There is hardly any factor in connection with the theory of the road which needs more material modification as civilization changes than this factor of gradient. The sharpest contrast in the whole of history is that which I have just mentioned: that of the railroads. Men suddenly found themselves possessed of a new instrument which enormously multiplied their power on the flat and yet was quite incapable of anything like the old gradients. Going level or on very slight gradients it could give them travel far more rapid and inexpensive than any that had been known before, but one in fifty bothered it badly, one in thirty was wholly unnatural, and the existing gradients of one in ten, eight, six, were out of the question. Further, the least inclination increased all difficulties, and the addition of inclination produced these difficulties in more than a geometrical progression. The result was the revolution whose effects we see about us everywhere: the tunnel, the cutting, the embankment. To-day, a couple of generations after that revolution, there comes the new problem of the internal-combustion engine, where the gradient again appears in a new light. The motor takes gradients far steeper than the rail. Its difficulties are not increased in the same ratio. But it cannot always deal with the horse road. Lynton and Lynmouth and their Devonshire valley form perhaps the best example of this in Great Britain. You have here terrible gradients which were just possible for the horse vehicle and are hardly possible for the motor vehicle, and you have the new road round by Watersmeet attempting partially, but not entirely, to solve the problem. A special case in this general category of gradients, and one much more complex than appears at first sight, is the case of the pass, or gap. Men have always naturally made for any notch in a line of hills to save themselves the effort of higher climbing. It began with foot travel, and has continued right on throughout the history of the Road. In high mountains provided with low passes the use of a saddle in the range was obvious and often necessary; but there were disadvantages even in that apparently unexceptionable rule. One was the question just dealt with of the double slope: the consideration of the other side--the most obvious pass from the one side did not necessarily lead to the best descent upon the other. Another was the conformation of many ranges, which is such that the approach to the ridge is much steeper at the _summit_ of a “col” or pass than it is by tracks to one side.