For most of the Middle Ages England had no fixed capital. The government went wherever the king went, and the king rarely stayed anywhere long. I turned four old printed itineraries into maps, day by day: where Henry II, John, Edward I and Elizabeth I slept, how far they rode, and how Westminster slowly won.
On 15 June 1215 King John was at Runnymede, the meadow by the Thames where he met the rebel barons. We know this from Magna Carta, but we could also know it from the dull business of government: letters sealed in the king's name were dated with the place where he was. In 1835 the archivist Thomas Duffus Hardy went through the rolls of John's reign and listed, day by day, every place those letters name. His index places John at Runnymede on that day, and for the next week he came and went between Runnymede and Windsor Castle, a few kilometres up the river.
Later scholars did the same for Henry II (Robert Eyton, 1878) and Edward I (Henry Gough, 1900), and John Nichols collected the progresses of Elizabeth I (1823). The books are out of copyright and scanned by the Internet Archive, but their tables are locked in messy OCR text. I wrote parsers to pull out the dates and places, matched 90% of John's entries and 93% of Edward I's to a point on the map, and checked the result against events whose dates are not in doubt.
The itinerant court was the normal way to rule a medieval kingdom. There was no ministry building to visit. The king's council, the small curia regis, "met constantly and travelled with the king from place to place", in Wikipedia's words. Royal power was personal, and to be governed you had to find the king.
Some of government did settle. The Exchequer, where sheriffs accounted for the king's revenue, had its home by Westminster Hall from the twelfth century. Clause 17 of Magna Carta required common pleas, suits between subjects, to be heard in "some fixed place", and the Court of Common Pleas sat in Westminster Hall. So the question these maps can answer is not whether kings travelled, but how much, where, and how that changed as Westminster became a capital.
Pick a monarch and press play. The crown marks where the king spent the night (the last place named that day), the line shows the last weeks of travel, and the faint circles are every place he slept, sized by the number of nights. Dates are Old Style (Julian), as the editors print them. Gaps in the record are gaps on the map: the crown waits where the king was last seen.
John's most frequent stopping places are a map of the Angevin realm: Rouen, the capital of Normandy, with 142 nights, then Winchester (138), the castle of Orival on the Seine, the hunting palaces of Woodstock and Clarendon, Windsor, Marlborough, and the archbishop's house at Lambeth. Westminster is only tenth. Across his reign I can put John in 421 different places for the night.
Edward I's list is headed by Westminster, with 841 nights, 15% of all the nights I can place. The rest of his top ten is a record of war: Rhuddlan and Conwy in north Wales, Caernarfon, Stirling, Berwick, Carlisle, and Lanercost Priory, where the sick old king spent 130 nights between 2 October 1306 and 28 February 1307 before his last march north.
The data show John on the road far more often than Edward. On consecutive days for which both places are known, John's court had moved on 42% of the time (more than 3 km between nights), against 23% for Edward I. A typical day on the road took John 35 km in a straight line (the median), and Edward 20 km. John's average stay in one place was 1.8 nights, Edward's 2.3.
In a typical well-recorded year (the median of years with at least 150 placed nights), that adds up to 5,046 km for John and 1,807 km for Edward, in a straight line. John's busiest year was 1204, the year Normandy fell, at 7,541 km. Some of John's restlessness is noise: misread dates in the old index make a few impossible jumps, and I left out the 221 entries that are far from every other record around them or clash with another place on the same day (see section 10). But the difference with Edward is too large and too steady to be an artefact.
Henry II, in the words of Wikipedia's article, "controlled more of France than any ruler since the 9th century Carolingians". Eyton's narrative gives me only 153 dated statements of where Henry II was, a sample rather than an itinerary, but it splits exactly: 47% of the placed ones are in France and 47% in England. Of the 17 Christmas courts Eyton places, 10 were held on the continent, at Bordeaux, Falaise, Le Mans, Bayeux, Nantes, Argentan, Angers and Caen.
For John the break is sudden. From his accession to his crossing from Barfleur to Hampshire in December 1203, 78% of his recorded days were spent across the Channel; from then to his death, only 13%. The exceptions are his expeditions: to Poitou in 1206 (81% of placed nights from June to December), Ireland in the summer of 1210 (78%), and Poitou again in 1214 (94%). Once Normandy was lost, the itinerant court became an English one.
Edward I's itinerary is his wars, almost to the month. In the first Welsh war (from July to November 1277) 51% of his placed nights were in Wales; in the conquest of 1282 and 1283, 67%; in 1284, the year of the Statute of Rhuddlan, 82%. From 1286 he stayed in Gascony for almost three years (99% of placed nights in France), and in the winter of 1297 to 1298 in Flanders (99% in the Low Countries). Then Scotland: 72% of his nights from March to September 1296, and 98% from May 1303 to August 1304, when he besieged Stirling.
About 36% of John's nights in England were within 10 km of a royal forest I could place, and 20% of Edward's. For John the share is highest in November (51%) and lowest in May (22%). This is a crude measure: forests were large, their bounds changed, and only 78 of the 137 I listed could be placed at all.
Christmas was the great feast of the court, held where the king chose to be. Of 14 Christmases I can place for John, 11 were in different places; Edward's 15 were in 14. Neither king kept Christmas in one fixed palace.
Only 8% of John's nights were in or near Westminster (Westminster, Lambeth, the Tower or the Temple); for Edward I it was 16%. Edward still spent most of his life elsewhere, and in war years he hardly came at all. But the Exchequer and the Court of Common Pleas were already fixed there. Government was settling in one place, and the king was becoming a visitor to it.
By Elizabeth's reign the courts and the Exchequer had long been fixed at Westminster. The summer progress survived, but it was a different thing: a tour of the counties near London to be seen, to stay with courtiers at their expense, and to be entertained. Nichols prints the household account of the progress of 1561, day by day in Latin, which I parsed: 37 days from 10 July 1561 to 18 August 1561, 12 stopping places from the Charterhouse in London through Havering, Ingatestone, New Hall, Colchester, Harwich and Ipswich to Hedingham, and only 187 km in a straight line. She moved on 10 days, about 16 km at a time, and stayed 2.5 nights on average.
For other years I have only Nichols's year-by-year summary in his preface, which names 53 stops in 16 years, among them Kenilworth in 1575, Bristol and Longleat in 1574, and Cowdray and Elvetham in 1591. Choose "Elizabeth I" on the map to see them. A progress went where the hosts were, not where the government needed to be.
Most days are blank. I have a record for 67% of the days of John's reign and 47% of Edward I's. A blank day is not a day of rest: it means no surviving document was dated there that the editor found. For Henry II I have only 153 dated statements, and for Elizabeth one summer plus a list of named stops.
The OCR is messy and so are my parsers. I parsed Hardy's index rather than his two-column table, which the OCR scrambles; for 2,273 of John's 5,387 entries the year had to be reconstructed from the order of the index (see methods). Gough's daily pages were read line by line, with each place accepted only if it closely matches a place his own index lists for that month. Of John's day-to-day moves, 5.2% exceed 100 km in a day, as do 1.0% of Edward's; some are Channel crossings, some are errors. I flagged and left out 221 of John's entries and 99 of Edward's that are far from every record around them or that clash with another place on the same day.
Old names are hard to place. I identified the commonest old and Latin forms by hand, pointing each to a named entry in a gazetteer, and matched the rest automatically. Some small French and Irish places are not in the gazetteers I could use and are left unplaced; a few are placed approximately (Freemantle at Kingsclere, the New Temple at the centre of London). Distances are straight lines between such points.
King John. T. D. Hardy, A Description of the Patent Rolls in the Tower of London, to which is added an Itinerary of King John (1835). I parsed the index of the itinerary: 643 place entries, each with county or French province, then calendar year, regnal year, month and days. Where the OCR separated a long entry's year labels from its month lines, I reassigned months to years with a small dynamic programme: years in printed order, months rising within a year, and the regnal years printed beside each year matching the regnal years implied by the dates (John's regnal year began on Ascension Day). Result: 5,387 place-days on 4,244 days.
Edward I. H. Gough, Itinerary of King Edward the First throughout his Reign (1900, 2 vols). I parsed Gough's index of places into 1,881 names with their county and the months he was there, then walked the daily pages month by month, recognising each day line by its expected day number and weekday (computed on the Julian calendar) and re-synchronising on each page's running head. 11,898 of the 12,652 day lines were found; 5,992 days carry a place matched to the index.
Henry II. R. W. Eyton, Court, Household and Itinerary of King Henry II (1878), a dated narrative. I kept a statement when a dated paragraph says the King is, keeps court, lands or stays at a place in Eyton's own index of places, or a charter is dated there, or Eyton's margin repeats the place: 153 statements, 30 flagged as uncertain ("probably", "c.").
Elizabeth I. J. Nichols, The Progresses and Public Processions of Queen Elizabeth (1823, vols 1 and 2): the Latin household account of the 1561 progress, and the stops named in the preface. The scan labelled volume 3 is in fact Nichols's Progresses of King James I and was not used.
Places. Manual identifications for the commonest names, then OS Open Names, Wikidata royal residences, castles and abbeys, and GeoNames towns, choosing between namesakes by the county the editor prints (a namesake outside it is rejected unless it is the only town of that name) or, failing that, by the places where the king is found within three days either side. Country of each point from GeoNames or the nearest OS settlement.