ÇATALHÖYÜK 1999 ARCHIVE REPORT

Heavy Residue Archive Report

Slobodan Mitrovic

About Heavy Residue

The main purpose of this work is to show that results acquired from heavy residue analyses have an important place in the process of studying archaeological evidence.

Firstly, it would be useful to explain what the heavy residue (HR) sorting and processing looked like this year and point to several possible problems an HR involved archaeologist could encounter.

The whole procedure starts with flotation. After samples are dried out, they are then sieved through 4mm, 2mm and 1mm sieves. When we get three separate fractions, the next step is to sub-sample 2mm and 1mm fractions to obtain 25% (or less) of a fraction and process them faster and more easily. 4mm fraction is sorted 100%.

The actual sorting is entrusted to local women, who sort the material with amazing routine. They are efficient, precise, as well as careful enough to ask when something is not clear. Also, women who are new at the job adjust to the system very quickly. In that way possible mistakes are reduced to minimum and do not affect analyses.

Once a supervisor has filled in all the labels, bagged the sorted material and put it in a mother bag, along with a completed HR sheet, a sample is to be weighed. This is where one should be most careful, because the weights are the main data for further research. The weights are entered into the database in order to get ‘gram per litre of soil’ values (density) for each category of archaeological evidence. Then the task is to compare densities of different units (samples), and in such a way to create patterns.

In order to check whether sub-sampling provides accurate enough values, a random sample is selected (5134.2.3604). It is first processed, recorded and weighed in a standard fashion: the smaller fractions are 25% sorted. The remainder is processed, as well, so in the end, we have the actual weight data for the whole sample. It can be seen from the numbers below that the weight values for 25% of a sample do not always match those for a whole sample, since they depend on the sampling procedure itself.

Although smaller fraction measurements are accurate only to a certain degree and likely to present a problem in some cases, at this level of research they seem not to influence results.

The focus of the analysis that will be discussed here is the series of several superimposed units (layers) from the SE corner of space 170, building 17, South area. They form three different features – F. 538, 541 and 542 – that represent fire installations and associated rake-outs, or a "dirty floor". Both ovens and rake-outs are associated with the series of 5020 units, i.e. floors of the house.

Rake-outs show high concentration of various remains due to the fact that they are, actually sweepings, rubbish from the other parts of a house that have been kept clean. It is expected that the graphs made by comparing densities from rake-outs will provide information about dwellers` activities, which took place inside the house and therefore left certain trails.

Mostly obsidian and bone material are used to make histograms concerning 4 mm fraction. For 2 mm and 1mm fractions of a sample, plant is included into, as well, although what we get in heavy residue is material that does not float. Beads and flint are taken into account where they are present.

Of course, the first and most obvious pattern that emerges from a histogram is the distinction between a clean and a dirty floor. Now, the units that attracted our attention most and yielded the most important data were the rake-outs (5021.28.3035, 5021.29.3043, 5021.30.3023, 5021.301.3471 – associated to F. 538; 5034.2.3054 - F. 541; 5038.2.3049 – F.542 respectively).

According to the stratigraphy matrix it can be assumed that F. 542 was, after a "short while", succeeded by F. 541, which was then replaced by F. 538. The last one was longest in use, repaired during the period and more elaborate. Units 5040.2.3065 and 5041.2.3064 are also rake-outs, but have no associated fire installation. It would appear that in the beginning the dwellers had used the area just as a dirty floor or an improvised fire-place (both samples are rich in flint), and later made an oven. The physical characteristics of the floor of the house are such that it was merely sloping towards SE corner and perhaps that caused the area to become "dirty".

As far as the rake-outs are concerned, their sample volumes differ very much. Unit 5021 had 4 samples of 5, 24, 30 and 4 litres, unit 5034 one sample of 3 litres and unit 5038 one sample of 8 litres. The stray sample 5021.301.3471 was found at a later date, so it is not included in all the graphs. Furthermore, the bone density values for F. 538 are a little disturbed since bits of a scapula, lifted during the excavation, ended up in these samples.

The rake-out unit 5021, both through quantity of material and volume of flotation samples, confirms that it was used for a longer period of time. Probably, after the oven (F. 538) was abandoned, it continued to serve as a dirty area. Rake-out 5034 (figure 2a and 2b), on the other hand, although comparatively smallest in volume, is percentually very rich in obsidian, flint and bone material, which could suggest that the dwellers were more productive during this phase. The rake-out 5038 shows particularly small density of material for the rake-outs. Even the sample 5039.2.3053 (single unit representing F. 542) shows higher values than 5038.2 for bone material, suggesting that the area around the oven was kept considerably cleaner in this phase.

Beads have been found, with one exception (unit 5037.2.3045), only in the rake-outs. Whole beads are present, as well as broken ones, as if they had not been carefully looked for but just swept up. Flint is usually present in all the rake-out units, but 5034 is, by far, the most abundant in this material (figure 3).

What is usually found in the rake-out samples is as follows: bone, obsidian, shell, plant, eggshell, flint, beads, etc. These remains show what was swept, i.e. what was being done inside a house. Thus we know beads were produced, obsidian worked on, flint used to start fires and food prepared. These data represent what was left after certain amount of garbage was disposed of (Russell and Martin, 1998).

Full interpretation and understanding of the series of units discussed here is not yet possible, since not all them have been looked into by specialists. Undoubtedly, combining all the analyses we can get will provide more complete information. Still, the preliminary scan of botanical remains (made by A. Kennedy) from HR samples 5021.28.3035, 5021.30.3023 and 5021.301.3471, suggests various cooking incidents. One find from 5038.2 is particularly interesting: a nicely worked bone tool, which was broken and discarded as no longer useful. The abundance of flint in 5034.2 may point to a season or seasons when fire was indispensable, whereas 5021.30 contains no flint whatsoever. In the sample 5021.30.3023 an articulated paw was found that could have represented an amulet (N. Russell). The higher concentration of obsidian in certain samples (5021.30, 5021.3471, and 5034.2 stick out) may suggest that our dirty area was used as a knapping place or inform of a different tool production rate through time.

Among other uses of the data, in one particular case HR density histograms helped the excavators redefine the stratigraphy matrix. The sample 5073.23 (space 170, building 17) had higher values for bone and obsidian. This record showed that the sample belongs to another structural whole and not to the clean floor, as was originally supposed.

Conclusion

The results presented here show a small-scale analysis of HR data, which will probably allow a large-scale one. Hopefully processing and studying of HR material will be widely applied and used as a starting point for further research at Çatalhöyük.