Page 140 - Kennemerland VOC ship, 1664 - Published Reports

R. PRICEETAL.:THEKENNEMERLAND
pectively, or the unclassifiable category, so lines of the latter are clear from what is known
they could not be expected to cluster satis- factorily (numbers 10, 18, 29, 30, 62, 81, 118, and 136). In passing, it can be noted
that the similarity suggested above between ingot 10 and class 4 is not reflected in this analysis. Of the remaining 20 ingots, six come from the poorly defmed group containing classes 6 and 9 (numbers 23, 45, 53, 67, 90, and 104), this being the largest group rep- resented among these residuals. Among the other 14 misplaced ingots there are represen- tatives of classes 1A (numbers 2, 25, and 69), 3 (number 80), 4 (numbers 39 and 46), 5A (numbers 12, 16, and 24), 5B (numbers 28, 49, and 74), and 8 (numbers 21 and 86). However, after allowance is made for the inad- equacies in our data, and especially lost stamps on several ingots, together with the standard weaknesses of any cluster analysis, a failure rate of about 19%seems acceptable.
Apart from demonstrating the relationship between the main shape classes, which may prove to be of historical significance, the prin- cipal result of this cluster analysis has been to show the close correlation between shape and stamps, suggesting that the latter are connected in some way with production. They may have been applied soon after casting to identify batches, or they may have been added by customers, middle-men or final recipients, t o denote suppliers. In view of the variable frequency of the marks, it seems likely that they were in fact applied at a number of these stages, the general feature being a relationship to the source of the ingots rather than any other consideration (e.g. weight, quality, price, or intended destination). It should be obvious, in view of the lack of relationship between shape class and ingot weights, that there is no correlation between the latter and stamps.
of the process of wrecking; the ship stove in her bottom on the Stack, depositing most of the ingots at once, and then drifted into the South Mouth, dribbling out a few more ingots which landed just to the west of the Stack (Price & Muckelroy, 1977: 199-201).
Casual inspection of Fig. 13 shows no very obvious patterning of classes across the site, although there are a few striking minor associ- ations, such as the concentration of all three members of class 3 in the middle part of the narrow gully. In order to achieve an objective assessment of any significant information contained in this map, a formal statistical test was applied. The sea-bed distribution can be broken down into seven discrete elements; the 37 ingots at the eastern end of the main gully, the seven in the narrow section below that, the 37 in the wider middle section which follows, the 12 lying lower down beyond the gap, the four in the western extension of the gully, the five lying to the south-west, and the 17 in the pile to the north-west. The occurrence of
examples of each class in each of these areas was compared to that which might be expected if every class were equally represented across the whole site, using the Chi-Squared statistic. The results are given in Table 2, showing that in
12 of the 14 classes there was no significant deviation from that expectation.
The two exceptions are interesting; one is class 2B, valid at the 5% level, and the other is the unclassifiable ingots, valid at 1%level. The former have a marked concentration in the upper part of the main gully (numbers 3 7 , 9 7 , and 98) and within the pile to the north-west (numbers 11 and 111). This is the group which was also unusual in having a low average weight, but, apart from noting that it is consistently the odd one out, it is difficult to deduce anything substantial from this observation. The unclassi- fiable ingots are clearly concentrated within the north-western pile (numbers 4, 8, 10,and
The final attribute of this assemblage to be
considered is its arrangement across the sea-bed
off Stoura Stack, as shown in Fig. 13, which
may shed some light on the way in which it was
loaded on board the Kennemerland. Certainly,
these extremely heavy objects cannot possibly
have been subject to any disturbance or re-
ordering in the 300 years between deposition
and recovery, so their distribution can only this consignment, were added at the end as reflect ship-board location and the manner of make-weights, packed in the uppermost layers, spillage from the wrecked vessel. The basic out- and were thus the last to drop out of the
18), with only two representatives elsewhere, number 62 in the lower part of the narrow gully, and number 81 to the south-west. It might be suggested that these ingots, not being part of the main production batches included in
21