There is an underlying assumption, based on historical and archaeological inference, that the trade during the late Saxon period was orientated around a particular town whose market supplied the local hinterland (Hodges 1989, 164). Distance from the market town is often considered to be the determining factor in the placement of these towns and markets and to be influential in the trade of goods. One of the standard methods for determining the importance of distance to the distribution of an artefact is through a statistical method called regression analysis (Cooper and Weekes 1983). Regression analysis assesses the correlation between the amount of pottery at each site and the distance that site was from the place of production, in this instance, the town of Lincoln (Figure 2). Regression analysis measures the slope of the line which best fits the dots on a scatterplot where the y-axis is the amount of pottery and the x-axis represents the distance from the kiln site. If this slope is steeply angled, it indicates that distance is an important contributing factor to the distribution. In this case, the regression slopes of the distributions are quite low (r-squared value = 0.052), indicating that distance was not an important factor in the distribution. However, the distances used in the regression were Euclidean, or as the crow flies. They do not represent travel distance. Therefore, the regression plots are not measuring the distances involved in transporting the pottery from town to settlement.

Figure 2: Regression diagram of Lincoln wares: Euclidean distances

Indeed, both the distribution maps and the regression plots do not give much information on the extent of travel along roads and/or rivers. Nor do they indicate the time spent in transit. Distance and time are fuzzy concepts. They do not react well to rigidly-defined criteria of analysis. It is impossible to say if a traveller took one road or another; or opted to take a boat part-way e.g. Medieval sources record a trip from Cambridge to York. The journey to York was accomplished by boat but the route back to Cambridge went via the roads (Barley 1936, 15-16). Furthermore, travel time differs between modes of transport. Walking is much slower than riding. Going downstream on a tidal river when the tide is moving out is much faster than poling upstream along a slow moving river. In the example above, it took the travellers to York two days to go up the Witham from the Wash to Lincoln (60km) and the same amount of time to travel from Lincoln to York along the Trent-Humber-Ouse route (120km) (Barley 1936, 15-16). So many factors, both social and environmental, could have influenced this difference, illustrating the difficulties of talking about precise travel times.

One can narrow the factors by considering the movement of particular people and/or artefacts. In this case, it is the distribution of pottery from the towns into the hinterland. There has been much discussion on the ways in which pottery was transported (foot, horse, cart or boat) and what the mechanisms were for that distribution (potter, purchaser or trader) (Kilmurry 1980; Hayfield 1985; Symonds 1999). For the purposes of this discussion, it has been assumed that travel times vary little between the terrestrial modes of transportation. While pottery was an important commodity, it would not have travelled by the medieval version of Federal Express in the satchel of a royal messenger at the speed of 30 miles or 48km per day (Stenton 1965, 255). Travel would have been much slower by ox and cart (a probable transportation method) (Langdon 1986, 49 & 114; 1987, 56). It would have been at a rate of approximately two miles or just over three kilometres per hour (Langdon 1987, 56). Travel along rivers is more difficult to characterise because of variances in current and tide which can affect the speed of the boat. While these factors could be modelled, an average of 4 miles or 6.4km an hour is used, taken from the time it took for Frederic Barbarosa who travelled at a speed of 30-40 miles (50-65km) per day along the Rhine from Sinzig to Aachen, a distance of 60 miles (100km) (Leighton 1972, 177).

In the medieval period, travel occurred via roads and rivers. Roman roads are the most easily mapped. Conversely, local paths are difficult to discern, especially for a region as large as a county. Much debate has ensued on the use of roads and rivers, not only for the medieval period (Barley 1936; Edwards and Hindle 1991; 1993; Langdon 1993), but also prehistorically (Sherratt 1996). There has been significant work done on the variety of roads and pathways in the Anglo-Saxon period (Hooke 1980; 1981; 1998; MacDonald 2001). Most of this work has relied on textual evidence in charters and boundary clauses. These records survive in significantly greater quantities in the south of England in comparison to the north. Indeed, due to the lack of written material for Lincolnshire in the Anglo-Saxon period, the only roads that can be discussed are Roman (Figure 3). Many Anglo-Scandinavian settlements in Lincolnshire from which pottery has been recovered are proximal to the Roman roads and so while local pathways no doubt existed, for the purposes of this study the Roman roads have been used to indicate the distance travelled between settlements.

Figure 3: Roman roads of Lincolnshire

Many of the arguments for the extents of navigation and the popularity of roads versus rivers have been based on textual evidence from the medieval period (Hindle 1976; Edwards and Hindle 1991; 1993; Langdon 1993). The extent of navigability for the Lincolnshire rivers were taken from the work by Edwards and Hindle on medieval rivers (Edwards and Hindle 1991, 131) (Figure 4). Navigability would have been influenced by physical alterations to the river such as the addition of fishing weirs and traps. Indeed, the Domesday notations for Torksey and Nottingham guaranteed the supply and transportation of the king's messengers along the Trent to York (Foster and Longley 1924, 11) and also decreed that:

"in Nottingham the River Trent and the dyke and the road to York are so protected that if anyone hinders the passage of ships, or if anyone plows or makes a dyke within 2 perches of the King's road, he has to pay a fine of £8" (Parker and Wood 1977, 280a.20).

Figure 4: Navigable Rivers

Distance is not that informative, in and of itself. It needs to be connected with time. Two hundred kilometres is a
long way by foot. At an average walking speed of five kilometres an hour it would take twenty hours or two to three days.
Alternatively, it is only two hours in a car travelling one hundred kilometres per hour. Terrain and social factors also
influence the length of a journey and the choice of travel route. The 11th-century historian and geographer, Adam of
Bremen, in his *Gesta Hammaburgensis ecclesiae pontificum*, comments on two ways to get to Trondheim in Norway. The
first is a five-day journey by boat from Ålborg in Denmark. The second is overland from Skåne, but he warns that
this route "is slower in mountainous country and because it is full of danger travellers avoid it" (*cf*.
Page 1995, 43). Lincolnshire is relatively flat and therefore terrain was not incorporated
into the GIS. Travel times and
distances were not precise enough to warrant the inclusion of such a subtle variant. However, this is not to deny the
importance of terrain in more hilly landscapes.

© Internet Archaeology
URL: http://intarch.ac.uk/journal/issue13/1/4.html

Last updated: Wed Nov 13 2002