At last, however, on March 30, one of the two new batteries in the gorge of the Picurina was able to open, and on the 31st the other followed suit, supported by a third supplementary battery (no. 7), planned under the left flank of the fort. The practice was excellent, but at first the effect was not all that had been hoped: the Trinidad and the Santa Maria bastions were solidly built and resisted well. On April 2, however, both began to show considerable and obvious injury, and it was clear that a few days more would ruin them. But there was one serious _contretemps_: the inundation between the Picurina and the fortress showed no signs of going down--it had been swollen by the rains of the 20th-24th, and could not flow away so long as the dam at the lunette of San Roque kept it back. While the water was held up, the breaches, soon about to develop, could only be got at by a narrow and curved route, between the inundation and the steep slope on which stands the Pardaleras. It had been intended that the assault should be delivered from the trenches, but this was impossible till the Rivillas should have fallen to its usual insignificant breadth and depth. Hence efforts were made to burst the dam at all costs, but neither did artillery fire suffice, nor a venturesome expedition on the night of the 2nd of April by the engineer Lieutenant Stanway and 20 sappers, who slipped down the ravine and laid powder-bags against the dam, despite of the French fire. The powder exploded, but did not do its work. For several days an attempt was made to sap down to the dam from the second parallel. But it cost so many lives at the head of the sap, and the zig-zags advanced so slowly, that on the 3rd of April the attempt was given up, and it was determined that the breaches must be assaulted from the west bank of the Rivillas only.