striking their masts.” Last of all, in recent photographs the stone bridges of China are steep whenever they are built with arches, so we can follow the Chinese gabled bridge from our own time to the thirteenth century. They came into use partly because they were convenient to shipping, and partly because they could be erected from low embankments. APPENDIX II STEEP ROMAN BRIDGES Young pontists are always eager to know whether the Romans built gabled bridges, setting an example both to the East and to the West. On this topic there is little evidence, for most of the Roman bridges were built of timber. At Rimini, in the famous bridge of Augustus, there is an ascent at each end over the abutment, and at Alcantarilla, near Utrera, in Andalusia, the Roman bridge may be described as hog-backed. It crosses the Salado, a tributary of the Guadalquivir. Recently Mr. Edgar Wigram visited Alcantarilla, and he writes to me as follows:-- “The Roman bridge there is most interesting, almost untouched by restoration, yet it remains serviceable. It is a hog-backed structure of two arches, each about thirty-five feet in span; the width between the parapets may be fifteen feet, but a swarm of bees happened to be merry on the bridge, so I did not try to take accurate measurements. The voussoirs and spandrils are of stone with hammer-dressed faces, while the soffits are formed with wedge-shaped blocks of concrete, and a certain amount of brick is found in the piers. Along the river on one side are remains of an embankment. A tower stands at one end of the bridge, placed centrally to it, so the road has to make a double turn to pass. One wall of the tower is destroyed, but the other three are still about half their original height. The lower courses are of big stone blocks, while the upper part of the faces are filled in with ‘tapia’ concrete; the angles (or at least the two which still remain intact) are grooved with a queer circular recess some twelve inches in diameter. What purpose these grooves can have served I do not know. They look as if they may have been intended to accommodate the hinge-posts of gates; but a gate hung in them would hardly swing through ninety degrees. If a second tower ever existed, its foundations do not appear above ground-level. At Córdova there is only one tower, and it stands in a very similar position. By analogy, then, we may suppose that a second tower was not built at Alcantarilla; yet the grooved angles seem to require a corresponding tower with corresponding grooves, if gates were ever swung from them. Perhaps the grooves formed pivots for some sort of defensive engine, such as the ‘iron hand’ of Archimedes, which seems to have been some sort of great grappling crane. The angles of a tower would be fit places to plant a weapon of this description; but we need help from an expert in ancient military engineering.”