tcablogmaster has long wondered if more information could be gained from the intriguing sequence of layers revealed by our excavations at Millhams. These layers represent the deposition of silt and sand following regular inundation of this marshy site by the adjacent river Avon, as revealed by the inclusion of tiny water-living mollusc shells in the layers. One analysis technique used to determine the presence of human occupation is phosphate analysis, since phosphate levels are high in human and animal waste, the inevitable accompaniment to human settlement. Since we are looking for a suspected house or dwelling on the site, this seemed an appropriate technique to try in order to see if variations in phosphate concentration as a function of layer depth or location could be deduced.
The first experiment involved a garden soil test kit, using a makeshift colorimeter app in a mobile phone to try to determine the comparative concentration of phosphate in a sample of water used to extract phosphate from 10ml sample of soil. Unfortunately the colour changes produced seemed to be very difficult to compare with the supplied colour chart and this method was discarded.
A second attempt at a qualitative test was an implementation of the so-called Eidt Ring Test using instructions on the SASSA website (Soil Analysis Support System for Archaeology) at http://www.sassa.org.uk/index.php/Analytical_Methods:Phosphate_Qualitative . All the chemicals required – hydrochloric acid, ammonium molybdate and ascorbic acid – were available on ebay and using the recipe, tcablogmaster concocted Reagent A (ammonium molybdate dissolved in dilute hydrochloric acid) and Reagent B (a solution of ascorbic acid in water). Deionized water readily available from motorists centres was used to make up the solutions.
Application of the test requires a tiny sample of soil (about 50mg) to be placed in the centre of a 90cm filter paper and 2 drops of Reagent A added. After 30 seconds, 1 drop of Reagent B is added. The development of the blue colour is then observed and the resulting pattern scored according to various, rather imprecise criteria, such as length of time for the colour to appear, its intensity and the radius of the ring. A number of tests were conducted on the 4 layers identified in Pit 52 – the topsoil plus 3 layers of silt or silty-sand – and the results are shown opposite. The topmost sequence shows the result of an onsite test – unreliable in many ways because it seemed that probably too much soil was used in the samples and also the filter papers kept blowing away! Subsequent test were done at home in the greenhouse with the bottom row probably the most representative since care was taken to measure the weight of the soil sample.
From these few results and in the absence of experience with the technique it is difficult to form any opinion of the relative phosphate levels. It looks as though the topsoil is the richest in phosphate it is difficult to say whether the difference is significant. Clearly the test is comparative but there is great difficulty of judging and recording the intensity of the reaction for the purposes of comparison – tcablogmaster would very interested in comments from anyone with greater experience of the technique!.
While preparing a report for our Annual General Meeting held in June 2017, tcablogmaster was intrigued by the statistics of our excavations at Millhams. During the period June 2016 to June 2017 the number of test pits completed exceeded fifty – actually up to 52 at the time of the AGM. At a conservative estimate that the dimensions of the each test pit are 1m x 1m x 1.5m, tcablogmaster reckons that’s in excess of 75 cubic metres of earth removed and then replaced. With consolidated topsoil generally reckoned to weigh about 1.7 metric tonnes /cu. m. so that’s over 130 tonnes of soil which we have moved twice, out and back.
A reminder to our members that the 2017 AGM of our society takes place on Wednesday 28 June 2017 in the Harbour View Room of Stanpit Village Hall in accordance with the notice sent to all members by post our email.
We have now turned our attention to the north of the Millhams garden area in which we have been working, with the excavation of two further test pits, 51 and 52. Pit 51 turned out to be largely inconclusive and has now been filled in although we have sampled the various layers revealed with a view to some further analysis. Pit 52 has been excavated to a depth approaching 1.5 to 2m and has revealed the usual set of layers of silt, sand and gravel although at the lowest gravel layer we have found this to be overlain with a large numbers of fragments of water-logged wood which we interpret as belonging to a former revetment of the bank of a water channel. The photograph shows the collection of the fragments recovered (the red/white scale has bands of 100mm). None of the wood was found in place so we assume that the revetment, if such it was, had been allowed to collapse over time. The wood consist mainly of wooden stakes of softwood showing evidence of bored holes which may have been used to nail the stakes to some horizontal planking – a small piece oak hardwood was also recovered in isolation.
The gravel layer revealed a modest concentration of the usual collection of medieval pottery and bone which we have come to expect on site. However it has proved very difficult to bottom out the pit due to the speed at which water enters at the lowest level. The flow appears to be associated with the tide, and near high tide we need to run our pump and generator continuously to keep pace with the water flowing in. When not pumped the pit will eventually flood to a depth of a metre or so.
A few weeks ago Mike Gill carried out a gradiometer survey of the north of site and we were intrigued by the large magnetic anomaly which was apparent. Our first plan was to site a new pit on where we believed the anomaly to be but this revealed no significant metallic objects although yielding similar layering and pottery finds to other areas of the garden. We then employed a metal detector to look around the site of our new pit and under the edge of the spoil heap we did indeed get a strong response from the detector. The response was found to come from a a couple of large nails and a spike-like object about 0.5m long, all in the top soil with the spike being buried vertically with its top about 0.1m below the ground surface.
As such it would seem to be unremarkable except that on closer examination it has a small wheel about 20mm in diameter mounted by a rivet in a slot towards the end of the spike which has been flattened into a leaf shape. We are puzzled as to its purpose – any suggestions? – from its situation we would estimate that it is probably of recent origin (within last 100 years).
The map on the left shows the magnetometer results obtained on 12 March 2017 as processed by Mike Gill using Snuffler software, georeferenced and plotted on our Millhams mapping. The light blue background shows the nominal area of the area surveyed which is a 20m-wide strip running approximately north-south for about 50m. The survey results in grey-scale are superimposed on the blue background which shows through in areas where no results could be obtained because they were inaccessible or overgrown. In the middle of the patch, there were several beehives which are intrepid surveyor approached with caution since their inhabitants were quite active in spite of the chilly weather and became more so as the day warmed up.
The results showed little in the way of structure which we could interpret either as archaeological or geological features. One striking magnetic anomaly is shown at the north-east of the survey grid which we believe may indicate a large buried ferrous object which may repay some excavation to determine its nature.
As our first activity of 2017, member Mike Gill carried out magnetometer survey on 12 March of the north garden area at Little Millhams. This area becomes heavily overgrown later in the year, and we have found the bees from the hives located there to become quite aggressive when we approach too close, so carrying out the survey early in the year seemed a good plan.
We borrowed a magnetic gradiometer made available by Bournemouth University as part of the LoCate project, and held by the New Forest National Park Authority at Lymington. A number of our members have been trained in the use of the gradiometer and Mike Gill, pictured here in action at Millhams, has obtained impressive results from the equipment on other sites.
With much anticipation we were able to view the results on site without delay. Although there were no obvious rectilinear features suggestive of structures, some aspects of the data will repay further study and perhaps some tweaking of the grey-scale display. We were able to complete the best part of 3 x 20m grid squares in the garden, constrained by the topography of the site, the watercourses and various fallen trees which, all in all , constituted a very successful day.