Sherds were gathered from the surface of the midden several months before excavation. There is a slight bias in the surface sample away from entirely white sherds, as towards the end the sample was becoming too big and interest concentrated on the red- and blue-striped sherds.
The site for a 1m x 3m trench was chosen where surface concentrations of sherds had been high, and positioned so as to avoid vegetation and animal burrows where possible. The trench location was approved by the City of Edinburgh Archaeological Service, who visited the excavation on the first day to monitor it. Plate 11 shows the trench early on during the excavation, and its woodland location can be clearly seen.
Plate 11: The trench early on in the excavation, looking south down the hill, showing the stoney layer in the top half of the trench, which was plotted before being removedInitially a team of four people - the author, Ken Dawson, Ian Mitchell and Allison Naismith began the excavation of the site, with Bob Will supervising. The author undertook most of the subsequent excavation alone, as the midden turned out to be much deeper than at first anticipated, and the excavation was therefore extended in time. Ken Dawson, Agathe Dawson and Jonathan Riddell assisted at various times, and Bob Will made supervisory visits.
It was decided that structural finds should not be recovered from the excavation as they were not going to be analysed. A large quantity of plate glass was therefore left in the soil. At first every find was plotted in plan, but the number of finds soon made this impractical, and so grid squares were set up with a labelled bag for each, and only finds which were considered to be especially good examples of their type, or date indicators such as marked clay pipes, were plotted in 3 dimensions. Every find which was plotted in 3 dimensions was recorded with description in the Finds Notebook.
A layer of stones was encountered in the top half of the trench, and it was planned before being removed. In the bottom half of the trench, which was on a steeper slope, the ash was reached a lot sooner. It was this bottom half of the trench which was dug to the full permitted depth of 1m.
In spite of the efforts made to avoid animal burrows in the positioning of the trench, several burrows cut the trench, and, added to the fact that the ash was relatively unmixed with soil and non-compact, neat excavation proved difficult. Section drawings (see Figure 8) were made of two sides of the bottom half of the trench and the sections were also photographed (see Plate 12). Punctured plastic was laid in the bottom of the trench before it was filled in, to mark the point where the excavation had reached. This was done because the deposits were becoming richer in large pieces of pottery. Since this remained the best position for a trench on this extensive midden, it was considered that any further excavation may wish to carry on where this excavation had finished.
The study of the pottery from the surface had already suggested categories into which most of the sherds could be placed, according to decoration and use. Much of the initial work involved categorising the finds using the categories already established, so that they could then be studied.
Examination of the finds catalogues in the appendices will give an idea of the categories employed. For example, a sherd could be in the broad functional category of 'crockery', its body type would then be specified, e.g. 'earthenware', the position of the sherd on the original article would then be stated, e.g. 'rim sherd', the surface decoration would be given, unless it fitted precisely into a named category, in which case only the category name would be given. Any discernible or unusual measurements would be noted, for example the height of the article if known, and finally an opinion would be given as to what type of article the sherd had come from, e.g. tankard.
Of all the sherds in the finds catalogues, only those coming from the main midden, and which are either in the broad category 'crockery' or 'serving crockery' are analysed in the main part of this dissertation. They are the categories containing the most artifacts, and the other categories and sites will be covered briefly in the appendices. The crockery is also the most easily identifiable as having come from the poorhouse since the majority belongs to sets with similar stripe patterns, and two of these sets include the crest with the words 'Edinburgh City Poor House', although after the excavation there can be no doubt as to the fact that the midden material is all from the poorhouse.
Within the two categories mentioned, decoration was analysed to a large extent, and a chronology based on decoration was attempted. The radius of sherds was measured where possible. The final analysis was principally concerned with examining different numbers or proportions of various articles of crockery according to decoration or use. The analysis was undertaken in order to answer the questions set out in the Aims.
The positions of individual items of crockery with different patterns in the trench were examined in order to investigate the possibility of establishing a chronological sequence based on pattern. In addition to this, working assumptions were stated and calculations were made on the basis of these assumptions to project back numbers of different types of crockery patterns over time.
The most successful and productive collection of potteries around the turn of the century was in Staffordshire. More locally, Glasgow had many potteries, and there were also potteries at Portobello and Preston Pans.
Blue and white transfer printed pottery was very popular and generally affordable, particularly Willow Pattern, and also Asiatic Pheasants. Large numbers of earthenware articles were made in these patterns, such as plates, vegetable dishes and meat dishes. Even cheaper than the transfer-printed wares were sponge-printed wares. Invariably unmarked, they were also produced in a variety of patterns. It is clear from middens from around the turn of the century that striped and banded crockery was also very common, though it does not tend to survive due to its lack of appeal in terms of the pattern.
Slightly more expensive wares included porcelain, and crockery with gold enamel. Relief-moulded crockery, and hand-painted pieces were also available.
Earthenware normally received a firing in the kiln before any decoration was applied. Fired earthenware which is unglazed in known as biscuit. If decoration was put on under the glaze, it usually only required one additional firing after the glaze was applied, known as the glost firing. For colours which were unstable at high temperatures, and particularly for gold, decoration was applied after the glost firing, and the ware then received a third, lower temperature firing after the decoration was applied. Overglaze decoration is often known as enamelling.