benchmark and at least one other existing benchmark and must show there is no change in their relative

elevations.

(2) All first-order lines are divided into 1- to 2-kilometer sections. Each section must be run

forward and backward. The two runnings of a section must not differ by more than ...3 millimeters √K,

where K is the length of the section in kilometers. On all sections that are 0.25 kilometers or less in

length and which require two or more setups in each direction, a discrepancy of not more than 2.0

millimeters between the backward and forward measurement is considered a satisfactory check. In the

case of a single-setup section, the checks between forward and backward runnings seldom exceed 1.0

millimeters, averaging 0.6 millimeters or less.

(3) When additional runs are made due to excessive divergence, the indiscriminate mean of all

measured differences in elevation for that section is computed, excluding obvious blunders. If any

measurement of the difference in elevation for that section is more than ...6.0 millimeters *√K *from the

indiscriminate mean, that observation is rejected. No rejection should be made on because of a residual

smaller than ...6.0 millimeters *√K*, unless there is some good reason for suspecting an error in that

particular measurement. In such cases, the reason for the rejection must be stated in the record. After

the rejection has been completed, the mean of all *remaining forward measurements *is computed and

compared with the mean of all *remaining backward measurements *to determine if the required accuracy

has been obtained. The *mean *of the *means *is then the final value.

(4) The method and equipment used in first-order levels are designed to yield a maximum of ...1

millimeter for the probable accidental error and ...0.2 millimeter for the probable systematic error in a

distance of 1 kilometer.

b. Second-Order Leveling. The uses of second-order leveling are also quite important, so the

criteria are only slightly less strict than those for first-order leveling. Second-order leveling is used to

subdivide nets of first-order leveling and to provide basic control for the extension of levels of the same

or lower accuracy. This order leveling is used to provide data for mapping projects, local surveys, and

special projects, which include the positioning of radar equipment and stellar camera pads. Second-

order leveling is also used for initial missile site surveys. Second-order levels are divided into two

classes, Class I and Class II.

(1) Class I is used in remote areas where the line must be longer than 40 kilometers due to the

unavailability of routes, for the development of additional or higher order networks, and for spur lines.

All lines must start on previously established benchmarks of first or second order. New levels must be

run between the starting benchmark and at least one other existing benchmark to prove that they have

not changed their relative elevations. Failure to check within the limit of ...8.4 millimeters *√K *(where K

is the distance between benchmarks in kilometers) may indicate that at least one or both of the

benchmarks must be tied in to prove the starting elevation. All Class I lines are divided into 1- to 2-

kilometer sections which are run both forward and backward. The discrepancy between these runs must

not exceed ...8.4 millimeters *√K*, where *K *is the length of the section in kilometers. When