line of communications.... I am bound, therefore, to make a pressing demand that your left wing may be kept in a position which makes the communication between our armies sure, so that we may be able, by uniting our disposable forces, to go out against the enemy with the assurance of success.’ [289] The letter is printed in Marmont’s _Correspondance_, iv. pp. 304-5. This was precisely what Marmont had intended to do. He was convinced, like Soult, that Wellington’s next move would be against Badajoz, and he placed Montbrun and the divisions of Foy, Brennier, and Sarrut about Talavera, Monbeltran, and Almaraz, precisely in order that they might be in easy touch with Drouet. On February 22 he wrote to his colleague explaining his purpose in so doing, and his complete acquiescence in the plan for a joint movement against Wellington, whenever the latter should appear on the Guadiana[290]. His pledge was quite honest and genuine, and in reliance on it Soult made all his arrangements. These, however, appear to have been rather loose and careless: the Marshal seems to have felt such complete confidence in the combination that he made insufficient preparations on his own side. No reinforcements were sent either to Badajoz or to Drouet, whose 12,000 men were dispersed in a very long front in Estremadura, reaching from Medellin and Don Benito on the right to Fregenal on the left. This is why Graham, when he moved forward briskly on March 17th, found no solid body of the enemy in front of him, but only scattered brigades and regiments, which made off in haste, and which only succeeded at last in concentrating so far to the rear as Fuente Ovejuna, which is actually in Andalusia, and behind the crest of the Sierra Morena. We may add that having been advised by Drouet as early as March 11th[291] that British troops were accumulating behind Elvas, Soult ought to have taken the alarm at once, to have moved back to Seville from Santa Maria by Cadiz, where he lay on that date, and to have issued orders for the concentration of his reserves. He did none of these things, was still in front of Cadiz on March 20[292], and did not prescribe any movement of troops till, on that day, he received Drouet’s more definite and alarming news that Wellington was in person at Elvas, and had moved out toward Badajoz on the 16th. Clearly he lost nine days by want of sufficient promptness, and had but himself to blame if he could only start from Seville with a considerable field-force on March 30. All that he appears to have done on March 11 was to write to Marmont that the long-foreseen hypothesis of a move of Wellington on Badajoz was being verified, and that they must prepare to unite their forces. Jourdan has, therefore, some justification for his remark that he does not see why Soult should have been before Cadiz, amusing himself by throwing shells into that place[293] as late as March 20th.