Our Friends Beneath the Sands

  A TENUOUS FRENCH PRESENCE in ‘the Red Island’ dated from 1643, when Cardinal Richelieu’s East Indies Company established Fort Dauphin in the far south as a trading post, but this was abandoned in 1674. During the eighteenth century French settlers on the Indian Ocean islands to the east – Macareigne, l’Ile Bourbon (Réunion) and l’Ile de France (Mauritius) – continued to trade with the eastern tribes through Tamatave, one of the few practical harbours on the east coast. The fall of the First Empire left France with only Réunion, and in 1817 the new British governor of Mauritius signed a treaty with Radama I, styling him King of Madagascar. (Thereafter Protestant missionaries had some success in converting the Hova nobility, and Christianity became the court religion.)3 Undiscouraged, in 1821 the Réunion French restored an earlier foothold on the eastern offshore island of Ste Marie, and in 1840 acquired that of Nossi Bé off the north-west coast, declaring a protectorate over the northern Sakalava tribes. King Radama II (r. 1861 – 63) signed a commercial treaty with Napoleon III, but his assassination by a group of aristocrats brought this to nothing. A remarkably adhesive prime minister, Rainilaiarivony, came to prominence in 1864; he was to be the husband of three successive queens, and behind their important ceremonial prestige he would manipulate effective power for thirty years, expertly juggling relations with the French and British.4

  In 1882 the land-hungry settlers of Réunion engineered the despatch of Admiral Pierre’s squadron to demand acceptance of a virtual French protectorate. Rebuffed, the admiral bombarded and occupied Tamatave, and put bluejackets ashore in the far north at Diégo Suarez. At that time Prime Minister Ferry was both preoccupied by the events unfolding in Tonkin and unwilling to provoke Britain, so he played for time. By the Convention of Tamatave signed on 19 December 1885, Imerina ceded Diégo Suarez to France and accepted a sort of phantom, unacknowledged protectorate for ten years (the French and Malagasy texts of this document varied significantly, and both sides signed it in bad faith). A resident-general was installed at the Hova capital (Le Myre de Vilers – the man who had sent Captain Rivière to Hanoi in March 1882); but while plantations were established on the east coast, and Malagasy Skirmishers were recruited at Diégo Suarez, for the most part the French officials were simply ignored. The deputies from Réunion, supported by Eugène Étienne’s colonial lobby, repeatedly pressed for military occupation, and in 1890 the diplomatic roadblock was cleared by an agreement giving France a free hand in Madagascar in return for a British monopoly in Zanzibar. Paris instructed Le Myre de Vilers to demand a new and much more tightly drawn treaty, and when this was refused by the government of Queen Ranovalona III the French began planning a military operation, which they anticipated would be quick and decisive.

  THE MOST OBVIOUS difficulty for a military force was the complete lack of roads in Madagascar. This was not a case of simple neglect but deliberate policy; it had distorted economic and military development, but was to the advantage of the court noblemen. In the 1820s the British had failed to persuade Radama I to build a road from the harbour at Tamatave up the eastern escarpment to Antananarivo, and periodic threats of French invasion made the continued difficulty of this obvious route to the capital an essential feature of Merina defence plans, such as they were.

  Since the island had roughly the population of modern Wales spread over an area larger than France, the main limitation on Merina prosperity was a shortage of manpower. Historically the rulers had relied upon slave-raiding, both within the island and on the African coast. Growing rice, the island’s staple food, was a labour-intensive activity that left few hands for work that might have increased wealth through trade; and in parallel, the lack of communications other than by rough tracks also put a premium on human porters. Except in European plantations on the east coast there was no wheeled transport, and there were taboos against using oxen as draft animals; virtually all movement of goods and passengers was by shoulder-poles, man-dragged sledges and palanquins. The Merina empire, ruled by an oligarchy of intensely competitive aristocrats clustered around the court in Anatananarivo, operated a system of forced state labour for all kinds of work, but alongside this a sophisticated commercial haulage business grew up to service all economic activity. Most of the porters were slaves owned by aristocratic syndicates based in the capital, who naturally profited by preserving the roadless status quo and keeping porterage rates high. It is calculated that by 1833 slaves already made up one-third of the total population of the Merina state, and two-thirds of the 50,000 inhabitants of the capital.5

  THE FRENCH WERE NOT IGNORANT of these conditions, and spent more than a year in planning. In August 1894 a joint commission of the War and Navy ministries called for an invasion force of at least 12,000 men, to be landed not at Tamatave – some 175 miles east and north of Antananarivo – but at Majunga in the north-west, about 50 miles further from the capital along a route avoiding the defensible eastern escarpments (see Map 8). From the jungle behind Majunga the south-eastwards march towards, into and across the Merina highlands would involve three main stages: in crow’s-flight terms rather than actual track miles, the first 100 miles up the Betsiboka river, through the lowlands and foothills, would end in a climb of about 2,000 feet on to the first shelf of the highlands – the Andriba plateau; some 30 miles further on, in the Ambohimenas hills, another climb up past Kinadji led to the highest pleateau, nearly 5,000 feet above sea level; after another 25 miles to Babay this central plateau dipped slightly to the high, fertile plain surrounding Antananarivo.

  Since the expedition would have to be supplied from Majunga throughout the march to the capital, the planners put great reliance on the fact that for the first 80-odd miles the Betsiboka was navigable, and in October 1894 the Navy Ministry ordered the construction of a flotilla of steam tugs and lighters drawing no more than 27 inches. That same month, from the French foothold at Diego Suarez, Lieutenant Aubé of the Naval Troops reconnoitred the route south as far as the Ambohimenas and made sketches (the only available maps of the whole island were in an unhelpful 1:1 million scale). It would be impossible to obtain locally the 18,000 – 20,000 porters needed to carry supplies for a 12,000-man force, but it was calculated that this requirement could be reduced by 75 per cent by the generous provision of voitures Lefèbvre – small two-wheeled iron carts, drawn by a mule and carrying a 440lb payload. Aubé reported that only moderate work was needed to make the tracks on the inland plateau passable for these, and no fewer than 5,000 carts were ordered.6

  In November 1894 it was suddenly announced that the whole expedition was to be organized by the Army (it is said that the war minister, General Mercier, got it by underbidding the Navy for funding by 30 million francs). While the Army had contributed troops for the basically Navy campaigns in Tonkin and Dahomey, it had not itself organized a full-scale overseas expedition since Mexico in the 1860s. This lack of experience would prove most telling in the decision that two-thirds of the infantry would be European, including four battalions of Metropolitan Line and Light volunteers.7

  Naval Troops officers argued that the men would not be acclimatized, the Majunga route was too long, and the expedition needlessly large – while the Hova army was estimated at some 40,000 men, its morale and readiness were uncertain. These arguments for a smaller, lighter and faster-moving force were dismissed by the Metropolitan generals as typical seat-of-the-pants Naval Troops improvisation (given the rivalry between the two services, the Navy’s near-monopoly of eye-catching overseas campaigns was resented). They insisted that this was a case for Army professionalism: the Merina government was known to have purchased about 25,000 breech-loading single-shot rifles, and had acquired modern artillery and British instructors.8 On 23 November 1894 the National Assembly voted the funds for the expedition, and on 11 December the Foreign Ministry informed friendly nations that a state of war existed between France and Madagascar.

  Command of the expeditionary force was given to Army General Duchesne, a veteran of both Tonkin and the débâcle on Formosa, w
ith Army General Metzinger and Navy General Voyron to command his two brigades. The expedition totalled some 15,400 officers and men, including about 10,400 infantry in thirteen 800-man battalions: four Metropolitan, three Naval Infantry, two Algerian, and one each Legion, Malagasy Skirmishers, Hausas and white Réunion volunteers. There was an Africa Light Horse squadron for scouting, and eight artillery batteries. Engineer and Train (transport) units were almost entirely Metropolitan, but the main logistic manpower would be provided by 5,500 hired Algerian muleteers-cum-cart drivers (‘Kabylies’), with 5,640 mules for the Lefèbvre carts. A Navy coastal and river flotilla of eight small warships was attached to the force, and thirty steamers were ‘hired in from trade’ to transport the expeditionary corps.9

  SMALL DIVERSIONS WERE PROVIDED, between December 1894 and April 1895, by marsouins landed at Tamatave and others patrolling from the Diégo Suarez garrison. Another few naval companies landed unopposed at Majunga on 15 January 1895, and during February they reconnoitred inland, encountering no resistance from vassals of the Merina empire. Serious movements would not be possible before the dry season began in May, but the advance element of the expeditionary corps proper came ashore with General Metzinger on 1 March 1895 – a mixed force of Algerians, marsouins, Malagasies and support troops – and began unloading engineer materials and stores.

  The plan called for Metzinger to establish the base camp, and then reconnoitre for 50 miles up the Betsiboka river during the tail-end of the rains in April. The second phase was to start when General Duchesne arrived with the bulk of the force in May. Metzinger’s 1st Brigade would then advance south to the confluence of the Betsiboka and Ikopa rivers – the limit of navigation – by the first week in June; from there they would build a short cart road to the site for an advanced supply base at Marololo. From Marololo southwards the third-phase advance would proceed in bounds, with the leading battalions improving the tracks towards the Andriba plateau that formed the northern ramparts of Imerina, followed by the second brigade escorting the supply-cart convoys. It was expected to take another six weeks – say, until early August – to get the whole force from Marololo up on to the Andriba plateau; thereafter, the activity of the Hova army would decide events.

  MAJUNGA LAY JUST INSIDE BOMBETOKE BAY, 25 miles short of the chaotic sandbars and mangrove swamps at the mouth of the Betsiboka river. An unsuspected coral reef kept large ships further from shore than had been anticipated, and heavy swells swamped some of the landing boats, making all unloading difficult and time-consuming. Building the planned wharf took months, so hastily landed crates were scattered over 300 yards of beach to lie unprotected until they were identified and collected. The ships carrying the disassembled sections of the gunboats and cargo lighters and the first 1,000 carts were late, arriving between 18 April and 7 May, and assembling and launching the boats in the choppy bay proved much harder than expected. This cancelled the riverborne element of the vanguard’s planned advance in April, and although gunboats were in operation by 1 May it seems to have been at least three weeks later before the first launch towed a pair of lighters up river. Thereafter a couple of lighters every two or three days seems to have been the norm, and this shortfall in the planned river transport capacity was to have malign consequences.10

  Seeking to replace the boats with hired porters, the Train only managed to recruit some 400 Sakalavas. About 7,700 porters would eventually be assembled by June, most of them from as far afield as Indochina, Somalia and even Algeria, which suggests a certain desperation. Crucially, the delay in getting the river supply line working meant that the heavy, narrow-wheeled Lefèbvre mule carts had to be used from the start instead of being kept for the harder ground of the uplands.11 They immediately became bogged down on the tracks through the swampy northern jungle, and in the absence of native labour the four Engineer companies had to be diverted from skilled work to road-making.

  The lords of the lowland jungle were countless billions of malarial mosquitoes. The sappers quickly became worn out by the stifling climate and by fever casualties, so survivors were relieved and saved for specialist bridging work, and many of the infantry had to take their place in road-gangs as soon as they came off the beach. Progress in building the cart road was also held up periodically by the need to bridge a succession of streams and rivers up to 100 yards wide and, in the eventual case of the Betsiboka near Marololo, 450 yards (the Merina empire relied on fords, which were often impassable at this season). In a vicious spiral, bridging materials were delayed because they had to compete with rations for the load-space on the few lighters getting up river; rations were needed for troops already weakened by the road-building that was intended to circumvent the problem of insufficient boat-space; yet the road they were building was not yet adequate to get the heavy bridging materials forward.

  Despite heavy rain and floods the advance guard began to venture inland from Majunga on 25 March, without serious resistance. During April, the 13th Naval and elements of the Algeria Marching Regiment landed, and on 28 April the bulk of the Legion battalion arrived aboard the Liban.

  THE TWO FOREIGN REGIMENTS had each been ordered, on 4 January 1895, to provide two companies for the 1st Battalion of the Algeria Marching Regiment (I/RMdA), which was officially created on 2 February under the command of Colonel Oudri of the 2nd Foreign. The 1st Battalion, formed during March under the command of Major Barre, had sailed from Oran at the beginning of April with 22 officers and 818 rankers. Private Silbermann recalled a difficult landing over the unfinished 160-yard wharf; he described Majunga as busy but filthy, with disreputable-looking European and Creole traders already erecting makeshift ‘booze-shacks’ in defiance of the harassed town commandant.12

  The Legion started marching up river on 1 May, part of three columns of the RMdA, marsouins and Malagasy Skirmishers. (The haphazard arrival of different troopships over a month, and the need to have this latter local unit at the spearhead – where it performed consistently well – distorted the theoretical brigade organization from the start.) The columns chased Hova troops out of a 6-mile line of hill positions astride the river in front of Marovoay without difficulty but, camping in the tall grass that night, Silbermann had to stay awake to keep a smoke-fire burning – the mosquitoes were worse than anything he had known in Dahomey. On 6 May, General Duchesne arrived at Majunga; his infantry would be complete by the 11th, but his artillery, service troops, mules, drivers and stores would arrive only in individual shiploads during the rest of May and June.13

  The expedition commander went forward to the head of the advance on 16 May to confer with Metzinger about the apparently elusive enemy. The Hovas and their famous modern artillery might be anywhere, and any general out of contact with a retreating enemy fears being lured into a trap. While the rest of the corps began to build up behind them, the advance guard had to keep pushing on through the marshy forest, trying to get a feel for the Hova dispositions and possible intentions. Between 13 and 17 May, Colonel Oudri led the légionnaires towards a strong position held by some 2,000 Hovas at Ambodimonti, to link up there with Lieutenant-Colonel Pardes’ Malagasies and Algerians. The turcos had an undemanding encounter with a Hova column, but when Oudri and Pardes met at Ambodimonti they found the trenches empty except for 2,000 abandoned shells for Krupp and Hotchkiss field guns and 10,000 rounds of small arms ammunition, including boxes of American .45ca1 for Gardner machine guns. The Hovas did not seem to be the men (or women) that Behanzin’s Fon had been in 1892, but so far their tactics were incomprehensible.14

  IN TRUTH, THE HOVAS HAD NO TACTICS, because their 37,000 armed men were not an army. Among the aristocracy, motivated entirely by factional rivalries, military commands were little more than titular honours. Any of the principles of modern soldiering that European instructors had managed to instil in a few literate young noblemen were shallow-rooted, since their culture provided no seedbed.The royal court took pride in its half-dozen batteries of modern artillery, but there were few trained gunners; the guns w
ere symbols of power rather than familiar tools, and the majority of them were kept at Antananarivo to overawe the population. The royal guards, too, were simply ‘anti-coup troops’, and the Hova aristocracy had never dared to equip and train a true army of their subjects. An actual majority of the common people of the Hova heartland were slaves, the rest subsistence farmers, and the rank and file belatedly conscripted in 1895 had no reason for military enthusiasm. They were also handicapped by a local warrior culture of brief raiding and pillage after ritual displays, rather than disciplined manoeuvre or sustained effort (in this they were clearly inferior to Behanzin’s smaller and ostensibly more primitive Dahomeyan army). Even after the exploratory French landing at Majunga in January, the ministers of old Rainilaiarivony’s government did little beyond making speeches, and the conscripts gathered in a large camp outside Antananarivo became increasingly restive and nervous. After visiting them, the London Times correspondent E. F. Knight wrote:They were a ragged lot, and discipline there appeared to be none . . . When paraded before the [prime minister] or some other great man, they used to raise cheers and brandish their arms, while their officers waved their swords . . . and simulated the slaughter of the foe. These were practically the sole manoeuvres, for the drill their European officers had taught them was now neglected as foreign trickery unworthy of Hova warriors.