An Experimental Study of Field Methods. 
551 
quire into the method of interval determination. The engineer 
in charge gives out the following facts: The interval was de¬ 
termined by a single observer, and made to depend on ten ob¬ 
servations, taken at a certain half hour of a July day, over 
ground, and in weather, that differed widely from average con¬ 
ditions met during the three years of field work. Assuming 
that the engineer followed his present advice and determined his 
interval in midday, the explanation for the excessive readings is 
plain to be seen. During summer months rod readings be¬ 
come a minimum at about 11 o’clock, so that in the common 
stadia equation D — KS 
K, 
) 
will be too large, since S, 
the rod intercept, is too small for work before 9.30 and after 2.30. 
The inevitable result of using too large a value of K (the inter¬ 
val) would be the excessive readings to which this work was sub¬ 
ject. If the interval had been determined at some morning hour 
the errors would have been of opposite sign, as has already been 
seen in the discussion of my own work. Other examples could 
be given. 
When a survey continues through widely different conditions, 
such as summer and winter, it would be wise to redetermine 
the interval at each decided change in season, taking care that 
the conditions of the determination may approximate those to 
be met in the field. For example, if the survey is to be over 
ordinary soil, do not determine interval on a stone curbing; if 
in the field, the rodman is not to be aided by a plumb line, do 
not suffer such aid in the determination; and above all, let each 
man who is to observe with the transit, determine an interval 
himself for his own use, for in this way you will eliminate the 
systematic errors due to his personal equation. 
PROOF THAT ACCUMULATIVE ERRORS CAN BE PREVENTED. 
In order to discover to what extent errors would accumulate 
with the adopted interval, while making a long stadia measure¬ 
ment, under actual field conditions, thirteen independent stadia 
measurements of the base line were made under the conditions 
of time, weather, and length of sight recorded in table 3. 
