Meanwhile, on the hill of San Miguel, a second battery (No. 2 on the map) was dug out behind the gorge of the dismantled hornwork, and trenches for musketry (_a.a_ on the map) were constructed on the slope of the hill, so as to bring fire to bear on the flank and rear of the lower defences of the Castle. The French heavy guns of the Napoleon battery devoted themselves to incommoding this work; their fire was accurate, many casualties took place, and occasionally all advance had to cease. A deep trench of communication between the batteries on San Miguel and the attack in front of San Pedro was also started, in order to link up the two approaches by a short line: the ground was all commanded by the Castle, and the digging went slowly because of the intense fire directed on it. Meanwhile battery No. 1 on San Miguel at last came into operation, firing with five howitzers (the 18-pounders originally placed there had been withdrawn) against the palisades and flank of the north-western angle of the outer _enceinte_ of the Castle. These inefficient guns had no good effect; it was found that they shot so inaccurately and so weakly that little harm was done. After firing 141 rounds they stopped, it being evident that the damage done was wholly incommensurate with the powder and shot expended [September 25]. This first interference of the British artillery in the contest was not very cheering either to the troops, or to the engineers engaged in planning the attack on the Castle. The guns in battery 1 kept silence for the next five days, while battery 2, where the 18-pounders had now been placed, had never yet fired a single shot. Wellington was now staking his luck on the mine, which was being run forward from the head of the flying sap.