How Many Hobbits? 3,000 Years of Middle Earth Population History

This is the third post in my series on the demography of Middle Earth. The first piece introduced some methods for estimating the population of Middle Earth in TA 3019, the year that most of the Lord of the Rings is set. The second piece provided more details, estimating TA 3019 populations for some of the major people groups of Middle Earth.

UPDATE 1: After wonderful feedback from fellow nerds, I have made some revisions, in particular to Elven populations. Lindon got bigger in the early period and got smaller after TA 1975 and especially after TA 2950, while the Woodland Realm got bigger in all periods. The Elven population of Belfalas, i.e. Edhellond, was also increased before its abandonment in TA 1951.

UPDATE 2: More awesome nerds! It was noted that my land area estimates were way too low, but my benchmark densities too high due to selecting less-than-ideal historical analogies. Combined with a lot of pushback on the population of Eriador, there have been substantial and in the main downwards revisions.

In this post, we’re going to turn back the lock and look at history. Tolkien’s writings on the history of the third age are really just a sketch; for much of his world, we have many centuries of the Third Age with no recorded events of note. Even where we putatively have a royal dynasty, we may be missing 75% of the names of kings and the dates of their rule. We get tantalizing hints of the depth and complexity of Tolkien’s world (a kingdom of Rhovanion whose princess married into the Gondorian royal line?), yet there are nonetheless enormous gaps.

Even so, there’s a lot we can say. Consider a region like Erebor and its surrounds (including Dale, Lake-Town, Esgaroth, and their hinterlands up to the eaves of Mirkwood). To understand how I estimate regional populations, Erebor is a clarifying example, because several key elements of its history are probably familiar to readers.

Erebor’s Example

Here’s my estimate of historic population:

Where does this come from?

Well, Tolkien reports that the first Dwarven colonists discovered Erebor in the Second Age, but they didn’t meaningfully settle it. Nonetheless, I assume over time there was probably some gradual Dwarven expansion there. It’s an obvious stop on the road from the Iron Hills (which we know was continuously settled) to Khazad-Dum. Especially since we know from the annals of the Gondorian kings that this region was pretty well packed with little human kingdoms and tribes and principalities, demand for various ores was almost certainly abundant, and Erebor’s rich deposits cannot have gone totally unexploited. Furthermore, when Thrain I led Durin’s Folk to Erebor after the Balrog drove them out of Khazad-Dum in 1980, it is implausible that they showed up to an un-mined mountain. Such a large population movement demands that some mines already existed. There must have been a long-standing Dwarven colony which had dug out considerable delvings, such that they could be repurposed to house a considerable share of the population of Khazad-Dum.

So population spikes in the late 1990s and early 2000s. But then we know Erebor is abandoned in TA 2210 when Thorin I leads his people into the northern, Gray Mountains, where they have relatives and one of the other Dwarven clans had long resided, and where mineral deposits were perhaps more abundant. So the Dwarven population of Erebor is greatly reduced until, surprise surprise, in TA 2589, the Dwarves are back, having been driven from the Gray Mountains by dragons. Their kingdom under the mountain lasts until TA 2770, when Smaug drives them out. They return under Thorin II Oakenshield, assisted by Bilbo Baggins, in TA 2941, and are greatly bolstered by Dwarven refugees from all around Middle Earth, including Dwarves not of Durin’s Folk from the east, where Sauron’s consolidation of rule over the Easterlings led to major disruptions and Dwarven migrations west.

But what about men?

Obviously, the land was fertile and well-watered with navigable waterways. Somebody lived there. But who were they, and how numerous? I opt to assume the population of men was not inconsiderable. With the Woodland Realm to their west, any settlement of men had one totally secure border. The dragons to the north cut down on northern orcish problems, but rarely ventured far south. The easterlings could be a problem, but in the wake of Sauron’s defeat in TA 0–2 we must imagine the area enjoyed relative security and peace on that front. The men who lived in this area were Northmen, descended from the “Greater Folk” of men, the ancestors of the Edain who ventured into Beleriand in the First Age, and Tolkien is clear that the Northmen are next to only the Edain in general capabilities and civilizational complexity.

We also know the area was politically sophisticated. The Easterlings invade Gondor in TA 490 and do not appear to be a disorganized migration or even a tribal confederacy. Several of their kings were the kings to whom Sauron gave rings, and were corrupted into the Nazgul. Three separate groups of Easterlings are reported to invade Gondor: generic “easterlings” TA 490–550, “Wainriders” in the 1800s, and Balchoth in TA 2509. The Wainriders and Balchoth both seem to have been heavy on cavalry, suggesting the area may have shifted more towards pastoralism in the period.

Regardless, fertile lands around Erebor, with protected western borders by the Elves, and easy trade access to the resources of the Iron Hills, there’s just no way the area wasn’t settled.

Which is weird, because in one place Tolkien states the city of Dale was founded in response to the Dwarven occupation of Erebor.

But elsewhere, Tolkien gives a different hint: in Unfinished Tales, Tolkien reports that defeated Northmen allies of Gondor fled north to Dale, where they merged with the men there, to whom they were kin. This implies that in TA 1856 when that story occurs, Dale already existed, had Northmen in it, and was a well-enough known and secure-enough polity that the retreating men felt confident going there. This same event led to another faction of Gondor’s Rhovanion allies fleeing west over the Anduin into the region around the Gladden fields: these men formed the nascent Eored, which, after two more migrations, would become Rohan.

Additionally, we can assume population growth was slower around major documented episodes of Easterling violence (likely spilling into the region and disrupting trade). Finally, the major population event, Tolkien’s account of the Great Plague of TA 1636 is explicit: the men of Rhovanion and around the eaves of Mirkwood were absolutely eviscerated by this plague.

Thus, to understand the mannish population of the Erebor/Dale region, we have some key elements. There must have been men there over most of the period in question; it’s an ideal settlement site, there’s just no way it was void of people. The 1636 plague must do extreme demographic damage, but in 1856 there still has to be a big enough population to support some kind of organized settlement in Dale. That means the 1635 population must be quite large indeed. Population growth during periods of Dwarven occupation of Erebor can be assumed to be robust, while the Balchoth episode (which stretched further north than Wainrider activities), can be assumed to have been very damaging. Then, in TA 2770, Smaug’s attack we know led to the obliteration of Dale and the decimation of the human population in the area. Yet not its complete erasure: Lake Town survives, and carries on long-distance trade all the way to Rhun, implying that many thousands of people must yet live around the area. Smaug’s long quite allows population rebound, and Dale’s reconstruction accelerates it further.

This is just one example of how I came up with these estimates. Every region gets its own treatment. But you can see the kinds of general outlines I’m working with. I won’t walk through every region in detail.

Population Through the Third Age

So with that kind of analysis as the basis, and making sure the final output figures match the detailed estimates we’ve come up with for TA 3019 in prior posts, what figures to we arrive at?

Here’s population over time for five major populations in northwestern Middle Earth: Men, Elves, Dwarves, Hobbits, and Orcs. We’ll start with just the combined population.

Middle Earth’s population boomed in the few hundred years after Sauron was defeated. But… the boom did not last. This is sort of peculiar because, as I have noted in prior posts, the lands of Middle Earth could probably sustain a population of 25–40 million people without terribly great difficulty. To understand what’s going on here, let’s look at population by people group.

How Many Babies do Gondorians Have

Population tapers off, however, for a few reasons. A you can clearly see, Men are the lion’s share of Middle Earth’s population. I assume their populations rise considerably after Sauron’s defeat in TA 0–2 as Gondor and Arnor thrive, Sauron’s oppressions end, and disruptive wars are mostly avoided. But then around the 400s, population growth ceases, and it ceases when population is just 10million or so. What’s going on here is a few things. First, by TA 490, Sauron’s influence is clearly at work: major Easterling invasions of Gondor are under way, not thoroughly beaten back until TA 550. Likewise, by TA 861, Arnor has splintered into three realms, and the uniform read of Tolkien’s limited textual evidence is that Arnor’s population was in decline, despite no recorded wars (though one Arnorian king, Valandur, is killed in violence in an unknown context in the 600s). Interestingly, Tolkien seems to suggest that more Dunedain settled in Arnor than in Gondor, which has led some fans to suggest Arnor was actually more populous than Gondor. I suspect that’s incorrect however: Arnor perhaps had more Dunedain, but Gondor was packed to the gills with Middle- and Low-Men, as is documented in continuing existence of the Pukel Men and the Men of the Mountains in TA 3019.

More generally, Tolkien is concerned with civilizational decline. In many places, he refers to groups declining, fading, withering, etc. Readers sometimes read this in very vague ways, but I take it very straightforwardly: as a Latin Rite Catholic with 4 children, Tolkien was commenting on the low birth rates normative among Elves, Dwarves, and the Dunedain, i.e. all the elite, and likely eventually among the peoples they governed. Tolkien explicitly says Elven fertility declined over the ages and by the Third Age was quite low. And while we cannot conclude too muchabout total family size from the Gondorian or Arnorian king lists, we can observe rates of childlessness. Numenorean law explicitly allowed for the accession of Queens when no son was available, so if the dynastic line is broken (except in one case) we can conclude the father was truly childless.

No Arnorian or Arthedainian king was ever childless. In fact, none of the Chieftains of the Dunedain were ever even childless. Aragorn is a pure unbroken male-line descendant of Elendil, which is precisely what makes him so unique. However we can assume that not very many of the Arnorian kings or Dunedain Chieftains had many children, because Aragorn is the last such male heir, meaning there weren’t many of these men who had multiple sons.

On the other hand, Gondorian kings are childless in several cases! Tarannon Falastur (r. TA 830–913) dies childless with a famously miserable marriage as well. Narmacil I (r. TA 1226–1294) is also childless. And finally, Earnur (r. 2043–2050), the last heir of Anarion, also died childless. Of 33 kings, 3 were childless.

Now, if we compare to English kings, across 59 English/British monarchs, 16 were childless, a much higher rate. However, their average age at death was 50, and 6 died before age 25. Moreover, England/Britain has had by my count 14 different dynasties across 1,137 years. In fact, by my count, childless monarchs in Britain have a 37% chance of the next ruler not being of their dynasty, whereas non-childless monarchs have just a 6% chance of the next ruler not being of their dynasty.

On the other hand, Gondor has zero dynasty changes among the kings, but one dynastic change (the end of the line of kings) after a childless king out of 3 total childless kings, implying a 33% chance of dynasty change for childlessness: about what we observe among English/British monarchs. Moreover, although Gondorian kings have lower rates of childlessness, a few other facts stand out: zero Gondorian kings died before maturity, and so their fertility rate should be higher than English kings. And yet, for the 13 Gondorian kings for which we can reasonably approximate total fertility (many we simply know if they had at least one son, giving us little guidance about total fertility), Gondorian kings averaged just 3 children each. By comparison, including 20th century kings, English males monarchs who did not die prematurely (i.e. more direct analogs for the Gondorian kings) averaged 5.7 children each, almost twice as high. Thus, whereas English monarchs produced 276 potential continuers of the royal lineage across 1137 years of rule, we can guess the Gondorian kings produced just 102 potential heirs across a total of 2053 years of rule.

In other words, Gondorian monarchs have surprisingly low fertility, despite, and perhaps because of, their great longevity. The average age at birth of a first son for Gondorian kings is 74 out of 28 kings with confirmable first birth dates for sons, while average age at death was 227. On the other hand, English monarchs tended to have a first son by age 22 for those who had sons, out of an expected lifespan of 50 years. Compared to their lifespan, Gondorian kings had kids at a fairly young age (33% of expected lifespan, vs. 44% for English monarchs), meaning many Gondorian kings coexisted with great-grandchildren, a rare experience for English monarchs. For example, Meneldil, the first king after Isildur, dies in TA 281: his son is born 42 years before the Third Age, his grandson in TA 48, his great-grandson in TA 136, and his great-great grandson 222. His great-great-great grandson is born in 310.

Just for transparency, here’s the figures I have for Gondorian monarchs. If you think I’ve estimated their fertility wrong, you can tell me!

What am I getting at here? Simple: since we know that in premodern societies elites tended to have more surviving children due to greater resource access and ability to secure mates, the fact that Gondorian kings have barely over 1/2 the birth rate of English kings is very striking. Now, maybe their infant mortality rates were lower: but that’s the point, they look more like modern fertility behaviors. Fertility is fairly low among the Men of Middle Earth.

We can point to other evidence too! Tolkien tells us a fair amount about the Plague of 1636. It killed a huge share of the human population (including a king of Gondor and all his kids, at least 3 kids from Tolkien’s grammar, possibly more). I show it killing 25% of men around northwestern Middle Earth. Tolkien is also adamant that populations never recovered. Minhiriath remained a wasteland. No sophisticated Rhovanion kingdom ever re-emerged except Dale. Osgiliath is never repopulated. Over 1,400 years later, great cities like Tharbad remain empty.

Is this plausible? For a comparison, here’s English population around the Middle Ages:

The Black Plague killed perhaps half of England’s population. But by 1650, 300 years later, England had recovered. A similar story exists around much of the world.

England recovered because in premodern societies, when population densities fall, it opens up more land per person, which tends to boost wages and living standards, resulting in more marriage, more fertility, and more child survival.

But that didn’t happen in Middle Earth! My working theory of why has two components:

1. Sauron- I think Sauron is cursing the fertility of Middle Earth, making life whither and leading to reduced fertility or elevated miscarriage. We strongly suspect he caused the Great Plague, so this is plausible.
2. Elite Norms- The Elves are the ultimate Elite of middle earth. Their culture is highly influential. They have very low birth rates. The Dwarves are the wealthiest of middle earth, another plausible elite class: they have low birth rates (more on this below). The Dunedain are clearly an elite: we have established their low birth rates. Middle Earth’s elites among the free people may be propagating small-family, anti-natal social norms, leading to population stasis, as occurred in e.g. Edo-period Japan or the early days of the Roman Empire.

Populations of men do rise somewhat during the Watchful Peace (2063–2470), supporting the idea that Sauron’s activity suppresses population growth. That’s not just my modeling: the only way to make the huge plague effects square with observed population in TA 3019 is to assume meaningful growth between the two, and yet population is clearly declining in TA 3019 in many places, so you really have to get growth earlier, and the Watchful Peace is the most plausible period for this.

Having discussed men, let’s turn to the other groups!

Elves, Hobbits, Dwarves, and Orcs, oh my!

Looking at the other peoples of Middle Earth, we notice first that the Elves are in decline. They are sailing west, and their fertility rates are very low. There’s comparative Elven stability TA 30–850 as the world was mostly peaceful, and again 2063–2470, but in periods of disorder and chaos, Elves flee across the Straight Road to Valinor. As a result, a group that was fairly large in TA 0 (almost 1.2 million elves!!) is shrinking over time: only around 300,000 remain in TA 3019.

Second, the Dwarves also decline. This one often surprises Tolkien lore fanatics, but Tolkien indicates that the Dwarves of Khazad-Dum were in decline even before the Balrog awoke. The dynamics here are apt enough if you’re read up on Dwarven sociology: the quasi-canon lore suggests 1/3 of Dwarven women never marry and that birth rates of those who do marry are not especially high. Combine that with repeated displacements and wars, and it’s rough to be a Dwarf.

Hobbits do grow over the Third Age. Hobbit population dynamics pre-1600 do not show a ton of growth, as one would expect from their small homeland and itinerant wanderings. These repeat migrations are not consistent with major long-run growth. But after settling the Shire in TA 1601, Hobbit populations boom, though the plague of TA 1634 interrupts that growth, as do the Days of Dearth later in the Third Age. Eventually, the Hobbits begin to hit what seem to be subsistence limits, at least within the framework of non-industrial, low-hours-worked lives that Hobbits prefer, which is what leads to colonization of Buckland and, after TA 3019, the Westmarch as well.

Finally, the orcs. Estimating orcish population is thorny. My method here was like this:

1. Identify major battles.
2. Make army size estimates for the Free Peoples using population estimates from the relevant populations with various assumptions about mobilization rates etc
3. Guess about how many orcs it takes to offset 1 Elf, or 1 Dwarf, or 1 Man, or 1 Hobbit, etc
4. Assume only 1/6 orcs can actually be deployed into combat efficiently
5. Make orcish populations 6x estimated army sizes needed to yield historically observed win/loss outcome of battles for which I can estimate the required orcish combatant count
6. Make orcish population rise/fall as necessary to hit those numbers

This is a crude approach. You can contest it. But it does yield some rising orcish populations over the Third Age, which I find plausible. Major booms-busts in orcish populations include the rise of Angmar post-TA 1250 (and then a decline with its fall in TA 1975), the rise of Dol Goldur after the mid 2nd millennium (and some decline 2063–2470), the rise of Misty Mountain orcs post-TA 1980 and especially post-2300, and then their decimation in the War of Dwarves and Orcs in the late 2700s, followed by a major rebound in the last two centuries of the age as Sauron’s activity in Mordor escalated rapidly.

On the whole, I think these estimates by people group are reasonable… but what about estimates by political entity?

Population by Polity

Middle Earth is full of different governments. Some are clear-cut and pretty well described: Thranduil’s Woodland Realm, Gondor, the Shire. Others are not as clear: the Woodmen and Beornings, the Dunlendings, various unnamed realms in Rhovanion, some allied to Gondor and some not. Much of Middle Earth is effectively ungoverned by any known organized polity as of TA 3019, and this was true for many regions in virtually every period.

Using all the available data on when territorial control changed for various polities (especially the Dunedain states) in combination with population trends by region and people group, we can make reasonable guesses at the population of various states at various times. So, for example, here’s Gondor and its various vassal states (note: because my map cuts off Umbar and most of Harad, Gondor and vassal populations are underestimated in periods of Gondorian supremacy in the south).

Gondor rose through natural growth and conquest until the period 1000–1450, then through civil war, defeat in external battles, and especially plague, it declined sharply through about 2063. There was a lull in the watchful peace before a decline in the last 5 centuries of the Third Age.

And here’s Arnor and its successor states:

Arnor experienced a boom in population for a few hundred years after Sauron fell, and then a steady decline, likely due to low fertility, after TA 500. When Arnor split up in TA 861 a few outlying regions were lost, and further losses occured over time. Arthedain was the clearly dominant contender of the three successor states, though Cardolan stans are welcome to send me mail if they like: I’d love to justify a bigger Cardolanian population! Convince me! Rhudaur, meanwhile, is clearly the junior state, which makes its vassalization by Angmar around 1350, and annexation in 1409, make sense. Arthedain’s absorption of Cardolan around 1409 gives it a momentary boost, but I assume Minhiriath was lost in that absorption. The plague hits Arnor’s southern regions especially, but after the plague population rebounds enough to venture a resettlement of plague-wrecked Cardolan. That doesn’t last however, and constant war with Angmar depopulates much of Arthedain in the 1900s, before final defeat in 1974.

Just for perspective, here’s all Dunedain realms together:

So the high water mark of the Dunedain was around TA 250–1000, and then they experienced a variety of disasters that felled them TA 1300–2000, the Gondor held steady after TA 2000.

Here are the Elven realms:

UPDATE: This graph is quite different from the graph I presented in the first version of this post. Lindon starts higher and ends lower, while the Woodland Realm is much higher in all periods.

To start with, my assessment of Avarin Elven populations in the east around Rhun are wildly speculative. Also, the volatility before TA 200 is also speculative. We know many Elves sailed west in the period, and many more simply moved west, presumably to Lindon. This was a period of great population reshuffling. I assume it benefits Lindon; you might quibble. Regardless, Lindon undergoes massive decline over the period as many Elves sail west. Lothlorien and Rivendell see a less several long-term decline after TA 100. The Woodland Realm has some volatility, and rather few losses due to Elves sailing west, but nonetheless declines over time.

You may be interested in my estimates of other realms of Men. Here are the Free Peoples:

Miscellaneous unnamed Rhovanion realms ebb and flow in response to Gondor’s expansion and contraction, wars launched by Easterlings, plague, and Sauron’s various machinations. The Woodmen and Beornings persist across the period, but diminished over time especially with the rise of Dol Goldur. The Eored/Rohan has some volatility during their pre-2509 migrations, but then their population grows explosively after they settle and form the realm of Rohan; evidently the Rohirrim have very high birth rates! Just, not Theoden.

Dale and Laketown we’ve already discussed in the discussion of Erebor.

If we look to the realms of evil men, we see:

Easterling populations are all over the place as their control of territory often ebbs and flows in different regions with different rises and falls of Easterling polities. Harad’s population is egregiously underestimated, as I can assess only near Harad. Dunland, meanwhile, is relatively flat over time, only rising as a meaningful polity a bit before the Rohirrim settle Rohan.

Here, I show the summed populations of all evil realms:

The power of evil ebbed and flowed before TA 2063 with various efforts by Sauron to secretly weaken his foes. Then, strangely, because the Watchful Peace allowed populations to rise everywhere, the powers of evil benefited as well. By TA 3019, evil realms could claim 3 million subjects.

And the realms of the free peoples?

By TA 3019, around 4million. Here are the two as a simpler graph:

Here you can see the basic story of the Third Age. The Free Peoples dominated the first 1,000 years without much serious contest. The second 1,000 years saw the Free Peoples decline through various machinations of Sauron: Angmar, Wainriders, Plague, possibly low fertility can be included in Sauron’s schemes. This period didn’t see an open buildup of the power of evil, but a carefully planned weakening of the Free Peoples. Then the third millennium saw the power of the realms of evil grow steadily.

UPDATE 1: To clarify, these are the populations of evil realms within my area of study. My map excludes most of Harad, all of Umbar, all of Khand, and almost everything east of the sea of Rhun. Doubtless Sauron has a huge population reservoir in those areas. However, distant vassals are less useful. It was hard enough for Sauron to muster his armies from nearby vassals; the ability of even a large vassal in very distant lands to contribute armies is limited, and the duration of any campaign would likewise be circumscribed. In sum, then: the actual population of evil realms is much higher than is shown here, though the trend may be similar. It’s also notable that there are Free Peoples elsewhere as well: apparently Sauron’s muster in the east was partly disrupted by the efforts to the two Blue Wizards who were organizing resistance to Sauron in the east.

Tolkien is clear that Sauron did not intend to attack just when he did in TA 3019. Perhaps he meant to wait a few more years, perhaps just a few weeks: he was rushed along by the discovery of the ring, which accelerated his plans. This graph points to an obvious conclusion. Though powerful realms of good like Lindon were far from Mordor, Sauron’s experience of the Second Age led him to be cautious of launching a war before he had a demographically suitable base. He didn’t quite have it in TA 3019, as evinced by the fact that not only was the Siege of Gondor broken, but his armies were smashed at Erebor and Lothlorien as well, while his ally Saruman was broken.

Had the ring not been destroyed, Sauron would have kept fighting, and plausibly a next invasion wave might have successfully decapitated Gondor even with Aragorn leading it. But also, plausibly Cirdan would have sent some elves, Lothlorien and the Woodland Realm too. Lothlorien won their fight, the Woodland Realm and Erebor hung in the balance, but a conventional military victory for the Free Peoples in the first round of fighting in the north was not impossible. Likewise, Gondor, given more time, would muster more troops from its westerly reaches, this process was already underway, which is why Aragorn is able to march to the Black Gate, because a few thousand more levies from the south showed up.

Point blank, in TA 3019, after Saruman’s defeat and Aragorn’s enlistment of the Army of the Dead, Sauron did not have a path to an immediate decisive victory. Erebor would only crack under a long siege, the advances of Dol Goldur against the Elves were defeated, the Corsairs of Umbar had been broken as a fighting force, Isengard and Dunland were subdued — Sauron still had vast military reserves from Harad, Khand, and Rhun, but they would have to once again fight at the fords of the Anduin, and against a better-led and better-prepared foe, and without some of their allies in the first round. Plausibly Sauron’s actual strategy would have been to harass Gondor with smaller forces and focus his main efforts on starving out Erebor then crushing the Woodland Realm, as these were by far the hardest polities for the Free Peoples to reinforce from their southerly bases.

A Note on Eriador

Eriador has been a major bone of contention in discussions about these numbers. Tolkien gives is a frustratingly large and yet too small amount of detail about it. Sources disagree. For the uninitiated, Eriador is this region:

It contains all of Arnor as well as Eregion and Angmar and some bits of land on the west and north we don’t usually count as Arnor. The realm of Arnor is about 270,000 square miles, and if you add in Eregion, Angmar, and odds and ends around the northern, eastern, and western fringes, you get something like 350,000 square miles for the region. The land is on the whole pretty fertile, including the buccolic lands of the Shire right in its heart.

How many people live in Eriador? In simple terms, what everybody agrees on is that Eriador was once populous enough to support a complex, sophisticated society with at least 5 known and significant urban centers (Fornost, Tharbad, Lond Daer, Annuminas, and Bree, though Lond Daer died out as Tharbad grew, and Anniminas died out as Fornost grew). Arnor was a major kingdom which left behind culture, language, infrastructure, and a genetic legacy on the people of Eriador. There’s also agreement that population declined pretty persistently through most of the Third Age, and by the 2900s was very low outside the shire. What we disagree on is exactly how low.

But… how big was it in the early period? And how small at the end?

As it happens, fans have gone to great lengths to offer an estimate! A publication called Other Minds designed to support Middle Earth roleplaying had a 2012 article by Thomas Morwinsky about population and urbanization in Eriador. Below, I line up his estimates vs. mine. I assume we are both excluding orcs.

As you can see, Morwinsky has a much higher estimate of the early population of Eriador, but by the 2900s is actually below my estimate (though we are pretty similar).

So who’s right?

Well, step 1 is just to switch to something more intuitive, like population density, which is easier to get a grasp on. Below I show my estimates of density, Morwinsky’s, and then also a range of historic estimates from 0–1000 AD in our world.

Morwinsky’s population densities fall in an intermediate range, around 15/mi until the 1400s, when they converge towards mine. Densely settled European polities like Germany, France, Italy, and Spain are appreciably more thoroughly settled than Morwinsky’s vision of Eriador in its prime, but it is much more settled than, say, Great Britain or Poland or Uzbekistan or Sweden. The years there by the way are literally 0–1000 AD. Obviously the timescales don’t actually overlap correctly, it’s just so you can see the variation.

On the other hand, my model suggests Eriador at its prime was only as densely settled as medieval Poland or Britain.

So who’s right?

Well, I can tell you that, having gone 18 rounds at this with other Tolkien fans, the texts are not going to settle the question. Tolkien’s sketch of Arnor is barebones thin in terms of population details. What we do know is the wars of the Second Age left Eriador greatly diminished, and that battlefield casualties were demographically significant, which suggests that the population must be small enough for battlefield casualties to take a chunk out of it. We also know that the watchtower of Amon Sul is not a huge fortification, and yet was jealously fought over by all 3 Arnorian successor states. We know that Arnor/Arthedain actually did not have the demographic heft to maintain two major urban centers simultaneously (Annuminas and Fornost), and the same seems to be tree of Cardolan (Tharbad and Lond Daer/an unnamed city near the downs). Rhudaur has no known urban sites at all.

To me that points towards Arnor’s density peaking rather low compared to medieval populations. I have a hard time visualizing even Eriador at its prime as a thick patchwork of farmlands and towns like France or Germany, not least since Tolkien describes much of it, like Minhiriath, as burn and shriveled wasteland even at the height of Arnorian rule.

Perhaps most strikingly though, I want to suggest that Arnor really is post-Roman Britain. Like Rome, the Dunedain were once a unified and dominant culture distinct in language and warfare, unchallengable. Then the empire divided (Arnor/Gondor:Rome/Byzantium). The theoretically senior and more culturally ancient realm (Arnor) nonetheless withered through demographic decay, barbarian threats, and fractious infighting. The last gasps of Arnorian/Roman rule made efforts at reunification with Gondor/Byzantium, but it was not to be. Ultimately Gondor(Earnur)/Byzantium(Justinian) do sweep in and crush Angmar/the barbarians, but it’s too late, and Rome cannot be restored. The Byzantium/Gondor carries on and has its glories, but in time the old Numenorean blood/Latin language is diluted with Middle Men/Greek language, and the kings/emperors are effectively replaced by stewards/magnate dukes. In time, in the north, legends abound of a secret king, Arthur/Aragorn, who one day will return to restore the fabled glory of his kingdom. A Reunited Kingdom, a New Rome. And, as a kicker, apparently (I have not read the letter myself but am trusting secondhand sources), in at least one letter, Tolkien apparently referred to Venice as Pelargir and to England as “the northern kingdom.” To me, this seals the deal. Arnor at its height should have the population density of Roman Britain. Here’s the density again:

You can see that my line is pretty much directly on the red line for Great Britain 100–400 AD, i.e. Roman Britain. So, at its peak, Arnor=Roman Britain.

Now, when Rome abandoned Britain to barbarian attacks, population fell — a lot. So too as Arnor declined. But Eriador’s collapse is far more total than post-Roman Britains. The men of Angmar and the hillmen of Rhudaur are nice stand-ins for barbarians at the end of Rome or vikings, but the orcs are something else — they eat the local population for dinner, literally. Moreover, I’ve already established that Middle Earth is a bizarrely low-fertility regime. Ruinous warfare, plagues, a foe whose food is you, that’s way worse than post-Roman Britain. So Eriador’s density I have bottoming out around 1.3 people per square mile, similar to Sweden, Norway, or Mongolia: largely empty and desolate places. I have it rising during the Watchful Peace because how can it not? There’s peace! No Angmar! Hundreds of years of no recorded threat! But then a decline post-2700 as new threats emerge.

After 2700, Morwinsky has a lower density estimate than me, although we are pretty close. So, why are we different? Morwinsky gives regional estimates, I have regional estimates, let’s compare. But to start, we must compare maps! Here’s his map of polities in 2915:

And, for reference, here’s my map of regions in Eriador:

What you will immediately notice is three of Morwinsky’s polities (Meach, Senana, and Saralainn) are actually not in Eriador. Does he count them as Eriadoran population?

Yep:

So we are gonna ignore Meach, Senana, and Saralainn. Those aren’t actually part of Eriador.

That lowers his estimated population for Eriador to just 369,000, vs my estimate for TA 2915 in Eriador of 627,000.

One other thing to note is that he only gives population figures for regions with recognized polities. He assumes zero population in the lands between these polities. Using his map, the 6 polities he describes have about 32,000 square miles of territory, or a bit under 10% of the land of Eriador. So he assumes a flat zero population through 90% of the territory of Eriador (an egregious act of Tom-Bombadil-Erasure). I’ll note here that if the parts of Eriador he does not describe merely had the population density of medieval Siberia, there would be another 30–40,000 people in Eriador vs. what Morwinsky counts. I think Tolkien’s stray references to isolated people, Boromir getting directions, travelers on the road, trolls eating villages, etc all implies that Eriador is at least as peopled as literal Siberia.

Now, what do I estimate?

Well, our only perfectly matched region is The Shire. For the others, I take his land areas, and assume that the areas he describes have 4x the density of the region in which they occur, basically concentrating population in the areas he outlines. So for example, in 2915, I estimate a population density for Rhudaur of 0.6 people per square mile, very very low. But for the 6,442 square miles of area Morwinsky assigns to the 3 polities inside Rhudaur (Noddfa Rhaglaw, Arcirbann, and Dor-en-Egladil), I assume their territories have a density of 2.4 people per square mile, meaning the residual 35,000 square miles of Rhudaur must have extremely low density, etc. I also make an exception for Bree since it’s a known urban site and take a ratio of 6 times the background density instead of 4.

Using this method, here’s what I get:

Morwinsky grants a lot more hobbits than I do. And to be honest, he might be right. The Shire is pretty densely settled. You can make an argument for 300,000 hobbits without much difficulty.

Morwinsky has a much smaller Breeland. I discussed Bree in the first post. An issue is when Tolkien says there are 100 “stone houses of the Big People,” many readers assume that means there are just 100 houses. But that isn’t so — there could be wooden houses! We also know there are hobbit holes! Bree is the settlement at the junction of two significant roads with a long history, a stably reproducing local ethnicity (Breemen, descended from refugee Gwaithurim from the south), and rich agricultural lands around it. Bree itself could easily have 4,000 residents, plus another few thousand in Archet, Staddle, and Combe, plus their hinterlands and the hobbits all up on the hill. My estimate here is a bit high maybe but 10–20,000 residents of Breeland writ large is plausible.

Morwinsky also puts a lot more people in the non-canon Rhudaur realms, and especially the non-canon Sonn in Minhiriath. Sonn is a big issue for me because he puts more people in Sonn than I have in the entirety of Minhiriath.

Then I show the residuals. Breeland is within Arthedain; Morwinsky assumes zero other people in Arthedain (even the fertile and well-watered Evendim Hills) or Rhudaur or Minhiriath. Ultimately our Minhiriath estimates are somewhat compatible — 40–50,000. Or Rhudaur estimates are not terribly far off — he’s 37,000, I’m 25,000. We have a major Shire disagreement.

So in every place where Morwinsky provides estimates except Breeland, he’s actually higher than mine!

Why do I have higher overall population?

Because of areas Morwinsky doesn’t describe. I assume the areas outside his boundaries in Rhudaur have a density of 0.26 per square mile, basically like the arctic circle tribal Siberian populations of the middle ages. I assume Minhiriath outside the coast has 0.5 people per square mile (considerably less dense than medieval Mongolia!), which basically means I assume there are some squatters in the ruins of Lond Daer, or riverfolk up towards Tharbad, or some farmers here and there in the ruin of the old woods.

Arthedain is the big difference here. I assume Arthedain outside of Breeland has a density of 1.7 people per square mile. Many people will object to this. But this is similar to the medieval population density of African deserts, the Arabian interior, or Sweden minus the populous southern tip. 1.8 people per square mile is lower than many nomadic regions had! It’s totally consistent with a mostly-abandoned territory where maybe some squatter communities live in Fornost or Annuminas, pastoralists roam the Evendim Hills, rangers camp in the woods, a handful of “wild hobbits” dig holes in the downs, a petty dwarf here or there looks for iron, etc. And I think we can assume a lot of even that population lives between Bree and the Shire along the road.

But that’s actually not all — between Rhudaur, Arthedain, Minhiriath, and the Shire, we actually pretty much agree! I have 383,000 people, Morwinsky has 369,000.

But then I have 125,000 people in Cardolan. Boromir gets directions from someone, Arthedain tried to resettle it, it’s a major road, Tharbad was only lately abandoned, the Swanfleet was attractive to the Stoors, the vales of the Brandywine are rich and prosperous land with easy access to trade with the Shire, the Blue Mountains, and Lindon… it’s just unfathomable that Cardolan is totally empty. Again, we aren’t talking about cities and civilizations here. Pastoralists, nomads, hunters, isolated farmsteads, fishers by the river, maybe a few sequestered villages here or there. At 125,000 people, Cardolan’s density is still just 2.2 people per square mile. That’s similar to medieval Algeria including the Sahara desert! Basically, my working assumption here is that Tharbad as a city is gone, but that its former residents have dispersed into pastoralist, squatting in some corners of the city, foraging in the Swanfleet, or maybe isolated farming communities well away from the road.

If you want to argue for Cardolan at 1 person per square mile or even 0.5 you could credibly do so — but that is still going to yield tens of thousands of residents of Cardolan. Many fans object to the idea of anybody in Cardolan, so even these estimates are objectionable. For myself, given how excellent, strategic, and infrastructurally dense Cardolan’s land is, I’m comfortable saying that Cardolan’s surviving Middle- and Low-men slightly repeopled during the Watchful Peace.

I could do these summaries everywhere, but it’s easier to just show a graph of my numbers for each of my regions, vs. Morwinsky’s numbers. Also, I supplement his numbers by assuming that his “uninhabited” areas have 0.1 people per square mile.

We know that the Lune and Baranduin valleys were the major population centers for Arnor. The Lune valley in particular is almost totally sheltered from attach by the Shire, proximity to the elves and dwarves, etc. I just cannot accept the idea that nobody lives there. Somebody lives there! Not many somebodies, but somebody. Likewise with the areas between the Shire and the region I call “Arthedain” and the regions I label as Lindon. It’s in purple on my map. I just can’t fathom that nobody lives in that secure, stable, fertile, economically connected zone. Not many people maybe, but that’s obviously where anybody would flee to if they could.

Then we have Arthedain and Cardolan which I’ve already described. I also assume Angmar has a small human population: hillmen migrating in, maybe migrating Lossoth, maybe a few surviving Angmarim who’ve repopulated a bit. One of my more controversial assertions is that Eregion is unlikely to be totally empty. It’s fertile, it’s full of elf treasure, it’s on the doorstep of Dunland which we know is a demographically busy area with a market economy and active trade. If you want to mentally zero out Eregion you can, but I just can’t bring myself to do so.

So we end up in this weird situation where Morwinsky actually has higher estimates for many of the areas he estimates, but then just totally zeroes out everything else. I have somewhat lower estimates for areas we both estimate, but assume that various kinds of diffuse and disorganized peoples are likely to occupy empty, fertile, secure lands, especially central, western, and southern Eriador.

Ultimately, your mileage may vary, but here I’ve at least laid out the assumptions informing these choices.