As to the first of these points, it will at once be observed that unless you have some strong compelling motive for driving a simple straight line you will, in a country of many rivers, avoid such a scheme and seek for the cheapest crossing of each water. You must seek a ford, or narrow, or a place with specially hard banks, and not merely take haphazard that part of the stream which lies on your direct line, and seeing that such numerous waterways involve also long numerous flats along the streams, valley floors subject to flood or formed of boggy soil, the tendency to diversion in a road system under such conditions is intensified by the marshes which abound in a country so watered. If you look at the Roman road system you will see how, for the considerations which I will deal with later, it usually; if not always, neglects special opportunities and takes the water as it comes, preferring an expensive straight line to a cheaper winding line; but everything done since the Roman road system has been affected by the perpetual consideration of the easiest river crossings inland, while the same influence has deflected the road round the greater estuaries and ports. [Illustration: _Part II_, Sketch II] The lowest bridge over a river is a point of transformation. It stops traffic from the sea going any higher. But to carry on your journey from the sea as far as possible is obviously an economic advantage, especially in early days of expensive and slow road traffic. Therefore a nation dealing with the sea and largely living through sea-trade casts its first bridge _as far up stream as possible_, and that is exactly what you find upon all the rivers of England for centuries. Even to this day the tendency to build bridges lower down than the old first bridge is checked, in spite of the very strong motive we have in the development of the railway system. Take a map (Sketch II), and look round the coast and see how true this is. The lowest old bridge of the Tyne was at Newcastle; of the Trent, I believe, at Gainsborough; of the Thames, of course, right up inland at London; of the Stour, at Canterbury; of the Sussex Ouse, at Lewes; of the Arun, at Arundel; of the Exe, at Exeter. The deep arms of Plymouth Sound were unbridged until the railway came; so Fowey river and the Fal, unbridged to this day; the Severn is not bridged at all till Gloucester, nor was the Dee till Chester. Now this had the effect everywhere of checking a direct road system and deflecting the ways everywhere to suit the convenience of the ports. And there again we find, for reasons which will be given in a moment, the Roman roads directly crossing estuaries, but every subsequent road system going round them. Take two examples. The Roman road to the north, which runs all along the ridge of Lincolnshire, strikes the Humber where that stream is from 2000 to 3000 yards wide, crosses by a ferry, and continues on the far side.