478 MOUND EXPLORATIONS. 



Plotting the figure carefully from these notes, and then drawing the 

 nearest possible coincident circle, we obtain results similar to those 

 obtained by the survey of the Observatory circle at ^Newark. This is 

 shown in PI. xsxvii. In this plate the solid black line of «hort chords 

 running along the middle of the wall marks the actual line of survey, 

 while the dotted line is the nearest approximate circle, the center of 

 which is at the intersection of the two designated diameters.' These 

 diameters are actually surveyed lines and relate to the line of chords. 

 The middle of that running from d to b is at the intersection, but the 

 middle of that running from a to c is about 2 feet from the intersection 

 toward c. 



As it is imxjossible to show this satisfactorily in a llgiire on the scale 

 given here, the attention of the reader is called to the following facts, 

 which he can verify iudependentlj' by making for himself the ftlat on a 

 larger scale from the notes given above relating to the High Bank 

 circle. 



(1) The chords forming the sides of the inclosed quadrilateral subtend 

 equal arcs of the surveyed line; that is to say, the distance along the 

 wall from a to b is equal to that ft-om b to c, also to that from c to d and 

 from dto a; the distance in each case beiug 830.4 feet or one-fourth of 

 the circumference according to the survey. As these chords are, respec- 

 tively 744, 746, 743, and 741 feet in length, showing an extreme varia- 

 tion of less than 3 feet from a medium and of but 3 feet from a triie 

 quadrant, we have an evidence of the close approximation to a true 

 circle. 



(2) The extreme difference between the various diameters (except at 

 the eccentric jjoint at the southeast, between the gateways) does not 

 exceed 8 feet, or a variation from the medium of 4 feet, and from that 

 of the true circle of little more than 5 feet. 



(3) A circle with a radius of 526 feet and center at the intersection 

 of the two given diameters varies at no point from the surveyed line 

 (except at the eccentric x>oint in the southeast) more than 6 feet; or, in 

 other words, both would fall on a wall only 6 feet wide. 



It is evident, therefore, that we have here a very close approximation 

 to a true circle. 



The Octagon at this jjoint (see PL xxxviii) differs from that at 

 Newark chieiiy in size and closer approximation to a square. The 

 variation from the usual form resulting from throwing the southern 

 gateway along the wall between the angles is readily accoiuited for by 

 the fact that there is here a somewhat abrupt depression which is 

 avoided by the curve given the wall. The field notes of the resurvey 

 are as follows, commencing at station 1 in the middle of the gateway 

 leading to the circle : 



' The actual scale on tlie plate (which was accidentally omitted) is 200 feet to the inch. 



