metres. The error of closure should not exceed 1:100 (Davies and Buckley 1987: 141). This means that, if the total length of your survey framework was 500 metres, an error of closure of 5 metres is the maximum acceptable error. Anything less than this is better of course, though it is unlikely that your error of closure would be less than 1.6 metres. *Archaeologists Field Handbook 29/6/04 5:24 PM Page 93

SITE SURVEYINGW93 FIGURE 4.2: When recording the information from a compass and pacing survey, it is standard practice to start at the foot of the page and proceed upwards, making sure that the figures for bearings and distances cannot be confused. A line is drawn across the page to signify the end of each leg, and details of features on either side of the route entered to the right or left of the central column of measurements depending on their location. *Archaeologists Field Handbook 29/6/04 5:24 PM Page 94

W94 THE ARCHAEOLOGIST’S FIELD HANDBOOK FIGURE 4.3: Plotting a compass and pacing survey

• Before you can correct this error, much less start plotting details on to your survey plan, you need to make sure that the error of closure for your traverse is within the acceptable limit. Add up the total length of your traverse (i.e. add the distances for each leg of your framework). Divide this number by your error of closure distance (i.e. the distance between Point A and Point A1 as plotted on the graph paper). If you arrive at a figure between 100 and 300 you are within the acceptable limits. • To correct for an error of closure, first draw a line on your plan between Point A and Point A1 (remembering that they should actually be the same point) and measure the distance between the two. Using a set square, draw lines parallel to this line through each point on your framework (you will need these lines later to correct the positions for each point) (see Figure 4.3). • Now you need to construct a special linear scale to illustrate the error of closure between your starting point and your end point. To do this, draw a straight line which is equal to the total length of your traverse. Obviously, because you are drawing this on to a sheet of graph paper, you will have to use a suitable scale to fit this line on the paper. The scale will have to be much smaller than that used for plotting the framework and will really have to be smaller than 1:1500 to work properly. A total traverse *Archaeologists Field Handbook 29/6/04 5:24 PM Page 95

SITE SURVEYINGW95 length of 300 metres, for example, could be drawn at a scale of 1:1000 (or 1 centimetre = 10 metres) as a line which is 30 centimetres in length. The left-hand end of this line represents your starting point (Point A), the right-hand end your end point (Point A1) (Figure 4.4). Make sure you mark off the position of each point in your traverse along this line. • On the right had side of this line, draw a perpendicular line equal to the length of your error of closure distance (remember to reduce it to the same scale you used to represent the total traverse length). In the example above, for instance, say the error of closure distance was 5 metres. To equal this distance at a scale of 1:1000, you would draw a perpendicular line 5 millimetres long at the right-hand end of your scale. • Join the top of this vertical line back to the starting point (so that you have a triangle) and, working from the left-hand side, draw vertical lines marking off the position of each point in your traverse. The top of these lines will become points B1, C1, D1, etc. These distances represent the corrections necessary to adjust the error of closure at each of these points. FIGURE 4.4: Correcting the error of closure *Archaeologists Field Handbook 29/6/04 5:24 PM Page 96

W96 THE ARCHAEOLOGIST’S FIELD HANDBOOK • You will notice that the length of the vertical lines increase as you move towards the right-hand end of the scale; this is because they indicate the cumulative error which increases for each leg of your survey. The length of these lines also indicates the distance of each plotted framework point from its correct location—simply use the length of each line to measure the correct location for each point on your plan. On the traverse map, draw lines through B, C and D, etc. that are parallel to the line you have already drawn through A and A1. Use the corrections established above to determine the positions of B1, C1, D1. Remember always to plot the corrected position for each point in the same relationship as that for your starting-point and end-point. In other words, if your end-point plotted to the southwest of your starting point (meaning that the correct location is really to the northeast), make sure that you plot the corrected position for every other point to the northeast also. Join A to B1, C1, D1 and then back again to A. Your traverse should now close (see Figure 4.4 on page 95). If not, you have serious problems and may have to start again. • You can now plot in the details of your plan, remembering always to include the names of the people who made the map, the date, site, a scale, a north arrow and a legend if you have used symbols (such as shading or cross-hatching) to depict detail. Because your compass readings will indicate magnetic north only, you should also use the degree of declination from a topographic map (see ‘Using a compass’ on page 47) to indicate true north.

USING THE BASELINE/OFFSET TECHNIQUE

This technique uses the same principles as a compass and pacing survey (i.e. measuring the location of features from an established baseline) and requires little more in the way of equipment than a couple of long and short tape measures. Because it requires measuring to and from the baseline, however, it requires the labour of more than one person.

• First lay out a baseline. A long tape measure is best for this, as the distance along the baseline can easily be read off the tape as the survey progresses. Give careful consideration to where you lay out your baseline—it should be aligned in such a way that most (if not all) of the features can be measured from it without having to lay out another one. If you are plotting the ruins of a building, for instance, the baseline would be best laid out through the centre and running down the long axis of the building, so that walls and other features on both sides can be measured from it. You can lay out the baseline on or above the ground (i.e. fixed to free-standing pegs for instance), but it must be kept straight and as horizontal as possible. Laying the baseline flat on the ground is simplest, because raised baselines tend to stretch and sag, although a *Archaeologists Field Handbook 29/6/04 5:24 PM Page 97

SITE SURVEYINGW97 perfectly horizontal baseline is not always possible. Once again, the decision of precisely how much variation from the horizontal is permissible comes down to how accurate you need your survey plan to be—obviously every deviation from the horizontal intro- duces an element of error, so try to keep this to a minimum. If you are drawing a plan of a rockshelter, for example, the floor may be quite undulating and you may need to locate the baseline some way from the floor of the shelter to keep it horizontal. • Remember, once the baseline is fixed, don’t move it until all of your measurements are complete. • Take a compass reading along your baseline (it doesn’t matter in which direction, but for consistency’s sake from the 0 metre point on the baseline is best) and note down the reading in your field notes. This will allow you to indicate the direction of magnetic north on the finished plan. • You can now begin to measure offsets from the baseline to the various features of the site. • The most important thing to remember about a baseline/offset survey is that all features must be measured at right angles to the baseline. You must keep all offsets at 90° to the baseline to ensure that you are measuring the shortest distance between the baseline and the feature. If your angle varies above or below 90°, then the distance you are measuring will also increase or decrease, giving you an inaccurate measurement. • Because they have the baseline extending to either side of them, the person standing at the baseline is the one most able to judge accurately when an offset is at the correct angle. For anything over 3 metres, you will have to use a different technique. There are three ways you can do this, all of which use basic principles of geometry to establish a right angle:

1. bisecting an arc; 2. 3–4–5 triangle; 3. triangulation.

FIGURE 4.5: Over a short distance (less than 3 metres), you can estimate a right angle for an offset reasonably accurately by eye, but only from the position of the baseline *Archaeologists Field Handbook 29/6/04 5:24 PM Page 98

W98 THE ARCHAEOLOGIST’S FIELD HANDBOOK METHODS FOR MEASURING RIGHT-ANGLED OFFSETS 1: Bisecting an arc

FIGURE 4.6: Bisecting an arc

To use this method, get one person to hold the end of a tape measure firmly on the point which is being measured, while you stand at the baseline and swing the other end of the tape measure over it in a short arc. As you swing the tape measure over the baseline, you will notice that the distance increases as the tape reaches either end of the arc, but lessens towards the centre of the arc. It is this shortest distance which you are looking for, because this will indicate when the tape measure is at a right angle to the baseline. If you are not confident that you can work this out by eye, then mark each end of the arc where it crosses the baseline, measure the length of this distance on the baseline and then divide it in half. This halfway point marks the corner of the right angle.

2: 3–4–5 triangle This method relies on the 3–4–5 ratio of a right-angled triangle (see Figure 4.7). Basic- ally if the measurements for each side of a triangle are always kept in units of 3, 4 and 5, or any multiple of these (for example 30 cm–40 cm–50 cm, 60 cm–80 cm–100 cm), then the angle between the two perpendicular sides of the triangle will always be 90°. For example, you want to lie out a second baseline perpendicular to your first. At the *Archaeologists Field Handbook 29/6/04 5:24 PM Page 99

SITE SURVEYINGW99 FIGURE 4.7: 3–4–5 triangle

offset point on the first baseline (where the second baseline will begin), you fix the end of one tape measure. You measure along the first baseline for a distance of 1 metre. This will be the base of the right-angled triangle. Holding the end of a second tape measure over this point, you give the ends of both tape measures to another person who moves away from the baseline until they are standing at roughly 90° to the offset point. You already know that for the triangle to be a right-angle these two measurements must equal 60 and 80 centimetres respectively, so keep adjusting both tape measures until you have them at the correct length. The point where they cross at the correct lengths is the other end of your right-angled offset.

3: Triangulation As its name suggests, this method uses triangles to plot the location of features. This is best done using two tape measures. Measure a known length on the baseline and hold or peg a tape measure at each point. Cross the other ends of the tape measures over the point you are measuring to and note down the length of all three sides of the triangle, as well as its position on the baseline. Because you know the length of all *Archaeologists Field Handbook 29/6/04 5:24 PM Page 100

W100 THE ARCHAEOLOGIST’S FIELD HANDBOOK FIGURE 4.8: Triangulation

three sides, as well as its precise location along the baseline, you will be able to plot this triangle accurately on to your survey plan. This method is best used to plot features which are a long way off the baseline, but ideally the angle at the apex of the triangle (i.e. at the measured point) should be kept as close as possible to 90°. It should definitely be no smaller than 40° and no larger than 140° (Hobbs 1983: 46). You may, of course, use more than one of these methods at the same site (for instance, for short measurements you might estimate by eye, for longer measurements you might prefer to use triangulation). It doesn’t matter: just make sure that you record each measurement in your field notes for when you come to draw up your site plan. Make absolutely sure that you always record your measurements in the same order—that is, reading along the baseline first, and then along the offset to the feature. If you confuse the order of these readings at any time, you will not be plotting features in their correct locations.

It is much easier to draw up a baseline and offset survey as you go along—in other words, to have two people measuring (one always at the baseline and one always at the features) while a third person simultaneously plots the position of each feature on to graph paper and constructs the plan. Obviously the draftsperson will need to work out a suitable scale for the plan before any measuring begins, and will also be the person responsible for literally ‘joining the dots’ or making the plan come together. As the draftsperson, you have the responsibility of keeping a sharp eye on how the plan is progressing—you are in the best position to notice if distances seem wrong or if features *Archaeologists Field Handbook 29/6/04 5:24 PM Page 101

SITE SURVEYINGW101 don’t plot where they should. Don’t FIGURE 4.9: Using a plumb-bob to establish your just blindly trust the measurements position on the baseline being given to you. The measurers have no overall scheme in front of them to see the relationships between each point, but you do. Evaluate each measurement in the context of the plan as it is progressing on paper and don’t be afraid to ask for measurements to be repeated or for extra measurements to be taken. In this sense, it is really the draftsperson who directs the survey, not the other way around. If your site is large or very complex and can’t be measured using a single baseline, you may need to establish a survey framework which is, in effect, a series of four baselines with the diagonals measured in to make the framework rigid (see ‘The basics’ on page 86). The baseline and offset technique is most accurate on level ground, but if there is any slope you will need to take care to keep your baselines horizontal. You will also have to take care with your offset measurements, because if any features are located considerably above or below the horizontal plane of the baseline (i.e. if they are on parts of the site that slope away from the baseline), you will not be able to simply measure along the ground to these points. If you did, you would be measuring up or down the length of the slope which would give you a longer measurement than would a properly horizontal offset. In this case you will need to hold the offset tape measure above the baseline so that it is roughly horizontal with the location of the feature being measured and use a plumb-bob dropped below the tape to find your exact position on the baseline (see Figure 4.9).

USING THE BASELINE AND OFFSET TECHNIQUE TO RECORD VERTICAL SURFACES

The baseline and offset technique is also useful if you are recording standing structures or other vertical archaeological features, or if you need to create a cross-section through *Archaeologists Field Handbook 29/6/04 5:24 PM Page 102

W102 THE ARCHAEOLOGIST’S FIELD HANDBOOK a site such as a rockshelter (see ‘Recording rockshelters’ on page 221). In these cases, instead of your offsets being measured from a baseline across to a feature, your offsets will be measured above or below the baseline. Once again, you will need to give careful thought as to where you place your baseline to ensure that all of the major measurements can be made from it. You also need to make sure that your baseline is kept horizontal and that all measurements above or below it are kept as close to vertical as possible.

A case study in recording a cross-section through a rockshelter FIGURE 4.10: Hypothetical use of the baseline and offset technique to create a cross-section of a rockshelter

The aim here was to produce a cross-section through the shelter to show the changes in the height of the ceiling and the slope of the floor of the rock shelter from front to back. This is important for looking at how people have used the spaces within the shelter and will help to make sense of the features recorded in the site plan. First, a horizontal baseline was laid out from the front to the back of the shelter. The front of the shelter was about 35 centimetres below the level at the rear of the shelter, so the baseline needed to be fixed to a ranging rod located outside the shelter. Because there were artefacts scattered across the floor of the shelter, the ranging rod had to be placed well outside the front of the shelter so as not to disturb any potential sub-surface deposits. The end of the tape measure at the rear of the shelter was simply weighted down with a large rock (checked out first to make sure it was not an artefact). A series of measurements was made above and below the baseline to record the changing levels in the roof *Archaeologists Field Handbook 29/6/04 5:24 PM Page 103

SITE SURVEYINGW103 and floor. For measurements of ceiling height, a retractable tape-measure was used as a plumb-bob (the end held up at the point on the ceiling being measured) and the other end simply dropped down immediately beside the baseline to obtain a reading of height. The same technique was used to record the slope of the floor at the front of the shelter.

A case study in recording the wall of a ruined building The aim here was to record the details of the construction techniques used in the wall and the extent of damage. A string baseline was established across the width of the wall about 80 centimetres above the level of the ground. Because the wall is only 1.5 metres high, this was simply a convenient height from which to measure. The baseline was fixed to the wall using nails inserted firmly into the mortar (making sure not to damage the wall in the process), and a tape measure pegged along the string. Measurements were made both above and below this baseline to the various features on the wall. FIGURE 4.11: Hypothetical use of the baseline and offset technique to create a vertical ‘plan’ (or elevation) of a wall *Archaeologists Field Handbook 29/6/04 5:24 PM Page 104

W104 THE ARCHAEOLOGIST’S FIELD HANDBOOK BASIC PRINCIPLES OF LEVELLING

Using the baseline and offset technique to measure changes in the height of a rockshelter is a form of levelling, or surveying rise and fall. It is only useful over small areas, however, and indicates a major problem with the basic methods of site surveying. They can only provide you with a reasonably accurate plan of the horizontal layout of a site; none of them can give you an accurate indication of how the land rises or falls, or how steep or flat the area containing the site is. To document this, you will need to be able to record levels: the changing height of the ground across the site (Drewett 1999: 66). The principle of levelling is very simple. It involves projecting an imaginary horizontal plane across the site and measuring the height of the ground above or below this. Surveyors refer to this as the line of collimation, and it is exactly the same in principle as the horizontal baseline you use when drawing vertical surfaces (see ‘Using the baseline and offset technique to record vertical surfaces’ on page 101). It can only be measured with the proper equipment, however—such as an automatic level (often called a ‘dumpy’ level), theodolite or Electronic Distance Measuring device (EDM). You will have seen surveyors using these and similar equipment—they are designed to be set level on a solid tripod, and the height at various points across the site read off a stadia rod (a telescopic staff with units of height marked in alternate red and black segments) through the telescopic lens of the level. If you look through the telescope, you will see one vertical and one major horizontal cross-hair. The horizontal cross-hair represents the line of collimation and, by reading the changing height of the staff along this line, you are reading whether the ground is rising or falling. FIGURE 4.12: Surveying the changing height of the ground across a site *Archaeologists Field Handbook 29/6/04 5:24 PM Page 105

SITE SURVEYINGW105 How to set up an automatic or ‘dumpy’ level • Step 1: Establish the location of the instrument. This is important, because ideally you want to be able to take as many readings as possible without having to move the instrument. Try to find a centralised spot from which all parts of the site are visible. Once you have decided on this, set up the level. First erect the tripod (the telescopic legs which form the base). If you are working as part of a large group, you should erect the tripod to the height of the shortest user. In any case, make sure the tripod is at a comfortable height for constant use, and lightly but firmly tamp the legs into the ground. Don’t make them immovable yet, however, because the next step will be to make sure that the instrument itself is perfectly level. Make sure that the head of the tripod (where you will be shortly attaching the instrument) is roughly level and doesn’t have an obvious tilt in any direction. Now attach the instrument. The level will have a base plate (see Figure 4.13 on page 106) that can be screwed into the top of the tripod. Don’t screw this in tightly yet, but make sure that the level is firmly fixed to the tripod and can’t slide off. Now you have to level the instrument. As long as you have only lightly screwed the instrument in place, the slightly convex surface that is the head of the tripod will allow you to slide it around in a tight circle. Note the effect this has on the levelling bubble (Figure 4.13) and see if you can get the instrument close to horizontal. All surveying equipment will have such a visual means for you to judge how level they are—usually a centrally located air bubble inside a marked ring on the circular base of the instrument. You need to get the dumpy at least close to being level at this stage (i.e. the bubble needs to be almost within the circle if not completely inside it), then tighten the screw holding the level in place. This probably will change the level of the instrument slightly, but don’t worry as you can adjust this next. If you can’t get the instrument anywhere near level at this stage you will have to rethink the positioning of the tripod legs and check whether any need lengthening or shortening. One of the ways to adjust the gross level of the instrument is to tamp the individual legs of the tripod more firmly into the ground. Keep an eye on the levelling bubble as you do this. Most automatic levels use a combination of three large foot screws (Figure 4.13) at the base of the instrument to make them perfectly level, although some may have an internal levelling mechanism. These foot screws are designed to be used in pairs (imagine a triangle underneath the instrument that can be raised or lowered slightly on each side). Align the telescope so that it lies parallel with one pair of foot screws and, using both hands, turn these two screws outwards (i.e. in opposite directions towards the edges of the instrument). Note the movement of the bubble as you do so. If you had the instrument approximately level in the first stage, you will eventually succeed in getting the air bubble exactly in the centre of the bull’s-eye circle. Sometimes this process is deceptively quick and simple, sometimes it seems as if you’ll never get it *Archaeologists Field Handbook 29/6/04 5:24 PM Page 106

W106 THE ARCHAEOLOGIST’S FIELD HANDBOOK FIGURE 4.13: The major components of an automatic level

right, but you must persevere. Swing the telescope through 90° so that it lies directly over the third foot screw and make sure the bubble is still perfectly centred. Swing the telescope over each foot screw and check that the bubble remains centred. • Step 2: Once the instrument is steady and level, mark the location with a permanent fixture such as a wooden peg. This must be located directly beneath the instrument, so you will need to tie a plumb-bob to the hook underneath the centre of the tripod head and position the peg directly underneath it. Mark the top of this peg with an indelible cross. This is your first survey station. • Step 3: Measure the height of the instrument above ground with a tape measure, and record this in your field notes. • Step 4: Begin taking readings. The first thing to do is to make sure the telescope is focused correctly. In most models, the eyepiece will be surrounded by a rotatable dial that focuses the cross-hairs (Figure 4.13). Use this to make sure the cross-hairs are sharply defined. Use the peep-sight along the top of the telescope to approximately align it with the staff and sight on the staff through the eye piece. The horizontal fine motion control knob (or pair of knobs) on the telescope’s circular base will allow you to shift the telescope incrementally left or right until it is perfectly aligned with the staff (Figure 4.13). Use the focusing knob on the right-hand side of the telescope’s body to bring both the cross-hairs and the face of the staff into perfect focus (Figure 4.13). The first reading you will have to take must be a backsight (BS) to the site datum to establish the height of the line of collimation. Hold the staff at the datum, take the reading at the central horizontal cross-hair and then add it to the known height of the datum. If the site datum has a value of 60 metres above sea level for instance, and the reading gives a height of 1.4 metres for the staff, then the height of the line of *Archaeologists Field Handbook 29/6/04 5:24 PM Page 107

SITE SURVEYINGW107 FIGURE 4.14: How to set up a level *Archaeologists Field Handbook 29/6/04 5:24 PM Page 108

W108 THE ARCHAEOLOGIST’S FIELD HANDBOOK collimation is 61.4 metres. If you don’t know the precise height above sea level of your datum then you will have to assign an arbitrary height to it (for instance, 100 metres. Don’t make it zero or you’ll end up with negative numbers). The sum of the first reading and the datum height is the value for the line of collimation (otherwise known as the height of the instrument [HOI]). An automatic level also has the facility to read degrees from north so that each reading can be aligned to a particular compass bearing. As part of your first backsight, and before you move the telescope for the next reading, take a compass reading on the staff along the same axis as the telescope to determine a bearing for the backsight. For example, this might be 270°. Rotate the large circular dial at the base of the instrument so that its reading in degrees aligns with your compass bearing (i.e. so that you can read 270° through the index window underneath the eye piece) (Figure 4.13 on page 106). When you’ve realigned the telescope for the next reading, you’ll be able to read the bearing for that position by simply reading the degrees shown in the index window. Once you have taken your backsight, all subsequent readings (to various features on the site) are called intermediate sights,orinter-sights (IS). For each inter-sight, subtract the reading at the central cross-hair from the height of the instrument to give you the reduced level (RL) for that spot. You can also use an automatic level to measure distance as well as height and so give you a plan of a site. Inside the dumpy’s telescope you will see two smaller cross-hairs above and below the major cross-hair denoting the line of collimation. If you subtract the height of the staff at the lowest cross-hair from the height at the highest cross-hair and multiply by 100, this will give you the distance from the level to the staff in metres. If you find that it is impossible to cover the entire site from one position (e.g. if there is heavy vegetation cover, or the site rises or falls too steeply for it all to be equally visible), then you will have to move the level to a new location. This follows the same setting-up principles as before, with one additional step. The last reading you were able to take from the present survey station will become your first foresight (FS). Don’t move the staff from this location while the level is being moved. The rule is that the staff must remain stationary while the level is being moved and the level must remain stationary while the staff is being moved (Casey 1972: 15). Move the level to the new survey station (obviously you will have chosen this carefully so that you can see new parts of the site from the new location, but are still within sight of the last foresight). Set it up again following steps 1–3 and then take a backsight to the location of the last foresight. If you don’t take this reading you will be unable to tie the different parts of your survey together in the final plan. Calculate the new line of collimation for the second survey station by adding the reading for the backsight to the reduced level for that spot (which you calculated from the previous foresight reading). Continue your survey. *Archaeologists Field Handbook 29/6/04 5:24 PM Page 109

SITE SURVEYINGW109 FIGURE 4.15: How to fill in a level booking sheet *Archaeologists Field Handbook 29/6/04 5:24 PM Page 110

W110 THE ARCHAEOLOGIST’S FIELD HANDBOOK How to fill in a level booking sheet This is relatively straightforward and lets you keep track of your readings as you go along. It is also the source data that you will use to plot your survey as a plan, so it is important that you do it correctly. A sample level booking sheet is included in Appendix 1. The first column on the sheet is a description of the location of the dumpy level (these are your survey stations). This assumes that you may have to move the instrument, so the various positions it is set up in can simply be designated 1-100 or A-Z. Note the height above sea level of your survey area (if you know it) in Column 8. If you don’t know this, choose an arbitrary level of 100 metres (so you don’t end up in negative numbers as the ground falls). This will be your first reduced level. Remembering that the first reading after setting up a survey station will always be a backsight (either to the site datum or to the point of the last foresight), this is entered in Column 2. There will only be one backsight per survey station, so don’t enter any further readings in this column unless you have just moved the level. Don’t make the common mistake of writing the foresights in this column—this will only confuse things later. The backsight will establish the line of collimation (i.e. how high the instrument is), which you calculate by adding it to the reduced level (in this case the site datum) to get a height in metres. Place this figure in Column 7. Align the dumpy to the correct compass bearing for the backsight (see ‘How to set up an automatic or “dumpy” level’ on page 105) and note the bearing in Column 11. The next readings you will take will all be inter-sights to the various features of the site you wish to plot. Place the staff on a feature and take a reading on the central (major) horizontal cross-hair. Place this figure in Column 3. The difference between this figure and the height of the instrument in Column 5 will give you the reduced level for that location. Record this figure in Column 8. Enter the value for the upper cross-hair in Column 4 and for the lower cross-hair in Column 5. Enter the bearing in Column 11.

You can use the upper and lower cross-hairs visible through the dumpy’s telescope to check the accuracy of your survey readings. The difference between the value of the lower cross-hair and the centre cross-hair should be the same as the difference between the value of the centre cross-hair and the upper cross-hair. For example, if the reading at the centre cross-hair is 1.275, and at the lower cross-hair 1.240, then the reading at the upper cross-hair should be 1.310. Each is 0.035 metresdistant from the centre cross- hair and they need to agrees to within 0.005 metres or better. If you can’t get your readings to agree, you’ll have to take them again and check for potential sources of error. Is the staff vertical? Are you reading the staff correctly? If you are new to using a dumpy level you should check your readings regularly to minimise errors. *Archaeologists Field Handbook 29/6/04 5:24 PM Page 111

SITE SURVEYINGW111 Calculate the horizontal distance to that spot, so that you will be able to plot it on to your plan later (subtract the value of the lower cross-hair from the value of the upper cross-hair and multiply by 100). Place this figure in Column 10. To check the accuracy of your survey, subtract the sum of the foresights from the sum of the backsights. Then subtract the last reduced level from the first. The answers should be the same. If you’re taking many readings and using several pages of the booking form, you can check each page separately by making sure that each begins with a backsight and ends with a foresight. If an inter-sight comes at the end of a page, enter it as a foresight on that page and as a backsight on the next (Hobbs 1983: 53).

Tips for successful levelling • The most important point is to make sure that the instrument is properly level before you use it. If it isn’t, then none of your readings will be accurate. • Once you have levelled the instrument, don’t kick or disturb the tripod. If you do, you’ll have to re-level the instrument before you continue, including taking a new backsight to the site datum and FIGURE 4.16: Using a spirit level to check calculating a new line of collimation. whether the stadia rod (levelling staff) is • If you are holding the staff, then you vertical have several responsibilities. First, you need to make sure that you are holding it upright. By looking through the telescope, the surveyor at the dumpy level will be able to see if you have tilted it to the right or left and can indicate to you in which direction you should move it. They won’t know whether you have tilted it forwards or backwards, however. To compensate for this (it can be very difficult to know whether the staff is truly upright when you are holding it), you can rock it slightly backwards and forwards so that the person at the dumpy can take the highest reading (which will be the horizontal). You can also use a small carpenter’s line level, held against the back of the staff to judge when it is upright (see *Archaeologists Field Handbook 29/6/04 5:24 PM Page 112

W112 THE ARCHAEOLOGIST’S FIELD HANDBOOK Figure 4.16 on page 111). Staffs for EDMs and total stations have a built-in air bubble so you can make sure they are level. Second, you will have to ensure that you have not rotated the face of the staff away from the telescope. Watch the direction in which the staff is facing and be prepared to adjust it if the surveyor can’t read it clearly. • If the surveyor can’t see you or the staff at all (say, if there is screening vegetation in the way) then you will have to move the staff slightly in one direction or the other until it becomes visible. This is where walkie-talkies are invaluable. Sometimes only a slight adjustment will be necessary, but take care that, as you move the staff, you keep it upright and vertical. This is most easily done by moving the staff in small steps (sometimes only centimetres) in a given direction until the surveyor tells you to stop. • To produce a contour plan, you need to grid the site and take spot height readings at each point on the grid. You then need to decide on the contour interval (the distance between contours) and join points of equal height across the site.

With the advent of more sophisticated forms of surveying equipment such as Elec- tronic Distance Measurement (EDM) instruments, it is possible to create extremely accurate site plans. An EDM uses a built-in transmitter to send out a laser beam towards a reflecting prism attached to a staff held at the position which is being plotted. The length of time taken for the beam to reach the prism and return to the EDM forms the basis for the calculation of distance. These measurements are all recorded digitally for later download into a computer. A total station has both a theodolite (to measure levels) and an EDM (to measure distance) as part of the one instrument. It can therefore give you both horizontal and vertical measurements as part of the same plan. As a final word, no survey will ever be perfect. You will need to decide on the level of accuracy you require before you begin—a decision which will largely depend on why you are recording the site in the first place and what it is you want to know. For any site survey, there are acceptable levels of error which you should take note of (see Table 4.1 on page 113). For a plan plotted at a scale of 1:100, for instance, it is perfectly acceptable for measurements to be taken to only the nearest 50 millimetres; for a plan at 1:1000, measurements need only be to the nearest 500 millimetres. As the scale becomes larger, the level of accuracy increases, so at 1:10 you should try to keep your measurements to the nearest 5 millimetres, for example. The levels of accuracy shown in Table 4.1 also will have a bearing on the final scale at which you can draw your site plan. Remember, if you are recording a large site, the entire plan of which will end up being drawn on a single A4 page, there is little point in recording things to the nearest millimetre. The relationship between the scale of your drawing and the smallest measurement you can literally draw is different from the standards for accurate measurement in Table 4.1. Even though a plan drawn at 1:100 scale should be *Archaeologists Field Handbook 29/6/04 5:24 PM Page 113

SITE SURVEYINGW113 TABLE 4.1: Acceptable levels of accuracy for site plans Scale of final plan Acceptable error 1:5 2 mm 1:10 5 mm 1:25 10 mm 1:50 20 mm 1:100 50 mm 1:250 100 mm 1:500 200 mm 1:1000 500 mm

measured to the nearest 50 millimetres, anything which is this small will only end up being 0.5 millimetres on your final plan. Given the difficulty of drawing something which is half a millimetre long, the standards for plotting are somewhat broader (see ‘Drawing horizontal surfaces (plans)’ on page 286 and Table 9.2 on page 287). If you draw the plan by hand, it is important to remember that you should never ink in original pencil drawings, as these form an essential element of the primary site archive (Drewett 1999: 177). Instead, you should use drafting film (not tracing paper) to redraw the plan (for more information see Chapter 9: Photography and illustration), or scan it and redraw it using a computer graphics program.

REFERENCES AND FURTHER READING

Casey, D.A. 1972, ‘Elementary surveying for Australian archaeologists’, in D.J. Mulvaney, (ed.), Australian Archaeology: A Guide to Field and Laboratory Techniques, AIAS Press, Canberra, pp. 5–21. Davies, Martin and Buckley, Krystal 1987, Archaeological Procedures Manual: Port Arthur Conser- vation and Development Project, Occasional Paper No. 13, Department of Lands, Parks and Wildlife, Hobart. Drewett, Peter 1999, Field Archaeology: An Introduction, UCL Press, London. Hobbs, D.R. 1983, ‘Surveying techniques useful in archaeology’, in G. Connah (ed.), Australian Field Archaeology: A Guide to Techniques, AIAS Press, Canberra, pp. 43–63.

USEFUL WEBSITES

The Land Surveyor Reference Page, www.lsrp.com, contains a range of online resources, including links to professional publications on the topic.

*Archaeologists Field Handbook 29/6/04 5:24 PM Page 115

CHAPTER FIVE BASIC EXCAVATION

TECHNIQUES

W W

W W W

WHAT YOU WILL LEARN FROM THIS CHAPTER

W How archaeologists excavate W Why careful control of the excavation process is important W How to dig W Standards for describing soil W How to use a Harris matrix to interpret stratification W Basic procedures for sieving, recording, bagging and labelling W Basic procedures for collecting on-site samples W Basic conservation measures for protecting excavated finds W Tips for surviving an excavation

The basic excavation toolkit 10–12 centimetre drop forged pointing trowel. Make sure that the neck and blade are cast in one piece, otherwise it will break Hand-shovel or dustpan Plastic buckets Secateurs Dental picks/plasterer’s tools

115 *Archaeologists Field Handbook 29/6/04 5:24 PM Page 116

W116 THE ARCHAEOLOGIST’S FIELD HANDBOOK Fine, soft paintbrushes (in a range of sizes for cleaning large areas and for reaching into small cavities) Hand-brush or whisk, for cleaning hard surfaces and brushing-up soil Spring balance (for weighing buckets) Carpenter’s string level (for making sure baselines are horizontal) Nested mesh sieves (in a range of sieve sizes—10, 5 and 2 millimetres are the most common) Plastic sheeting, to cover the site or lay out deposits Knee pads (or foam squares) to make excavation more comfortable Soil pH test kit Munsell soil colour chart Recording forms Clipboard Artefact tags (aluminium or polyethylene) Nails (for securing tags to the walls of trenches if necessary) String and line level (for marking the edges of trenches and for establishing base lines for drawing sections) Spikes or tent pegs or chaining arrows, for marking out the trenches Mattock, pick or shovel (for removing turf or for backfilling) Drawing equipment (see Chapter 9) Photographic equipment (see Chapter 9) Tape measures, in a range of sizes Ziplock® plastic bags, in various sizes

Optional Wheelbarrows A stepladder (you may need to stand on this to photograph trenches from above) A hand sprayer (you may want to wet the walls of the trenches to observe differences in soil colour) Artefact processing equipment (plastic basins, drying trays, toothbrushes) General site equipment (chairs, folding tables, canvas or shade cover)

THE BASICS

The aim of any archaeological excavation is to try to understand what may have happened at a site in the past by carefully excavating the various material remains which make up that site. Excavating requires both care and patience since, to be able to understand the sequence of activities at a site, archaeologists must slowly strip away each soil layer in *Archaeologists Field Handbook 29/6/04 5:24 PM Page 117

BASIC EXCAVATION TECHNIQUESW117 succession. The basic principle on most excavations is that each new soil layer is removed completely before proceeding on to the next—in other words, excavation proceeds horizontally first (i.e. by removing all traces of one soil layer first), and vertically second (when excavation of the next soil layer is begun). This kind of careful ‘layer-cake’ excavation is called the stratigraphic or context system of excavation. Although not all digs will adopt this process, it is the most common system followed by archaeologists (for more information, see ‘The principles of excavation’ on page 121). A context is any discrete archaeological entity on a site, such as a post-hole, a depositional layer, a rubbish pit or an erosion event (Drewett 1999: 107), and is the same thing as a ‘layer’, a ‘feature’ or a ‘stratigraphic unit’. Under the context system, the goal of any excavation is not to dig as deeply as possible as quickly as possible, but rather to be sure that all of the information from each soil layer, or stratum, is kept together so that it may be analysed later as a discrete entity. If you think of a site which has had many successive activities performed on it over hundreds of years, all of these activities may have left separate and distinct evidence behind them, trapped in different layers of the soil. If archaeologists were to simply mix up all of these layers, they would be unable to separate the information from each activity, unable to date the sequence of activities and thus unable to reconstruct exactly what had happened there in the past. One of the most important things to realise is that, while all excavation aims to retrieve as much information as possible, in the process it literally destroys the site forever. No excavated site can ever be put back: once excavated, it is gone forever and only exists in your recording forms, field notes, reports, photographs, publications and archives. For this reason, you must ensure that your excavations are always conducted according to the highest possible standards. As soon as you begin an excavation, it is your ethical responsibility to ensure that it is done professionally, up to and including the standards you adopt for analysis, reporting, archiving, conservation and curation. In the end, it is better to go for less done well, rather than more done badly: ‘At the end of the day a well excavated, well recorded, fully published 1 x 1 m unit does considerably less harm than a huge, badly controlled, unpublished ...excavation.’ (Drewett 1999: 97)

Almost everyone recognises an archaeologist as someone who excavates. All excavation or any other form of collecting artefacts, however, is inherently destructive and can never be repeated. It is for this reason that an excavation permit must be obtained before excavation or collection can commence. Excavation should be a last resort—if you can get all the answers you want through non-destructive recording techniques, then DON’T excavate. It is a primary ethical responsibility of all archaeologists that excavation is only undertaken by professionals (or under professional supervision), and never without proper *Archaeologists Field Handbook 29/6/04 5:24 PM Page 118

W118 THE ARCHAEOLOGIST’S FIELD HANDBOOK research and planning. It is for this reason that excavation, or the collection of artifacts, is the one aspect of archaeology which is regulated in all states and territories by legislation governing who may undertake such activities and how they should be conducted. Removing artefacts from their location in a site or the landscape without the proper legal permissions is, in most cases, both unnecessary and harmful. While there may be some instances where it is unavoidable (such as when the place is in imminent danger of being destroyed, or when Indigenous custodians insist on collecting items from sites in their country), in general it is both illegal and unethical to remove artefacts. If you find a particularly interesting or unusual artefact, by all means sketch it or photograph it (or both), but leave it where you found it.

Professional recording standards require that an archaeologist be able to control the excavation process as much as possible. Not all archaeological evidence is in the form of three-dimensional artefacts. Much evidence for past human behaviour will be subtle and perhaps not even instantly recognisable (such as stained soil from decomposed timber posts indicating the original location of buildings, or slightly darker layers of soil indicating the presence of charcoal from cooking fires). This evidence will be lost if care is not exercised during the excavation process, particularly with regard to changes in soil colour, type or texture as the excavation progresses (see ‘The principles of excavation’ on page 121). Any change in the physical characteristics of the soil could be of great importance, so you should never under-estimate what’s happening under your trowel (see ‘Using a trowel and brush’ on page 141). The main aspect of this control is to enable the archaeologist to pinpoint where each and every piece of evidence comes from. Archaeologists must be sure of the exact location of each artefact within the site (see ‘Recording in three dimensions’ on page 130). The positioning of artefacts horizontally across the site can tell archaeologists how different parts of the site were used and what activities were performed there; the positioning of artefacts vertically through the site can tell archaeologists what happened when and the order of events in the past. This need to be absolutely sure of the provenance of all artefacts is what drives archaeologists to keep all the evidence from a particular context together throughout the digging, sieving, sorting, bagging, cleaning and analysis process (see Figure 5.1 on page 120). It is also why archaeologists tend to excavate carefully selected portions of a site in such a way that the boundaries of each trench are very specific. Usually they do this by adopting a standard grid system and stringing out their excavation trenches to exact dimensions (See ‘Laying out a site grid’ on page 125). Trenches are usually laid out in multiples of 1 metre, and sometimes divided into smaller squares of 25 or 50 centimetres within each metre square; these are called quadrats. *Archaeologists Field Handbook 29/6/04 5:24 PM Page 119

BASIC EXCAVATION TECHNIQUESW119 A consistent and logical numbering system is also imperative for each trench and for each context or layer within each trench (see ‘Laying out a site grid’ on page 125), so that all the contents of a particular context can be recorded and labelled consistently. This numbering system is like a library catalogue: it is what ties the archaeological features and structures, the artefacts, their locations and the physical descriptions of the various soil strata together so that, in the end, the site may be analysed in a meaningful way (see ‘Labelling trenches’ on page 127 and ‘Labelling finds and samples’ on page 150). You should give considerable thought to the numbering/labelling system of your site before you begin excavating to ensure that numbers are not repeated and that all trenches and contexts are numbered consistently and logically. Public health and safety on an excavation are also your responsibility. If you’re excavating in a public place, you may have to erect mesh fencing or some other form of barricade to protect the public from any accidents and, for the same reason, backfilling the site after the completion of excavation is also your job. The location of your excavation trenches and their ultimate depth may also be a health and safety issue. In New South Wales, for example, occupational health and safety regulations require that any excavation over 1 metre in depth must be shored, and there may also be regulations for how closely you can excavate to a standing structure (in South Australia, this is no closer than 2 metres). You will need to refer to the occupational health and safety legislation in your state when planning archaeological fieldwork. To avoid the dangers of deep, narrow trenches, you can cut the sides back in a series of steps, although this may create problems for recording. If you are going to be excavating on a deep or difficult site, you may need to consult with a civil engineer. You are also responsible for backfilling the trenches and returning the area to some semblance of its former appearance once the excavation is completed (i.e. ensuring that no open holes are left behind and removing your rubbish). Before you backfill, it is common practice to place plastic or weed mat on the bottom of the trench so that future archaeologists will know the mixed material put back into the trenches is the result of a past archaeological excavation, and not some bizarre ritual practice of past people! Plastic or weed mat will also indicate clearly where your excavation ended, in case you or another archaeologist should decide to re-open your trenches. Archaeologists sometimes include readily date- able items (such as coins with the current year) in the base of the trench before they backfill as a hint to future archaeologists. In essence, all excavation requires careful planning (see ‘Designing your research’ on page 3). Think about what aspects are important to the success of your project and make sure you seek out and record those aspects during excavation. One of the key things to realise is the importance of maintaining flexibility in your methods. Every site is different, so you will need to carefully weigh up the pros and cons and tailor the excavation methods to suit your particular research questions, the time available and the individual nature of the site. *Archaeologists Field Handbook 29/6/04 5:24 PM Page 120

W120 THE ARCHAEOLOGIST’S FIELD HANDBOOK Once you have carefully thought through your research aims and worked out how excavation can best serve them, all excavation follows a fairly standard process:

• Lay out trenches. • Photograph entire site before commencing (see ‘Photographing excavations’ on page 279). • Excavate (this is in reality a complex series of steps in its own right—see Figure 5.1). • Photograph entire site upon completion. • Backfill. • Analysis. • Reporting. • Conservation/curation of recovered artefacts. FIGURE 5.1: A flowchart to follow when recording an excavation unit, or context. This sequence should be repeated for all units *Archaeologists Field Handbook 29/6/04 5:24 PM Page 121

BASIC EXCAVATION TECHNIQUESW121 THE PRINCIPLES OF EXCAVATION

Sites can be created over long or short periods of time—they can even be the remains of a single event—but excavation assumes that the order in which the different parts of the site have been laid down will reflect the sequence of events which occurred at that site in the past. This is known as the Principle of Superposition, and it simply assumes that more recent deposits will be laid down on top of older ones. While this is not always straightforward, it is the fundamental basis for using stratification, or the way in which the structure of the soil is divided into different layers or deposits, to interpret what happened at a site. You should note the distinction between stratification (the process of sedimentary layering and its observed result) and stratigraphy (the archaeologist’s interpretation of the stratified layers, in words or drawings). There are often complicating factors to this, of course, such as when various natural processes deposit or remove material from a site (such as wind depositing silt or water eroding the site, or rodent burrows churning and mixing up the soil), or when later events remove or alter evidence of previous events. There are three other fundamental principles which describe these possibilities:

• The Principle of Association, which presumes that items found together in the same deposit are of essentially the same age. This must be applied with caution, however, as some items may be looked after for a long time before they are finally thrown away (such as a treasured tool or family heirloom), making them much older than the other materials associated with them. • The Principle of Reversal, which allows for those rare cases when deposits have been removed from the site and re-deposited in reverse order. This usually takes place as a result of major construction activity or digging, when large quantities of earth are removed and then redeposited upside down. • The Principle of Intrusion, which states that an intrusion must be more recent than the deposits through which it cuts (Barber 1994: 85). A rubbish-pit dug into the ground or an underground oven are both intrusions cutting into the older deposits around them.

All archaeological excavation is based around these three simple principles. Together they imply that, through the careful removal of the layers which make up a site, and a detailed description of their texture, colour and contents, an archaeologist can reconstruct the sequence of events (both human and natural) which took place at that site in the past. Put simply, by analysing the stratification of a site, it is possible to work out the whole story of human and natural processes that the site has to tell. *Archaeologists Field Handbook 29/6/04 5:24 PM Page 122

W122 THE ARCHAEOLOGIST’S FIELD HANDBOOK Because stratification is so important, most excavation proceeds by carefully stripping each stratigraphic layer from a site in turn. This is why, when you are excavating, it is important to describe and take note of any changes in soil colour, texture or appearance as you dig. It is also why you should not hurry. Stratigraphy is closely linked to the process of excavation itself in that, as you remove each layer, you form opinions about how the different strata in a site were laid down and how they relate to each other:

The archaeologist’s golden rule is to excavate one layer at a time—and nothing in that layer should escape his or her detection. It isn’t possible to read significance into a layer or level until you know how it lies, how it was formed, what its composition is, and what its relationship is to the layers above and below it. (Joukowsky 1980: 171–72)

For this reason, excavation should proceed horizontally first and vertically second. In other words, you should finish excavating one unit or layer completely before you begin the next. A site which has distinct layers or contexts is said to be stratified. Not all sites are like this, however, and therefore not all excavation proceeds according to the stratigraphic system. Some sites have no visible stratification, while at others the more significant cultural layers might be covered by large quantities of later, less significant, debris, or the entire site may have resulted from a single event of a known date (such as a load of rubbish being dumped). In such cases, excavating according to stratigraphic levels would be fruitless. Even if you choose not to excavate the site according to stratigraphic layers, you still need to be able to control the removal of soil from the site if you want to draw any meaningful conclusions about the vertical or horizontal location of artefacts. In this case, a site is often dug in arbitrary levels which can still provide sufficient vertical and horizontal control. These levels are sometimes referred to as spits, or units, and can be of any thickness depending on the overall depth of the site and the degree of resolution you want to achieve in locating the artefacts. Keep in mind that whatever arbitrary depth you choose as the standard for your levels, this will be the finest degree of resolution that you will be able to achieve in your analysis and results. If several different activities at a site have only left behind them 2 centimetre layers or deposits, which you choose to excavate in 5 centimetre spits, then any one of these spits is likely to contain information from at least two different activities. Because you have chosen to group them together in one arbitrary spit, you will not be able to separate these in your excavation or in your subsequent analysis or conclusions. Once again, flexibility is the key: some sites may require a combination of arbitrary and stratigraphic excavation, such as if you find thick homogenous layers within an otherwise stratified site (generally at the beginning or end of the excavation) which can be removed by arbitrary levels. *Archaeologists Field Handbook 29/6/04 5:24 PM Page 123

BASIC EXCAVATION TECHNIQUESW123 APPROACHES TO EXCAVATION

Once you have decided on the most appropriate method of excavating your site, the next important decision to make is which parts of the site you are going to excavate. This is essentially a sampling decision, similar in principle to deciding how you are going to selectively survey an area (see ‘Developing a suitable sampling strategy’ on page 66). As with a surface survey, if you can’t dig the entire site, you will have to make some meaningful decisions about which part or parts of the site are most likely to give you the information necessary to answer your research questions. There is an ethical element to this decision as well, of course, in that some archaeologists argue you should always leave part of a site intact so that future generations of archaeologists, who may bring with them new and better methods, will still be able to retrieve some in situ information. There are two aspects to this decision:

• deciding how much of the area to excavate; • deciding where to place your excavation trenches.

How much? There are two ways to approach the excavation of any site, which depend on how much of the site you wish to excavate: the trench system or the open-area system. The trench system is concerned with obtaining a cross-section through the site and tends to excavate relatively narrow portions of the site to sufficient depth (it is often referred to dis- paragingly as the ‘telephone box’ approach). Because it aims to dig deeply, it is well suited to answering chronological questions, such as the sequence of dates or the dates for earliest occupation, and for indicating the richness of the deposits and revealing the stratification of the site, since the walls of the trench preserve the stratigraphic profile of the excavation unit until the very end of the excavation. Because the trench system excavates narrow vertical slices through a site, however, it cannot expose spatial information across the site. The open-area system was developed in response to this need for horizontal information and exposes large expanses of the site, often to only a relatively shallow depth. It is thus quite successful at revealing information about activity areas or site structure, although it may not be so successful for establishing a sequence of dates (although it may be, if you dig deeply enough). Each system has different pros and cons, and it is important to realise that each is designed to recover different types of information. They can, of course, be used in tandem to complement each other, or be combined with other methods such as mechanical excavation or surface stripping to answer many different kinds of question. The choice of which to employ will depend on the time and resources available to you and the particular set of research questions you are asking. *Archaeologists Field Handbook 29/6/04 5:24 PM Page 124

W124 THE ARCHAEOLOGIST’S FIELD HANDBOOK Where? Intra-site sampling is the choice of where to excavate within a site. Remembering that vertical trenches are designed to recover chronological information and horizontal, open-area excavations are designed to recover information about spatial activity areas, obviously the selection of excavation areas and the methods of excavation you adopt will influence the kind of data you can collect from a site. If you are investigating a rockshelter which has been inhabited by people for several thousands of years, for example, you might want to know about the chronology of the site (i.e. the sequence of its occupation over time), or about the use of space within the site (i.e. were people using different parts of the site for sleeping, eating, cooking, etc.?). Excavating the site through a system of vertical trenches is usually the best means of obtaining enough stratigraphic information to determine chronology, but it is unable to tell you anything meaningful about how space was used across the rockshelter (because you haven’t excavated enough of it). If you think of the rockshelter as analogous to your house, how representative would a shallow vertical ‘slice’ through your house be for reconstructing all of your day-to-day activities? Similarly, a shallower, open-area excavation will be better suited to recovering just this kind of spatial information, but may not necessarily give you enough depth to obtain an adequate chronology. In reality, of course, obtaining information about chronology and the sequence of occupation is just as important as finding out about the patterning of activities across space, so in deciding where to place your excavations and how much of the area you are going to excavate you will need to think carefully about what it is you want to know and how best you will be able to find this out:

• What kind of data do you want to recover? • What sampling approach best fits this goal? • What sample size will best answer your questions (i.e. how much information will you need)?

In reality, you will only be able to decide where to excavate after you have completed an intensive surface-survey of the area in question. Careful surface-survey may reveal discrete activity areas or clusters of certain elements or features, or the patterning of artefacts across the ground surface may be able to tell you about differential erosional/ depositional patterns across the site. Any indication that there might be more archaeological evidence below the ground surface will be a useful guide as to where to place your trenches. You may also wish to excavate a seemingly bare area of ground to test whether what you observe on the ground surface is really a good indication of what exists below the surface. Examining the ground carefully before you dig can also tell you what kind of *Archaeologists Field Handbook 29/6/04 5:24 PM Page 125

BASIC EXCAVATION TECHNIQUESW125 tools you will need to do the job properly (Joukowsky 1980: 172). Researching the recent history of your site may also give you some idea of what took place in the recent past at least, and therefore the kind of deposits you might encounter.

Sub-surface sampling In making these decisions, the archaeologist is effectively trying to assess what might lie below the ground surface. There are more scientific ways of doing this than simply guessing, of course, all of which can be employed to help you decide on where it would be best to dig. This is called sub-surface sampling. Sometimes small-scale subsurface sampling will be necessary to determine the horizontal extent of a site. Soil cores, auger holes or shovel-test pits can be dug at intervals across the site to obtain broad stratigraphic or spatial information. Soil cores are commonly 3–5 centimetres in diameter, with auger holes slightly larger at 10–15 centimetres in diameter. Shovel-test pits are usually slightly larger than the width of a shovel blade (i.e. between 25 and 50 square centimetres) and have the distinct advantage of allowing you to examine the stratigraphy in the sides of the pit, something which is not possible for auger holes or cores. Com- parison between different methods of subsurface sampling suggests that shovel-test pits are the most effective means for ‘seeing’ beneath the ground surface, although they are quite labour-intensive (Hester, Schafer and Feder 1997: 57–59). As with any form of sampling, the best results can only be obtained after careful forethought—you will not only have to consider where you are going to place your test pits, but also how close together they should be. Bear in mind, however, that any form of subsurface sampling will destroy a small portion of the site as it is dug, so it is always wise to limit the effects of this, particularly if you are going to conduct a larger-scale excavation later. Because all forms of subsurface testing will destroy a portion of the archaeological deposit, they must also be conducted under the auspices of an excavation permit. Other forms of sub-surface testing, such as magnetometry or the use of ground- penetrating radar, have the advantage of being non-invasive, but can only be conducted by specialist operators with appropriate equipment.

LAYING OUT A SITE GRID

Once you have made your decision about where to excavate, the next step is to lay out your excavation trenches. This may seem pedantic, but is important in terms of being able to maintain control over the positioning of artefacts and features across the site. Archaeologists often use a grid which, for convenience, divides the site into 1 metre grid units (often called ‘squares’). An excavation trench can be any multiple or fraction of *Archaeologists Field Handbook 29/6/04 5:24 PM Page 126

W126 THE ARCHAEOLOGIST’S FIELD HANDBOOK these grid units, which are related to each other through a universal numbering system (see Figure 5.2). The two axes of the grid are labelled the ‘X’ and the ‘Y’ axis. Note that it is standard practice to begin numbering coordinates along these axes at a relatively high number, such as ten, to allow for the possibility of extending a trench if necessary but to avoid going into negative numbers when doing so. You don’t actually need to string out every grid unit across the site; however, to avoid having to measure them again, you should mark the ends of the X and Y axes with stable and relatively permanent pegs. If you want to excavate a trench that is larger than 1 square metre, but don’t wish to string out the boundary of every 1 metre grid unit within it, it is simplest to mark the corners of each unit along the edges of the trench with masking tape attached to the trench string. You do have to formally string out each excavation trench within your site, however. The string outline of the trench serves as a constant reference point for measuring depth and for ensuring that the sides of the trench are kept straight. You must ensure that all of your excavation trenches are kept square (meaning horizontally rectilinear and vertically straight-walled), because it is this which allows you to control the excavation process. As a first step, establish where the edge of your trench will be, measure its length and then mark the two corner points with pegs hammered lightly but firmly into the ground.

FIGURE 5.2: A hypothetical grid across an excavated site. Any of the squares in between the already excavated trenches can be opened and assigned X and Y coordinates from the established grid. *Archaeologists Field Handbook 29/6/04 5:24 PM Page 127

BASIC EXCAVATION TECHNIQUESW127 Because trenches are usually laid out in multiples of 1 metre with 90° corners, the next step is to mark the opposite corners. Establish a right angle from one established peg to a third corner peg using triangulation (see ‘Using the baseline and offset technique’ on page 96). Install a fourth corner peg by the same method and then measure all four sides and the diagonals to be sure that the trench is square. Table 5.1 provides diagonal measurements for excavation trenches up to 15 metres square, on the principal that the diagonal of a square will always measure 1.414 times the length of the sides. This is related to Pythagoras’ theorem, which states that in a right-angled triangle, the square of the hypotenuse is equal to the sum of the squares of the opposite sides. In practice, even though you have carefully measured the first side of the trench, and then triangulated the opposite two corners from this, a trench will rarely be perfectly rectilinear on the first go. Inevitably you will have to keep adjusting the positions of your pegs until the trench is actually square. The only way to do this is to keep measuring the sides and the diagonals until you have them right (see Figure 5.3 on page 129). Once you have strung out your trench, make sure you don’t trip over the string or pegs during excavation. For this reason, you may wish to tie flagging tape or some other brightly coloured material to the pegs to make them easy to see (and avoid). Remember that the corners will be your permanent reference points for excavating and measuring, so it is very important not to disturb them.

LABELLING TRENCHES

All excavation trenches within the grid of your site must be numbered logically so that all finds and descriptions of deposits can be tied securely to their place of origin. There must be no confusion about this, so deciding on a logical labelling system is a first priority. Every artefact or sample which comes out of a trench, and every description of a layer or a feature, must be able to be described so that you know precisely where it came from, otherwise it will be impossible to draw any meaningful conclusions about your site. Even though you have X and Y coordinates for every grid unit, these would be very time- consuming to transcribe on to every bag or recording form, so it is best to simplify the system by assigning a number to each trench. At the very least, a labelling system should contain a site prefix or code for the site (such as two or three letters which denote its name), a unique number or letter for each trench, and a consecutive and unique number for each context or layer. Individual artefacts or samples within each context can also be numbered if required. This string of letters and numbers will be the code that is written on every bag containing an artefact or sample (see ‘Labelling finds and samples’ on page 150), and on every recording form. *Archaeologists FieldHandbook29/6/045:24PMPage128

TABLE.1: Diagonal measurements for archaeological trenches Length (m) 1 2 3 4 5 6 7 8 9 101112131415 1 1.414 2.236 3.162 4.123 5.099 6.083 7.071 8.062 9.055 10.050 11.045 12.042 13.038 14.036 15.033 2 2.828 3.606 4.472 5.385 6.325 7.280 8.246 9.220 10.198 11.180 12.166 13.153 14.142 15.133 3 4.243 5.000 5.831 6.708 7.616 8.544 9.487 10.440 11.402 12.369 13.342 14.318 15.297 4 5.657 6.403 7.211 8.062 8.944 9.849 10.770 11.705 12.649 13.601 14.560 15.524 5 7.071 7.810 8.602 9.434 10.269 11.180 12.083 13.000 13.928 14.866 15.811 Width (m) 6 8.485 9.219 10.000 10.817 11.622 12.530 13.416 14.318 15.232 16.155 7 9.899 10.630 11.402 12.207 13.038 13.892 14.765 15.652 16.553 8 11.314 12.042 12.806 13.601 14.422 15.264 16.125 17.000 9 12.728 13.454 14.213 15.000 15.811 16.643 17.493 10 14.142 14.866 15.620 16.401 17.205 18.028 11 15.556 16.279 17.029 17.804 18.601 12 16.970 17.692 18.439 19.209 13 18.385 19.105 19.849 14 19.799 20.518 15 21.213 *Archaeologists Field Handbook 29/6/04 5:24 PM Page 129

BASIC EXCAVATION TECHNIQUESW129 FIGURE 5.3: The sequence for laying out an excavation square *Archaeologists Field Handbook 29/6/04 5:24 PM Page 130

W130 THE ARCHAEOLOGIST’S FIELD HANDBOOK Stephen Sutton’s tip for stringing out an excavation trench Even after you’ve strung out your trench, it’s possible for the string to move (or, more likely, for some unfortunate person to trip over it or kick the pegs out of alignment). To make it easier to reconstruct the precise location of your original corners, mark the place where the strings cross with a permanent, easily visible pen line. In this way, if the string is ever moved, you’ll be able to see its original position easily and replace it accordingly.

RECORDING IN THREE DIMENSIONS

Because archaeologists want to know the exact location of artefacts and features, both horizontally and vertically, they record these things according to their position both along the length and the width of the trench (their position on the X and Y axes of the site), but also their position through the depth of the deposit. This is called the Z axis (see Figure 5.4 on page 131). The most accurate way to measure depth is to use a dumpy level or EDM to plot individual finds, although this technology is not always easy to come by. If you don’t have access to it, you will have to use an old-fashioned tape measure. If you use this low-tech solution, make sure that you always measure from a fixed horizontal baseline (the string of the trench is fine provided you use a carpenter’s string level to make sure it really is horizontal and hasn’t been disturbed during excavation). Any alteration or movement in this baseline will affect all your subsequent measurements. On many sites, recording the precise position of every individual artefact in three dimensions is not always practical and is not necessarily going to provide you with more accurate or ‘better’ archaeological information, unless there is good evidence that the artefacts are still in their area of primary use or discard (Drewett 1999: 143). Recent research (Balme and Beck 2002) suggests that simply plotting artefacts to within a 25 or 50 centimetre quadrat (i.e. this artefact was recovered from Trench 1/context 10/quadrat a) will provide sufficient spatial resolution to answer most questions about *Archaeologists Field Handbook 29/6/04 5:24 PM Page 131

BASIC EXCAVATION TECHNIQUESW131 FIGURE 5.4: Using X, Y and Z coordinates to plot the location of an excavated artefact

activity areas at a site. Any special or unique finds can still be plotted individually if necessary. Whether you are more concerned with chronological or spatial questions will largely be determined by the type of site you are excavating. Historical sites, for example, are usually formed over a relatively short period of time, and may be well documented; therefore, detailed information about chronological change is not always important. The emphasis in these sites may be more on spatial questions—changes in patterning horizontally, across space—rather than chronological issues.

RECORDING THE EXCAVATION PROCESS

Just as with a surface survey, there are several complementary aspects to an excavation which must be recorded consistently and in detail as the excavation progresses. These will be essential pieces of information in the final jigsaw puzzle which will be your archaeological analysis. The core aspects to record throughout any excavation are:

• the soil or deposits (for each context or spit); • any features encountered; • the process of excavation itself. *Archaeologists Field Handbook 29/6/04 5:24 PM Page 132

W132 THE ARCHAEOLOGIST’S FIELD HANDBOOK DESCRIBING DEPOSITS

The four main elements to record when describing soils are:

• colour; • texture; • consistency; • coarse components or composition.

It is important that you standardise such descriptive information as much as possible, so that other people reading your report will know precisely what you mean. If you let everyone describe a variable such as colour in their own way, for example, you will end up with as many descriptions as there were workers. In reality, of course, none of these attributes can be assessed in the field with any kind of scientific accuracy, and the standards employed by soil scientists and archaeologists vary considerably. Your goal should simply be to try to record as much descriptive information in as standard a fashion as possible. Colour should be recorded using a Munsell Soil Colour Chart, which provides an internationally recognised standard against which to assess soil colour and makes the process far more objective than would otherwise be possible. The basic divisions of soil texture are sand, silt and clay, defined in terms of the size of their mineral particles (sand = 0.06–2 millimetres; silt = 0.002–0.06 millimetres; clay = less than 0.002 millimetres). For on-site purposes, however, it is best to use a more general measure which can be estimated by hand-texturing, such as whether or not the soil will hold its shape when damp. The overall rule of thumb is that clay coheres, silt adheres and sand does neither. Thus clay will be sticky and plastic, silt will have particles that are invis- ible to the naked eye, and sand will have a visible gritty feel when moistened (Roskams 2001: 178). A general test which you can apply to all types of soil is to roll it into a ball and test its malleability (Figure 5.5 on page 133). If the soil can be rolled into a sausage shape that still holds its shape when bent into a ring, then it is largely clay; if it breaks when bent into a ring, it is largely silt. Sand, of course, cannot be rolled into any shape. Consistency measures the degree of compaction of the soil and whether or not it holds together. Variations in compaction across a deposit can be important, as different activities on the site will have affected the consistency of the soil in different ways. To assess consistency, take a slightly moist cube of soil and try to crush it between your thumb and forefinger:

• If it cannot be moulded into a cube at all, it is a loose soil. • If it crushes easily (if there is no resistance), it is a weak soil. • If low pressure is required to crush it, it is a friable soil. *Archaeologists Field Handbook 29/6/04 5:24 PM Page 133

BASIC EXCAVATION TECHNIQUESW133 FIGURE 5.5: Sediment composition flow chart (after Museum of London 1990)

• If greater pressure is required to crush it, it is a firm soil. • If it cannot be crushed at all, it is a compact, or hard soil. • If it is bound together with a substance other than clay, it is a cemented soil. (Roskams 2001: 180)

The assessment of the coarse component is an estimate of the size of visible particles within the soil and of the proportion of different grain sizes in the deposit. When assessing the size of the visible particles, use the general guide in Table 5.2 on page 134 (Museum of London 1990). When assessing the proportion of different grain sizes (composition), you can estimate the percentage of inclusions and their approximate grade size using the chart in Figure 5.6 on page 134. If possible, you should also estimate the degree of sorting that is visible in the deposit. This is an assessment of the frequency with which particles of the same size occur and will give you some idea of how the deposit was laid down. A deposit in which all of the *Archaeologists Field Handbook 29/6/04 5:24 PM Page 134

W134 THE ARCHAEOLOGIST’S FIELD HANDBOOK TABLE 5.2: A guide to estimating the size of visible particles Description Particle size Clay/silt Not visible to the naked eye Fine sand 0.02 mm–0.06 mm Medium sand 0.06 mm–0.20 mm Coarse sand 0.20 mm–2.00 mm Fine pebbles 2 mm–6 mm Medium pebbles 6 mm–20 mm Coarse pebbles 20 mm–60 mm Cobbles 60 mm–100 mm

FIGURE 5.6: Estimating the percentage of inclusions (after Museum of London 1990)

particles are of very similar size, for example, indicates that it has been well sorted prior to deposition (by water or wind, for instance).

DESCRIBING CULTURAL FEATURES

Obviously you also have to record any cultural features (i.e. the artefacts) you encounter during excavation. For example, the presence or absence of stone artefacts or charcoal within a rock shelter deposit is clearly an essential component. In historical archaeology, when you describe other features in the excavation (such as walls, timbers, or various materials), it is important to outline their methods of construction using standard terminology so that others will understand precisely what you mean. Detailed standards *Archaeologists Field Handbook 29/6/04 5:24 PM Page 135

BASIC EXCAVATION TECHNIQUESW135 FIGURE 5.7: Estimating the degree of sorting (after Museum of London 1990)

for describing timber, brick and other remains are contained in ‘Describing structural components’ on page 178.

DESCRIBING THE EXCAVATION PROCESS

An excavation is described essentially by the information you record in your field notes, detailing the day-to-day running of the dig and any problems or successes (see ‘Keeping a field journal’ on page 70). Your details of the sequence of events for the day and the decision-making process can be invaluable for sorting out problems in later analysis, or for clearing up any unwitting mistakes made on recording forms or artefact bags.

RECORDING SECTIONS

One of other main ways in which archaeologists record stratigraphic information from a site is by drawing and photographing the vertical walls of their trenches. These walls are called sections or profiles, and the aim is to represent both the visible soil layers and the discernible archaeological features as informatively as possible. Stratification is *Archaeologists Field Handbook 29/6/04 5:24 PM Page 136

W136 THE ARCHAEOLOGIST’S FIELD HANDBOOK what allows the archaeologist to place events at a site into chronological order. Because the Principle of Superposition rests on the assumption that the deposits at a site are laid down in sequence over time, it is also the basis for relative dating (see ‘The principles of excavation’ on page 121). This is the process of putting things into order from earliest to latest, but without assigning any specific dates to the things themselves. Absolute dating, on the other hand, uses a particular technique or process to assign a specific date to something (such as when radiocarbon dating is used to date a piece of bone). Both forms of dating are essential to archaeological research for obvious reasons. Without some knowledge of when things happened, it is impossible to properly interpret a site. For this reason, it is important that your photographs and section drawings represent the vertical sequence of layers or contexts visible in the wall of the trench as accurately as possible (see ‘Photographing excavations’ on page 279 and ‘Drawing vertical surface (sections)’ on page 289). The principle of drawing a section is exactly the same as that for drawing a site plan: using baselines with offset measurements to plot features (see ‘Using the baseline and offset technique’ in Chapter 4 and ‘Drawing horizontal surfaces (plans) in Chapter 9 on page 286). Remember that the scale at which you draw the section will determine how much detail you can include in it. The hardest part of drawing a section is deciding where one layer or context ends and the next begins. Sometimes, lightly wetting the profile with water from a hand sprayer will bring out distinctions in soil colour, but another way around this problem is to use different symbols to identify a distinct boundary as opposed to an indistinct or uncertain one. You can draw a section either cumulatively—that is, as each context is excavated—or at the end once all excavation is complete. If you draw your section cumulatively, make sure that you include all the necessary coordinates to allow the individual drawings to be fitted back together later. If you draw your sections at the end, you will probably have to refer back to your excavation notes to know which contexts you are seeing and drawing in the profiles. A certain amount of annotation will be necessary to make your section drawings intelligible to someone else, but try not to reproduce all of the information from the context or excavation recording sheets. For more information on how to draw a section, see ‘Drawing vertical surfaces (sections)’ on page 289.

INTERPRETING STRATIGRAPHY

Current archaeological analysis does not rely solely on drawn sections for interpretation. A Harris matrix is probably the most commonly used means of making sense of archaeological stratigraphy, although there are debated alternatives. Harris matrices work best for the kind of deposits for which they were developed (i.e. complex historical archaeological *Archaeologists Field Handbook 29/6/04 5:24 PM Page 137

BASIC EXCAVATION TECHNIQUESW137 sites with structures). They are not necessarily very useful for pre-contact Indigenous sites, as these tend not to have the same kinds of deposits. A Harris matrix makes it possible to represent a complete three-dimensional stratigraphic sequence for a site in a single two-dimensional diagram. Its great virtue is that it enables all contexts excavated at a site to be shown simultaneously, not just those contexts which appear in the section (recent research has indicated that up to 40 per cent of recorded contexts will not show up in any section [Bibbey 1993: 108]). According to Harris, there are only three stratigraphic events which are possible on a site (Brown and Harris 1993: 10):

• deposits (the context or layer, which can be either natural or cultural, horizontal or vertical [i.e. a structure such as a wall]). This is the result of any event which acted to place evidence at a site, such as layers of debris or the construction of a wall; • interfaces between one layer and the next (i.e. the surface of the context); • cuts, such as pits, wells, graves, etc., which are dug through earlier layers and which can be defined as stratigraphic units in their own right. A ‘cut’ is essentially something which has happened on the site to remove evidence, rather than to deposit it. It is therefore a ‘negative’ feature, but it is important to record it in the same way as you would a ‘positive’ feature.

At the time of excavation, a decision must be made as to the nature of each context: is it a cut or a cultural or natural deposit? There is no room here to include all of the interpretive information which can be encoded into a Harris matrix, or to explore the various alternative schemas which have been developed in response to its perceived flaws. We have simply included the basic principles of the Harris matrix as one means of attempting to define relationships between different strata on an archaeological site. A Harris matrix is best produced as excavation progresses and each context must be added to the matrix at the time of its excavation. This is the only time that inconsistencies or unclear relationships between contexts can be sorted out through further excavation/investigation if necessary. Interpreting stratigraphy is rarely straightforward, and you should never assume that you will be able to remember stratigraphic relationships between units. If you are working on a large site you may only be excavating a small part of the total sequence anyway, so your observations of how contexts relate to each other will be vital to the overall understanding of the site. The essential basis of the matrix is very simple: a number, always written or drawn inside a consistent rectangular box, is assigned to each context or unit. Horizontal and vertical relationships between contexts are represented by horizontal or vertical lines drawn between boxes to represent the sequence. There are only three possible chronological relationships between any two contexts (see Figure 5.8 on page 138). *Archaeologists Field Handbook 29/6/04 5:24 PM Page 138

W138 THE ARCHAEOLOGIST’S FIELD HANDBOOK FIGURE 5.8: The stratigraphic relationships of a Harris matrix

• Relationship A: 1 is later in date than 2 (in other words, 1 overlies 2) 1

2 • Relationship B: 1 and 2 are identical in date

1 2

• Relationship C: 1 and 2 have no direct temporal relationship

1 2

The elegance of the Harris matrix is that it reduces all possible forms of stratigraphic connections to one of these three relationships and then uses these relationships to build a complete chronological sequence for the site. You can combine these three stratigraphic relationships in a variety of ways (see Figure 5.9 on page 139). Constructing a Harris matrix requires the excavator to think logically about the relationships between contexts and ensures that all contexts, regardless of size or supposed importance, are included in the final analysis (Bibbey 1993: 106–7). When establishing matrix sequences:

• First, look for correlations across the sequence (i.e. horizontal correlations). This means looking for deposits which are of the same date, or for deposits which may once have been part of a single continuous unit, but which have since been cut by later intrusions (Relationship B). This last is quite difficult to do and direct correlations between units must be inferred with care. It is for this reason that Harris matrices can only be produced as excavation progresses, as decisions about where each context fits within the overall sequence will often be based on similarities or dissimilarities between physical characteristics (colour, texture, inclusions, etc.), surface level, or the nature or date of recovered artefacts. The physical characteristics of the units in question will probably be the best guide as to whether or not two units are linked. This is one of the key reasons that it is imperative to record all descriptive information about excavation units using consistent terminology (see ‘Describing deposits’ on page 132). • Second, decide on the associations of successive units (i.e. vertical correlations). This means deciding on the sequence of deposition for the site—what is above or below each context (Relationship A) ? *Archaeologists Field Handbook 29/6/04 5:24 PM Page 139

BASIC EXCAVATION TECHNIQUESW139 FIGURE 5.9: Representing stratigraphic relationships in a Harris matrix (after Brown and Harris 1993) *Archaeologists Field Handbook 29/6/04 5:24 PM Page 140

W140 THE ARCHAEOLOGIST’S FIELD HANDBOOK Where a stratigraphic sequence exists, it makes sense to interpret it from the earliest elements upwards in the same order in which it has developed (Roskams 2001: 247). Constant use by archaeologists has amended the basic Harris matrix in many ways to depict a wide range of complementary information. By altering the form of the lines

FIGURE 5.10: The basic principles of the Harris matrix can be adapted to record a wide variety of complementary information (after Brown and Harris 1993) *Archaeologists Field Handbook 29/6/04 5:24 PM Page 141

BASIC EXCAVATION TECHNIQUESW141 between boxes, the shape of the boxes or their relative positions, a matrix can be made to depict the strengths of linkages between strata, different types of debris resulting from different activities, relative or absolute periods of time, and even construction sequences for standing structures (see ‘Recording standing structures’ on page 176).

USING A TROWEL AND BRUSH

You use a trowel both to define the extent of a deposit and to remove it to expose the underlying layers. The nature of the deposit (its texture and consistency) will determine the most appropriate trowelling techniques. The main decision you will have to make is whether to use the point or the edge of the trowel to remove the soil. This will really depend on the depth of the layer, how compacted the soil is and how large or fragile the artefacts contained within it are. If the deposit is loose or sandy, for example, then it is probably easier to scrape the soil away with the edge of the trowel; if it is hard and compacted, then the only option may be to try and break it up carefully with the point of the trowel. On the other hand, if you are excavating a site containing relatively large and fragile artefacts (such as mollusc shells), then removing the deposit in ‘chunks’ rather than scraping it away and risking slicing through the fragile organics may be your only option. If you are scraping, always use your trowel with the edge of the blade parallel to the ground and pull it towards you. You will probably need to keep your weight over the trowel to do this effectively, which may prevent you from sitting down as you dig. If you adopt this method you will have to work so that you pull the excavated material towards you on to the unexcavated portion in front of you, rather than pushing it away on to the already exposed surface. On some sites with fragile artefacts, you may be better off to loosen a small area of soil in a corner of the square closest to you with the trowel first, and then use the vertical face of this pit to open up the remainder of the grid unit. In this case, you can use the tip of the trowel both to loosen the dirt and at the same time to turn it over, and you will work away from yourself. For either technique, remember to always keep your dustpan and bucket handy so that you can continually remove the excavated soil from your unit as you dig. Once you have filled up your bucket, you will probably have to sieve the contents. As the sieves may be located some distance from the actual trenches, make sure that you do not make the buckets too heavy to carry (see ‘Sieving and sorting’ on page 146). Archaeologists use brushes to clean the soil from delicate or fragile finds and to tidy up the base of each soil layer after it has been excavated and before it is photographed (see ‘Photographing excavations’ on page 279). Small brushes and excavation tools (such as dental picks or plasterer’s tools) are best for excavating in small cavities. If you come *Archaeologists Field Handbook 29/6/04 5:24 PM Page 142

W142 THE ARCHAEOLOGIST’S FIELD HANDBOOK FIGURE 5.11: Use the edge of the trowel to pull sediment towards you and onto the dustpan. Any fragile or delicate artefacts will require careful brushing to avoid damage.

across a discolouration in the soil or other unusual feature, it may be best to brush this down with a hand-held brush until you can determine what it might be, rather than to keep trowelling and run the risk of damaging it. Soft, hair-bristle brushes (like those used for typical household cleaning) are best for sweeping up loose dirt. Whatever technique you adopt, you must always ensure that you work slowly and carefully, and take careful note of any apparent changes in the texture, colour or nature of the deposit. The goal of all stratigraphic excavation is to define as far as possible the limits of each stratigraphic layer as it is being removed and to remove each separately and in its entirety before beginning work on the next. As the excavator, it is your job to ensure that underlying strata are not cut into prematurely, thus destroying any opportunity for you or anyone else to observe vital relationships between layers. In essence, you are only removing the uppermost layer until something different appears. This is simple to say, but not always easy to do. Some differences between stratigraphic contexts will be very subtle (such as when one area of soil has a different texture or particle size to other areas around it); others will be readily apparent (such as a layer of fill which is a different colour to that underneath it); There is no easy answer here, and being able to distinguish between different layers will largely be a matter of experience. Sometimes you can literally ‘feel’ the difference as you dig when one area is harder or easier to dig than those around it. As a general rule, whenever you encounter any noticeable change in the soil (colour, texture, hardness, inclusions), stop and assess the situation carefully before you proceed. *Archaeologists Field Handbook 29/6/04 5:24 PM Page 143

BASIC EXCAVATION TECHNIQUESW143 If you find you’ve made a mistake (and cut too deeply for example, or removed a layer which you thought was uppermost but which turned out to underlie another adjacent layer) the best response is to stop and record the situation fully (with drawings and notes) before you proceed. The most important thing to remember when excavating is always to work systematically. The best advice is to go slowly and methodically, and always ask for advice if you are unsure. You should already be aware by now of the necessity for archaeologists to control the excavation process. One of the many ways in which they do this is by trying to dig their trenches as squarely as possible—that is, by trying to remove the soil within their trenches in careful blocks. This is because they wish to know precisely where every artefact and feature comes from within the site: obviously if the wall of a trench has been undercut (i.e. if it slopes away from the excavator, into the adjoining area), then any artefacts which are recovered from that undercut zone by rights come from the next square. One of the main aims of good excavation, therefore, is to keep the sides of your trench vertical and the walls and floor square (unless you are digging according to the context system, in which case the base of each context will be following a stratigraphic layer which is highly unlikely to be square). If you are excavating many small contexts, you don’t need to be too pedantic in keeping your walls absolutely vertical; they can be trimmed in a single operation before photographing the trench at the end of each context. Remember to keep the soil from the trimmings and any artefacts it contains separate from the other contexts. You have no way of knowing whether that soil and those artefacts came from the top of the trench, the middle or the base—all you know is that they came from the trench wall.

FIGURE 5.12: The best way to ensure that the sides of your trench are kept vertical is to stand or kneel directly above them and trim from the top down. It is easy to undercut the walls if you try and trim them with the wall in front of you since you can’t see whether you are cutting truly vertically or not. *Archaeologists Field Handbook 29/6/04 5:24 PM Page 144

W144 THE ARCHAEOLOGIST’S FIELD HANDBOOK In the end, the degree of accuracy which is necessary in your excavation technique will depend on the research questions that are being asked: if your research absolutely depends on knowing exact stratigraphic divisions between layers (such as if you are investigating activity areas, or precise changes in use over time), then you will have to proceed with great care; if you are only interested in broad divisions of use or chronology, then a less precise excavation strategy may be adopted—sometimes even extending to excavating the site using arbitrary layers (see ‘The principles of excavation’ on page 121).

Excavation etiquette • Never undercut a trench wall, even if you can see an interesting artefact in the wall of the trench. • Never pull an artefact, stone or other feature out of the wall of the trench. • Never pull an object out of the ground. Excavate around it until you have reached its base and then remove it in one piece. • If an object or unusual feature is uncovered, a good rule is to leave it in place (in situ) until the area is completely excavated and the object can be removed carefully in its proper stratigraphic context. • When working in a rockshelter, avoid wearing boots with big tread as the tread can disturb small artefacts and the fine silty deposits built up in these places. Tennis shoes or some other sturdy footwear with fine tread is best, although at some sites you may be asked to dig bare foot. • Don’t walk on newly excavated areas (particularly someone else’s!) unless it is absolutely unavoidable (i.e. if you need to clean up that area prior to photographing or drawing it). • When you are excavating, always move backwards across the trench to avoid kneeling on the freshly excavated surface. • If weeds or roots are present, cut them with secateurs—don’t pull them. • Don’t trip over the string or the pegs or disturb them in any way during excavation. • Never step too close to the edge of a trench as you may run the risk of collapsing the wall and becoming the most unpopular person on the dig. If you are excavating in a deep trench, get someone to help you in and out, so that neither of you puts your full weight on the edge of the trench. • Never sit on the edge of a trench (for the same reason). • Be willing to take your turn at a variety of tasks. Excavation requires sieving, sorting, cleaning and backfilling as well as actual digging, and no one wants to be restricted to one task all the time. • Bear in mind that tempers get frazzled, particularly on long or arduous digs, and you may have to maintain goodwill even in the face of seemingly overwhelming obstacles. *Archaeologists Field Handbook 29/6/04 5:24 PM Page 145

BASIC EXCAVATION TECHNIQUESW145 It does not help anyone if you sulk, and a temper tantrum can destroy everyone’s morale for the whole day. • Different people have different physical tolerances and may work at different paces, so be patient if someone works more slowly than you do, or needs more guidance. • If you are a trench or site supervisor, be supportive in the way you provide advice to people who are learning to excavate. Try to remember that people aren’t born with excavation skills and that some people will need more specific instructions than others. • If you’re supervising other workers make sure that you brief them as fully as possible before the dig begins. Make clear what your aims are and what you want to know. This will help them to understand the importance of following particular procedures and guide them in terms of what to look for. Update these briefings regularly so that no one feels lost. • If you are camping with other workers during the dig, remember to be tolerant of others, particularly after a hard day’s work. Don’t hog the shower and do your fair share of the household tasks (like cooking or cleaning). Believe it or not, the main sources of irritation on archaeological fieldwork expeditions is who gets priority for the showers and who should be doing the cooking! • Most importantly, fieldwork is a job, not a holiday, so make sure you are always on time and promptly back at work after breaks.

Val Attenbrow’s tips for excavating shell middens • Identify stratification and excavate accordingly. Shell middens vary widely in size, composition and complexity. They range from deposits which are homogeneous throughout to deposits which are finely stratified and may contain, for example, hearths, lenses of specific shell species and/or tool manufacturing events, as well as animal bones and stone artefacts. Excavation should proceed in a manner that ensures any stratification is identified so that excavation units can reflect stratigraphic boundaries. Where a midden appears to be homogeneous, excavators still need to take account of how the deposits may have accumulated—for instance, whether there is a slope in the particular areas being excavated. If the surface of the midden (or an exposed face) suggests materials accumulated on a sloping surface, then the orientation of excavation units should not be horizontal but on a slope that approxi- mates the way the deposits accumulated. This will prevent excavation units from cross-cutting any ‘layers’, even if they are not visible during excavation. *Archaeologists Field Handbook 29/6/04 5:24 PM Page 146

W146 THE ARCHAEOLOGIST’S FIELD HANDBOOK • Choose the most appropriate sieve mesh size. If the midden is composed mostly of whole shells, then a 5 millimetre mesh may be sufficient, unless numerous small (less than 5 millimetre) species are also present. Where stone artefacts are present and fish bones are abundant, then nested 5 and 2 millimetre sieves are recommended, although the shell analysis to identify species composition may be based on materials retained in only the 5 millimetre sieve. In deposits where fish bone is present but sparse, the addition of a sieve with 1 millimetre mesh may be necessary to recover the bones. Depending on the species, the addition of a 1 millimetre sieve where fish bones are abundant may enable more bone to be recovered (i.e. more fragments, greater weight), but it may not necessarily enable a greater number of species to be identified. • Look for shell tools and manufacturing debris. It is often the case that only samples of the shell component of middens are analysed to determine the species composition; sometimes ‘excess’ shell is discarded in the field. In such situations, these deposits should be inspected prior to discard for any shells that may have been used as tools or for shell that is the debris from the manufacture of shell tools. Both bivalves and gastropods were used as scrapers and cutting implements—for example, in processing plant foods—and were also used as adzes to make wooden artefacts. In addition, along the eastern coast of Australia, shell fish-hooks were made from oval pieces cut from Turbo shells. These ‘blanks’ may also be present, along with whole and/or broken hooks. • Make a reference collection. Make a collection of whole shells of all visible species, large and small, from the shoreline adjacent to the excavated midden. This can be used as a reference collection to aid in the identification of midden species. Collect several specimens of each species, particularly gastropods, so that some can be broken up to examine the internal surfaces and shape of the column. If an analysis of meat weights and so on is to be carried out, collect live specimens. Bivalves and gastropods will stay alive for some weeks or can be put into preservative to keep them longer.

SIEVING AND SORTING

Most artefacts from sites are not recovered during the process of digging, but in the subsequent process of sieving the excavated soil. In the interests of stratigraphic control, it is important to sieve each excavated soil layer separately and to keep all artefacts and samples from each layer distinct through careful labelling and bagging. Never allow your bucket of excavated soil to become mixed with buckets from another grid unit or another *Archaeologists Field Handbook 29/6/04 5:24 PM Page 147

BASIC EXCAVATION TECHNIQUESW147 layer—and never allow the artefacts you recover from your bucket to become mixed with others unless they all come from the same context and will be bagged together anyway. Most sieving is done by hand (i.e. by emptying your bucket into a hand-held sieve, shaking it to remove the loose soil, and then sorting through the material trapped in the sieve for artefacts). This is one of the most time-consuming jobs on site, and is usually where a backlog will build up if there are not enough sieves. For this reason, make sure that you have enough sieves (and enough sievers!) to keep the process moving. On some large sites, it may be possible to use mechanical sieves to speed the sieving process, although you will still need the same amount of time to sort through the sieved material for artefacts. Hand-held sieves come in a variety of mesh sizes (2, 3, 4 or 10 millimetres are all standard), and can be ‘nested’ together (i.e. with one fitted over another) so that you can sieve through two mesh sizes simultaneously. When using nested sieves, always remember that the larger mesh size fits over the smaller one. The decision of which mesh sizes to use will depend on the nature of your site and what questions you want answered. A small mesh size will obviously make a big difference to excavating a shell midden which may contain many small and delicate fish bones, but may be redundant on an historical site containing relatively large fragments of glass and ceramics. On sites with clay soils, wet sieving may be the only way to retrieve artefacts. Immers- ing hand sieves containing excavated material into large containers of water can sometimes be effective, although at some sites a pressurised stream of water may be FIGURE 5.13: Sorting through sieved mat- necessary to break down the soil and reveal erial for artefacts artefacts. Wet sieving is a much more labour- intensive process than dry sieving and can damage artefacts, particularly if you have to push the soil through the sieve to break it down. All artefacts removed during wet sieving will have to be allowed to dry before they are bagged. The other major decision in relation to sieving is where to sieve. This may sound trivial but, as all of the soil that goes through the sieves will ultimately have to be put back into the trenches, you should consider very carefully where to place it. Sieve piles can become very large very rapidly, and have a habit of spreading widely at the base, particularly if people continually walk on them. When deciding where to place your spoil heaps: *Archaeologists Field Handbook 29/6/04 5:24 PM Page 148

W148 THE ARCHAEOLOGIST’S FIELD HANDBOOK • Use flat ground not too far from the trenches. (Think how far you can expect people to regularly carry heavy buckets full of soil. Making sure that the spoil heaps are as near as possible to the trenches will also help you when you have to backfill.) • Think about whether or not you will have to clear the ground surface of vegetation first to ensure that you don’t lose any soil. • Think carefully about where you’re likely to excavate, particularly in terms of allowing yourself the option of extending trenches. In other words, don’t place your spoil heaps where you may later want to dig a trench!

Once you have removed all the artefacts from the sieve, they will have to be bagged and labelled to keep track of them. The precise system for tracking and labelling artefacts will depend on the preferences of the excavator or site supervisor, but as a general principle all artefacts will be grouped together by context/stratigraphic unit and placed in clearly labelled finds trays or bags. Any special finds which require immediate conservation should be bagged and labelled separately, and treated immediately (see ‘Conserving finds on site’ and ‘Labelling finds and samples’ on page 150). For detailed information on recommended conservation treatments for excavated materials, see Museum of London (1990) and Watkinson and Neal (1998).

Mike Morwood’s tips for protecting rock art when excavating Because dust is highly abrasive, it can be extremely damaging to painted rock art panels, particularly in the confined space of a rock shelter. To reduce dust during the excavation process:

• Place a screen or curtain (calico is suitable) between the excavation area and the art surface and leave a gap of at least 30 centimetres between the screen and the rock art surface to allow the air to circulate. • ‘Carpet’ the ground surface of the shelter or place wooden planks parallel to the edges of the excavation area to reduce dust stirred by traffic. • Because sieving is the greatest generator of dust, place the sieves on plastic sheeting and contain the spoil heaps on this sheeting. • Sieve downwind of the rock shelter, erect a screen between the sieving area and the general excavation area, and place calico ‘skirts’ around the base of the sieves. • During backfilling, place the spoil into hessian or plastic bags and stack the bags in the trenches. Complete the backfilling by placing a layer of dirt over the bags. (Morwood 1994: 10–12) *Archaeologists Field Handbook 29/6/04 5:24 PM Page 149

BASIC EXCAVATION TECHNIQUESW149 Sometimes artefacts will be sorted into different classes on site (e.g. glass, ceramics, bone, metal, etc.) before being bagged but, once again, the complexity of this process will depend on the size of the site and the preferences of the excavator. If you can, get as much as possible of the basic processing (cleaning, washing, gross sorting) done on site. Remember that for every day you spend in the field, whether surveying or excavating, you should ideally allow three days in the lab or office to process or write up the results— longer if both are necessary (i.e. if you first have to analyse the artefacts and then write up the report). When cleaning artefacts, it is always best to take a gentle approach. Ceramics, glass and stone artefacts can be cleaned with a soft toothbrush and water (Drewett 1999: 145). Other materials, such as metal, bone or shell, are often highly friable and shouldn’t be washed unless it is obvious they won’t be damaged. Most metals are not stable and will deteriorate rapidly when exposed to air. You should use only a dry brush to clean these kinds of fragile materials. Bear in mind, when excavating stone artefacts, that washing will remove any potential residues from the surface of the artefact (see ‘Recovering artefacts with residues and use-wear’ on page 219). If you are worried about damaging any fragile or unusual artefacts, then store them responsibly (see ‘Conserving finds on site’ on page 150) and seek professional advice before you clean them. When cleaning, make sure you keep the overall dig recording system intact. In other words, clean each bag of artefacts separately and make sure that every artefact goes back into the labelled bag from which it came. Never separate the contents of any bag from its label or context identifier.

Anita Smith’s tips for a successful excavation • Carry your own little kit of the following essentials (and have your name on them!): tweezers, pencil, soft brush, photo scale, magnifying glass, compass, line level, clinometer, baby torch, note pad—it all fits into a bum bag. • Sitting on the sieves may not seem glamorous, but it’s where you learn heaps from the more experienced people around you. • If you are unsure—ask, ask, ask and keep asking! There’s nothing worse than realis- ing you’ve discarded the only fragment of non-local stone in the entire site because you didn’t think it was important. • Fieldwork is part of the process, not simply a means to an end. It offers insights into the archaeological material that can never be gained from lab analysis alone. • Take notes, notes and more notes—everyone is diligent on the first day but after that they tend to taper off. Keep writing them regardless of the conditions; they are invaluable when you come to analyse material months or even years later. *Archaeologists Field Handbook 29/6/04 5:24 PM Page 150

W150 THE ARCHAEOLOGIST’S FIELD HANDBOOK CONSERVING FINDS ON SITE

You don’t have to be specially trained as a conservator in order to care for excavated finds. A general rule of thumb is that organic materials should be stored in an environment similar to the one in which they were found: e.g. finds from a damp environment are best kept damp; those from a waterlogged environment are best kept wet; and those from a desiccated environment are best kept dry. Even ordinary soils will hold some moisture content, however, which means that artefacts may have to be dried out before they can be bagged. For this reason avoid placing metal or fragile organic materials directly into sealed plastic bags because the moisture cannot escape and the resulting ‘sweat’ may cause the finds to deteriorate. Instead store them in hard polythene boxes with silica gel to keep the interior of the box dry (Drewett 1999: 146) or in perforated polythene bags to allow moisture to escape. Alternatively you can thoroughly air-dry objects before storage. If you are dealing with a compound object (i.e. one that is made up of more than one type of material, such as a shoe or a hafted tool) then do not separate the components. If you are trying to reassemble broken artefacts (such as ceramic vessels), make sure that any chemical compounds used to hold the pieces together will not damage the artefacts and that the process is reversible. Basi- cally, common sense will always be the best on-site guide—if an artefact is fragile or friable, don’t wash it; handle and store it carefully. Seek professional curatorial advice before doing anything you are unsure of (see ‘Managing archaeological collections’ on page 156).

LABELLING FINDS AND SAMPLES

This may seem trivial, but in fact the labels which you attach to artefacts or samples need to be durable and legible for a very long time: archaeological artefacts are of little value if there is no record of their origin. The most durable labels are made from plastic and aluminium. Because they deteriorate easily, paper and card labels are not widely recommended. When you write on these labels you should use permanent markers (such as Artline 70 pens) so that the writing is as durable as the labels themselves. You can also use basic black ballpoint biros because, although their ink is less permanent, writing with the pen will make an indelible impression in the surface of the label which can still be read even if the ink has faded. When labelling a find or a sample, be sure to include on the label: • the name of the site, or the site prefix; • the excavation trench from which the object came; • the context or layer from which the object came; • the date; • a basic description of the contents (e.g. glass, metal, soil sample). *Archaeologists Field Handbook 29/6/04 5:24 PM Page 151

BASIC EXCAVATION TECHNIQUESW151 Jane Balme’s tips for excavating bone Observing small differences in the appearance of bones during excavation can pay enormous dividends in the quality of information gathered during excavation. The presence of clusters and alignments of bones can tell you about post-depositional processes. Articulated bones indicate a lack of disturbance and hence provide information about the spatial distribution of activities.

• Observations aren’t worth having unless you communicate them to other members of the team and write them down. • A sample of animal bones is useless without labels containing information about their source. This includes their location within the site (square and excavation unit) as well as the mesh size of the sieve from which they were taken. The best way to make sure that you don’t forget some information is to be consistent in your labelling. Write down the information in the same order each time. • Label the sample immediately and never leave any samples lying around unlabelled. • Condensation in plastic bags will cause paper labels to disintegrate. You can reduce condensation by putting a few pin pricks in the bag, but you will need other labels too. Permanent marker on the outside of the plastic bag is good but, because this can scratch off, make sure that you pack the bags in boxes to reduce this problem. • Patience is definitely a virtue when dealing with archaeological bones. You will need patience to pick each bone individually from the samples and lots more when you do the identifications in the laboratory.

COLLECTING SAMPLES IN THE FIELD

Archaeologists often collect samples of materials found during excavation for analysis in the laboratory. This may include samples of charcoal, wood or bone to use for radiocarbon dating, soil for extracting pollen or seed samples, or even small flakes of ochre from rock art for dating or determining the source of the ochre. It is not enough, of course, that you just collect anything that you like in any way that you like. The best samples are those which are found in situ, are properly described and recorded, and which can be linked to archaeologically meaningful features (such as living floors, hearths, specific occupational periods, etc.) (Hester, Shafer and Feder 1997: 323). When collecting samples, you must follow certain procedures to ensure that your samples don’t become useless for later analysis. In particular, you must be aware of: *Archaeologists Field Handbook 29/6/04 5:24 PM Page 152

W152 THE ARCHAEOLOGIST’S FIELD HANDBOOK • the proper methods for collecting different kinds of samples to avoid contamination; • the different quantities of each kind of material which will be sufficient for proper analysis.

Collecting to avoid contamination The major kinds of samples collected by archaeologists are:

• soils with a high proportion of silica, such as sands, or other suitable minerals, such as quartz, feldspar or zircon, for thermoluminescence (TL) dating; • artefacts, provided they have been kept from sunlight, for TL dating; • charcoal, for radiocarbon dating; • soils, for botanical analysis.

Tips for collecting soil samples for thermoluminescence dating Samples that will be submitted for TL dating need to be kept away from sunlight. Thermoluminescence is used to date sediments that have been long buried and will give you a date for when the mineral grains within the sediment were last exposed to sunlight.

• You can collect a sample by inserting plastic piping within the wall of the excavation, but you will need to ensure the sample is not contaminated by grains which fall as the piping is inserted. • Only do this out of direct sunlight (i.e. collect the sample at night, or shade the collection area). • Place your sample directly into an opaque or semi-opaque, water-tight container and put the container immediately into a black plastic bag or other light-tight storage (Byrne 1997). • Make absolutely certain that your TL samples are not exposed to any sunlight at all once they have been collected and keep their exposure to fluorescent light to a minimum (less than ten minutes) (Byrne 1997). • If you must examine the sample, use a torch or a bulb light (Byrne 1997). • You should aim to collect two samples from each level, making sure they are not situated close together, and that each is approximately 0.5 kilograms in weight. This would be roughly equivalent to a lump 10 centimetres across (Byrne 1997). • It is best to collect samples for TL dating as far as possible away from the bedrock, as small grains of bedrock can work their way through the soil profile over time. This is especially important for friable soils, such as areas with sandstone outcrops. *Archaeologists Field Handbook 29/6/04 5:24 PM Page 153

BASIC EXCAVATION TECHNIQUESW153 • Make sure you only collect TL samples which are deeper than 20 centimetres below the ground surface. If they are any closer to the surface, they are likely to be inaccurate (Byrne 1997). • The laboratory which dates your TL samples will also need to know the moisture content of the surrounding soil. You will need to provide them with details of burial conditions, location of the samples, a description of the surrounding deposits, and a rough estimate of how the average water content of the soil relates to the probable content of the samples supplied (Byrne 1997). • If there is any information about seasonal or long-term variations in rainfall, or evidence that the water table might have been anywhere near the context of your samples, the lab will want to know that too (Byrne 1997).

When collecting samples to use for radiocarbon dating, take particular care with the handling and packaging of samples to avoid contamination. Because any form of modern carbon will contaminate an archaeological sample, don’t use paper or cloth bags to hold samples and don’t use cotton wool or tissues as packing materials for samples (Gillespie 1986: 5). The best storage containers are strong polythene bags, aluminium foil and small glass phials. While you are still in the field, pick out all obvious foreign matter from your samples (stones, plant roots and leaves, loose soil or sand) and make sure that you have an adequate quantity for analysis (see Table 5.4 on page 157). Any organic material can be dated with radiocarbon techniques, including:

TABLE 5.3: Some of the range of organic materials that can be radiocarbon dated Textile Charcoal Wood Shell–marine, river, estuarine Sediments/soils/peats Ice cores Plant material (e.g. seeds) Bone Antler Leather Coprolites (fossilised faeces) Paper Pollen Hair Fish remains Parchment Coral Insect remains Avian eggshell Horn

Samples which will be submitted for botanical analysis need to be protected from contamination by modern seeds which can easily blow into a site. Once the sample has been collected, it is impossible to distinguish between contemporary pollen and aged pollen, as they look identical. *Archaeologists Field Handbook 29/6/04 5:24 PM Page 154

W154 THE ARCHAEOLOGIST’S FIELD HANDBOOK Alice Gorman’s tips for collecting samples for radiocarbon dating There are two types of radiocarbon (C14) dating: standard radiometric determinations and AMS (Accelerator Mass Spectrometry), which can be used for samples too small for standard dating. AMS costs more than standard dating, however, so contact the lab for the latest prices before you begin. You will need approximately 500 milligrams for a standard radiocarbon date, but only 5–10 milligrams for an AMS radiocarbon date. When collecting samples:

• Handle them as little as possible. • In preference, pick them up using the point of the trowel or a pair of tweezers. You can use gloves, but be advised that some forms of disposable gloves contain a dusting of corn flour which, as an organic powder, may contaminate your samples. • Wrap each sample in aluminium foil and clearly label it, so that it can easily be distinguished. You can place your sample directly into a plastic bag, but make sure you tell the lab what you have done. • Remember to exclude all modern carbon from your sample, so never include a cardboard or paper label in the bag with your carbon sample. • If you must use a cardboard or paper label, double-bag the sample (put the sample inside one bag, seal it, then place the sample and the label inside another sealed plastic bag). • For advice on sample collection and packaging, don’t hesitate to contact the laboratory.

Sample treatment • Samples should be dry, as bacterial activity in wet samples can affect the final age determination. If you must dry out your sample, use a low-temperature oven which has never been used for radioisotope experiments. Cover the samples lightly with perforated foil and heat at a temperature of at least 40ºC until dry. • Record details of any treatment, such as drying, to submit with the sample.

Documentation • It is important to have full documentation for all samples. Much of this information is routinely recorded for any archaeological procedure; however, some laboratories have specific requirements, so check with them before collection. The kinds of information needed may include: —Collection: date, sample weight, grid references, latitude and longitude, depth of sample and stratigraphic position, stratigraphic relationship to other samples submitted. Was the sample sealed in a recognisable horizon, or sealed in a localised *Archaeologists Field Handbook 29/6/04 5:24 PM Page 155

BASIC EXCAVATION TECHNIQUESW155 feature such as a grave or pit? How secure is the stratigraphic context? Was the sample wet or dry when collected? Can any more material be collected? Did the sample came from a surface or excavated deposit? —Treatment and storage: if the sample was wet, how was it dried, did you use any chemical treatments or preservatives, was the sample cleaned? —Estimated age: this helps the lab to select the appropriate instrument for measurement as well as enabling them to contact you at an early stage if the estimated age seems to be significantly different from the measured age. —Environment: geological, archaeological, palaeoenvironmental, associated cultural, palaeobotanical or other material; perhaps also site sketches and photographs. —Taphonomy, or how the sample got to where you found it: the factors that are relevant here are other, natural, activities which may have affected the carbon content of your sample, such as visible root-penetration in the collection area, evidence of leaching or humus penetration in the soil profile, etc. —Contamination: any other carbonaceous material in the horizon, such as calcium

carbonate (CaCO3)-bearing rocks in the catchment, potential sources of non- contemporaneous carbon, etc. —Nature of the sample: for shells, for example, note whether they are marine or freshwater; note the family/genus/species if known for wood, charcoal, shell, seeds etc.; note the type of bone, e.g. femur.

Submission • Most labs have web pages and submission forms online. If you are sending samples overseas, there may be customs regulations you have to follow. • Make sure you keep copies of all of your submission forms, in case anything is lost or needs to be checked. • Turn-around times vary from four to six weeks to a few months, so check before you send your samples in. • Some sample types require additional pre-treatments to remove contamination, such as dilute acid/alkali treatment for decomposed wood and charcoal, or for peat and lake sediments. Check the website and be aware that you may need to include the cost of pre-treatments in your budget.

For other organic samples (such as bone or shell) • When submitting shells, clean off all soil, sand and debris and air-dry the samples prior to packaging. • Cleaning should only be done with a brass, steel or nylon brush. Never use animal- bristle brushes or organic-fibre brushes. *Archaeologists Field Handbook 29/6/04 5:24 PM Page 156

W156 THE ARCHAEOLOGIST’S FIELD HANDBOOK • Make sure your shell sample consists of one species only, and preferably of large, single shells rather than fragments. • When dating shell samples, make sure that you identify the shell species to the laboratory to avoid inaccurate dates. Shell species in the ocean will absorb less C14 from the atmosphere than lacustrine or riverine species and may return widely different carbon dates. • Shell samples should also be tested for secondary recrystallisation using x-ray diffraction (XRD), so that rogue carbons are not being dated. • Try to avoid cleaning bone samples.

All samples must be properly recorded before collection—it is essential that basic information about possible associations, any evidence of disturbance, the method of collection, the handling and storage procedures you followed, the depth and position of the sample within the excavation unit, and the condition of the sample when collected are all noted at the time of collection. For example, if you handle a charcoal sample with your fingers, you will need to inform the laboratory of this, as the oil on your skin can contaminate the sample. You will need to give details of all the collecting and bagging procedures to the laboratory when you submit your samples, as you may not get accurate dates without this information.

Appropriate quantities for samples This will vary according to how well preserved the material is and the particular technique that is going to be used to analyse it. The best rule is to collect as much of a sample as possible, and to collect more than one sample of material from the same layer so that the results can be cross-checked for discrepancies or anomalies. Table 5.4 on page 157 provides a summary of appropriate quantities to collect of various materials.

MANAGING ARCHAEOLOGICAL COLLECTIONS

What happens once you have collected your artefacts or excavated your site, analysed your results and written your report? You are likely to be left with boxes of excavated material, some of which may be highly fragile or fragmentary, and a large paper and photographic archive. At this point you will have to make a decision about what to do with this material. As a professional, it is your responsibility to make sure this collection is taken care of for the long term, either by turning it over to a recognised authority (a state or local museum or keeping place) or by returning it for reburial on site. Before you do anything *Archaeologists FieldHandbook29/6/045:24PMPage157

TABLE 5.4: Suggested sample quantities Material Minimum weight (g) Optimum weight (g) Roughly Equivalent quantity for standard dating for standard dating equivalent to for AMS dating Charcoal (clean) 2–5 20–30 A handful 5 mg (min); 50 mg (recommended) Charcoal (dirty) 3–5 2–50 Two handfuls Charcoal which is 5–10 50–100 1/2 bag*—four handfuls grey-brown, either hard or more (mineralised) or soft and smeary (leached) with no defined wood structure Charcoal which is finely 1000 5 full bags* spread throughout the deposit (but which ‘smears’ on excavation, rather than being seen as definite lumps) Wood 5–7 30–100 A piece approx. 30cm long 5 mg (min); 100 mg 2 cm wide 2 cm thick (recommended) Cloth or paper 3–5 30–100 A 30 cm 30 cm piece Shell 5–40 100–200 Varies with the type of shell 15 mg (min); 100 mg (recommended) Bone 20–100 100–200 1 mg (min); 30 mg (recommended) Charred bone 200–500 1500–3000 Collagen 100–300 800–3000 Organic (sediment) soil 500 1500–2000 2 medium sized bags* 10 mg (recommended) Seeds 7–10 100 Grass/leaves 5 35–50 Flesh, skin or hair 5 45 Peat 15 100 1/4–1/2 bag* Dung 7 30 * ‘Bag’ = 15 cm 30 cm plastic bag *Archaeologists Field Handbook 29/6/04 5:24 PM Page 158

W158 THE ARCHAEOLOGIST’S FIELD HANDBOOK with a collection, however, you will have to weigh up your ethical responsibilities (see ‘Archaeologists and their profession’ in Chapter 1 on page 15) and make sure that you are aware of your legal obligations under state legislation (which may require you to turn the collection over to a state repository). You will also have to consult with the relevant authorities and other interested parties. While it would be ideal to turn all excavated collections over to an official repository for permanent care, in reality no museum is likely to want a collection of broken glass, fragments of domestic ceramics and miscellaneous rusted metal. You need to find out what the repository’s collection policy is before you begin excavating, because if they agree to accept the collection, they will probably require you to follow particular methods when documenting and storing the archive.

Richard Robins’ tips on the proper care and management of excavated collections If you intend to collect and store any archaeological material, the management of the collection begins at the inception of your research project. As archaeological excavations or surface collections are essentially destructive activities, it is incumbent on you to ensure that the material is preserved in the best possible way. From a research perspective, these collections serve the important function of providing a check on the work of the original collector. From a site-management perspective, well-made and managed collections may reduce the necessity to undertake further collection or excavation on a site. There are three essential elements to proper collection management: organisation, conservation and storage.

Collection planning (organisation) As a first step, consult with staff at the repository where your collection will eventually be housed. Most state museums now have strict depositional requirements and they will be able to tell you the necessary procedures and requirements for the deposition of material. If a museum has a computerised database, use it. It will be a useful tool to coordinate and track the collected material, including the paper archive, and if you use the museum’s official registration system from the outset, you will avoid unnecessary duplication and minimise handling of the artefacts. One of the golden rules of collection management is always to consult with museum staff and conservators before you begin fieldwork. The other is to make sure you obtain some preliminary estimates of the time and costs required to manage and store your collection so that the final figure doesn’t come as a nasty shock.

Conservation Conservation is the most under-rated aspect of any archaeological excavation, so *Archaeologists Field Handbook 29/6/04 5:24 PM Page 159

BASIC EXCAVATION TECHNIQUESW159 you should devise a plan to manage your archive prior to undertaking a collection or excavation. You will need to work out some preliminary costs for conservation, particularly in terms of estimating the time that will be needed to process the material. Prior to collecting any material, you should consult with someone with conservation expertise. They will give you an idea of what you can and can’t do. The following are some of the basic rules that should be followed:

• Immediately following excavation, maintain an environment around the artefact which is as similar to its burial conditions as possible. • The less done to an object the better, so keep handling to a minimum. • Ensure that techniques, particularly those using chemicals, are reversible. It is also helpful to know in advance the kind of laboratory tests you might use and any implications these may have for the way in which you collect or store your samples. The wrong treatment during excavation may make your artefacts or samples useless for analysis. • Document everything done to the artefacts. • Anticipate field conditions and plan accordingly—prepare a field kit with the correct storage materials.

Storage Storage is not only an organisational issue, but also a conservation one. It is essential to use materials that will not decay or damage artefacts or aspects of them, such as blood residues or starch.

Do not use: • matchboxes; • plastic bags with twist ties; • coloured plastic containers; • plastics with PVC; • acidic paper or cardboard; • high-acid tissue; • paper towels; • newspaper; • glass containers; • rubber bands; • pressure-sensitive tapes.

Do use: • clear Ziplock® plastic bags; • clear plastic containers; *Archaeologists Field Handbook 29/6/04 5:24 PM Page 160

W160 THE ARCHAEOLOGIST’S FIELD HANDBOOK • acid-free cardboard boxes; • acid-free tissue paper; • polythene or polyether foam; • Gortex®; • unbleached muslin; • bubble pack without PVC; • silica gel; • natural fibre cloth bags.

Martha Joukowsky’s tips for excellent excavating • Don’t hurry. Excavation requires a calm and purposeful approach. • Remember that all excavation proceeds horizontally first and vertically second. • Be systematic. When a new layer or anything unusual is encountered, stop. Clean off the remaining soil so that none of the material from one layer will be mixed with the one below it. • All grid units should be excavated in the same direction. • Ideally, excavation should proceed from the uphill side of the trench to the downhill side, so that newly excavated areas won’t be trampled. • For the same reasons, always clean higher surfaces before lower ones. • Always keep the trench wall swept clean so that the area being excavated is not contaminated by falling dirt or debris. • Always clean the top of the trench by moving the dirt back from its edges, so that the dirt doesn’t fall into the trench and contaminate other layers. • Always keep the trench walls vertical by cutting sharp right-angled corners at the bottom and by regularly trimming the walls as you go along to ensure they are kept straight. • Don’t wait until the earth dries out before you trim the walls, as there is greater risk of a cave-in when the earth is dry. • Keep the wall trimmings separate to the material excavated from the body of the trench. • Always be on the lookout for soil discolourations around features. This could be an important clue to interpreting those features. • Make sure you record sterile layers in the same way as other layers. • Make sure you record all descriptive information as objectively as possible. • Don’t swing tools (such as picks, mattocks or shovels) higher than your shoulder. *Archaeologists Field Handbook 29/6/04 5:24 PM Page 161

BASIC EXCAVATION TECHNIQUESW161 Don’t cut too deeply with them either, as you may damage buried artefacts. • Only fill buckets two-thirds full with soil. Soil, particularly if it has a high clay content, can be very heavy. (Joukowsky 1980: 172–75)

REFERENCES AND FURTHER READING

Balme, Jane and Beck, Wendy 2002, ‘Starch and charcoal: useful measures of activity areas in archaeological rock shelters’, Journal of Archaeological Science 29: 157–66. Barber, R. 1994, Doing Historical Archaeology: Exercises Using Documentary, Oral and Material Evidence, Prentice Hall, New Jersey. Barker, Philip 1998, Techniques of Archaeological Excavation, Routledge, London. Byrne, D. (ed.) 1997, Standards Manual for Archaeological Practice in Aboriginal Heritage Manage- ment, NSW NPWS, Sydney. Bibby, David I. 1993, ‘Building stratigraphic sequences in excavations: an example from Konstanz, Germany’, in E.C. Harris, M.R. Brown, and G. J. Brown (eds), Practices of Archae- ological Stratigraphy, Academic Press, New York, pp. 104–21. Brown, M.R. and Harris E.C. 1993, ‘Interfaces in archaeological stratigraphy’, in E.C. Harris, M.R. Brown and G.J. Brown (eds), Practices of Archaeological Stratigraphy, Academic Press, London, pp. 7–22. Davies, Martin and Buckley, Krystal 1987, Archaeological Procedures Manual: Port Arthur Conser- vation and Development Project, Occasional Paper No. 13, Department of Lands, Parks and Wildlife, Hobart. Drewett, Peter 1999, Field Archaeology: An Introduction, UCL Press, London. Gillespie, Richard 1986, The Radiocarbon Users’ Handbook, Oxford University Committee for Archaeology, Monograph No. 3, Oxford. Harris, E.C., Brown, M.R. and Brown, G.J. (eds) 1993, Practices of Archaeological Stratigraphy, Academic Press, London. Hester, T.R., Shafer, H.J. and Feder, K.L. 1997, Field Methods in Archaeology, 7th edn, Mayfield Publishing Co, Mountain View, California. Joukowsky, Martha 1980, A Complete Manual of Field Archaeology, Prentice-Hall, Englewood Cliffs, New Jersey. Museum of London 1990, Archaeological Site Manual, 2nd edn, Department of Urban Archae- ology, Museum of London. NSW Department of Planning and the Heritage Council of New South Wales 1993, Historical Archaeological Excavations: A Code of Practice, Government Printer, Sydney. Polach, H., Golson, J. and Head, J. 1983, ‘Radiocarbon dating: a guide for archaeologists on the collection and submission of samples and age reporting practices’, in G. Connah (ed.), Australian Field Archaeology: A Guide to Techniques, AIAS, Canberra, pp. 145–52. Roskams, Steve 2001, Excavation, Cambridge Manuals in Archaeology, Cambridge University Press, Cambridge. *Archaeologists Field Handbook 29/6/04 5:24 PM Page 162

W162 THE ARCHAEOLOGIST’S FIELD HANDBOOK Watkinson, D. and Neal, V. 1998, First Aid for Finds, Rescue/UK Institute for Conservation Archaeology Section, London.

USEFUL WEBSITES

For details of pre-treatment for different classes of excavated material see: www.c14dating.com/pret.html. Free digital checkers for Munsell soil charts are available at: www.munsell.com/. A history of the Harris matrix and various publications on its uses can be found at: www.harrismatrix.com/history.htm. The Australian Institute for the Conservation of Cultural Material Inc (AICCM) maintains a state-by-state listing of professional conservators at www.aiccm.org.au/aiccm/people. Chris Clarkson and Sue O’Connor

6 An Introduction to Stone Artifact Analysis

Introduction

Perhaps the best place to begin this chapter is by stating what it is not. This chapter is not a “cookbook” of methods and techniques for aspiring stone analysts confronted with an assemblage of stone artifacts for the ﬁrst time, or those seeking to bolster their work with the latest literature or new techniques. Nor is it an exhaustive overview of analysis conducted over the past few decades that details their strengths and weaknesses, and points to major theoretical stumbling blocks or methodological advancements. Readers in search of such critical reviews are directed to Odell (2000, 2001b) for a global perspective (but with a North American focus), Dibble (1995) for a view of emerging continental schools of thought, and Hiscock and Clarkson (2000) for a review of pressing issues in Australian lithic studies. Rather, this chapter aims to arm the student of lithic technology with a set of principles to guide the construction of their research design, alert them to the philosophical underpinnings of various kinds of stone analysis, point to simple but frequently overlooked issues of data management, provide an overview of some common laboratory techniques, and provide case studies and suggested readings that offer insight into both the process of actually doing stone analysis and drawing meaningful conclusions from the results. It takes a “question and answer” format, in the hope that some frequently asked questions might be addressed in a straightforward manner.

There are a number of very good reasons why archaeologists study stone An overview artifacts. Primary among them is the fact that stone artifacts are usually the most durable and often numerous remains of past human activities, and in Why study stone many cases constitute the only surviving trace of people that lived hundreds, artifacts? thousands, and even millions of years ago (in the case of our recent hominid

AN INTRODUCTION TO STONE ARTIFACT ANALYSIS 159 ancestors). Because they survive under conditions that typically destroy most other human creations and castoffs, stone artifacts are ubiquitous in the landscape. A second reason for studying stone artifacts is that for most of human history stone tools played a vital role in our day-to-day survival, in shaping the physical world to our various needs, and even in signifying to others our identity and place in the world. They therefore constitute a vast and invaluable record of the enormous diversity of strategies people have devised to make a living, solve common problems, communicate, and to live and compete with one another. As this chapter deals exclusively with methodological issues and laboratory techniques, it offers little discussion of the sorts of theoretical frameworks that might adopt these techniques in addressing the “big questions” in archaeology. The potential for lithic analysis to engage with disciplinary theory, however, now seems far more practicable than at any time in the past. For instance, a great deal of thought has been given to the place of technology as an integral aspect of cultural variability, adaptation, and change (Lemonnier 1986; Pfaffenberger 1992; Bleed 1997; Schiffer & Skibo 1997), the social and evolutionary mechanisms giving rise to technological innovation (van der Leeuw & Torrence 1989; Bamforth and Bleed 1997), the behavioral and physical factors governing variability in both individual artifacts (such as fracture mechanics and the effects of reduction intensity) (Dibble & Whittaker 1981; Cotterell & Kamminga 1987; Dibble & Pelcin 1995; Pelcin 1997a, 1998; Shott et al. 2000; Macgregor 2005) and whole assemblages (such as patterns of artifact procurement, transport, use and discard) (Binford 1979; Shott 1989; Torrence 1989; Nelson 1991; Kuhn 1995), the symbolic role of stone in communicating social, political, and ideological relationships or differences (Ingold 1990; Sinclair 1995; Wurz 1999; Harrison 2002), the role of social agency in stone artifact manufacture and use (Dobres 2000; Sinclair 2000), the technological signatures of various mechanisms of trade and exchange (see, among many more, Renfrew et al. 1968; Ericson & Earle 1982; Zeitlin 1982; Torrence 1986; Peterson et al. 1997; Torrence & Summerhayes 1997; Specht 2002), as well as stone artifacts as markers of gender (Gero 1991; Sassaman 1992; Dobres 1995; Walthall & Holley 1997). Most recently, archaeologists have begun to explore technological variability using formal optimality models drawn from evolutionary ecology (Bright et al. 2002; Brantingham 2003; Ugan et al. 2003). Many of these studies are moving toward the development of new theoretical approaches for explaining assemblage variation.

What are stone artifacts? A stone artifact is any piece of rock modified by human behavior, whether intentionally or unintentionally. Although this definition could properly be applied to extreme and even ridiculous cases, such as humanly modified land- scapes, aqueducts, or open cut mines, it is most often used to signify portable, chipped, ground, or pecked stone objects created by a single or small group of individuals, and usually in the context of hunter–gatherer, pastoralist, early agricultural, or other nonindustrialized societies.

160 CHRIS CLARKSON AND SUE O’CONNOR Most people are familiar with the simplest form of stone artifact manufacture How are they made? commonly portrayed in depictions of our early ancestors banging two rocks together. While this is, generally speaking, the way most stone artifacts were made, there is nothing simple about controlling the process to the degree that allows artifacts of specific shapes to be accurately and repeatedly produced from a block of stone, as was achieved by prehistoric artisans with sometimes startling finesse. The symmetry and regularity of some of the highest known forms of flintknapping can be astounding, as seen for instance in the fluted Folsom points of north American Paleo-Indians, the Solutrean points of Upper Paleolithic Europe, the flint daggers of the Danish Neolithic, the Gerzian ripple-flaked knives of Late Stone Age Egypt, or the obsidian eccentrics and polyhedral blades of Mayan and Aztec artisans (Figure 6.1). In reality though, most stoneworking tended to be far less sophisticated than these examples suggest (in terms of the precision and investment of labor), and literally involved the striking of flakes of varying shapes and sizes from a block of stone (a core), using a stone pebble (a hammerstone), or some hard object (an indentor) such as a piece of bone, antler, or hard wood. Re- moving a flake from a block of stone creates a positive scar or ventral surface, on the flake, and leaves behind a negative flake scar on the core. The opposite side to the ventral surface on the resulting flake is called the dorsal surface. Cores are artifacts that possess only negative flake scars. Flakes that have had other flakes removed from their surfaces after they were struck from the core are called retouched flakes. Because flakes can be removed from the dorsal surface of a flake before or after it is struck from a core, the term “retouched flake” is reserved only for artifacts that show clear signs of flakes having been detached after the creation of the ventral surface, and hence scars must either derive from or modify the ventral surface in some way to be treated as retouch. The term “nucleus” will be used in the following discussion to refer to any body from which flakes have been removed, whether flakes or cores. The process of fracture propagation that underlies flaked stone artifact manufacture is complex, and the effects of various core morphologies on the fracture path are not well understood, even by engineers. Yet it is the fracture path that ultimately determines the morphology of flakes and cores, and archaeologists have therefore begun to try to understand this process. Due to the complexity of this subject, readers are directed to a number of papers that provide detailed overviews of fracture mechanics for archaeologists (Cotterell & Kamminga 1977, 1987; Phagan 1985), as well as more narrowly focused experimental investigations (Dibble & Whittaker 1981; Phagan 1985; Dibble & Pelcin 1995; Dibble 1997; Pelcin 1997a,c, 1998; Shott et al. 2000; Macgregor 2005). Without delving into the details, it is possible to briefly describe some of the main principles and the most common fracture features that result. First of all, only a limited number of stone types are well suited to making flaked stone artifacts, and these generally possess three qualities: they are elastic,

AN INTRODUCTION TO STONE ARTIFACT ANALYSIS 161 (a) (b) (c) (d)

2cm 2cm 2cm 2cm

(e)

(f)

2cm

Figure 6.1 Examples of some of the highest known achievements in stone artifact manufacture: (a) a fluted Folsom point; (b) an Upper Paleolithic Solutrean point; (c) a Danish Neolithic flint dagger; (d) a Late Stone Age Egyptian Gerzian ripple-flaked knife; (e) a Mayan chert eccentric; (f) Aztec obsidian pressure blades and cores (from Whittaker 1994, copyright © 1994, by permission of the University of Texas Press).

in that they will temporarily deform when force is applied to them; they are brittle, in the sense that they will fracture if the applied force exceeds the capacity of the material to deform elastically; and they are isotropic, meaning they are equally susceptible to fracture in any direction and will not preferen- tially fracture along particular planes.

162 CHRIS CLARKSON AND SUE O’CONNOR Cryptocrystalline or amorphous silicates (such as chert, chalcedony, and flint), monocrystalline or microcrystalline silicates (crystal quartz and “milky” quartz), acrystalline silicates (such as glass and obsidian) and some larger-grained and less homogeneous materials such as silcrete and quartzite all possess these qualities to varying degrees and are commonly employed in flaked stone artifact manufacture (Cotterell & Kamminga 1987; Kooyman 2000). In most forms of flaking, force is directed into the platform (i.e., any surface receiving force) of a nucleus with an indentor (any object imparting force to a nucleus) using one of four techniques: striking the nucleus at high velocity with either a hard indentor such as a hammerstone (hard hammer percussion) or a soft indentor such as a piece of wood, bone or antler (soft hammer percussion); slowly applying pressure through a process called dynamic loading (pressure flaking); striking a positioned punch (indirect percussion); or applying compressive force by placing the nucleus on an anvil and striking it from above (the bipolar technique) (Cotterell & Kamminga 1987; Kooyman 2000). Skilled flintknappers observe that in most flaking, force is generally directed into the nucleus using both an inward and outward motion (Crabtree 1972a; Whittaker 1994), creating both “opening” and “shearing” stresses in the nucleus (Figure 6.2a). Fracture occurs when stresses within the nucleus reach a critical threshold and break the molecular bonds that hold the nucleus together. The most common form of fracture is known as conchoidal fracture, which begins from preexisting flaws in the surface of the nucleus close to the point of impact and creates what is known as a Hertzian cone, as illustrated in Fig- ure 6.2b. The Hertzian cone propagates in a circle around the contact area and expands down into the nucleus in a cone shape at an angle that is partly dependent on the angle of the applied force. If the nucleus is struck close to the edge, only a partial cone will be visible on the flake (Figure 6.2b). Whether or not a fracture will continue to propagate through the core once a cone is formed (i.e., rather than just leaving an incipient cone in the nucleus), depends on whether the force of the blow is sufficient to accelerate and overcome the inertia of the material that is to be removed. Once fracture is initiated, a number of counteracting stresses created by the magnitude and direction of force (tensile, bending, and compressive stresses) will influence the path that it then takes through the core. In conchoidal fracture, the path will typically first head into the core before diving back toward the free face, creating the bulb of force, and then stabilizing on a path that is more or less parallel to the free surface. Conchoidal flakes (i.e., those with Hertzian initiations) often retain a ring crack at the point of force application (PFA), and an eraillure scar just below the point of percussion on the bulb of force (Figure 6.3). Undulations in the fracture path also often leave compression waves on the ventral surface of flakes. Fissures radiating out from the point of percussion are also often found on the ventral surfaces of flakes, but are most often seen on fine-grained materials.

AN INTRODUCTION TO STONE ARTIFACT ANALYSIS 163 (a) Opening force Shearing force Direction of force

Platform Opening force Shearing force

Fracture Exterior platform angle Nucleus

(b)

Complete Hertzian cone formed by spherical indentor near the center of a rectangular prism, and a partial cone formed near the edge

Platform

Lip Bulb of Free force surface Fracture

(d) Feather Step Hinge Plunging (outrépassé)

(i) (ii)

Figure 6.2 Types and features of fracture initiation and termination: (a) fracture forces; (b) Hertzian cones; (c) fracture initiations; (d) termination types (after Cotterell & Kamminga 1987; Andrefsky 1995).

Force eventually exits the nucleus either gradually and at a low angle, creating a feather termination, or more rapidly and at around 90 degrees, creating a step or hinge termination (Figure 6.2d). Not all fractures follow this path, however, and the fracture path sometimes travels away from the free surface and exits on the other side of the nucleus, creating a plunging or outrépassé termination (Figure 6.2d). Pelcin’s (1997c: 1111) controlled experiments have

164 CHRIS CLARKSON AND SUE O’CONNOR Platform Proximal Ring crack (PFA)

Cone of force Bulb of force Erraillure scar

Dorsal Ventral

Medial Fissures

Lateral margins

Compression waves

Distal

Figure 6.3 Fracture features often found on the ventral and dorsal faces of a conchoidal ﬂake (reproduced courtesy of the Trustees of the British Museum).

shown that when all other variables are held constant, increasing platform thickness will produce regular changes in termination type from feather through to hinge terminations, as the force becomes insufficient to run the length of the free face. The direction of force is also often implicated as a determinant of either hinge or step terminations, but this proposition has not been tested under controlled circumstances. Others have suggested that thick platforms and inward-directed force are more likely to produce outrépassé terminations, given sufficient force to initiate a fracture (Crabtree 1968; Phagan 1985: 237, 243). Less commonly, fracture will initiate behind the point of percussion, creating a bending initiation, which dives rapidly toward the free face without forming a Hertzian cone, and leaves a pronounced “lip” on the ventral edge of the platform (Figure 6.2c). Bending initiations are most commonly formed on nuclei with low angled platforms and have a fracture surface that often resem- bles a diffuse bulb, even though no bulb is present (Cotterell & Kamminga 1987: 689). Although it has long been thought that bending initiations are typically produced by soft hammer and pressure flaking, Pelcin (1997c: 1111) found that bending initiations were repeatedly created on cores with low platform angles when blows were placed relatively far in from the edge, suggesting that their frequent association with soft hammer and pressure flaking is more likely a factor of the common use of these techniques in knapping cores with low platform angles (e.g., bifaces) than it is of either force or indentor type. Pelcin (1997b) was also able to show that soft hammer flakes were on average longer and thinner than hard hammer flakes, and that this technique was therefore better suited to bifacial thinning than hard hammer percussion.

AN INTRODUCTION TO STONE ARTIFACT ANALYSIS 165 Hence the association between soft hammer/pressure and bending initiations is likely to be coincidental rather than causally linked. Compression fractures created by bidirectional forces produce a wedging initiation that results in flattish fracture surfaces without a bulb of force (Fig- ure 6.2b; and see Cotterell & Kamminga 1987). Because compression fractures are typically initiated by particles driven into existing percussion cracks, flakes created through this process often exhibit battered or crushed platforms with cascading step scars on the platform edge (Cotterell & Kamminga 1987). Bipo- lar cores and flakes that have been rested on an anvil most commonly display this form of initiation. Because the anvil on which the nucleus is supported can also act like a hammerstone, bipolar flakes can at times exhibit platform and initiation features at both ends, such as crushing, dual bulbs of force, and bidirectional compression waves. When nuclei are stabilized on an anvil, problems of inertia – or the probability of a blow moving the core rather than detaching a flake – can be dramatically reduced. This technique is therefore ideally suited to working very small cores (Hiscock 1982). Recent controlled fracture experiments have revealed that the closer the Hertzian cone is to the edge of the nucleus, and the lower the external platform angle (EPA), the less material needs to be accelerated away from the core, and hence the less force will be required to initiate a fracture (Speth 1974, 1981; Dibble & Whittaker 1981; Dibble & Pelcin 1995; Pelcin 1997a–c). The more these variables are reduced, however, the smaller the resulting flake will be. This relationship is illustrated in Figure 6.4a, and can be seen to be a simple result of changing core geometry. Alternatively, increasing platform

(a) (b) Increasing platform thickness Increasing platform thickness

5145123 4 23

Low exterior High exterior platform angle platform angle

Increasing length

Figure 6.4 The effects of increasing or decreasing platform angle and platform thickness: (a) low exterior platform angle; (b) high exterior platform angle.

166 CHRIS CLARKSON AND SUE O’CONNOR angle and striking further from the edge requires greater force input to initiate a fracture, but also results in larger flakes (Figure 6.4b). Increasing force input by too much can result in longitudinal splitting of the flake or crushing of the platform edge. At some point, increasing EPA and/or platform angle will reach a threshold at which the amount of force required to detach a flake will exceed the inertia of the nucleus itself, and will result in moving the nucleus rather than detaching a flake (Phagan 1985: 247). At this point, force requirements can be reduced by decreasing EPA, platform thickness or both, or by stabilizing the core on an anvil. Most recently, Macgregor’s (2005) experiments have demonstrated that removing some of the mass of the free face (such as might occur through overhang removal, for instance) allows a blow to be placed further from the platform edge (given the same amount of force) than would have been possible were it not removed (thereby detaching a larger flake). Furthermore, Macgregor found that the morphology of the free face directly affected the morphology of the resulting flakes. His experiments demonstrated that features such as large preexisting step or hinge terminations on the free face will decrease the viable platform area at which fractures can be successfully initiated. In the case of preexisting step and hinge fractures, more force and the placement of blows further into the nucleus was required to successfully remove a preexisting step or hinge termination without adding another one. It can be expected then that as more step and hinge terminations build up on the dorsal surface, it will become increasingly difficult to remove them from the free face, as the viable platform area will become too small and the amount of force required too excessive to strike off a flake without shattering the platform, adding new step terminations, splitting the flake longitudinally, creating an outrépassé termination, or failing to initiate a flake altogether. A recent study by Pelcin (1997a) also demonstrated that varying the shape of the free face morphology affected the dimensions of the resulting flakes. His findings confirm the observations of flintknappers that setting up ridges running the length of the core face aids the production of longer, thinner, and more parallel-sided flakes (Crabtree 1972a: 31; Whittaker 1994: 106). Thus, a number of trade-offs exist between the interdependent variables of platform size, platform angle, core inertia, force input, and nucleus morphology that knappers must manipulate to gain control over the fracture path and to extend the reduction of raw materials. A large number of strategies were employed in the past to modify force variables, rectify problematic morphologies, and prevent prematurely damaging the nucleus. Some of these strategies are listed in Table 6.1. These focus on variables that are under the direct control of the knapper and tend to be visible archaeologically. As should be apparent by now, fracture mechanics plays a preeminent role in shaping each individual artifact. It is important to keep this in mind when inferring the meaning of variation in flake and core form. While different forms could be interpreted as having stylistic or functional meaning,

AN INTRODUCTION TO STONE ARTIFACT ANALYSIS 167 Reference(s)

Dibble and Whittaker (1981), Phagan (1985: 237), Dibble and Pelcin (1995)

Phagan (1985: 237)

Speth (1974, 1981), Dibble and Whittaker (1981), Dibble and Pelcin (1995), Pelcin (1997a–c)

Dibble and Whittaker (1981), Phagan (1985: 237), Dibble and Pelcin (1995)

Negative effect

Reduces control over the fracture path by complicating the platform morphology

May result in smaller ﬂakes; if blow is placed too close to the edge, the platform may shatter

Reduces the size, mass, and inertia of the nucleus

May create step or hinge terminations to the free face

Increases force requirements, removes mass more quickly from the nucleus, increases platform angle by removing more material from the platform end of the nucleus

May result in smaller ﬂakes; if blow is placed too close to the edge, the platform may shatter

Reduces control over the fracture path by complicating the platform morphology

Positive effect

Reduces EPA and force requirements by removing ﬂakes from the platform surface

Reduces force requirements by reducing the amount of mass that must be accelerated

Creates a new platform with lower angles

Increases platform angle and strength; allows blows to be placed further in from the edge, creating larger ﬂakes

Reduces chances of platform crushing and results in bigger (heavier) ﬂakes

Reduces force input requirements by reducing the amount of mass that must be accelerated

Reduces EPA and force requirements by removing ﬂakes from the platform surface

Strategy

Faceting

Decrease platform area of intended ﬂake

Core rotation

Overhang removal

Increase platform thickness of intended ﬂake

Decrease platform area of intended ﬂake

Faceting

Common problems, solutions, and negative effects of various stoneworking procedures.

Table 6.1

Problem

High platform angles (excessive force requirements)

Low platform angles (decreased ﬂake size and increased platform crushing)

Low nucleus inertia

168 CHRIS CLARKSON AND SUE O’CONNOR Phagan (1985: 247)

Hiscock (1982, 1996), Cotterell & Kamminga (1987)

Phagan (1985: 247)

Speth (1972: 38), Phagan (1985)

Macgregor (2005)

Less control over force delivery

May shatter the nucleus through excessive force

Harder to initiate fracture

Flake may terminate abruptly if insufﬁcient force; excessive force may result in a plunging termination or shattering the ﬂake

Flake may terminate abruptly if insufﬁcient force to overcome the irregularity; excessive force may result in a plunging termination or shattering the ﬂake

Can increase curvature of the free face, resulting in more curved ﬂakes; new platforms can encounter irregularities left by knapping from previous platforms

Increases the inertia of the nucleus by supporting it against a larger object

Increases the inertia of the nucleus by resting it on an anvil and imparting a compressive force

By increasing the speed of force input (e.g., by using a faster swing and/or lighter indentor, or a longer indentor that enables greater leverage from the wrist), force can be imparted to the nucleus faster than its inertia can be overcome through movement

Increases the coefﬁcient of friction and creates micro-ﬂaws in the surface

Removes projections, irregularities or preexisting step or hinge terminations from the free face by removing larger, thicker ﬂakes

Removes problematic features gradually

Projections, irregularities, or preexisting step or hinge terminations are removed from the free face from the opposite end

Stabilize core

Bipolar technique

Increase speed of force input

Grinding and/or faceting

Increase platform thickness of intended ﬂake, platform angle, and force input

Position blow to left or right

Rotate nucleus

Insufﬁcient platform friction

Poor free face morphology

AN INTRODUCTION TO STONE ARTIFACT ANALYSIS 169 Reference(s)

Crabtree (1968, 1972b: 60), Kobayashi (1985), Pelcin (1997b)

Dibble (1997)

Crabtree (1968, 1972b), Phagan (1985), Pelcin (1997a), Andrefsky (1998), Kooyman (2000)

Negative effect

Thinner ﬂakes will have a greater chance of transverse snapping due to “end shock”

Greater chance of longitudinally splitting the ﬂake

Increases the probability of transverse breaks due to “end shock”

As above

Positive effect

Results in thinner bulbs, and hence thinner ﬂakes, creating higher cutting edge to weight ratios

Produces a higher cutting edge to weight ratio while minimizing increases in platform angle

Produces longer, thinner ﬂakes with a higher cutting edge to weight ratio

Use several of the strategies listed above, such as core rotation, stabilization or bipolar working, preparing the platform (faceting and overhang removal), changing the indentor type (e.g., soft hammer), and adjusting the platform size (increasing or decreasing platform the thickness and width)

Strategy

Soft hammer technique

Increase platform thickness of intended ﬂake relative to thickness

Setup arises on core face

Extend reduction of nucleus

)

cont’d

(

Table 6.1

Problem

Flakes have insufﬁcient cutting edge for weight

No replacement raw material

170 CHRIS CLARKSON AND SUE O’CONNOR they might just as well relate to the methods employed in working various raw materials, to create flakes of different shapes, to prolong reduction, or to overcome certain difficulties. Pecking and grinding are quite different manufacturing processes to flaking. Pecking involves either dislodging grains or small pieces of material from the surface of a nucleus, or creating small and intersecting impact pits (incipient cones of force) over the surface of the nucleus until a specific shape is attained (Crabtree 1972a). Grinding, either on or with an abrasive material, likewise gradually wears away the surface of an artifact and usually results in the formation of many parallel striations (sometimes microscopic) aligned in the direction of the grinding motion that may blur preexisting fracture features or polish high points on the surface of the artifact.

Above all, the recognition of fracture features and the various techniques How do you recognize employed by past knappers to rectify problems or improve their control over different techniques? the fracture path requires experience. Replicative flintknapping also provides a rapid way of improving your identification skills by generating large numbers of flakes and cores showing a range of features created using known techniques. Flintknapping can also provide a means of generating hypotheses about how an assemblage might have been created, although analogical arguments of this kind do not provide tests in themselves of the various procedures used in the past. Only the archaeological record itself can provide such tests (e.g., refitting and attribute analysis; Schindler et al. 1984: 176). It is often difficult to develop consistent sets of criteria to reliably identify specific procedures, as the case of soft hammer percussion discussed above demonstrates, but fortunately the recognition of some of the most common techniques is quite straightforward. A list of some of the commonly employed features used to identify various techniques, compiled from the observations of archaeologists and flintknappers, is presented in Table 6.2 (see, among many others, Crabtree 1972a; Cotterell & Kamminga 1987; Ahler 1989; Hayden & Hutchings 1989; Whittaker 1994), although such features should be used with extreme caution. Entire assemblages should also provide a better “feel” for the use of dominant techniques than should individual specimens (Kooyman 2000: 78). According to replicative flintknappers (Crabtree 1972a; Newcomer 1975; Whittaker 1994), hard hammer techniques more frequently produce pronounced bulbs of force, compression waves and ring cracks, and expanding flake margins, whereas soft hammer technique produces more diffuse bulbs, flatter fracture surfaces, and narrower flakes. These observations are borne out to some degree by controlled experiments (Cotterell & Kamminga 1987: 686; Pelcin 1997b), although it is difficult to know how well controlled observations translate to archaeological assemblages in which a wide range of variables have presumably varied freely (Dibble 1997: 151). Studies of replicative flintknapping debitage, or the by-products of flaking, have produced arguments

AN INTRODUCTION TO STONE ARTIFACT ANALYSIS 171 N/A

N/A

pecked

Present

N/A

Variable Variable

N/A

Absent

ground implement implement

Present

Abrasion: Pecking:

Crushed

Flat/pronounced N/A

Flat

Absent

ﬂake

Crushed

Absent

Platform

Bipolar

None/pronounced N/A

Rare

Facetted

Pressure

Diffuse

ﬂake

Thin and parallel Parallel

Pronounced

Variable

Much smaller

Small

Subdued and

May be present

Rare/deep

Platform

Absent

widely spaced

Flaking

Diffuse

Soft

Tend to feather

hammer ﬂake

(20–60%)

Thin and expanding

Facetted/crushed

Pronounced

to plano-convex

Subdued and widely spaced

Rare

Thinner than hard hammer

Variable size, tend

Rare (5–10%)

Platform

Absent

1%) Common Common

Thicker than soft hammer

Variable

Pronounced

Variable

N/A

Hard hammer ﬂake

Tend to large size and triangular

May be present

closely spaced

Common (95%)/shallow Less common/shallow

Common (60–80%)

A list of features and their supposed frequency in various forms of stone artifact manufacture.

Table 6.2

Feature

Platform

Bulb of force

Compression waves Pronounced and

Bulbar ﬁssures

Erraillure scar

Ring crack

Bending Initiation Rare (

Shape

Platform scarring Variable

Ventral curvature Variable Thickness

Termination

Striations

Impact pitting

172 CHRIS CLARKSON AND SUE O’CONNOR both for and against feasible identification of soft hammer working in archaeological assemblages (Mewhinney 1964; Touhy 1987). Pressure techniques can sometimes be quite distinctive and recognizable on retouched implements (Akerman & Bindon 1995). Likewise, some analysts believe they can recognize the flakes produced during pressure flaking from a combination of size, thinness, bending initiations, high ventral curvature, and a complex platform and dorsal morphology. As flintknappers point out, however, these same features can be created by percussion flaking, and cannot be considered diagnostic of any one technique in and of themselves (Touhy 1987; Whittaker 1994). Bipolar flaking also presents difficulties for consistent identification (Jeske & Lurie 1992). Crushing of the platform edge, together with a flattish fracture surface and a battered distal end are the usual criteria employed in identifying bipolar flakes, although not all flakes removed from bipolar cores possess these features (Cotterell & Kamminga 1987), and some possess platform features at both ends, or crushing in addition to fully formed Hertzian initiations. Negative scars can sometimes also appear on the ventral surfaces of bipolar flakes, directed from either end as a result of the crushing blow. Bipolar flakes also are not easily separated from bipolar cores, but the presence of a single flat scar on one face may serve as a guide, whereas bipolar cores may tend to exhibit a number of scars on all faces. The identification of stoneworking techniques such as overhang removal, faceting, core rotation, retouching, and burination is generally more straightforward. Figure 6.5 illustrates the characteristics of overhang removal and faceting. Overhang removal can be identified by the presence of a series of

Flake Core Faceting Faceting

(Platform view)

Overhang removal Overhang removal

FLAKE CORE (Dorsal and core face view)

Figure 6.5 Platform features indicative of various preparatory techniques (arrows indicate the location of blows).

AN INTRODUCTION TO STONE ARTIFACT ANALYSIS 173 smallish scars initiated from the platform surface onto the dorsal surface of flakes or the face of cores. Overhang removal is performed by rubbing or gently tapping the edge of the core to remove the lip remaining after previous flake removals. Faceting looks much like overhang removal, but is oriented in the reverse direction, with smallish flake scars initiated from the dorsal surface onto the platform surface of cores and flakes. There is no real size cutoff between overhang removal or faceting flakes and other dorsal flake scars, and most analysts either employ an arbitrary cutoff (we use 15 mm), or simply use their intuition. Attempts have also been made to identify the distinctive features of overhang removal and faceting flakes so that they may be identified in archaeological assemblages (Newcomer & Karlin 1987). Core rotation is identified simply by the presence of a number of platforms on cores, or by the existence of truncated flake scars that originate from a point at which a platform no longer exists. Core rotation can also be detected by the presence of redirecting flakes that preserve old platform edges on their dorsal surfaces at different orientations to the current platform (Figure 6.6a). Not every rotation of a core will result in a redirecting flake, and many rotations simply result in striking cortical flakes or flakes with complex platform morphologies (see “How do you measure flake reduction?”). Some of the potential uses of these three stoneworking techniques are listed in Table 6.1. Retouching is also easily identified if flake scars can clearly be seen to initiate from or modify the ventral surface (Figure 6.6c), but in cases where flaking is initiated from the dorsal surface without clearly modifying the ventral surface, it is often hard to be sure whether it is retouch or preexisting dorsal scars that are present. A classic case of this problem occurs in Australia, where redirection flakes with old steep platform edges on their dorsal surfaces are misidentified as backing retouch. The key to the proper identification of retouch therefore is to locate the actual point of initiation of scars in order to determine whether they were formed before or after the creation of the ventral face. Lateral spalling of the margins, or burination, is another form of retouching that can be misidentified as preexisting dorsal scarring or old platforms (Figure 6.6b). Bifacial reduction is recognized on cores and flakes as flaking that is directed from either side of the platform edge or lateral margin (Figure 6.6e). Modern flintknappers have identified a set of criteria that they believe can be used to consistently recognize the debris resulting from reduction of bifacial cores and bifacially retouched flakes. These include the high prevalence of bending initiations, pronounced curvature along the percussion axis, low platform angles, faceted and or ground platforms, and complex dorsal scar patterns that remove a portion of the opposite margin (Bordes 1972; Crabtree 1972a,b; Touhy 1987; Patterson 1990; Whittaker 1994: 196). Once again, it is unclear what proportion of bifacial debitage displays some or all of these features. A large number of recent studies have employed a range of techniques, such as mass and attribute analysis on replicated debitage (Patterson & Sollberger

174 CHRIS CLARKSON AND SUE O’CONNOR (a) (b) Burinate retouch

Redirecting ﬂake Burin spall Old burin scars

Dorsal Ventral

(d)

Dorsal Ventral

(e)

Dorsal Ventral

Figure 6.6 Various reduction techniques and forms of retouch: (a) platform redirection; (b) burination; (c) dorsal retouch; (d) ventral retouch; (e) bifacial retouch.

1978; Patterson 1982, 1990; Stahle & Dunn 1984; Ahler 1989; Odell 1989; Shott 1996; Austin 1997; Steffen et al. 1998), breakage patterns (Sullivan & Rozen 1985; Baulmer & Downum 1989; Prentiss & Romanski 1989), or a combination of these (Morrow 1997a; Bradbury 1998), to try to differentiate the various techniques used to create archaeological assemblages (such as hard and soft hammer, pressure, bifacial reduction, and core versus ﬂake reduction), but with varying degrees of success (Prentiss & Romanski 1989; Shott 1994; Prentiss 1998; Bradbury & Carr 1999). A ﬁnal technique discussed here is heat treatment. Although the manner in which heat treatment works is still not well understood, this technique often

AN INTRODUCTION TO STONE ARTIFACT ANALYSIS 175 results in improvement of the strength (Purdy & Brooks 1971) and ﬂaking properties of stone (Crabtree & Butler 1964). Heat treatment is often discussed in the technological literature, but discrimination between deliberate and accidental thermal alteration is almost impossible, and requires careful attention to the context of heat application and the range of assemblage elements affected. When properly executed, thermal alteration can cause ﬁne-grained materials to acquire a “greasy luster” or change color. Alteration to the homo- geneity of the stone can also be seen directly via electron microscopy when samples of the same stone with and without heating are compared (Purdy & Brooks 1971; Flenniken & White 1983). Excessive or rapid heating and cooling can result in the formation of pot lid scarring, crenated fractures, crazing, spalling, and color alteration, but the presence of these features does not necessarily imply that heating was unintended.

Analyzing Stone The first step in any analysis should be to determine what it is that you are Artifacts trying to find out, and what analytical techniques will provide the answers. We use the word “should” because no project can ever anticipate the full Research design range of possibilities that will eventuate, and as new problems may spring up What are you trying to in the course of the analysis, these may require a different set of techniques or find out? even the development of novel methods.

How do you build your The questions can come from many sources; they may spring from the questions? imagination fully formed, or coalesce gradually as you digest the disciplinary literature and analyze its strengths and weaknesses. Good questions stand to shed new light on important issues in archaeology and can be answered through empirical research (i.e., stone artifact analysis) that can be undertaken within the time frame available (Odell 2001a).

Are some analyses more This depends entirely on whether a good match exists between the questions meaningful than others? that you are setting out to answer and the methods and data used to address them. Beyond this, there is no “right” way to analyze stone that will guarantee more meaningful results. The philosophical position that is taken, however, often leads us to choose various forms of analysis over others for the particular advantages that they offer. The following section on classification provides an example of one such situation in which our underlying ontological positions, or “views of reality,” may influence the sorts of data we collect and the types of classifications we employ.

Classifying an Classification in archaeology, as in all fields, really only serves two purposes. assemblage of The first is to structure our observations into a limited set of groupings that stone artifacts can be said to be alike in a defined way. Grouping our observations in this way allows our results to be compared, contrasted, and explained. The second Why classify? purpose is to provide a set of terminological conventions, usually a set of

176 CHRIS CLARKSON AND SUE O’CONNOR named groupings or “classes,” that allows us to communicate about the world in a simpliﬁed and understandable fashion (Lyman et al. 1997: 15).

There are three basic rules on which successful classifications are based. The Are there rules of first is that classifications should be based on sets of variables whose import- classification? ance and means of combination is somehow determined from a body of theory. The second is that there must be recognizable similarities and differences between the phenomena being observed in relation to the variables on which the classification is based (Hill & Evans 1972; Dunnell 1986; Bailey 1994: 232). The third rule is that the classification must be exhaustive, or in other words, it must encompass all of the observed variation. Many classifications fail on these three counts, and particularly in the case of exhaustiveness. For instance, many classifications adopt the use of “miscellaneous” categories in which to place specimens that do not meet any of the classificatory criteria, rather than revising the classification to include unique objects. Obviously, the variables employed in a classification, as well as their means of combination, are of prime importance in determining its utility for a particular research design, its comparability to alternative systems, its sensitivity to variation, and its suffi- ciency as an exhaustive and unambiguous description of variation.

There are numerous forms of classification, ranging from ad hoc folk classi- What are the different fication to systematic forms, and a potentially infinite range of variables upon types of classification? which to base any system of division. We can usefully distinguish three elements of classification, all of which find their way into lithic classification to some degree. A first principle relates to the criteria used to assign objects to a particular class, and it is possible to differentiate between monothetic and polythetic class construction. In monothetic class construction, objects belong to a certain class only if they possess all of the specified attributes (or properties) that define that class (see Figure 6.8 below). The implication of this type of classification is that an object can be assigned to a particular class according to the presence of any single attribute, because it is assumed that if it possesses one, it must possess them all. Polythetic classification, on the other hand, is better suited to dealing with variation in that it requires that an artifact possess only one or more of the total number of defining properties to belong to a class, and that no artifacts possess all of them. The implication of this form of classification is that a single property does not always provide an accurate basis on which to assign an artifact to its proper class, and classification must instead take into account the total combination of attributes and their overall weighting in the system. Polythetic classifications require explicit definition of each defining property, so that different analysts do not accidentally produce different classifications. To give an example of the kinds of classes that each system might produce, as well as the ways in which properties can be combined to form distinct

AN INTRODUCTION TO STONE ARTIFACT ANALYSIS 177 Retouch properties Angle Edge shape Feature Location Surface Ventral

Specimen Distal Low High Nosed Distal point Distal and proximal Dorsal only Ventral only Bifacial Pronounced bulb Bulbar trimming Straight edge Concave edge Convex edge Notched Lateral Distal and lateral High-angled concave and nosed end scraper 1 X X X X 2 X X X X 1 3 X X X X Low-angled convex and notched side and end scraper

4 X X X X MONOTHETIC 5 X X X X 2 6 X X X X Point 7 X XXXX 8 X X XX X 3 9 X XX X X X Tula 10 X XXX X X POLYTHETIC 11 XX X X X X 4 12 XXX X X X X

Retouched ﬂakes

Figure 6.7 An illustration of some typical classifying variables and their means of combination under monothetic or polythetic classiﬁcation.

classes, two monothetic and two polythetic classes are shown in Figure 6.7. The first three specimens are assigned to the monothetic class “High-angled concave and nosed-end scrapers,” on the basis that all members possess high edge angles, concave edges, nosed projections, and distal retouch. In this system, the presence of any one of those features will identify the artifact as belonging to that class, as they are mutually exclusive and do not occur in any other monothetically defined class. The second class is also a monothetic class with a different set of attributes that are also held in common by all its members. In contrast, the next two classes are constructed using a polythetic system in which not all properties are held in common by all members, although at least one property is held in common by all (e.g., a distal point for Points, and both pronounced bulbs and dorsal bulbar trimming for Tulas). Each system has its own strengths and weaknesses. For instance, monothetic classifications are simple and straightforward to design and implement, but suffer difficulties when dealing with variation and complexity. It must be

178 CHRIS CLARKSON AND SUE O’CONNOR acknowledged that monothetic classes are also high-level abstractions in the sense that they impose rigid boundaries around phenomena that may in fact form a continuum. Polythetic classifications are better able to deal with variation, but may embody too much flexibility. Unless the various defining properties are rigorously defined and weighted, there is great potential for each researcher to come up with a different set of assignments. A second principle of classification is that objects can either belong exclusively to a certain class and no other, or they can belong to many classes, sometimes with a “membership weighting.” This division again determines how variation is dealt with. Exclusivity means that variation is suppressed to fit unique objects into a limited set of classes, as is the case for Classes 1 and 2 in Figure 6.7. Alternatively, overlapping classifications mean that variation is allowed expression, and that unique objects are recognized for their potential to fit into any number of classes, depending on which attributes are given prominence in the classificatory scheme. For example, Classes 3 and 4 in Fig- ure 6.7 share features in common with specimens found in other classes, and we could theoretically assign them to all of the classes with which they overlap – if we were to find this useful for some purpose. By placing a weighting on specific variables, however, it would be equally possible to narrow the range of classes into which they fit. Tinkering with the choice and weighting of variables allows the degree of overlap between each type to be expanded or narrowed. A third principle centers on whether classes have some sort of structure imposed on them, such as a hierarchical arrangement of the sort seen in Linnean biological classification. Unordered classifications impose no precedence or structure and treat each class is as though it is “on the same level.” In stone artifact classification, hierarchical classification is best seen in classifications that attempt to order each stage of the reduction process in terms of the chronological sequence in which it takes place (Hiscock 2005). For instance, all of the specimens in Figure 6.7 also belong to the higher-order grouping “Retouched Flakes.” Andrefsky (1998: 65) provides an example of the way in which either monothetic and polythetic methods could be used in the construction of a hierarchical classificatory system.

Classifications can be undertaken by manually allocating objects to a class What are the different using a set of variables whose importance is deduced from theory, or using methods of classification? statistical techniques that find clusters within the data. It has sometimes been claimed that these techniques can provide an objective means of “discovering” natural types (Spaulding 1953), but as Dunnell (1971) points out, while statistical techniques may indeed derive attribute clusters from empirical data that are of utility for certain problems, they cannot discover types with an independent reality, as the robustness of class divisions ultimately rests on the value and weighting of the attributes employed. Statistical types are therefore always constructed at some level.

AN INTRODUCTION TO STONE ARTIFACT ANALYSIS 179 How do you choose Choosing between classificatory systems is not straightforward, but depends between classificatory systems? on the sorts of data that you want to collect and the types of questions being addressed. Our underlying views of reality may also sway our decision to use one form of classification over another, as these tend to favor a certain depictions of the world over another. Most classifications in use today are built around one or other of two alternative views of the world that have important consequences for the way things are classified. The first was discussed at least as long ago as Plato’s time and is today called “essentialism.” This idea holds that the world is divided into real, discontinuous, and immutable “kinds.” This notion underlies most typological constructions, which hold that artifacts, particularly retouched implements and certain types of flakes and cores, can be separated into discrete and mutually exclusive kinds (Dunnell 1986; Dibble 1995; O’Brien 1996; Lyman et al. 1997; Hiscock 2002b, 2005). In the context of stone artifact manufacture, essential forms are often thought of as “mental templates,” or combinations of traits that are favored by the maker. Variation is seen as a consequence of the imperfect realization of the conceptually perfect form, and is usually attributed to differences in raw material properties or individual skill levels (Dunnell 1986). In practice, individual artifacts are usually assigned through comparison with the “type specimens,” or sets of artifacts that exem- plify the ideal forms for each class. The essentialist metaphysic lends itself to the use of mutually exclusive, unordered, monothetic classes of the sort typically employed in most typologies. An alternative view of reality is called “materialism,” and holds that all phenomena are unique, often arranged as continuums, and that “kinds” are illusory and imposed on reality rather than extracted from it. Materialist classifications therefore set out to find ways of depicting variation as well as central tendency, and treat observational units as units of measurement rather than real kinds. In archaeological classification, the materialist metaphysic has been particularly embraced by evolutionary and processual schools. The processual school has argued the position that there is no natural, single, or “best” typology and no inherent meanings to be discovered in an assemblage of artifacts (Hill & Evans 1972). Rather, the meaning imposed on archaeological phenomena derives from a priori problems, hypotheses, and other interests (Hill & Evans 1972: 252). Hence processualists encourage the selection of attributes that are derived from the discipline’s problems and that will lead to classifications that are useful in addressing those problems. Evolutionary archaeologists make the additional claim that most phenomena are in a state of constant change (as in cultural phenomena and artifacts themselves), and that classifications may be enhanced by somehow factoring time, distance, and/or historical relatedness into their formulations (Lyman & O’Brien 2000). This can be clearly seen, for instance, in the changes that take place in the form of an artifact as reduction continues (Dibble 1995; Hiscock & Attenbrow 2002, 2003; Clarkson 2005).

180 CHRIS CLARKSON AND SUE O’CONNOR By this stage, we hope we have convinced you that classification requires How do you build your some thought, that no classification is “real” or fixed, and that it is most useful own classification? if approached as a tool for measurement, description, and problem-solving. Important points as far as stone artifacts are concerned are that different levels of classification can exist, that the same artifacts can be assigned to different classes according to the weighting and combination of variables used, and that all classifications will create a certain level of abstraction and ambiguity, but that this can be reduced by being explicit about the choice of variables and the weightings given to each. This also increases the ease with which each classification can be replicated by other researchers (Andrefsky 1998: 62). Building your own problem-oriented classifications therefore requires attention to these factors. Once they have been dealt with, however, virtually any set of groupings based on a potentially infinite range of variables is conceivable. Another approach that we would advocate in classifying stone artifacts is that some categories should always remain exclusive, whereas others might be allowed to overlap, as in the case of the three chronologically separate and mutually exclusive categories of cores, flakes, and retouched flakes. Classifica- tions that set out to describe and order manufacturing processes and/or products should generally seek to keep these classes distinct. Unfortunately, many classifications blur these classes, and draw an initial division within assemblages between “tools” and “debitage” (e.g., Andrefsky 1998). “Tools” are all those artifacts believed to represent the intended “end-products” of the process, while “debitage” constitutes all the waste left over from tool production, use, and maintenance (Dibble 1995; Hiscock 2005). These divisions are based on propositions that cannot be verified empirically, and it is therefore safest when building classifications to start with basic observational categories, and if other higher-level categories are required, to build on them as required.

The selection of variables to record and measure in an analysis is clearly one of Choosing attributes the most important decisions that you will make. As Hiscock and Clarkson to record and (2000) state: measure

What attributes should . . . the most crucial consideration must be the analytical power of the attribute you choose? and its relevance to the questions posed . . . [T]he application of a single stand- ardized method of analysis, including the use of a standard set of attributes, is not an appropriate response because different observations will be needed for each new question and in each archaeological context. However, for any particular question there may be a number of relevant attributes, and it is valuable to also consider the power of equally relevant variables.

Our advice for choosing the most powerful attributes is to read widely within the technological literature and identify attributes that help address the questions you have posed. Compiling a table of justiﬁcations and references to

AN INTRODUCTION TO STONE ARTIFACT ANALYSIS 181 successful uses of each attribute can also provide a useful starting point, as in Table 6.1. Phagan (1985) provides a fairly extensive list of attributes (and some justiﬁcation for each) commonly employed in the analysis of ﬂakes and cores, as does Clarkson and David (1995). Attributes relevant to the recording of retouch are detailed by Hiscock and Attenbrow (2002, 2003, 2005), Clarkson (2002a, 2005), Hiscock and Clarkson (2005), and Dibble (1995).

What is a “basic” In the broader scheme of things, a basic analysis, if such a thing could be said analysis? to exist, would probably try to incorporate some description (whether quantitative or qualitative) of the size, shape, level of reduction, raw material, and technological and typological category for each artifact in an assemblage, in the hope that the broadest range of questions possible might be addressed. It may be possible to cover each of these aspects and yet still record only a small number of variables. As stated above, it will be the power of each attribute to address each criterion that will determine how streamlined the analysis can be.

Managing data Because laboratory analysis is slow and painstaking, there is a good argument to be made for reducing data handling time by entering information straight How should you record into the computer as it is gathered. Although some archaeologists still prefer your attributes? to use spreadsheets, there is no doubt that a database provides a far superior means of entering, storing, and retrieving data about individual specimens. Computer data entry may not be practical in some ﬁeld situations, where the use of recording forms may still be the most suitable option.

When do you need to use Statistics are typically used in stone analysis to provide a means of seeking statistics and what independent confirmation that the patterns observed in the data are not sim- statistics are most useful? ply a result of the vagaries of sampling (i.e., random effects), small sample size, or the result of a complex interaction of several variables that makes the important variables or patterns difficult to determine. It is difficult to advise what statistics should be used to analyze the data, as each question and analysis lends itself to different techniques and tests. Never- theless, several tests tend to be used over and again in lithic analysis. These include chi-square tests, t-tests, Spearman’s rho, and regression analysis. These are all basic techniques for working with the kinds of data that archaeologists use in lithic analysis. A good introduction to all of these techniques, as well as a demystification of concepts such as significance and sampling, and useful suggestions for identifying and working with skewed populations (which most lithic assemblages tend to be), is provided for archaeologists by Drennan (1996). These basic tests are all that is typically needed to make comparisons between assemblages, confirm ordering in the data, and determine whether a relationship between two variables is strong and significant. Statisticians can be helpful in identifying the techniques that best address your questions, and in navigating and interpreting the complex world of

182 CHRIS CLARKSON AND SUE O’CONNOR multivariate statistics, but you should be capable of performing most simple tests yourself with the help of statistics software.

As stoneworking is a reductive technology, the measurement of the degree Measuring extent of to which this process has progressed often forms the basis of many modern reduction analyses. Quantifying the extent of reduction allows estimations to be made of the amount of time and energy invested in the production of an artifact, Why measure reduction? the level of departure of the observed form from its original form, the amount of material likely to have been created as a product of the process, and the position in the sequence at which changes in manufacturing strategies took place and their likely effects on artifact morphology. At a higher interpretive level, many archaeologists see measures of reduction as critical to the testing of behavioral models that hypothesize the place of stone artifacts in broader systems of time budgeting, mobility, and land use. Consequently, measures of reduction have come to be associated, at least implicitly, with discussions of risk, cost, and efﬁciency in past technological systems (Bleed 2001). These discussions build on the assumption that the differential distribution of sequential steps and stages through space and time will reﬂect aspects of planning, land use, ecology, and settlement and subsistence patterns affecting people’s daily lives (Nelson 1991; Kuhn 1995). Measures of reduction are consequently fast becoming a central component of lithic analysis, and also form the basis of the European chaîne opératoire approach (Leroi-Gourhan 1964; Meignen 1988; Roebroeks et al. 1988; Geneste 1990). This approach places emphasis on the technical “choices” that people make during lithic reduction between a variety of possible solutions and the context of these decisions within broader cultural values and social relations (Lemonnier 1986; Dobres 2000).

Both fracture mechanics and basic engineering principles would suggest How do you measure core that striking more and more mass from a core will affect its size and geometry, reduction? which will have direct consequences for the nature of force input, the viability of different reduction strategies, and the size and morphology of the flakes produced over the sequence. We can speculate, for instance, that the gradual reduction of cores will result in more flake scars and less cortex, that continued use of a platform will result in a decrease in platform size, and that as more mass is struck from a core, the size of the core and resulting flakes might also decrease. If cores are rotated during this process to create fresh platforms once old ones become damaged or unproductive, cores should begin to preserve signs of former flaking on the platform surfaces as well as indications of the existence of old platforms. To provide an example of the sorts of procedures that can be used to track morphological changes and the use of different technological strategies over the sequence of core reduction, a number of variables are plotted against increasing numbers of core rotations in Figure 6.8. These changes are

AN INTRODUCTION TO STONE ARTIFACT ANALYSIS 183 80 60 40 Number of scars 20 0 90 80 60 % multi-scarred platforms 40 20 0 60

50 % step/hinge 40

30 100

90 Final platform angle

80 80

70 % overhang removal 60

50 60

40 % cortex 20

0 100

75 % platform perimeter used 50

25 40 30 20 Number of cores 10 0

0 1 2 3 ≥4 Number of rotations 184 CHRIS CLARKSON AND SUE O’CONNOR Figure 6.8 Changes in core morphology over the sequence of reduction. documented from a set of 87 small, locally occurring, river-rounded chert cobbles found at a site near Wollongong on the southern coast of New South Wales, Australia. This diagram shows that many core characteristics show an increase over the sequence of reduction, while others decrease. For instance, as might be expected, the number of scars found on cores increases with each rotation, as does the percentage of platforms that have more than one conchoidal scar (resulting from former use as a core face). The percentage of scars found on the core showing step and hinge terminations also increases as core rotation proceeds, as does the external angle of the last platform used on the core. Overhang removal increases early on and remains high throughout the remainder of the reduction sequence. Overhang removal was presumably used to strengthen the platform to better receive the forceful blows required to remove flakes from small cores with increasingly high- angled platform edges. In contrast, cortex diminishes at a fairly consistent rate throughout the sequence, indicating that similar amounts of material were likely removed from each platform with each rotation. The used portion of the platform edge first increases and then decreases, as irregularities left on the core face and platform by previous rotations reduce the usable platform perimeter. The number of cores from each stage of reduction also clearly indicates that most cores were abandoned in early stages of reduction, although a small proportion continued into later stages of reduction, by which time cores were heavily rotated and generally lacking cortex, or were subjected to bipolar reduction. Changes in core morphology over the reduction sequence are illustrated as a reduction flowchart in Figure 6.9, which depicts a number of the ways of flaking small spherical nodules followed at the site. While archaeologists have sometimes used this type of chart to illustrate normative reduction sequences through which most forms are argued to pass, this chart ascribes frequencies to each stage in each sequence as determined from the assemblage itself. Reduction begins with a single flake removed from a cortical platform. In the left-hand sequence (Sequence 1), new platforms are always created from the previous flaked surface via 90 degree core rotations. In the middle sequence (Sequence 2), new platforms are always created from cortical surfaces. In the right-hand sequence (Sequence 3), a single large scar is removed from each surface, which then becomes the platform for the next single large flake removal. Also illustrated in Figure 6.9 are late-stage rotated and bipolar cores with and without cortex, which represent the very end stages of all sequences. From this diagram, it can be seen that Sequence 1 was most commonly followed at the site, but that Sequences 2 and 3 also formed common alternatives. The mapping of reduction sequences in this way allows variation as well as the central tendency to be explored, and also demonstrates that core reduction was a highly variable process, with knappers responding to the results of each successful or unsuccessful blow in a flexible fashion, in which the options

AN INTRODUCTION TO STONE ARTIFACT ANALYSIS 185 5%

10%

32% 5% 3%

10% 4% 2%

8% 1% 1%

2% 1% 1%

Late stage with cortex Late stage, no cortex Bipolar, no cortex Bipolar with cortex

2% 10% 2% 1%

Figure 6.9 The reduction ﬂowchart and the frequencies with which various core reduction sequences were employed at Sandon Point.

186 CHRIS CLARKSON AND SUE O’CONNOR for rotation, discard, or strategy switching (such as to bipolar reduction) were appraised at various points along the way. This simple case study indicates that a number of variables are likely to be useful measures of reduction intensity, at least for this type of core reduction. Used in combination, these attributes would provide a reasonably sound basis on which to infer the level of reduction reached by any individual core. This could be done by dividing the continuum into a number of intervals (such as early, medium, and late), or by using a continuous ranking system. Assigning each core its own degree of reduction then allows the intensity of core reduction to be traced across space and time. It should be kept in mind, however, that different forms of reduction might well require the use of different variables than those employed here.

Lithic analysts employ various means of assigning flakes from archaeological How do you measure flake assemblages into reduction stages, but most of these tend to involve comparison reduction? of archaeological specimens with experimentally produced assemblages. To avoid this analogical approach, our case study ranks flake reduction according to simple and universal changes in flake morphology that are deduced from the analysis of changing core morphology presented above, as reflected in dorsal and platform scar morphology. This type of analysis is called diacritical analysis (Sellet 1993), and aids in the construction of hypothetical reduction models. In this case, diacritical analysis allows changes in flake morphology to be examined for Sequence 1 of the pebble core reduction sequence illustrated in Figure 6.9. The reduction process can be modeled by examining stages in flake scar superimposition on the platform and dorsal surfaces of flakes and the stages of decortication present. Nine stages of flake production can be envisaged. As in Figure 6.9, the first phase involves the creation of an initial flake scar on the core to serve as a platform for the next stage of reduction. This results in the production of Stage 1 flakes that possess primary cortex (i.e., 100 percent) on all surfaces. The second stage involves the rotation of the core through 90 degrees, so that flakes can be struck from the first scar. These Stage 2 flakes will possess a single conchoidal scar on the platform and primary cortex on the dorsal surface. Stage 3 flakes result from continued reduction of this second face, and will have single conchoidal platforms but only secondary cortex (<100 percent) remaining on the dorsal surface. Stage 4 flakes result from the final stages of reduction on this face, and will possess the same type of platform, but will have no cortex remaining on the dorsal surface (tertiary cortex). At some point, reduction is likely to end on this second face as nonfeather terminations increase, or platforms become unproductive due to high platform angles. This will either result in discarding the core, or the formation of another platform by rotating the core again. Redirection flakes, or flakes that remove the edge of an old platform, are sometimes created by this process, and these are here labeled Stage 5 flakes. At this early stage of reduction,

AN INTRODUCTION TO STONE ARTIFACT ANALYSIS 187 redirection flakes should preserve cortex on one or more of their surfaces (and this is the criteria used for determining Stage 5 redirection flakes). The process of reduction then begins anew on the third face, with cortical flakes with multiple scarred platforms (as a result of striking from previously flaked surfaces) produced first (Stage 6), followed by flakes with secondary cortex (Stage 7), and then by flakes with no cortex (Stage 8). Once this face also begins to encounter difficulties for further reduction, the core may be discarded or rotated again, and another redirection flake may be produced. As cores enter increasingly later stages of reduction, cortex is likely to have been entirely removed from all surfaces, and hence redirection flakes from this stage would show no cortex on their surfaces. These flakes are assigned to Stage 9 (or 3 to N rotations). Although not included in this hypothetical reduction model, another option for knappers is to switch to the use of a bipolar reduction technique once cores become too small to continue freehand percussion, and this was frequently undertaken at the site. Figure 6.10 maps out the sorts of changes in flake characteristics that accompany each of the stages of Sequence 1 reduction as deduced from dorsal and platform scar patterns. Interestingly, these changes are largely cyclical, with the gradual increases or decreases in characteristics taking place throughout the first phase of flake removals (Stages 2–4) often repeated in the second phase (Stages 6–8). The basic series of changes is as follows: flakes are at first rather squat but become increasingly elongate toward the end of each phase as parallel ridges become more common, platform area decreases, platform preparation becomes common, the proportion of nonfeather terminations increases, and the size of flakes measured by width and thickness decreases (weight does not show a sequential decrease due to differences in initial nodule size). Redirecting flakes stand out in terms of their larger size and their apparent nonconformity to the trends otherwise seen for most characteristics. This is not surprising, since striking off old platforms often requires delivering large amounts of force to the core from a sub-optimal platform. Hence, striking off old platforms is unlike other forms of flaking, and the resulting flakes are often distinctive. The last line shows the frequency with which flakes at each stage of reduction are found in the assemblage, and indicates that the greatest proportion of flakes belong to early stages of reduction. This is consistent with the pattern seen for cores, which were rarely taken into later stages of reduction.

How can you explore Archaeologists are often interested in the process of blank selection – or the blank selection? selection of a subset of ﬂakes from the total population produced at a site – in which prehistoric knappers selected appropriate ﬂakes for further use, retouching, and transport away from the site. Blank selection is of interest as it has the potential to inform us about design considerations (such as tool performance, reliability/maintainability, suitability to prehension and hafting, and multifunctionality), a range of environmental and cultural constraints

188 CHRIS CLARKSON AND SUE O’CONNOR 80 60 % nonfeather 40 Terminations 20 0 60

% elongate 40 flakes 20 0 80 60 % overhang 40 removal 20 0 14 12 10 Thickness 8 6 4 33 Key 30 27 Platform creation flake Width 24 21 Early stage 1R reduction 18 15 Mid-stage 1R reduction 160 Late stage 1R reduction 120 Platform area 80 Early stage redirection flake 40 Early stage NR reduction 0 Mid-stage NR reduction 60 Late stage NR reduction 40 Number Late stage redirection flake 20 0 Stage 123456789

Figure 6.10 Changes in ﬂake morphology over Reduction Sequence 1.

(functional, material, technological, socioeconomic, and ideological; Hayden et al. 1996), and the level of standardization in the production system, both in terms of overall flake production and selection from the larger pool of flake variation. An example of one approach to examining the pool of variation in flake forms produced at the same site near Wollongong, and the range of blank shapes selected for various forms of retouching, is shown in Figure 6.11. Here, two measures of flake shape are plotted against one another to illustrate the

AN INTRODUCTION TO STONE ARTIFACT ANALYSIS 189 60

20 Backed artifacts

0 “BLADES” Flakes Retouched ﬂakes −20 Angle of the lateral margins (degrees) −40

−60 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 Elongation

Figure 6.11 A graph illustrating the range of variation in ﬂake shape employed in different forms of retouching.

spread of flake shapes found at the site. Plotted on the y-axis is the angle of the retouched margins expressed in degrees, with 0 indicating parallel-sided margins, positive values indicating contraction of the margins, and negative values indicating expansion of the lateral margins (Clarkson & David 1995). The x-axis plots elongation (length/width), with values ranging from very squat flakes (i.e., values of 1) through to extremely elongate flakes (i.e., values of 4.5). The resulting graph shows a wide spread of flake shapes produced at the site, with the majority proving to be squat flakes of widely varying marginal angles that likely derive from earlier stages of core reduction (see above). Retouched flakes clearly represent a much smaller subset of the total range of flake shapes, while backed artifacts (large and small symmetrics and asymmetrics) (Hiscock 2002a) make up an even narrower range still. Overall, retouched and backed flake shapes mirror the broader pattern of squat flakes with variable marginal angles produced at the site, although backed artifacts tend more frequently to be parallel-sided with contracting margins. The graph also indicates that flake production and blank selection were far from stand- ardized, suggesting that few design considerations affected the types of blanks chosen for further modification. The greater constriction of variation seen in backed artifacts, however, points to tighter constraints on the design of these implements than was the case for retouched artifacts more generally, and these may be related to hafting requirements, functional efficiency, potential for multifunctionality, or all of these factors.

190 CHRIS CLARKSON AND SUE O’CONNOR Retouched flakes are most commonly the subject of detailed lithic analysis, How do you measure but until recently few techniques existed to measure the amount of time and retouch? labor invested in their production. More recently, a number of measures have been proposed and tested that offer a means of measuring reduction for a number of different forms of retouching (Kuhn 1990; Clarkson 2002b; Hiscock & Clarkson 2005). The Index of Invasiveness, for instance, provides a measure of retouch coverage over both the dorsal and ventral surfaces of an artifact that is suited to measuring unifacial and bifacial retouch (Clarkson 2002b). It is best suited to the measurement of artifacts that tend to become more invasively retouched over the sequence of reduction. The Index of Invasiveness calculates intensity of retouch as a value between 0 and 1 by estimating the extent of retouching around the perimeter of a flake as well as the degree to which it encroaches onto the dorsal and ventral surfaces (for procedures, see Clarkson 2002b). Another measure, the Geometric Index of Unifacial Reduction (GIUR), was developed by Kuhn (1990) and calculates edge attrition as the ratio of the height of retouch to the maximum thickness of the flake. This technique is designed for the measurement of unifacial retouch, as the name suggests, but is also best suited to the measurement of steep and marginal retouch. Again, a score of 0 indicates no retouch, while a score of 1 indicates that retouch height is equal to flake thickness. Used in conjunction, these two techniques are capable of describing almost any form of retouching, and of quantifying the degree to which retouch in steep and marginal or acute and invasive. Although retouched flakes (commonly termed “scrapers”) have often been treated as stylistically irregular artifacts shaped simply to meet immediate needs (Hayden 1979; White & O’Connell 1982), recent studies (Clarkson 2002a, 2005; Hiscock & Attenbrow 2002, 2003, 2005) have demonstrated that this group of implements can display marked internal consistency when examined in light of increasing retouch intensity, with a regular series of changes noted to the shape, extent, and type of retouch found on their margins as retouch increases. These changes can be depicted using a number of indices of retouch extent, shape, and type. These are the percentage of the perimeter of an artifact that is retouched, the curvature of the retouched edge, the angle of retouch, and the invasiveness of retouch. The percentage perimeter of retouch is calculated by dividing length of retouch by the perimeter of the flake. The angle of the retouched edge is calculated as the mean of several edge angle measurements taken at regular intervals along the retouched edge. Edge curvature is measured by dividing the maximum diameter of retouch by the total depth of retouch (Clarkson 2002a). Negative results indicate concave edges, while positive ones indicate convex edges. An example of the power of these measures of flake reduction is illustrated in Figure 6.12 by plotting the morphological changes that occur in a population of 128 retouched flakes from the same site near Wollongong as retouching

AN INTRODUCTION TO STONE ARTIFACT ANALYSIS 191 50 100 (a) (b) 80

40 60

40 30 % margin retouched % margin retouched 20

20 0 0.25 0.50 0.75 1.00 –0.4 –0.2 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 Kuhn Index Index of Edge Curvature

90 0.3 (c) (d) 80 70 0.2 60 50 40 0.1

30 Index of Invasiveness Mean edge angle (degrees) 20 10 0.0 0.25 0.50 0.75 1.00 0.25 0.50 0.75 1.00 Kuhn Index Kuhn Index

Figure 6.12 Changes in ﬂake morphology as retouching increases: (a) the percentage of the margin retouched; (b) the edge curvature; (c) the mean retouched edge angle; (d) the Index of Invasiveness.

increases. Figure 6.12a, for instance, indicates that retouching usually starts out covering only a short section of the lateral margins, but as Kuhn’s GIUR index increases, eventually extends to cover around 50 percent of the margin. As retouch spreads around the margin, the curvature of the edge also changes, as shown in Figure 6.12b. Retouch usually starts out slightly concave, but then straightens out before finally becoming quite convex. Similarly, Figures 6.12c and 6.12d indicate that retouch generally starts out quite low-angled and marginal, but ends up steep and more invasive. This sequence of changes accounts for much of the variation in retouched flake morphology, and underlies the differences in form that have sometimes been formalized as elaborate scraper typologies (such as that of McCarthy et al. 1946). Much like the situation for cores argued above, once these changes in flake form are documented and the reduction sequence understood, individual artifacts can be assigned their positions in a particular reduction sequence. An understanding of reduction sequences also enables the construction of classifications that divide the continuum in sensible places, rather than jumbling together artifacts on the basis of features that are disconnected from the mechanisms actually creating that variation. The resulting classes could be treated as

192 CHRIS CLARKSON AND SUE O’CONNOR measurement units (in this case of intensity of reduction) that expressed temporal and historical relationships (in this case departure from an original form), thereby meeting some of the expectations of materialist classiﬁcatory systems. Other studies have also explored stage and continuum models in artifact reduction, and some of these employ quite different measures of reduction intensity to those employed here (Barton 1988; Gordon 1993; Marcy 1993; Neeley & Barton 1994; Dibble 1995; Holdaway et al. 1996; Morrow 1997b; Yvorra 2000; Lamb 2005; Law 2005).

This can be a big problem for any analyst working on assemblages that have Dealing with been subjected to intense heating and trampling, extensive reduction and use, difficult or some other post-depositional process. However, even broken and damaged assemblages artifacts can preserve information relevant to the reconstruction of the overall What if most of your manufacturing system, and assemblage attrition can also reveal much about artifacts are broken or site formation, disturbance, and occupational intensity (Flenniken & Haggarty damaged? 1979; Mallouf 1982; Hall & Love 1985; Hiscock 1985; Jung 1992). Figure 6.13 illustrates a range of fragment types that are commonly found in flaked stone assemblages. Different kinds of technological information can be recovered from each fragment type. For instance, transversely broken fragments such as proximal pieces preserve information about the platform surface, the presence and type of platform preparation, and the platform angle, while distal pieces preserve information about the flake termination type. Medial pieces can be informative about the cross-section of the flake, while all of these fragments may reveal something of the original width and thickness of the flake, its dorsal scar morphology, or the presence, type, and amount of cortex. Longitudinally split flakes may preserve original length and thickness as well as some features of the platform and termination. Once transverse and longitudinal fractures occur together on the same artifact, however, information loss increases dramatically. Surface fragments and flaked pieces often yield little technological information at all. Combining the relevant information from each type of fragment with that gained from complete flakes may be a useful way of increasing the number of observations if assemblages are highly fragmented, especially for categories that suffer high breakage rates and may be underrepresented in the assemblage (such as very thin artifacts or retouched flakes). The recovery of information from broken artifacts is often a necessary means of increasing sample size (see “What if you only have a small number of stone artifacts?”). One problem that results from fragmentation can be to accurately determine how much flaking took place at a site. In cases where, for instance, an assemblage contains only a few heavy artifacts, or where post-depositional processes such as burning and trampling have caused severe artifact fragmentation, simple measures such as weight or number of artifacts can be unreli- able. Hiscock (2002b) has recently explored this problem and suggested some

AN INTRODUCTION TO STONE ARTIFACT ANALYSIS 193 Longitudinal fragments Left Right

Marginal Surface (pot lid)

Transverse and Transverse and longitudinal Transverse longitudinal

Left proximal Proximal Right proximal

Left medial Medial Right medial

Left distal Distal Right distal

Figure 6.13 An illustration of the range of ﬂake fragment types found in many assemblages.

counting procedures that may assist in the quantification of numbers of flaking events and in assessing the effects of breakage and weathering. Hiscock’s technique allows a number of different indices to be calculated, including the Minimum Number of Flakes (MNF) present in an assemblage. This is derived by dividing the assemblage into raw material types and then adding the number of complete flakes of each raw material to whichever is the greater number of proximal or distal fragments, the greater number of left or right fragments, and the greater number of left or right proximal or distal fragments. MNF, or

194 CHRIS CLARKSON AND SUE O’CONNOR better still MNA (i.e., Minimum Number of Artifacts, including both flakes and cores), provides a superior means of assessing and comparing the intensity of flaking within or between sites than weight of raw numbers. Shott (2000) also offers various techniques (MNT, ETE, and TIE) for estimating the original number of artifacts in an assemblage. To determine how serious a problem fragmentation is, a number of simple techniques exist that provide quantitative assessments of artifact fragmentation. The simplest index is calculated by dividing the number of broken fragments and flaked pieces by the number of complete artifacts. A more sophisticated measure of fragmentation is obtained by dividing the number of broken fragments and flaked pieces by Hiscock’s (2002b) MNF or MNA statistics. A measure of information loss, on the other hand, can be calculated by dividing the total number of fragments from which little or no technological information can be derived (such as surface fragments, marginal fragments, flaked pieces, and longitudinally and transversely broken fragments) by those that provide a great deal more technological information (such as proximal, medial, distal, right, and left fragments). These two techniques can be particularly useful in assessing the severity of attritional processes and their effects on the information potential of a site.

This dilemma is quite common in archaeology and was central to the long- What should you do if you running naturefact/artifact debate, in which claims were made for very old can’t tell artifacts from natural rocks? but dubious artifacts (Schnurrenberger & Bryan 1985; Peacock 1991). Pseudo- artifacts (often called eoliths or geofacts) can result from a host of natural processes, such as “natural soil movements, glaciation, wave action, high velocity water movement, gravity (such as alluvial fans or steep inclines), rapid temperature changes, internal pressure (such as starch fractures and pot lids), exfoliation, tectonic movements, diastrophism, solifluction, foot trampling, and other unintentional activity caused by nature” (Crabtree 1972b: 78). As well as the difficulties faced in differentiating natural fractures from artifacts, certain types of stone, such as vein quartz, fracture in such a way that identification of artifacts becomes difficult. Weathering of artifact surfaces can also have severe effects on assemblages and can obscure or obliterate the diagnostic features of stone artifacts (Hiscock 2002b: 251). In cases in which natural fracture is common, or stone type or weathering renders artifacts difficult to distinguish from nonartifactual rocks, it may be helpful to try to determine the presence or absence of the fracture features identified above for each specimen, as well as those for grinding or pecking. Peacock’s (1991) analysis of natural and artifactual stone assemblages found many of these features to be reliable indicators of human manufacture. The compilation of lists of fracture features for each piece will enable the ranking of specimens in degrees of certainty (i.e., number of features present), with the aim of rejecting those lower in the rank order and accepting those higher up. Additionally, certain fracture features tend to occur more commonly among

AN INTRODUCTION TO STONE ARTIFACT ANALYSIS 195 naturally fractured pieces (due to the type and magnitude of forces) than among humanly derived assemblages (such as obtuse platform angles, edge rounding, and micro-edge and -ridge fracturing) (see Schnurrenberger & Bryan 1985, ﬁg. 5.1, table 5.1). The context of ﬁnds also provides a key to whether or not artifacts would be expected in a given location (e.g., rockshelters and middens versus graded tracks and garden beds).

What if you only have a Sample size can have a profound effect on the sorts of information retrieved small number of stone from an assemblage, and sampling is therefore a key issue in understanding artifacts? stone artifact assemblage patterning through time and space. The probability of drawing a representative sample of the original population drops dramatically as sample size decreases, and this is particularly true in lithic assemblages, where many objects of interest, such as retouched ﬂakes, tend to be rare (i.e., less than 5 percent of the assemblage) and therefore unlikely to turn up in small samples. The diversity of an assemblage – or the number of different elements found in the assemblage – is also sensitive to assemblage size, as the number of assemblage elements present cannot exceed the number of artifacts in the assemblage, and hence small assemblages will always have low diversity. It is therefore generally worthwhile enlarging the sample at the expense of other components of the research (e.g., number of attributes recorded) in order to obtain a representative sample of the parent assemblage.

How can you overcome Unfortunately, there is no magic minimum number that will always over- sample size effects? come sample size effects, as every assemblage or region is likely to be different. The only general rule as far as stone artifacts are concerned is likely to be “more is better” (within practical limits, of course). It is possible to determine something like a minimum sample size for each assemblage, however, by studying the effects of increasing sample size by repeatedly drawing random samples from a large assemblage and examining the variation and deviations in percentages of classes, mean characteristics, and so on between each sample and the parent assemblage. Of course, this can defeat the purpose, since a large sample must already have been collected in order to perform such a test, although it may be helpful in determining minimum sample sizes that should ideally be obtained from sites within a region or survey area. A number of assemblage descriptions are immune to sample size effects. Assemblage richness, for example, provides a measure of assemblage composition that is not affected by sample size, and mathematically determines the number of classes that are expected to occur for a given sample size (Leonard & Jones 1989). Assemblage richness is calculated by dividing assemblage diversity, however measured (e.g., raw materials, or classes), by sample size. Differ- ences in assemblage richness between populations can be determined by plotting diversity against sample size (using a scatterplot) and comparing the gradients of slope for each population (higher gradients mean richer assemblages). Rich- ness therefore provides a more robust comparative measure of assemblage

196 CHRIS CLARKSON AND SUE O’CONNOR complexity than does diversity. Assemblage richness also has real interpretive applications. For instance, archaeologists often interpret differences in the richness of raw materials and assemblage elements as a reflection of the range and diversity of activities carried out at a site (Binford 1979; Shott 1989). Theoretically, it is possible to determine how large a sample is needed before all classes recognized in an analysis will be present in an assemblage (or even a number of each class), by increasing sample size until the line of best fit flattens out, such that further increases in assemblage size no longer yield new classes. Obviously, the number of classes used in the analysis will affect the point at which this takes place. Such a technique could be used, say, to identify the ideal sample size to be recovered from a site threatened with destruction.

Archaeology increasingly draws on specialist information from ﬁelds outside When do you need of its own core knowledge and practice, and the incorporation of such tech- specialist nical analyses (usually involving measurement), particularly from the physical archaeometric sciences – often called archaeometry – can play a vital role in stone artifact analyses? analyses. Common examples of applications include the use of geochemical or elemental analyses to identify stone types, their age, origins, and alterations to the composition of artifacts due to time, weathering, and heat.

While the range of materials used to make stone artifacts often tends to be Determining the fairly limited (depending on local geology), the ability to differentiate reliably type and flaking between materials depends on a combination of expertise and experience, properties of stone and archaeologists should solicit the help of geologists for identification of all How do you identify but the most common types. A useful first step in identifying the range of raw different raw material materials in an assemblage is to consult a geological map of the region. Unfor- types? tunately, geological maps are rarely drawn at the scale that archaeologists desire, and usually do not pinpoint small sources of flakable stone. A second step might involve sorting artifacts into broad material categories (based on texture, grain size, and color). Without significant experience in stone identification, a next step would be to take a sample of the various types to a geolo- gist for identification. Once materials have been identified, it is often useful to know the original size in which they were available and their shape and original cortex, as this may influence the types of reduction strategies used. Often, archaeologists tackle these questions by looking for cores at earlier stages of reduction, to determine the size, shape, and nature of cortex for each raw material type.

The ethnographic literature often points to a range of cultural factors that How do you determine determined the types of stones selected to make various implements, and whether it is good- or poor-quality stone? these were often strongly inﬂuenced by associations with ancestral beings, totemic afﬁliations, or powerful substances (Gould 1968; Jones & White 1988; Taçon 1991). In the absence of information about social and ritual value,

AN INTRODUCTION TO STONE ARTIFACT ANALYSIS 197 however, archaeologists often look to observable measures of the properties of stone that might have inﬂuenced the choices of prehistoric knappers. These might help to quantify the ease with which stone fractures, how well it holds its fracture path, its suitability for particular functions, or its durability. The quality of raw materials has also been given prominence in modeling raw material selection in times and places where greater demands are placed on toolkit performance (Goodyear 1989). Materials testing laboratories offer a number of tests that can be useful in determining how suited certain materials are to conchoidal fracture. These include tests of tensile strength, compressive strength, elasticity, and hardness.

Sourcing stone Sources of stone with good ﬂaking qualities are not evenly distributed in the artifacts landscape and were often so highly valued that they were exchanged over many hundreds of miles. Sourcing stone artifacts from archaeological sites can Why is sourcing therefore provide insights into past exchange networks or interaction spheres, important? changes in territoriality, or changing access to stone resources.

How do you find out which Sourcing analysis is to some extent a misnomer because, as Shackley (1998b: source an artifact 261) points out, “nothing is ever really ‘sourced’. The best we can do is pro- came from? vide a probable fit to known source data.” How good this fit is will depend on how good the source data are, which in turn will rely on the location and sampling of potential sources. If no source localities have been mapped or sampled, archaeologists may have to do the field source sampling of geological outcrops themselves. Luedtke (1992) and Shackley (1998b: 262) outline useful hierarchical procedural guidelines for the archaeologist embarking on an artifact sourcing study. In the past, sourcing proceeded by visual inspection and comparison of stone artifacts and samples of stone from potential sources, either in the form of hand specimens or using microscopic characterization of the structure and texture of ground thin sections mounted on slides. Today, archaeologists routinely use chemical composition analyses such as X-ray fluorescence spectrometry (XRF), neutron activation analysis (NAA or INAA), inductively coupled plasma– atomic emission spectrometry (ICP–AES), and proton-induced X-ray – proton- induced gamma ray emission (PIXE/PIGME). Quantitative statistical analyses are then used to determine the best fit from the results of each analysis. While in some respects this eliminates qualitative assessments and allows vast numbers of specimens to be analyzed, sourcing studies will still only be as good as the field sampling that underpins them.

What is an adequate In order to have conﬁdence in the match of artifact to source, it is necessary to sample? have an adequate sample of reference material from potential source locations as well as an adequate sample of the artifacts from the archaeological assemblage with which to compare against the reference material. In theory, the

198 CHRIS CLARKSON AND SUE O’CONNOR more homogeneous the material, the smaller is the sample size that should be necessary to be confident that a representative sample has been obtained. Conversely, the greater the heterogeneity of a geological outcrop, the larger is the sample size required to obtain an accurate representation of its variability. While obsidian is relatively homogeneous, cherts and other secondary siliceous sediments can be extremely heterogeneous even within a single small outcrop. When undertaking field sampling, it is also necessary to keep in mind that secondary depositional contexts (such as stream beds and moraines) may have been equally important sources of stone as primary geological sources (such as quarries), and that to ignore them in field sampling can result in serious misassignments (Shackley 1998a: 6). These in turn may result in major interpretive errors where stone from an archaeological site is argued to have been traded or exchanged when in fact the raw material could have been derived from a secondary source close to the site.

Material culture, technology, and technological strategies were vital in the Conclusion operation of all cultural and social processes in the past. For much of human history, stone artifacts constitute a large part of the record of what humans accomplished, how they behaved, and how they interacted with one another. Happily, archaeology now seems set on investing a great deal more effort into rethinking and advancing lithic studies. Many old and current debates require the development of new analytical frameworks for their resolution and these are also likely to be aided by new advances in archaeometry. New practitioners in lithic studies will witness and take part in the development of new applications that will engage with disciplinary theory to an unprecedented degree. This chapter has attempted to provide a glimpse of the range of approaches employed in the subject today, and a baseline from which you may begin to explore the diversity of stone artifact analysis for yourselves.

The authors would like to thank Brit Asmussen for early comments on the Acknowledgments structure of this chapter, Peter Hiscock, Sophie Collins, and Oliver Macgregor for their helpful advice on later drafts, and Wal Ambrose for advice on recent advances in archaeometry. Navin Ofﬁcer Heritage Consultants are thanked for access to the assemblages used in our case studies.

Ahler, S. A. 1989: Mass analysis of ﬂaking debris: studying the forest rather than the trees. References In D. O. Henry and G. H. Odell (eds.), Alternative Approaches to Lithic Analysis. Archae- ological Papers of the American Anthropological Association, 1, 85–118. Akerman, K. and Bindon, P. 1995: Dentate and related stone biface points from northern Australia. The Beagle, 12, 89–99. Andrefsky, W. 1998: Lithics – Macroscopic Approaches to Analysis. Cambridge: Cambridge University Press.

AN INTRODUCTION TO STONE ARTIFACT ANALYSIS 199 Austin, R. J. 1997: Technological characterization of lithic waste-flake assemblages: multivariate analysis of experimental and archaeological data. Lithic Technology, 24, 53–68. Bailey, K. D. 1994: Typologies and Taxonomies: An Introduction to Classification Techniques. California: Sage. Bamforth, D. B. and Bleed, P. 1997: Technology, flaked stone technology, and risk. In C. M. Barton and G. A. Clark (eds.), Rediscovering Darwin: Evolutionary Theory in Archaeology. Archaeological Papers of the American Anthropological Association, no. 7. Washington, DC: American Anthropological Association, 109–40. Barton, C. M. 1988: Lithic Variability and Middle Paleolithic Behavior: New Evidence from the Iberian Peninsula. BAR International Series, S408. Oxford: British Archaeological Reports. Baulmer, M. F. and Downum, C. E. 1989: Between micro and macro: a study in the interpretation of small-sized lithic debitage. In D. S. Amick and R. P. Mauldin (eds.), Experiments in Lithic Technology. Oxford: British Archaeological Reports, 110–16. Binford, L. R. 1979: Organizational and formation processes: looking at curated technologies. Journal of Anthropological Research, 35, 255–73. Bleed, P. 1997: Content as variability, result as selection: toward a behavioral definition of technology. In G. A. Clark and C. M. Barton (eds.), Rediscovering Darwin: Evolutionary Theory in Archaeology. Archaeological Papers of the American Anthropological Associ- ation, no. 7. Washington, DC: American Anthropological Association, 95–104. —— 2001: Trees or chains, links or branches: conceptual alternatives for consideration of stone tool production and other sequential activities. Journal of Archaeological Method and Theory, 8(1), 101–27. Bordes, F. 1972: A Tale of Two Caves. New York: Harper and Row. Bradbury, A. P. 1998: The examination of lithic artifacts from an Early Archaic assemblage: strengthening inferences through multiple lines of evidence. Midcontinental Journal of Archaeology, 23, 263–88. —— and Carr, P. J. 1999: Examining stage and continuum models of flake debris analysis: an experimental approach. Journal of Archaeological Science, 26, 105–16. Brantingham, P. J. 2003: A neutral model of stone raw material procurement. American Antiquity, 68(3), 487–509. Bright, J., Ugan, A. and Hunskar, L. 2002: The effect of handling time on subsistence technology. World Archaeology, 34(1), 164–81. Clarkson, C. 2002a: Holocene scraper reduction, technological organization and landuse at Ingaladdi Rockshelter, Northern Australia. Archaeology in Oceania, 37(2), 79–86. —— 2002b: An index of invasiveness for the measurement of unifacial and bifacial retouch: a theoretical, experimental and archaeological verification. Journal of Archaeological Sci- ence, 29, 65–75. —— 2005: Tenuous types: “scraper” reduction continuums in Wardaman Country, northern Australia. In C. Clarkson and L. Lamb (eds.), Rocking the Boat: New Approaches to Stone Artifact Reduction, Use and Classification in Australia. British Archaeological Reports International Monograph Series. Oxford: Archaeopress, forthcoming. —— and David, B. 1995: The antiquity of blades and points revisited: investigating the emergence of systematic blade production south-west of Arnhem Land, northern Aus- tralia. The Artefact, 18, 22–44. Cotterell, B. and Kamminga, J. 1977: The mechanics of flaking. In B. Hayden (ed.), Lithic Use-Wear Analysis. New York: Academic Press, 97–112. —— and —— 1987: The formation of flakes. American Antiquity, 52(4), 675–708. Crabtree, D. E. 1968: Mesoamerican polyhedral cores and prismatic blades. American Anti- quity, 33, 446–78.

200 CHRIS CLARKSON AND SUE O’CONNOR —— 1972a: The cone fracture principle and the manufacture of lithic materials. Tebiwa, 15, 29–42. —— 1972b: An Introduction to Flintworking. Occasional Papers of the Idaho State University Museum. Idaho: Idaho State University Museum. —— and Butler, B. R. 1964: Notes on experiments in flintknapping: 1. Heat treatment of silica materials. Tebiwa, 7(1), 1–6. Dibble, H. 1995: Middle Paleolithic scraper reduction: background, clarification, and review of evidence to date. Journal of Archaeological Method and Theory, 2(4), 299–368. —— 1997: Platform variability and flake morphology: a comparison of experimental and archaeological data and implications for interpreting prehistoric lithic technological strategies. Lithic Technology, 22(2), 150–70. —— and Pelcin, A. 1995: The effect of hammer mass and velocity on flake mass. Journal of Archaeological Science, 22, 429–39. —— and Whittaker, J. 1981: New experimental evidence on the relation between percussion flaking and flake variation. Journal of Archaeological Science, 6, 283–96. Dobres, M.-A. 1995: Gender and prehistoric technology. World Archaeology, 27(1), 25–49. —— 2000: Technology and Social Agency. Oxford: Blackwell. Drennan, R. D. 1996: Statistics for Archaeologists: A Commonsense Approach. New York: Plenum Press. Dunnell, R. C. 1971: Systematics in Prehistory. New York: The Free Press. —— 1986: Methodological issues in Americanist artifact classification. In M. B. Schiffer (ed.), Advances in Archaeological Method and Theory. New York: Academic Press, 35–99. Ericson, J. E. and Earle, T. K. 1982: Contexts for Prehistoric Exchange. New York: Academic Press. Flenniken, J. J. and Haggarty, J. 1979: Trampling as an agency in the formation of edge damage: an experiment in lithic technology. Northwest Anthropological Research Notes, 13, 208–14. —— and White, J. P. 1983: Heat treatment of siliceous rocks and its implications for Australian prehistory. Australian Aboriginal Studies, 1983(1), 43–8. Geneste, J. M. 1990: Dévelopment des systèmes de production lithique au cours de Paléolithique moyen en Aquitaine septentrionale. In C. Farizy (ed.), Paléolithique moyen recent et Paléolithique supérieur ancien en Europe. Nemours: Mémoirs de Musée d’Ile de France 3, 203–14. Gero, J. M. 1991: Genderlithics: women’s roles in stone tool production. In J. M. Gero (ed.), Engendering Archaeology: Women in Prehistory. Oxford: Blackwell, 163–93. Goodyear, A. C. 1989: A hypothesis for the use of crypto-crystalline raw materials among Paleoindian groups of North America. In C. G. Ellis and J. C. Lothrop (eds.), Eastern Paleoindian Lithic Resource Use. Boulder, CO: Westview Press, 1–9. Gordon, D. 1993: Mousterian tool selection, reduction, and discard at Ghar, Israel. Journal of Field Archaeology, 20(2), 205–18. Gould, R. A. 1968: Chipping stones in the outback. Natural History, 77(2), 42–8. Hall, J. and Love, W. 1985: Prickly Bush, a site with backed blades on the Brisbane River: a pilot study towards the measurement of site “disturbance.” Queensland Archaeological Research, 2, 71–81. Harrison, R. 2002: Archaeology and the colonial encounter, Kimberley spearpoints, cultural identity and masculinity in the north of Australia. Journal of Social Archaeology, 2(3), 352–77. Hayden, B. 1979: Paleolithic Reflections: Lithic Technology and Ethnographic Excavations among Australian Aborigines. Canberra: Australian Institute of Aboriginal Studies.

AN INTRODUCTION TO STONE ARTIFACT ANALYSIS 201 —— and Hutchings, W. K. 1989: Whither the billet flake? In D. S. Amick and R. P. Mauldin (eds.), Experiments in Lithic Technology. Oxford: British Archaeological Reports, 235–57. ——, Franco, N. and Spafford, J. 1996: Evaluating lithic strategies and design criteria. In G. H. Odell (ed.), Stone Tools: Theoretical Insights into Human Prehistory. New York: Plenum Press, 9–50. Hill, J. N. and Evans, R. K. 1972: A model for classification and typology. In D. L. Clarke (ed.), Models in Archaeology. London: Methuen, 231–74. Hiscock, P. 1982: A technological analysis of quartz assemblages from the south coast. In S. Bowdler (ed.), Coastal Archaeology in Eastern Australia. Canberra: Department of Pre- history, Research School of Pacific and Asian Studies, Australian National University, 32–45. —— 1985: The need for a taphonomic perspective in stone artefact analysis. Queensland Archaeological Research, 2, 82–95. —— 1996: Mobility and technology in the Kakadu coastal wetlands. Bulletin of the Indo- Pacific Prehistory Association, 15, 151–7. —— 2002a: Pattern and context in the Holocene proliferation of backed artefacts in Australia. In R. G. Elston and S. L. Kuhn (eds.), Thinking Small: Global Perspectives on Microlithization. Archaeological Papers of the American Anthropological Association (AP3A) no. 12, 163–77. —— 2002b: Quantifying the size of artefact assemblages. Journal of Archaeological Science, 29(3), 251–8. —— 2005: Looking the other way: a materialist/technological approach to classifying tools and implements, cores and retouched flakes, with examples from Australia. In S. McPherron and J. Lindley (eds.), Tools or Cores? The Identification and Study of Altern- ative Core Technology in Lithic Assemblages. Pennsylvania: University of Pennsylvania Museum, forthcoming. —— and Attenbrow, V. 2002: Early Australian implement variation: a reduction model. Journal of Archaeological Science, 30(2), 239–49. —— and —— 2003: Morphological and reduction continuums in eastern Australia: measurement and implications at Capertee 3. Tempus, 7, 167–74. —— and —— 2005: Reduction continuums and tool use. In C. Clarkson and L. Lamb (eds.), Rocking the Boat: Recent Australian Approaches to Lithic Reduction, Use and Classification. British Archaeological Reports International Monograph Series. Oxford: Archaeopress, forthcoming. —— and Clarkson, C. 2000: Analysing Australian stone artefacts: an agenda for the twenty first century. Australian Archaeology, 50(1), 98–108. —— and —— 2005: Measuring artefact reduction – an examination of Kuhn’s Geometric Index of Reduction. In C. Clarkson and L. Lamb (eds.), Rocking the Boat: Recent Australian Approaches to Lithic Reduction, Use and Classification. British Archaeological Reports Inter- national Monograph Series. Oxford: Archaeopress, forthcoming. Holdaway, S., McPherron, S. and Roth, B. 1996: Notched tool reuse and raw material availability in French Middle Paleolithic sites. American Antiquity, 61, 377–87. Ingold, T. 1990: Society, nature and the concept of technology. Archaeological Review from Cambridge, 9, 5–17. Jeske, R. J. and Lurie, R. 1992: The archaeological visibility of bipolar technology: an example from the Koster site. Midcontinental Journal of Archaeology, 18, 131–60. Jones, R. and White, N. 1988: Point blank: stone tool manufacture at the Ngilipitji Quarry, Arnhem Land, 1981. In R. Jones (ed.), Archaeology with Ethnography: An Australian

202 CHRIS CLARKSON AND SUE O’CONNOR Perspective. Canberra: Department of Prehistory, Research School of Pacific Studies, Australian National University, 51–87. Jung, S. 1992: Trample damage of stone flakes as an index of occupation intensity: a case study from Magnificent Gallery. Queensland Archaeological Research, 9, 26–8. Kobayashi, H. 1985: The accidental breakage of backed blades. Lithic Technology, 14, 16–25. Kooyman, B. P. 2000: Understanding Stone Tools and Archaeological Sites. Calgary: University of Calgary Press and University of New Mexico Press. Kuhn, S. 1990: A geometric index of reduction for unifacial stone tools. Journal of Archae- ological Science, 17, 585–93. Kuhn, S. L. 1995: Mousterian Lithic Technology. Princeton: Princeton University Press. Lamb, L. 2005: Backed and forth: an assessment of typological categories and technological continuums. A case study from the Whitsunday Islands, Central Queensland Coast. In C. Clarkson and L. Lamb (eds.), Rocking the Boat: Recent Australian Approaches to Lithic Reduction, Use and Classification. British Archaeological Reports International Monograph Series. Oxford: Archaeopress, forthcoming. Law, W. B. 2005: Stone artefact reduction, mobility, and arid zone settlement models: a case study from Puritjarra Rockshelter. In C. Clarkson and L. Lamb (eds.), Rocking the Boat: Recent Australian Approaches to Lithic Reduction, Use and Classification. British Archae- ological Reports International Monograph Series. Oxford: Archaeopress, forthcoming. Lemonnier, P. 1986: The study of material culture today: toward an anthropology of technical systems. Journal of Anthropological Archaeology, 5, 147–86. Leroi-Gourhan, A. 1964: Le geste et la parole 1: technique et langage. Paris: Albin Michal. Leonard, R. D. and Jones, G. T. (eds.) 1989: Quantifying Diversity in Archaeology. Cambridge: Cambridge University Press. Luedtke, B. E. 1992: An Archaeologist’s Guide to Chert and Flint. Archaeological Research Tools, no. 7. Los Angeles: Institute of Archaeology, University of California. Lyman, R. L. and O’Brien, M. J. 2000: Measuring and explaining change in artifact variation with clade-diversity diagrams. Journal of Anthropological Archaeology, 19, 39–74. ——, —— and Dunnel, R. C. 1997: The Rise and Fall of Culture History. New York: Plenum Press. McCarthy, F. D., Brammell, E. and Noone, H. V. V. 1946: The stone implements of Australia. Memoirs of the Australian Museum, 9, 1–94. Macgregor, O. 2005: Abrupt terminations and stone artefact reduction potential. In C. Clarkson and L. Lamb (eds.), Rocking the Boat: Recent Australian Approaches to Lithic Reduction, Use and Classification. British Archaeological Reports International Monograph Series. Oxford: Archaeopress, forthcoming. Mallouf, R. J. 1982: An analysis of plow-damaged chert artifacts: the Brookeen Creek cache (41H186), Hill County, Texas. Journal of Field Archaeology, 9(1), 79–98. Marcy, J. L. 1993: Aperçu sur les stratégies de producion des raclois du niveau. In Riencourt-lès-Bapaume (Pas-de-Calais): un gisement du Paléolithique moyen, dirigé par Alain Tuffreau. Éditions de la Maison des Sciences de l’Homme, coll. Documents d’Archéologie Française, no. 37, 87–94. Meignen, L. 1988: Variabilité technologique au proche orient: l’exemple de Kebara. In J. P. Rigaud (ed.), L’homme de Néandertal. Liège: Université de Liège, 87–95. Mewhinney, H. 1964: A skeptic views the billet flake. American Antiquity, 30, 203–5. Morrow, T. A. 1997a: A chip off the old block: alternative approaches to debitage analysis. Lithic Technology, 22, 51–69. —— 1997b: End scraper morphology and use-life: an approach for studying Paleoindian lithic technology and mobility. Lithic Technology, 22(1), 51–69.

AN INTRODUCTION TO STONE ARTIFACT ANALYSIS 203 Neeley, M. P. and Barton, C. M. 1994: A new approach to interpreting late Pleistocene microlith industries in southwest Asia. Antiquity, 68, 275–88. Nelson, M. C. 1991: The study of technological organization. Archaeological Method and Theory, 3, 57–100. Newcomer, M. H. 1975: “Punch technique” and Upper Paleolithic blades. In E. H. Swanson (ed.), Lithic Technology: Making and Using Stone Tools. The Hague: Mouton, 97–102. —— and Karlin, C. 1987: Flint chips from Pincevent. In G. D. G. Sieveking and M. H. Newcomer (eds.), The Human Uses of Flint and Chert: Proceedings of the Fourth International Flint Symposium, Held at Brighton Polytechnic, 10–15 April 1983. Cambridge: Cambridge University Press, 33–6. O’Brien, M. J. 1996: Evolutionary archaeology: an introduction. In M. J. O’Brien (ed.), Evolutionary Archaeology: Theory and Application. Salt Lake City: University of Utah Press, 1–15. Odell, G. H. 1989: Experiments in lithic reduction. In D. S. Amick and R. P. Mauldin (eds.), Experiments in Lithic Technology. BAR International Series, 528. Oxford: British Archae- ological Reports, 163–98. —— 2000: Stone tool research at the end of the millennium: procurement and technology. Journal of Archaeological Research, 8(4), 269–331. —— 2001a: Research questions R us. American Antiquity, 66(4), 679–85. —— 2001b: Stone tool research at the end of the millennium: classification, function and behavior. Journal of Archaeological Research, 9(1), 45–100. Patterson, L. W. 1982: Replication and classification of large sized lithic debitage. Lithic Technology, 11, 50–8. —— 1990: Characteristics of bifacial-reduction flake-size distribution. American Antiquity, 55, 550–8. —— and Sollberger, J. B. 1978: Replication and classification of small size lithic debitage. Plains Anthropologist, 23, 103–12. Peacock, E. 1991: Distinguishing between artifacts and geofacts: a test case from Eastern England. Journal of Field Archaeology, 18(3), 345–61. Pelcin, A. 1997a: The effect of core surface morphology on flake attributes: evidence from a controlled experiment. Journal of Archaeological Science, 24, 749–56. —— 1997b: The effect of indentor type on flake attributes: evidence from a controlled experiment. Journal of Archaeological Science, 24, 613–21. —— 1997c: The formation of flakes: the role of platform thickness and exterior platform angle in the production of flake initiations and terminations. Journal of Archaeological Science, 24, 1107–13. —— 1998: The threshold effect of platform width: a reply to Davis and Shea. Journal of Archaeological Science, 25, 615–20. Peterson, J., Mitchell, D. R. and Shackley, M. S. 1997: The social and economic contexts of lithic procurement: obsidian from Classic-period Hohokam sites. American Antiquity, 62, 231–59. Pfaffenberger, B. 1992: Social anthropology of technology. Annual Review of Anthropology, 21, 491–516. Phagan, C. J. 1985: Lithic technology: flake analysis. In R. MacNeish, A. Nelken-Terner, C. J. Phagan and R. Vierra (eds.), Prehistory of the Ayacucho Basin, Peru. Michigan: The University of Michigan Press, 233–81. Prentiss, W. C. 1998: The reliability and validity of a lithic debitage typology: implications for archaeological interpretation. American Antiquity, 63(4), 635–50.

204 CHRIS CLARKSON AND SUE O’CONNOR —— and Romanski, E. J. 1989: Experimental evaluation of Sullivan and Rosen’s debitage typology. In D. S. Amick and R. P. Mauldin (eds.), Experiments in Lithic Technology. Oxford: British Archaeological Reports, 89–99. Purdy, B. A. and Brooks, H. K. 1971: Thermal alteration of silica materials: an archaeological approach. Science, 173, 322–5. Renfrew, C., Dixon, J. E. and Cann, J. R. 1968: Further analysis of Near Eastern obsidians. Proceedings of the Prehistoric Society, 34, 319–31. Roebroeks, W., Kolen, J. and Rensink, E. 1988: Planning depth, anticipation and the organization of technology: the “archaic natives” meet Eve’s descendants. Helinium, 28, 17–34. Sassaman, K. E. 1992: Lithic technology and the hunter-gatherer sexual division of labor. North American Archaeologist, 13, 249–62. Schiffer, M. B. and Skibo, J. M. 1997: The explanation of artifact variability. American Antiquity, 62, 27–50. Schindler, D. L., Hatch, J. W., Hay, C. A. and Bradt, R. C. 1984: Thermal alteration of Bald Eagle jasper. American Antiquity, 49(1), 173–7. Schnurrenberger, D. and Bryan, A. L. 1985: A contribution to the study of the naturefact/ artifact controversy. In M. G. Plew, J. C. Woods and M. G. Pavesic (eds.), Stone Tools Analysis: Essays in Honor of Don E. Crabtree. Albuquerque: University of New Mexico, 133–59. Sellet, F. 1993: Chaîne opératoire; the concept and its applications. Lithic Technology, 18(1), 106–12. Shackley, M. S. 1998a: Current issues and future directions in archaeological volcanic glass studies. In M. S. Shackley (ed.), Archaeological Obsidian Studies: Method and Theory. New York: Plenum Press, 1–14. —— 1998b: Gamma rays, X-rays and stone tools: some recent advances in archaeological geochemistry. Journal of Archaeological Science, 25, 259–70. Shott, M. J. 1989: On tool-class use lives and the formation of archaeological assemblages. American Antiquity, 54, 9–30. —— 1994: Size and form in the analysis of flake debris: review of recent approaches. Journal of Archaeological Method and Theory, 1, 69–110. —— 1996: Stage versus continuum models in the debris assemblage from production of a fluted biface. Lithic Technology, 21(1), 6–22. —— 2000: The quantification problem in stone-tool assemblages. American Antiquity, 65(4), 725–38. ——, Bradbury, A. P., Carr, P. J. and Odell, H. O. 2000: Flake size from platform attributes: predictive and empirical approaches. Journal of Archaeological Science, 27, 877–94. Sinclair, A. 1995: The technique as symbol in late glacial Europe. World Archaeology, 27(1), 50–62. —— 2000: Constellations of knowledge: human agency and material affordance in lithic technology. In M.-A. Dobres and J. Robb (eds.), Agency in Archaeology. London: Routledge, 197–212. Spaulding, W. G. 1953: Statistical techniques for the discovery of artifact types. American Antiquity, 18, 305–13. Specht, J. 2002: Obsidian, colonising and exchange. In S. Bedford, C. Sand and D. Burley (eds.), Fifty Years in the Field. Essays in Honour and Celebration of Richard Shutler Jr’s Archaeological Career. Monograph 25. Auckland: New Zealand Archaeological Associ- ation, 37–50. Speth, J. D. 1972: Mechanical basis of percussion flaking. American Antiquity, 37, 34–60.

AN INTRODUCTION TO STONE ARTIFACT ANALYSIS 205 —— 1974: Experimental investigations of hard-hammer percussion flaking. Tebiwa, 17, 7–36. —— 1981: The role of platform angle and core size in hard-hammer percussion flaking. Lithic Technology, 10(1), 16–21. Stahle, D. W. and Dunn, J. E. 1984: An analysis and application of the size distribution of waste flakes from the manufacture of bifacial stone tools. World Archaeology, 14, 84–97. Steffen, A., Skinner, E. J. and Ainsworth, P. W. 1998: A view to the core: technological units and debitage analysis. In A. Ramenofsky and A. Steffen (eds.), Unit Issues in Archae- ology. Salt Lake City: University of Utah Press, 131–46. Sullivan, A. P. and Rozen, K. C. 1985: Debitage analysis and archaeological interpretation. American Antiquity, 50, 755–79. Taçon, P. S. C. 1991: The power of stone: symbolic aspects of stone use and tool development in western Arnhem Land, Australia. Antiquity, 65, 192–207. Torrence, R. 1986: Production and Exchange of Stone Tools: Prehistoric Obsidian in the Aegean. Cambridge: Cambridge University Press. —— 1989: Re-tooling: towards a behavioral theory of stone tools. In R. Torrence (ed.), Time, Energy and Stone Tools. Cambridge: Cambridge University Press, 57–66. —— and Summerhayes, G. 1997: Sociality and the short distance trader: Inter-regional obsidian exchange in the Willaumez Peninsula region, Papua New Guinea. Archaeology in Oceania, 32(1), 74–8. Touhy, D. R. 1987: A comparison of pressure and percussion debitage from a Crabtree obsidian stoneworking demonstration. Tebiwa, 23, 23–30. Ugan, A., Bright, J. and Rogers, A. 2003: When is technology worth the trouble? Journal of Archaeological Science, 30, 1315–29. van der Leeuw, S. E. and Torrence, R. 1989: What’s New? A Closer Look at the Process of Innovation. London: Unwin Hyman. Walthall, J. A. and Holley, G. R. 1997: Mobility and hunter–gatherer toolkit design: analysis of a Dalton lithic cache. Southeastern Archaeology, 16, 152–62. White, J. P. and O’Connell, J. F. 1982: A Prehistory of Australia, New Guinea and Sahul. Sydney: Academic Press. Whittaker, J. 1994: Flintknapping: Making and Understanding Stone Tools. Austin: University of Texas Press. Wurz, S. 1999: The Howiesons Poort backed artefacts from Klaises River: an argument for symbolic behavior. South African Archaeological Bulletin, 169, 38–50. Yvorra, P. 2000: Exploitation de l’analyse quantitative des retouches pour la caractérisation des industries lithiques du Moustérien: application au faciès Quina de la vallée du Rhône. BAR International Series, no. 869. Oxford: British Archaeological Reports. Zeitlin, R. N. 1982: Toward a more comprehensive model of interregional commodity distribution: political variables and prehistoric obsidian procurement in Mesoamerica. American Antiquity, 47, 260–75.

206 CHRIS CLARKSON AND SUE O’CONNOR

Letting the past serve the present--some contemporary uses of archaeology in ... Ian C Glover Antiquity; Sep 1999; 73, 281; Academic Research Library pg. 594

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 25 Hoabinhian Food Strategy in Viet Nam Nguyen Viet

n 1982, influenced by Chester Gorman's excavation in the Spirit Cave (Thailand), II began writing "Micro Studies of Archaeology" to better understand Hoabinhian food strategies (200 ib). Besides using traditional digging methods, all sediments from Xom Trai and Con Moong caves were carefully selected through sieving and sorting using a microscope. These methods brought us thousands of different meal remains of animals, as well as plants, which allowed us to carry out a quantitative research to determine Hoabinhian food strategies in Viet Nam.

Some New Interpretations on the Hoabinhian Concerning Food Exploitation, Chronology, and Evolution

Radiocarbon Dating for Classic Typical Hoabinhian

In some of my recent publications (Viet 1989, 2000, 2001a, 2001b), I have shown believable evidence of early Hoabinhian sites circa 20000 bp. This was mainly based on the results of my reexcavating in 1986 and 1987 the most important Hoabinhian sites in Viet Nam, i.e., Xom Trai, Lang Vanh, Hang Muoi, Con Moong, and Sung Sam. Nearly a hundred samples were collected for radiocarbon dating, and half of these were processed in the radiocarbon dating laboratory in former East Berlin. Radiocarbon samples from many different Hoabinhian sites in Viet Nam were reexamined and remeasured. Working in the laboratory in Berlin, we dated and identified a classic and typical Hoabinhian

442 Hoabinhian Food Strategy in Viet Nam 443

Ul Ul VI Z ~ . <

...

Figure 1. The shaded portions represent the areas where Hoabinhian techno-complex finds were concentrated in Mainland Southeast Asia (after Moser 20;01)

... 444 Nguyen Viet system that existed about ten thousand years before the traditionally accepted time depth. The Muong Vang Valley in Hoa Binh province, where three typical Hoabinhian sites (Xom Trai, Lang Vanh, Hang Muoi) are located, is recognized as the homeland of Hoabinhian in Viet Nam. The tool kits of the Hoabinhan here are very typical of the so-called classic Hoabinhian recovered and determined by Madaleine Colani more than seventy years ago. At Xom Trai Cave, we collected a total of thirty-two organic samples, fifteen of which were charred fragments ofJuglans spp. nut covers. The rest were shells of the stream snail Antimelania spp. Almost all radiocarbon dates confirmed the first Hoabinhian occupation at circa 18500 bp. In the same valley, the Hoabinhian caves Xom Tre and Lang Vanh, located around 2 km and 4 km respectively south ofXom J Trai were contemporaneous with Xom Trai. Their 14C dates for upper stratigraphic layers were at 16000 to 17000 BP (Viet 1989, 2001). The radiocarbon dates for Hoabinhian sites in Muong Vang Valley support a very early Hoabinhian occupation I in this area. They represent the early typical Hoabinhian.

Radiocarbon Dating for Untypical Hoabinhian c f Almost all untypical Hoabinhian sites in Viet Nam have radiocarbon date ranging from 13000 to 8000 BP. The reexcavation at Con Moong in 1986-1987 by this author brought to light many reliable radiocarbon dates. These 14C ages confirmed the existence of a deeper Hoabinhian chronological sequence. Except for some sites (like Hang Muoi), most sites contained a poor and untypical Hoabinhian tool kit. They represented a Hoabinhian developing in more difficult conditions at later phases, during which an increasing population forced them to enlarge their food gathering areas farther away from their homeland. Because ofthis, two Hoabinhian subsistence systems developed. One was richer than, for example, the homeland Muong VangValley. It exisred in a moist and cool environment. Another one was poorer, particularly in tool kit development, and lived in warm, even hot conditions. The relationship between the two Hoabinhian systems will be discussed again below.

Hoabinhian Evolution

The earliest Hoabinhian evidence in Viet Nam might be found in the Tham Khuong Cave in the province of Lai Chau, where we have 14C and Electron Spin Hoabinhian Food Strategy in Viet Nam 445

Resonance dates circa 32000 bp. However, we have relatively few materials on this group. We believe that similar earlier Hoabinhian sites can also be found in the Hoa Binh province. Till now, the earliest Hoabinhian occupations in this province were discovered only in the Muong Vang Valley. Their 14C dates range from 18500 to 16000 BP. The Muong Vang Valley is recognized as the best Hoabinhian locality, where, with a paradisiac condition, man could easily get naturally formed pebbles suitable for tool making and could collect abundant food throughout the year. Occupants of the area belonged to the richer Hoabinhian, whose living area became the homeland of Hoabinhian culture (Viet 2001a). The Hoabinhian distribution shows that later Hoabinhian sites representing the poor Hoabinhians (Con Moong, Hang Dang, Sung Sam, Dong Cang, Mai Da Dieu, and Hang Chua) were located outside the center of the Hoabinhian stone source. There is increasing evidence of Hoabinhian sites at the transition period from the Pleistocene to the Holocene. The paradisiac condition in the "gold valley" at Muong Vang may have caused populations to increase, resulting in the need to enlarge the foodgathering areas. The case of Hang Muoi is a good example for later Hoabinhian population growth in the homeland region. Expanding the search for food into more difficult areas resulted in the creation of poor Hoabinhian sites such as Con Moong, Sung Sam, and Dong Cang. They belong to the enlarging Hoabinhian later system. This enlarging process began circa 13000 bp. Of the Hoabinhian enlarging movements, there were some Hoabinhian groups whose main food strategy was gathering mollusks from swamps and lakes, as the open-air shell middens nearby traditional Hoabinhian valley regions indicate. This is what happened in the case of the Da But site. The first Hoabinhian occupation at Da But happened perhaps around 8000 to 9000 hp. Here, the Hoabinhian tried a new way of living in new environmental conditions that corresponded to the warm and rainy Holocene climate. However, ~ the Hoabinhian in Da But didn't live in caves or rockshelters, but in cottages or simple houses. Wood or bamboo working became more intensive than before. Edge-polishing tools and ceramics were widely used. Whereas nomadic life in the mountains resulted in the absence of burial fields, the dead in Da But were usually buried near cottages or houses. This is evidence for a settled way of life. Since the Da But period, the Hoabinhian culture had the opportunity to adopt a Neolithic way of life (Viet 2002). 446 Nguyen Viet

Paleoenvironment

Changing Sea Level during Hoabinhian

The Hoabinhian existed for around fourteen thousand years (from circa 20000 to 6000 bp). This time range corresponds to the changes of th~sea level and the global climate from the Last Glacial Maximum (LGM) to the present climate and sea level. When the Hoabinhian occupied the Muong Vang Valley, the sea level was at around 150 m to 120 m. The temporal shoreline was around 200 m to 400 m away from the Hoabinhian homeland (Chen and Wengiang 1985; Ty 1985; Vie~ 2001a). According to Nakada and Lambeck (1988) the sea level rose relatively slowly during 18000 to 20000 BP, but became rapid during 12000 to 6000 BP. This is supported by new data from the Fairbanks Curve. In the Malacca Strait, Gegh, Kudrass, and Streif (1979) determined that the sea level, corresponding to -80 m in 6000 BP, rose to its highest level (+ 6 m) in 12000 BP. The same situation was recorded in Japan, New Zealand, and Australia (Nakada and Lambeck 1988) as well as in the East China Sea (Chen and Wengiang 1985). The 14C dates ofsome inundated shorelines in Hong Kong show that the sea level got to its present level at about 6000 BP (Meacham 1978). The boring hole ND1 in North Viet Nam shows that the sea during 12000 BP was at a -45 m level (Shigeko 2001) . Radiocarbon dating at Ha Lung Cave (North Viet Nam) states that before 6000 BP, freshwater snails were collected in this region, which at present is 10-20 m under water (Viet 1989). More recent research has .confirmed that the highest sea level (+4 m to +5 m) during the middle Holocene in

Viet Nam occurred during 6000 to 5000 BP (Viet 2002; Thuan 2002; Tiep 1998). The changes in sea level during the last 20,000 years led to a marked reduction of li;'ingsurface for the Hoabinhian (Bellwood 1985). However, it happened in Viet Nam in only some traditional Hoabinhian food exploiting areas, like Ha Long Bay. The reduction ofthe Hoabinhian living surface occurred only in the middle Holocene. Until now, we do not have evidence for a migration wave due to rising sea levels from the "lower Hoabinhian" into traditional Hoabinhian mountainous areas before 6000 BP.

Mean Temperature during Hoabinhian

The temperature during 20000 to 15000 BP is regarded as the lowest during the last 20,000 years. According to CLIMAP (1981), the mean sea surface "temperature Hoabinhian Food Strategy in Viet Nam 447

(SST) in South China Sea was around 2 °C lower than at present. Based on pollen data . from the Leizhou peninsula (Southern China), Zheng and Lei (1999) estimated that the temperature during the LGM in a mountain forest altitude of800 m was around 4 °C to 6 °C lower than it is today. The lapse rate here was 0.61 °C per 100 m. After Zhang (1984), we can estimate that the mean temperature in North Viet Nam at LGM was 0-5 °C in January and 25°C in July. Prell (1982) presented an isomap SST for the West Sunda Sea, in which North Viet Nam and South China Sea belong to the region with an SST of around 24°C in August (around 4 °C lower than at present). The pollen record from Taiwan suggested that the mean temperature in 24000 -14000 BP was 2 °C to 6°C lower than at present (Tsukada 1966). In Japan, pollen data confirmed that the mean temperature was red~ced from 8 °C to 9 °C in northern Hokkaido, 7.7 °C to 8.7 °C in Northern Honshu, 7.2 °C to 8.4 °C in Central mountains, 6.5 °C in Chugoku district, and 5 °C to 6 °C in Kyushu (Tsukuda 1983). In the Yangtze Delta, pollen data showed the mean temperature was below 5°C at LGM (Lin et al. 1989). For the case of southern Sunda, Verstappen (1975) proposed it was 3 °C to 5 °C cooler, while Peterson (1969) estimated that the mean temperature at Niah Cave region during the Wurm II glacial was 5 °C to 7 °C lower than today, with rainfall reduced by 30 percent. The pollen data taken by Stuijts (1984) in West Java (Indonesia) yielded evidence for a mean temperature of 5 °C to 6 °C during 16800 to 10870 BP, which is lower than at present. According to Frenzel (1968), almost the entire surface ofthe northern Hoabirihian region was covered by broad leaf evergreen forest during the Wurm Glacial. According to Aigner, this forest in China corresponds to a climate with a temperature that was about 5°C lower than today. In Viet Nam, the pollen data recorded from the Nguom rockshelter circa 23000 BP and from Xom Trai cave (18500 - 16600 BP) frequently included species such as juglans spp., querous spp., salix spp., carpinus spp., lithocarpus spp., and minica spp. These plant types corresponded to the cool forest vegetation (or montane forest II) that still exists in altitudes of 1,800 m to 2,400 m in Taiwan and central Sumatra (see Tsukada 1966 and Morley 1982). The presence of these species in lower altitudes in Viet Nam suggests a reduction ofmean temperature around 5 °C to 8 °C lower than at present. 1 The paleontological records support this cooling. AiluropodaJossils circa 18000 BP were found in Viet Nam at the Phung Quyen rockshelter and Pongo bones circa 23000 BP were recovered at the N guom rockshelter. The macro plant remains collected from some Hoabinhian caves show a clear change in the kind of flora exploited by Hoabinhian inhabitants during 18500 to 8000 BP. 448 Nguyen Viet

In Xom Trai Cave, 90 percent of total plant remains collected were charred shells, probably belonging to a walnut (jugLans) or gam (gnetum sp) species which can usually be seen today in high mountain forests. These charred fragments were also found in Lang Vanh, Mai Da Dieu, and Dong Cang. They were used as ideal materials for radiocarbon dating. In Xom Trai, Lang Vanh, and Mai Da Dieu, radiocarbon dates from these charcoal show ages ranging from 20000 to 16000 BP. Such plant remains were totally absent in the later Hoabinhian sites. In Taiwan, Tsukada (1966) mentions a particular event in pollen history: "near the top ofzone M,2 JugLans cathayensis (Dode) suddenly appears, but decreases before the M/R zonal boundary (ca. 12000 - 10000 BP-NV)" and 'JugLans cathayensis, associated with Cunninghamia, may give a reliable estimate of the mean temperature (lower by 2 °C to 6 0C) between ca. 25000 and ca. 14000 BP" (Tsukada 1966, 547). JugLans in Jihyuetans pollen history represented cool climate at LGM. Presently, this genus commonly grows in cool-temperatured forests in Taiwan at altitudes from 1800 m to 2400 m. About fifteen thousand years ago, sea temperatures began to increase, corresponding to the rise of sea levels. From the data of core V35-5 in South China Sea (in 7 °12'N-112 °5'E) and core SONNEN 50-37KL, in 18 °55'N-115 °45'E, which are located not far from the east margin of the Sunda Shelf (in 7 °12'N-112 °5'E), Broecker discovered the abrupt termination of the last glacial age, in which distinct phases of sedimentation and productivity of foraminifera organisms during 18000 to 6000 BP were recognized. During Phase I (from 18000 to 13000 BP), the sedimentation rate was rapid while the productivity of foraminifera organisms was slow. In Phase II (from 13000 to 6000 BP), it is the reverse (Broecker et al. 1990). According to Flenley (1979), all pollen data, whether from the northern part ofTaiwan (Tsukada 1966), the southern part ofEast Malaysia and Sumatra (Morley 1982, Maloney 1995), or further southeastward from New Guinea (Flenley 1979), demonstrated a climatic rapid amelioration during 14000 to 12000 BP. Morley, Fleriley, and Tsukada recorded a peak, with temperatures around 2 °C warmer than at present around 10000 to 8000 BP in Sumatra, around 8000 BP in New Guinea and New Zealand, and around 6000 BP in Taiwan. The climate also changed in Viet Nam during 12500 to 8000 BP (Viet, Kohl, and Goersdorf 1988). At Con Moong, we recovered a different assemblage ofplant remains from cultural sediment 2. Hoabinhian Food Strategy in Viet Nam 449

1. The earlier assemblage: From the lower layers, with radiocarbon dates of 12500 to 10500 BP, it was determined that humans collected nuts. Thousands of nutshell fragments determined as probable Aleurites spp. or Castanopsis spp. were recovered. Today, such plants grow mainly in the mountain regions of North Viet Nam. Castanopsis spp., according to Yen (1977), presently grows in high Thailand mountain forests, while in a higher latitude, in Taiwan, for example, this genus is found in warm temperature forests 500-1,800 m in altitude from sea level (Tsukada 1966). In New Guinea, this genus is frequently seen in lower mountain forests mountainforest I at 1,500-1,800 m altitude (Flenley 1985; see Morley 1982). This kind ofnut fragments have frequently been associated with Celtis spp. stone cores in dried form. Celtis spp. cores are relatively common in the Hoabinhian sites. On a block of cemented sediment brought by M. Colani from My Te Cave3 we _recognized eight Celtis spp. cores, which were stuck on a broad mammalian bone. Celtis spp. is also recorded at Nguom, Xom Trai, and Con Moong. According to Yen (1977), dried Celtis spp. cores were found in Spirit Cave, Thailand. Celtis spp. is one ofthe characteristic species ofthe submontane forest in Java-common at altitudes of 1,000-1,400 m (Morley 1982). In some higher altitudes, in Taiwan, for example, Celtis spp. commonly grows in subtropical rain forests at altitudes below 500 m (Tsukada 1966). Celtis spp. pollen was frequently recorded at Zengpiyan Cave (Northern Kwangzi) in layers 9000 -7000 BP (Wang 1989). 2. The later assemblage: In the upper level of the Con Moong Cave (from layers A4b to A2), Canarium spp. fruit stones were commonly uncovered, while _ there was a marked absence ofJuglans spp. The radiocarbon dates associated with such plant remains showed a time depth of 10500 to 8500 BP. Cannarium spp. may characterize the beginning of the Holocene in Viet Nam as well as in Southeast Asia. Some of these kinds of plants were determined by Yen as Canarium spp., or more confidently by Thin and Thuong (1987) as Canarium tonkinensis (Engl.). Canarium was recorded in Dong Cang, Hang Muoi, Hang Doi, and the upper level of Mai Da Dieu. In the Southeast Asian Hoabinhian context, charred fragments of Canarium spp. were documented in Spirit Cave, Tham Pa Chan, Banyan Valley Cave (Yen 1977), and Ongbah Cave (Soerensen 1988). It's notable that almost all Canarium spp. remains found in Southeast Asia are younger than 1lO00 BP'4 This species is very common in Mainland Southeast Asia until now. The seemingly sudden explosion of Canariumspp. in the vegetable menu ofCon Moong dwellers in 9500 - 8500 BP marked a warming phase in Viet Nam, which is

------~~------450 Nguyen Viet

supported by evidence in Taiwan (Tsukada 1966), Sumatra (Morley 1982), and New Guinea (Flenley 1979,1985). The COHMAP explained that the warming phase may have been the result ofa decrease in the distance between the earth and sun, especially after 9000 BP (COHMAP 1988,1044,1048). In Taiwan, Tsukada (1966) estimated that this warmer phase was 2 °C to 3 °C higher than at present during the peak of6000 BP. In Sumatra around 10000 to 8600 BP, the mean temperature increased, possibly by 2°C (Morley 1982).

Precipitation during Hoabinhian

Ha van Tan (1985) suggested that the rockfall layer at N guom rockshelter reflected a dry and cold climate before 23000 BP. Hoang Du and Nguyen Duc Tung (1977) took their pollen data collected from the lowest layer in the 1975 1976 Con Moong excavation, to represent a dry climate. This layer has a 14C age of 12,000-11, 000 years ago. The above-mentioned op inions represent the generally accepted view by almost all quaternary researchers in Viet Nam. These views accept that the late Pleistocene climate was generally dry and cool. This general conception was supported by climate models proposed by COHMAP (1988) for the case of the LGM climate. However, through our research, we have discovered many evidences that suggest a humid climate with strong rainfall during LGM in Viet Nam. The relative frequent presence~ of mollusk remains in the cultural deposits from 23000 to 18600 BP in the Nguom rockshelter (Long and Phong 1986), as well as the consequent increase ofmollusk remains during the Hoabinhian period, tells ofa long dry period covering the late Pleistocene. The stratigraphy ofsome late upper Pleistocene (Q3) sites in Southern China reflects an identical palaeontological event where some representative species ofthe Ailuropoda-Stegodon fauna became rare. In contrast, mollusk

" shells visibly increased. Based on mollusk data from Nguom and Tham Khuong, we can estimate that in about 30000 BP, the climate in some ofthe regions, in Tham Khuong (1200 m altitude, 21 °35N), for example, was not too dry and not too cold. The presence of many shells of the land snail Cyclophorus spp. in the lowest layer ofTham Khuong Cave supports this. From 30000 BP to 23000 BP, however, the climate probably became very dry and cooP Studying the surfaces of Melania spp. shells collected from Xom Trai, we found that they were consistently destroyed by being smashed on stream stones ~ This is Hoabinhian Food Strategy in Viet Nam 451 evidence of the role of floods caused by strong rainfall. The fast-running floodwater pushed the snail shells against the stones, smashing them in the process. This phenomenon suggests strong rainfall periods in Muong Vang Valley during 18500 to 6000 BP. It is contrary to the traditional theory, which estimates a dr-y-period during the LGM in Viet Nam. In the climatic amelioration, rainfall increased in almost all regions of the world. COHMAP (1988) points our that there was 20-30 percent more rainfall then, compared to present levels, particularly during 9000 to 6000 BP. The lake level data from China supports this opinion. In the northern Hoabinhian region, the pollen data from Zengpiyan exhibited a mild-humid to warm-humid climate during 9000 to 7000 BP. At Con Moong Cave in Viet Nam, we observed that at different archaeological layers, the nature of food remained different. This observation also demonstrates two distinct moist phases that corresponded to the increase of the land snails Cyclophorus spp. and Ranguna Spp.6at 12000 - 10000 BP, which peaked at about 10500 BP and 9500 - 9000 BP and also 9300 BP (Viet, Kahl, and Goersdorf 1988). These phases also corresponded to the abundance ofcharred remains of the plants Castanopsis spp. and Canarium spp. which were used as food.

Hoablnhlan Food Remains

Our knowledge of the Hoabinhian menu became more detailed with the excavations in the Mea Hong Son area (Northwest Thailand) led by Chester Gorman in the 1960s. Gorman was also the first researcher to address this question. Before Gorman, there was a common notion that prehistoric organic remains hardly survived under tropical conditions. But this is no longer correct in the case of Hoabinhian sites, where human habitats were mainly caves or rockshelters. The excavation of plant remains, fossils of small animals, even insect eggs and exoskeletons at Xom Trai, and Con Moong contributed to this change in thinking. Gorman (1970, 1971, 1977), and then Yen (1977), collected Hoabinhian-like plant culture remains at the Thailand sites ofSpirit Cave, Tham Pachan, and Banyan Valley Cave. The 1977-1979 excavations at Khao Talu, Men, Petch Kuha, and Heap Caves (West Thailand) provided more plant remains, which Pyramarn investigated (1989). In the excavations at Ongbah Cave (West Thailand), Sorensen (1988) recorded 452 Nguyen Viet

some Canarium spp. fragments. In Viet Nam, Hoabinhian plant remains were found in many sites: Xom Trai, Con Moong, Dong Cang, Mai Da Dieu, Hang Doi, MyTe, Sung Sam, Nguom, Lang Vanh, and Hang Muoi. However, systematic recovery was only carried out at the Xom Trai and Con Moong site excavations (Viet 2000a). In 1971, Gorman presented a list ofHoabinhian animal bone remains that consisted of thirty-two species from twenty-six sites. Pookajorn (1988) presented a comparative list of animals found at the Ban Kao and Mea Hong Son Hoabinhian caves. In Viet Nam, Long (1984) presented a complete picture ofanimal species collected from almost all Hoabinhian excavations in Viet Nam. From the lists mentioned above, we can recognize an abundance ofanimal and plant materials from Hoabinhian sites, but it is still not substantial enough for a significant paleoeconomic study. The discoveries from some excavations where animal bone recoveries were best documented, such as in Xom Trai in 1986, Con Moong in 1987, Dong Can in 1987, and Mai Da Dieu in 1989, are not yet completely published. The Hoabinhian economic picture remains, therefore, only very general as sketched over thirty years ago by Gorman (1971, 1977), Dunn (1970), Glover (1977), and Higham (1977). More recently, Surin Pookajorn (1988) and Gerd Albrecht (2000) undertook a more detailed study of this topic through an ethnoarchaeological approach. But the archaeological materials, especially the botanical and animal remains, were not documented and analyzed paleoeconomically. I would like to discuss some of the data on Hoabinhian food that I and my collaborators have documented and analyzed during the last twenty years:

General Context of Hoabinhian Food Collecting in Viet Nam

The Hoabinhian lifestyle in Viet Nam had two main characteristics: Mountainous valley food exploitation and limestone cave or rockshelter housing. There are four cases that have been carefully researched: Con Moong, Xom Trai-Lang Vanh, Dong Can, and Sung Sam. At Con Moong and Sung Sam, the Hoabinhian people occupied caves at high altitude (about 30 to 50 m over valley niveau). The valleys here were small and marked by the absence ofstreams. Humans must have had difficulty getting water during the dry season. Meal remains from Con Moong cultural sediments showed only products occurring during warm or rainy seasons (Summer) like land snails (CycLophorus spp.), rock crabs (Ranguna spp.), and charred stones of the Canarium spp. fruit. The limestone mountain here accounted for more than 60 percent of each gatherer's food foraging catchments (5 km radius from the cave). The same situation is seen also in Sung Sam Cave. However, the products gathered from Hoabinhian Hoabinhian Food Strate~ in Viet Nam 453

sediment in Dong Can Cave were mainly products of the winter season. This area's topography is characterized by a few limestone mountains (source of Cyclophorus), and earthen ferlarit humus hills (from which castanopsis or canarium fruits can be gathered), but there were more water sources like big streams even during the dry season. Bones of big stream fishes and of tortoises dominated the food remains examined. Ninety percent of the measured fish vertebra bore evidence that they were killed during winter. Very few snail shells were found here. In Xom Trai Cave and Lang Vanh rockshelter of Adam Garden, Muong Vang Valley, remains of food from all seasons were found. The catch basin occurred here with only 0.03 percent of limestone but 5 percent water surface, 40 percent valley surface, and about 55 percent ferlarit humus earthen hills. The main food ofthe Xom Trai and Lang Vanh inhabitants were stream snails (Melania spp.). Around forty seven thousand snail shells and 5 kg ofmammal bone remains could be counted from each cubic meter, corresponding to 78.5 kg ofedible meal from snails and 35 kg from mammals.? This means that the Hoabinhian people that inhabited the resource-rich Muong Vang Valley did not have to move seasonally. The existence of three contemporary Hoabinhian sites (Xom Trai, Lang Vanh, Xom Tre) in the same hunting and gathering area could mean that they were places of rest for the same Hoabinhian group. The moving period was perhaps every three or five years, when the natural food resource condition became bad as a result of overexploitation. Depending on the food resources, the Hoabinhian hunters and gatherers moved seasonally, staying in each location for only a short time. Only in very few cases did the Hoabinhian settle in one place for the entire year. The general food resources of Hoabinhian were gathered from the following environmental conditions:

Limestone rock-mountains (delivering land snail Cyclophorus spp. and some small mammals)

,. Different mountain water sources such as streams, small rivers, valley swamps, and lakes (providing snail melania spp., angulyagra spp., bivalve corbicula spp., and fish) Valley earthen surfaces and felarit humus hill surfaces (delivering nuts and fruits, roots, fungi, vegetables, wild cereals, and many wild mammals)

The traditional Hoabinhian food strategy was oriented on collecting snails and other plant products. They only occasionally hunted big mammals. 454 Nguyen Viet

The Hoabinhian Strategy of Food Exploitation on the Basis of a Common Broad-spectrum Food Exploitation (Hunting, Fishing, and Coliecting)B

From the quantitative data oftotal analyzable food remains excavated at Xom Trai Cave (1982, 1986) and at Con Moong Cave (1986), the determined orientation is not a subsistence based on hunted prey, but rather on collected food. A quantitative research on animal food sources at Xom Trai exhibited the notable domination of Antimelania spp. snails. Over 60 percent ofedible meat weight from animals belonged to stream snails (Viet 1990b). This number has great significance because from it, we have an idea ofhow much time Xom Trai inhabitants spent for collecting and preparing snails daily. Our experimental study shows that with a relatively abundant source, a thirty-year-old woman had to spend nearly two hours to get around 5 kg of stream snails (corresponding to around 2000 calories) and around two more hours preparing and consuming them. The focus on nonvertebrateanimal resources, especially snails (gastropoda),9 is an important feature of the Hoabinhian food gathering habit. This orientation differs in each distinct ecological niche or environmental condition. For example, in some sites, the shells of stream snails dominated the evidence gathered (cases ofXom Trai, Hang Muoi, Lang Vanh) but in others, land-snail shells dominated the findings (cases of Phung Quyen, Sung Sam, Con Moong). The environmental differences of aquatic resources and limestone rock surfaces determined the percentage of mollusk sources in the food composition in each Hoabinhian site. The difference in orientation can also be seen in the use ofplants as food. Over 90 percent ofthe plant remains excavated at Xom Trai (1982) belonged to the Juglans spp. nut. The same percentage ofthe plant remains collected at Con Moong (1986) belonged to Canarium spp. (at the upper layers) and to Castanopsis spp. (at the lower layers). The statistic calculations ofanimal bone fragments showed the predominance ofturtle in the hunting ofXom Trai cave dwellers: 26 percent oftotal bone fragments belong to this species. The same situation can be seen in the Con Moong and Dong Cang caves. The quantitative research shows the local differences ofHoabinhian food collecting. For example, in Dong Can, the Hoabinhian people consumed mainly fish, tortoise, and Canarium spp. nuts. But people in Con Moong mainly consumed land snails (Cyclophorus spp.), stream snails (Melania spp.), rock crabs (Ranguna spp.), nuts (Castanopsis spp.), and Celtis spp. fruits in the lower layers and Canarium spp. fruits in the upper layers. Hoabinhian Food Strategy in Viet Nam 455

We may estimate that snails, fruits, and nuts made up the aaily components of theHoabinhian diet. These orientations were not in contrast to the broad-spectrum hunting and collecting strategy, but they characterized the Hoabinhian food strategy.

Seasonality

A question arises as we research more about Con Moong: whether the cave dweller occupied this place in winter when today, almost all natural streams inside a 2 km radius are nearly totally dry. The majority of the food remains found demonstrates that they were collected mainly during summer, like the Canarium spp: (July-October), land snail (Cyclophorus spp.) (April-October) and rock crabs (Ranguna spp.) (April October), for example. The measurements of growth lines on some shell samples in the layers B3a and B4a (circa 12000 BP) showed that they were collected directly after a hungry period at around the months of March or April (Viet, 1990). In contrast, research on the growth lines on fish bones-a main food item of the Dong Cang inhabitants-shows that the occupations happened only in winter during November to February (Viet, Cuong, and Lam 2000). Gorman had suggested the possibility of seasonal occupation in Northwest Thailand Hoabinhian (Gorman 1971,313). The data on food remains from the Xom Trai Cave didn't exhibit this seasonality. The food sources in this region permitted long-term occupation by Hoabinhian inhabitants. However, the migration pattern of the Phi Tong Luang as described by Pookajorn (1988) points out a possible explanation of the contemporary existence of some sites located near each other (for example, case of Xom Trai, Lang Vanh, and Xom Tre). Maybe these three sites belong to only one Hoabinhian group. This group migrated, not seasonally, but only every three to five years, depending on the availability of regional food sources.

The Changes of Hoabinhian Food Strategy in Da But Culture

The Da But culture existed in northern Viet Nam during the middle Holocene sea transgression. Radiocarbon dates ofthis culture suggest a period ofexistence from the fifth to sixth millennium BCE to about the end ofthe third millennium BCE. As presented above, the Hoabinhian food tradition focused on mountainous shellfish and some nuts and fruits. The Dabutian adds to this tradition a new shellfish species: 456 Nguyen Viet the bivalve Corbicula spp. that lived in swamps and lakes. In some Hoabinhian caves, especially in the east border of the limestone rock massif Thanh Hoa, Hoa Binh, Ninh Binh, and Da Phuc sites we could see the increasing percentage ofvalley swamp shellfish (bivalve Corbicula spp., snail Angulyagra spp.). During 10000 to 8000 BP, rainfall increased in the continent. Under the influence of sea transgression, many coastal deltas were formed. It led to the formation of many swamps and lakes, both breeding grounds for mollusks. In the early Holocene, almost all swamps and lakes were freshwater. Compared to Hoabinhian tr;lditional catchments, the newly formed coastal deltas had larger aquatic surfaces. Thus, swamp shellfish population developed quicker than land and stream snail populations. This allowed the Dabutian to settle in a particular place for a longer period of time. They lived mainly by collecting'the bivalvia Corbicular spp. However, in the upper layers ofDabut, Koekkenmoedingmari'ne shellfish remains increased, an evidence of a rise in sea levels. In another Dabutian case, the people lived mainly by fishing. This happened at the Go Trung site, where thousands of marine fish bone fragments and many stone net weights were unearthed. In another site in Con Co Ngua, the Dabutiancollected many marine shellfish species aside from the Corbicular spp. This indicates cultural adaptation to higher sea levels. The change in subsistence strategy led to a change in the Dabutian lifestyle. The use ofedge-polishing axes to work with wood for building open-air housing (house or shelter building) increased, as well as pottery for cooking Corbicular spp. The dead were buried in a burial field in squat-sitting positions. They may have domesticated dogs, pigs., and water buffaloes, as suggested by excavated , bones in Dabut sediments. However, the Dabutians still did not depend on the production of their own food. They were basically gatherers, collectors, and hunters. Changes in food strategy made the Dabutian the most developed group in early and middle Holocene in middle northern Viet Nam. They were the first to possess the characteristics of a Neolithic society.

Endnotes

1. The temperature changes depend upon altitude, latitude, and position of sites from the sea. Normally, the higher the altitude and latitude or the more far off the sea, the cooler the tempreture. Rind and Peteet (1985) point out that around 18000 bp, the mean temperature in New Guinea was below 27 DC (nearly sea surface temperature) to 22.4 °C (surface air temperature) to 11 DC (at 2.5 km temperature) and required a surface 4 km lapse of7.5 DC per Hoabinhian Food Strategy in Viet Nam 457

1000 m, while Flenley (1985), based on the pollen data in different altitudes of New Guinea highland supposed a lapse rate of 0.8 °C per 100 m at 18000 BP. The standard lapse rate at . p~esent is 0.5 °C (in Taiwan) or 0.6 °C per 100 m (in Viet Nam, Southern China, and New Guinea)-see Tsukada 1966; Zheng and Lei 1999; Tum and Lin 1978; and Flenley 1989. Around 18000 BP, Nguom rock shelter and Xom Trai Cave around 400 to 500 km far from sea (at -100 m isovath).

2. Zone M existed in the time span of >47000 to circa 12000 BP.

3. This fragment is stored in the Museum ofVietnamese History, Hanoi.

4. The earlier Canarium spp. discoveries were reported in Sri Lanka with a 14C date of more than 12000 BP. This probably means that the Holocene began earlier in places southward to the equator.

5. In Vietnam, I have not yet found evidence for the warm phase of 28000 BP as it has been mentioned by Flenley (1985) and some Chinese researchers.

6. Ethnoarchaeological research confirmed that the land snail Cyclophorus spp. and rock crab Ranguna can be collected only in cool and moist conditions. Today, they go out only when it is cool (daily, early morning or evening) and moist (rainy season) . The collecting season is March-April-May and August-September-October. It is very difficult tb gather such food during June and July ( hot) as well as during November to February (dry).

7. Compare this to the Con Moong site, where only 10,000 shell remains were recovered, and at the Sung Sam Cave, where only 9,000 shells per one cubic meter sediment, and around ' 3 kg of mammal bone for each site were also recovered.

8. Many researchers recognize a broad-spectrum of hunting and gathering Hoabinhian strategy (Gorman 1971 ; Tan 1982). That was the basis of the Hoabinhian subsistence strategy, and it might be characteristic of all cases of the late Palaeolithic cultures in this region.

9. It is notable that fish remains as well as bivalves, occurred in very few quantities in the Hoabinhian deposits in north Vietnam. The greatest collection offish bones in a Hoabinhian site was found at Dong Cang Cave; however, it accounts for only 10 percent of the total weight ofvertebrate animal bone.

References

Aigner, ]. S. 1981. Archaeological remains in Pleistocen( China. Muenchen: Verlag ,C. H. Beck.. 458 Nguyen Viet

Albrecht, G. 2000. Recent Mani settlements in Satun Province, Southern Thailand. Journal of Siam Society 86: 161-99.

Bellwood, P. 1985. Prehistory ofthe lndo-Malayasianarchipelago. London: Academic Press.

Broecker, W. S., M. Andree, M. Klas, G. Bonani, W Wolfli, and H. Oeschger. 1988. New evidence from the South China Sea for abrupt termination of the Last Glacial Period. Nature 333: 156-57.

Chen Yijian, P. G., and Jiao Wengiang. 1985. Radiocarbon dates from East China Sea and their geological implications. Quaternary Research 24.

Chinh, H. X. 1989. Hoa Binh culture in Vietnam. Hanoi.

CLIMAP. 1981. Seasonal reconstruction of the earth's surface at Last Glacial Maximum.

COHMAP. 1988. Climatic changes of the last 18,000 years: Observations and model simulations. Science 241: 1043-52.

Du, H., and N. D. Tung. 1977. Pollen record of the Con Moong Cave. Khao Co Hoc l.

Dunn, F. L 1970. Culture evolution in late Pleistocene and Holocene SEA. American Anthropology 72: 1041-54.

Flenley, J. R. 1979. The equatorial rain forest: A geological history. London: Butterworths.

----. 1985. Quaterrtaryvegetational and climatic history ofIsland Southeast Asia. Modern Quaternary Research in Southeast Asia 9: 53-63.

Frenzel, B. 1968. The Pleistocene vegetation of Northern Eurasia. Science 161: 637-49.

Gegh, M. A., H. R. Kudrass, and H. Streif. 1979. Sea level changes during the last Pleistocene and Holocene in the Strait of Malacca. Nature 278: 441-43.

Glover, 1. 1977. The late Stone Age in Eastern Indonesia. WorldArchaeology 9: 42-6l.

Gorman, C. 1970. Excavation at Spirit Cave, North Thailand. Asian Perspectives 13.

----. 1971. The Hoabinhian and after: Subsistence patterns in Southeast Asia during the last Pleistocene and early recent periods. World Archaeology 3: 300c20.

----. 1977. A priori models and Thai prehistory. In Origins ofagriculture, ed. C. A. Reed, 321-56. The Hague: Mouton.

... Hoabinhian Food Strategy in Viet Nam 459

Higham, C. F. W 1977. Economic changes in prehistoric Thailand. In Origins ofagriculture, ed. C. A. Reed, 385-412. The Hague: Mouton.

Kajalem, D . 1989. Mesolithic exploitation of wild in Sri Lanka: Archaeobotanical study at the cave site of Beli-lena. In The evolution ofplant exploitation, ed. D. R. Harris and G. C. Hillman, 269-80. London.

Llnsheng, Z. 1984. Reconstruction of the climate of the late Pleistocene Ice Age in East China. In The evolution ofthe East Asian environment, ed. R. O. White. Hong Kong: University of Hong Kong Press.

Lin Jingxing, Z. S., Qui Jinbo, Wu Biaoyun et al. 1989. Quaternary marine transgressions and paleoclimate in the Yangtze River Delta Region. Quaternary Research 32: 296-306.

Long, V. T 1984. Hoabinhian mammal remains in Vietnam. Khao Co Hoc 4.

Long, V. T, and N. T Phong. 1986. Mollusc remains from Nguong Sediment. New Discoveries ofArchaeology in Vietnam, 1386.

Maloney, B. 1986. Environmental archaeology at Khok Phanom Di, Central Thailand. Circaea 4.1: 8-10.

----. 1995. Evidence for the younger Dryas climatic event in Southeast Asia. Quaternary Science Review 14: 949-58.

Meacham, Wed. 1978. Sham Wan, Lamma Island: An archaeological site study. Journal Monograph. Hong Kong Archaeological Society.

Morley, R. J. 1982. A Paleoecological interpretation of 10,000 years: Pollen record from Danau Padang, Central Sumatra, Indonesia. Journal ofBiogeography 9: 151-90.

Moser, J. 2001. Hoabinhian: Geographie und Chronologie eines steinzeitlichen Technokomplexes in Siidostasien. AVA-Forschungen Band 6.Linden soft:Koln.

Nakada, M., and K. Lambeck. 1988. The melting history of the late Pleistocene Antartic ice sheet. Nature, 333: 36-40.

Peterson, R. M. 1969. Wuerm II climate at Niah Cave. Sarawak Museum Journal 13: 34-35.

Pookajorn, S. 1986. Ethnoarcaheological research on the hunting group in Thailand in comparison with the Hoabinhian technocomplex. World Archaeology Congress, Southampton, 1386. 460 Nguyen Viet

----. 1988. Archaeological research of the Hoabinhian culture or techno-complex and its comparison with ethnoarchaeology of the Phi Tong Luang, a hunter-gatherer group of Thailand.

Prell, WI. 1982. Initial reports ofthe deep sea drilling project. Washington D . c.: Government Printing Office.

Pyramarn, K. 1989. New evidence on plant exploitation and environment during the Hoabinhian (Late Stone Age) from Ban Kao Caves, Thailand.

Rind, D., and D. Peteet. 1985. Terrestrial conditions at the last glacial maximum and CLIMAP. Sea-surface temperature estimates: Are they consisten? Quarternary Research 24: 1 22.

Shigeko, H. et al. 2001. Geomorphology of the Red River Delta and their fluvial process of geomorphic development, Northern Vietnam. In Long climate change and environment change ofthe lower Red River Delta, 71-92.

Sorensen, P. 1988. Archaeological excavations in Thailand. Scandenavian Institute of Asian Studies Occasional Paper No. 1. Copenhagen.

Stuijts, I. 1984. Palynological study ofSitu Bayongbong West Java. Modern Quaternary Research in Southeast Asia 8: 17-28.

Tan, H. V 1985. Late Pleistocene climate in Southeast Asia: New data from Vietnam. MQRSEA 9: 81-86.

----. 1998. Archaeology ofVietnam, Band 1- Stone Age.

Thin, N. N., and T. T. Thuong. 1987. Plant remains from excavations 1986-1987 in Hoabinhian caves Con Moong and Xom Trai. New Discoveries ofArchaeology in Vietnam, 1981.

Thuan, D. V 2002. Holocene shorelines in Vietnam.

Tiep, N. T. 1998. The Holocene sea levels in Vietnam. Geology 3.

Tsukada, M. 1966. Late Pleistocene vegetation and climate in Taiwan. Proceedings ofthe National Academy ofSciences 55: 543-48.

----. 1967. Vegetation in subtropical Formosa during the Pleistocene glaciation and the Holocene. Palaeography, Palaeoclimatology, Palaeoecology 3: 49-64 .

.. Hoabinhian Food Strategy in Viet Nam 461

----. 1983. Vegetation and climate during the last maximum in Japan. Quaternary Research 19: 212-35.

Ty, Luu. 1985. The ancient shorelines in Pleistocene and Holocene in Vietnam. Khao Co Hoc 2.

Verstappen, H. T. 1975. On paleo-climates and landform development in Malesia. Modern Quaternary Research in Southeast Asia 1: 3-35.

Viet, N. 1988a. Excavations at Hoabinhian Xom Trai Cave (North Vietnam). Second International Conference ofthe Association ofSoutheast Asian Archaeologists in Western Europe, Paris, 1988.

----. 1988b. Unterschiedliche Entwicklungen in der Steinzeit Nord-Vietnams. Continuity and Discontinuity in the Evolution ofMen, Wittenberg, Germany, 1988.

----. 1989. Radiocarbon Datierungen fuer die Ur- und Fruehgeschichte Nord-Vietnams, University of Berlin.

----. 1990a. Mollusc remains in the prehistorical sites in Vietnam. Khao Co Hoc 1: 33-67.

~--- . 1990b. The upper Palaeolithic rockshelter ofTham Khuong (North Vietnam) Radiocarbon and ESR datings and typology. XIV Indo-Pacific Prehistory Association Congress, Yogiakarta, Indonesia, 1990.

----. 1993. Hoabinhian environment in Southeast Asia. Human and Environment in the Palaeolithic (International Symposium), Neuwied, Germany, 1993.

----. 2000a. Hoabinhian plant remains. New Discoveries ofArchaology in Vietnam, 2000.

--- -. 2000b. New interpretations of Hoabinhian chronology. New Discoveries of Archaeology in Vietnam.

----. 2001 a. Further studies on the Hoabinhian archaeology in Vietnam. Journal of Southeast Asian Archaeology 21.

----. 2001 b. Micro studies of Hoabinhian archaeology in Vietnam. International Conference ofHundred Years Researching Vietnamese Archaeology, Hanoi, 2001.

----. 2002. Da Bat Culture: A special Hoabinhian adaptation during Middle Holocene sea transgression. Colloquium ofDabut Culture, Hanoi, 2002. 462 Nguyen Viet

Viet, N., and H. H . Nga. 1982. Re-excavation at XomTrai Cave (Ha Son Binh Province). New Discoveries ofArchaeology in Vietnam, 1982: 43-47 . .

Viet, N., G. Kohl, and]. Goersdorf. 1988. Klimasaenderungen von End-Pleistocene to Holozene in Nordvietnam. Applied and historical climatology, Elbingerode, Germany, 1988.

Viet, N.,-T. T. Cuong, and L. C. Lam. 2000. Seasonality in Hoabinhian food strategy. New discoveries ofarchaeology in Vietnam, 2000.

Wang, L. 1989. A quaternary sporo-pollen analysis of Zengpiyan Cave Site, Quilin. Acta Anthropologica Sinica 8: 69-76.

Yen, D. E. 1977. Hoabinhian horticulture: The evidence and the question from Northwest Thailand. In Sunda and Sahul, ed.]. Allen . London: Academic Press.

Zhang, D. 1984. Synoptic-climatic studies of dust fall in China since historical time. Scientia Sineca (Series B) 17: 825-36.

Zheng, Z., and Z. Q. Lei. 1999. A 400,00 year record of vegetational and climatic changes from volcanic basin, Leizhou Peninsula, Southern China. Palaeography, Palaeoclimatology, Palaeoecology 145: 339-62.

,. ARTICLE

WRITING ARCHAEOLOGY ANALYSES AND ARCHAEOLOGICAL ARGUMENTATION

Keith W. Kintigh

Keith W. Kintigh is a professor of anthropology at Arizona State University.

his document presents some ideas about writing that I’ve you have submitted the article for publication and the reader is developed over several years of reading term papers, the- a reviewer, you can imagine the consequences. ses, dissertations, and professional articles. It highlights T Consider an analogy. As a buyer of an airplane, I’m interested some common problems and makes some positive suggestions. in a machine that actually flies and its performance characteristics. I’m not interested in the design process that led to the airplane or in seeing all the designs that failed to fly. In an article, The Sequence of Analyses Mode of Argument you are selling a product (an argument for a conclusion), not a In presenting the results of a quantitative analysis, authors (and design process. As a consumer, it is not particularly important especially student authors) sometimes describe, analysis by to me whether you went through a lot of design, redesign, and analysis, the sequence of steps that they went through. In cari- unsuccessful testing, or whether divine inspiration led you cature, the model is: directly to a great result. It’s the final performance that counts. I doubt that the sequence-of-analyses mode is ever an effective Here’s my problem...My data set is...and here are my style of presentation. The problem is similar to the problem variables...First I did...analysis and it showed...Then I with site reports that describe the data room-by-room, hearth- did...and that showed...Then...Then...In conclusion, by-hearth, pot-by-pot, bone-by-bone, or whatever. In the first the reader can plainly see... place, my eyes glaze over almost immediately (I suspect that I The underlying idea is that, at the end, the conclusions will rise am not alone in this). Second, even with a serious effort, the out of the mist and will seem warranted by the analyses. This reader usually can not process the information for lack of ade- approach may be favored (perhaps unconsciously) because it quate guidance on what to focus on. In a quantitative presenta- shows the reader how much the writer suffered to achieve these tion, no one ever wants to read the results of eight cluster analy- fabulous insights. It also has the appearance of being even- ses at multiple clustering stages. (I don’t even want to read this handed, letting the data speak through the analysis rather than if you’re talking about my data.) This style of presentation is dif- having the author’s ideas imposed on the data. This dramatic ficult to follow because the reader never knows where the argu- pulling together of all the narrative threads is an excellent ment is going, and thus can’t focus on the critical aspects of the approach for a mystery novel, but not for a scholarly article. analysis. It puts an enormous burden on the reader to fully com- prehend the data and absorb the results of each analysis, in I hasten to add that this mode of presentation may well describe order to remain poised for the denouement. Most readers will a perfectly appropriate intellectual process that ultimately led to not accept this burden, and they should not be expected to. a reasonable conclusion. Doing analysis is an inherently messy business. Most analyses (and certainly all of mine) involve lots Suggested Mode of Argument of trial-and-error, sidetracks, and dead ends. I’ve never kept track of the numbers, but it would not surprise me if I discard- Ordinarily in an article (this applies to theses and dissertations ed 100 analyses for every one that appears in an article. as well), the objective should be to argue convincingly one or a relatively small number of points. At the outset, you may not In an article, however, one must display the wisdom of hind- actually know what those points are. However, by the time you sight in presenting the really interesting results. This compact are writing the last draft, you must have the key points clear in form is dictated partly by limitations of publication space but your head, and you must make them clear to your reader. Your more importantly by the limits of the reader’s patience and objective in writing should be to guide the reader to understand time. If the reader has to wait until page 23 to get to the won- your points and toward being convinced by the argument and derfully interesting result, the reader will likely never get there. If data that support them.

September 2005 • The SAA Archaeological Record 33 ARTICLE

The point of which you want to convince the reader may well not not know what to make of it, the reader probably will not know be your original hypothesis. That is, you might have started a either. By sticking to a few points in an argument with a strong- mortuary analysis intending to show that such-and-such a ly developed logical structure, the reader is more likely to really group could be fairly characterized as a complex chiefdom. follow where you are going. That said, you might include some However, your analysis may lead you to believe something dif- of the interesting but less-relevant analyses in an accessible elec- ferent or to believe that the evidence is ambiguous. A well- tronic archive that could be used and referenced by someone presented case that the evidence is ambiguous on an important who wished to explore the issue in more detail. issue can be an important contribution. In this case, the conclu- Once you have made your case, you need to attend to obvious sion you are trying to convince people of is that there is a balance objections. For example, let’s assume that you want to argue for of evidence on both sides of an argument or that the data are so clear-cut status differences in a cemetery that has considerable ambiguous that they do not permit us to rule out key alternative time depth. Showing large differences in the richness of graves explanations. that cross-cut age and gender would not be sufficient. An obvi- To make a convincing argument is not to say that you only pres- ous objection would be that the division of grave richness is not ent evidence that supports your point. Your responsibility is to due to social classes but to changing burial practices over time. fairly present the evidence relevant to your conclusions. While Even though you may have looked at that hypothesis and had you never present every possible analysis that bears on your good grounds for rejecting it, a brief argument with the relevant question, your responsibility is to present the critical analyses data would almost certainly be worth presenting. Absent that and let the chips fall where they may. If you want to convince the argument, readers will find your analysis not wrong, but uncon- reader of a particular conclusion despite apparently strong evi- vincing. The problem is knowing how many possible objections dence to the contrary, you should present your grounds for dis- to address. A lot of that depends upon the limits of the forum in missing that evidence. which the argument is being presented. Dissertations can address a substantial range of possible objections. Brief articles Your abstract or introductory paragraph should first sketch your can probably only address a few quite obvious ones. problem and an outline of your conclusions. The idea is to cap- ture the reader’s attention and make them want to read the rest. Papers presenting quantitative results that I consider successful Ordinarily, you would then state and justify the problem and tend to follow this general formula. Remember, the idea is to provide any necessary background. Next, you might present bring quantitative methods to bear on a substantive problem, not the what you are going to conclude in more detail and outline how other way around. I’m not trying to cram everyone into a writing you are going to get there. The idea, of course, is to assist the straightjacket here. There is a great deal of latitude within the reader in following your argument in every way possible. general framework I am suggesting. But at each stage, it should be clear why you have done the particular analysis, what you If you are up front about it, there is nothing wrong with pre- were looking for, and what it in fact shows. While it is appro- senting your argument in a quasi-deductive style, even though priate to present analyses that both support and disconfirm your your thought processes went the other way. The argument will hypotheses, it is not appropriate to present analyses that, how- often be of the form: “Here is my conclusion and here is the evi- ever interesting, do not contribute to the specific arguments of dence that supports it.” I am not suggesting that you mislead your paper. people about what you are claiming. Indeed, I’m happy for you to admit that these conclusions were developed inductively. You The Professional Audience might, in a paragraph, sketch the kinds of analysis that you tried in getting to your conclusions. However, if you do this, make If you are writing to a general professional audience, it is prob- sure that when you are done, it really contributes in some way ably reasonable to assume that the reader understands percent- to the paper. If not, lose it. From there on, you should be pre- ages, correlation, simple probability, how to read a contingency senting an argument that directly relates analyses to your con- table, and perhaps a bit about principal components (or factor) clusion. It just works better this way. analysis and cluster analysis. When you use less-common but still fairly standard techniques like discriminant analysis, multi- In the course of writing, you will usually find that to really make dimensional scaling, or correspondence analysis, you should the point convincingly, you will need to do additional analyses to remind the reader in one or a few sentences what the technique plug logical holes in your argument or to provide additional con- is doing for you. If you can’t state it succinctly, you probably firmation or disconfirmation. This happens to me all the time. don’t understand it well enough to be using it. If you use more With the suggested mode of presentation, it is almost always the obscure methods, such as Local Density Analysis or Koetje’s case that you will present a small fraction of the analyses that analysis of concentration, you must explain them sufficiently you executed. Each analysis that you present should be relevant that the reader can follow what you are arguing. If it is impor- to your argument. If some result looks interesting, but you do tant enough to use the method, it’s important enough to explain

34 The SAA Archaeological Record • September 2005 ARTICLE

it fully. My review article on spatial analysis (Kintigh 1990) does this for a number of methods. THINGS TO REMEMBER ABOUT No matter what analysis you use, make clear what variant you used (e.g., stepwise or simultaneous multiple regression). Make STATISTICS (WHATEVER ELSE YOU clear what variables you used and why you selected them and what FORGET) cases you used and why you selected them. Also make clear any transformations of the data that you did. Note that “standardize” George L. Cowgill has a marked and an unmarked sense; if you mean percentages, say so and say what they are percentages of. If you mean z-scores, George L. Cowgill is a professor of anthropology at Arizona State University. say you did a z-score standardization or something equivalent.

Precision and Completeness 1. Statistical analysis is not a way to arrive at certainty; it is a powerful aid in discerning what your data suggest, and how strongly they Writing with precision is essential. This is not any more true for suggest it. This is often done better by an estimation approach than quantitative arguments than for nonquantitative ones. Don’t use by hypothesis testing. “demonstrated” to discuss arguments that are better character- 2. Look at your data first, through simple tables and pictures. Often ized by “we have a slight hint that...” Nontrivial claims of this tells you everything important. If not, it will tell you what is sen- “proof” are rarely if ever justified in archaeology. sible or not sensible to do next. Do not rush to apply advanced techniques while overlooking the messages of simple methods. When you present tables and figures, it is your job to help the reader as much as possible. Make sure that table rows and 3. If you must do a hypothesis test, report the actual probability level columns and figure axes are adequately labeled. It is never obtained and don’t treat some arbitrary level such as five percent as a acceptable to have cases or variables numbered from 1 to N in talisman that tells you what to think. tables and figures only accompanied by a key. Doing this forces 4. It’s not the sampling fraction that matters; it’s the size of the sam- the reader to take many extra steps just to follow your argument. ple. For example, a well-chosen sample size of 100 that is one percent of a large population can tell you a lot, but a sample of 10 that If you are discussing data where percentages are the relevant is 20 percent of a small population tells you less. concepts, do not provide a table of counts. Instead, give a table of percentages with a total N on which each row or column of 5. One can wring useful suggestions from very small samples. How- percents is based. If you are trying to show differences, think ever, at least for artifact collections, it is usually desirable to get hard about what graphic would best make the case. Most people 300–400 cases if possible; smaller collections are often too small for satisfyingly accurate estimates of interesting proportions or other are not good at seeing patterns in tables of numbers. Stacked properties. Do not trust the maxim that 100 or so cases will general- box or dot plots, histograms, Ford diagrams, and scatter plots ly be enough. are simple devices that communicate well (although box plots still require a brief explanation for some audiences). 6. Proportions, percents, and ratios represent something relative to something else. Proportions are fractions, with a numerator and a In general, and especially in theses and dissertations, if you are denominator. When you write, always report the denominator. When presenting significant data, if it is at all possible, provide an you read, always ask yourself whether you understand what denomi- appendix with the raw data so others can re-analyze them or use nator is implied. Often you will find that the denominator is unclear them for other purposes—and by all means, make the data or inappropriate. available electronically. You have nothing to hide, right? 7. “Frequency” should always mean the count of something, rather Acknowledgments. This article began as a class handout several than the ratio of something to something else. It is often used to mean both. years ago. George Cowgill’s suggestions, some of which I have paraphrased, have greatly improved this article. Over the years, 8. If you are worried about data quality, reducing data to I have incorporated useful suggestions from other colleagues at “present/absent” only makes the problem worse unless you are sure Arizona State University; I regret that I have lost track of the that absence in the sample unambiguously implies absence in the specifics and cannot credit them individually. relevant population. But a category that is scarce but present in the population will be totally absent in many random samples from that References Cited population, and the chance that it is absent in any one sample is strongly dependent on the size of that sample. Together with sam- Kintigh, Keith W. pling vagaries, this makes “presence/absence” a very unstable statis- 1990 Intrasite Spatial Analysis: A Commentary on Major Methods. tic. If you want to be conservative, use something like “way below In Mathematics and Information Science in Archaeology: A Flexi- average,” “about average,” and “way above average.” ble Framework, edited by Albertus Voorrips. Studies in Modern Archaeology 3:165–200. HOLOS-Verlag, Bonn.

September 2005 • The SAA Archaeological Record 35 Peter White

15 Producing the Record

A discovery dates only from the time of the record of it, and not from the time of its being found in the soil. Pitt Rivers (1898: 28)

Introduction

I recently received a book catalog entitled Archaeology: Method, Theory and Practice. Among the four pages of books on method, there were books about survey, sampling, excavation, dating, GIS, electronic databases, and other things. Elsewhere, there were books on interpretations, public outreach, museum presentation, and working with communities. But there wasn’t a single book on the nuts and bolts of writing anywhere in the 24-page catalog. I found this surprising. Most of the end results of archaeological research, its reporting, reasoning, and interpretation, are produced as writing, either in hard copy or electronically. It is by our written work that the public, our teachers, and our peers evaluate us. So we should be paying attention to how we write and how we present our ﬁndings and results. The purpose of this chapter is to help you move from ﬁeld, laboratory, and library research and analysis to a written account of aspects or all of it. It presents some simple guidelines concerning organizing the structure of written works and using effective language. This chapter is about how to think about and present the content of your work. It is not about the philosophy of archaeology, how to interpret your material, or the technology of publishing. I have tried to focus on what will be useful to the most likely readers of this chapter, namely undergraduate and postgraduate students. But even senior professors need to think about the organization and language of their writing, so I hope some of them will also read this.

410 PETER WHITE Before starting to write, ask yourself: First Decisions

1 What is the purpose of this writing? What do I want to write about? 2 Who am I writing for? Who is my audience?

The answers to these questions will determine what goes into the piece – and what gets left out.

In most student contexts, what you are writing about will be determined by What do I want to the question or topic set for you. So the first rule is read the question! Read it write about? and make sure that your planned paper covers what is being asked. In other situations, you will have more scope to define your own topic. It may be a report on an excavation or survey, an analysis of data, or an interpretation of a period, place, or phenomenon. But you need to start out by deciding what you are writing about. Unless you are simply reporting what you did on a day-to-day basis, every piece of writing needs a point. What is the basic message you are trying to communicate? Is it that this site covers a certain period, that one argument about this pottery type is right and another wrong, or that you favor a religious rather than an economic interpretation of this city layout? To anyone who has been working on a problem or data set, this advice may seem rather curious. You have been slaving away looking for sites, measuring artifacts, constructing a philosophical position, counting bones, reviewing the origins of agriculture or whatever, so of course you know what you’re going to write about. But if you just start writing it is actually surprisingly easy to stray away from your topic. There are sidelines that you find interesting, or you perceive the need to qualify or explain some aspect that is not really central to your theme. Novelists say that they are often surprised by how some minor character develops importance as a story proceeds; the same can be true in archaeological writing. Conversely, it is not difficult to forget aspects of the topic that are really relevant to a full understanding of the data or argument. So answer this question first. Write out a plan of what this piece of writing must contain.

Deciding who you are writing for is particularly important for determining Who is my both the length of the piece and the style of your language, but to some extent audience? it also affects the content. You should consider the genre within which you should write. An essay or thesis will need to be written in academic language, explaining your methods fully, using technical terms as appropriate, and documenting your statements with full referencing. A report to a local community that has given you permission to work with them or on their land will need to be in plain English, with technical terms avoided or explained. Reporting the

PRODUCING THE RECORD 411 results of contracted research to a local authority or private developer will require clearly spelled out recommendations for (in)action, which you can justify on the basis of the research. In each case, the approach and language will be different and you should decide how you are going to write before you begin. What readers get out of your text is, of course, only partly within your control. Every reader brings his or her own knowledge, attitude, and viewpoint to any text. This conditions what a reader makes of your work and how they use it. But by thinking about your intended audience, you can try to make it user-friendly for them and less liable to misinterpretation of what you want to tell them.

Structure Half a century ago, the famous English archaeologist Sir Mortimer Wheeler wrote that “Plain and effective writing requires a plain and effective structure” (1956: 215). He was writing about excavation reports, but the structure that he advocates for these has been used widely by academics, especially most scientists, for much of the past century and is advocated by all the main writing gurus. While this formula is most applicable to scientific writing, it can be modified and simplified for almost any kind of archaeological writing, since it embodies a basically narrative structure (see, e.g., Pluciennik 1999). The formula is as follows: • introduction (aims) • background • methods • results • discussion and conclusions.

This basic formula clearly separates actions from thoughts, discoveries from deductions. It organizes the material logically so that a reader can understand how the work proceeded. On the other hand, it lets a reader who doesn’t necessarily want to read the whole piece use it to satisfy his or her particular requirements. Using this formula effectively implies that you should divide your work into appropriate sections, each with a heading. For long pieces, you may also need subsections. Don’t be afraid of these. Section and subsection headings help to guide readers through your text, showing where they have reached the end of one topic and are embarking on another. I will now discuss each part of the structural formula. This discussion is oriented toward report writing, but the basic structure can be adapted to almost anything more organized than stream of consciousness. The most usual modiﬁcation among archaeologists will be to say little about methods and to combine results, discussion, and conclusions. This might be expected in a paper on the origin of language or problems of ontology in archaeology. You are not discarding the formula but modifying it to suit your needs.

412 PETER WHITE This should tell the reader broadly what this piece of writing is about: Introduction what problem was being researched, what you are aiming to demonstrate, and what conclusions were reached. The introduction is the most difficult part to write and will usually need to be totally rewritten after you have finished a first draft of the whole work. This is because what you thought you were going to say when you started often isn’t quite what you end up saying.

Here, you review the theoretical and/or practical situation when you began Background the work. If a bulldozer had dug up part of a skeleton, which you then excavated, what had happened before your arrival? If your study is about seasonality in shell middens in Sydney, or California or the Orkney Islands, what did we know about the general problem and work in the speciﬁc area when you began? For a piece on archaeological ethics, you might need to outline the ethical code of the World Archaeological Congress and refer to codes of other disciplines. The point of this section is to identify the relevant context for your research. It is not to display the full extent of your knowledge about the topic. Ask yourself: What does a reader need to know to understand the context in which I am working? Stick to the answer!

What goes in here should be obvious: say what you did and how you did it. Methods Where did you survey? How much did you dig? What recovery methods were used? What analyses were undertaken, using what protocols? In a scientiﬁc report, this section is where you demonstrate that you knew what you were doing; or, in a public report, it is where you explain how and why you went about it the way you did. This is the section that anyone who doubts your results or conclusions will read to check that your methods were appropriate and correctly used.

What did you find out? This section will need clear descriptive prose, as well Results as tables and figures, so that the correctness, strength, and importance of your results can be readily assessed. This is where other people will look for material for their research, to see whether what you found is consistent with what is already known and with what they found. You therefore need to be very careful to say what you actually mean. You also need to ensure that your story is consistent. Working on manuscripts as editor of an archaeological journal, I regularly find that numbers given in the text and tables don’t agree, distances or positions stated are not what the map shows, stratigraphies don’t match, tables are inconsistent in totals or units and, only too often, that people just can’t add up! Errors of these kinds may be simple slips of the keyboard, but they don’t inspire confidence among readers in your research.

PRODUCING THE RECORD 413 Conclusions This part of the document will depend a great deal on the overall purpose and audience. An essay, thesis, or archaeological paper will require your opinion about what has been achieved. This may include considerable discussion about what you understand by the results and their implications (Where do we go from here?). A mitigation (contract) report will include a series of recommendations for action, phrased in plain English. Attached to the basic structure there are usually three add-ons – an abstract, references, and acknowledgments.

The abstract An abstract summarizes the text. It is for readers who don’t have access to, or time to read, your text. An abstract is a short version of the paper, not a description of it – so avoid phrases such as “We describe . . .” or “We then analyzed . . .” It answers several questions: Why did you start? What did you do? What did you ﬁnd? What do the ﬁndings mean? (O’Connor 1991: 70). It should do all this in 100–200 words, as in the following example from a paper entitled “Japanese Jomon sherds in artifact collections from Mele Plain on Efate in Vanuatu” (Dickinson et al. 1999: 15):

Fourteen cord-marked sherds collected from the surface of yam gardens on Mele Plain near Port Vila on the island of Efate in Vanuatu are unlike any known prehistoric wares from Paciﬁc islands. Based on typological analysis, comparative petrographic study of sand tempers supplemented by microprobe analysis of characteristic microperlite sand grains and thermoluminescence dating of key examples, we believe these sherds were made during the Early Jomon period in Aomori Prefecture near the north tip of Honshu. Although we are unable to specify the time or means of their introduction into Vanuatu, the sherds cannot represent indigenous pottery and are not part of the Oceanian ceramic legacy.

References Unless something different is speciﬁed, a list of references (bibliography) should contain all and only those works included in your paper. Although there are now various computer programs for organizing these and presenting them in a consistent manner, the “GIGO” maxim still operates. If you put Garbage In, you will get Garbage Out. In other words, you are responsible for the original accuracy of the citations. Getting references right is not easy. You should check the accuracy of every reference – even your own publications. The reference that is wrong or left out in someone else’s work is always the one you want! Don’t do this to your readers.

Acknowledgments Thank those organizations and individuals who gave you substantial or timely help in the form of permissions, grants, materials, or advice. Especially in archaeology, it is highly unlikely you have worked all on your own and people like to be acknowledged, while institutions often require it.

414 PETER WHITE . . . easy writing’s vile hard reading Writing Sheridan (1771)

It is not easy to write simple and clear English. Texts that are easy to read have usually been difﬁcult to write and are the product of several drafts. Most writers who are known for their good writing, whether academics or novelists, tell of rewriting each paragraph many times – although, as with many skills, ability often improves with practice. It also helps if you have decided what you want to say and who you are writing for. There are a number of steps in the actual production of good written work:

1 Start by writing an outline of what you want to say. A one-page summary of your aims, methods, results, and conclusions will help you to work out what to write. 2 Next, just write it. Start at the beginning and write according to the plan. Don’t agonize over each word and sentence – you can do that next time round. Try to get down what you think is a reasonably complete draft. But perhaps you, like many writers (including myself ) sometimes get a “threshold complex,” when you just can’t start – when no sentence seems right. If so, start in the middle with the easy bits, which are usually the methods and results. You know how you went about your work and what you found, so write about those topics. Once that’s in hand, the discussion and conclusions should follow. Then you can go back and write your introduction. 3 Having got some words down, look at the structure of what you have created. Does it follow the plan you made? Does the introduction cover the whole paper? Are the methods and results in logical order? Does the discussion mirror the results? Shuffle sections, paragraphs, and sentences until the structure and content are clear. Make a note on the text of what is missing at that place. 4 Rewrite. In this second draft, pay particular attention to the first section – the aims. This is where readers will start, so it needs to encourage them to persist. It is also, in my experience, often the worst part of any piece, loaded with generalities, irrelevancies, references, and inappropriate personal detail. The first section should map out the ground ahead and the goal to be reached. 5 When you think you have written something reasonable, put it away, if you have time, for at least a week or longer. Do something else. Then re- read it. You will be surprised at how it has changed! What you thought was organized and elegant will have become badly structured and poorly expressed. Rewrite it again. Be ruthless! 6 This time, after rewriting, don’t bury it, but give it to several friends to read. Try to choose a range of people – a colleague who knows what you’re talking about, your mother who doesn’t, or a friend who you think

PRODUCING THE RECORD 415 writes well. Choose people who you can expect to be honest with you, and who you know will be prepared to put some time into reading your work carefully and critically. But be prepared for shocks. Your friends will almost certainly ﬁnd faults. Your reaction is likely to be hostile, to defend your writing right down to every comma. Don’t. Swallow your pride and listen to your friends. It is quite likely that they will be right.

When you think you have a moderately decent draft, there are three aspects that you need to consider specifically. These are your spelling and grammar, which I consider together, your use of quotations, and your documentation or referencing. First, now that most people use a word processor with spelling and grammar checkers, the number of errors in a piece of written work might be expected to have fallen. But I am not sure it has. It is clear that many writers, especially students, are very casual about their spelling and know little about grammar. This is sometimes justified by arguing for “free expression.” But poor presentation in these matters suggests that you are illiterate – and thus probably also incompetent. I suggest that you should never let any of your work escape into the hands of others until you have checked it for spelling and grammatical errors. If you know that you are weak in this area, get help from someone who is better trained. You should also realize that no spelling or grammar checkers can check for improperly used words. “For” and “four” are both correctly spelled, but have quite different meanings; the same is true of “today” and “toady,” “grave” and “gravy,” and many other words. You actually need to read your work before letting it loose on the world. This is often best done with a friend, one of you reading it aloud to the other. Boring and slow it may be, but it can save you a lot of embarrassment, especially if your typing or grammar is not perfect. Secondly, consider your use of quotations from the work of others. The use of long chunks taken directly from other authors, even with acknowledgment, should be avoided except in quite specific circumstances. Many students use lengthy quotes in the belief that “experts” have said it better than they can, but the point of student essays, theses, and academic papers is to demonstrate that you are sufficiently on top of a topic to write about it yourself. Of course you will have to draw on data and even opinions of other workers in the field, but what readers will want to know is your contribution and how you make use of work already done. These thoughts about quotations lead directly to the third topic, referencing. When you use information that originally appeared in other works, this needs to be acknowledged. Failure to do so can be seen as plagiarism – in other words, stealing. Among students and academics, referencing is necessary to avoid any suggestion of plagiarism as well as to show what you have considered in producing your work. There are various systems of referencing, and examples of the three common ones are given in the Appendix. The

416 PETER WHITE choice of system that you use may be dictated by your lecturer, department, or journal, but whichever it is, successful use of it should aim for clarity, accuracy, and consistency.

This chapter cannot cover the writing of good English – to do that would need Language the whole of this book. What I can do here is point out a few simple, obvious do’s and don’ts:

1 Does each sentence say what you intend?

One clue [to the disappearance of Neanderthal peoples] is that they lived chieﬂy during hard winters when death rates must have been high. (Pfeiffer 1969: 194)

This is simple and direct, but nonsense.

This powder was then placed into an 0.5 mm capillary which was sealed. (Blau et al. 2002: 814)

How did they place the powder into a sealed tube?

2 Is each sentence simple and clear?

None of these [the distributions of archaeological remains] can be understood, however, without taking into account the ecological nature of the habitat exploited by the toolmakers and it is especially important for the prehistorian to know what main plant and animal resources were present and so available for use by the hunting/gathering populations during the Pleistocene when the biomes may often have differed signiﬁcantly from those of the present day in response to ﬂuctuations in humidity and temperature. (Clark 1975: 606)

Why not:

None of these [the distributions of archaeological remains], however, can be understood without taking account of the toolmakers’ habitats. It is especially important to know what main plant and animal resources were available to Pleistocene hunters/gatherers. These resources may often have differed signiﬁcantly from those now present, since biomes respond to ﬂuctuations in humidity and temperature.

3 Are you taking responsibility for your work by using the ﬁrst person and the active voice?

• “During the excavation it was noted that the stratigraphy of layers 7 and 8 was dissected in places by several pits” is a complicated way of

PRODUCING THE RECORD 417 saying “While excavating, I noted several pits which cut through layers 7 and 8.” Similarly, I could rewrite Blau et al. (2002: 814) quoted above: “We sealed the powder into an 0.5 mm capillary.” • Using “It was done” instead of “I did” is not more objective. Rather, it’s a cop-out – a way of saying “No, it wasn’t me – I’m not responsible. What I did was dictated by some higher authority.”

4 Is each sentence complete?

• “Whereas the second sample of points came from layers 6 and 7” is not a complete sentence, because “whereas” shows that it requires the previous sentence to understand the contrast being made. • When you read each sentence it should make sense on its own, without requiring any reference to previous or following sentences.

5 Are you using the right words?

• I have been faced with people’s “rolls” (instead of “roles”) in “sedimentary,” rather than “sedentary,” communities, which were “effected,” rather than “affected,” by drought. If long words bother you, use a dictionary. • The Oxford English Dictionary, Merriam–Webster’s Collegiate Dictionary, and the Macquarie Dictionary are among some of the online dictionaries (see www.oed.com, www.m-w.com, and www.macquariedictionary.com.au).

Examples similar to the above problems can be found in many reports. You will usually ﬁnd that it is readers of your work, not you, who notice your bad writing. Having others read your work for you is the best way of avoiding these problems.

Writing for Much of what I have outlined above about structure and approach applies to Publication papers written for publication as well as to term papers or theses. But there are several speciﬁc aspects to submitting works for publication that are often overlooked or ignored, to which I shall now turn.

Target carefully In this case, deciding who you are going to write for involves the journal or publisher as well as the audience. Study the form, structure, and content of the papers in the journal in which you want to publish, or of the books produced by the publisher you are thinking of approaching. Is what you want to write compatible with their current output? If not, can you rewrite it to be compatible or would you be better off approaching someone else? Look around. Different venues work in different genres and you should consider your options within these genres.

418 PETER WHITE If you are trying to write an archaeological paper based on a thesis or report, Start afresh don’t think that you can just paste together chunks extracted from the previous work, or even necessarily follow its format. Put your previous work away and start again. Write it out of your head, following the steps suggested above. There will be time enough to get the data right afterwards, when you can go back to the original. But if you want a paper to be clear and convincing, it must be written as a separate piece of work.

Read, mark, digest, and use the Instructions to Authors of the journal or pub- Follow instructions lisher you have chosen. These will cover such matters as acceptable length of article, form of submission (e.g., disk or hard copy), bibliographic style, the acceptable number, size, and style of illustrations, presentation of radiocarbon dates, and so on. If you have a problem, look at recent issues of the journal, or consult the editor. Instructions to Authors can be very precise – those for American Antiquity cover 22 pages – but failure to follow them will put you immediately offside with the editor, who may simply return your manuscript, unread.

Plan pictures, tables, and graphs while you are writing. These will be rough Think about drafts, but doing this concurrently will help you to decide what goes in them illustrations and and what goes in the text. You need to decide what is the function of each tables illustration. Is it to present an impression or provide a large amount of information? Answering this will help you to decide how to structure it. Illustra- tions should not duplicate the text, but should vividly elucidate it or amplify the data. Students, in particular, often put pictures into their essays, but then don’t refer to them or use them. This is a nonsense – the pictures become just irrelevant decoration. As Moser (1998) points out, pictures are persuasive in their own right, so you need to ensure that they interact with the text, each adding to the other. Use them or lose them. As with the text, don’t even think that you can just pluck illustrations out of a previous work. Most graphics drawn up for theses or reports are on A4 paper and nearly always include very large amounts of unused space. Very few journals or books will allow you to be this wasteful: look at the illustrations that they publish. What you need to produce are graphs and tables that fit well onto a page or into a column of the publication. You can get a sense of this by photo-reducing your intended illustrations to the likely final size. This will also tell you whether your lettering is too small (the usual problem) or you have tried to put too much in. Your aim should be to produce graphs and tables that communicate. O’Connor (1991, chs 3 & 4) has useful advice about preparing tables and figures (see Figure 15.1). However, for really elegant graphic and map presentations that avoid “chart junk” and maximize the data– ink ratio, Tufte (2001) reigns supreme. Many computer-drawn graphics are

PRODUCING THE RECORD 419 (a)

Attacks Drug “B” 20

10 200 Weight Kgms g /mls

5 100 Plasma Con. µ 0 0

01234 Months

(b)

20 Drug B

15 Plasma 10

concentration (mg/ml) 0 200

100 Weight (kg) 0 Frequency of attacks

1st 2nd 3rd 4th End of month record

Figure 15.1 Obscure and clear data presentation. (a) Several sets of data have been superimposed to save space; data are measured in unrelated units, unit labels are wrong, and arrows are unnecessary. (b) The same data, separated into three components and using no more space. (Reprinted from O’Connor, 1991, Writing Successfully in Science, HarperCollins Academic, with permission.)

clumsy and inelegant, sometimes even obscuring the data (Figure 15.2). Re- member KISS – Keep It Simple, Stupid – to help you to choose the best way of presenting your data consistent with your aims in presenting them. Illustrations such as maps and photographs should be essential to the paper. Use them to make basic or key points clearer, especially those that are hard to describe in words. Excellent examples are descriptions of making stone tools or pottery, where a set of clear pictures can replace mounds of hard-to-read text. Similarly, a paper discussing the difference between shell mounds, middens,

420 PETER WHITE Artefact count 120

100

0 P6 P4 xxx xxix P2 xxxi xxxii xxxiii xl xxxv xxxiv xli xxxvi xlii

xxxvii Time periods xxxix xxxviii Trenches

Figure 15.2 This is probably one of the world’s ugliest and most misleading graph forms. I suggest that you should never use it. It crams in a great deal of data, but note how hard it is to work out the actual value of a particular count. For example, the gray column in xxxiii/P3 is actually a higher count than the white column in xxxiii/P6, while the tall black column xxxiii/P4 is of lower value than the gray column xxxv/P3. It is also easy for tall columns in front rows to obscure shorter columns in the rear rows (created in Microsoft® Excel by Trudy Doelman). and cheniers will usefully use a photograph and/or drawing of each, which will bring out their differences. In the case of maps, consider what the reader needs to know: For instance, are all the contours or stream courses really necessary? Once again, KISS is a good guide.

Your bibliography should contain all and only those items referred to in your Reference paper. Present them in the precise style chosen by the publisher. Journals and efﬁciently book publishers are often quite fussy about this. If their style is not the same as the one used by your bibliographic computer program, well, tough: you will have to change, for they won’t (see Appendix: Referencing).

Proof reading is critically important in publication. A work will come to the Read the proofs editor or publisher on ﬂoppy disk, CD, or by email. Translating it onto paper, carefully and especially formatting into a journal or book, allows errors to occur. This is notably the case with symbols such as Greek letters or unusual abbreviations. But lines or blocks of text can be lost or transposed, captions wrongly placed, apostrophes re-symbolled, and so on. The reason is, of course, not “computer error,” but human error. It is people who use the formatting programs, decide

PRODUCING THE RECORD 421 on the layout, and push bits of a document around. So even if you send off a perfect document (unlikely!), you need to check it through thoroughly when it comes back to you in proof. By proof stage it is too late to correct the errors that you overlooked earlier, unless they are very small-scale and can be changed without disturbing the layout. But at proof stage you can correct technical errors, and you should use standard proof markings and follow any editorial instructions as to pen color and so on. I give some of the most frequently used prooﬁng symbols at the end of this chapter.

Conclusion Writing is by far the most frequent method of communicating archaeology. Even if your ﬁnal output is a talk or lecture, video, CD, or TV program, the structure and organization will have involved writing. Clear, logical writing is the pathway to an intelligent and intelligible result. The development of a well-written piece involves understanding the steps to be taken. First, you must know what the point of the piece is – what you want to get across. At the same time, you need to think about your readership and how to make your writing suitable for them. Next, plan the structure. Aims, background, methods, results, and conclusions is a well tested formula, although it can be modiﬁed to suit your particular circumstances. Be clear to yourself why you are modifying it. Then write, rewrite, rest, try out on your friends, and rewrite. Check your logic and language, your grammar and spelling, and your references for clarity, accuracy, and consistency. Plan your tables as you go along, so that they illuminate your text: a picture or graph may be worth a thousand words, but you have to use it to make it so. If you are writing for publication, follow the instructions – because all else fails. Finally, as with any sport or skill, good writing requires practice, as well as help from trainers and courage to admit that you can do better. No pain, no gain.

Acknowledgments I thank the editors for inviting me to write this chapter, of which this is the third version (amended, with thanks to the editors); thanks also to Robin Torrence for helpful comments (second draft) and Trudy Doelman for graphs.

Appendix: getting Most of the world now uses SI units. Even where this is not so, most academics things right and many businesspeople work with them. The three base SI units are the meter (m), the kilogram (kg), and the second (s). Often, multiples of these, SI units such as the millimeter (mm) and the gram (g), are more appropriate working units. Units should be used consistently within any piece of work – don’t jump

422 PETER WHITE between meters and millimeters, or between grams and kilograms. Decide which is best for your work and use it. The symbols for these units are set by the International Organization for Standardization and should be used in the proper way, such as “kg” rather than “Kg,” and “g” rather than “gm” or “gms.” Symbols never take a plural (“gms” is wrong) but names do (“grams” is ﬁne).

There is no absolute international standard for the presentation of radiocarbon Radiocarbon dates (14C) dates, but leading international journals such as Antiquity and American Antiquity largely agree on the information needed and the form in which this should be given. These forms are derived from the resolutions of the Inter- national Radiocarbon Conference and they should be what we all use. A common format will allow accuracy to be checked and improvements in the radiocarbon dating process incorporated. Full and accurate presentation of known information is especially important when dealing with dates produced over past decades, since these have different levels of precision and accuracy, and may well need revision. Start with the uncalibrated age supplied by the laboratory. Give the laboratory abbreviation and sample number, the age in years bp, the one sigma (symbol = σ) standard error, and the material dated. The abbreviation “bp” means “before present”: international agreements set this date at ad 1950 and this is what the laboratory will use. Never change the bp date supplied by the laboratory. Calibrated dates should now always be used for interpretation and discussion, except for ages older than 24,000 years (Stuiver et al. 1998). Calibrated dates are determined by using a calibration program that gives results in terms of date ranges along with associated probabilities. Calibrated dates are solar years and must therefore always be presented as cal bc/cal ad (cal bce/cal ce is also acceptable). The calibration program used (e.g., CALIB 4), the probability for each date range, and the material dated (if not already stated) should be given, along with any necessary corrections made, such as the Southern Hemi- sphere offset.

• The author–date (Harvard) system. “Russell’s (1990) analysis and my fieldwork Referencing (White 1993, 1994) have shown that the data are incomplete.” The references are given in alphabetical order of author, as in this book, although the format used here is closer to the Vancouver standard form. • The documentary-note system. “Russell’s1 analysis and my2 fieldwork have shown that the data are incomplete.” In this case, each numbered reference will contain as many publications as necessary (and may contain ex- planatory notes). • The Vancouver system. “Russell’s1 analysis and my2,3 fieldwork have shown that the data are incomplete.” The reference list will give each publication in numerical order.

PRODUCING THE RECORD 423 Prooﬁng symbols Symbol Meaning What to do

Add text here Write the symbol at the location in the text; write the text to be added in the margin with the symbol

Delete text Put a horizontal line through the text to be deleted; put the symbol in the margin

Change letters or text Put a slash through the letters to be changed; write the changes followed by “/” in the margin

Change to capitals Underline the letters to be changed; write “caps” in the margin

Change to lower case Put a slash through the letters; write the symbol in the margin

Transpose letters or words Use the symbol looped under and over letters or words to be exchanged; put the symbol in the margin

Close up Write the symbol at the space to be eliminated and in the margin

Change punctuation to this Put a slash through the incorrect punctuation in the text; write : the correct punctuation encircled in the margin (in the example, the new punctuation is a colon)

The complete range of prooﬁng symbols is given in the Style Manual for Authors, Editors and Publishers (2002) and the Chicago Manual of Style (1993) (see below).

Further reading You will find a range of books about writing in any large library or bookstore. There are books on writing science and popular science, writing in psychology, the humanities, or fiction, and on how to write a report or a term paper. I think that the most useful, basic, general account is O’Connor (1991). What she has to say is applicable to any kind of scientific writing and most archaeological writing. There is, however, a developing range of forms of writing archaeology that I have not discussed here at all. Approaches such as multiple stories in multiple voices and narration in fragments are enlivening the field considerably, as Joyce (2002) demonstrates. An excellent example is Schrire (1995). However, you will notice that such writing still needs structure, clarity, and precision. For much more detailed and precise advice on technical aspects of writing (especially books), from when to use parentheses to proof reading, the Chicago Manual of Style (1993), now in its fourteenth edition, continues to be the basic guide, especially for readers in the United States:

• Chicago Manual of Style, 14th edn. Chicago: The University of Chicago Press, 1993.

424 PETER WHITE The Chicago Manual is rivaled in clarity and comprehensiveness by the Style Manual for Authors, Editors and Publishers (2002). While its chapter on the legali- ties of publishing is directed to Australian conditions, in all other aspects it is internationally applicable. It is also elegant and easy to use:

• Style Manual for Authors, Editors and Publishers, 6th edn. Sydney: John Wiley, 2002.

Blau, S., Kennedy, B. J. and Kim, J. Y. 2002: An investigation of possible ﬂuorosis in human References dentition using synchrotron radiation. Journal of Archaeological Science, 29, 811–18. Clark, J. D. 1975: A comparison of the Late Acheulean industries of Africa and the Middle East. In K. W. Butzer and G. L. Isaac (eds.), After the Australopithecines. The Hague: Mouton, 605–60. Dickinson, W. R., Sinoto, Y. H., Shutler, R. Jr et al. 1999: Japanese Jomon sherds in artifact collections from Mele Plain on Efate in Vanuatu. Archaeology in Oceania, 34, 15–24. Joyce, R. A. 2002: The Languages of Archaeology. Oxford: Blackwell. Moser, S. 1998: Ancestral Images: The Iconography of Human Origins. Ithaca, NY: Cornell University Press. O’Connor, M. 1991: Writing Successfully in Science. London: HarperCollins Academic. Pfeiffer, J. 1969: The Emergence of Man. New York: Harper and Row. Pitt Rivers, Lieutenant-General 1898: Address to the Archaeological Institute of Great Brit- ain and Ireland. In Excavations at Cranborne Chase, near Rushmore, on the borders of Dorset and Wilts, vol. IV. Privately printed, 5–30. Pluciennik, M. 1999: Archaeological narratives and other ways of telling. Current Anthropo- logy, 40, 653–78. Schrire, C. 1995: Digging through Darkness. Johannesburg: Witwatersrand University Press. Stuiver, M., Reimer, P. J., Bard, E. et al., 1998: INTERCAL98 radiocarbon age calibration, 24,000 – 0 cal B.P. Radiocarbon, 40, 1041–84. Tufte, E. R. 2001: The Visual Display of Quantitative Information, 2nd edn. Cheshire, CT: Graphics Press. Wheeler, M. 1956: Archaeology from the Earth. Harmondsworth: Penguin.

PRODUCING THE RECORD 425