CHAPTER V

MARMONT TAKES THE OFFENSIVE. JULY 1812

On July 2nd Wellington had arrived at the end of the first stage of his campaign. He had cleared the French out of the whole of southern Leon as far as the Douro, had taken the Salamanca forts, and had beaten off with ease Marmont’s attempts to meddle with him. All this had been accomplished with the loss of less than 500 men. But the success, though marked, was not decisive, since the enemy’s army had not been beaten in the open field, but only manœuvred out of the considerable region that it had evacuated. The most tangible advantage secured was that Marmont had been cut off from Madrid and the Army of the Centre: he could now communicate with King Joseph only by the circuitous line through Segovia. All the guerrilleros of Castile, especially the bands of Saornil and Principe, were thrown on the Segovia and Avila roads, where they served Wellington excellently, for they captured most of the dispatches which were passing between King Joseph and Marmont, who were really out of touch with each other after the Marshal’s retreat from the Tormes on June 27th.

But till Marmont had been beaten in action nothing was settled, and Wellington had been disappointed of his hope that the Army of Portugal would attack him in position, and allow him to deal with it in the style of Bussaco. The Marshal had retired behind the Douro with his host intact: it was certain that he would be joined there by Bonnet’s division from the Asturias, and very possible that he might also receive succour from the Army of the North. The junction of Bonnet would give him a practical equality in numbers with the British army: any considerable reinforcement from Caffarelli would make him superior in force. And there was still a chance that other French armies might intervene, though hitherto there were no signs of it. For it was only during the first fortnight of the campaign that Wellington could reckon on having to deal with his immediate adversary alone. He was bound to have that much start, owing to the wide dispersion of the French, and their difficulty in communicating with each other. But as the weeks wore on, and the enemy became more able to grasp the situation, there was a growing possibility that outlying forces might be brought up towards the Douro. If Marmont had only been defeated on June 21st this would have mattered little: and Wellington must have regretted more and more each day that he had not taken the obvious opportunity, and attacked the Army of Portugal when it placed itself, incomplete and in a poor position, beneath the heights of San Cristobal.

Now, however, since Marmont had got away intact, everything depended on the working of the various diversions which had been prepared to distract the other French armies. One of them, Sir Home Popham’s, had succeeded to admiration, and had so scared Caffarelli that not a man of the Army of the North was yet in motion toward the Douro. And this fortunate expedition was to continue effective: for another three weeks Marmont got no succours from the army that was supposed to constitute his supporting force by the instructions of the Emperor and of King Joseph. But Wellington--not having the gift of prophecy, though he could see further into the fog of war than other men--was unable to rely with certainty on Caffarelli’s continued abstinence from interference. As to Soult, there were as yet no signs of any trouble from Andalusia. The Duke of Dalmatia had somewhat reinforced D’Erlon’s corps in Estremadura, but not to such an extent as threatened any real danger to Hill, who reported that he could keep D’Erlon in check on the Albuera position, and was not certain that he might not be able to attack him at advantage--a move for which he had his chief’s permission[453]. If only Wellington had been fortunate enough to receive some of Soult’s letters to King Joseph, written in the second half of June, he would have been much reassured: for the Marshal was (as we shall see) refusing in the most insubordinate style to carry out the orders sent him to move troops northward. Two minor pieces of intelligence from the South were of no primary importance--though vexatious enough--one was that Ballasteros had ventured on a battle at Bornos on June 1, and got well beaten: but his army was not destroyed. The second was that General Slade had suffered a discreditable check at Maguilla on June 11th in a cavalry combat with Lallemand’s dragoons. But neither of these events had much influence on Soult’s general conduct at the time, as we shall show in the proper place.

[453] See Wellington to Lord Liverpool, June 25. _Dispatches_, ix. pp. 253-4, and to Hill, ix. pp. 256-7, and again to Lord Liverpool, ix. pp. 261-2.

There remained one quarter from which Wellington had received information that was somewhat disturbing. An intercepted letter from King Joseph to D’Erlon showed that the latter had been directed to move towards the Tagus, and that the King himself was evidently thinking of bringing succour to Marmont, so far as his modest means allowed[454]. But since this projected operation seemed to depend on assistance being granted by Soult, and since it was doubtful in the highest degree whether Soult would give it, Wellington was not without hopes that it might come to nothing. ‘I have requested the Empecinado,’ he writes to Lord Liverpool, ‘to alarm the King for the safety of his situation about Madrid, and I hope that Marshal Soult will find ample employment for his troops in the blockade of Cadiz, the continued operations of General Ballasteros, and those in Estremadura of Lieut.-General Hill, whose attention I have called to the probable march of this corps of the Army of the South through Estremadura.’ As a matter of fact Soult prevented D’Erlon from giving any help to the King or Marmont; but a contingency was to arise of which Wellington, on July 1st, could have no expectation--viz. that, though refused all help from the South, Joseph might come to the desperate but most soldier-like determination to march with his own little army alone to the Douro, in order to bring to bear such influence as he possessed on what was obviously a critical moment in the war. The King and Jourdan were the only men in Spain who showed a true appreciation of the crisis: but they made their move too late: the fault was undoubtedly Soult’s alone. However, on July 1st, Wellington was justified in doubting whether any danger would arise on the side of Madrid. Joseph could not move the Army of the Centre to the Douro, without risking his capital and abandoning all New Castile. As late as July 11th Wellington suspected that he would not make this extreme sacrifice, but would rather push a demonstration down the Tagus to alarm central Portugal, a hypothesis which did not much alarm him[455]. The King and Jourdan knew better than to make this indecisive move, and marched where their 14,000 men might have turned the whole course of the campaign--but marched too late.

[454] See Wellington to Lord Liverpool, June 18. _Dispatches_, ix. p. 241, and June 25, p. 253. There was also in Wellington’s hands an intercepted letter of Joseph to Soult of May 26, distinctly saying that if Marmont is attacked in June, D’Erlon must pass the Tagus and go to his help. This is in the Scovell ciphers.

[455] Wellington to Hill, July 11. _Dispatches_, ix. p. 281. The idea that Joseph might operate on his own account begins to emerge in the correspondence on the 14th. _Dispatches_, ix. p. 283.

There was still a chance that Suchet might be helping the King--this depended entirely on an unknown factor in the game, the diversion which Lord William Bentinck had promised to execute on the coast of Catalonia. If it had begun to work, as it should have done, by the second half of June, there was little chance that any troops from the eastern side of Spain would interfere in the struggle on the Douro. But no information of recent date was yet forthcoming: it was not till July 14th that the vexatious news arrived that Lord William was faltering in his purpose, and thinking of plans for diverting his expeditionary force to Italy.

The situation, therefore, when Marmont went behind the Douro on July 1st, had many uncertain points: there were several dangerous possibilities, but nothing had yet happened to make ultimate success improbable. On the whole the most disappointing factor was the conduct of the Army of Galicia. It will be remembered that Wellington had arranged for a double diversion on Marmont’s flank and rear. Silveira, with the militia of the Tras-os-Montes and D’Urban’s Portuguese cavalry brigade, was to cross the Esla and besiege Zamora. Santocildes, with the Army of Galicia, had been directed to attack Astorga with part of his force, but to bring the main body forward to the Esla and overrun the plains of northern Leon. Silveira had but a trifling force, and the task allotted to him was small: but on July 1st he had not yet reached Zamora with his infantry, and was only at Carvajales on the Esla[456]. On the other hand D’Urban’s cavalry had pushed boldly forward in front of him, had swept the whole north bank of the Douro as far as Toro, and reported that all the French garrisons save Astorga, Zamora, and Toro had been drawn in--that Benavente, Leon, and all the northern plain were unoccupied. On July 2 D’Urban was at Castronuevo, north of Toro, right in the rear of Marmont’s flank--a very useful position, since it enabled him to keep up communication between Silveira and the Galicians, as well as to report any movement of the French right. Moreover, though his force was very small, only 800 sabres, it was enough to prevent any foraging parties from Marmont’s rear from exploiting the resources of the north bank of the Douro. Some such appeared, but were driven in at once, so that the Marshal had to live on his magazines and the villages actually within his lines: in the end these resources would be exhausted, and the old choice--starvation or dispersion--would once more be presented to the Army of Portugal[457].

[456] By no fault of his own, according to D’Urban. The orders for him to move were, by some delay at head-quarters, only forthcoming on June 8th. Only two of the four Tras-os-Montes militia regiments were then mobilized, and it took a long time to collect the rest and the transport needed for moving across the frontier.

[457] D’Urban’s manœuvres on both sides of the Douro are detailed at great length in his very interesting diary, and his official correspondence, both of which have been placed at my disposal. He worked on both sides of the Douro, but went definitely north of it after July 1.

But as a military body neither D’Urban’s 800 horse nor Silveira’s 4,000 militia had any threatening power against Marmont’s rear. They might almost be neglected, while the real pressure which Wellington had intended to apply in this quarter was not forthcoming. He had hoped that, by the time that he and Marmont were at close quarters, the Army of Galicia would have been taking a useful part in the campaign. It was not that he intended to use it as a fighting force: but if it could have appeared in the French rear 15,000 strong, it would have compelled Marmont to make such a large detachment for the purpose of ‘containing’ it, that he would have been left in a marked numerical inferiority on the Douro.

Unfortunately the Galicians moved late, in small numbers, and with marked timidity. They exercised no influence whatever on the course of the campaign, either in June or in July. Yet after Bonnet evacuated the Asturias and went off eastward on June 15th, the Army of Galicia had no field-force of any kind in front of it. The only French left in its neighbourhood were the 1,500 men[458] who formed the garrison of Astorga. Castaños, who had moved up to Santiago in June, and assumed command, did not take the field himself, but handed over the charge of the troops at the front to Santocildes. The latter sat down in front of Astorga with his main body, and only pushed forward a weak division under Cabrera to Benavente, where it was still too remote from Marmont to cause him any disquiet. The siege of Astorga was only a blockade till July 2nd, as no battering-train was brought up till that date. First Abadia, and later Castaños had pleaded that they had no means for a regular siege, and it was not till Sir Howard Douglas pointed out a sufficient store of heavy guns in the arsenal of Corunna, that Castaños began to scrape together the battering-train that ultimately reached Astorga[459]. But this was not so much the weak point in the operations of the Galician army, as the fact that, of 15,000 men brought together on the Orbigo, only 3,800 were pushed forward to the Esla, while the unnecessarily large remainder conducted a leisurely siege of the small garrison of Astorga. Wellington had reckoned on having an appreciable force, 10,000 or 12,000 men, at the front, molesting Marmont’s flank; this would have forced the Marshal to make a large detachment to keep it off. But not a man appeared on the east bank of the Esla, and the operations of D’Urban’s small brigade were of far more service to the main army than that of the whole of the Galicians. Marmont ignored the presence of the few thousand men pushed forward to Benavente, and was justified in so doing. Meanwhile Santocildes, with an optimism that proved wholly unjustifiable, sent messages that Astorga would be taken within a few days, and that he would then move forward with his main body. As a matter of fact the place held out till the 18th of August.

[458] Two battalions of 23rd Léger and one of 1st Line from Thomières’s division.

[459] For the curious story of their ignorance of their own resources see Sir Howard Douglas’s _Life_, pp. 156-7.

Wellington, therefore, was building on a false hypothesis when he wrote to Lord Bathurst, on July 7, that he was surveying all the fords of the Douro, and waiting till the river should have fallen a little and made them more practicable. ‘By that time I hope that the Army of Galicia under General Santocildes will have been able to advance, the siege of Astorga having been brought to a conclusion[460].’ Two days later he added, ‘it would not answer to cross the river at all in its present state, unless we should be certain of having the co-operation of the Galician troops[461].’ His delay in making an attempt to force the line of the Douro, therefore, may be attributed in the main to the tiresome conduct of Santocildes, who played to him much the same part that Caffarelli played to Marmont.

[460] _Dispatches_, ix. p. 274.

[461] Ibid., ix. p. 276.

While remaining in this waiting posture, Wellington placed his troops opposite the various passages of the Douro, on a line of some fifteen miles. His left, consisting of the 3rd Division, Pack’s and Bradford’s Portuguese, and Carlos de España’s Spaniards, with Le Marchant’s and Bock’s heavy dragoons, lay near the point where the Trabancos falls into the Douro, holding the ford of Pollos, where the favourable configuration of the ground enabled them to be sure of the passage, the enemy’s line being perforce drawn back to some distance on the north bank. It was always open to Wellington to use this ford, when he should determine on a general advance. The Light, 4th, 5th, and 6th Divisions, forming the right wing, lay opposite Tordesillas, with Rueda and La Seca behind them. Their front was covered by Alten’s cavalry brigade, their right (or outer) flank by Anson’s. The reserve was formed by the 1st and 7th Divisions quartered at Medina del Campo, ten miles to the rear. The whole could be assembled for an offensive or a defensive move in a day’s march.

Marmont was drawn up, to face the attack that he expected, in an almost equally close and concentrated formation: his front, extending from the junction of the Pisuerga with the Douro near Simancas on his left, to the ground opposite the ford of Pollos on his right, was very thickly held[462]; but on the 5th he rightly conceived doubts as to whether it would not be easy for Wellington to turn his western flank, by using the ford of Castro Nuño and other passages down-stream from Pollos. He then detached Foy’s division to Toro and the neighbourhood, to guard against such a danger: but this was still an insufficient provision, since Toro is fifteen miles from Pollos, and a single division of 5,000 men would have to watch rather than defend such a length of river-line, if it were attacked in force. Therefore when Bonnet, so long expected in vain, arrived from the North on July 7th, Marmont placed him in this portion of his line, for the assistance of Foy. He still retained six divisions massed around Tordesillas, whose unbroken bridge gave him a secure access to the southern bank of the Douro. With this mass of 35,000 men in hand, he could meet Wellington with a solid body, if the latter crossed the Douro at or below Pollos. Or he might equally well take the more daring step of assuming a counter-offensive, and marching from Tordesillas on Salamanca against his adversary’s communications, if the allies threatened his own by passing the river and moving on Valladolid.

[462] An interesting dispatch from D’Urban to Beresford describes the information he had got on the 5th by a daring reconnaissance along Marmont’s rear: there was not that morning any French force west of Monte de Cubillos, six miles down-stream from Pollos.

A word to explain the tardiness of Bonnet’s arrival in comparison with the earliness of his start is perhaps required. He had evacuated Oviedo and Gijon and his other posts in the Asturias as early as June 14th, the actual day on which Wellington commenced his offensive campaign. This he did not in consequence of Marmont’s orders, which only reached him when he had begun to move, but on his own responsibility. He had received correct information as to the massing of the allied army round Ciudad Rodrigo, and of the forward movement of the Galicians towards Astorga. He knew of the dispersed state of Marmont’s host, and saw the danger to himself. Should the Marshal concentrate about Salamanca, he could never join him, if the whole Army of Galicia threw itself between. Wherefore not only did he resolve to retreat at once, but he did not move by the pass of Pajares and Leon--the obvious route to rejoin the Army of Portugal. For fear that he might be intercepted, he took the coast-road, picking up the small garrisons that he had placed in one or two small ports. He reached Santander on the 22nd, not molested so much as he might have been by the bands of Porlier and Longa (whose haunts he was passing), because the bulk of them had gone off to help in Sir Home Popham’s raid on Biscay. From Santander he turned inland, passed Reynosa, in the heart of the Cantabrian Sierras, on the 24th June, and arrived at Aguilar del Campo, the first town in the province of Palencia, on the 29th. From thence he had a long march of seven days in the plains, before he reached Valladolid on the 6th, and reported himself at Marmont’s head-quarters on the 7th of July. He brought with him a strong division of 6,500 infantry, a light field-battery, and a single squadron of Chasseurs--even 100 sabres[463] were a welcome reinforcement to Marmont’s under-horsed army. It was an odd fact that Bonnet’s division had never before met the English in battle, though one of its regiments had seen them during the last days of Sir John Moore’s retreat in January 1809[464]. For the three years since that date they had always been employed in the Asturias.

[463] Ninety-four to be exact. See 28th Chasseurs in table of Marmont’s army in Appendix.

[464] The 122nd Line had been in Mermet’s division, in January 1809, but they had been in reserve at Corunna, and had not fired a shot in that battle.

The arrival of Bonnet brought up the total of Marmont’s infantry to 43,000 men, and his guns to 78. The cavalry still remained the weak point: but by a high-handed and unpopular measure the Marshal succeeded, during his stay on the Douro, in procuring nearly 1,000 horses for the dismounted dragoons who were encumbering his dépôt at Valladolid. In the French, as in the British, Peninsular army it had become common for many of the junior officers of the infantry to provide themselves with a riding-horse; most captains and many lieutenants had them. And their seniors, _chefs de bataillon_ and colonels, habitually had several horses more than they were entitled to. Marmont took the heroic measure of proclaiming that he should enforce the regulations, and that all unauthorized horses were confiscated. He paid, however, a valuation for each beast on a moderate scale--otherwise the act would have been intolerable. In this way, including some mounts requisitioned from doctors, commissaries, and suttlers, about 1,000 horses in all were procured. The number of cavalry fit for the field had gone up by July 15th from about 2,200 to 3,200--a total which was only 300 less than Wellington’s full strength of British sabres. It occurs to the casual observer that the horses, having never been trained to squadron drill or to act in mass, must have been difficult to manage, even though the riders were competent horsemen. This may have something to do with the very ineffective part played by the French cavalry in the next fortnight’s campaigning.

A quaint anecdote of the time shows us General Taupin, an old Revolutionary veteran, with all the officers of his brigade called together in a village church. ‘He ascended the pulpit and thundered against the abuse of horses in the infantry: he would make an end of all baggage carried on mules or asses, but most especially of the officers’ riding-horses. “Gentlemen,” he cried, “in 1793 we were allowed a haversack as our only baggage, a stone as our only pillow.” Well--it was a long time since 1793: we were in 1812, and the speaker, this old and gallant soldier, had _six_ baggage mules himself[465].’

[465] _Mémoires_ of Lemonnier-Delafosse of the 31st Léger, pp. 177-8.

During the first ten days after the deadlock on the Douro began, the French were much puzzled by Wellington’s refusal to continue his advance. Foy, the ablest of them, noted in his diary that he must conclude either that the enemy was not numerous enough to take the offensive--his strength might have been over-valued--or else that he was waiting for Hill to bring up his corps from Estremadura. This last idea, indeed, was running in the brains of many French strategists: it obsessed Jourdan and King Joseph at Madrid, who were well aware that Hill, marching by Alcantara and the passes of the Sierra de Gata, could have got to the Douro in half the time that it would have taken his opponent, D’Erlon, who would have had to move by Toledo, Madrid, and Segovia. But the simple explanation is to be found in Wellington’s dispatch to Lord Bathurst of July 13. ‘It is obvious that we could not cross the Douro without sustaining great loss, and could not fight a general action under circumstances of greater disadvantage.... The enemy’s numbers are equal, if not superior, to ours: they have in their position thrice the amount of artillery that we have, and we are superior in cavalry alone--which arm (it is probable) could not be used in the sort of attack we should have to make[466].’ He then proceeds to demonstrate the absolute necessity of bringing forward the Army of Galicia against Marmont’s rear. Its absence was the real cause of the deadlock in which he found himself involved. All offensive operations were postponed--meanwhile the enemy might receive reinforcements and attack, since he had not been attacked. ‘But I still hope that I shall be able to retain, at the close of this campaign, those acquisitions which we made at its commencement.’

[466] Wellington to Bathurst. _Dispatches_, ix. p. 284.

Meanwhile Marmont, having had a fortnight to take stock of his position, and having received reinforcements which very nearly reached the figure that he had named to King Joseph as the minimum which would enable him to take the offensive, was beginning to get restless. He had now realized that he would get no practical assistance from Caffarelli, who still kept sending him letters exaggerating the terrors of Sir Home Popham’s raid on Biscay. They said that there were six ships of the line engaged in it, and that there was a landing-force of British regulars: Bonnet’s evacuation of the Asturias had allowed all the bands of Cantabria to turn themselves loose on Biscay--Bilbao was being attacked--and so forth. This being so, it was only possible to send a brigade of cavalry and a horse artillery battery--anything more was useless to ask[467]. This was written on June 26th, but by July 11th not even the cavalry brigade had started from Vittoria, as was explained by a subsequent letter, which only reached Marmont after he had already started on an offensive campaign[468]. As a matter of fact, Caffarelli’s meagre contribution of 750 sabres[469] and one battery actually got off on July 16th[470]. Marmont may be pardoned for having believed that it would never start at all, when it is remembered that a month had elapsed since he first asked for aid, and that every two days he had been receiving dispatches of excuse, but no reinforcements. He had no adequate reason for thinking that even the trifling force which did in the end start out would ever arrive.

[467] Caffarelli to Marmont, in the latter’s _Mémoires_, iv. p. 417.

[468] Ibid., pp. 421-2.

[469] He sent finally only two regiments, not three as he had originally promised.

[470] Caffarelli to Marmont, in the latter’s _Mémoires_, iv. p. 425, announcing their departure.

Nor, as he demonstrates clearly enough in his defence of his operations, had he any more ground for believing that Joseph and Jourdan would bring him help from Madrid. They resolved to do so in the end, and made a vigorous effort to collect as large a force as was possible. But the announcement of their intention was made too late to profit Marmont. The dispatch conveying it was sent off from Madrid only on July 9th[471], and never reached the Marshal at all, for the two copies of it, sent by separate messengers, were both captured by guerrilleros between Madrid and Valladolid, and came into Wellington’s instead of into Marmont’s hands. This was a consequence of the insecurity of the communication via Segovia, the only one route open when the Army of Portugal retired behind the Douro. On July 12th the last piece of intelligence from Madrid which Marmont had received was a dispatch from Jourdan dated June 30th--it had taken twelve days to get 150 miles, which shows the shifts to which its bearer had been exposed. This letter is so important, as showing what the King and Jourdan opined at the moment, that its gist is worth giving.

[471] Original is in the Scovell ciphers. It seems to be unpublished.

Jourdan begins by complaining that on June 30 the last dispatch from the Army of Portugal to hand was sixteen days old, of the date of June 14th. It is clear, then, that no copies of the reports sent by Marmont on June 22 and June 24 had got to Madrid--a circumstance to be explained by the fact that Wellington had them instead of their destined recipient[472]. Jourdan then proceeds to say that he is informed that Wellington has 50,000 men, but only 18,000 of them British. ‘The King thinks that if this is so, you are strong enough to beat his army, and would like to know the motives which have prevented you from taking the offensive. He charges me to invite you to explain them by express messenger.’ In the South it was known that Hill, with 18,000 men, was advancing on June 18th against D’Erlon. That officer was to be reinforced from Seville, and was probably at close quarters with Hill. The King had sent orders that D’Erlon was to move northward into the valley of the Tagus, if Hill marched up to join Wellington. But, it being probable that the order would not be very promptly executed, ‘his Majesty would like you to take advantage of the moment, when Wellington has not all his forces in hand, to fight him. The King has asked for troops from Marshal Suchet, but they will never be sent. All that His Majesty can do at present is to reinforce the garrison of Segovia, and order its governor, General Espert, to help the garrison of Avila, if necessary, and to supply it with food.’

[472] They are both in the Scovell ciphers, and quoted above, p. 370.

This letter, which clearly gives no hope of immediate help for the Army of Portugal from Madrid, and which might be taken as a direct incitement to bring Wellington to action at once, must be read in conjunction with the last epistle that Marmont had received from the same quarter. This was a letter of the King’s dated June 18. The important paragraph of it runs as follows:--

‘If General Hill has remained with his 18,000 men on the left (south) bank of the Tagus, you ought to be strong enough to beat the English army, more especially if you have received any reinforcements from the Army of the North. You must choose your battlefield, and make your best dispositions. But if Hill joins the main English army, I fancy they are too strong for you. In that case you must manœuvre to gain time. I should not hesitate to give you a positive order to defer fighting, if I were certain that Count D’Erlon and his 15,000 men, and a division from the Army of Aragon, were on their way to you: for on their arrival the English army would be seriously compromised. But being wholly uncertain about them, I must repeat to you that if General Hill is still on the south side of the Tagus, you should choose a good position and give battle with all your troops united: but if General Hill joins Lord Wellington, you must avoid an action as long as possible, in order to pick up the reinforcements which will certainly reach you in the end[473].’

[473] Joseph to Marmont, June 18, in Ducasse’s _Correspondance_, ix. pp. 28-39.

I think that there can be no doubt in the mind of any honest critic that on the strength of these two dispatches from his Commander-in-Chief, Marmont was justified in taking the offensive against Wellington, without waiting for that help from Madrid which the King had not offered him. Hill being far away, and Wellington having no more than his own seven divisions of Anglo-Portuguese, Marmont is decidedly authorized to bring him to action. The sole factor which the second Madrid dispatch states wrongly, is the proportion of British troops in the allied army: Jourdan guesses that there are 50,000 men, but only 18,000 British. As a matter of fact there were 49,000 men at the moment[474], but about 30,000 were British. This made a difference, no doubt, and Marmont, if he had been determined to avoid a battle, might have pleaded it as his justification. But he was not set on any such timid policy: he had wellnigh attacked Wellington at San Cristobal on June 21st, when he had not yet received his own reinforcements. When Bonnet had come up, and the British had obtained no corresponding addition to their strength, he was eager to take the offensive, and Joseph’s and Jourdan’s dispatches distinctly authorized him to do so.

[474] Two battalions, the 1/38 and 1/5th, joined before the battle of the 22nd, bringing up the total force by 1,500 bayonets more.

After the disaster of Salamanca, Napoleon drew up an indictment of Marmont, of which the three chief heads were:

(1) He took the offensive without waiting for reinforcements which were to join him.

(2) He delivered battle without the authorization of his Commander-in-Chief.

(3) He might, by waiting only two days longer, before he committed himself to a general action, have received at least the cavalry and guns which he knew that Caffarelli had sent him[475].

[475] See the letter of Clarke to Marmont enclosing the Emperor’s indictment, in Marmont’s _Mémoires_, iv. pp. 453-4.

The very complete answer to these charges is that:

(1) When the Marshal took the offensive he had no reason to suppose that any reinforcements were coming. Caffarelli had excused himself: the King had promised succour only if Hill joined Wellington, not otherwise. Hill had never appeared: therefore no help was likely to come from the southward.

(2) He had clear permission from Joseph to give battle, unless Hill should have joined Wellington.

(3) The succours from Caffarelli, a weak cavalry brigade and one battery, were so small that their arrival would have made no practical difference to the strength of the army. But to have waited two days for them, after the campaign had commenced, would have given Wellington the opportunity of concentrating, and taking up a good position. It was only after the manœuvring had begun [July 15th] that this little brigade started from Vittoria, on July 16th. The Army of Portugal had already committed itself to offensive operations, and could not halt for two days in the midst of them, without losing the initiative.

From his own point of view, then, Marmont was entirely justified in recrossing the Douro and assuming the offensive. He had got all the reinforcements that he could count upon: they made his army practically equal to Wellington’s in numbers: in homogeneity it was far superior. If he had waited a little longer, he might have found 12,000 men of the Army of Galicia at his back, setting all Old Castile and Leon aflame. Moreover Astorga was only victualled up to August 1st, and might fall any day. He could not have foreseen King Joseph’s unexpected march to his aid, which no dispatch received before July 12th rendered likely. His misfortune (or fault) was that he undervalued the capacity of Wellington to manœuvre, his readiness to force on an offensive battle, and (most of all) the fighting value of the Anglo-Portuguese army.

It cannot be denied that Marmont’s method of taking the offensive against Wellington was neat and effective. It consisted in a feint against his adversary’s left wing, followed by a sudden countermarch and a real attack upon his right wing.

On July 15th Foy and Bonnet, with the two divisions forming the French right, received orders to restore the bridge of Toro, to drive in Wellington’s cavalry screen in front of it, and to cross to the south bank of the Douro. At the same time the divisions of the French centre, opposite the fords of Pollos, made an ostentatious move down-stream towards Toro, accompanied by the Marshal himself, and those on the left, near Tordesillas, shifted themselves towards Pollos. Almost the whole French army was clearly seen marching westward, and the two leading divisions were actually across the river next morning, and seemed to be heading straight for Salamanca by the Toro road.

Wellington was deceived, exactly as Marmont had intended. He drew the obvious conclusion that his adversary was about to turn his left flank, and to strike at Salamanca and his line of communications. It would have been in his power to make a corresponding move against Valladolid, Marmont’s base. But his own line of communications meant much more to him than did Marmont’s. There was a great difference between the position of an army living by transport and magazines, and that of an army living on the country by plunder, like that of the French marshal. Wellington had always been jealous of his left wing, and as early as July 12 had drawn up an elaborate order of march, providing for the contingency of the enemy crossing the Douro at Toro and the ford of Castro Nuño. If his entire force seemed on the move, the whole British army would make a corresponding shift westward--if only a division or two, the mass transferred would be less in similar proportion. He had no idea of defending the actual course of the river: in a letter written a few days later to Lord Bathurst, he remarked that ‘it was totally out of my power to prevent the enemy from crossing the Douro at any point at which he might think it expedient, as he had in his possession all the bridges [Toro and Tordesillas] and many of the fords[476].’ His plan was to concentrate against the crossing force, and fight a defensive action against it, wherever a good position might be available.

[476] See _Supplementary Dispatches_, xiv. p. 68.

There were two reasons for which Wellington regarded a genuine offensive move of Marmont by Toro and Castro Nuño as probable. The first was that he had received King Joseph’s dispatch of July 9th, captured by guerrilleros, which gave him the startling news that the King had resolved to evacuate all New Castile save Madrid and Toledo, and to march with his field-force of some 14,000 men to join the Army of Portugal[477]. Wellington wrote to Graham (who was now on his way home) early on the 16th, that either the Galicians’ approach on his rear had induced Marmont to collect his troops near Toro, or he had heard that Joseph was gathering the Army of the Centre at Madrid, and was threatening the allied left ‘in order to prevent us from molesting the King.’ It was clear that if Wellington had to shift westward to protect his line of communications, he could make no detachment to ‘contain’ King Joseph, who would be approaching from the south-east. Another letter, written an hour or so later, says, ‘these movements of Marmont are certainly intended to divert our attention from the Army of the Centre (which is collecting at Madrid), if he knows of this circumstance, _which I doubt_[478].’ The doubt was well grounded.

[477] See _Dispatches_, ix. p. 294.

[478] Wellington to Clinton, July 16, 7 a.m. _Dispatches_, ix. p. 291.

That the whole movement on Toro was a feint did not occur to Wellington, but his orders of the 16th, given in the evening, after he had heard that two French divisions were actually across the Douro on his left, provide for the possibility that some serious force may still remain at Tordesillas and may require observation.

The orders direct the transference of the great bulk of the allied army to a position which will cover the road Toro-Salamanca. They were issued in the evening to the following effect. The reserve (1st and 7th Divisions) was to march from Medina del Campo to Alaejos beyond the Trabancos river, and subsequently to Canizal and Fuente la Peña behind the Guarena river. The left wing, which was watching the fords of Pollos (3rd Division, Bock’s cavalry, Bradford’s and Carlos de España’s infantry), to Castrillo on the Guarena. Of the right wing the 6th Division and two regiments of Le Marchant’s horse were to move on Fuente la Peña, the 5th Division on Canizal. Alten’s cavalry brigade was to follow the 1st Division. This left the 4th and Light Divisions and Anson’s cavalry still unaccounted for. They were set aside to act as a sort of rearguard, being directed to move westward only as far as Castrejon on the Trabancos river, ten miles short of the concentration-point on the Toro road, to which the rest of the army was ordered to proceed. It is clear (though Wellington does not say so) that they would serve as a containing force, if the enemy had left any troops at Tordesillas, and brought them over the Douro there, or at the fords of Pollos.

All these moves were duly executed, and on the morning of the 17th Wellington’s army was getting into position to withstand the expected advance of the enemy on Salamanca by the Toro road. This attack, however, failed to make itself felt, and presently news came that the two divisions of Foy and Bonnet, which had crossed the Douro at Toro, had gone behind it again, and destroyed their bridge. What Marmont had done during the night of the 16th-17th was to reverse the marching order of his whole army, the rear suddenly becoming the head, and the head the rear. The divisions to the eastward, which had not yet got near Toro, countermarched on Tordesillas, and crossed its bridge, with the light cavalry at their head. Those which had reached Toro brought up the rear, and followed, with Foy and Bonnet, at the tail of the column. This was a most fatiguing march for all concerned, the distance from Toro to Tordesillas being about twenty miles, and the operation being carried out in the night hours. But it was completely successful--during the morning of the 17th the vanguard, consisting of Clausel’s and Maucune’s divisions and Curto’s _chasseurs à cheval_, was pouring over the bridge of Tordesillas and occupying Rueda and La Seca, which the British had evacuated fifteen hours before. The rest followed, the two rear divisions cutting a corner, and saving a few miles, by crossing the ford of Pollos. This was a safe move, when the cavalry had discovered that there were none of Wellington’s troops left east of the Trabancos river. By night on the 17th the bulk of the French army was concentrated at Nava del Rey, ten miles south-west of Tordesillas. In the afternoon Wellington’s rearguard, the 4th and Light Divisions, and Anson’s cavalry had been discovered in position at Castrejon, where their commander had halted them, when he discovered that he had been deceived as to his adversary’s purpose. The rest of the British army had concentrated, according to orders, in the triangle Canizal-Castrillo-Fuente la Peña, behind the Guarena river and in front of the Toro-Salamanca road.

[Illustration: THE SALAMANCA CAMPAIGN]

Wellington’s first task was to drawback his rearguard to join his main body, without allowing it to become seriously engaged with the great mass of French in its front. This he undertook in person, marching at daylight with all his disposable cavalry, the brigades of Bock and Le Marchant, to join the force at Castrejon, while he threw out the 5th Division to Torrecilla de la Orden to act as a supporting échelon on the flank of the retiring detachment. The remaining divisions (1st, 3rd, 6th, 7th) took up a position in line of battle on the heights above the Guarena, ready to receive their comrades when they should appear.

The charge of the rearguard this day was in the hands of Stapleton Cotton, the senior cavalry officer with the army, who outranked Cole and Charles Alten, the commanders of the 4th and Light Divisions. He had received no orders during the night, and his last, those of the preceding afternoon, had directed him to halt, till his chief should have discovered the true position and aim of the French army. Wellington explained, in his next dispatch home, that the various details of intelligence, which enabled him to grasp Marmont’s whole plan, did not reach him till so late on the 17th that it was useless to send Cotton orders to start. They could only be carried out at dawn, and he himself intended to be present with the rearguard before the sun was far above the horizon. He arrived at seven o’clock in the morning, in time to find his lieutenant already engaged with the French van, but not committed to any dangerous close fighting. Cotton had, very wisely, sent out patrols before daylight to discover exactly what was in front of him; if it was only a trifling body he intended to drive it in, and advance towards La Nava and Rueda[479]; if Marmont was in force he would take up a defensive position at Castrejon, and wait for further orders.

[479] See report of one of the officers commanding patrols, Tomkinson of the 16th L.D. in the latter’s _Memoirs_, p. 180.

The patrols soon ran into French cavalry advancing in force, and were driven back upon Anson’s brigade, which was drawn up on a long front in advance of the village of Castrejon. On seeing it, the enemy brought up two batteries of horse artillery, and began to play upon the scattered squadrons. Bull’s and Ross’s troops[480] were ordered out to reply, and did so with effect, but the total strength of the French cavalry was too great, and Anson’s regiments had presently to give way, though not so much owing to the pressure on their front as to the sight of a large column of French infantry turning the left of their line, and marching on Alaejos, with the obvious intention of getting round to their left rear and molesting their retreat towards the Guarena, where the main body of the British army was awaiting them.

[480] Belonging one to the cavalry, the other to the Light Division.

Wellington was involved in person in the end of the cavalry bickering, and in no very pleasant fashion. He and Beresford, with their staffs, had arrived on the field about seven o’clock, in advance of the two heavy cavalry brigades, who were coming up to reinforce Cotton. He rode forward to the left of the skirmishing line, where two squadrons, one of the 11th and one of the 12th Light Dragoons, were supporting two guns of Ross’s troop, on high ground above the ravine of the Trabancos river. Just as the Commander-in-Chief came on the scene, a squadron of French cavalry, striking in from the flank, rode at the guns, not apparently seeing the supporting troops. They met and broke the squadron of the 12th Light Dragoons, which came up the hill to intercept them. ‘Some of Marshal Beresford’s staff, seeing this, and conceiving the guns to be in danger, rode up to the retiring squadron calling “Threes about[481]!”’ This unfortunately was heard by the supporting squadron of the 11th, who, imagining the order to be directed to themselves, went about and retired, instead of advancing to relieve their broken comrades above. Therefore the mass of pursuers and pursued from the combat on the flank, came hurtling down on the guns, and on the head-quarters staff just behind them. Wellington and Beresford and their followers were swept away in the rout, and had to draw their swords to defend themselves. Fortunately the misdirected squadron of the 11th soon saw their mistake; they halted and turned, and falling on the scattered and exhausted French dragoons drove them back with great loss; few, it is said, except their _chef d’escadron_, who showed uncommon gallantry, got away[482]. It was a dangerous moment for the allied army--a chance thrust in the _mêlée_ might have killed or disabled Wellington, and have thrown the command into the hands of Beresford or Stapleton Cotton.

[481] Tomkinson, p. 188.

[482] Compare Tomkinson’s narrative of this incident (pp. 180-1) with Napier’s vivid and well-told tale (iv. pp. 254-5). Both agree that the French were inferior in numbers to the two squadrons, and that there was deplorable confusion.

Wellington had no sooner detected the flank movement of Marmont’s infantry towards Alaejos, than he ordered the 4th and Light Divisions to retire towards the Guarena, covered by G. Anson’s brigade, while Bock’s and Le Marchant’s heavy dragoons, farther to the left, drew up in front of the infantry of the turning column, and detained it, retiring, when pressed, by alternate brigades. Marmont’s whole army was now visible, moving on in two long columns, of which the more southern followed the 4th and Light Divisions, in the direction of Torrecilla de la Orden, and tried to come up with their rear, while the other, passing through Alaejos, made by the high-road for Castrillo on the Guarena, where the British reserves were posted.

There was a long bickering fight across the eight miles of rolling ground between the Trabancos and the Guarena, not without some exciting moments for Wellington’s rearguard. After passing Torrecilla de la Orden, and picking up there the 5th Division, which had been waiting as a supporting échelon to cover their southern flank, all the British infantry had to march very hard, for troops diverging from the northern French column got close in upon their right, and, moving parallel with them, bid fair to reach the Guarena first. In the retreat the 4th Division moved on the right, and was therefore most exposed, the Light Division next them, the 5th Division farther south and more distant from the turning column of the French. The cavalry pursuit in the rear of the retreating force was never really dangerous: it was held off by Le Marchant’s Heavy and Anson’s Light Dragoons without any great difficulty, and the 5th and Light Divisions only suffered from some distant shelling by the French horse artillery. But the 4th Division, though covered from the pursuit in their direct rear by Bock’s German squadrons, found a dangerous point about a mile on the near side of the Guarena, where two batteries from the French turning column had galloped forward to a knoll, commanding the ground over which they had to pass, and opened a teasing fire upon the flank of the brigades as they marched by. General Cole, however, threw out his divisional battery and all his light companies to form a screen against their attack, and moved on, protected by their fire, without turning from his route. The covering force fell in to the rear when the defiling was over, and the division suffered small loss from its uncomfortable march[483].

[483] See Vere’s _Marches and Movements of the 4th Division_, p. 28. Napier’s statement that the Light Division was more exposed than the 4th or 5th during the retreat, seems to be discounted by the fact that it had not one man killed or wounded--the 5th Division had only two (in the 3rd Royal Scots), the 4th Division over 200; and though most of them fell in the last charge, a good number were hit in the retreat.

Wellington allowed all the three retreating divisions to halt for a moment on the farther side of the stream, at the bottom of the trough in which it runs. ‘The halt near the water, short as it was, gave refreshment and rest to the troops, after a rapid march over an arid country in extremely hot weather[484].’ But it could not be allowed to last for more than a very few minutes, for the pursuing enemy soon appeared in force at several points on the heights above the eastern bank of the Guarena, and many batteries opened successively on the three divisions, who were of necessity compelled to resume their march up the slope to the crest, on their own side of the water. Here they fell into position on Wellington’s chosen defensive fighting-ground, the 4th Division forming the extreme northern section of the battle array, by the village of Castrillo, the Light and 5th Divisions falling in to the line of troops already drawn up in front of Canizal, while the 1st and 7th Divisions were extended to the south, to form the new right wing, and took their place on the heights of Vallesa, above the village and ford of El Olmo.

[484] Vere’s _Marches and Movements of the 4th Division_, p. 28.

Some anxious hours had been spent while the retreat was in progress, but Wellington was now safe, with every man concentrated on an excellent position, where he was prepared to accept the defensive battle for which he had been waiting for the last month. It seemed likely at first that his wish might be granted, for the French made a vigorous attack upon his left wing, almost before it had got settled down into its appointed ground. It would appear that General Clausel, who commanded the more northerly of the two great columns in which the French army was advancing (while Marmont himself was with the other), thought that he saw his chance of carrying the heights above Castrillo and turning the allied left, if he attacked at once, before the 4th Division had been granted time to array itself at leisure. Accordingly, without wasting time by sending to ask permission from his chief, he directed a brigade of dragoons to outflank Cole’s left by crossing the Guarena down-stream, while Brennier’s division passed it at Castrillo and assailed the front of the 4th Division. Clausel’s own division advanced in support of Brennier’s.

This move brought on very sharp fighting: the turning movement of the French dragoons was promptly met by Victor Alten’s brigade [14th Light Dragoons, 1st Hussars K.G.L.], whose squadrons had been watching the lower fords of the Guarena all day. Alten allowed the hostile cavalry to cross the river and come up the slope, and then charged suddenly, in échelon of squadrons, the left squadron of the 1st Hussars K.G.L. leading[485]. The enemy had only begun to deploy when he was attacked, Alten’s advance having been too rapid for him. The two French regiments (15th and 25th Dragoons) were, after a stiff fight, completely routed and driven downhill with great loss, till they finally found refuge behind a half-battery and an infantry battalion which formed their supports. General Carrié, commanding the two regiments, was taken prisoner by a German hussar, having got cut off from his men in the flight. The French lost in all 8 officers and more than 150 men, of whom 94 were prisoners--mostly wounded. How sharp the clash was may be seen from the fact that Alten’s victorious brigade had not much fewer casualties--the 14th Light Dragoons lost 75 killed and wounded, the German hussars 60[486]. But no doubt some of these losses were suffered not in the cavalry combat, but a little later in the day, when Alten charged the French infantry[487].

[485] Brotherton of the 14th L.D. says with the _right_ échelon advanced (Hamilton’s _History of the 14th_, p. 107), but I fancy that the German Hussars’ version that the _left_ échelon led is correct, as the right squadron of their regiment would have been in the middle of the brigade, not on a flank. See narrative in Schwertfeger, i. pp. 368-9.

[486] These are the official returns. The regimental histories give only 45 and 56 respectively.

Martinien’s lists show six casualties in officers in the two French regiments, and two more were taken prisoners, General Carrié and a lieutenant of the 25th Dragoons.

[487] Brotherton says that the first two squadrons which charged the French dragoons made no impression, and that it was the impact of the third, led by himself, which broke them.

While this lively fight was in progress on the flank, Brennier’s division had crossed the Guarena in a mass, and on a very short front, apparently in three columns of regiments, battalion behind battalion. They were ascending the lower slopes below Cole’s position, when Wellington, who was present here in person, suddenly took the offensive against them, sending W. Anson’s brigade (3/27th and 1/40th) against them in line, with Stubbs’s Portuguese (11th and 23rd regiments) supporting, in columns of quarter distance. The French division halted, apparently with the intention of deploying--but there was no time for this. The line of Anson’s brigade enveloped both the hostile flanks with its superior frontage, and opened fire: after a short resistance the French gave way in great disorder, and streamed down to the Guarena. As they fled Alten let loose part of his brigade against their flank: the horsemen rode in deep among the fugitives, and cut off 6 officers and 240 men as prisoners. Clausel had to bring up a regiment of his own division to cover the broken troops as they repassed the river; it suffered severely from Cole’s artillery, losing 6 officers killed and wounded, and many men[488].

[488] This was the 25th Léger.

The attempt to take liberties with Wellington’s army, when it had assumed the defensive on favourable ground, had thus failed in the most lamentable style, and with very heavy loss--at least 700 men had been killed, wounded, or taken in Marmont’s army that day, and all but a few scores belonged to the four infantry and two cavalry regiments which Clausel sent to attack the heights by Castrillo[489]. The corresponding British loss that day was 525, including about 50 stragglers taken prisoners during the retreat from the Trabancos to the Guarena, because they had fallen behind their regiments--foot-sore infantry, or troopers whose horses had been shot. The cavalry, which had so successfully covered the long march across the open, had a certain amount of casualties, but the only units that had suffered heavily were the four regiments--horse and foot--that dealt with Clausel’s attack, who lost 276 men between them.

[489] The exact figures, save for officers, are as usual missing. But Martinien’s invaluable lists show that of 41 French officers killed, wounded, or taken that day, 35 belonged to the four infantry regiments (17th and 25th Léger, 22nd and 65th Line) and the two cavalry regiments (15th and 25th Dragoons) which fought at Castrillo.

Wellington must have felt much disappointment at seeing Clausel’s offensive move at Castrillo unsupported by the rest of the French divisions, who were lining the farther bank of the Guarena parallel with the whole of his front. But Marmont, unlike his venturesome subordinate, nourished no illusions about the advisability of attacking a British army in position. He made no move in the afternoon; in his memoirs he points out that the infantry was absolutely exhausted, having been continuously on the march for three days and one night.

This day had been a disappointing one for the French marshal also. He had failed to cut off Wellington’s two detached divisions, so that all the advantage which he had obtained by his marches and countermarches between Toro and Tordesillas was now exhausted. The allied army had succeeded in concentrating, and was now drawn up in his front, covering Salamanca and its own line of communications in a very tenable position. Napier truly remarks that, since the attempt to isolate and destroy Cotton’s detachment had miscarried, Marmont had gained no more by his elaborate feint and forced marches than he would have obtained by continuing his original advance across Toro bridge on the 16th. He had got the whole Anglo-Portuguese army arrayed in a defensive position in front of him, on the line of the Guarena, instead of somewhere in the neighbourhood of Fuente Sauco, a few miles farther east.

On the morning of the 19th July it seemed as if a new deadlock was to bring the campaign to a standstill, for the two armies continued to face each other across the Guarena, Wellington hoping rather than expecting to be attacked, Marmont looking in vain for a weak point between Castrillo and Vallesa, where it would be worth while to try a forward thrust. While he was reconnoitring, his weary infantry got a much-needed rest. At about four o’clock in the afternoon, however, the whole French army was seen falling into column, and presently edged off southward till it lay between Tarazona and Cantalapiedra. Wellington thereupon made a corresponding movement, evacuating Castrillo to the north, and extending his line of battle beyond Vallesa to the south. There was a little distant cannonading across the valley of the Guarena, and some of the shells set fire to the vast fields of ripe wheat which covered the whole country-side in this region. The conflagration went rolling on for a long way across the plain, leaving a trail of smoke behind.

The situation on this evening had nothing decisive about it. It was clear that neither side intended to fight save at an advantage. Marmont had shown himself more cautious than had been expected. Wellington had at this moment every motive for risking nothing, unless the enemy proved more obliging than he had shown himself hitherto. He had reasons for self-restraint at this moment of which his adversary knew nothing. The first was that he was aware (from intercepted dispatches) of King Joseph’s intention to march from Madrid to join the Army of Portugal: with a possible 15,000 men about to appear on his flank, he must look to the future with care. The second was that he had received a few days before the untoward news that Lord William Bentinck’s long-promised expedition to Catalonia might not ever take place. The Commander-in-Chief in Sicily wrote that he had found new opportunities in Italy, which it might be his duty to seize. His troops had been embarked, but they were not to be expected for the present off the coast of Spain. This was a disheartening piece of intelligence: Wellington had been told to count upon this support both by Bentinck himself and by the Home Government. If it should fail, Marshal Suchet, left undisturbed by this diversion, might send considerable reinforcements to Madrid[490].

[490] For dismay expressed by Wellington at this news see dispatches to Henry Wellesley dated Rueda, July 15, and to Lord Bathurst (_Dispatches_, ix. pp. 285 and 287).

As a matter of fact he did not--being, like Soult, a general of much too self-centred a type of mind to help a neighbour if he could avoid it. Only one regiment of the Valencian army ever got to Madrid, and that came too late for King Joseph’s purpose. But so far as Wellington could guess on July 19, it was quite possible that Suchet might find 10,000 men, to add to the disposable 15,000 of the Army of the Centre.

There was also the possibility that D’Erlon, obeying the orders which King Joseph kept sending to him, might make up his mind to cross the Guadiana and Tagus, and come north by Arzobispo and Madrid. If so, Hill was to make a parallel march by Alcantara, and would certainly arrive many days before D’Erlon. This was a mere possibility; there were good reasons for holding that Soult might forbid any such move; and till D’Erlon started northward, Hill must remain behind to contain him. The problem was not pressing: it could not develop for many days[491].

[491] See Wellington to Hill, _Dispatches_, ix. p. 290.

On the other hand there was news that the Galicians were at last on the move. Santocildes had been prevailed upon to leave a smaller force to besiege Astorga, and had come down with a second division to join Cabrera at Benavente. This force, advancing up the Douro valley, would find absolutely no enemy in front of it, and must obviously disturb Marmont’s operations, since it might be at the gates of Valladolid, his base and storehouse, in a few days. He would then be forced to detach a division or so to save his dépôts, and he could not spare even a brigade if he wished to continue on the offensive. Certain intelligence that there was not a Frenchman left behind on the Douro, save the trifling garrisons of Toro, Zamora, and Tordesillas, had been brought in by General D’Urban. That officer, after conducting a very daring exploration round the rear of Marmont’s army, almost to the gates of Valladolid, had recrossed the Douro by Wellington’s orders at the ford of Fresno de Ribera, and fell in upon the left flank of the allied army near Fuente Sauco on July 18th[492]. For the rest of the campaign his 700 sabres were at Wellington’s disposal[493]. His report showed that Marmont’s rear was absolutely undefended, and that the Galicians could march up the Douro, if desired, without finding any opponents: it would be perfectly possible for them to cut all Marmont’s communications with Valladolid and Burgos, without being in any danger unless the Marshal detached men against them.

[492] Not July 17th, as Napier says. D’Urban’s diary proves that he recrossed the Douro on the 18th.

[493] He left one squadron near Zamora, to serve as covering cavalry for Silveira’s militia, who remained waiting for Santocildes’s advance, which they were to observe and support. His force was therefore reduced to 700 men.

The 20th of July proved to be a most interesting day of manœuvring, but still brought no decisive results. Early in the morning the whole French army was seen in march, with its head pointing southward, continuing the movement that it had begun on the previous day. Marmont had made up his mind to proceed with the hitherto unsuccessful scheme for turning his adversary’s right wing[494], in the hope of either cutting him off from his communication with Salamanca, or of catching him with his army strung out on too long a line from continuous and rapid movement. The character of this day’s march differed from that of the 19th, because the single well-marked Guarena valley ceased after a time to separate the two hostile armies. That little river is formed by three tributaries which meet at and above the village of El Olmo: each of them is a paltry brook, and their courses lie along trifling irregularities of the broad tableland from which they descend. It is only after their junction that they flow in a deep well-marked valley, and form a real military obstacle. Of the three brooks, that which keeps the name of Guarena lies most to the east: up its right bank and towards its source Marmont’s march was directed. Wellington’s parallel movement southward, on the other hand, was directed along the left bank of the Poreda, the middle brook of the three. Between them there was at first a narrow triangular plateau, on which neither party trespassed save with cavalry scouts.

[494] He adds in his _Mémoires_, iv. pp. 251-2, that if he had not succeeded in getting ahead of Wellington’s van, he had a counter-project of trying to get round his rear, but the British marched so exactly parallel with him that he got no chance of this.

After a few miles of marching Marmont ordered his advanced guard to cross the Guarena, which they could do with ease, no British being near, save a few cavalry vedettes. He then turned the head of his column south-westward, instead of keeping to his original direction due south. Having crossed the Guarena he came in sight of the British column marching on the other side of the Poreda brook from Vallesa. The movements of the two armies tended to converge, the point on which both were moving being the village of Cantalpino. It seemed likely that the heads of the marching columns must collide, and that a combat, if not a general action, would ensue. Each army was marching in an order that could be converted into a battle line by simply facing the men to right or to left respectively. Wellington had his troops in three parallel columns, the first one, that nearest to the French, being composed of the 1st, 4th, 5th, and Light Divisions, the second, which would have formed the supporting line if the army had fronted and gone into

## action, contained the 6th and 7th and Pack’s and Bradford’s brigades:

the 3rd Division and España’s Spaniards formed a reserve, moving farthest from the enemy. The light cavalry were marching ahead of the column, the heavy cavalry and D’Urban’s Portuguese brought up their rear. Marmont was clearly seen to be moving in a similar formation, of two columns each composed of four infantry divisions, with Curto’s _chasseurs_ ahead, and Boyer’s dragoons at the tail of the line of march[495].

[495] Marmont describes the formation (_Mémoires_, iv. p. 252) as ‘gauche en tête, par peloton, à distance entière: les deux lignes pouvaient être formées en un instant par _à droite en bataille_.’

The day was warm but clouded, so that the sun did not shine with full July strength, or the long march which both armies carried out would have been brought to an end by exhaustion at a much earlier hour than was actually the case. As the long morning wore on, the two hostile forces gradually grew closer to each other, owing to the new westward turn which Marmont had given to his van. At last they were within long artillery range; but for some time no shot was fired, neither party being willing to take the responsibility of attacking an enemy in perfect order and well closed up for battle. Either general could have brought on a fight, by simply fronting to flank, in ten minutes; but neither did so. Marmont remarks in his _Mémoires_ that in his long military service he never, before or after, saw such a magnificent spectacle as this parallel march of two bodies of over 40,000 men each, at such close quarters. Both sides kept the most admirable order, no gaps occurred in either line, nor was the country one that offered advantage to either: it was very nearly flat, and the depression of the Poreda brook became at last so slight and invisible that it was crossed without being noticed. The ground, however, on which the French were moving was a little higher than that on which the allies marched[496].

[496] There is an excellent description of the parallel march in Leith Hay, ii. pp. 38-40, as well as in Napier.

The converging lines of advance at last almost touched each other at the village of Cantalpino: the light cavalry and the 1st Division, at the head of Wellington’s front (or eastern) column of march had just passed through it, when Marmont halted several batteries on a roll of the ground a few hundred yards off, and began to shell the leading battalions of the 4th Division, which was following closely behind the 1st. Wellington ordered Cole not to halt and reply, nor to attack, but to avoid the village and the French fire by a slight westerly turn, to which the other divisions conformed, both those in the first and those in the second line[497]. This amounted to the refusing of battle, and many officers wondered that the challenge of Marmont had been refused: for the army was in perfect order for fighting, and in excellent spirits. But Wellington was taking no risks that day.

[497] This swerve and its consequence are best stated in Vere’s _Marches of the 4th Division_, p. 30.

The slight swerve from the direct southerly direction at Cantalpino made by the allied army, distinctly helped Marmont’s plan for turning its right, since by drawing back from its original line of movement it allowed the enemy to push still farther westward than his original line of march had indicated. This meant that he was gradually getting south of Wellington’s vanguard, and would, if not checked, ultimately arrive at the Tormes river, near the fords of Huerta, from which he would have been edged off, if both armies had continued in their original direction. During the early afternoon the parallel move continued, with a little skirmishing between cavalry vedettes, and an occasional outbreak of artillery fire, but no further developments. The baggage in the English rear began to trail behind somewhat, owing to the long continuance of the forced marching, and D’Urban’s Portuguese, who shepherded the stragglers, had great difficulty in keeping them on the move. A few score sick and foot-sore men, and some exhausted sumpter-beasts, fell behind altogether, and were abandoned to the French[498].

[498] Marmont says that if he had possessed a superior cavalry he could have made great captures, but he dared attempt nothing for want of sufficient numbers: he alleges that he took 300 stragglers--certainly an exaggeration as the British returns show very few ‘missing.’ _Mémoires_, iv. p. 233.

Late in the afternoon the armies fell further apart, and all save the outlying vedettes lost sight of each other. This was due to the fact that Wellington had made up his mind to settle down for the night on the heights of Cabeza Vellosa and Aldea Rubia, where Marmont had taken up his position a month before, when he retired from before San Cristobal. This was good fighting-ground, on which it was improbable that the French would dare to deliver an attack. The 6th Division and Alten’s cavalry brigade were detached to the rear, and occupied Aldea Lengua and its fords.

This had been a most fatiguing day--the British army had marched, practically in battle formation, not less than four Spanish leagues, the French, by an extraordinary effort, more than five. When the camp-fires were lighted up at night, it was seen that the leading divisions of the enemy were as far south as Babila Fuente, quite close to the Tormes and the fords of Huerta: the main body lay about Villaruela, opposite the British bivouacs at Aldea Rubia and Cabeza Vellosa. An untoward incident terminated an unsatisfactory day: D’Urban’s Portuguese horse coming in very late from their duty of covering the baggage-train, were mistaken for prowling French cavalry by the 3rd Division, and shelled by its battery, with some little loss of men and horses. The mistake was caused by a certain similarity in their uniform to that of French dragoons--the tall helmets with crests being worn by no other allied troops[499].

[499] The heavy cavalry in the British army were still wearing the old cocked hat, the new-pattern helmet with crest was not served out till 1813. The light dragoons were still wearing the black-japanned leather headdress with the low fur crest: in 1813 they got shakos, much too like those of French _chasseurs_.

The net result of the long parallel march of July 20th was that Marmont had practically turned Wellington’s extreme right, and was in a position to cross the Upper Tormes, if he should choose, in prolongation of his previous movement. The allied army was still covering Salamanca, and could do so for one day more, if the marching continued: but after that limit of time it would be forced either to fight or to abandon Salamanca, the main trophy of its earlier campaign. There remained the chance of falling upon Marmont’s rear, when his army should be occupied in crossing the Tormes, and forcing him to fight with his forces divided by the river. If this offensive move were not taken, and the parallel march were allowed to continue, the next day would see the armies both across the Tormes, in the position where Graham and Marmont had demonstrated against each other on June 24th. Wellington could not, however, begin his southward move till he was certain that the enemy was about to continue his manœuvre on the same plan as that of the last two days. If he started too early, Marmont might attack the San Cristobal position when it was only held by a rearguard, and capture Salamanca. Till an appreciable fraction of the French were seen passing the Tormes it was necessary to wait.

It appeared to Wellington that his adversary’s most probable move would be the passage of the Tormes by the fords at and just above Huerta. That he would abandon his previous tactics, and attack the British army, was inconsistent with the caution that he had hitherto displayed. That he would continue his march southward, and cross the river higher up, was unlikely; for the obvious passage in this direction, by the bridge of Alba de Tormes, was commanded by the castle of that town, which had been for some time occupied by a battalion detached from Carlos de España’s division. Wellington looked upon this route as completely barred to the French: he was unaware that the Spanish general had withdrawn his detachment without orders on the preceding afternoon. This astonishing move of his subordinate was made all the worse by the fact that he never informed his chief that he had taken upon himself to remove the battalion. Indeed Wellington only heard of its disappearance on the 23rd, when it was too late to remedy the fault. He acted on the 21st and 22nd as if Alba de Tormes were securely held. It would appear that Carlos de España thought the castle too weak to be held by a small force, and moved his men, in order to secure them from being cut off from the main army, as they clearly might be when the French had reached Babila Fuente. But the importance of his misplaced act was not to emerge till after the battle of Salamanca had been fought.

At dawn on the 21st Wellington withdrew his whole army on to the San Cristobal position[500], and waited for further developments, having the fords of Aldea Lengua and Santa Marta conveniently close if Marmont should be seen crossing the Tormes. This indeed was the move to which the Marshal committed himself. Having discovered at an early hour that Alba de Tormes was empty, and that there was no allied force observing the river bank below it, he began to cross in two columns, one at the fords of Huerta, the other three miles higher up-stream at the ford of La Encina. Lest Wellington should sally out upon his rear, when the greater part of his army had got beyond the Tormes, he left a covering force of two divisions in position between Babila Fuente and Huerta. This, as the day wore on, he finally reduced to one division[501] and some artillery. As long as this detachment remained opposite him, Wellington could not be sure that the French might not attack him on both sides of the Tormes.

[500] Napier says that this move was made on the night of the 20th, under cover of the smoke of the already-lighted camp-fires of the army. This is contradicted by Vere’s journal of march of the 4th Division, by Leith Hay’s Journal [’at daylight we marched to the Heights of San Cristoval’], by Tomkinson’s diary, and D’Urban, Geo. Simonds, and many others who speak of the move as being early on the 21st.

[501] This was the division of Sarrut.

The defile of the French army across the fords naturally took a long time, and Wellington was able to allow his weary infantry some hours of much-needed rest in the morning. Only cavalry was sent forward at once, to form a screen in front of the hostile force that was gradually accumulating on the near side of the fords. In the afternoon, however, when the greater part of the French were over the water, nearly the whole allied army received orders to cross the Tormes, and occupy the heights to the south of it. It moved practically in battle order, in two lines, of which the front passed by the ford of Cabrerizos, the second by that of Santa Marta. Only a reserve, now consisting of the 3rd Division and D’Urban’s Portuguese horse, remained on the north side of the river near Cabrerizos, to contain the French force which was still visible at dusk on the slopes by Babila Fuente. Till this detachment had disappeared, Wellington was obliged to leave a corresponding proportion of his men to contain it, lest the enemy might try a dash at Salamanca by the north bank. Marmont made no such attempt, and in the morning it was obvious that this rearguard was following the rest of his army across the Tormes.

During the night the French advanced cavalry were holding Calvarisa de Ariba on their left and Machacon on their right: the infantry were bivouacked in a concentrated position in the wooded country south of those villages. The British cavalry screen held Calvarisa de Abaxo[502], Pelabravo, and the height of Nuestra Señora de la Peña, close in to the corresponding front line of the enemy’s vedettes. The infantry were encamped in two lines behind the Ribera de Pelagarcia, the ravine, which runs north from Nuestra Señora de la Peña to the Tormes, between Santa Marta and Cabrerizos. This was Graham’s old position of June 24th, and excellent for defence. The right was on well-marked high ground, the centre was covered by woods. Only the left, near Santa Marta, was on lower slopes.

[502] That the British cavalry were still at dawn so far forward as Calvarisa de Abaxo is shown by Tomkinson’s diary (p. 185), the best possible authority for light cavalry matters. The 4th Division camped in the wood just west of Nuestra Señora de la Peña (Vere, p. 31), the 5th on high ground in rear of Calvarisa de Ariba (Leith Hay, p. 45), the 7th a little farther south, also in woody ground (diary of Wheeler of the 51st).

About an hour after nightfall the hills where French and English lay opposite each other were visited by an appalling tempest. ‘The rain fell in torrents accompanied by vivid flashes of lightning, and succeeded by instantaneous peals of thunder:’ writes one annalist: ‘a more violent crash of the elements has seldom been witnessed: its effects were soon apparent. Le Marchands brigade of cavalry had halted to our left: the men, dismounted, were either seated or lying on the ground, holding their horses’ bridles. Alarmed by the thunder, the beasts started with a sudden violence, and many of them breaking loose galloped across the country in all directions. The frightened horses, in a state of wildness, passing by without riders, added to the awful effect of the tempest[503].’ The 5th Dragoon Guards suffered most by the stampede--eighteen men were hurt, and thirty-one horses were not to be found. Another diarist speaks of the splendid effect of the lightning reflected on the musket-barrels of belated infantry columns, which were just marching to their camping-ground. Before midnight the storm had passed over--the later hours of sleep were undisturbed, and next morning a brilliant sun rose into a cloudless sky[504]. The last day of manœuvring was begun, and the battle which both sides had so long avoided was at last to come.

[503] Leith Hay, ii. p. 46.

[504] Diary of Green of the 68th, p. 98.

SECTION XXXIII: