There are plenty of mentions of battles and skirmishes in surviving early medieval documentary sources, ranging from Bede's Ecclesiastical History to Brittonic poetry such as Y Gododdin. Unfortunately, there are no detailed descriptions, and certainly no equivalent of accounting records or payment rolls that would show how many men were involved. Can we make even an approximate estimate of the likely numbers involved?
Old English original:
Ðeofas we hatað oð .vii. men; from .vii. hloð oð .xxxv.; siððan bið here.
--Laws of Ine, available online
Modern English translation:
13. §1. We use the term "thieves" if the number of men does not exceed seven, "band of marauders" [or "war-band"] for a number between seven and thirty-five. Anything beyond this is an "army" [here]
--Regia Anglorum website
Ine was a king of the West Saxons. He succeeded to the throne in 688 when his predecessor Caedwalla abdicated and went on pilgrimage to Rome, and ruled for 37 years according to Bede (Ecclesiastical History Book V Ch. 7). Ine's law code thus dates to the period between 688 and 725.
Never have sixty swordmen in a set fight
Borne themselves more bravely
--Translated in Alexander (1991)
The 'sixty swordmen' are followers of the Danish king Hnaef, attacked in a hall at night by their enemies. The date is uncertain (if indeed the incident is historical and not legendary). However, one of the participants is the (legendary?) warrior Hengest. If (a big if) he is to be equated with the Hengest who came to Britain at Vortigern's invitation in the mid-fifth century, the date of the Fight at Finnsburh would be somewhat before Hengest's removal to Britain.
A.D. 784. This year Cyneheard slew King Cynewulf, and was slain himself, and eighty-four men with him.
--Anglo-Saxon Chronicle, available online
An earlier entry in the Chronicle for the year 755 refers to the same incident (there positioned at the beginning of King Cynewulf's reign, whereas here it is in its more or less* correct chronological position at the end of his reign). Unlike the Chronicle's usual laconic entries, the 755 entry reads like a condensed form of a saga. Cyneheard, who was a kinsman of King Cynewulf and had some grievance against him, had killed Cynewulf and made a bid to become King of the West Saxons himself. The 755 entry makes it clear that all Cyneheard's followers chose to fight and die with him, indicating that 84 men was the size of Cyneheard's army.
Three Faithful War Bands
The War-Band of Cadwallawn, when they were fettered; and the War-Band of Gafran son of Aeddan, at the time of his complete disappearance; and the War-Band of Gwenddolau son of Ceidiaw at Ar(f)derydd, who continued the battle for a fortnight and a month after their lord was slain. The number of each one of the War-Bands was twenty-one hundred men
--Red Book of Hergest, available online
Three hundred gold-torqued men
Three hundred spirited horses
That charged with them
The thirty and the three hundred
Alas! They did not return
--Y Gododdin, stanza B1.9, translated by Koch (1997)
The twelfth battle was on Mount Badon in which there fell in one day 960 men from one charge by Arthur; and no one struck them down except Arthur himself, and in all the wars he emerged as victor.
--Historia Brittonum ch. 56, available online
It seems fairly clear that early medieval armies were not large, numbering a few dozen or a few hundred. All the early English sources give figures of less than 100:
These are broadly consistent with the sort of numbers that might have constituted one or a few ship's crews of the period. The large Sutton Hoo ship buried in Mound 1 had between 20 and 40 oars - as the tholes had only survived in places along the gunwale the exact number is unknown - and Martin Carver suggests the most likely number is 28, seven pairs each fore and aft of the mast (Carver 1998 p.171). As there would presumably have been people on board who did not row (e.g. the helmsman) the total crew would be more than the number of oarsmen. A ship's crew is a natural unit for a warband; in Beowulf, the hero and his followers sail in a single ship to Denmark to fight the monster Grendel. Taking Martin Carver's estimate of 28 rowers for the Sutton Hoo ship and adding on a few non-rowers would come to something resembling the 35-man upper limit for a warband in Ine's law code. A smaller ship would be consistent with a smaller warband. More than one ship, or a big ship carrying a significant number of extra warriors, would qualify as an army.
Men of very high status, such as kings, who controlled large amounts of resources, may have maintained larger warbands. A royal warband may have been expected to be larger than an 'average' warband, consistent with the numbers of 60 and 84 given for the Fight at Finnsburh and Cyneheard. A royal warband may have been indistinguishable from an army.
If warbands were typically 35 men or fewer, this does not preclude engagements involving larger numbers. Larger armies could have been assembled from multiple warbands, co-operating (to a greater or lesser degree) under a chief leader. Bede's Ecclesiastical History gives several hints that could be consistent with this type of 'modular' organisation. For example, at the Battle of Winwaed in 655, Bede describes Penda's army as thirty times greater than the army of Oswy and Alchfrid of Northumbria and "...comprised of thirty battle-hardened legions under famous commanders" (Book III Ch. 24). This is consistent with the idea that Oswy and his son Alchfrid between them commanded a single warband, presumably the royal retainers, while Penda had assembled a group of thirty warbands each of similar size. At the Battle of Degsastan in 603 Aethelferth of Bernicia defeated Aidan of Dal Riada, but "...Aethelferth's brother Theodbald and all his following were killed" (Book I Ch.34). This is consistent with Theodbald leading an independent or semi-independent unit, whose fortune in the battle differed from that of the unit(s) under Aethelferth's command.
The Brittonic sources suggest somewhat larger numbers, at 300 and 2100, respectively. Both are multiples of 300, which may indicate that they owe as much to poetic convention as to a muster roll. Nevertheless, they presumably give an idea of the sort of numbers that sounded reasonable to the intended audience. Y Gododdin presents elegies to fallen heroes from a variety of kingdoms, which would be consistent with a 'modular' alliance made up of multiple independent warbands under joint leadership, as suggested above. The Three Faithful Warbands triad claims that 2100 men comprised a single warband, a far larger number than any of the other sources. However, the Triads exist in a medieval manuscript which may have had numbers mis-copied over the years, or the numbers may have been increased to reflect medieval poets' ideas about the expected size for an army.
Arthur's (legendary?) battle at Badon has by far the largest number in any of the sources, at 960 casualties, implying that unless the casualty rate was 100% - surely an impossible feat even for the (legendary?) King Arthur - the number of participants was presumably considerably higher. The number itself may reflect poetic convention, miscopying or plain exaggeration. However, it is roughly three times the number of the doomed company of Y Gododdin, which would be consistent with Badon being regarded (in poetic convention and popular culture at least) as a battle of unusual size and importance. There is no indication of the size of the Arthurian force that inflicted the 960 casualties - assuming that Arthur had an army with him, and was not a superhero who despatched them all single-handed.
The limited documentary evidence available is consistent with fairly small numbers for armies in the early medieval period, of the order of magnitude of a few dozen or a few hundred, perhaps organised as independent or semi-independent warbands which could be assembled into a larger army under a common leader as occasion demanded.
Alexander M (translator). The earliest English poems. Penguin Classics, 1991, ISBN 978-0-140-44594-7.
Anglo-Saxon Chronicle, available online
Bede, Ecclesiastical history of the English people. Translated by Leo Sherley-Price. Penguin Classics, 1968, ISBN 0-14-044565-X.
Historia Brittonum ch. 56, available online
Koch JT. The Gododdin of Aneirin. Text and context from Dark-Age North Britain. University of Wales Press, 1997. ISBN 0-7083-1374-4
Laws of Ine, available online in Old English
Red Book of Hergest, available online
*'More or less' because the 755 entry says that Cynewulf reigned 31 winters, which would place the fatal fight in 786 rather than 784.