1834.] On the Measurement of the Ilahy Gfuz. 361 



which give 1127 slabs, equal to feet 1501 00.8 in. of the tape, or adding 

 its error, to 18036.8 inches, and the Ilahy guz, 32.012 inches. 



To this length, however, it appears to me that a correction for the 

 mortar should be applied, as it is more probable that the marble slabs 

 were cut to half a guz, than laid down by that measure ; and from 

 various observations, I assume this quantity at 0.03 inch between each 

 slab, or 0.06 inch, for the guz, leaving 31.952 inches for the deduced 

 length. I am the more satisfied that this correction is necessary, from 

 having observed in several places, larger slabs of marble laid down, with 

 grooves cut in them to match the junctions of the small slabs, and at 

 the end of all of which a space had been left and filled up with mortar, 

 equivalent to the accumulated excess of length caused by the mortar 

 between the slabs of half a guz, and proportional to their nurnbers. 

 I measured some of the larger slabs, and found their lengths as follows : 



guz deduced from each stone 

 in. in. 



A stone of four guz, 128.4 32.2 



A stone of two guz, 64.0 32.0 



A square stone of 1 guz, 31.8 31.8 



Another, 31.8 31.8 



the average of the whole giving exactly 32.0 in the Ilahy guz. 



Assuming then 32 inches to be a very near approximation to the 

 real Ilahy guz, it is worthy of observation that almost every principal 

 apartment, or part of the building, has been planned in an integral 

 number of guz : thus each face of the inside of the principal octagon 

 under the central dome is 24 feet = 288 inches, which is exactly equal 

 to nine guz of 32 inches, and the north-east boorj on the river terrace is 

 16 feet in diameter = 192 inches = six guz of 32 inches. I am there- 

 fore of opinion, that the average length of the Ilahy guz was, at the 

 time of the building of the Taj, as nearly as can be now ascertained, a 

 minute fraction below 32 English inches. 



The two first data, chosen by Mr. Halhed, viz. the measure of an 

 average of barley-corns, or of Musoori pice, appear to me less liable 

 than most others to error, and the mean of these is 31.93 inches. It 

 is probable that the length 2460 barley corns, or 400 pice, divided by 

 10, would give a still nearer approximation. 



[We think the argument deduced from the measurement of the apartments 

 being integrals of the guz to be conclusive, for it is the native mode of laying 

 down the. ground plan of a building always to divide the paper off into squares of 

 some unit of length: — See the plan of the Visvesvur Shiwala in Prinsep's Illus- 

 trations of Benares, first series. In that building the unit was also one guz, but 

 it was the mairrUtree guz, of about 26 inches, still commonly employed in the 

 town. From similar measurements of other ancient buildings, Muhammedan and 

 Hindu, might thus be obtained with tolerable accuracy the value of linear mea- 

 sures of different periods. — En.J 

 2 A 



