by Hew Strachan
The Germans fell back west of Taveta to the hills of Reata and Latema. This position was well prepared but it was 19 kilometres long, too extensive for the troops available, and Lettow kept his main concentration to the rear at Himo. Smuts was determined to follow up as fast as possible, but owing to the need to consolidate his rear, had only three battalions available to attack. The main hills rose to 330 metres, and an attempt to seize the nek between the two during the course of 11 March was unsuccessful and costly. The bulk of the artillery was still being brought up, and the efforts of forward observation officers to direct its fire were hampered by the bush. Tighe, commanding the 2nd division, decided to use the cover of night to attack with the bayonet. Lettow, meanwhile, concluded that the major threat lay to his left, in the north, and at 5.30 p.m. sent two companies thither. But at 7.50 p.m. he received reports that Kraut’s positions on the Latema side of the nek had been broken. At 9.30 p.m. he lost telephone contact with Kraut, and by 10.30 Kraut was reported as in full retreat. Lettow therefore ordered the whole position to be abandoned, and his forces to take up a new line backing onto the River Ruvu, facing north, with their left on Kahe.
In reality, although Kraut had ordered a retreat at 10 p.m., one company had not received the order and Latema could have been held. Moreover, there was comparable confusion in the British command. Smuts had never intended the frontal attack to precede Deventer’s envelopment to the north. Nor did he appreciate that some of his troops had reached the summits of both Reata and Latema in the course of the night. On the morning of 12 March he ordered the 2nd division to fall back, while Deventer’s outflanking move to the north— Lettow’s original fear—took effect. Deventer pushed on towards Moshi to link up with the 1st division on 14 March.
Smuts had opened the door into the northern part of German East Africa. But he had not inflicted a major defeat on the Schütztruppen. The blame was laid on Stewart, commanding the 1st division. The advance west of Kilimanjaro, although no longer the major thrust, began from Longido on 3 March, so giving Stewart three days to get across the German line of retreat before Deventer’s and Tighe’s attacks took effect. Stewart’s progress was slow, manifesting an undue concern about problems of supply. But the real difficulty was that Smuts’s plan did not make clear in which direction he thought the Germans most likely to withdraw. Stewart’s advance would have its greatest and most immediate effect if Lettow planned to fall back to the west, from Moshi to Arusha. Moreover, Deventer’s move to the north of the Taveta Hills rested on a similar assumption. As Lettow planned to fall back down the northern railway, an advance on Moshi by Stewart and Deventer could only shoulder the Germans in the direction which they already planned to follow. Stewart would have had to reach Kahe by 12 March to have fulfilled Smuts’s hopes of true envelopment. It must, therefore, be presumed that Smuts, given his predisposition to see Lettow as a guerrilla, imagined that the German commander would seek the interior, would show the same disinclination to use railways and harbours as Smuts did himself, and would be confirmed in that tendency by the demonstrations around Ngulu. On this basis Lettow could only be expected to fight if not threatened with envelopment; wide turning movements by mounted troops would only keep the Germans moving, and were therefore a way of avoiding battle, not of seeking it.217
Lettow’s new position was a strong one which sustained the threat to Taveta and to the railway line from Voi, but which also gave further opportunities for a British tactical success. His right flank rested on the River Lumi and Lake Jipe, his left on the Pangani, as the Ruvu became after Kahe. Any British frontal attack would be channelled by the crocodile-infested rivers flowing north-south into the Ruvu. But Lettow had both the Ruvu and the Pare Mountains to his back. His principal line of withdrawal, the railway, lay behind his left, while his own inclinations were to concentrate for a counter-attack on his right.
Smuts’s plan was to attack frontally with the 1st division while sending Deventer’s mounted brigade from Moshi, west of the railway, to Kahe in order to cut off the German retreat. The difficulties confronting the 1st division in its attack caused it to move against Kahe more than against the centre. On 21 March Deventer was unable to find crossings on the Pangani. Eventually part of his command swam the river and took Kahe Hill. Deventer then pushed back north on Kahe, while sending two dismounted squadrons south to cut the railway below Kahe. The Germans had already abandoned Kahe, and were in positions south of the Pangani strong enough to check Deventer’s relatively weak command from working round their southern flank to the railway. Sheppard, Stewart’s successor in command of 1st division, did not know that Deventer had control of Kahe, and at 4.45 p.m. ordered his men to dig in 3 or 4 kilometres to the north. Thus the South Africans failed to push in the attack at the vital moment. To the south, Deventer’s two detached squadrons were blocked by impenetrable and seemingly endless bush. However, their efforts did not go unobserved. They were reported to Lettow as threatening Kisangire. A determined German counter-attack might have regained Kahe and Kahe Hill. But the danger to Kisangire decided Lettow that he should withdraw there himself, pulling his troops south of the Ruvu during the night of 21/2 March. Thus the British gained command of the Ruvu, and so secured their communications from Voi to Moshi. But once again they had failed to trap the German troops.
The March-May rainy season came late in 1916. Smuts was therefore lucky to have got as far as he had. But when the rains did arrive they were the heaviest for some years. All operations on the northern railway were suspended. Else-where the effects were less severe. In the west the Belgians advanced into Ruanda and Urundi; in the south-west a force under Brigadier-General Northey began its push from Northern Rhodesia; and to the south Portugal declared war on Germany in March. Tombeur’s plan of a year previously, that of a number of converging but independent thrusts along the circumference of the German colony, thus found practical application. Co-ordination was admittedly non-existent. Smuts had no direct line of communications to Tombeur; Northey was answerable directly to the Colonial Office, not even to the War Office; no joint commander was appointed nor planning conference held. But the momentum of the allied onset was not lost.
Smuts’s transport and supply services had banked on the rainy season for a moment’s pause and consolidation. However, Smuts was unhappy at the prospect of inactivity on the northern front until June. He assumed that the Germans did not intend to fall south of the central railway. Thus neither Northey’s column nor any Portuguese effort promised immediate effects. The pivot of Lettow’s resistance seemed to be Tabora. Schnee had transferred his capital there, given the danger to Dar es Salaam; it was the Schütztruppen’s major recruiting centre; and its inland position on the central railway played to German strengths and British weaknesses in matters of supply and communication. Intelligence gathered by Deventer, when he took Lol Kissale on 6 April, confirmed this analysis of German intentions; it pointed to plans for the defence of Ufiome and Kondoa Irangi, in order to bar the western route to the central railway.218 Smuts was committed to accepting Belgian co-operation in the west, but on the assumption that Tombeur’s operations would be secondary, not primary. However, if Smuts did nothing during the rains, the Belgians not the British would spearhead the advance on Tabora. Britain’s credit in Africa would not be fully restored; South Africa’s war aims might be forfeit to Belgium’s.
Smuts briefly reconsidered the amphibious option favoured by Smith-Dorrien. Given his assumptions about the significance of Tabora, Tanga seemed remote and unimportant. Dar es Salaam, by virtue of its position at the head of the central railway, gave directly onto Tabora, but heavy seas and the unhealthy coastline helped confirm Smuts in his predisposition against such a scheme. Communication difficulties precluded the formation of a major British concentration at Mwanza for a direct drive on Tabora. Instead, Smuts opted to march on Kondoa Irangi. Although supported in his decision by Meinertzhagen, the conception and execution were essentially Boer. Afrikaner settlers in the north as
sured Smuts that the rains to the south and west would not be as severe as those around Kilimanjaro. His forces were reorganized into three divisions, two of them exclusively South African and each composed of an infantry brigade and a mounted brigade: mobility was emphasized over firepower. Smuts’s aim was less the defeat of Lettow than the occupation of territory. The lessons of South-West Africa were being applied in circumstances that were totally different.
Success in their first campaign had made the South Africans heedless of the needs of horse management. The remounts they had employed had been shattered by the punishing marches imposed upon them. But the campaign had not collapsed; railways and motor-transport had kept some form of supply going, and disease in itself was not a major threat. In East Africa mechanical transport was not so readily available, and great areas, embracing a zone up to 400 kilometres inland or until the altitude reached 1,000 metres, were home to the tsetse fly. Ultimately all horses in East Africa would succumb to the fly. The task of the veterinary services was to keep it at bay for long enough to enable the advance to take effect. This could be achieved in two ways, prophylactically or operationally. The first was undermined by the inability of regimental veterinary officers to establish their authority over the decentralized commands of Boer units, whose squadron and troop leaders had powers of discretion. Arsenic powder could prolong the horse’s life, but it was issued in pills and soldiers gave it to their mounts whole, with the result that it was left unconsumed at the bottom of the feed-bag. The second was the casualty of Smuts’s staff arrangements. The veterinary staff was ‘always miles away’ from the commander-in-chief and his chief of staff. The British knew— as a result of maps obligingly provided by German veterinarians before the war—the locations of the worst tsetse areas, but this intelligence was not incorporated in the campaign plan. Equine wastage in 1916 ran at 100 per cent per month, when the veterinary services reckoned it could have been half that.219
The newly constituted 2nd division, commanded by Deventer, left Lol Kissale on 8 April. Its mounted brigade, in the van, reached Kondoa Irangi on 19 April, and the infantry brigade arrived by the end of the month. The weather contradicted the predictions: nearly 18 inches of rain fell at Moshi in the last fortnight of April. The supplies dumped at Longido and destined for Deventer’s men became bogged on muddy roads. Tsetse fly annihilated the mounted brigade’s mobility: of 3,894 horses issued to the brigade by 23 May, 1,639 had died since March and 718 were unfit for service.220 Fortunately Kondoa Irangi itself was a fertile area.
However, the weather was not partial in its effects. The ‘short’ rains, those that fell in November, had proved light in 1915, and accordingly the harvest yields in the Kilimanjaro area had been low. Throughout February and March the German troops in the north were threatened with starvation. Food was brought from the depots on the central railway. To bridge the gap between Kilosa and the light railway at Handeni, an extra 20,000 porters were required. Then came the heavy rains of late March. Although the nine-day hike between the railways was organized in relays, both shelter and provisions proved inadequate. About 20,000 porters died, and many others fell prey to dysentery and respiratory illnesses.221 As the intelligence captured at Lol Kissale revealed, Lettow was already planning to regroup around Kondoa Irangi in order to ease his supply problem before Deventer began his main advance.
Smuts’s strategy thus conformed to Lettow’s intentions. It allowed him to use his railway communications and the availability of interior lines to maximum advantage. Leaving Kraut and ten companies (2,400 men) in the Pare Mountains, Lettow brought the strength facing Deventer’s 3,000 rifles to eighteen companies (4,000 rifles) and six guns (two of them heavy guns from the Konigsberg). Deventer began to look dangerously vulnerable. For the first time in East Africa the Germans countered artillery with artillery.222 But the ground before Kondoa Irangi was open and exposed. Furthermore, German reconnaissance was inadequate, and failed to spot British positions south of the town. Lettow’s decision to attack on the night of 9/10 May, reflecting the lack of cover for a daylight advance, did not take account of his shortage of intelligence. The Germans were checked with 35 per-cent losses. Thereafter, Lettow remained unwontedly passive. Deventer’s position was entrenched, and in early June was supplemented with heavy artillery.
In concentrating south of Kondoa Irangi, Lettow decided to abandon the northern railway. If Smuts wished to end the campaign quickly, he should have used the advent of dry weather to reinforce Deventer and then risked heavy casualties in a battle for the Dodoma-Kilimatinde section of the central railway. Such counsels were not lacking at his headquarters.223 But Smuts now insisted that the northern railway was vital and that the Pare Mountains should be cleared. Given his previous assumptions about the importance of Tabora and the irrelevance of Tanga, this was perverse; it confirms the supposition that the avoidance of major battle was not Lettow’s strategy but Smuts’s. Deventer was to be supported indirectly, through an advance on Mombo and then Handeni: Lettow was to be outmanoeuvred, not outfought.
The rains in the north abated in the second week of May. Smuts decided to follow the line of the Pangani rather than of the railway. The east bank of the river was apparently undefended. The railway, on the other hand, marked a succession of German points of resistance, and its destruction by Kraut could delay Smuts’s progress if he allowed its repair to dominate the tempo of his advance. Smuts therefore organized his forces in three columns, the river column, the centre column (to follow the railway), and the eastern column (to move from Mbuyuni to the Ngulu Gap and then through the Pare Mountains). The centre column would threaten the German front, while the river and eastern columns moved to its flanks. Smuts’s aim was speed, not battle. Convinced that Lettow intended to reconcentrate on the northern railway, and reckoning that it would take him fifteen days to do so, Smuts wanted to get to Handeni and its railhead as soon as possible.
Kraut’s inclination was to fight. But the nature of the British flanking moves ensured that combat rarely occurred. The river column found its momentum slowed by thick bush. Thus, it threatened the Germans without ever endangering them. Kraut’s offensive spirit took second place to common sense, Schnee’s admonitions to prudence, and the availability of the railway line for rapid retreat. Only if the centre column had fixed the Germans in frontal assaults could Smuts’s succession of enveloping moves along the Pangani have had any effect.224
Kraut’s force was, of course, never designed to give battle. Nor did Lettow plan to reconcentrate at Handeni. Smuts’s advance was therefore as fast as he could have hoped. Beginning on 21 May, in ten days his troops covered 208 kilometres and reached Bwiko. Smuts then directed the river column on Mkalomo, assuming that Kraut would hold Mombo against the centre column.
Kraut intended to do so but was manoeuvred out anew, and so fought a brief action at Mkalomo on 9 June. The same happened at Handeni. Finally, the Germans took up defensive positions on the River Lukigura, north of the Nguru Mountains. Smuts hoped to take Kraut in the rear from the west, but the poverty of the maps and the difficulty of the terrain once again betrayed the ambition of the manoeuvre. Smuts had advanced 400 kilometres in under five weeks. The major constraint on his progress was not enemy action but his line of communications. The nearest railhead was Bwiko, 144 kilometres to his rear. His men were on half rations, restricted to a diet of hard-tack and mealie flour. A halt was called.
Smuts imagined that his advance to the Lukigura had eased the pressure on Deventer at Kondoa Irangi.225 He now envisaged the two columns advancing in tandem towards the central railway, pinning the Germans between Morogoro and Dodoma.226 In reality, Deventer’s lack of movement throughout May and June was due not to Lettow’s attentions but to health and supply problems. Moreover, Lettow had no intention of being tied to the central railway. At the end of April he and Schnee had agreed to abandon their original intention of withdrawing west to Tabora. Instead, they proposed to plunge south to Mahenge. Lettow used the veterinary su
rveys which Smuts had spurned, deliberately posting his men so as to draw mounted troops into fly-infested areas. Thus tsetse and extended communications would slow the British pursuit.227 Lettow began to move his troops away from the Kondoa Irangi front on 20 June, over three weeks before Deventer was ready to advance.228 He joined Kraut at Turiani in the Ngurus, his aim being to cover the evacuation of supplies from the central railway southwards.
On 14 July Deventer resumed his advance, directing his main column on Mpwapwa, and smaller columns on Dodoma and Kilimatinde. By the end of the month Deventer was astride the central railway and pushing eastwards on Kilosa to link up with Smuts.
Smuts spent July in the shadow of the Ngurus and the Königsberg’s guns. The main German positions were on Mount Kanga, facing the Lukigura. Smuts was informed that the valleys of the Mjonga and the Lwale, running north-south behind the Kanga massif, were practicable for troops. He therefore planned that one column (Sheppard’s) should engage the Germans on Kanga, while two more marched first west and then south following the valleys through the hills, converging on Turiani, thus cutting off the Germans’ line off retreat. Sheppard set of on 7 August, but found his progress through the bush and along the mountain slopes too slow. He therefore had to retrace his steps and skirt the mountains to the east, following the line of the Lukigura. Meanwhile, the advance of the two western columns, begun on 5 and 6 August, proved equally laborious. The 2nd mounted brigade, heading the Mhonda valley column, reached the Mjonga on 8 August. But its position was unobservable from the air owing to the thickness of the cover, and it had no wireless. Sheppard’s column was at this stage back at its start position, and the main bodies of the other two were only just beginning their southward ascent into the hills. Although the 2nd mounted brigade was in the right position, in isolation it lacked the strength to cut the German communications. Lettow made good his escape through Turiani to the Wami.