68 THE UNITED STATES GEOLOGICAL SURVEY. [bull. 227. 
coated with hot asphalt, and over the top of the post is riveted a 
bronze tablet similar to that first described. These posts are set in the 
earth with only one foot of their length projecting above the surface, 
and wherever possible a flat stone is placed at the bottom as a suitable 
rest. 
In order that these bench marks should be so marked as to fulfill 
the apparent intention of Congress and supply the popular demand by 
showing approximate elevations above mean sea level, it was decided 
that they should not be given ordinary serial numbers, but that the 
approximate elevation to the nearest foot should be stamped upon 
them, and these figures should constitute their numbering. It was 
soon found, however, that the assumed elevation above sea level as 
determined for a given point might be changed during a season or 
within a few seasons, and that this might change b} T several feet the 
relative elevations stamped upon two bench marks adjacent one to the 
other. To satisfy both conditions, therefore — the popular demand for 
sea-level heights and the scientific demand for exact differences between 
adjacent bench marks — the following method was adopted: 
The elevation above mean sea level to the nearest foot, as deter- 
mined at the time of running the levels, is stamped upon the bench 
mark, and, in addition, an abbreviation of the name of the central 
datum point. Thus, for an elevation depending on North Creek, the 
mark would be, say, "1728 N. C." on tablet or iron post. This datum 
being subsequently found to be in error about 5 feet, later bench marks 
established in that neighborhood are marked as though resting upon 
a different datum, that of Albany, from which the check was obtained, 
and the next bench mark to that above cited may be stamped "1810 A." 
An engineer running between these two and finding them out accord- 
ing to his levels by 5 feet, will at once realize that the letters "N. C." 
or the letter "A." carries some special meaning, and will doubtless 
write to the office of the Geological Survey to inquire concerning it, 
when the desired information will be furnished. 
The primary leveling is done by first-class methods, but not of such 
quality as would classify it as precise leveling. In order to check and 
eliminate the errors incident to such work, lines of levels run within 
an area under topographic survey are so arranged as to close back on 
themselves in circuits or polygonal figures, these lines being run by a 
single levelman and rodman; the closure of various circuits, then, 
leads at once to the detection of any gross error. Unusually long or 
important lines are run b} 7 a levelman and two rodmen in such manner 
as to make a duplicate rodded line. 
In order to establish as nearly uniform grade of accuracy for the 
work as practicable, general instructions have been issued to the level- 
men, and are printed in the "Instructions governing the topographic 
branch." The allowable limit for closure error for good work is set 
