On the night of the 4th and 5th of July the defences of Monjuich appeared in such a ruinous condition that Commandant Fleury, the engineer officer in charge of the advanced parallel, took the extraordinary and unjustifiable step of assaulting them at 10 p.m. with the troops--two companies only--which lay under his orders, trusting that the whole of the guards of the trenches would follow if he made a lodgement. This presumptuous attack, made contrary to all the rules of military subordination, was beaten off with a loss of forty men. Its failure made Verdier determine to give the fort three days more of continuous bombardment, before attempting to storm it: the old batteries continued their fire, a new one was added to enfilade the north-western bastion, and cover was contrived at several points to shelter the troops which were to deliver the assault, till the actual moment of the storm arrived[35]. But three hundred yards of exposed ground still separated the front trenches from the breach--a distance far too great according to the rules of siegecraft. The Spaniards meanwhile, finding it impossible under such a fire to block the breach, which was now broad enough for fifty men abreast[36], threw up two walls of gabions on each side of it, sank a ditch filled with chevaux-de-frise in front of it, and loopholed some interior buildings of the fort, which bore upon its reverse side. [35] Napier says (ii. 250) that ‘the breaching fire ceased for four days before the assault,’ and that this caused the failure. The statement is in direct contradiction of Vacani (iii. 277) who states that Verdier on the contrary ‘proseguì per tre giorni il vivo fuoco della sua artilleria,’ and of Belmas (ii. 530) who makes the same statement. [36] See Alvarez’s letter in Belmas’s Appendix, no. 15, where he says that the breach had this breadth since July 3.