With the railways it is otherwise. The economic effort required for the construction of the track is such that the added expense of clearing the forest is a much smaller fraction of the whole, and the type of bed which has to be established for the track partially solves (but only partially) the question of upkeep. Even the railway can be overcome by the vigour of tropical vegetation, but it has a better economic basis in the densely wooded country than has the Road. One of the most curious facts in the history of roads due to the obstacle of wood is the deflection of the Roman Road through this cause after the decline of civilization. One can find many instances of this even in England, light as is the afforesting of this country, and small as are the districts affected. Thus the Great North Road making for Stamford is a broad, unmistakable way raised high above the neighbouring country, and looking like some great double rampart, from the crossing of the Welland for miles to the north and west. It approaches a small patch of wood on a hill and disappears. It remains lost for a mile after its destruction by the wood, and is not found again in anything like its earlier sharpness of outline till Stamford is reached. That is because the upkeep through the wood became too difficult in the Dark Ages, and men turned the obstacle by developing a new road round it. Another very clear example is to be found on the Stane Street north of Eartham, where the great _Nore Wood_ through which the Roman road was driven usurped it in later times, overgrew it, and deflected the modern road round by Duncton Hill. We have here probably not so much a case of keeping down the new growth as of the wetness of the track when artificial material ceased to be used, and of the difficult going thus made between the trees. The occasional fall of trees across the road left unremoved, and the danger in such times from any close cover must not be neglected. But, whatever the cause, woodland perpetually deflects a Roman road after the breakdown of the old civilization. It deflects it almost as often as does the marsh of a river valley. (b) The obstacle to the making of a road due to intermittent vegetation is one which plays no part in our system, and is unknown to our climatic conditions. Nor is it of any great effect save in a few special highly characterized regions of the world. The track, once established, can commonly keep down even the riot of spring vegetation in open land. Such exceptions as there are, due to the exceptional development of grasses, affect no part of the world where communications need high development. The factor exists, but needs no more than a mention. ii The last main cause modifying the trajectory of a road is the relative proximity of material for its construction, using the word “proximity” in the wider sense to include all economic effort: what to-day we call the “cheapness” of the material.