66 OFF TO DERBENT
larger blocks, however, are of still greater proportions, so that Ibn Fakih (903 a.d.) said that it would take fifty men to lift them.^ All the blocks are carefully set; and some of the old- est accounts of them speak of their being bound together by cramps of iron,^ so that they must have formed a perfect breastwork in the days when artillery was not known.
The composition of the stone itself is noteworthy, as its basis is a peculiar calcareous stone filled with mussel shells, charac- teristic of the so-called Aral-Caspian formation, and found also in the mountains near by.^ It must be added that in some places the nature of the stone varies slightly.
We walked a considerable distance along the top of the bul- warks, which were high enough, I noticed, to cause my guide giddiness when the path led near the edge; and then we descended at a place where there was a breach in the wall, probably due to an old gate, and were assisted down by the helping hand of a Russian soldier who chanced to be strolling along the rampart.
While standing on the wall, I thought of the old picture of Derbent, sketched by Olearius three hundred years ago and re- produced above. The outline of the two parallel rows of rampart was still as distinct as then, but the Dubaru, or ' Double Wall,' which once ran midway across the town, connecting them and shutting off the lower quarter, had long since disappeared.* Olearius's description of the city, in connection with his draw- ing, is so good that I repeat it, though the flavor of its old- fashioned German is somewhat lost in my literal translation.
'The city is divided into three parts. The Uppermost (A) is the Castle on the mountain, where the Governor hath his dwelling. 'Twas
1 Ibn Fakih, ed. De Goeje, Bill. » See Trautschold, Ueher die Naph- Geog. Arab. 5. 291. tha-quellen von Baku, in Zt. d.
2 Ibn Fakih, 5. 291. Eichwald, Deutschen Geologischen Gesellschaft, Beise, 1. 127, states that the Persians 26. 269, Berlin, 1874 ; and compare asserted that a cement of ' Hirsch- Olearius, p. 377, and Hanway, 1. 370. hornleim(!)' [hartshorn glue?] orig- * See Kazem-Beg, p. 101, n. 1 and inally bound the blocks together. n. 2.
�� �
