LIEUT.-COL. A. R. CLARKE ON STANDARDS OE LENGTH. 
457 
If in the results we have now obtained we substitute the value of 2Ki, we get finally 
for the Pulkowa copy of the klafter, 
K'=2-07403658g±-00000041; 
and for the two lengths of the Milan copy, 
K" .3 =2*07401462^ ± -00000045, 
K'i. n =2 -07402990p + -00000051. 
In a pamphlet by M. Struve, entitled “ Vergleichungen der Wiener Masse mit mehreren 
auf der Kaiserl. russischen Hauptsternwarte zu Pulkowa befindlichen Masseinheiten,” 
1850, we find it stated that the length of the Pulkowa copy of the Austrian klafter, as 
determined by Professor Stampfer, is 
0 1 -00029±0 1 -00020 
shorter than the legal or standard Vienna klafter, the unit here being the “ line ” or 
part of the klafter, that is 2400 millionths of a yard. Hence if be the true 
length of the klafter, the Pulkowa copy at 61 0, 25 is 
K' = H4— 0-70 + 0-48, 
the unit here being, as usual, the millionth of a yard. The date of the certificate is 
April 1849. 
Professor Stampfer also compared the Milan copy of the klafter with the legal or 
standard Vienna klafter, and with this result, dated October 1856, 
K " >3 =H4— 0 ! -0058 + 0'0005 . . . 10 comparisons, 
K5. n =»+0 1 -0000±0-0004 ... 10 comparisons, 
the bar being at 61 0, 25. This, expressed in millionths of a yard, is 
K +3 =^-13-92 + 1-20, 
Ki'. n =fe- 0-00 + 0-96. 
It appears from this that the difference of the klafters on the two surfaces of the 
Milan copy, as determined at Southampton and at Vienna, are in tolerable accordance, 
being 15-28 in the one case and 13*92 in the other. 
But the discordance between the Milan and Pulkowa copies is considerable. Taking 
the length I . II for example, the difference of the two bars as determined at Southampton 
is 6 ‘ 68 , while the Vienna comparisons imply a difference of —0-70. 
But it is difficult to imagine any constant error in the observations made here, as not 
only the frequent alterations of the adjustments and daily shiftings of the bars makes 
this very improbable, but an examination of the tables of errors of the different series of 
comparisons shows no trace of such error. 
