Soult’s army was at this time cantoned with its right wing about Almanza, Yecla, and Fuente la Higuera, and its left wing—now about to become its advanced guard—in and around the large town of Albacete. A detachment from this wing had been for the last ten days attacking the isolated rock-fortress of Chinchilla, the only inland stronghold which was held by the Spaniards in the kingdom of Murcia. It was a Gothic donjon on an inaccessible cliff, only formidable because of its position, and manned by a trifling garrison. It might have held out indefinitely, having a resolute governor, a certain Colonel Cearra. But on October 9th, in a frightful thunderstorm, lightning struck the donjon, killed 15 soldiers, wounded many more, set the place on fire, and disabled the governor[120]. The garrison capitulated in sheer dismay, and the use by the French of the high road between Albacete and Almanza was no longer incommoded by the existence of this petty fortress.

King Joseph with the Army of the Centre marched out from Valencia on October 17th, and had his head-quarters at Requeña on the road to Cuenca on the 19th. On the 23rd he reached that ancient and much dilapidated city, and found it already in the hands of Drouet, who had arrived there on the 20th with Barrois’s division and the cavalry brigade of Avy from the Army of the South: he had expelled from it Bassecourt’s 3,000 Murcian troops. Soult had started on the 15th from Albacete, and had sent off Drouet’s detachment from San Clemente to Cuenca, while he himself marched by Belmonte on Tarancon and Santa Cruz de la Zarza, which he reached on October 25th. He had not got into real touch with the enemy till, on the last-named day, the cavalry on the right of his advance came into contact with Freire’s Murcian horse in front of Tarancon, and those on his left ran into the vedettes of Long’s British dragoons in front of Ocaña. During the time of his advance these troops had been retiring in front of him, from Consuegra, Toboso, Almonacid, Belmonte, and other places in La Mancha, where they had been providing a long screen of posts to observe his movements. They had, by Hill’s orders, retired from the 18th onward before the French cavalry, without allowing themselves to be caught up. It was only immediately in front of the Tagus that they slackened down their pace, and allowed the French to discover them. There was a smart skirmish in front of Ocaña on October 25th, between Bonnemain’s brigade and the 9th and 13th Light Dragoons and 10th Portuguese cavalry. The allied squadrons were pushed back towards Aranjuez with the loss of some 30 men, Erskine, who was in command of the cavalry division, refusing to make a stand or to bring up his reserves. His management of the troops was (not for the first time) much criticized by eye-witnesses[121], but it must be remembered that Hill had directed him not to commit himself to a serious action. On the same day Freire’s horse were turned out of Tarancon by Perreymond’s chasseurs.

The position of the commander of Wellington’s detached corps in front of Madrid had become a very responsible one between the 15th, when Soult’s advance began, and the 24th, when the enemy came up to the line of the Tagus and developed his attack. Fortunately Hill was in close touch with his chief: so well was the line of communication between them kept up, that it only took two days for a letter from Burgos to reach Madrid—and vice versa. When the army had come back from the Arlanzon to the Douro, the time became even shorter. Wellington received at Cabezon on the evening of October 27th dispatches that Hill had written on the morning of the 26th[122]. This contrasts wonderfully with the slow travelling of French correspondence—Souham got at Briviesca on October 17 a letter written by King Joseph at Valencia on October 1st. It had been obliged to travel by the absurdly circuitous route of Tortosa, Saragossa and Tudela. Truly the guerrilleros made concerted movements of French armies singularly difficult.

On the 17th October Hill had already got off his first letter of alarm to Wellington, saying that Soult was certainly on the move; by the 19th he knew that there was a column moving upon Cuenca, as well as the larger force which was advancing by San Clemente and Belmonte. He asked for orders, but meanwhile had to issue his own, in consonance with earlier directions received from Wellington. These presupposed two conditions which had not been realized—that the fords of the Tagus would be impassable, and that Ballasteros’s Andalusian army would already have crossed the Sierra Morena to Alcaraz and be lying on Soult’s flank. But their general directions were still practicable: the line of the Tagus was to be defended unless the enemy were in overwhelming strength: if (contrary to Wellington’s expectation) the whole French force in Valencia should advance, and its numbers prove greater than Hill could hope to check, he had been directed to evacuate Madrid, and to fall back beyond the Guadarrama, in order to join his chief on the Adaja, south of Valladolid, in Old Castile[123]. The first thing necessary was to discover the strength of the enemy—all accounts sent in by the Spaniards agreed that it was very great, and in particular, that the column going by Cuenca was no mere detachment, but a solid and considerable force. Meanwhile Wellington, even as late as October 12[124], had been informing Hill that his design of marching down on Madrid with three divisions, when the siege of Burgos should be either successfully concluded or else abandoned, was still retained. Any morning a dispatch might come to say that the commander-in-chief, with 15,000 men, was on his way to Valladolid; and therefore the army in front of Madrid must be ready for him, concentrated and in marching order. For if he came in person with such a reinforcement, Soult and King Joseph could be fought and beaten, whenever they made their appearance.

On October 15th, when the French advance had actually begun, the allied troops in New Castile were disposed with an outer screen, mainly consisting of Spanish troops, and a central nucleus of Hill’s own Anglo-Portuguese placed in cantonments between Madrid and the Tagus. Bassecourt was at Cuenca with 3,000 men; Elio, with Freire’s Murcian horse and a weak division of infantry—5,000 men in all—was watching the high roads from Albacete and Requeña to Madrid, in front of Tarancon. Penne Villemur’s cavalry with Morillo’s infantry—3,500 men at the most—lay across the great chaussée from Andalusia, about Herencia and Madridejos[125]. These troops formed the outer screen—not taking account of the Empecinado, who was (as usual) on the borders of New Castile and Aragon, worrying Suchet’s garrisons in the latter kingdom. Behind Penne Villemur, and south of the Tagus, were Long’s British and H. Campbell’s Portuguese cavalry brigades, in La Mancha, at Toboso, Villacanas, and other places. All the rest of the British troops were north of the Tagus, in the triangle Madrid-Toledo-Fuentedueñas, as were also D’Urban’s Portuguese horse and Carlos de España’s Spanish infantry division. When the advance of Soult and King Joseph developed itself, Hill drew everything back behind the Tagus, save Bassecourt’s division at Cuenca, which being evicted from that place by Drouet on the 20th did not retire towards Madrid, but went up into the mountains, and ultimately by circuitous routes rejoined the Alicante army.

On the 25th October, when Soult’s advanced cavalry had driven Long and Freire from Ocaña and Tarancon, Hill had his whole force, British, Portuguese and Spanish, arrayed in what he intended to be his preliminary fighting position along the Tagus. The extreme right was formed by Skerrett’s 4,000 men from Cadiz, who had got up to the front just in time to take their share in the fighting. They lay at Toledo and Añover. Then came the four brigades of the British 2nd Division, two of them at Aranjuez—which was held as a sort of tête de pont south of the river—and two at Colmenar de Orija. The line beyond them was prolonged by Penne Villemur and Morillo about Belmonte de Tajo and the fords of Villamanrique. Elio and Freire, who had retired across the bridge of Fuentedueñas after being driven out of Tarancon, was in charge of the upper Tagus from that point to Sacedon. Behind this front line lay the reserves—the 3rd and 4th Divisions close together at Valdemoro and Cienpozuelos, behind Aranjuez; the Light Division at Arganda; Carlos de España at Camporeal; Hamilton’s Portuguese division at Chinchon. Of the cavalry, Long and Campbell’s Portuguese, after being turned out of Ocaña, had fallen back on Aranjuez: D’Urban’s Portuguese were at Arganda, Slade’s brigade at Morata, Victor Alten’s at Getafe[126]. One march would concentrate the whole of the Allies, horse and foot—save Elio and Skerrett’s detachment alone—to defend the passage of the Tagus at either Aranjuez or Fuentedueñas, the two crossing-places which Hill judged that Soult would take into consideration, when he attempted to force the line of the river. About 36,000 men would be available, of whom 28,000 were Anglo-Portuguese and 8,000 Spaniards.

Soult, however, kept perfectly quiescent in front of Aranjuez and Fuentedueñas on the 26th-27th. He had still his cavalry to the front, but his infantry divisions were only coming up in succession: Conroux’s in especial, being still in quarantine owing to the yellow fever, was very far behind. But it was not merely the late arrival of his rear that kept Soult motionless: he was waiting for the Cuenca column to bring pressure to bear upon Hill’s flank, and did not intend to commit himself to any important engagements until the whole French army was in line. He expressed to King Joseph his opinion that the Allies were drawn out upon too long a front, and that a bold thrust at Aranjuez would probably succeed, when the attention of Hill should be drawn away to the East by the appearance of the Cuenca column in the direction of Fuentedueñas. Meanwhile he proceeded to make his preparations for attacking Aranjuez on the 28th.

Such an attack was never delivered, because Hill, on the evening of the 27th, made up his mind that he must not fight upon the Tagus[127]. For this determination there were three causes. The first was that the river still remained so low, the autumn rains having been very scanty hitherto, that it was fordable in many places. The mere breaking of the bridges at Aranjuez and Fuente Dueñas did not make it impassable, as Wellington and Hill himself had expected would be the case by the end of October. Secondly, if the line of the Tagus were forced at any point, the troops strung out along it had a very dangerous retreat before them, owing to the fact that the Tajuna, a stream not much smaller than the Tagus itself in this part of its course, runs behind it and parallel to it at a distance of only eight or ten miles. The number of spots where the Tajuna could be crossed, by fords, bridges, or ferries, were very few, and it was to be feared that bodies of troops abandoning positions on the Tagus, and retreating to the next line, might find themselves pressed against the Tajuna at impassable sections of its course, and so might be destroyed or captured if the enemy pursued with vigour. Thirdly—as Soult expected—the movement of the King and the column from Cuenca had now begun to exercise pressure on Hill’s mind. He had already moved two British brigades of the 2nd Division to Fuente Dueñas, replacing them at Aranjuez by Skerrett’s force, which left Toledo. But what if the King should cross the Tagus not at Fuente Dueñas but above it, where the river was only observed by Elio’s Murcians? They certainly could not stop him, and the whole Tagus line would be turned.

Hill’s resolve was now to defend not the Tagus but the line, running North and South, of the Henares and the Jarama (the river formed by the union of the Tajuna and Manzanares), from Guadalajara to the point near the Puente Larga where the Jarama falls into the Tagus. On the 28th Skerrett evacuated Aranjuez, and all the other troops fell back in similar fashion. This position left the allied army still covering Madrid, and with a safe retreat to the passes above it, should things go ill. The new disposition of forces was as follows: Toledo had been handed over to the partida of El Medico, since no French reconnaissances had come in this direction, and it was clear that the enemy had no serious intentions on this flank. The extreme right wing of the army was formed by the 4th Division, now once more under General Cole, who had come up, cured of his wound, from the Salamanca hospital. It lay at Añover, behind the point where the Jarama flows into the Tagus, with its flank covered by the Gunten river and some of Long’s dragoons. Next in line, six miles to the North, was Skerrett’s force, holding the Puente Larga, the main passage over the Jarama river, two miles north of Aranjuez. Beyond him were the 3rd Division and Hamilton’s Portuguese about Valdemoro and St. Martin de la Vega. Then came the Light Division at Alcalá de Henares[128]: Carlos de España’s and Morillo’s Spaniards were in their company. Elio’s Murcians were directed to fall back on Guadalajara. So much for the infantry: the cavalry was kept out in front, with orders to keep a line of vedettes on the Tagus till they should be driven in, and then to hold the course of the Tajuna in a similar fashion, before breaking its bridges and falling back on to the Henares and Jarama, the real fighting line. But nothing was to be risked, and the main body of each brigade was to keep itself in front of a practicable crossing, by which it could retreat when the enemy should have shown himself in force. ‘Sir Rowland,’ wrote his Quartermaster-General, ‘wishes you to keep the posts on the Tajuna, and those in front of it (on the Tagus), as long as you can with safety. Cover the line of the Henares as long as you can.’[129] Slade’s, Long’s, and Campbell’s Portuguese squadrons had the right, covering the river bank from Aranjuez to Villamanrique with their vedettes, D’Urban, Victor Alten, and Penne Villemur held the left, from Villamanrique up stream.

On Oct. 28th the French cavalry, having detected the disappearance of Hill’s infantry, crossed the Tagus both at Aranjuez and Fuente Dueñas in force, whereupon the allied horse retired behind the Tajuna and broke all of its bridges. Soult at once commenced to repair the bridges of Aranjuez, and brought an infantry division forward into the town on the 29th, but made no serious effort to feel Hill’s position behind the Jarama and Tajuna, being determined not to involve himself in heavy fighting till King Joseph and the column from Cuenca were up in line. The head of the Army of the Centre, however, reached Fuente Dueñas this same day, and began to pass[130], meeting (of course) with no opposition. But the reconstruction of the bridge took some time, and D’Erlon’s infantry was not across the Tagus in any force till the next day. The King himself rode to Ocaña, conferred there with Soult, and made arrangements for a general forward movement upon the 30th. There would have been heavy fighting upon the 30th-31st, if Hill had been permitted to make a stand on his chosen position with the 40,000 men whom he had placed in line between Alcalá and Añover. He had now all his troops concentrated except Elio’s Murcians, who lay out in the direction of Guadalajara with no enemy in front of them. But on the morning of the 29th Hill received a dispatch from Wellington, dated from Cabezon on the night of the 27th, which upset all the arrangements made hitherto. The important paragraph of it ran as follows: ‘The enemy are infinitely superior to us in cavalry, and from what I saw to-day very superior in infantry also. We must retire, and the Douro is no barrier for us. If we go, and cannot hold our ground beyond the Douro, your situation will become delicate. We certainly cannot stand against the numbers opposed to us in any situation, and it appears to me to be necessary that you, as well as we, should retire. The only doubt which I entertain is about the road which you should take, and that doubt originates in the insufficiency of this army to stop the army opposed to it for a sufficient time to allow you to reach the Adaja. I propose to remain on the Pisuerga to-morrow (October 28) and as long as I can upon the Douro, and then to retire by Arevalo. God knows whether I shall be able to remain on either river!; and if I cannot, your retreat should be by the valley of the Tagus. If I can remain, we should join as arranged by previous letters. If I can remain on the Pisuerga to-morrow, I shall pass the Douro on the 29th, and shall probably be able to prevent the enemy from crossing in force till the 1st November, in which case I shall reach Arevalo on the 3rd. You will not receive this letter till the 29th. You will arrive at the Escurial, probably on the 31st, at Villacastin on the 2nd, at Arevalo on the 4th.... If I should not be able to hold my ground either on the Pisuerga or the Douro, I shall apprise you of it at the first moment, and shall suggest your line of retreat.... Your march, as proposed (i.e. via the Guadarrama) at least as far as Villacastin, would be secure, whereas that by Talavera, &c., would not, till you shall cross the Tagus. Do not order the bridge at Almaraz to be taken up or destroyed, till you are certain you do not want it.’[131] The dispatch ended by directing Hill to bring on with him Carlos de España’s, Morillo’s, and Penne Villemur’s Spaniards, but to order Freire, Elio, and Bassecourt to join Ballasteros by the route of Toledo, while the Empecinado had better go to his old haunts in the mountains beyond Guadalajara.