up another tank in the town behind the orphanage. This was the plan of His Excellency van Mydregt, although it was never put down in writing. Meantime care must be taken that the slaves and other native servants of persons residing in the Castle do not through laziness throw the dirt which they are supposed to carry away from the fort on the opposite bank of the moat, and thus raise a space which the Company would much rather lower, and gradually and imperceptibly prepare a suitable place for the battery of an enemy. I have had notices put up against this practice, under date July 18, 1695, and these must be maintained and the offenders prosecuted. Considering the situation of the Castle and the present appearance of the moat, I think that the latter is already sufficiently deep if always four or five feet water be kept in it. In order to do this two banks would have to be built, as the moat has communication in two places with the river, while the river also touches the fort at two points. This being done I think the moat could be kept full of water by two or three water mills driven by wind and pumps, especially during the south-west monsoon or the dry season, when an attack would be most likely to occur, and there is always plenty of wind to keep these mills going both by night and day. A sluice would be required in the middle of these banks so that the water may be let out whenever it became offensive by the river running dry, to be filled again when the water rose. It would have to be first ascertained whether the banks could really be built in such a way that they would entirely stop the water in the moat, because they would have to be built on one side against the foundations of the fort, which I have been told consist of large irregular rocks. An experiment could be made with a small mill of the kind used in Holland in the ditches along bleaching fields. They are quite inexpensive and easily erected and not difficult to repair, as they turn on a dovetail. The late Commandeur Anthony Paviljoen also appears to have thought of this plan even before this Castle was built, when the Portuguese fort was occupied by the Company, as may be seen from his instructions of December 19, 1665. [56] This would, in my opinion, be the course to follow during the south-west monsoon, while during the north-east monsoon there is usually so much rain that neither the salt river nor the water mills would be required, while moreover during that time there is little danger of an attack. These three plans being adopted, the banks of the moat could be protected by a wall of coral stone to prevent the earth being washed away by the water, as the present rocky bed of the moat is sufficiently strong to serve as a foundation for it. The moat has already been dug to its proper breadth, which is 10 roods.