260 IRREGULAR VARIATIONS OF PRESSURE. 
Another element found very important at the outset of the reductions was the non-periodic 
variation of pressure. Fortunately, several good barometers were observed sufficiently near 
this part of the line to give an approximate correction for many days. The importance of at- 
tention to this source of erroris very great in lines determined by successive differences, and 
every mode of correction and avoidance of it was employed throughout. 
The mode previously employed, or whether indeed any mode was employed, to obviate er- 
rors arising in this manner, is not given by Nicollet and others who have made determinations 
in the interior ; but the results on this line show careful attention to it to be indispensable. 
At Napoleon, for the 24th of June, a non-periodic correction of —.117 was applied to the 
mean of the barometric readings, obtained from comparison with full records at Lebanon, 
Glenwood, and Memphis, Tennessee, and at New Orleans. 
A less correction was applied at Little Rock ; at Fort Smith the correction —.027 was ap- 
plied to the mean of all the readings, and beyond this the corrections given below : 
July 28th 十 .030 
July 29th 十 .050 
July 30th 十 .080 
July 31st 十 .070 
August 1st 十 .030 
August 3d / 一 .080 
August 4th —.080 
August 5th —.040 
August 7th —.020 
August 8th......... —.060 
August 9th i —.050 
Further correction in this manner was impossible; but the more important points were de- 
termined from observations carried through two or more days, and such points were made bases 
for reference of minor camps and intermediate stations. 
The line was carried forward by reference of Camp No. 2 to the sea-level, assuming 30.050 
inches as the mean pressure at the gulf with the barometer corrected to 32?, and the air tem- 
perature at 64? ; and by subsequent reference of the principal camps to each other, correcting 
the differences to the station having the greatest number of observations, and correcting inter- 
mediate lines, determined by minor camps and single stations, proportionally, for the error of 
their terminus. 
At Albuquerque, a reference of the mean of its numerous observations was again made to 
sea-level, and the elevations eastward to Laguna Colorada were determined by differences from 
this point. The error resulting from non-periodic variation is mainly avoided in this manner, 
and the lines from these different bases agree very nearly at their terminus. 
Beyond Albuquerque, the elevations are determined by mutual and cross reference of princi- 
pal camps to each other, to Leroux's spring, Camp No. 91. At the mouth of Bill Williams’ fork 
a new base was determined, with which the elevations eastward were compared as far as Camp 
No. 91, correcting the principal camps by direct reference to the sea. The same elements were 
taken ES sea-level as before, viz: 30.050 inches as barometric height at 32?, and the air tem- 
perature at 64°. This check by direct reference to the sea, gives but slight corrections on the 
determinations by consecutive comparisons from the Colorado, and the lines agree very nearly 
at their terminus at Camp No. 91. The differences for single stations and minor camps are, 
however, sometimes considerable. "These are corrected by taking departures from each princi- 
pal camp, and correcting the elevations proportionally from their point of meeting midway. 
West of the Colorado, the line is carried forward by differences from the mean of the three 
first camps at the point of departure, otherwise as before, and its terminus at the Pacific 1s with- 
out important error. 
