552 
Smith—Stadia Measurements . 
A glance at this table will show the actual error on each 
street expressed in meters and also fractionally; the fractional 
error on each complete measurement of the base (2,217 meters); 
and also the accumulated error of the thirteen measurements. It 
will be seen that the average error of the base line measure¬ 
ments was 1 in 2,250, and that the longer the measurement, 
the greater became the relative accuracy, owing to the compen¬ 
sation of errors. 
Thus, beginning with an error of 1 in 2,100 on the first 
measurement, the error decreases, till at the end of eighteen 
miles the accumulated error was but 1 in 10,400. The total 
time spent in measuring these eighteen miles was fourteen 
hours. 
CONCLUSION. 
A somewhat wide correspondence with experienced stadia 
engineers, indicates that at present, hardly any two engineers 
agree as to the proper time for determining the stadia interval, 
there being almost as many opinions as there are hours of the 
day. If the conclusions of this paper are correct, all are indi¬ 
vidually wrong, and yet collectively right. Determine your in¬ 
terval not at any one hour , but during many , so as to approximate 
as closely as time allows, the average conditions for the field work. 
The degree of accuracy which will result will correspond closely 
to the degree of such approximation. 
Many engineers will say that the present method is accurate 
enough, and so it is for a large class of work; but for a much 
larger class of work it is not accurate enough. The stadia 
method is such a convenient and economical one to use, that it 
would seem a step of progress to widen its field of usefulness. 
This can be done by using additional care in determining the 
interval. 
The chief objection to this added care, both in the original 
interval determination and also to any subsequent ones rendered 
necessary by decided change in field conditions, will doubtless 
be the cost of repainting the rods in the latter case and the 
cost of such added care in both cases. The first objection would 
be entirely removed by not incorporating the interval in the rod 
