It's a very good question that should remind us of the fragility of many of the ballpark estimates we too easily take for granted - traditionally a quarter or a third, now often a half or even sometimes more.

There are no general death registration systems until the mid-16th century when population is already bouncing back from the low, so that historians have had to rely on scattered returns for particular groups - the aristocracy or religious communities, or more representative local returns usually of taxpayers or taxable households or tenants (i.e. peasant households) or inheritance charges on particular manors for which record survive.

A sample of manorial counts has been worked up by Campbell who finds a sharp drop in England from 4.8m in 1348 to 2.6m in 1351, with a low of 1.9m c.1450 - though it should be noted that there's a severe drop in sample size in the 14th and 15th centuries after the relative abundance of the 13th.

For general populations elsewhere there's even less to go on, though local tax counts where available provide a guide to the relative impact of successive epidemics. England's experience may have been exceptional, though a recent paper dismisses the long-held notion that the southern Netherlands remained relatively lightly scathed, instead noting the long-term impact of successive waves, with rural-urban migration masking falls in town populations.