No precisely dateable material has been found to establish exactly when the Romans first built a fort at the place they called Vindolanda, but pottery finds indicate that it was in the mid to late 80s – during that post-Agricolan rethink of strategy when the Romans abandoned Caledonia. This fort was, of course, timber-built, and it was the first in a series of timber forts on the site. The second, built c.92, was double the size. The third timber fort, c.97, was apparently a refurbishment of the second. The next major rebuild, the fourth timber fort, which was evidently larger still, was in c.105. The end of the fourth fort is loosely dated round-about 120. However, a late modification to its layout was the addition of a very large timber courtyard-building, superior in style and construction to the norm, and it has been suggested that this “palatial building” was constructed to accommodate Hadrian’s retinue, on the emperor’s visit to Britain, which was in 122, and that the fort was re-built to make space for the legionaries drafted-in to construct Hadrian’s Wall.[*] Occupation of this fifth timber fort probably ended in the mid to late 120s, when the garrison moved to new accommodation on the Wall. Vindolanda was not abandoned, however, though the fort was reduced in size. Development continued, in stone, and it was occupied until the end of Roman rule in Britain, and beyond.[*]

Vindolanda is famous for the hundreds of writing tablets – slivers of wood, written on with pen and ink (chance survivors of the rubbish tip) – that have been unearthed since 1973. The tablets are associated with the first four timber forts, and have survived by being sealed in oxygen-free conditions deep below the level of the final stone buildings. One tablet (Tab.Vindol. 154), found in the ditch of the first fort (mid to late 80’s–c.92), is evidently a strength report of the unit based at Vindolanda at that time:

18th May, net number of the 1st Cohort of Tungrians,
of which the commander is Julius Verecundus the prefect,
752, including 6 centurions,
of whom there are absent:
Guards of the governor46
At the office of Ferox at Coria337
including 2 centurions (?)
At London1 centurion (?)
including 1 centurion
including 1 centurion
At (?) …1 (?)
Total absentees456
including 5 centurions
Remainder, present296
including 1 centurion
Of these: 
Suffering from inflammation of the eyes10
Total of these31
Remainder, fit for active service265
including 1 centurion

Virtually two thirds of the unit, the 1st Cohort of Tungrians (Cohors I Tungrorum),[*] is away from base – almost half at Coria, i.e. Roman Corbridge, a dozen miles to the east. (In fact, Vindolanda’s first fort was rather small for the whole cohort.[*]) And more than a tenth of those on base are unfit for duty – whether the “wounded” were the victims of battle or accident cannot be gauged. Some tablets show the kind of everyday duties that would occupy the average squaddie:

24th April, in the workshops 343 men.
Of these: shoemakers, 12
builders to the bath-house, 18
for lead …
saw-makers (?)
… hospital …
to the kilns …
for clay …
plasterers …
for … tents (?) …
for rubble …
Tab.Vindol. 155 (2nd or 3rd fort)
7th March
sent to build a guest-house for Marcus, the medical orderly, builders, number 30 [*]
to burn stone, number 19 (?)
to produce clay for the wattle fences of the camp …
Tab.Vindol. 156 (3rd fort)

One tablet attributed to the third fort (c.97–c.105), seemingly the end of a longer document, has a comment concerning the natives’ fighting style:

… the Britons are unprotected by armour (?). There are very many cavalry. The cavalry do not use swords nor do the Brittunculi mount in order to throw javelins.
Tab.Vindol. 164

The derogatory nickname Brittunculi, which is translated along the lines ‘wretched little Brits’, is known only from this tablet.

Most tablets, by far, have been found in third-fort contexts – particularly amongst the leftovers from bonfires. It is clear that the garrison was the, part-mounted, 9th Cohort of Batavians (Cohors VIIII Batavorum equitata).[*] Probably the best known of the Vindolanda Tablets is a birthday party invitation to Sulpicia Lepidina, wife of Flavius Cerialis, a prefect of (i.e. commander of) the 9th Cohort of Batavians during the third-fort years, from Claudia Severa, wife of Aelius Brocchus, apparently commander of a unit stationed at a nearby fort (perhaps at Coria, but not certainly known):

Tab.Vindol. 291. The tablet was folded, along a central score, and would have been tied closed. The address is on the back: “To Sulpicia Lepidina, wife of Cerialis, from Severa.”
Claudia Severa to her Lepidina greetings. On 11th September, sister, for the day of the celebration of my birthday, I give you a warm invitation to make sure that you come to us, to make the day more enjoyable for me by your arrival, if you are present (?). Give my greetings to your Cerialis. My Aelius and my little son send him (?) their greetings. I shall expect you, sister. Farewell, sister, my dearest soul, as I hope to prosper, and hail.[*]
Tab.Vindol. 291

The 9th Cohort of Batavians left Britain, never to return, in about 105.[*] Vindolanda would then seem to have been (mainly) garrisoned, once again, by the 1st Cohort of Tungrians.[*] The 1st Cohort of Tungrians may have moved to the nearby fort of Vercovicium (Housesteads), on Hadrian’s Wall, when it was first commissioned in the mid to late 120s – the unit certainly moved to Housesteads at some stage, as attested by several undated inscriptions, and the Notitia Dignitatum places it at Housesteads around the year 400.[*] At this time – near the end of Roman rule in Britain – Vindolanda was garrisoned by the 4th Cohort of Gauls.[*] The commanders of the garrisons at Housesteads, Vindolanda, and the other forts categorized as “along the line of the wall” in the Notitia Dignitatum, were under an officer with the title dux Britanniarum (Duke of the Britains), whose command evidently comprised all Roman forces in, what is now, northern England.[*]

Robin Birley Vindolanda: a Roman Frontier Fort on Hadrian’s Wall (2009), Chapter 8.
(Vindolanda is not, of course, actually on Hadrian’s Wall – it is a mile south of the Wall.)
Tabulae Vindolandenses. See Vindolanda Tablets on the Roman Inscriptions of Britain website.
The timber fort phases underlie the later phases of stone fort and attached civilian settlement (vicus). There were evidently nine forts (the last three, starting c.160, stone) – the second built on top of the demolished and levelled remains of the first; the third on top of the demolished and levelled remains of the second; and so on. The lowest level can be up to 7 metres below current ground level.
The Tungrian homeland was in modern-day Belgium.
According to Robin Birley*, the first fort was no bigger than 3 acres, but a size nearer 5 acres would be expected for the full cohort. Robin Birley (who was, for 36 years, Director of Excavations for the Vindolanda Trust) suggests that the larger second fort was built (c.92) when the detachment at Corbridge was reunited with the rest of the unit at Vindolanda.
* Vindolanda: a Roman Frontier Fort on Hadrian’s Wall (2009), Chapter 5.
Notitia Dignitatum (Register of Dignitaries): a list of Roman civil and military posts, compiled round-about 400. (See The Roman Army in Britain Part II.)
The highlighted reading is suggested by Bowman and Thomas in an Appendix to Tabulae Vindolandenses III (2003). Originally, in Tabulae Vindolandenses II (1994), the authors had: “sent with Marcus, the medical orderly, to build the residence”.
The 1st Cohort of Tungrians was seemingly based in Britain for more than three centuries – as a name anyway, it’s hard to imagine there could have been anything remotely ‘Tungrian’ about it by the end.
written in 1937, by W. H. Auden (1907–73).
Over the heather the wet wind blows,
I've lice in my tunic and a cold in my nose.
The rain comes pattering out of the sky.
I'm a Wall soldier, I don't know why.
The mist creeps over the hard grey stone.
My girl’s in Tungria; I sleep alone.
Aulus goes hanging around her place,
I don't like his manners, I don't like his face.
Piso’s a Christian, he worships a fish;
There'd be no kissing if he had his wish.
She gave me a ring but I diced it away;
I want my girl and I want my pay.
When I’m a veteran with only one eye
I shall do nothing but look at the sky.
The highlighted closing sentences are written in a different hand from the rest of the letter – undoubtedly that of Severa herself. Hers is said to be the earliest known writing in Latin by a woman.
As published in Tabulae Vindolandenses II (1994), A.K. Bowman and J.D. Thomas had the reading “for … wagons” here. However, in an Appendix to Tabulae Vindolandenses III (2003) the authors are sure this initial reading was incorrect, and suggest “saw-makers” or possibly “sawmen”.  “24th April” (viii Kal. Mai.) is also an amendment – previously it was thought to be “25 April” (vii Kal. Mai.).
Incidentally, this tablet could belong to the second fort or the third, it is not possible to place it in either with certainty – Robin Birley Vindolanda: a Roman Frontier Fort on Hadrian’s Wall (2009), Chapter 6.
By the end, Roman Britain was divided into five provinces, though where the boundaries between them were is not certainly known. The Duke’s area of command, which, to judge by the Notitia, was the territory north of a Humber–Mersey line, would appear to have encompassed more than one of them, hence ‘of the Britains’.
The Batavians’ homeland, in modern terms, was in the Netherlands. Incidentally, the 9th cohort’s numeral is invariably written as VIIII on the Vindolanda Tablets, but the abbreviation CIXB (Cohors IX Batavorum) is incised on leather off-cuts that have also been preserved.
Evidently, both the 1st Cohort of Tungrians and the 9th Cohort of Batavians were stationed at Vindolanda during the period of the second fort (c.92–c.97). It would appear that this larger fort was built for the Tungrians, but that they were then replaced by the Batavians, who subsequently made the third-fort improvements and, presumably, changes required by their cavalry.
It is clear from many tablets, and the leather off-cuts, that the 9th Cohort of Batavians garrisoned the third fort, but there are a couple of tablets (Tab.Vindol. 263 & 311) that make mention of the 3rd Cohort of Batavians. Whilst the whole or part of the 3rd cohort, or individuals from it, may have been present at Vindolanda at some stage, it is believed that they were never actually based there.
Tab.Vindol. 180 & 181 are from a fourth-fort context (c.105–c.120+). 180 records a distribution of wheat to “the legionary soldiers”. 181 names, amongst others, “the Vardullian cavalrymen” as owing money. These latter were presumably a detachment of Cohors I Fida Vardullorum civium Romanorum, a part-mounted (equitata) cohort, whose homeland was in northern Spain, and who are attested, by a diploma (CIL XVI, 43), to have been in Britain by 98.
When the 1st Cohort of Tungrians were first based at Vindolanda (c.85–c.92), the strength report (Tab.Vindol. 154) demonstrates that the unit was on-its-way to becoming milliary, i.e. nominally ‘one thousand strong’. By 103, according to a military diploma of 103 (RIB 2401.1), it was milliaria, but in a diploma of 122 (CIL XVI, 69) it is not milliaria – perhaps a sign of heavy Roman losses in warfare at the start of Hadrian’s rule. By the time the cohort was involved in building the fort at Castlecary, on the Antonine Wall, it was evidently at milliary strength once again (RIB 2155). Incidentally, inscriptions show that the commanders of both the Tungrian cohorts in Britain (the 1st was an infantry unit, the 2nd was part-mounted) were prefects even when the units were milliaria (e.g. RIB 1580 and RIB 2104), and the commander of the 9th Cohort of Batavians, which may well have been milliary whilst stationed at Vindolanda, was a prefect not a tribune.
Dating references in Tab.Vindol. 581 certainly place the 9th Cohort of Batavians at Vindolanda until 30th April 104, and probably until 16th July 104. (An argument that this latter date relates to 105 is convincingly countered by Bowman and Thomas in Tabulae Vindolandenses III, freely available online.) The unit was in Raetia (comprising parts of: Austria, Switzerland, Germany) by 116 (RMD 155).
The notion, mainly based on tile-stamps from Buridava, that the 9th Cohort of Batavians fought in Trajan’s second Dacian war (105–106) – which resulted in Rome’s annexation of Dacia (modern Romania, more or less) – has been shown to be mistaken. The unit would appear to have transferred from Britain directly to Raetia. See: Florian Matei-Popescu ‘On the Presence of the Cohort IX Batavorum Milliaria Equitata in Moesia Inferior’, Acta Musei Napocensis 41-42/1 (2007), freely available online.
A military diploma (CIL XVI, 55) provides testimony that the 3rd Cohort of Batavians, mentioned in a couple of Vindolanda tablets, had transferred to Raetia by 107. It seems reasonable to suppose that both the 3rd and the 9th Cohorts transferred to Raetia c.105 – presumably it was the withdrawal of these and other units from Britain that obliged the Romans to abandon, for the time being, southern Scotland. Incidentally, both the 3rd Cohort, in the diploma of 107, and the 9th, in the diploma of 116, are described as milliaria, i.e. ‘one thousand strong’.
(See The Roman Army in Britain.)
Corpus Inscriptionum Latinarum
Roman Inscriptions of Britain
Roman Military Diplomas
The commanders of all cohorts named in the Notitia are tribunes, “even though it seems unlikely that any of the cohorts had an operational strength near to 500, let alone 1,000, men”, writes Michael G. Jarrett: “Details of the change from prefects to tribunes are not clear, nor are the dates at which it was taking place; but on the evidence currently available it is reasonable to suppose that after c.240 the title tribunus cannot, on its own, be taken as evidence for a milliary cohort.” (‘Non-Legionary Troops in Roman Britain: Part One, the Units’, Britannia Vol. 25, 1994.)