Is there any concrete evidence to suggest that the bridge with criss-cross piers has gone through many phases of change, of growth or of decadence? Yes. At Archangel, in North Russia, the criss-cross piers are more primitive; instead of being arched they are upright and stiff; but as the bridge is nearly a quarter of a mile long, and as it is taken down every spring (before the ice breaks up noisily, and the Dwina thunders into a raging torrent), crude workmanship in a hurried routine is excusable. The main point is that a bridge akin to the Gaulish type and to the variation in Kashmir exists in North Russia. And another variation is met with at Bhutan, in India. Brangwyn has drawn it, and we shall study it later in a page on gateway-towers (p. 272). In the highlands of Eastern Kurdistan, the borderland of Asiatic Turkey and Persia, travellers find a bridge akin to the Bhutan variety. An excellent book on these highlands has been published,[20] and its authors, very generously, have written for me some valuable notes on the bridges. Before I quote them in full, let me ask you to remember that in Eastern Kurdistan timber is uncommon; hence the criss-cross bridge has been evolved into another sort of primitive structure--a third cousin, several times removed. A Kurdistan bridge is built as follows: “A site is selected, if one can be found, where two immovable and flat-topped masses of rock face one another across the stream to be bridged: an abutment of unhewn stones is built on these, solid, until a height has been reached sufficient to be safe from any flood. “Then a bracket of four or more rows of poplar trunks is constructed on each abutment; short stout trunks form the bottom row, and those of each succeeding one are naturally longer than the preceding. Unless the bridge is unusually wide in the footway four poplars are enough to form a row, and the butts of the trees, which are kept shore-wards, are weighted down with big stones as counter-weights to hold them in place. “The top of each row of trunks projects perhaps five feet beyond the preceding one, so that when a bracket of four rows is completed, it may project perhaps twenty feet over the stream. “When the corresponding bracket has been completed, two long poplar trunks are slung by withies from bracket end to bracket end, a footway of withy hurdles, resting on faggots, is laid down over all, and the bridge is complete. The length of this centre span is of course limited by the height of the poplars available. I should think fifty feet the extreme possible. “If the width of the river makes it necessary, one or more piers of stone,--I have seen as many as three,--are erected in midstream, preferably on rock foundations. Each of these carries a bracket on each side, but this double bracket is usually made of ‘whole trunks’ and these naturally need no counter-weighting.