I

The British government had long considered Ireland a potential security risk. During the late nineteenth and early twentieth centuries, information on internal threats had been collected by the efficient R.I.C. and its sister organisation, the D.M.P. Political policing had featured in the R.I.C. remit and was augmented by detective special branches in the R.I.C. and D.M.P.Footnote ¹³ However, Paul McMahon has argued that by 1914 ‘The zeal and efficiency of Irish police forces in intelligence work had declined because of complacency, bureaucracy, and the expectation that major police reform would accompany Home Rule’.Footnote ¹⁴ From that vulnerable position, the Irish police were easy targets for the I.R.A., which enforced a popular boycott of the R.I.C. and began to assassinate aggressive members of both forces (especially detectives) from 1919 onwards. By early 1920, the R.I.C. intelligence network was in a shambles, while in the capital D.M.P. detectives from the ‘G Division’ had been targeted by Michael Collins’ I.R.A. intelligence ‘Squad’. Charles Townshend explained, ‘A steady sequence of shootings brought G Division to the point of paralysis’.Footnote ¹⁵

The Irish insurgency posed many challenges to British military intelligence as it emerged from the First World War. Recent achievements in cryptography and signals intelligence did not threaten an I.R.A. organisation which rarely communicated by telegraph or radio.Footnote ¹⁶ Counter-intelligence successes against German agents in Britain could not be replicated amid a more hostile civilian population in Ireland.Footnote ¹⁷ Proficiencies in aerial observation, front line cartography and field reconnaissance were of little use against republican guerrillas who did not hold fixed positions and seldom concentrated their forces. However, First World War military intelligence had developed effective systems regarding captured enemy documents and interrogations of prisoners of war, which were quickly applied to Ireland.Footnote ¹⁸ Overall, while many basic elements of military intelligence remained unchanged, the British Army had to adapt to a new operating environment and different type of enemy in Ireland. ‘From early 1920 the army began to develop its own intelligence system’, wrote Paul McMahon, continuing ‘the army had come late to the show, and lacked the basic knowledge at first, but out of necessity gradually built up a wide-reaching intelligence system’.Footnote ¹⁹

In April 1920, the British Army appointed General Nevil Macready as commander-in-chief in Ireland and he quickly recognised the shortcomings in intelligence. The government offered Macready unified command over both the army and the police in Ireland, which were operating independently of each other. Because Macready wanted the R.I.C. to remain autonomous, he rejected this offer and continued the crown forces’ awkward dual authority command structure. The new head of the R.I.C., General Hugh Tudor, sought someone to fill the new position of chief of Combined Intelligence Services, who would establish a central command of police intelligence, create a new clandestine service, and improve coordination with military intelligence.Footnote ²⁰ Tudor chose his wartime comrade, Colonel Ormonde Winter, an army artillery officer, lacking both policing and intelligence experience. While Winter controlled all of police intelligence in Ireland, he focused much of his attention and resources on Dublin.Footnote ²¹ As a result, the British Army assumed greater responsibility for crown force intelligence outside the capital.Footnote ²²

British Army units in Ireland were organised into four division-strength units by 1921, the 1st, 5th and 6th divisions, and the Dublin District.Footnote ²³ Headquartered inside Cork city's Victoria Barracks and operating on the southern part of the island, the 6th Division had an approximate strength of over 20,000 soldiers by June 1921, representing 35.7 per cent of the British Army troop total in the country. The division fielded twenty-seven army battalions (or equivalents) organised into four brigades (the 16th, 17th and 18th infantry brigades, and the Kerry Brigade), with most units raised in England.Footnote ²⁴ The 6th Division's operating boundary ran from County Clare on the west coast to County Wexford on the east coast, including the entire province of Munster and the Leinster counties of Kilkenny and Wexford.

The 6th Division's troops faced many of the I.R.A.'s most innovative and aggressive units in some of the most disturbed parts of Ireland. Republican guerrilla fighters hid their identity and obscured their organisation and operations from both the crown forces and the general population. The I.R.A. received significant logistical assistance from the female republican organisation, Cumann na mBan, which managed its own parallel structure of geographic units.Footnote ²⁵ The 6th Division faced an estimated 58,681 I.R.A. members, organised into twenty-one self-governing brigades, representing 51 per cent of the I.R.A.'s national strength.Footnote ²⁶ R.I.C. reports of 1,093 armed I.R.A. attacks on crown forces from March 1920 to July 1921 show that 650, or 59 per cent of the national total, occurred within the 6th Division area.Footnote ²⁷ Identifying and disabling the perpetrators of these attacks was a major challenge for crown force intelligence.

Army intelligence efforts in Ireland were modest in 1918. At that time, cursory two-page intelligence reports were drawn up for the Southern District (one of three in Ireland) by the district intelligence officer (Captain Maunsell). Every month he provided impressions of public opinion and major war-related developments, with information seemingly gathered from the press, legal proceedings and R.I.C. reports. While the Irish chief secretary complimented Maunsell for his efficiency, the system was inadequate to meet the more rigorous challenges which emerged from 1919. ‘These officers were not specialists, and the results achieved were limited’, concluded the British intelligence historian Anthony Clayton.Footnote ²⁸ After the 6th Division formation was formally established in Ireland during November 1919, new intelligence personnel were appointed at brigade and divisional headquarters.Footnote ²⁹ About this time, the army began to collaborate more closely with police information networks, also recruiting civilians ‘for intelligence purposes’. In west Cork in late 1919, the 1st Battalion of the Essex regiment appointed an intelligence officer (I.O.), and other battalions in the 6th Division subsequently followed this example.Footnote ³⁰

In January 1920, the army's Irish command instructed its division commanders to draw up arrest lists of I.R.A. officers and other prominent republicans. This was a watershed in the building of the 6th Division blacklist discussed below. The R.I.C. assisted in a joint roundup of republican suspects at the end of the month, but the underwhelming operation netted just fifty-four arrests across Munster, illustrating the steep decline of police intelligence. In early 1920, Irish-born Captain Campbell Kelly was appointed 6th Division intelligence officer. While the army noted his ‘exceptional local knowledge’, presumably owing to his Irish background, this was Kelly's first intelligence position. Lower down the divisional structure, intelligence postings were made slowly. Although the 17th Brigade (covering Cork city and the south and west of the county) quickly appointed an I.O., the same positions remained vacant within both the 16th Brigade (headquartered in Fermoy) and the 18th Brigade (headquartered in Limerick) until mid 1920, while the Kerry Brigade still lacked an I.O. in September. Delays in filling intelligence posts suggest a lack of urgency within the 6th Division headquarters.Footnote ³¹

Jim Beach has observed that during the First World War, the British forces lacked a unified espionage system of agents.Footnote ³² The same issue seems to have arisen in Ireland during 1919‒21. A Home Office report of secret service funding distribution in 1920‒21 reveals a striking imbalance of financial assistance across Ireland. Ormonde Winter's intelligence operation in Dublin received £29,800, while the director of naval intelligence, Hugh Sinclair, was issued £25,000 from a total outlay of £77,000. General Strickland's 6th Division received a paltry £2,200, though that amount was higher than the sums for the 1st and 5th divisions.Footnote ³³ This suggests that Winter's Dublin network depended more on secret service agents and paid informants than the army units in the provinces. The 6th Division appears to have utilised its own personnel and those of the R.I.C., as well as a network of ex-soldiers and sympathetic civilians.

II

The post-war reflections of Major Arthur E. Percival, stationed with the Essex regiment (1st Battalion) in west Cork during 1920‒21, provide personal insights of an active intelligence officer. In January 1920, Percival's unit lacked information on its adversaries, and found the R.I.C. of little assistance except as guides and sources of local gossip. However, by systematically collecting information, Percival developed an understanding of his district and its population, which helped him generate a census-like profile of each household to determine residents’ political sympathies. He also attempted to list the I.R.A. order of battle, and identify the names of I.R.A. officers, their physical descriptions and where they lived. Percival noted the persistent challenge of detecting I.R.A. prisoners who wore civilian garb and feigned innocence when captured, and were often inadvertently released.Footnote ³⁴ Nonetheless, Percival made rapid advances in building a blacklist in his battalion area. By the autumn of 1920, it contained 168 names largely from the Bandon Valley area.Footnote ³⁵ Major Percival's status as one of his battalion's senior officers (he outranked both his brigade and division intelligence officers) probably allowed him to funnel greater battalion resources into intelligence. He steadily built his network between January and September 1920, and believed this early start was invaluable. Some of his information came from captured I.R.A. company membership rolls, but he also managed to turn a few I.R.A. prisoners who provided valuable inside information.Footnote ³⁶ Additional intelligence came from civilian residents, including some members of the local Protestant community, to a greater extent than I.O.s received in other districts.Footnote ³⁷ The overall effort produced solid results, with a number of I.R.A. companies in the Essex operating area suffering some of the highest arrest levels in the 6th Division.Footnote ³⁸

Anthony Clayton described the battalion intelligence officers operating in Ireland: ‘Some of these were very effective, but the work of many was amateurish and impeded by other unit duties.’Footnote ³⁹ A more limited and typical British intelligence organisation was maintained by 2nd Lieutenant John Basil Jarvis, of the 1^st Battalion, Oxfordshire and Buckinghamshire Light Infantry (Ox and Bucks). Educated at the Royal Military Academy Sandhurst, Jarvis was appointed the Ox and Bucks I.O. based in Limerick city, just five months after being commissioned. Aged 20, he possessed neither intelligence experience nor prior knowledge of Ireland. His surviving 1920 diary shows Jarvis worked part-time on intelligence; his low military rank and inexperience suggests his superiors prioritised conventional soldiering over information-gathering, an attitude that appears to have been relatively common among British military forces. By September 1920, Jarvis joined raids in Limerick, wrote district intelligence summaries for his brigade intelligence officer, examined captured documents and liaised with the R.I.C. He made numerous journeys into nearby County Clare on a motor bike, led foot and bike patrols, and joined raids by lorry-borne troops.Footnote ⁴⁰

There is no doubt that some advances were made. By late 1920, a 6th Division photographic bureau had been established in Cork city to record the images of I.R.A. prisoners. The divisional intelligence staff expanded to three officers and two clerks, while each of the division's four brigades were allocated two I.O.s, a clerk and a photographer, which made coordination easier. At this stage the 6th Division I.O. reported some forty-five agents in his network.Footnote ⁴¹ They probably resembled agents described by an army intelligence officer in the 5th Division area, whose contacts included railway porters, shopkeepers and bartenders.Footnote ⁴² The I.R.A. believed the crown forces exploited their ex-servicemen network for intelligence purposes, and frequently assassinated ex-soldiers as suspected civilian informants.Footnote ⁴³ For example, the ex-soldier, John Dwyer of Castletown-Kenneigh, was picked up by the I.R.A. in January 1921 but mistook the republicans for British soldiers. He then revealed his status as an informant to the British army in Bandon receiving a weekly payment of £5 (this remuneration can be compared to ten shillings a day paid to seventy five agents in June 1920 which had been approved by the War Office for intelligence work in Ireland).Footnote ⁴⁴ The Secret Service fund referenced above allowed the 6th Division to make cash awards not exceeding £50 for information given, which were authorised by the division commander, Major General Sir Peter Strickland.Footnote ⁴⁵ Although the documentation is poor, it also appears that additional information was gathered by imported British secret service agents in Munster cities. For example, during late April 1921 Wilfred Ewart noted the presence in Cork city hotels of government agents and ‘stray Englishmen’.Footnote ⁴⁶

The government extended martial law to the entire 6th Division area on 5 January 1921. This brought unified army command over both police and military intelligence operations under the divisional intelligence officer, Captain Kelly. In Cork city and its suburbs, at least, greater co-operation was obtained by amalgamating the Hampshire Regiment battalion I.O. office and the R.I.C. district commissioner's office in April 1921.Footnote ⁴⁷ This ‘Local Centre’ was directed by one of Cork's most successful intelligence officers, Lieutenant Alfred Koe, who transferred to the R.I.C. For example, during an army raid on Cork City Hall in August 1920, Koe recovered incriminating documents used to prosecute Lord Mayor Terence MacSwiney.Footnote ⁴⁸ The I.R.A.'s First Southern Division intelligence officer, Florence (Florrie) O'Donoghue, took particular interest in these developments, noting to Michael Collins that Koe ‘has taken over an office in Moore's Hotel Cork (recently commandeered by police). He has a staff of men and women clerks, and works under the divisional commissioner. Apparently this is one of the new information centres.’Footnote ⁴⁹ As a garrison battalion in the division's hottest intelligence frontier, Cork City, Koe's Hampshire Regiment (2nd battalion) enjoyed special prominence within 6th Division intelligence structure. The battalion's commanding officer, Lieutenant-Colonel C.N. French, spent much of his time at the War Office on ‘special duties’, and ran a conference in October 1921 for all intelligence officers in Ireland.Footnote ⁵⁰

The British army had first recommended creating the army/R.I.C. liaison office system in March 1920. By the Truce (11 July 1921), though, ‘Local Centres’ only existed in Cork city, Limerick city and Clonmel, with the last-mentioned office established just days before the cease-fire.Footnote ⁵¹ Nonetheless, the 6th Division official history claimed that had hostilities resumed, ‘Local Centres’ would have been extended across the 6th Division. In October 1921, further optimism was voiced in Chester at the intelligence officers’ conference, which concluded that ‘the Intelligence Organisation in the autumn of 1921 was very complete’.Footnote ⁵² However, in most of the 6th Division area, police and military intelligence integration remained elusive. Moreover, smoother collaboration with the R.I.C., did not offer the army a silver bullet for its Irish problems. Lieutenant Grazabrook, the Gloucester regiment I.O., noted in April 1921 that his battalion worked successfully with the Kanturk and Newmarket R.I.C., but found the police in places like Banteer unreliable and their information useless.Footnote ⁵³ Lieutenant Jarvis, who finished the war in charge of a small detachment in Tulla, County Clare, reflected gloomily: ‘lack of information was a severe handicap to the detachment of the regiment.’Footnote ⁵⁴ A memoir from the South Staffordshire regiment (2^nd Battalion) based in Cork city recalled that British soldiers were ‘very much handicapped by a total lack of intelligence, except in those less affected districts where the R.I.C. were still able to operate’. It also noted that paid informers had almost ceased to exist and the situation only began to improve in March 1921.Footnote ⁵⁵ In short, British military intelligence capabilities were enhanced by organisational innovations and the deployment of additional specialist personnel, but at a glacial pace which was inadequate for the dire situation facing much of the division area. British post-conflict intelligence assessments subsequently over-estimated their impact.

III

Army information was derived from a combination of civilian informers, captured republican documents, and I.R.A. prisoners and deserters. Intelligence from civilians largely dried up owing to the I.R.A.'s fierce and lethal campaign against suspected civilian informants, which according to the 6th Division official history, ‘had the result of making information very hard to obtain.’Footnote ⁵⁶ Seized republican documents were particularly important, including those captured from I.R.A. general headquarters in Dublin. According to an army report, ‘these documents were not only the foundations on which the I.R.A. List and Order of Battle were built, but each seizure usually led to further raids and the capture of more documents’.Footnote ⁵⁷ While a limited number of informants could be found within the I.R.A. ranks, their numbers were few and their positions low in the rebel hierarchy.Footnote ⁵⁸

More significant information appears to have come from I.R.A. prisoners. During the First World War, British Army intelligence developed structures and techniques to secure and exploit information from captured enemy personnel.Footnote ⁵⁹ Prisoner interrogation systems appear to have been quickly adopted in the 6th Division, including the prompt and systematic screening and questioning of republican prisoners from a centralised division ‘cage’ in Victoria Barracks.Footnote ⁶⁰ Captain Kelly likewise planted disguised intelligence officers amid captured republican prisoners, which he claimed led to the elimination of an entire I.R.A. flying column — Kelly was referring here to the destruction of the East Cork (4th Battalion, Cork No. 1 Brigade) flying column at Clonmult.Footnote ⁶¹ British leverage over prisoners increased after the government began to execute I.R.A. prisoners captured with arms. Indeed, 6th Division military courts accounted for fourteen of the twenty-four republican executions during the War of Independence.Footnote ⁶²

Michael Occleshaw has noted that the British archives rarely describe British army prisoner-of-war interrogations during the First World War.Footnote ⁶³ The same holds for Ireland in 1919‒21, even though the British army's official history admitted that its best information came from I.R.A. deserters and prisoners.Footnote ⁶⁴ Ormonde Winter acknowledged that some information was obtained from prisoners but did not explain the circumstances. In contrast, Republican contemporary documents and post-conflict accounts provide ample testimony of British physical and psychological abuse of captured I.R.A. personnel in the 6th Division area.Footnote ⁶⁵ These included severe beatings, mock executions and extended periods of maltreatment, often intended to extract information, with numerous army intelligence officers directly named. When torture allegations did become public, such as the 1920 case of west Cork I.R.A. leaders Tom Hales and Pat Harte, they were denied. Ormonde Winter wrote that Tom Hales ‘was induced to impart the names of most of his officers, and in order to indemnify himself from the results of this information he accused the Crown forces of torture in their endeavour to make him speak’.Footnote ⁶⁶ British propaganda conceded that Hales and Harte were beaten in custody, but claimed they were not tortured.Footnote ⁶⁷ Such refutations appear to have been made to cover up practices referenced by numerous prisoners. Paul McMahon concluded that, ‘Some prisoners gave information to escape punishment; others gave information unwittingly to skilled interrogators; torture, though officially deemed counter-productive, was used in many cases’.Footnote ⁶⁸

The British were also aware that the telephone and mail systems were rich sources of information for I.R.A. intelligence (for example, the government attempted to purge disloyal postal workers).Footnote ⁶⁹ Those same communication networks were likewise exploited by British I.O.s. For example, the I.R.A. detected a telephone operator, Miss Carroll, at Killarney railway station phoning Lieutenant Sherwood (the brigade I.O.), who frequently visited her house. Lieutenant Casse, the intelligence officer for The Buffs regiment in Fermoy, was similarly observed calling to the Ballyhooley Post Office postmaster and his daughter, Miss Margaret O'Toole. The I.R.A. further noted his close friendship with the postman in Fermoy, John Maguire, who later bid Casse a hearty farewell at the train station when he departed Ireland with his regiment.Footnote ⁷⁰ Such relationships allowed army I.O.s to challenge republican dominance of civilian communications networks, yet they appear to have been wholly dependent on the uneven capabilities of individual intelligence officers.

Intelligence duties outside the office could be risky. The army sent some of its I.O.s into the Irish countryside where they were easily identified. Lieutenants Rutherford and Brown left Fermoy on motorbikes dressed in mufti, but they were captured, executed and buried in Rusheen in mid Cork by I.R.A. volunteers.Footnote ⁷¹ The I.R.A. shot and secretly buried Lieutenant Vincent at Watergrasshill, County Cork in May 1921, after he unwittingly called at a republican household while gathering information. At the nearby Fermoy railway station, republicans collecting Vincent's kit found a camera and a notebook naming his R.I.C. and civilian contacts.Footnote ⁷² In November 1920, Captain Thompson, the Manchester Regiment I.O. at Ballincollig Barracks, was intercepted on his motorcycle, executed, and dumped in a turnip field near Cork city.Footnote ⁷³ The same month, the I.R.A. shot off his motor bike Lieutenant Hambleton, the Northamptonshire Battalion intelligence officer, near Nenagh, county Tipperary.Footnote ⁷⁴ The 6th Division I.O., Captain Kelly, escaped multiple I.R.A. assassination attempts in 1920 and 1921, and his I.R.A. counterpart Florrie O'Donoghue distributed his photo with the instructions that ‘constant lookout should be kept for this man, and every effort made to get him’.Footnote ⁷⁵ Numerous other military intelligence officers became known locally for their part in home raids or interrogations, and their names remained embedded in republican memory decades after the conflict's conclusion.Footnote ⁷⁶ While the 6th Division area did not experience a spectacular equivalent of Bloody Sunday, I.R.A. attacks on suspected crown force intelligence assets in the south were more numerous and sustained than in Dublin.

The intelligence war was certainly not entirely one-sided in the 6th Division area, as the British scored some notable victories. Individual I.R.A. companies were crippled by well-coordinated military raids, such as the disaster at Caherguillamore in East Limerick during December 1920, in which five Volunteers were killed, many more captured and, in the words of historian Tom Toomey, ‘the Bruff battalion … practically ceased to exist’. The British Army reported 148 arrests during the raid; the I.R.A.'s Bruff company reported eleven arrests during the war and Grange company fifteen.Footnote ⁷⁷ The information that led to the arrest of the Cork 1 Brigade commander Terence McSwiney and much of his staff in August 1920 came from a letter seized in the post by army intelligence officers in West Cork.Footnote ⁷⁸ A raid on Rahanisky House near Cork city netted much of the Active Service Unit of the First Battalion, Cork 1 Brigade (the agent who provided the information, Bride McKay, had direct contact with Captain Kelly). Most of the intelligence successes noted in the 6th Division's official history were centred in County Cork, reflecting the intense intelligence focus there from the 6th Division and 17th Brigade headquarters at Victoria Barracks, Cork city.Footnote ⁷⁹ However, a post-conflict army appraisal also noted that the 17th Brigade intelligence officer was preoccupied with the Cork city I.R.A., to the detriment of his work in rural areas.Footnote ⁸⁰

Peter Hart argued that by this stage of the conflict, most British army battalion I.O.s in County Cork had at least one informer in the I.R.A. ranks,Footnote ⁸¹ and ‘most battalions and Auxiliary companies had a very accurate picture of the organisation and men they were fighting’.Footnote ⁸² Likewise, William Sheehan referred to a robust British Army intelligence operation that by July 1921, which ‘knew the IRA order of battle in Cork ‒ the structure of the Southern Division, the composition and location of its battalions and companies’.Footnote ⁸³ Undoubtedly, by the July Truce, British intelligence military had improved significantly. However, did crown forces have a good picture of critical I.R.A. personnel by the summer of 1921 across the entire 6th Division? The answer to this question requires deeper analysis of intelligence outcomes.

IV

Shortly before the Truce, the crown forces reached their operational peak, making it possible to assess the nature, scope and quality of British efforts by analysing its intelligence products in this period. These included the 6th Division weekly intelligence summary, which outlined the general insurgency situation and summarised developments in each I.R.A. brigade area. Based on a First World War model, the weekly reports were shared with I.O.s throughout the division along with top national police and army leaders, and they were also used to brief the British cabinet.Footnote ⁸⁴

These updates show both the centrality of army information gathering and its uneven intelligence assessment in the closing stages of the conflict. For example, a report in late April 1921 quotes General Strickland as acknowledging that his martial law area, ‘contained the counties where the rebels had made most headway. And that the situation was rapidly becoming worse.’Footnote ⁸⁵ He further conceded that incoming information was limited in the 6th Division, except in Cork city.Footnote ⁸⁶ The following week, a different report tracked I.R.A. flying columns in the 6th Division and detailed military operations and I.R.A. attacks on crown forces. The British identified a rebel concentration in the neighbourhood of Headford Junction in the I.R.A.'s Kerry No. 2 Brigade area. However, little information was forthcoming on the I.R.A. in its adjoining Kerry No. 1, Kerry No. 3 and West Limerick brigades. The same report optimistically assumed the Cork No. 1 Brigade ‘was in a bad way’ owing a lull in armed activity. It did not detect that the Cork 1 Brigade was preparing to launch a series of coordinated attacks in May, which caused heavy casualties among the Crown forces.Footnote ⁸⁷ While these reports provided an abundance of localised detail, they lacked an analytical overview which might have helped the Cabinet better formulate policy.

Army battalion I.O.s working at a local level provided the foundations from which the divisional reports were compiled. The Grazabrook diary (held at the Gloucester Regiment Museum) contains numerous single paged weekly intelligence summaries produced by the Gloucestershire Regiment (1st Battalion) stationed in Kanturk and dated from April 1921 onwards. The 1st battalion intelligence officer wrote up the summary based on information collected from captured documents, R.I.C. reports, and prison and civilian interrogations. The material was assessed, organised, and then shared with other battalion and brigade I.O.s. The Gloucester battalion intelligence officer also received important information from his Brigade I.O., Lieutenant F.C. Sherwood. The Grazabrook diary contains a surviving 6th Division weekly intelligence summary for the week ending 13 June 1921, which was largely compiled from numerous battalion intelligence summaries such as that produced by the Gloucester I.O.Footnote ⁸⁸ Intelligence officers within the R.I.C. Auxiliary Division operated a similar system in which the company I.O. was expected to prepare a ‘Weekly Intelligence Summary’.Footnote ⁸⁹

The assorted army battalion intelligence reports gave 6th Division leaders an overview of developments across the division, though the picture was often incomplete. For example, the 6th Division weekly intelligence summary for the week ending 13 June 1921 outlines various British military operations against the I.R.A. in the division area. It also noted ‘more than usual’ I.R.A. attacks on crown forces. The intelligence summary explained that Cork 1 Brigade commander Sean Hegarty had escaped his pursuers in the Clydagh Valley, in contrast to the army's capture of Cork No. 2 Brigade commander Sean Moylan. It also stated that ‘the rebels’ in north Cork around Cullen were ‘avoiding contact with Crown forces’.Footnote ⁹⁰ This division report written in the last weeks of the war contained some solid information about fugitive I.R.A. leaders and guerrilla concentrations but was overly optimistic about the general situation. Such reports reflected improving army information-gathering but inadequate analytical structures, a problem that had also plagued British Army intelligence during the First World War.Footnote ⁹¹

V

A more critical and assessable British Army intelligence product were the 6th Division ‘Blacklists’ used to identify important I.R.A. members for arrest and detention. Military intelligence officers compiled these lists from captured documents, R.I.C. data, local informants and open sources such as newspapers.Footnote ⁹² A complete blacklist dated from June 1921 can be found in the papers of Lieutenant John Jarvis (noted above), naming 847 republicans wanted for arrest in the British Army's 6th Division area.Footnote ⁹³ Distilled from the military's various intelligence assets, it gives historians a valuable tool to analyse the state of British military intelligence in the south of Ireland during the final phase of the conflict.

Individual entries were organised alphabetically in each of the eight counties within the 6th Division, providing every suspect's name, age, place of residence, occupation, I.R.A. rank, and a brief physical description. Leading republicans tended to get more space. For example, the entry for the notorious I.R.A. officer Dan Breen reads:

However, most entries were far more cryptic, such as: ‘Bartholemew Murphy, Templebryan, North Cork. Cork: Adjutant “On the run”. Farmer Age 30.’

Even prior to the First World War, the British Army emphasised and organised the collection and analysis of captured documents, practices that were followed in Ireland.Footnote ⁹⁴ Numerous blacklist entries refer to captured I.R.A. documents provided by Ormonde Winters’ centralised Raids Bureau, a national intelligence office which summarised and redistributed seized documents to British I.O.s across the country. For example, from County Kerry:

Here, ‘Mulcahy's documents’ refers to I.R.A. chief of staff Richard Mulcahy, whose papers were captured in Dublin during November 1920 and February 1921, while the ‘Eileen McGrane's documents’ were a significant set of Michael Collins’ records seized in the flat of his assistant Eileen McGrane.Footnote ⁹⁵ While such Dublin raids could yield valuable insight into the provincial I.R.A., the time needed to index and distribute the captured documents reduced their immediate operational value. The Raids Bureau system moved far too slow for British units in the field.

Greater centralisation extracted information from different intelligence sources and branches, which collectively surface in the blacklist. For example:

Much of this detail was correct. But the profile lacked a physical description, even though O'Donoghue was a pallbearer in Tomás MacCurtain's and Terence MacSwiney's funeral processions during 1920, which were witnessed by tens of thousands of spectators. It also omits O'Donoghue's former role as Cork 1 Brigade Intelligence Officer and current position as adjutant and intelligence officer of the I.R.A.'s First Southern Division.Footnote ⁹⁶

Numerous entries were derived from suspects’ previous arrests, though the utilisation of prison records differed significantly between counties. The variable quality of entries is also apparent in those noted to be ‘on the run’ (fugitives who did not sleep at home for fear of arrest). For example, in County Cork ninety-nine suspects were deemed on the run or roughly 31 per cent of all Cork entries. Of these, forty-six were situated in the Cork 3 Brigade area (west Cork), an imbalance best explained by the 1920 Essex regiment blacklist, which recorded this precise information because its intelligence officer recognised its importance.Footnote ⁹⁷ In other counties, the number of identified I.R.A. fugitives was far lower. In addition, the British may have not successfully merged their army and police suspect lists. The R.I.C. Auxiliary Division company stationed in Dunmanway in late 1920 evidently used the early Essex list referenced above to build its own local blacklist. Revealingly, many of the new I.R.A. names added to this Dunmanway list by Auxiliary intelligence officers, did not find their way onto the 6th Division blacklist, which suggests problematic integration between army and police systems even late in the conflict.Footnote ⁹⁸

The 6th Division Blacklist reveals significant regional discrepancies. Table 1 lists the number of suspects identified in each county. The second column shows that well over a third of suspects (38 per cent) were located in Cork, as compared to relatively few identified in Waterford, Wexford and Kilkenny. However, the imbalance is less striking after accounting for the I.R.A.'s overall numeric strength in Cork. The third column provides percentages of I.R.A. members identified on the blacklist out of the total I.R.A. strength in each county, which demonstrates how relatively few volunteers ended up on the blacklist. These range from a low percentage of 0.7 per cent of I.R.A. members in Kerry to a high of only 2.1 per cent in Tipperary. Column four in Table 1 checks the number of women identified on the list as a share of total of suspects, per county. The I.R.A. integrated female activists (usually members of Cumann na mBan) in its communications, safehouse and intelligence networks.Footnote ¹⁰⁰ Perceptive British military intelligence officers understood the critical role played by women in the I.R.A. organisation and thus identified important female activists. Women comprised a significant portion of the blacklist in counties Tipperary (26 per cent), Cork (17 per cent) and Waterford (16 per cent). However, no women are returned for Counties Wexford, Clare and Limerick, showing a less sophisticated British appreciation of the republican organisation there.

Table 1: 6th Division BlacklistFootnote ⁹⁹

On the surface, the blacklist's descriptive detail seems impressive, but alternative ways are needed to judge its reliability. One method is to assess how many leading I.R.A. figures were named on the list. To do so, we constructed an ‘IRA Brigade Leadership Sample’ for nineteen I.R.A. brigades operating in the British 6th Division area, broken down by county.Footnote ¹⁰¹ (In the I.R.A. hierarchy, a neighbourhood or village comprised a company; a district unit made up of several companies was called a battalion; a number of battalions formed a brigade.) For each I.R.A. brigade, we identified all members of the brigade headquarters staff at the time of the 11 July 1921 Truce, along with the four senior officers of every battalion (using the ranks of commander, vice-commander, adjutant and quartermaster). Our officer lists come from the I.R.A. nominal rolls, produced during the 1930s as part of a government military pensions scheme and now digitally available via the Military Archive's Military Service Pensions Collection.Footnote ¹⁰³

The I.R.A. brigades in the British Army's 6th Division area represented a nominal strength of 58,681 I.R.A. volunteers. Our I.R.A. brigade leadership sample totals 650 officers across that area, of whom only 146 are identified on the blacklist: at 22.5 per cent, this accounts for less than a quarter of the I.R.A.'s leading officers in the 6th Division area (see Table 2). The quality of the 6th Division Blacklist varied greatly between counties. For example, much of the I.R.A. elite in Wexford and Waterford remained undetected, while British military intelligence showed greater awareness of the I.R.A. leadership in counties Cork, Clare and Kerry. Even in those three counties, roughly two-thirds of senior I.R.A. officers were either unknown to the British or not recognised as important. This further demonstrates the uneven and deficient state of British military intelligence across the 6th Division.

Table 2: I.R.A. Brigade leadership sample identified in 6th Division BlacklistFootnote ¹⁰²

The I.R.A. brigade leadership sample also reveals a robust guerilla organisation battling the 6th Division at the end of the War of Independence. The M.S.P.C. nominal rolls list four senior brigade officer positions (commander, vice-commander, adjutant and quartermaster) for nineteen I.R.A. brigades.Footnote ¹⁰⁴ Of those seventy-six staff positions, only two were vacant on 11 July 1921.Footnote ¹⁰⁵ Of four senior battalion leadership positions (commander, vice-commander, adjutant and quartermaster) reported for 111 I.R.A. battalions in the 6th Division area, accounting for 444 senior battalion positions, just ten were vacant at the time of the Truce. This indicates that the brigade and battalion levels of the I.R.A. hierarchy continued to function despite British counterinsurgency measures, which again questions the effectiveness of intelligence-led operations targeting the I.R.A. leadership.

Limited army intelligence successes, such as those achieved against the Cork No. 3 Brigade, did not necessarily severely damage the I.R.A. The British identified 47 per cent of the Cork No. 3 Brigade leadership, easily the 6th Division's best performance. However, all four of the Cork No. 3 Brigade senior staff positions were filled, along with twenty-seven of twenty-eight battalion senior staff positions. In addition, the R.I.C. reported thirty-eight armed attacks on the crown forces in the Cork No. 3 Brigade area during 1921, including twenty-one from May to 11 July.Footnote ¹⁰⁶ Overall, the I.R.A. campaign in the 6th Division area grew during the final stage of the War of Independence. Using police reports of armed attacks on the crown forces (R.I.C. ‘Outrages against Police’ daily/weekly reports) in the 6th Division area, totals can be provided in three-month increments.Footnote ¹⁰⁷ From April through June 1920, there were seventy-two armed attacks on the crown forces in the 6th Division area. The figure jumped to 103 attacks from July through September 1920, and then levelled off at ninety-seven from October through December 1920. From January through March 1921, there were 128 attacks in the 6th Division; that total spiked to 189 from April through June 1921. These figures show that the 6th Division was clearly not defeating the republican insurgency in 1921, which we again link to inadequate British military intelligence.

The I.R.A.'s durability is further corroborated by arrest data acquired from M.S.P.C. unit roll returns (see Table 3). The British government used republican internment arrests and court-martials as a metric to measure its counter-insurgency success. General Macready provided these figures weekly to the British cabinet from early 1921,Footnote ¹⁰⁸ while Dublin Castle included internment arrest totals in its weekly press updates of the British campaign.Footnote ¹⁰⁹ Arrest returns have been retrieved from 896 I.R.A. companies across Ireland, using the term ‘arrest’ to describe those I.R.A. members imprisoned at the time of the July 1921 Truce. These 896 companies comprise 64,136 members, or roughly 55 per cent of the total I.R.A. volunteers in Ireland.Footnote ¹¹⁰ A sample of 2,990 prisoners from those units has been created, which accounts for 48.78 per cent of the estimated 6,129 republicans jailed at the Truce of 11 July 1921.Footnote ¹¹¹

Table 3: Sample of I.R.A. membership in companies returning arrest data organised by county. Estimated percentage of arrested members per the total returned by county.Footnote ¹¹²

I.R.A. arrests often followed from good intelligence, making that metric critical to any appraisal of British intelligence effectiveness. The arrest data sample confirms that I.R.A. units in the 6th Division area suffered very minor losses in terms of imprisonment. These range from a low of 1.55 per cent of sampled members in County Kerry (73 arrested out of 4,722 members), to a high of 6.4 per cent of sampled members in County Wexford (65 arrested out of 1,016 members). While some companies took heavy casualties, they were the exception rather than the rule. Of the 428 sampled I.R.A. companies in the 6th Division area, 230 (53.7 per cent) had either zero members or one member imprisoned by the July 1921 Truce. Across the 6th Division area, the arrest data shows that a mere 3.76 per cent (1,240) of the I.R.A.'s 32,986 sampled members were imprisoned by July 1921. This was lower than the national share of 4.82 per cent arrested across all of Ireland. However, it differed sharply from the Dublin city I.R.A., which saw arrests of 13.01 per cent (565) of 4,343 sampled members jailed in the same period, revealing considerably more successful outcomes for the British forces there.

Table 4 shows the number of internment arrests made by each of the British Army's four divisions. Using the troop strength in each army division, it then calculates the average number of arrests per British soldier. Once again, the results suggest the 6th Division underperformed compared to the other three British Army division formations in Ireland.

Table 4: Internment arrests by British Army divisions by the TruceFootnote ¹¹³

VI

As seen in tables 3 and 4, British success in the Dublin District compared to the 6th Division, is striking. One possible explanation for the discrepancy is Dublin's difficult urban operational landscape, which may have constrained I.R.A. mobility and made members more easily accessible to searches and random stops. However, the urban environment by itself does not explain Dublin's high arrest rates, which far exceeded those in Munster's major urban centres. For example, in Limerick city, the combined arrest rate for four companies was 2.07 per cent of members; in four Waterford city companies the arrest rate was 4.9 per cent; in Cork city, the combined arrest rate for ten companies was 7 per cent.Footnote ¹¹⁴ The presence there of 6th Division G.H.Q., a large military garrison, and relatively good army/R.I.C. cooperation appears to have yielded somewhat better results for the British in Cork city, yet I.R.A. losses remained far lower than in Dublin.

Another factor in the 6th Division's lower arrest rate was it faced more aggressive and better armed I.R.A. units than elsewhere.Footnote ¹¹⁵ The persistent threat of I.R.A. attack hindered British mobility in many parts of the divisional area. The I.R.A. in the 6th Division also accounted for 62 per cent of suspected civilian informers killed throughout the country in 1920–21, while it secretly executed crown force intelligence personnel (and suspected personnel) on a larger scale than elsewhere.Footnote ¹¹⁶ Taken together, I.R.A. lethality reduced both the submission and collection of information in many areas.

However, we flag two critical features of the British counter-insurgency campaign in Dublin, which produced the highest arrest rate per soldier among the four British divisional formations in Ireland. As mentioned earlier, British police intelligence in Dublin fell under the direct supervision of Ormonde Winter, who allocated much of his organisational resources to the capital. In addition, the Dublin's military garrison was much larger and more concentrated than the rest of the country. This is evident when considering the Dublin District ratio of British troops compared to local I.R.A. members than the 6th Division area. In the latter area, 20,018 soldiers faced a nominal I.R.A. strength of 58,681 members, while in Dublin city, 11,412 British soldiers faced an estimated 5,316 I.R.A. members.Footnote ¹¹⁷ Rounding those figures up, in the 6th Division there were roughly two British soldiers for every six I.R.A. members, compared to a Dublin District rate of two British soldiers for every one I.R.A. member. The relative troop saturation in Dublin differed fundamentally from the British army's thin spread across the 6th Division. Overall, the British invested far more intelligence assets and troop resources in Dublin than in its southern theatre, which explains the superior results achieved.

VII

In the wake of the British Army's demoralising retreat from Ireland in 1922, its leaders struggled to explain their inability to defeat the I.R.A. In the intelligence sphere, a lack of unified command was blamed for many crown forces shortcomings. Writing in 1925, the former commander of British force in Ireland, General Nevil Macready, argued ‘it would have been better to have relied on a purely military organisation, to be placed at the service of the police, instead of attempting a dual organisation, which, with the best goodwill on both sides never worked altogether smoothly owing to a diversity of system and the lack of unity of control.’Footnote ¹¹⁸ However, this article has shown that unity of command by itself would not have produced the outcome sought by the British administration. Under martial law in the 6th Division area, the British Army largely controlled both police and military intelligence, but its results were not any more impressive than elsewhere.

The 6th Division's post-conflict assessment concluded that a ‘fairly efficient’ intelligence system was built across the division.Footnote ¹¹⁹ But such innovations were only gradually implemented in Cork city, arrived quite late elsewhere and collectively made only limited inroads against the I.R.A. by the time of the Truce. I.R.A. brigade and battalion leadership positions in the division area remained filled. Armed attacks on the crown forces in the 6th Division area peaked during the war's last three months. Our analysis of the Division blacklist reveals a low level of identification of the I.R.A. leadership cadre across the 6th Division (Table 2). These findings are further corroborated by our I.R.A. arrest data sample (Table 3), which confirms that the 6th Division achieved lower arrest levels than the other three British divisional formations in Ireland. While some districts within the 6th Division produced high arrest rates, such performances were exceptional; more districts experienced few or even no arrests. Only a small fraction of I.R.A. members in the 6th Division area were jailed by the Truce of 1921, amounting to 1240 of 32,986 members, or merely 3.76 per cent of sampled members.Footnote ¹²⁰ This evidence leads us to conclude that the British counter-insurgency campaign in the 6th Division area had not succeeded by the closing stages of the conflict. We contend that poor intelligence was a critical element in this failure, and resulted from an inadequate investment of troops, resources and specialist intelligence personnel in the 6th Division area. As a point of comparison with the British campaign, one can consider much more impressive results achieved by the national army during the Irish Civil War (1922‒3), when during less than a year of fighting it imprisoned over 13,000 anti-Treaty I.R.A. members (roughly double the British total) and so thoroughly defeated the I.R.A. that it called a unilateral ceasefire in 1923.Footnote ¹²¹

The tendency of Irish War of Independence intelligence historians to view Dublin in isolation has skewed the overall national intelligence picture, as has the focus on Michael Collins. Our findings regarding British Intelligence in the 6th Division area show much more varied regional outcomes. We also believe the evidence discussed here justifies a reappraisal of the Anglo-Irish intelligence war in Dublin, including a more critical view of Michael Collins and his celebrated I.R.A. intelligence network. The high rates of I.R.A. incarceration in Dublin compared to the 6th Division area may also help explain why the subsequent Anglo-Irish Treaty received significant support from the I.R.A. leadership in the capital, compared to its near unanimous rejection by the Munster I.R.A. leadership. Finally, we advocate new studies of British intelligence efforts in the British Army's 1st and 5th divisions for comparative purposes, and to help complete the overall national picture.

While the 6th Division intelligence system enacted numerous improvements and innovations, these were more impressive to their creators than the relatively weak results warranted. British intelligence failings in its southern theatre followed a problematic pattern within Ormonde Winter's Irish intelligence system. Writing internally in April 1921, Lieutenant General Macready remarked that ‘everything seems to point to the view that Winter has not got the right method, and we here very much doubt whether he will ever get it’.Footnote ¹²² Macready's negative assessment broadly concurs with our qualitative and quantitative evidence regarding the 6th Division intelligence system during the Irish War of Independence. It also helps explain why the British government ultimately lost faith in its ability to secure a military victory in Ireland and focused instead on securing a negotiated political settlement with their Irish adversaries in the summer and autumn of 1921.