28 A N C I E N T M N U M E N T S . 



bed, which, before the construction of this work, kept the course indicated l)y the 

 dotted Hue .r. It was turned by the builders from its natural channel into the 

 ditch, along which it still runs for a considerable distance ; but at one place it 

 has broken over the wall, obliterating it for nearly two hundred feet. It is dry at 

 most seasons of the year ; and, unless much swollen by the rains, keeps the course 

 of the ditch, terminating in a deep gully, formed by the tlow of water from a 

 copious and unfailing spring. This gully is made to answer as a ditch, for the 

 space yet intervening, to the edge of the terrace. It is fifteen feet deep, by sixty 

 or seventy wide. In several other instances, this artificial change in water-courses 

 has been observed. 



The gateways of this work are six in number ; one opening into the smaller 

 enclosure to the east, two upon the table lands, one to the spring first mentioned, 

 and two others towards the creek. Two considerable springs occur within the 

 walls. It is not necessary, however, upon the hypothesis already advanced in 

 respect to this work, to suppose its ancient population wholly dependent upon these 

 sources for their supply of water ; inasmuch as it is very evident that many centu- 

 ries have not elapsed since the creek, now one hundred rods distant, washed the 

 base of the terrace upon which it stands. Indeed, until recently, and until 

 prevented by dykes above, the creek at its highest stages continued to send a 

 |)ortion of its waters along its ancient channel. 



The slight wall along the terrace bank is composed chiefly of smooth, water- 

 worn stones, taken from the creek, and cemented together by tough, clayey earth. 

 The wall of the square is wholly of clay, and its outlines may be easily traced by the 

 eye, from a distance, by its color. It appears, as do the embankments of many 

 other works, to have been slightly burned. This appearance is so marked, as to 

 induce some persons to suppose that the walls were, in certain instances, originally 

 composed of bricks partially baked, but which have in process of time lost their 

 tbrm, and subsided into a homogeneous mass. That they have in some cases been 

 subjected to the action of fire, is too obvious to admit of doubt. At the point z 

 in the lower wall of the square, stones and large masses of pebbles and earth, much 

 burned, and resembling a ferruginous conglomerate, are turned up by the plough. 

 May not this feature be accounted for by supposing the walls to have been originally 

 siu-niounted by palisades, which were destroyed by the action of fire ( Such a cause, 

 however, seems hardly adequate to produce so striking results. 



The broken table land upon which the main work extends, forms natural bastions 

 at T and .V, which have gateways opening to them. At the point marked C in the 

 embankment, a ([uantity of calcined human bones are observable. 



Such are some ot" the features of this interesting work; and if then- detail has 

 been tedious, it may be urged in extenuation of sucii minuteness, that descrip- 

 tions have hitlierto been quite too vague and general. Minute circumstances are 

 often of the first importance in arriving at correct conclusions. The comparative 

 slightness of tlie wall and the absence of a ditch, at the points possessing natural 

 detences, — the extension of the artificial defences upon the table lands overlooking 

 and conuuanding the terrace, — the facilities afforded for an abundant suj)ply of 

 water, as well as tliL' large area enclosed, with its mysterious circles and sacred 



