The Black Death reached Sicily in 1347. For the next three centuries plague kept coming back, to the same towns, every decade or so. This is that history mapped year by year: 6,848 recorded outbreak years in 1,172 towns, from Messina in 1347 to Porto in 1900, and the order in which plague finally left.
Plague is a disease of rats and their fleas that spills over into people. In 1347 it arrived in the Mediterranean on ships from the Black Sea, and within four years it had killed perhaps a third or more of the people of Europe. Then it stayed. For three hundred years and more a town could expect plague to return in most generations, and it did, until, region by region, it stopped. The map below shows every outbreak year in one of the great inventories of that history, Jean-Noël Biraben's, as digitised in 2012 by Ulf Büntgen and colleagues.
Each dot is a town; each outbreak is a town and a year in which plague was recorded there. Press play to watch the Black Death spread out of the Mediterranean ports, then the waves that followed it. Switch view to see the first wave on its own, when each town last saw plague, or how often it came. Click any town for its full list of years.
Data: Büntgen U, Ginzler C, Esper J, Tegel W, McMichael AJ (2012), digitised from Biraben J-N (1975 to 1976), EnviDat, ODbL. Towns at their present-day positions; country outlines are modern (Natural Earth). Base map tiles © Esri.
The inventory's first outbreaks are in 10 Mediterranean towns in 1347: Messina, Catania, Syracuse and the other Sicilian ports, Reggio across the strait, Naples, Genoa and Cagliari in Sardinia. In 1348 plague is recorded in 110 towns, from Alexandria, Cairo and Tunis to Seville, Bordeaux, London and Bristol. By 1349 it is in Paris and the German lands; by 1350 it has reached Danzig, Stralsund and the eastern Baltic. In all, 195 towns have their first recorded outbreak between 1347 and 1353, and 59% of those in Europe were reached by the end of 1348.
How far from Messina, and when
Each dot is a town with its first recorded outbreak in 1347 to 1353 (195 towns), by year and straight-line distance from Messina. Dots are spread sideways within each year so they do not overlap. Hover for the town.
How fast did it travel? This needs care. The inventory gives years, not days, and the first surviving record of plague in a town is not the day it arrived. Ships also carried plague in great leaps: in 1348 it was already 2,000 km from Messina, in England, while towns much nearer were not reached until 1349 or 1350. Fitting a line through distance and year for the European towns gives a speed of about 176 to 214 km a year, depending on whether distance is measured from Messina or from the nearest town infected in 1347 and how the line is fitted. That is 0.5 to 0.6 km a day as the crow flies: an average over fast sea passages and slow movement overland, not the pace of any single journey. The year-by-year reach is clearer: the ten towns of 1347 all lie within 885 km of Messina (Genoa is the farthest); by the end of 1348 nine in ten of the towns reached lay within 1,974 km, and the farthest, Dundalk, 2,421 km away.
The Black Death was the beginning, not the peak, of this record. Outbreaks are recorded in almost every decade from the 1340s to the 1660s, and the busiest decade is the 1630s, with 570 town-years in 300 towns, when plague swept France and northern Italy while the Thirty Years War raged in the German lands. The single busiest year is 1630 (119 towns). Partly this is a record of record keeping: town archives fill out in the sixteenth and seventeenth centuries, and 68% of all outbreak years fall between 1500 and 1669.
Outbreak years per decade, by region
Town-years with plague recorded, by decade and region. Hover a bar for the numbers.
After the 1660s the bars collapse. Only 7.6% of the record comes after 1667, and from the 1720s onwards 86% of it is in North Africa, Anatolia, the Levant and Egypt.
London has the most outbreak years in the inventory, 132 between 1348 and 1667, ahead of Algiers (112), Paris (89) and Toulouse (88). Most towns appear far less often: 548 of the 1,172 (47%) have a single outbreak year, and the 20 busiest towns hold 21% of all records. A big number usually means a big town with good records, and an author who read them: London's bills of mortality, kept week by week from the sixteenth century, are among the best plague records anywhere.
The 20 towns with most outbreak years
Distinct years with plague recorded, 1347 to 1900. Bars coloured by region. Click a bar to see the town on the map.
When plague struck a town it often stayed for more than one year, so I join consecutive outbreak years in a town into one episode (4,375 episodes in all, 30% of them lasting two years or more). The gap from the start of one episode to the start of the next in the same town has a median of 13 years; half of all gaps are between 6 and 31 years, and 56% are 15 years or less. The gaps shorten over time, from a median of 18.5 years for episodes starting in the fourteenth century to 11 in the seventeenth. In London, with 33 episodes, the median gap is 7 years.
Years between outbreaks in the same town
3,203 gaps between the starts of successive plague episodes in a town, in two-year bins; the last bar is 60 years or more.
Some of this regularity is real: plague needed time for a town to refill with people who had not been exposed, and it came back with trade. Some of it is not. A long gap can simply be a gap in the archives.
Plague did not leave Europe all at once. Using the last year in which three or more towns in a region record plague, it left Italy first (1657), then Britain and Ireland (1667), Iberia (1681), Poland and the Baltic (1712), central Europe (1713) and France (1722). In North Africa it lasted to 1821 and in Anatolia, the Levant and Egypt to 1834, a century after western Europe.
When plague left each region
Line: first to last recorded outbreak. Diamond: the year by which 95% of the region's outbreak years had happened. Large dot: the last year with plague in three or more towns. Ring: the last record of all. Sorted by the large dot.
The famous last epidemics are all here. London's Great Plague of 1665 is recorded in 57 towns across Britain that year, and London's last recorded year is 1667. The Great Plague of Marseille, brought by a ship from the Levant in 1720, is recorded at Marseille in 1720, 1721 and 1722 and in 10 towns of Provence and the south in those years. Messina, where the Black Death landed, was struck again in 1743. Last records in a region can be strays: a single year at Marseille in 1786, Noicattaro near Bari in 1815 and 1816, Mallorca in 1819, Porto in 1899 and 1900 (part of the third pandemic, which spread from Asia by sea). One famous epidemic is missing: Moscow's plague of 1771 is not in the digitised inventory at all, and Russia appears only twice, so the east is far less well covered than the west.
The last recorded outbreak in each country
Modern countries, sorted by the last year with plague recorded anywhere in them.
This is a map of where plague was written down and where those writings survived and were read, not a complete map of plague. France alone holds 40% of the outbreak years, and France and the United Kingdom together 58%: Biraben was a French demographer and historian, and his inventory is fullest for France. The Low Countries, Scandinavia, Bohemia beyond Prague, Hungary, the Balkans and most of Russia have few or no records, though plague certainly reached them. Constantinople, one of the great plague cities, is absent. The inventory records that plague was present in a year, not how many died, so a year with a handful of cases counts the same as one that killed half a town. Some years in the source are approximate, and a few towns were placed by an automatic geocoder in the wrong place; I corrected five that were clearly wrong (Prague, Milan, Noja, Graz and Sciacca) and list them in the data.
London has 132 plague years in the record. Constantinople has none.
Data. The file published with Büntgen and others (2012) lists 6,929 rows of town, year and coordinates. Its latitude and longitude columns are swapped; I checked the fix against the file's own projected coordinates (ETRS89 LAEA), which agree to within a few metres for every row. Two rows for Eisleben were broken by a comma in the name and are repaired, one row with the year 9999 is dropped, and 80 rows that repeat a town and year already listed are merged, leaving 6,848 town-years.
Places. Each town is placed in a modern country (Natural Earth) and one of eight regions. Modern borders are a convenience: in 1500 there was no Germany or Italy, and Algiers, Tunis, Tripoli, Cairo, Smyrna and Beirut were in the Ottoman world for most of the period.
Measures. An outbreak is a town-year. An episode joins consecutive outbreak years in a town; recurrence is the gap between episode starts. The first wave is every town whose first record falls in 1347 to 1353; distances are great-circle kilometres from Messina or from the nearest town infected in 1347, and speeds are slopes fitted by Theil-Sen and least squares. A region's end is shown three ways (the last record, the last year with three or more towns, and the year by which 95% of its records had happened), because a single late record can mislead.