SECTION XIX THE PORTUGUESE CAMPAIGN OF 1810. THE PRELIMINARIES: JANUARY–AUGUST 1810 CHAPTER I

THE MILITARY GEOGRAPHY OF PORTUGAL

The continual existence of Portugal down to the present day in face of the persistent hostility and immensely superior force of its neighbour Spain seems at first sight to be one of the most inexplicable phenomena in modern history. It appears all the more astounding when we remember that the lesser kingdom was once conquered, and held down for sixty years, by the greater power. Few states have won back and maintained their independence in such masterful fashion as did Portugal, in the long ‘War of Independence’ that followed the insurrection of 1640 under the house of Braganza. But intense national spirit and heroic obstinacy on the part of the smaller people are not sufficient to account for the survival of the Portuguese kingdom as a separate entity. Its geography, which at the first sight seems hopelessly unfavourable to its defence, turns out on investigation to be eminently suitable for resistance against an attack from the east. On a first glance at the map it appears as if Portugal was composed of no more than the lower valleys of three great Spanish rivers, the Douro, the Tagus, and the Guadiana, so that the state which owns three-quarters of the course of each of these streams has but to send down its armies from the uplands of Leon and New Castile, to conquer the narrow land which lies about their estuaries. But nothing can be more deceptive than the map, when the Iberian Peninsula is in question. As we observed in our earlier volume[150], the rivers of Spain and Portugal are not highways, or lines of communication,[p. 154] but barriers—torrents sunk in gorges cut deep below the level of the face of the land. The chief roads, with few exceptions, avoid, instead of courting, the neighbourhood of the great streams. The leading routes which descend from Spain into Portugal in no case follow the lines of the Douro or the Tagus. Though the coast-plains, which form the heart of the kingdom of Portugal, its most wealthy and populous regions, lie about the mouths of those rivers, it is not by descending their banks that conquest or trade arrives most easily at its goal. As a matter of fact, Spain and Portugal turn their backs upon each other: the smaller realm looks out upon the sea; her strength and wealth lie upon the Atlantic coast: the inland that touches Spain is rugged and unpeopled, in many parts a mere waste of rock and heath. Nor, on the other hand, do Leon and New Castile look towards Portugal: the real ports of Madrid are Valencia and Alicante, not Lisbon, and that not from political reasons, but simply because those are the points where the sea can be reached with the minimum of mountain and desert to be passed through. The way down from the central tableland of Spain to the Mediterranean is less difficult than the way down to the Atlantic. Hence comes the fact that the high-roads leading from Spain into Portugal are so surprisingly few, and that the two main alternative routes from Madrid to Lisbon run, the one much further north, the other much further south, than might have been expected. There is not now, and never has been, any straight road down the Tagus between the two capitals, obvious though the line looks upon the map. The two main gates of Portugal are at Almeida and Elvas; at Alcantara, which appears the natural point of approach, there is but the most miserable of posterns—as Junot discovered in November 1807, much to his discomfiture. Marshal Berwick had made the same experience in 1705, during the War of the Spanish Succession[151].

In the old wars between Spain and Portugal the whole land frontier of the smaller kingdom was exposed to attacks from the larger. But the circumstances of 1810 differed from those of[p. 155] 1705 or 1762 or 1801, in that the subsidiary campaigns in the extreme north and south, which had always accompanied the main clash along the frontiers of Beira and Alemtejo, could not on this occasion take place. The French were no longer in possession of Galicia, from which the Spaniards had been wont to demonstrate against Oporto, nor, at the other extremity of the line, had they a firm grip on Huelva, and the Condado de Niebla, from which alone an attack could be directed against the remote southern province of Algarve.

Portugal presents three sections of frontier to an invader coming from the side of Spain. The northernmost, that from the mouth of the Minho to Miranda-de-Douro, was not within the scope of operations in 1810. It can only be approached from Galicia; that province was not subdued, nor had the French any intention of dealing with it till after they should have dealt with Portugal. An invasion of the Tras-os-Montes and the Entre-Douro-e-Minho would have been an objectless operation: they would fall of themselves if once Lisbon were captured and the English expelled from the Peninsula. A move against Oporto by some flanking division of the invading army might have been conceivable, but such an attempt would be made, if made at all, from the south of the Douro, through northern Beira, and not through the mountains of the Tras-os-Montes.

There remain two other sections of the Portuguese frontier: the one from the Douro to the Tagus, and the other from the Tagus to the Guadiana. Both of these were accessible to the French in 1810, since they were in possession alike of the plains of Leon and of La Mancha, and of northern Andalusia. It was open to them to choose one or the other front for attack, or to attack both at once. Lisbon being the objective, it was clear that an attack on the northern or Beira frontier possessed a paramount advantage over an attack on the southern or Alemtejo frontier. A successful advance north of the Tagus brings the invader directly to the gates of Lisbon; one south of the Tagus brings him only to the heights of Almada, where he is separated from the Portuguese capital by the broad estuary of the Tagus. Napoleon’s power, like that of the devil in mediaeval legends, ended at the edge of the salt water; and in face of the naval strength which the[p. 156] English always maintained at Lisbon, a victorious French army camped on the heights of Almada would be almost as far from final success as when it started from Spain. The 1,900 yards of strait which protected the Portuguese capital could not be crossed. The most that the invader could accomplish would be to worry the ships in the port, and the lower quarters of the city, by a distant bombardment, if he could bring up heavy guns from Spain[152]. For nearly twenty miles inland from Lisbon the estuary of the Tagus expands into a broad brackish lagoon four to eleven miles broad, a complete protection against any attack from the east. Only at Alhandra does this inland sea contract, and for some further miles northward from that point the eastern bank of the Tagus is formed by broad salt-marshes (lezirias) cut up by countless channels of water, and practically inaccessible. It is only at Salvaterra, thirty miles north of Lisbon, that the Tagus assumes its ordinary breadth, and becomes an ordinary military obstacle. From that point upwards an invader from the Andalusian side might endeavour to cross it, and it presents no more difficulties than any other broad river. But, though even Rhines and Danubes may be passed in the face of an enemy, the operation is not one which a prudent general courts, and the Tagus is broad, absolutely bridgeless, and fickle in the extreme in its alternations of high and low water. To fight one’s way from the valley of the Guadiana in order to meet such a problem at the end does not seem inviting. And even if the Tagus is passed, there are still thirty miles of road, including some formidable defensive positions, between the invader and Lisbon[153]. Yet there was one contingency under which an advance on the left bank of the river might be advantageous to the invader, and so possible was this contingency that Wellington from the very first had declared that he thought it probable that the French would move troops in that direction. If the[p. 157] Anglo-Portuguese army were drawn away to the Beira frontier, between Tagus and Douro, in order to resist a front attack delivered from the plains of Leon, and if it became involved in an active campaign somewhere far to the north, on the line of the Coa, or the Mondego, or the Alva, a subsidiary French force, striking south of the Tagus from the direction of Spanish Estremadura, might give dreadful trouble. If it could cross the Tagus anywhere between Abrantes and Salvaterra, it might get between the Anglo-Portuguese army and its base, and either fall upon its rear or capture Lisbon. For this reason Wellington, so far back as October 1809, had made up his mind that, if the French had an army on foot anywhere in the direction of Badajoz and Elvas, he must leave a considerable proportion of his own forces to watch them, and to defend, if need be, the line of the lower Tagus[154]. As long as the enemy had not yet subdued Badajoz and the neighbouring fortresses, and while there was still a strong Spanish army in that quarter, the need for precaution was not so pressing. Nevertheless, all through the summer of 1810 Wellington kept Hill with one English and one Portuguese division at Portalegre, south of the Tagus, though he withdrew this detachment when Masséna marched on Coimbra. Matters were much more perilous after the battle of the Gebora and the fall of Badajoz in February 1811. From that time onward, all through 1811 and 1812, nearly a third of the Anglo-Portuguese army was kept in the Alemtejo, first under[p. 158] Beresford, then under Hill, in order to guard against the possible stab in the back from the French army of Andalusia.

But the attack south of the Tagus was in Wellington’s, and also, we may add, in Napoleon’s conception[155], only a secondary operation. The main invasion was almost inevitably bound to take place on the Beira, not on the Alemtejo frontier. Between Fregeneda, where the Portuguese border line quits the Douro, to the pass of Villa Velha on the Tagus there is a distance of somewhat more than 100 miles. The division between Portugal and Spain does not lie along any well-marked natural feature, such as a mountain range or a broad river—though two small sections of the frontier are coincident with the insignificant streams of the Elga and the Agueda. It is rather drawn, in a somewhat arbitrary and haphazard fashion, through the midst of the desert upland, where Spain and Portugal turn their backs to each other. For the only piece of flat plain-land on the whole border is that from the Douro to Almeida, a mere ten or twelve miles, and immediately behind the Coa, only three or four miles from Almeida, the mountains begin. The rest of the frontier runs through thinly-peopled, barren highlands, from which the Coa, the Mondego, the Zezere, and the Pon?ul fall away towards Portugal, and the Agueda and the Alagon towards Spain. The mountains are not, for the most part, very high—the culminating peak of the Serra da Estrella is only 6,540 feet—but they are singularly rugged and scarped, and much cleft by ravines, along whose sides the few roads crawl miserably, in constant precipitous dips and rises. This broad belt of upland, one long series of defiles for an invader, is some hundred miles broad, and does not cease till Coimbra, on the one side, or Abrantes, on the other, is reached: only then does the plain-land begin, and the country-side become fertile and thickly peopled. Only from those points onward is it possible for an army to live on the local produce: in the upland it must carry its food with it; for a single division would exhaust in a day the stores of the poor villages of the mountains; and the small[p. 159] poverty-stricken towns of Guarda, Celorico, Sabugal, Penamacor, and Idanha have few resources. Castello Branco and Vizeu are the only two places in the upland where there is a valley of some breadth and richness, which can supply an army for many days. In this simple fact lies the explanation of the difficulties of the Portuguese campaigns of 1810 and 1811. Both the invader and the defender must bring their food with them, and protracted operations can only be kept up by means of incessant convoys from the rear. The campaign not infrequently became a starving-match, and the combatant who first exhausted his provisions had to retire, and to disperse his divisions in search of the wherewithal to live. Thanks to Wellington’s providence it was always the French who were forced to this expedient.

The Beira frontier is divided into two sections by the range of mountains which crosses the border at right angles, half way between Douro and Tagus: it is known as the Sierras de Gata and de Jarama while in Spain, as the Serra da Estrella when it reaches Portugal. Its central ganglion lies between the high-lying towns of Sabugal and Penamacor in Portugal and the pass of Perales in Spain. From this point run off the great spurs which separate the valleys of the Pon?ul, the Zezere, the Agueda, the Coa, and the Alagon. An invader must make his choice whether he will advance into Portugal south or north of the Serra da Estrella: to attempt to do so on both sides of the range would be risking too much, if there is an enemy of any strength in the field, since the columns to the right and to the left would be hopelessly separated, and liable to be beaten in detail. In the whole Peninsular War there was only one invasion made by the southern route, that of Junot in the winter of 1807-8. It was successful because it was absolutely unopposed. Nevertheless the French lost many men, had to leave their artillery behind them, and only arrived with the shadow of an army at Abrantes. It is true that Junot chose absolutely the worst path that could be found between the Serra da Estrella and the Tagus—the pass of Rosmarinhal, close above the latter river—and that he would have fared not quite so badly if he had marched from Zarza on Idanha and Castello Branco. But even at the best this region is most inhospitable: there are points where water is not procurable on stretches of eight or ten miles, others where the main[p. 160] road is so steep that a six-pounder requires not only a dozen horses but the assistance of fifty men to get it up the slope. Except in the immediate neighbourhood of Castello Branco, the country-side is almost uninhabited[156]. The whole ‘corregidoria,’ which took its name from that town, and extended from the Elga to the Zezere, had only 40,000 souls in its broad limits—it was forty miles long by thirty broad, the size of a large English county. Junot’s experiences served as a warning to his successors, and no French army during the rest of the war endeavoured to cross this corner of Portugal when advancing on Lisbon. Castello Branco was seized once or twice by a raiding force, but it was never used as the starting-point of an army making a serious attempt to advance towards the Portuguese capital.

There remains to be considered only the section of the frontier between the Douro and the Serra da Estrella, the front on which Masséna’s great blow was delivered in the autumn of 1810. It was a region which had from the earliest times been the battle-ground of the Spanish and Portuguese. Half a dozen times since the Middle Ages armies from the plains of Leon had invaded the Beira on this front. Such campaigns always began with a siege of Almeida, the sentinel-fortress pushed out in front of the mountains to face the Spanish Ciudad Rodrigo. Almeida generally fell—it is too advanced a position for safety, and (as Dumouriez remarked in his military study of Portugal) its value would have doubled if it had only been placed upon the west instead of the east bank of the Coa, close to the friendly mountains, and not on the outskirts of the perilous plain. But rare, indeed, were the occasions on which the Spaniard succeeded in piercing the broad belt of tangled upland beyond Almeida, and appearing at the gates of Coimbra or Oporto.
Map of Central Portugal

Enlarge  CENTRAL PORTUGAL

There are four lines of further advance open to an invader who has captured Almeida. The first, a march on Oporto via Pinhel and Lamego, may be mentioned only to be dismissed from consideration. It is of no use to an army which aims at Lisbon, and proposes to conquer Portugal by a blow at its heart. Masséna, whose directions were to drive the British into the sea in[p. 161] the shortest and most effective fashion, could not have contemplated such a secondary object as the capture of Oporto for a moment. There remain three other roads to be investigated.

(1) The road north of the river Mondego, by Celorico, Vizeu, Bussaco, and Coimbra. (2) The corresponding and parallel road south of the Mondego, from Celorico by Chamusca, Maceira and Ponte de Murcella to Coimbra. (3) The road which, striking south from Celorico, crosses the headwaters of the Zezere, by Belmonte and Fund?o, and then, climbing the Serra de Moradal, descends to Castello Branco, and from thence reaches Abrantes by the Sobreira Formosa. It may be remarked, by the way, that nothing in all the geography of Portugal seems more astonishing than that there should not be a fourth alternative road, one down the long valley of the Zezere, which, running in a straight line from Belmonte to Abrantes, looks on the map as if it ought to be a main artery of communication, and seems to indicate the obvious road to Lisbon from Almeida, since a straight line drawn between these points would run along the river for some forty miles. But as a matter of fact there was neither a first nor a second-rate road down the Zezere: the only towns on its course, Covilh?o and Belmonte, lie hard by its sources, and its central reaches were almost uninhabited. The only good line of communication running near it is a by-road or duplication of the third route mentioned above, called the Estrada Nova, which, leaving the upper Zezere at Belmonte, keeps high up the side of the Serra de Moradal, and rejoins the Castello Branco road at Sobreira Formosa. This route was much employed by Wellington in later years, as a military road from north to south, usable even when Castello Branco was threatened by the French. But in 1810 he had ordered it to be rendered impassable, and this had been done by making several long cuttings at points where the track passed along precipices, the whole roadway being blown or shovelled down into the gulf below[157]. The French were, of course, unaware of this, and[p. 162] Masséna is said by his confidant Foy to have taken the Estrada Nova into serious consideration, and to have decided against it because of the necessity for forcing the passage of the Zezere when the defiles were passed, and for laying siege to Abrantes[158]. A far more practical objection was its extreme wildness: it runs along an absolutely uninhabited mountain-side, and the neighbourhood is destitute not only of food but of water for great sections of its length. This Masséna ought to have known, if his Portuguese advisers had been competent. Apparently he was wholly unaware of its character, just as he was necessarily ignorant of the fact that his prescient adversary had blasted away huge sections of it, so that it was absolutely impassable for guns or wagons, as also that earthworks had been carefully constructed to cover the point where it debouches on to the Zezere.

The Castello Branco road, therefore, with this dependent by-road, the Estrada Nova, was practically left out of consideration by the Marshal. There remains the choice between the two northern routes, Celorico-Vizeu-Coimbra, and Celorico-Chamusca-Ponte de Murcella-Coimbra. Both traverse rough ground—but ground less rough than that to be found on some parts of the Castello Branco road. Along both there is an intermittent belt of cultivated land, and not unfrequent villages. Both are intersected by many good military positions, on which a defending army can offer battle to an invader with advantage. In especial, the northern road strikes and climbs the granite ridge of Bussaco with every disadvantage for the attacking side, and the southern road is contracted into a difficult defile at the passage of the Alva near Ponte de Murcella. On the whole,[p. 163] however, this last is the better line for advance:—the strongest testimonial to the fact is that Wellington expected Masséna to take it, and erected at the passage of the Alva almost[159] the only earthworks, save those of the lines of Torres Vedras, which he constructed in his preparations for the reception of the invader. When he first heard that the Marshal was moving forward from Almeida to Celorico, and was clearly aiming at the Mondego valley, he announced that he should endeavour to stop the invader on the Alva[160], not apparently thinking it at all probable that Masséna would move by Vizeu and the north bank of the Mondego. On realizing that this was really his adversary’s design, he observed with some exultation that, while there were certainly many bad roads in Portugal, the enemy had taken decidedly the worst of those open to him[161]; moreover, he had committed himself to attack the heights of Bussaco, the most formidable position in the whole of northern Portugal. How the French commander came to make this choice we shall discuss in its proper place. Suffice it to say that Wellington had not realized how bad was Masséna’s information, how worthless his maps, and—what is most surprising of all—how entirely destitute of local knowledge were the Portuguese traitors—Alorna, Pamplona, and the other renegade officers—whom the Emperor had sent as guides and advisers to the Marshal. And in truth, the unsuspected ignorance of Masséna and his advisers added an incalculable element of chance to the problem set before Wellington. He was obliged to make his plans on the hypothesis that the enemy would make the correct move: and not unfrequently the enemy, for reasons which the English general could not possibly foresee, made the wrong one.

The French invasion was bound to commence with a preliminary clearance of the outlying fortresses still in the hands of the Spaniards. These to some extent protected the Portuguese frontier in 1810, though they had been built with the express purpose, not of protecting, but of threatening it, and had never[p. 164] before been attacked by an enemy coming from the east. Only three of these fortresses were of any importance—Astorga, Ciudad Rodrigo, and Badajoz. The other strongholds of Spanish Estremadura, Alcantara (which had stood a siege in the War of the Spanish Succession), Albuquerque, Olivenza, were either not in a state of defence at all, or were hopelessly antiquated, and little suited to face modern artillery and modern siegecraft.

Astorga lies so far to the north that it might have been neglected without much peril to the French scheme of invasion. But the Emperor had ordered that it should be reduced before the great enterprise began: it gave the Spanish army of Galicia a foothold in the plains of Leon, from which it might operate against Masséna’s rear, if he should pass it by. Its capture, too, was considered a matter of small difficulty, for it was but a mediaeval walled town, to which some hasty outworks had been added during the last year. It will be remembered that when Moore passed that way in January 1809, Astorga had been treated by both sides as an open town, and no attempt had been made to garrison or defend it. Since then La Romana had repaired its dilapidated enceinte, stockaded its suburbs, and armed it with guns brought from Ferrol. As late as January 11, 1810, Napoleon seems hardly aware of this fact: in a dispatch of that date he orders Loison to make his head quarters there, evidently under the impression that it is not held by the Spaniards, or at least that it is a place which they will evacuate at the first appearance of a serious attack[162]. It is only in March that he writes to Junot that Astorga must be besieged and taken, in order to occupy the attention of the Galicians and to thrust them back into their mountains[163].

Ciudad Rodrigo was a more serious business. It was a regular fortress, though only one of the second class; its prestige, as the only Spanish stronghold on the Portuguese frontier, was great. It commands the whole southern stretch of the plains of Leon, being the only place out of the control of an invader who is superior in cavalry, and therefore master of the defenceless Tierra de Campos. There was also a small Spanish army depending upon it, and clinging to the skirts of the Sierra de Gata. This was the division[p. 165] of Martin de la Carrera, which had been left behind when the greater part of Del Parque’s Army of the Left marched down into Estremadura in January 1810, in order to replace there the troops which had gone off to the defence of Cadiz. It was clear that Ciudad Rodrigo must be taken, and Martin de la Carrera brushed away or destroyed, before any serious attempt to invade Portugal was begun. If the Emperor thought that such a remote place as Astorga was worth his notice, it was obvious that he would regard Ciudad Rodrigo as absolutely indispensable to his designs. It was for its reduction that he gave Masséna the great battering train of fifty heavy guns, with 2,500 artillerymen and sappers, which was assigned to him, independent of the artillery of the three corps of the Army of Portugal.

Badajoz, far to the south, in Spanish Estremadura, stood to the defence of Southern Portugal exactly as Ciudad Rodrigo to the defence of Northern Portugal. It possessed also in Elvas a counterpart to Almeida. But Badajoz is immensely larger and stronger than Rodrigo, just as Elvas is infinitely more formidable than Almeida. The two fortresses on the frontier of Leon are small places crowning mere mounds set in a plain. Badajoz and Elvas have towering citadels set on rugged hills, and overlooking the whole country-side. They have also strong detached forts on dominant positions: the circuit of ground that must be taken up by an army that intends to besiege them is very large, and at Badajoz there is a first-class river, the Guadiana, which cuts in two the lines which the assailant must occupy. It may be added that based on Badajoz there was a whole Spanish army of 15,000 men, not a mere division of 3,000, like that which lurked in the mountains above Rodrigo. Noting the strength of Badajoz and Elvas, the Emperor had made up his mind that they should be observed and ‘contained’ by troops from the Army of Andalusia, but not attacked till Lisbon had been conquered and the English expelled from Portugal. ‘Les Anglais une fois battus et rembarqués, Badajoz et Elvas tombent d’eux-mêmes,’ he wrote in a holograph minute addressed to Masséna, just before the advance across the Portuguese frontier began[164]. It was only when the invasion had been brought to a standstill before the lines of Torres Vedras, that he came to the conclusion that pressure must[p. 166] be applied south of the Tagus, to distract Wellington, and that Soult, as a preliminary to an attack on the Alemtejo, must capture Badajoz and Elvas, and disperse the Spanish Army of Estremadura. The idea came to him tardily: thanks to Wellington’s careful starvation of the main French army, Masséna was forced to retreat into Leon when Soult had only recently captured Badajoz, and had not yet shown a man in front of Elvas. The scheme was hatched, like so many of Napoleon’s Spanish plans, about three months too late for effective realization. As late as November 1810 the orders to Soult are merely to demonstrate against Badajoz, and to hinder the departure of La Romana’s army for the lines of Torres Vedras, but not to besiege the Estremaduran fortress. Probably it was Foy’s information as to the existence and strength of the unsuspected lines of Torres Vedras, which reached him late in November, that made the Emperor realize the advisability of that secondary attack by the south bank of the Tagus which Wellington had foreseen and taken means to meet a full year before. Fortunately a capable general on the defensive always knows his own weak spots long before they are discovered by the enemy. By the time that Soult had at last captured Badajoz and the small dependent places—Olivenza, Albuquerque, Campo Mayor—Wellington had got rid of Masséna from the neighbourhood of Lisbon, and was preparing to chase him home across Northern Portugal. Within a few weeks of its surrender by the traitor Imaz, Badajoz was being besieged by a detachment of the British army, and Soult had his hands full, as he strove at once to hold down Andalusia and to relieve the beleaguered fortress.