See Maps 25 and 26

IT IS appropriate at this stage to glance back at events as they had appeared to the Duke of Aosta. After the capture of the frontier posts in Kenya and the Sudan in July 1940, and their success in British Somaliland in August, the Italian High Command in East Africa settled down to a defensive policy, qualified only by Marshal Badoglio's request that the Viceroy should study the possibility of moving into the Sudan at the same time as Graziani resumed his offensive and advanced to the Suez Canal. The arrival of the 5th Indian Division in Port Sudan and its move forward were of course noted by the Italian Intelligence, who greatly overrated the numbers of 'Anglo-Indians'. The British attack on Gallabat early in November made no particular impression on the Viceroy, nor, in his view, on Ethiopian opinion either, but by the end of the month he judged that further British attacks were imminent: from Gallabat towards Gondar, and from the Kassala area towards Agordat and Keren. He expected that these would begin almost simultaneously, and be followed later by offensives into Galla-Sidamo and against Kismayu, and perhaps also by an attempted landing in British Somaliland as a preliminary to an advance on Harar.

Faced with all these threats the Viceroy decided that the system of holding the reserves centrally would have to be modified, for although the Italians were on interior lines the distances were very great and motor fuel and tyres were already seriously short. He decided therefore to send forward some of his reserves nearer to the northern front, where he expected the first attacks to occur. If other fronts were also attacked they would have to look after themselves as best they could, playing for time and avoiding decisive engagements; he hoped that if the main attack were held, the lesser ones could be dealt with later. He accordingly allotted four colonial brigades and various other units from his reserve to General Frusci's northern sector and brought two more brigades to a central position at Dessie.

He then ordered the commanders in Eritrea, Amhara, and Gojjam—Generals Frusci and Nasi—to organize areas of resistance which


were to be defended to the end. Mobile forces, chiefly colonial brigades, were to be concentrated between these areas and the frontier and were to hold positions from which they might issue to attack the enemy's flanks and rear. A retreat would have a very bad effect on the loyalty of the native troops and their conduct could not be foreseen, whereas the effect of checking the enemy's advance early would undoubtedly be encouraging and would mean that resistance might be prolonged for some months—into the rainy season. General Pinna was to continue to control the Air Force, but his units were to be dispersed throughout the sectors in order to be in close touch with the sector commanders.

The dispositions in the vital areas at the end of December were broadly as follows. In Eritrea General Frusci had three colonial divisions, three colonial brigades, and garrison and other troops. At Gallabat and Metemma he had three colonial brigades, and in and north of Gondar five Blackshirt battalions and numerous native bande. In Gojjam, General Nasi had four colonial brigades (twenty-three battalions in all) and several bande. In Shoa, around Addis Ababa, were the Savoia and Africa Divisions, which with other units amounted to twelve Metropolitan and Blackshirt battalions and twelve colonial battalions. Two more colonial brigades were in reserve at Harar. The ten colonial brigades, and Blackshirts and bande in the southern and Juba sectors, were regarded as a necessary minimum and could not be drawn upon.

By now the Viceroy was becoming increasingly pessimistic. He felt that his power to defend the Empire was declining; resources were being used up without being replaced, while economic difficulties were causing widespread discontent and a large part of the forces had to be ready to keep internal revolt in check. The morale of the British, on the other hand, had been stimulated by the success of their offensive in Cyrenaica, and their forces in the Sudan and Kenya were being rapidly strengthened. They were sure to attack before long, and the Viceroy expected the first blow to fall in the northern sector. He had confidence in General Frusci, however, and felt that he would be able to resist successfully; if he gave way 'everything would crumble'.

On 11th January the Viceroy explained that the British mechanized troops would have such an advantage on the flat ground that he wished to withdraw most of his forces from the Sudan-Eritrea frontier to the eastward of Tessenei, Sabderat, and Keru; none of the ground west of Agordat and Barentu would hamper mechanized action sufficiently. Next day Mussolini approved the Viceroy's proposals, expressed confidence in his leadership, and reminded him that the destiny of Africa would be decided by events in Europe. The Viceroy thereupon gave orders for the evacuation of Kassala,


Tessenei, and Gallabat-Metemma. These operations were to be supported by the Air Force.

The withdrawal was, on the whole, successful, though the 41st Brigade's action at Keru was disappointing. A disquieting feature was that continual retreating had greatly disheartened the forces, especially the native troops, and the Viceroy ordered that there was to be no withdrawal from Agordat and Barentu, and that a regiment of his best troops—the Savoia Grenadiers—was to be sent from Addis Ababa to Keren to help 'to close the gap absolutely'. On the other fronts all was comparatively quiet. An attack on Kismayu might soon occur, but he decided not to, hold this place and ordered everything of use to the enemy to be removed from it. At the same time he ordered General De Simone to hold the line of the Juba to the last; however, after General Cunningham's first successful attack he told him to disengage and fall back to cover the Jibuti railway. In the west, the front between Lake Rudolf and the Blue Nile was quiet. In Gojjam there was always the chance of a flare-up if the Italians suffered many reverses, and the Viceroy decided to simplify the system of command by creating a new western sector under General Nasi to include the whole of Amhara and Shoa, thus leaving General Frusci responsible only for meeting the British advance in Eritrea, while General Nasi dealt with the Patriots.

By the middle of February the Duke of Aosta thought he had discerned the aims of the British, which were to capture the Italian colonies of Eritrea and Somalia and reconquer British Somaliland with their own forces, while helping the ex-Emperor with money, arms and aircraft to recapture Ethiopia. He protested to Rome that the forecasts which had been thought so pessimistic were being proved accurate. He had sent all the reinforcements he could spare to the northern sector, and realized that nothing could now reach the Empire from outside except aircraft, and for these he asked, if only to enable him to carry out essential reconnaissance. Out of 188 aircraft at the middle of February, only 67 were serviceable; a month later these figures had dwindled to 39 out of 150.

The Duke believed that so long as any Italians held out anywhere the Empire could not be called lost. He proposed therefore to give ground where resistance would be useless and gradually to concentrate his troops and Italian civilians in districts whose inhabitants were loyal, and to hold out there until the end of the war. He himself would fight and govern to the end. These plans were approved generally by Mussolini, who laid down that the Viceroy's object should be to gain time and await the outcome of events in Europe; he was not to abandon the capital unless absolutely compelled, because this would be politically equivalent to losing the Empire. As for aircraft, the loss of Benghazi had made it more difficult to fly


them to East Africa, but they would continue to be sent. In the past two months 20 aircraft had arrived in this way, of which eight were fighters.

The Viceroy could now do little more than scrape up reinforcements for Keren and Massawa and hope for the best. On 16th March he wrote to Mussolini bemoaning his many difficulties, notably the lack of transport, anti-tank weapons, and aircraft, and the poor quality of the Italian officers serving in the Empire. The white troops varied—when led by good officers they were excellent —but the bulk of his forces were natives, who would follow strong masters for the sake of prestige and reward, and only the simplest could now believe that the Italians were stronger than the British. They regarded an order to retreat as a proof of weakness, and would desert sooner than obey if retreat meant leaving their own countryside and abandoning their families to unknown dangers. He commented bitterly on the poor spirit of the Italian civilians, who seemed unwilling to make the sacrifices imposed on them by the war. In response to his request for a forecast of the probable course of the war elsewhere, he was informed that the Axis powers would probably occupy Egypt by autumn 1941 and that the existence of organized forces in the Empire would obviously be a help towards conquering the Sudan and British East Africa. Therefore everything possible should be done to prolong the resistance of the remaining forces.

On 24th March the Viceroy ordered De Simone to resist as long as possible on the Awash, while Nasi's forces in Gojjam were to prevent the rebels from advancing into Shoa. Everything useful was to be removed from Addis Ababa, and all troops (other than those required for protecting the Italian civilians from the natives) were to move north and create a southern front for General Frusci. There would be three fortified areas of final resistance, in Eritrea, Gondar, and Galla-Sidamo. Once more the Duce approved, and exhorted the Viceroy to resist to the utmost.

As soon as Gazelle Force could get clear of the mines and demolitions to the east of Agordat, they pressed on boldly along the Keren road.1 On the afternoon of February 2nd the leading troops reached the foot of the wall of hills that shuts off the Keren plateau from the valley below; a barrier pierced only by the Dongolaas gorge, which carries on its eastern slopes the road, and on its western the light railway. A length of the road, well inside the gorge, was found to be completely blocked, and Gazelle was brought to a halt. It took the British nearly eight weeks to break through this barrier of hills, which was resolutely defended by some of the best Italian troops.


The battlefield of Keren
Map 29. The battlefield of Keren

To east and west of the gorge a number of peaks rise steeply to more than 2,000 feet above the floor of the valley. The names given to the features on the west were, in turn, Sanchil; Brig's Peak, Saddle, Hog's Back, and so on westward to Samanna. To the east the hills were slightly lower, though no less severe. Dologorodoc stood on a jumble of under-features and was surmounted by a concrete field-work—Fort Dologorodoc; farther east, Falestoh, joined to its northern neighbour, Zeban; eastward again, Zelale, or The Sphinx; and between Falestoh and Zelale was a tangle of rugged ridges across which a track led northwards to Keren over the precipitous Acqua Col. Sanchil and Brig's Peak were of particular importance because they gave observation all the way to Keren. Cameron Ridge was a halfway house to Sanchil and the other peaks to the westward.

On this formidable position the Italians had already deployed one colonial brigade and the three battalions of 11th Grenadier Regiment of the Savoia Division from the general reserve at Addis Ababa, a sure sign that the High Command fully intended to hold the plateau. In addition, on the route from Karora were the two brigades of the 1st Colonial Division, whose commander, General Carnimeo, was placed in command of the whole front. In the back area were the remains of three brigades that had been at Keru and Agordat. One brigade arrived next day from the general reserve at Addis Ababa, and in the Arresa area were the two brigades which had withdrawn through Barentu, and the one from Um Hagar.

Much as General Platt disliked the idea of a head-on collision, there was no other practicable way of reaching Asmara. The track through Arresa was useless for vehicles, while the route through Karora, which approached Keren from the north, might be of use for a diversionary move, but would not support the traffic of a large, force. Therefore, although the 7th Indian Infantry Brigade Group from the Port Sudan area was ordered to advance by this rout; it was imperative to try to open the main road through the Dongolaas gorge.

The commander of the leading brigade of the 4th Indian Division, Brigadier R. A. Savory, failed to find any way round to right or left of the position and decided to go for Sanchil and Brig's Peak, whose capture would have obvious advantages. From the afternoon of 3rd February the 11th Indian Infantry Brigade, with increasing artillery support, struggled hard for three days to secure a footing.2 They actually reached both peaks, but lost them to counter-attacks, and after stiff fighting were left in possession of Cameron Ridge and some features to the west of it. Meanwhile 'I' tanks had probed the gorge and confirmed that the road-block was impassable.


There is a sameness in the result of these early attacks, and of many later ones, which is understandable in the light of the conditions. The British artillery could bring down observed fire on the forward slopes of the hills, but targets beyond the crest were much more difficult to hit unless observation parties could establish themselves on the top. Aircraft were used for directing artillery fire, but they were not always available, the flying conditions were bad, and the technique was not highly developed. Moreover, if the targets were close behind the crest they were very difficult for guns, or even bomber aircraft, to hit. On the other hand, the whole of the British area was in full view from one point or another on the hills, so that to avoid observed artillery fire or to preserve secrecy it was necessary to move almost entirely by night.

The attack demanded great exertions of the infantryman, for the weather was hot and the hillsides waterless, and he had to clamber up rocks and over huge boulders, plod across shale, and often tear a way through thorn scrub and spear grass. Weapons, ammunition, and the minimum of other necessary equipment made, on a steep slope, an oppressive load. During the last hundred feet or so of rocky scramble the supporting artillery had to lift its fire, and an alert defender could man the crest from a covered position behind it and throw showers of small grenades among the exhausted climbers. When, in spite of all this, the attacking troops reached the crest, they were often too exhausted and depleted in numbers to withstand the counter-attacks which were usually made promptly and boldly, supported by the accurate fire of hidden mortars. Moreover, it was a laborious and costly task to supply the foremost troops with rations, ammunition, and water, for everything had to be carried up by men, so that perhaps a quarter of a battalion's numbers might be acting as porters at a time when every infantryman was badly needed. Perhaps worst of all was the heart-breaking labour of bringing the wounded off the hillside.

By 7th February, however, although the task was obviously hard, there was still no reason to doubt that the 4th Indian Division would be able to capture Keren, for on many previous occasions the Italians had started by resisting stoutly, only to give way when hard pressed. In any case, there was not enough transport to maintain both the 4th and 5th Indian Divisions in their present positions. The 5th, less its 29th Indian Infantry Brigade, was therefore withdrawn from the Barentu area to within easy reach of railhead at Kassala, and given the task of training intensively for mountain warfare.

The next phase of the battle was conducted by General BeresfordPeirse, who had now two Indian Infantry Brigades—the 5th and the 11th. He started with a night attack by the 5th (Brigadier W. L. Lloyd) against a new part of the front, Acqua Col. Desertions by


Italian colonial troops in this area had suggested that it might be a weak spot, and a rapid success here might lead to cutting off all the enemy west of the gorge. The difficulties were great, for no one knew the ground in detail, although it was clear that bad going, as well as the need for secrecy, would prevent the use of vehicles, and this meant a very long carry for the men.

The attack was made by the 4/6th Rajputana Rifles, and the other two battalions stood ready to push through along the track that led to Keren.3 The leading companies gained their objectives, but with such losses that they were too weak to defeat the counterattacks which followed. To the inevitable confusion of a battle in the dark in wild country was added a breakdown of signal communications. Next day the plan was given up, though covering positions on this eastern flank continued to be held.

General Beresford-Peirse decided, with General Platt's approval, to renew the attack on February 10th, this time on both fronts. The 11th Indian Infantry Brigade was first to capture Brig's Peak; then the 5th was to take Acqua Col, and if they were successful the 29th (Brigadier J. C. O. Marriott)—on loan from the 5th Indian Division—would pass through and exploit towards Keren. The outcome was that on the left Brig's Peak was twice taken and twice lost, while Saddle and Hog's Back were won and had to be given up. On the right, a gallant attempt to capture Acqua Col just failed, in the course of which Subedar Richpal Ram of the 4/6th Rajputana Rifles won a posthumous Victoria Cross. This battalion suffered 123 casualties from a strength already depleted by the previous attempt, while the other battalion engaged, the 4/11th Sikhs, lost over 100. General Beresford-Peirse considered that with this rate of losses he was not justified in renewing the attack; the enemy was fighting extremely well in a strong position, and seemed to be lavishly equipped with mortars and machine-guns. General Platt agreed with him and had further reconnaissances made to see whether the road to Arresa could be made into a main route. He came to the conclusion that its best use would be to support a small force which might divert a much larger enemy force from the Keren front. The main effort would still have to be made at Keren and both divisions would have to be used. The 4th Indian Division, in spite of steady daily losses, would have to maintain its present positions while the preparations for the large operation were being made.

The time required for these preparations depended mainly upon the number of vehicles that could be put on to the 180 mile journey forward of railhead, after the needs of troop-carrying and current daily maintenance had been met. The plan was to deliver into the


forward area enough petrol, food and stores for fourteen days, and a large quantity of ammunition. Railhead was pushed forward to Tessenei, to reduce road distance, but even so every possible vehicle was used to its utmost. In the forward area, too, there was much to be done; dumps of all kinds to be stocked, tracks to be made, signal communications to be laid, and the light railway put into running order, principally for removing the wounded; these tasks were all the harder for having to be done by night.

All this time the 7th Indian Infantry Brigade Group had been making its way south from Karora. Brigadier Briggs had now under his command one field battery diverted from Keren, one British and one Indian battalion, and the Free French Brigade d'Orient, formed of the 14th Battalion of the Foreign Legion from Egypt and the 3rd Battalion of the Chad Regiment from French Equatorial Africa. This force had occupied Mersa Taclai on the Red Sea, through which was established an improvised sea line of communication, and on 23rd February fought a successful engagement with the enemy at Cub Cub about 45 miles north-east of Keren.

With the capture of Agordat and Barentu, British fighter aircraft were able to reach the enemy's main air bases in Eritrea; attacks were at once made on Gura and Azoza airfields, and others were reconnoitred. During the first half of February efforts were mainly concentrated on supporting 4th Indian Division by offensive patrols, tactical and photographic reconnaissances, and bombing attacks on gun areas, dumps and communications—especially the railway between Asmara and Keren. Italian bomber activity was confined to the forward area, but fighters made several low-flying attacks on British airfields and at Agordat they destroyed or damaged thirteen aircraft.

From mid-February to mid-March, the period of preparation for the attack by both divisions, determined efforts were made to weaken the enemy in the Keren area and to cripple his Air Force. Eight Wellesleys of No. 223 Squadron one> night dropped 10,700 lb of bombs on the Caproni workshops at Mai Edega;4 two days later Wellesleys and Hurricanes attacked them again by day, after which all the main buildings had been extensively damaged. Hurricanes carried out several offensive patrols as far as Gura and Asmara, meeting little opposition. Asmara and its airfield were heavily attacked on 19th February, Massawa airfield on the 21st, and Makalle on the 23rd. The last was a particularly successful attack, made at the long range of 150 miles, and caused eight aircraft to be burnt out and others damaged. This sustained offensive practically drove the Italian Air Force out of Eritrea, and reduced it to almost complete inactivity. In addition, as H.M.S. Formidable (Nos. 826


and 829 Squadrons, Albacores) passed up the Red Sea5 her aircraft made several attacks on Massawa, and as a pendant to all these operations British aircraft dropped thousands of pamphlets in Amharic and Tigrean script, which, according to prisoners' statements, caused many Italian colonial troops to desert.

By 1st March General Platt had decided upon his outline plan. The attack was to be begun by 4th Indian Division to the left of the Dongolaas gorge, followed by 5th Indian Division on the right. The objective on the left was the line of peaks from Sanchil to Samanna; and for the 5th Indian Division Fort Dologorodoc was eventually chosen, with exploitation towards Falestoh and Zeban. General Heath elected to attack from west of the gorge, because only there could any covered assembly positions be found. An important point was that the bulk of the artillery could support first one division and then the other without changing position, and it was thought that the gently falling ground behind Fort Dologorodoc promised good fire effect to the artillery and aircraft. Particular attention was paid to providing good communications between the various Army headquarters and the Royal Air Force. Finally, in order to keep the enemy stretched as much as possible, the 7th Indian Infantry Brigade Group, now only fourteen miles north-east of Keren, was to make a further advance, while on the Arresa front a mobile detachment consisting of Skinner's Horse and units of the Sudan Defence Force was set up with orders to simulate a much larger force.

On 8th March the 5th Indian Division—now rejoined by its third Brigade, the 9th (Brigadier F. W. Messervy)—began to move into the forward area as unobtrusively as possible. Though the Italians must have expected a renewal of the attack they do not seem to have had any particular suspicions, for they did nothing to interfere; their chief concern at the time was with the harassing effect of the British air attacks. Their patrols showed some enterprise to the west of the main positions, and the newly arrived 51st (Middle East) Commando was sent up to counter their activities.

Early on the morning of 15th March a stream of Blenheims, Wellesleys and Hardys of Nos. 14, 47, 223 and 237 Squadrons began an intense bombardment of Italian gun areas and infantry positions, and at 7 a.m. the artillery of both divisions and 212th and 233rd Medium Batteries R.A. opened fire in support of 4th Indian Division's attack. This was made by 11th Indian Infantry Brigade reinforced by two battalions of the 5th Indian Infantry Brigade and 2/5th Mahratta Light Infantry. Sanchil and Brig's Peak were taken —but not for long; Hog's Back and Flat Top were taken and held; and on the extreme left the weakened 5th Indian Infantry Brigade


won a foothold on Mount Samanna and was then checked. Nevertheless General Platt and General Heath considered that a sufficient diversion had been caused, and at 10.30 a.m. 2nd Battalion The Highland Light Infantry, under command of the 9th Indian Infantry Brigade, attacked across the gorge towards Dologorodoc. They were soon stopped by fire, which was especially severe in enfilade from the lower slopes of Sanchil, and it was decided not to renew the attack until after dark. That night the 4th Indian Division made another attempt to take Brig's Peak and Sanchil without success, and Hog's Back and Flat Top were held against counter-attacks. On the right of the gorge, however, 3/5th Mahratta Light Infantry and 3/12th Frontier Force Regiment of 9th Indian Infantry Brigade secured the outlying features on the south-west of Mount Dologorodoc soon after midnight. Brigadier Messervy took the opportunity to send his reserve battalion, 2nd West Yorkshire Regiment, at the Fort from the south, and by 6.30 a.m. they had captured it with very slight loss. The enemy made several counter-attacks, which were all repulsed.

The next night, March 16th/17th, General Platt put in his last reserve—4/10th Baluch Regiment and 3/18th Royal Garhwal Rifles —to make yet another attempt on Sanchil and Brig's Peak; once again the Italian resistance was too strong. The same night, on 5th Indian Division's front, an attempt to exploit the success at Fort Dologorodoc was made by passing the 29th Indian Infantry Brigade (1st Worcestershire Regiment, 3/2nd Punjab Regiment and 6/13th Frontier Force Rifles) through to assault Falestoh and Zeban. The ground proved to be very difficult, the barrage was lost, and during the morning the leading troops were pinned down short of their objectives by intense fire and virtually isolated. The Royal Air Force came to the rescue at very short notice by dropping ammunition and rations packed in improvised containers. During the night these troops were withdrawn from their untenable position.

There was, however, one encouraging outcome. The attack had made it possible for the engineers to examine the road block, and they estimated that 48 hours' work would be required to clear it for tracked vehicles.6 The task, viewed from a distance, had seemed much more formidable, and the idea of attempting it had been discarded. Even now it could not be done until covering positions had been secured on each side of the gorge. General Platt ordered 5th Indian Division to seize these positions, clear the block, and then pass a force of tanks and Bren carriers through to Keren. The date was fixed as 25th March, to give time for the necessary preparations.

Between 18th and 22nd March the Italians attacked Fort Dologorodoc no less than seven times and suffered damaging losses.


General Lorenzini, who was respected by both sides as an inspiring leader, was killed on this front. General Frusci reported daily to the Viceroy that his troops were suffering heavily, especially from air attacks and artillery fire. By 20th March the strength of units had been reduced by a third, and thereafter Frusci's reports show increasing anxiety. General Platt's forecast was being amply borne out, for he had said: 'It is going to be a bloody battle . . . against both enemy and ground. It will be won by the side which lasts longest.'

Early on 25th March the attack up the gorge took place: 9th Indian Infantry Brigade on the right, 10th on the left. A good deal of artillery, mortar, and machine-gun fire was met but both brigades reached their objectives beyond the block and captured nearly 500 prisoners, many of them Bersaglieri. The Sapper and Miner Field Companies of 5th Indian Division, working in rotation and hampered by mortar and shell fire, cleared a passage through the roadblock by the afternoon of the 26th. A force of 'I' tanks and Bren carriers, known as Fletcher Force, was ordered to make its way through to Keren at dawn.7

During the afternoon there had been a marked lessening of Italian activity. The Italians had seen with anxiety their failure to prevent the clearing of the road, and on 26th March General Frusci reported that the situation was critical and that it would no longer be possible to stop the British tanks. That night the Italians gave up the struggle and skilfully withdrew the bulk of their troops and guns, leaving only a screen. Next morning, as the British tanks picked a way across the road-block, white flags appeared on Sanchil and soon the 4th Division's patrols were moving forward across the ground they had fought so hard to gain. By 7 a.m. the air reported that the enemy was clearing Keren, and Colonel Fletcher was ordered to pursue. The tanks entered Keren an hour later and moved on to the Habi Mantel gorge where they awaited the other vehicles which had been delayed by the difficulties of passing the Dongolaas gorge.

At Keren the British had 536 killed and 3,229 wounded; the Italians, according to General Frusci, had 3,000 killed. The battle was not the last of the campaign, but it was decisive, for although the Italians managed to withdraw a large amount of their artillery, and some of their infantry, they never fought with the same determination again. They had seen that General Platt could not manoeuvre


them out of Keren, and, although his diversionary action was as successful as could be expected, they can never have doubted that the vital point was at Keren. They knew that they must hold the hills astride the Dongolaas gorge and retake any that they lost. This simple plan succeeded until their casualties became more than they could bear.

It is not possible to say when General Frusci felt that the battle was lost, but it is worthy of note that he seems to have judged that the last two battalions of the Savoia Grenadiers arrived from the general reserve too late to influence the action. They were used in rearward positions where they achieved nothing. It is remarkable, too, that the Italians never tried to drive the British from Cameron Ridge or to interfere with their preparations for the set battle. Even a limited counter-attack or one or two raids might have caused a lot of delay, and General Platt expected that they would attempt something of the sort.

An active hostile Air Force would have immeasurably increased the hazards and difficulties facing the British. The Army had therefore good reason to be grateful to the Royal Air Force, which by its bold and sustained operations drove the enemy's air bases farther and farther back and caused heavy losses in aircraft, results to which the squadrons from Aden and the South African squadrons farther south contributed their share. The consequence was that Italian aircraft interfered very little during the vulnerable stages of the build-up and deployment, and played a negligible part in the critical phase of the battle. By 22nd March the Italians had only 37 serviceable aircraft in the whole of East. Africa, which is the measure of the Air Force's success.

To General Wavell belongs the credit for allotting to the Eritrean front the two divisions most likely to adapt themselves quickly to the conditions and the best able to work in double harness. Many of the commanders and troops had had experience of mountain warfare, but not against an enemy who possessed aircraft and artillery and numerous mortars, machine-guns, and grenades. This formidable weight of fire made progress very slow, but the enemy, though fighting with determination, was suffering heavily also. And as in every hard fought battle there were many instances of great gallantry and tenacity on both sides.

The British commanders, having seized the initiative, never lost it and stuck doggedly to their task, supported by the resolution, endurance, and courage of the regimental officers and men. They were devotedly served by the whole administrative machine: food, water, and ammunition were sufficient—nearly 200,000 rounds were delivered to the batteries during March—and the arduous task of evacuating the wounded and sick was well done. The achievements


of the 4th and 5th Indian Divisions, who formed the bulk of the British force, drew a generous tribute from the Prime Minister.

As soon as Keren was captured the Air Force began to harass the Italian columns retreating to Asmara and towards Gondar. The winding road from Keren to Asmara was well suited to delaying tactics, and the enemy had made ready to oppose the British advance near Ad Teclesan. He realized that this was the last ditch before Asmara, and that when Asmara fell Eritrea was lost.

On 31st March, after a day-long fight, the Teclesan position, fell to the 9th Indian Infantry Brigade, with 460 prisoners and 67 guns. Next day the Italian authorities in Asmara surrendered the town.

Our naval and air forces were at last rewarded for the long watch on Massawa. On the evening of 31st March three of the remaining six large Italian destroyers had put to sea. Their plan, it is believed, was to raid the Gulf of Suez and then scuttle themselves. The Leone grounded shortly after sailing; next morning she sank and the operation was cancelled. Aircraft had reported the stranded Leone and had seen another destroyer returning to Massawa; the third reached port unseen. On 2nd April all five remaining destroyers sailed in company to attack Port Sudan and then sink themselves. Their departure was reported by aircraft from Aden and at dawn on the 3rd four of them were seen some twenty miles east of Port Sudan by Swordfish of Nos. 813 and 824 Squadrons, which had flown from the Eagle at Alexandria to Port Sudan a few days before. These Swordfish, reinforced by five Blenheims of No. 14 Squadron, sank the Sauro and Manin. The hunt went on, and the Pantera and Tigre were next found close inshore south of Jeddah, where they were being abandoned. Here Wellesleys of No. 223 Squadron from Port Sudan and H.M.S. Kingston destroyed them. The missing destroyer Battisti had developed a defect during the previous night and had been scuttled. All six destroyers were thus accounted for. The Italians, however, had one success before Massawa fell. The cruiser Capetown was torpedoed by one of the remaining Italian motor torpedo boats, and after being towed to Port Sudan she had to go to Bombay for permanent repairs.

The 4th Indian Division was already under orders to go back to Egypt, leaving the 7th Indian Infantry Brigade Group under the command of General Heath. On General Platt's orders he sent light mobile forces southward to Adowa and Adigrat, and turned to his main task of capturing Massawa. His first action was to send an ultimatum to Admiral Bonetti, commanding at Massawa, by the unusual means of ringing up his headquarters on the undamaged


telephone line. If the ships in Massawa were scuttled, the British would take no responsibility for feeding or evacuating the Italian population in Eritrea or Ethiopia, to which the reply was that the Duke of Aosta had ordered resistance and no guarantee could be given regarding the ships.

The 7th Indian Infantry Brigade Group soon reached the northern part of the Massawa defences, and by 5th April the ring was closed by 10th Brigade and the Brigade d' Orient. Admiral Bonetti was called upon to surrender, but after consulting Rome he refused. Early on 8th April an advance by 7th Indian Infantry Brigade Group was resolutely opposed, but a simultaneous attack by 10th Indian Infantry Brigade and B Squadron 4th Royal Tank Regiment bit deeply into the western defences and the Free French troops broke into the south-western sector. Bombers of Nos. 47 and 223 Squadrons co-operated throughout, principally by attacking the Italian guns, which were very active. Early in the afternoon Admiral Bonetti surrendered with 9,590 men and 127 guns. Much damage had been done to the harbour, all the ships had been scuttled, and a great deal of equipment and supplies had been dumped in the sea. Nevertheless a sea supply line was quickly organized; by 27th April the Massawa-Asmara railway was cleared for traffic; and by 1st May the main line of communication for 5th Indian Division was running through Massawa.

All organized opposition in Eritrea was now over; all the Italian warships had been accounted for, and the handful of aircraft remaining in Ethiopia presented no danger. The strategic object of the East African campaign, which was to remove the threat to shipping through the Red Sea, had thus been attained. It was not long before this victory had a most valuable sequel, for on 11th April President Roosevelt declared the Red Sea and Gulf of Aden to be combat zones no longer. This meant that ships of the United States could now enter these waters and lighten the task of the overtaxed British shipping.

A few days after the loss of Keren the Viceroy saw that there was no longer any hope of retaining a hold on Eritrea, and decided upon a centre of resistance at Amba Alagi instead. On the 30th he telegraphed to Mussolini that the very rapid crumbling, forecast a long time ago, had set in. 'It only remains for us to resist wherever we can, and for as long as we can for the honour of the flag.' The Italian forces in Ethiopia were now concentrating in three centres; one in Galla-Sidamo, under General Gazzera, one at Gondar under General Nasi, and one on the Asmara-Dessie road at Amba Alagi, to which the Duke himself went on 3rd April, and where he united


the remainder of his own reserve troops with the remnants of General Frusci's Eritrean Army.

General Platt had only enough troops to fight one serious engagement at a time, and in order to remove all threat to the Sudan frontier he proposed to deal first with General Nasi. He began, therefore, to consider how to tackle the formidable Wolchefit Pass on the road to Gondar, but at this moment came General Wavell's order to clear the Dessie-Asmara road as quickly as possible, so that it could be used for the transfer of forces northward. General Platt accordingly gave the task of watching Wolchefit to a Motor Machine-Gun Group of the Sudan Defence Force and a host of Patriots under Major B. J. Ringrose, and ordered 5th Indian Division to destroy the enemy force at Amba Alagi.

The commander of the 5th Indian Division was now Major-General A. G. O. M. Mayne, who had replaced General Heath on his promotion to command a Corps in Malaya. In spite of the prompt arrival of a nucleus of the British Military Administration,8 there were many duties in Asmara and Massawa for which troops were still required, and the force available for further operations against the Italians was no more than the mobile Fletcher Force, the 29th Indian Infantry Brigade, two battalions (one from 9th Indian Infantry Brigade, one from 10th), one Commando, two field regiments and one medium battery. In the air there was No. 237 (Rhodesian) Squadron (Gladiator, Lysander and Hardy), supplemented by a detachment of Wellesleys of No. 47 Squadron. A large force of Ras Seyoum's Patriots, led by Lieutenant-Colonel W Rankin of the Sudan Defence Force, was sent to work round behind Amba Alagi, and by 27th April had captured Socota.

Approaching from the south was the 1st South African Brigade Group, which had left Addis Ababa on 13th April in consequence of General Wavell's order to General Cunningham to help to secure the use of the road between Addis Ababa and Asmara. After making its way past extensive demolitions the Brigade came under shell fire from a prepared position at Combolcha, a few miles south of Dessie. Here a strong force, with ample artillery and good observation, was holding a line of hills to the east of the road and roughly at right angles to it. To the west of the road the ground was exposed and marshy. Running towards the enemy's position was another line of hills, roughly parallel to the road. Brigadier Pienaar decided to establish himself on these hills and then make a plan for attacking the main position.

This entailed a methodical advance by the infantry, while the twelve field guns of 4th Field Brigade and one section of 1st Medium


Brigade S.A.A. proceeded to gain the mastery over the far more numerous Italian guns. Progress by the infantry was slow, for everything had to be manhandled over steep and rough going, but by the afternoon of 19th April the 1st Duke of Edinburgh's Own Rifles completed their long advance by occupying a hill which overlooked the main Italian position. The enemy at once tried to retake this hill, but failed with considerable loss, and many deserters began to come in. Early on the 22nd Brigadier Pienaar made his attack. He had been joined by a large band of Patriots under Lieut. A. G. S. Campbell, of The Black Watch, which proved most useful in locating and harrying the enemy's flanks. The 1st Royal Natal Carbineers moved across country and attacked the enemy's left flank, which they proceeded to roll up, and the 1st Transvaal Scottish captured the remaining hills. Much was due to the efficiency of the artillery support. By the early afternoon the enemy was in full retreat.

The results of this successful action were remarkable. Dessie fell after little further opposition. In all over 8,000 prisoners, of which a large proportion were Italians of the Africa Division, were captured together with 52 guns and large quantities of other weapons, lorries, stores, and petrol. Nineteen burnt-out aircraft were found on Combolcha airfield. The South African casualties were nine killed and thirty wounded, a tribute to Brigadier Pienaar's careful plan and a proof of the saying that hard work saves casualties. The physical exertion of carrying heavy loads up steep hills at an altitude of well over 6,000 feet was very great, especially for men who had been lorry-borne for so long. The further advance towards Amba Alagi was delayed for some days by elaborate road blocks, but Campbell's scouts were able to push ahead, and at Quoram joined forces with Rankin's Patriots.

Meanwhile General Cunningham was finding that his operations were being continually hampered by requests from Italian civilians for protection against marauders. He therefore invited the Duke of Aosta to receive another communication about the safety of Italian women and children. The Duke agreed, and General Cunningham explained to an envoy that only if the Italian forces unconditionally laid down their arms could he be responsible for the protection, feeding, and evacuation of Italian nationals. The Duke considered this proposal to be an attempt to extort surrender by threats, and on 20th April rejected it.

The stronghold in which the Duke had installed his headquarters was situated halfway between Asmara and Dessie, where the road climbs by a series of hairpin bends to cross a steep mountain barrier at the Toselli Pass, nearly 10,000 feet high.9 To right and left of the road stretched a tangle of precipitous rocky heights, dominated by


The Battle of Amba Alagi
Map 30. The Battle of Amba Alagi

Amba Alagi—11,186 feet—less than a mile from the Pass itself. It was here that General Frusci had assembled a force estimated at four or five thousand men, with perhaps thirty guns. (This was in fact an under-estimate.) The neighbouring peaks had been fortified, a quantity of barbed wire put up, and cave shelters and gun pits constructed; water was sufficient but not abundant, and there was food for three months.

While his troops were moving forward, General Mayne made his plan. He thought the Italian position to be very wide for the number of its defenders, and by stretching them hoped to create a weak spot. There were two possible lines of advance apart from the main road. On the east a track led to the rugged Falaga Pass, seven miles from the Toselli Pass; between them were the features Gumsa, Twin Pyramids and Triangle—all fortified. On the west a spine of defended hills (Pyramid, Whale Back, Elephant, Middle) led to Little Alagi and Amba Alagi itself. General Mayne decided to demonstrate against the centre and east and to attack from the west, a decision which showed great confidence in his troops, for it committed them to a long approach over almost pathless hills. He considered that the enemy would be unlikely to expect a force of any size to attack from this direction.

Fletcher Force, which had been probing towards Falaga Pass, made such good progress, and was receiving so many deserters from Italian colonial troops, that on May 2nd General Mayne ordered Colonel Fletcher to increase the pressure. An attempt was accordingly made to capture the Pass, but without success. The same day, 3rd May, 3/18th Royal Garhwal Rifles made the demonstration against the centre, and early on 4th May 6/13th Frontier Force Rifles and 3/2nd Punjab Regiment of 29th Indian Infantry Brigade, supported by timed artillery concentrations, took Pyramid, Whale Back and Elephant. A further advance over the bare ridge to Middle Hill was deferred until dawn next day. This hill was duly taken under a barrage, but daylight disclosed strong wire obstacles across the narrow neck which led onward, and 1st Worcestershire Regiment were pinned down by machine-gun fire.

There seemed little chance of further progress from Middle Hill, but as Fletcher Force had discovered that the enemy's eastern flank did not extend far beyond the Falaga Pass, General Mayne decided to transfer his main effort to that side. He quickly organized a brigade group, gave Colonel Fletcher the headquarters of 9th Indian Infantry Brigade, and ordered him to work round the eastern flank on May 7th with a view to attacking Falaga Pass.

To add to the enemy's uncertainties, 6/13th Frontier Force Rifles (29th Brigade), after a long advance by night in the rain, made a surprise attack, and occupied two outlying hills (Khaki and Centre)


to the south of the Toselli Pass. The same day, May 8th, General Cunningham placed the 1st South African Infantry Brigade Group (less one battalion) at General Platt's disposal, retaining responsibility only for its supply. By May 10th the leading South African troops were only twenty miles away, with Rankin's Patriots and Campbell's scouts close by. These irregulars proved to be a rather mixed blessing, because they were joined by many who were not amenable to control and whose main object was loot—British or Italian. By this time the 3/12th Frontier Force Regiment (9th Brigade) had taken the Falaga Pass and had reached Gumsa, capturing a large number of prisoners and obtaining excellent observation for the artillery. The net was tightening, but the way from Gumsa to Amba Alagi was seen to lie along a saw-tooth ridge, of which each tooth was defended—a very uninviting prospect. Rain added its difficulties, particularly to the exacting business of supplying the troops on the heights. The nights were bitterly cold.

General Mayne now decided to attack the features Twin Pyramids and Triangle from the south, and on 11th May he flew to arrange with Brigadier Pienaar for the operation to begin on the 13th. While the preparations were going forward the Patriots made a private and independent attack on Twin Pyramids, and most gallantly and astonishingly captured them. Encouraged by this, they tried, again independently, to storm Triangle, but this time were bloodily repulsed. Nevertheless, the attempt assisted the South Africans, whose attack, methodically pressed against stiff opposition throughout 14th May, together with pressure by the 9th Indian Infantry Brigade, induced the Italians to abandon Triangle that night. It was not necessary to attack Fort Toselli, for the enemy's endurance was exhausted.

Negotiations for an armistice began on the 16th and were conducted at Amba Alagi by Colonel D. Russell of the divisional staff, after the Italian envoy, General Volpini, had been murdered by marauders while on his way to General Mayne's headquarters. The Italians were anxious to escape humiliation and asked that the Viceroy and his troops should be allowed to remain at Amba Alagi, under arms but non-belligerent. This novel form of parole was refused, and a little later the Italians had to accept surrender with the honours of war. On 19th May the garrison, about 5,000 strong, marched past a Guard of Honour and was then disarmed, and next day the Duke with his personal staff made his formal surrender to General Mayne before being conducted to General Platt.

Amba Alagi differed from Keren in that there was far more scope for manoeuvre. The features of the battle were the flexibility of the British plan and the cross-country mobility of the troops over very difficult ground and in appalling weather. After the ring was closed


the only question was how long it would take to overcome the defence, but as the perimeter shrank and the bombardment by air and artillery became more concentrated the morale of the garrison seems to have sunk rapidly. The Viceroy's anxiety about the recovery and care of the wounded (which he reported to Rome on 13th May) was no doubt shared by his men, who were alarmed also by the activities of the irregulars. Indeed, many Italians surrendered in order to avoid falling into their hands.

Of the 325 Italian aircraft in East Africa when the war began, and the 75 which were flown in at various times, there now remained a negligible number in a serviceable state. The Italian records show 250 as lost on all the East African fronts by the middle of March; the corresponding British losses were 138. Of the 350,000 or so men with which the Italians entered the war there now remained the garrisons of the defended posts in and around Gondar, swollen by fugitives from the north, and seven colonial divisions in GallaSidamo which included the remains of General De Simone's force. These would have to be rounded up, but everything else had been destroyed, captured, or dispersed.

The main reasons for this catastrophe are easily seen. First and foremost is the fact that in June 1940 Italy was not ready for war, but was thrown into the conflict as a matter of political expediency. It was assumed that the Axis had practically won the war already and were certain to do so before very long. There was no strategic plan for the sort of war in which Mussolini, to his surprise, soon found himself engaged. The results of this miscalculation upon the campaign in the Western Desert have already been seen; in East Africa they were much more serious, because the territory was cut off from Italy by sea and land and became a wasting asset, since steps had not been taken in time to build up enough stocks.

The Duce's initial instructions to the Viceroy were perfectly consistent with the basic idea of a short war. There was no need to attack French Somaliland, for France was on the verge of collapse. It was good defensive strategy to deprive the British of the main gateways into the Empire and it was tactically easy to do so. And when, after a few exasperating weeks, French Somaliland passed safely into Vichy hands, the opportunity of seizing the weakly held British Somaliland was too attractive to miss, even though no strategic advantages were to be gained. The truth is that with the exception of French Somaliland there were no worth-while objectives within easy reach of the Italian land forces, and for a venture farther afield there was not the necessary administrative backing,


nor, with the possible exception of the Savoia Division, were the higher formations sufficiently well trained.

For these reasons it was no doubt sound to adopt a defensive strategy on land, although this must have been a disappointing decision because it was known that the British had very few forces in Kenya and the Sudan, and the time to strike at them was before they became stronger. This strategy would have been still sounder had it included the intention of making determined air attacks upon British shipping in the Red Sea. But virtually nothing was done to interfere with the arrival of British reinforcements in convoys, at the ports of Mombasa and Port Sudan, or on the railways leading from them. Moreover it was not long before dwindling stocks of spares, fuel, and reserve aircraft were acting as a brake on all enterprise, which was further curbed by the inability of the ground organization to maintain the number of aircraft in the country. A quarter of these indeed were in workshops when the war began. Lack of enterprise in the air was matched at sea, and the destroyers and submarines based at Massawa were inactive. Thus the strategic asset of Italian East Africa's geographical position on the flank of Britain's vital sea-route was thrown away.

In these circumstances the role of the Italian forces in East Africa could only be to hold out against the British for as long as possible. To do this it was of course necessary to keep internal unrest in check, which was a commitment that grew with every British success. The Duke of Aosta was certainly in an unenviable position, not made any better by the knowledge that whatever he did the war would be decided somewhere else. He maintained a realistic outlook, kept his head, and as was to be expected set a good example, but the Italian failures in Greece and the Western Desert made it all the harder for him to keep up the morale of the Italian nationals, and through them that of the loyal population and the native troops. By early February 1941 the local situation had become much worse, for the simultaneous British advances into Eritrea, Somalia and Gojjam had shown up the weaknesses of the Italian colonial forces, and the danger of internal disorder had greatly increased. There were not enough Italian national troops to stiffen all the fronts, and it cannot halve been easy to decide where best to use them. For instance, while one of the regiments of Savoia Grenadiers took part in the whole of the decisive battle of Keren, the other was used neither there nor in the difficult country in front of Harar, but was held back at Addis Ababa until it was too late to use it effectively anywhere.

But, although the Italian forces were not equal to the dual task of resisting the British and maintaining internal order, the military commanders should have appreciated that the British too had their difficulties and that these could be turned to good account to impose


delay. For example, the Italians might have known that the British advance on Kismayu would involve administrative problems which the capture of an undamaged port would greatly ease. Yet the port facilities at Kismayu were not destroyed, although there was plenty of time and the High Command had ordered it to be done. A huge stock of petrol was left at Mogadishu as a gift, and another British problem was solved. It was well known that a landing from the sea could be a very hazardous undertaking, yet the British were allowed to land and take Berbera almost unopposed. The impression of British speed and invincibility led to a general loss of heart, which was manifested by wholesale desertions by the colonial troops.

Even when allowance is made for the inferiority of the Italian equipment, and for the depressing effect of being always on the defensive, there remains the impression that on many occasions positions were abandoned before they had caused anything like the maximum of delay, and that a more aggressive defence would have achieved far better results. In Gojjam the Italians were actually the better equipped—for example, Gideon Force had no artillery at all —but here also there was a failure to grasp the military essentials; the colonial brigade at the south of Lake Tana should never have allowed itself to be contained by a small detachment, leaving Wingate free to attack Debra Markos against big odds and open the way to Addis Ababa from the north. Weakness of purpose of this sort more than offset some very gallant acts of junior leadership.

The British succeeded largely because they exploited the enemy's weaknesses. Until these were disclosed, the British commanders, though they saw clearly what had to be done, were inclined to doubt whether they had the resources or the time to do it. Once they began, they sought and took their opportunities, and showed firm purpose coupled with sound judgment in matters of time, space, and resources. Above all they were ready to take risks.

The campaign involved men of many races in a wide variety of operations, in the course of which they had to advance vast distances over all sorts of country, in extremes of climate, with seldom even the barest of amenities. Not only must the administration have been excellent, but the training which enabled plans to he quickly translated into action had obviously been carried out on sound lines.

The British forces in East Africa, their main task over, were drastically reduced. The 5th South African Infantry Brigade had already sailed from Mombasa for Egypt; the 2nd was awaiting shipping at Berbera; the 1st was on its way to Massawa— also for shipment to Egypt. The 4th Indian Division had gone; the 5th was under orders to follow. Withdrawals of air forces began directly after the capture of Massawa, and during April three fighter, three


bomber, and one general reconnaissance squadron left for Egypt. These were Nos. 94 and 203 Squadrons R.A.F. from Aden, No.1 Squadron S.A.A.F., and Nos. 14 and 223 Squadrons R.A.F. from the Sudan, and Nos. 2 and 24 Squadrons S.A.A.F. from Kenya. No. 237 (Rhodesian) Squadron and No. 12 (Bomber) Squadron S.A.A.F. followed at the end of May. For the present, Ethiopia was under British military guidance and control, while British Somaliland, Italian Somaliland, and Eritrea were under British Military Government. The forces remaining behind to deal with the Italian armies were the 11th and 12th African Divisions. To support them were: in the Sudan, No. 47 (Bomber) Squadron; at Aden, No. 8 (Bomber) Squadron; and in Kenya No. 3 (Fighter), No. 11 (Bomber) and Nos. 40 and 41 (Army Co-operation) Squadrons of the South African Air Force.

With the collapse in East Africa, Italy suffered the third of the major defeats which followed her disastrous decision to enter a war for which she was not ready and for which there was by no means universal enthusiasm. By early in February 1941, Cyrenaica had been lost, and with it a vast number of men and a mass of material. By the middle of March the last unaided attempts to defeat the Greeks had failed. On both these fronts the Germans were preparing to intervene, and the British were soon to learn—if they did not know it already—that the means and methods that had done so well against the Italians were not good enough against the Germans.



1 See page 401, Map 29 and Photo 39.

2 11th Indian Infantry Brigade consisted of 2nd Cameron Highlanders, 1/6th Rajputana Rifles, 3/14th Punjab Regiment, with 3/1st Punjab Regiment attached.

3 1st Royal Fusiliers and 4/11th Sikhs.

4 See Photo 38.

5 On her way to replace H.M.S. Illustrious.

6 See Photo 41.

7 One squadron less one troop of 4th Royal Tank Regiment, and about 50 Bren carriers of units of 4th and 5th Indian Divisions, all under the command of Lieutenant-Colonel B. C. Fletcher, Highland Light Infantry, who had been in temporary command of 10th Indian Infantry Brigade until March 21st, when Brigadier T. W. Rees took over.

8 For a full account see Lord Rennell of Rodd: British Military Administration in Africa 1941 to 1947.

9 See Map 30 and Photos 42 and 43.


Previous Chapter (XXII) * Chronology

Transcribed and formatted for HTML by David Shepherd for the HyperWar Foundation