The harvest mice (Micromys Minutus) in ‘Manawydan’ (the ‘Third Branch of the Mabinogi’)


Harvest mouse nest by kind courtesy of Derek Crawly from the Mammal Society.

Harvest mouse nest by kind courtesy of Derek Crawley from the Mammal Society.

Species mentioned: A swarm of very hungry harvest mice.

Source: ‘Manawydan’ the ‘Third Branch of the Mabinogi’.

Date of Source: c.1000-1250.

Highlights: An angry sorcerer takes revenge by turning his whole family into harvest mice to try and eat his enemy’s crops. Amazingly this doesn’t work out as he planned, and one harvest mouse is sentenced to be hanged as a thief.


THE TEXT AND ITS DATE (scroll down for the animals)

The ‘Four Branches of the Mabinogi’ are the most well read and important pieces of Middle Welsh prose literature. They have been widely available in translation for more than a century now together with various other pieces of medieval Welsh literature (look for any book called The Mabinogion, but don’t trust Lady Charlotte Guest’s edition.) The four branches can be read independently but there is some continuation between the branches and some characters which appear in multiple branches.

The date of the texts is uncertain.  All four texts appear in full with only minor differences in the White Book of Rhydderch (dated c.1350) and the Red Book of Hergest (dated c.1382). However there are also fragments of ‘Branwen’ and ‘Manwydan’ in Peniarth 6 (dated c.1250) This gives us a TAQ (terminus ante quem) for the texts – they must have been first composed earlier than the mid-thirteenth century. Various dates of composition have been given since then from 1000-1250 A.D. Most scholars are comfortable with a date in the middle of this range c.1100.

The most recent scholar to try and establish an authoritative date for the text was Simon Rodway (2007). He looked at the ending of the third person singular past tense, which is ‘-ws’ or ‘-wys’ early on and becomes ‘-awdd’ later on. There are three important problems with his approach (i) scribes often modernised texts as they copy them, (ii) the % of relevant verbs using the ‘-wys’ ending changed only 8% between 1100-1250, and we already know the texts were written prior to 1250 and (iii) authors were probably capable of using archaic language, which makes dating hard to assert on this standard alone.

The date-range of 1000-1250 A.D. is sufficiently clear for our purposes.



Early on in the story the four main characters are placed under a curse. The curse is set by a sorcerer called Llywd Cil Coed. Imagine him like a stereotypical evil wizard (i.e. a walking plot device).  He wants revenge for his friend, who in the ‘First Branch of the Mabinogi’ tried to steal a woman who didn’t want him. He was helped in this endeavour by the main character’s gambling problems, and only saved because the woman in question (Rhiannon) was clever enough to come up with a way around it.

This friend’s story ended shortly later when he was tied up in a sack and beaten like a piñata. This game is called badger-in-the-bag, from which we may assume that the ordinary victim was not a human. I mention the story because it illustrates the way humans have justified the persecution of animals (sport). If badger-kind had only produced evil sorcerers like humankind, perhaps sports like this would never have become popular.

The curse tightens around the characters as they go through the story. At first all the humans and domestic animals in Dyfed (south-west Wales) disappear. Next two of the companions are cursed and cannot move from the place. Finally the two remaining companions try farming and plant three fields of wheat. In the harvest season a horde of mice come to eat all of their crop, one field per night. The main character keeps watch over the field on the third night and witnesses the destruction:

  • He went to keep watch over the field. While he was doing so, towards midnight, he heard the loudest noise in the world. He looked. There was a huge army of mice – they could not be counted or measured. The next thing he knew, the mice were making for the field, and each one was climbing up along a stalk and bending it down, and breaking the ear and making off with the ears, and leaving the stalks behind. And as far as he knew there was not a single stalk there without a mouse to it. And they ran away carrying the ears with them. Then, enraged and angered, he charged in among the mice. He could no more keep his eye on one of them than on the gnats or the birds of the air. But he could see that one was very fat, and unlikely to be able to move quickly. He went after that one and caught it, and put it in his glove, and tied the mouth of the glove with the string, and kept hold of it and made for the court. (Davies,2007, p.42)

By the end of the story the mice are revealed to be shape-changed humans, and the heavy one is revealed to have been pregnant as a human.



Is it possible to identify the species of mouse imagined by the author? For a linguist alone it is not. The Welsh term used ‘llygod’ is still the basis for the modern words for a mouse (llygoden), rat (llygodyn fawr), shrew (llyg) and vole (llygoden gota), and at the time of Manawydan these names had not separated.

A second difficulty with a species level identification is that the idea of a swarm or plague of devouring rodents was a very popular one even in the pre-medieval period (Davies, 2006, p.239). Similar stories can be found in Pliny’s ‘Naturalis Historia’, VIII:56. and the Bible, 1 Samuel: 6:4-6. Rodents do occasionally swarm, but accounts of this were very popular, and most authors probably copied the motif into their own stories without ever having seen it happen.

However in this case it is possible that the account was based on (exaggerated) real-life experience of animal-human conflict on an agricultural farm. There are five important species indicators in the text:

  1. Diet – the creatures eat wheat.
  2. Social behaviour – the creatures travel like an army and all attack the wheat at the same time. They avoid him as a swarm.
  3. Time of day – the creatures only come at night time.
  4. Eating pattern – the creatures climb to the tip of the stalk, making it bend slightly before breaking off the ear and leaving with it.
  5. Size – able to fit into (one finger?) of his glove.

Based solely on the creature’s diet and approximate size, we can confine the search to mouse, rat or vole (meaning the animal intended is not a shrew, mole, rabbit, hare or weasel).

Here are the mice, rats and voles found in Britain, together with notes on their suitability. Source of ecological information is the Handbook of British Mammals (Harris & Yalden, 2008):

Species Likelihood Notes
Harvest mouse(M. minutus) Most Likely Specialises in arable land. Historically rare in Wales. Does eat wheat above ground level as described. Semi-territorial. Does not significantly affect harvest.
Wood mouse(A. sylvaticus) Likely Commonly found in agricultural fields. Found at high densities (-200/ha in winter) Nocturnal. However tend only to eat shed seeds.
Field vole(M. agrestis) Unlikely Not common in agricultural land. Eat stalks and leaves of crops, not seeds. Semi-territorial. Too big for a glove finger.
Ship rat(R. rattus) Unlikely Rarely found outdoors but love cereal grains. Present at high densities. Too big for a glove
House mouse(M. domesticus) Unlikely Rarely found outdoors, but favours grain as a food. Present at high densities.
Bank vole(M. glareolus) Unlikely Rare in fields, although found in margins. Main diet supplied from trees (seeds, bark, leaves). Too big for a glove finger.
Yellow-necked mouse(A. flavicollis) Impossible Found mainly in ancient woodland. Not associated with agricultural areas.
Water vole(A. terrestris) Impossible Rarely strays more than 2m from water. Eats mainly grass.
Hazel dormouse(M. avellanarius) Impossible Species does not leave woodlands, and spends most of life high above ground level.
Edible dormouse(G. glis) Impossible Species not introduced until later.
Brown rat(R. norvegicus) Impossible Species not introduced until later.


There are two likely species, the harvest mouse and wood mouse. Both species sometimes (i) eat cereals, (ii) are generally nocturnal, and (iii) would fit in one finger of a glove.

Although neither mouse species steals whole heads of grain like the mice in the story, the feeding pattern described here more probably belongs to the harvest mouse. Wood mice are heavier and they rarely eat growing grains. When they do attack crops, they gnaw through the stalk half way up rather than trying to reach the tips like harvest mice (Heroldová & Tkadlec, 2011). The harvest mouse leaves the stem intact, climbs to the top of it and eat grains there leaving a recognisable sickle feeding pattern (Harris & Yalden, 2008 p.122).

Harvest mouse

Harvest mouse by kind courtesy of Derek Crawley from the Mammal Society.

On the other hand, the feeding pattern may resemble that of a harvest mouse, but the swarming does not. The damage caused by rodents to crops on the field is not usually significant, except under certain circumstances (like swarms). This element of the story seems the most fantastic, and the least likely to be inspired by a living model. Most probably the story was inspired by a popular legend about swarming rodents, but informed by contemporary observations of (tiny and relatively benign) harvest mice eating grain. If that is the case then a harvest mouse interpretation fits well.



Harvest mice distribution in Britain, copyright National Biodiversity Network.

Harvest mice distribution in Britain, 10km resolution. Copyright National Biodiversity Network. Use allowed for educational purposes.

Ecologically, the importance of a harvest mouse in Dyfed, south-west Wales is high. At the moment there are very few records of mice in Wales except along the coast lines. This record suggests that historically harvest mice may have been present in the area. However there is a caveat to this interpretation. Although the story is set in Dyfed, it could easily have been composed somewhere else entirely. Even if it was composed in Dyfed, the author might be basing the account on observations made elsewhere in Britain. This is a tentative probability, not a strong basis for definite conclusions about presence/absence.

From a literary point of view, this nature-sensitive reading has provided a new interpretation to the text. It suggests that the story does not need to be read as only a literary motif, but may have been influenced by natural observation of mice. This is contrary to the pattern for medieval artwork which does not seem to be commonly based on the real world (Raye, 2013).

Historically, it suggests that medieval land-users were aware of mice. We can only hope that farmers were more tolerant of mice than the hero of our story; Manawydan takes the mouse he has caught and attempts to have it hanged as a thief. This scene is comical but may be symptomatic of the real issue of human-wildlife conflict at an early date.

© Lee Raye 2014.



Davies, S. (trans. 2007) The Mabinogion. (Oxford World’s Classics)

Harris, S. and Yalden, D.W. (ed. 2008) Mammals of the British Isles, Handbook, 4th ed. (The Mammal Society, Southampton).

Heroldová, M. & Tkadlec, E. (2011) Harvesting behaviour of three central European rodents. Crop Protection 30:82-4.

Raye, L. (2013) Lions in Insular British Artwork: 650-1000 A.D. Journal for Late Antique Religion and Culture 7:72-89.

Rodway, S. (2007) The Where, Who, When and Why of Medieval Welsh Prose Texts: Some Methodological Considerations. Studia Celtica 41:47-90.





Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s