GREEK HORIZONTAL CURVES IN MAISON CARREE AT NIMES. 575 



would naturally ap[>ear equal, the differeuce of size indicates to the 



eye a deflection in the line of the building, or, in other words, the 



spectator appears, in so far, to be nearer to the large capital than he is 



to the smaller one. 



I. 



We will now specify some of the maximum cases of irregularity 

 according to the measurements of Penrose, which are given in feet and 

 decimals of a foot. The curve of the Parthenon entablature on the 

 flanks, about 228 feet in length, is 0.307 (decimals of a foot). At the 

 sides of the building it is 0.171 in something over 100 feet. (The flattest 

 curve in Greek art is the entasis of the Ercchtheium columns, which is 

 0.0195 in 21 feet.) The Parthenon columns lean 0.22S in 30 feet, an 

 inclination of one unit in 150 units. In other words, as the columns 

 lean to the center of the building they would, if sufficiently prolonged 

 in height, meet at a height of 5,856 feet above the level of the pavement. 

 The ant?B have a forward lean of 1 unit in 82, and the acroteria and 

 the antefixte have a forward lean of 1 in 25. A maximum devia- 

 tion in spacings of the metopes is 0.325; the measurements of these 

 spaces being 4 feet and over. The maximum deviation in intercolumnar 

 spacings is over 2 feet, but this amount of deviation is only found at 

 the angles where the columns next the corner are that much nearer the 

 corner. At these points the spacings narrow from 8 feet and a decimal 

 to G feet and a decimal. Aside from the angle columns the maximum 

 intercolumnar deviation on the north flank is 0.136, in measurements 

 which are all over 8 feet with decimal variations. A maximum devia- 

 tion in the diameters of columns (of corresponding lines and sizes) is 

 0,23 in measurements giving diameters of 5 feet and a decimal. A max- 

 imum deviation in size of the capitals is 0.312 in measurements of 6 feet 

 and a decimal. 



These instances will give an idea of the amount of actual irregulari- 

 ties according to actual measurement, and we will add that instances of 

 two adjacent measurements being equal are almost absolutely unknown. 

 We can occasionally trace some scheme in the variations by comparing 

 two halves of one end, or one side of the building, but when such a 

 scheme appears it does not repeat itself in any two different series of 

 measurements on one side or one end of the building. For instance, 

 in the metopes of the east front the spaces widen from the angles 

 toward the center, but this does not hold of the intercolumnar spacings, 

 where the only perceptible scheme is that which makes the corner 

 intercolumuiations narrower by 2 feet and a fraction. 



That all these remarkable deflections and irregularities were intended 

 has been proven by masonry measurements and masonry observations. 

 Penrose places the maximum deviation, due to error or carelessness in 

 the Parthenon masonry, at one-flftieth of an inch. The two ends of the 

 building are of equal width within that fraction. The difference of 0.02 

 (incli decimal) in 101 feet points out 'Hhe degree of error which may 



