Anyone new to this blog or geophysics in archaeology is recommended to read the material on the “Geophysical survey in archaeology” page.
Firstly, apologies to anyone waiting on the Durobrivae report. I am trying to resolve a small problem with the GPR data and will post something soon. We did get some good results, if not quite so spectacular as the temple we found last year. The “tumulus” is proving very intriguing.
Back in April 2014 we surveyed part of a site in Little Hadham. We had always intended to go back, but never quite managed to get our act together. Last weekend we finally managed to plan another three days at the site, working in the field to the west of the road. We were mainly intending to undertake magnetometry (Fig. 1), but as we had enough people we also did some Earth Resistance survey (Fig. 2).
The three days were about as different as you could get. The first day was quite nice, the second day wet, drizzly and foul (we had to keep trowelling the wheels of the mag clear of mud) and the last day was absolutely glorious.
The results from the magnetometry survey were excellent (Fig. 3). The features are relatively subtle, however. The image below is clipped to +/- 1.5nT. In others words, all readings above 1.5nT are plotted solid black, and all readings below -1.5nT are plotted white. At Gorhambury, I clip the images at +/- 5nT. The pottery kilns are Verulamium have very strong values of -15 to +150nT.
As can be seen from Fig. 3 there are lots of mainly linear features, some very straight, and some quite sinuous. We are clearly dealing with a multi-period site. The faint striping running west-nor-west to east-sou-east are a result of the harrowing of the field, made more visible by the extreme clipping of the image. To make the discussion easier, I have labelled up the figure.
One of the first things to note is how different all the linear features are. The one indicated by blue arrows is quite straight and for some of its length, at least, very magnetic (-6nT to +11nT). The one labelled with red arrows is, however, very sinuous and only faintly more magnetic than the background (about +/- 1nT). That ditch seems to continue as indicated by the green arrows, which in places seems to break up into a series of linked “blobs”, either patches of more magnetic material dumped in the ditch, or perhaps pits within the line of the ditch.
The strength of the magnetic values is dependent on two things: firstly, the source of the magnetism. Soils may be strongly magnetically enhanced by burning or intense occupation, for example, or may only be weakly magnetically enhanced if they contain just slightly more rotted organics than the background. Secondly, size can also be a factor. A large feature can contain more magnetically enhanced soil than a very small, shallow feature.
Features C and D are very straight, and are unlikely to be pre-Roman but they could be Roman or later. I wonder if E could be a drove-way leading up from the valley to the west? We need to do some work in the archives and see how much the field systems in this area have changed.
We have two circular features: A and B. My initial quick thought is that these are both round barrows. The majority of barrows are Bronze Age, but we do get barrows in the Roman and Saxon periods too. Their location on a ridge with excellent views would support their interpretation as barrows. The fact that the two features looked so different worried me, and then I remembered a lecture I used to give on aerial photography. Could this be a windmill? A quick comparison with an image published by Wilson (2000, Fig. 58) strongly suggested this interpretation. Of course windmills also want to be up high! We’ll come back to the putative windmill below.
We initially decided to use the Earth Resistance meter over a patch of the field where the farmer had noted it was difficult to plough, and where there were a large number of flints on the surface. Over the three days we completed 14 20x20m grids.
As can be seen from Figure 6, we have detected the ring ditches quite clearly, and some of the linear features. The stripes are plough scars. This makes it quite difficult to see what is happening in many places. TerraSurveyor can apply a Two-dimensional Fast Fourier Transform (2D FFT) to the data to try and removing striping such as this. Using a 2D FFT doesn’t always help, but in this case the results were excellent (Fig. 7).
To aid discussion I have labelled-up the plot as before (Fig. 8).
The two circular features (A and B) show clearly in the res data. Feature B appears to have taken a “bite” out of the high resistance area F. This area is where the flints were on the surface, and rather than being a building, it seems more likely we are dealing with a pocket of flints in the periglacial drift geology. The ditch which runs ENE–WSW just above the letter F shows very clearly where it has cut through these flints. The ditch shown with green arrows in both Figs, 4 and 8 is quite clear. What is intriguing is that ditch C, which appears only faintly to the west of the road in Fig. 4, and then peters out, clearly continues as shown by the pink arrows in Fig. 8. We must never forget that ditches show well in mag data because they act as “traps” for more organic, magnetically enhanced soils. If the fill is not magnetically enhanced, as may happen when one moves away from occupation sites, we may not be able to detect them.
The res survey is particularly pleasing from a teaching point of view. Not only has the 2D FFT worked very well, but this is the exception to my usual statement that “res is less good at picking-up ditches and pits…”.
When I first emailed the people helping on the survey and said I thought feature B was a windmill, Peter Alley immediately pointed out that the lane which runs through the site is “Millfield Lane.” Well done Peter! This is what Wilson (2000, p. 108) says:
Medieval post mills stood on crosstrees whose foundation-trenches formed a cross measuring about 10m wide overall. The crosstrees were usually embedded in, or set in the top of, a low mound surrounded by a ditch. The higher the mound, the broader its ditch, but the less likely that the timbers have penetrated the subsoil. Crop-marks of windmill-mounds thus fall into two groups: those with proportionally broad ditches that usually display no central cross… and those with modest ditches (2 – 3m wide) and a cross within. The ditch is ordinarily 25m in diameter; it may have two even three entrances.
Our feature is almost exactly 25m in diameter but the longest part of the surviving crosstree foundation is slightly less at about 8.5. The ditch is between 3 and 4m wide. The “ditch”, however, does not look like a classic ditch feature in the mag data, and it may have other origins.
I checked the book by Howes (2016) but he only discusses the smock mill known from elsewhere in the parish which was burnt down in 1981 (pp. 132–4).
Jim West wrote:
I have been looking at windmill design to try to identify what would create the large circle (dia about 20m) in the mag results. Thoughts so far:
The windmill was probably a post mill, i.e., the whole of the upper structure rotated on a single post. The post was often supported on a cruciform base rather than set in a hole. This design was in use for several hundred years until c.19 when the more powerful smock mills were introduced.
An example of a trestle base (this one is on brickwork which seems to be a modern improvement)
with the upper structure it looks like this (different mill)
These mills had to rotated to face the wind which was done by pushing on the long beam (threaded through the steps) on right in the image above. This beam (or tailpole in miller speak) had to be long enough to get the leverage to move 15-35 tons of mill. Most of the larger mills had a wheel on the end of the tailpole, usually with an iron rim.
Some later mills used wind power to rotate the mill. The example below gives some idea of the length of the tailpole.
My initial conclusion is that the large circle in the results is a record of the arc of the tailpole; there may have been a surface laid to reduce rolling resistance because the width of the “path” is too great for a wheel rut and with a prevailing SW wind in the UK more wear would be expected in the NE sector.
The results also show what could be one of the beams of the trestle; the beam at right angles to it is less clear but I think supports the ides of a post mill on a trestle.
Having overlaid the res and mag data, the feature suggested by the res does seem a little larger than that suggested by the mag, although they do overlie to some extent. The outer ring, therefore, may be a complex mix of the outer ditch (likely to be more irregular) and the sweep of the tail pole (which would be a perfect circle).
Ruth examined some of the historical evidence:
I have been looking at the old maps I have access to, to see if I could work out when the lane was named ‘Millfield Lane’, the house ‘Millfield Cottage’ and see if I could find any mills in the area. Working back from the modern OS maps, the house only became identified on the map as ‘Millfield Cottage’ between 1960–80. It was previously ‘Millfield Houses’. [NB: you can browse through old maps on the National Library of Scotland website]. I looked back through 20th century maps and back to 1870’s and the name of the lane was consistently Millfield Lane, but there was never a mention of a windmill at that site. On the 1880 Shire view map, there is a windmill (corn) just to the NW of Little Hadham. It is beside Mill Common. The Bryant Map of 1822 has a windmill drawn at, what is probably, the same place as the 1880 map – but none in our field.
Dury and Andrews, 1766 [NB: available to view online here] does record windmills as there is one at Hadham Lordship, but none in the area of Hadham Ford, Berry Green, Hadham on Ash and Green Street. Rowe and Williamson (2013, p. 261) mention that there was a mill in Little Hadham built 1786–7, which is likely to be the one in the Bryant Map and the 1880 OS map. The mill must post-date the map by Dury and Andrews (1766) [NB: available to view online here]. The construction of 1786–7 must have been a ‘new build’.
It may also be the one I found mentioned in Wikipedia entry – Hertfordshire Windmills, Little Hadham, which gets listing from Moore (1999). Only two are mentioned: the first a smock mill that was built in 1786 and burned down in 1981 that matches the location and description of the one at Mill Common mentioned by Rowe and Williamson.
The second entry is that listed by Moore (1999) and dates to before 1700. Moore (p. 77) notes that ‘… just north of Bury Green there is a house today, which O.S. maps name Millfield Houses. Fields on both sides of the road in this position are names ‘Mill feelde’ on a map dated 1588 but it is possible that the field could have been names from a horse mill situated nearby in medieval times. There is no doubt that there was a windmill in medieval times and possibly two sites. The only miller’s name found was in the 1587 Muster Roll Richard Howell – myller.’ There are a couple of references to a 13th century mill, but Moore was unable to show that these refer to this site.
The Domesday Book does not list any sort of mill within Little Hadham.
There is clearly some more historical work to follow-up on. It would be good to see the 1588 map mentioned by Moore which is at HALS, and I’d like to see the references in Holt’s 1988 book cited by Moore.
Hopefully, it won’t take us three-and-a-half years to return to this fascinating site.
Historic England (2011). Mills. Available online.
Holt, Richard (1988). The Mills of Medieval England. Blackwell.
Howes, Hugh (2016). Wind, Water and Steam. The story of Hertfordshire’s mills. Hatfield: University of Hertfordshire Press.
Moore, Cyril (1999). Hertfordshire Windmills and Windmillers. Bishops Stortford: Windsup Publishing.
Rowe, Anne and Tom Williamson (2013). Hertfordshire: a landscape history. Hatfield: University of Hertfordshire Press.
Wilson, D. R. (2000). Air photo Interpretation for Archaeologists, second edition. Stroud: Tempus.