An inherent defect of an ordinary weir is that it obstructs the passage of floods. The obstruction may or may not be of consequence. Sometimes it is of great consequence. Attempts have been made to partially remedy the evil by placing the weir obliquely to the stream, thus giving it a greater length. At ordinary water-levels the flow over the crest of the weir is normal to its length, or nearly so. Supposing that the water has to be held up to a given level, the crest of the weir must be higher, because of its greater length, than if it were normal to the stream. In a flood the water has a high velocity and flows over the weir in a direction nearly parallel to the axis of the stream, so that the effective length of the weir is not much greater than if it were normal to the stream, and, its crest being higher, it obstructs the flood as much. Oblique weirs are usually made as in fig. 31. If made in one straight line, there might be excessive action on the bank at the lower end. If the weir is lengthened, not by being built obliquely but by a widening of the stream at the site, the crest has to be raised and nothing is gained. The only arrangement by which a weir can be made to hold up water when a stream is low and to let floods pass freely, consists in having part of the weir movable, _i.e._ consisting of gates, shutters or horizontal or vertical timbers, which can be withdrawn to let floods pass, and can be manipulated to any extent so as to regulate the amount of water passing. A familiar instance of a movable weir is the one which is usually placed across a mill stream, the wooden gates working in grooves in the masonry. Above a weir in Java, 162 feet long, there was a great accumulation of shingle in the bed of the river, and the head of a canal taking off above the weir became choked. The crest of the weir on the side away from the canal was raised 5¼ feet and the crest sloped gradually down, a length of 43 feet on the side next the canal remaining as it was. This was quite successful. It was practically a contraction of the river near the canal off-take, and this must have caused scour, so that the bed became lower than the floor of the canal head and the shingle was not carried in. The shingle, however, is said to have been carried over the weir (_Min. Proc. Inst. C.E._, vol. clxv.). A lock is an adjunct to a weir, used when navigation has to be provided for. The lock may be placed close to the weir or it may be in a side channel, the upstream end of the lock being about in a line with the weir. Locks have already been discussed in CHAP. IX., _Art. 3_. Frequently a “salmon ladder” has to be provided. It consists of a series of steps or a zigzag arrangement so that the velocity of the water is not too great for the fish to ascend.