DETERMINATION OF ALTITUDES BY BAROMETER. 117 



army stationed at these posts. They agreed very well in all important abnormal changes with 

 Dr. Hammond s Fort Reading observations, and with those made at permanent camps on our 

 route; the only difference being that the oscillations at San Diego were not so great as those at 

 Fort Reading and Benicia. Hence, as we were never more than four hundred and fifty miles in 

 a direct line from Fort Reading, it is fair to suppose that the abnormal oscillation over the 

 whole region traversed during our survey, was practically the same as that measured by the 

 stationary barometer at the fort. 



The method taken to form a table of corrections for abnormal oscillations was this : all the 

 observations taken at Fort Reading, San Diego, and Benicia, during the time that we were in 

 the field, were reduced to what they would have been, had the temperature of the mercury been 

 32 Fahr. They were then corrected for the horary variation ; those at Fort Reading by the 

 tables deduced from our own observations ; those at Benicia by one constructed by Lieutenant 

 W. P. Trowbridge, Corps of Engineers, from a set of observations taken there by the Medical 

 Department of the army ; and those at San Diego by one kindly furnished me by Lieutenant J. 

 G. Parke, United States Topographical Engineers, constructed from observations taken on his 

 recent survey in that vicinity. The observations thus corrected were plotted, forming curves 

 which represented the abnormal oscillations alone at the different places, and which, as already 

 stated, were found to exhibit a remarkable correspondence. As both San Diego and Benicia 

 were south, and the country surveyed by us north, of Fort Reading, I considered the curve 

 constructed from observations taken there preferable to a mean of the three. I, however, used 

 the other two to detect errors of observation, and sometimes, when the surveying party was at 

 a considerable distance from Fort Reading, to determine the approximate velocity of a storm. 

 Observations were taken at 6 and 9 p. m. and 6 a. m. at our camps, and, when we remained 

 stationary, at many other hours during the day. These observations, corrected as above stated 

 also furnished an excellent check, by showing the direction of the abnormal curve where we 

 were. It was found to agree remarkably with that at the permanent stations. I then made a 

 laborious examination of the Fort Reading curve, carefully correcting it by the above mentioned 

 checks. Three times, where the observations had been interrupted for a few days, the blank 

 was filled by reference to the other curves. The resulting curve represented the abnormal 

 oscillations affecting the observations on our route, and to form a table of corrections, it only 

 remained to find its mean reading, and to take the difference between this and each of its three- 

 hourly readings, affecting the result with the positive sign when the mean was the greater, and 

 with the negative when it was the less. The values for intermediate hours were found by 

 interpolation. The correction from this table, applied with its sign to any observation taken 

 at the corresponding day and hour on the survey, eliminated the error arising from abnormal 

 oscillation, by reducing it to the mean reading for the period for which the table was con 

 structed. The importance of this correction may be seen from the fact that several times it 

 affected the computed height of a station between two and three hundred feet, and, as it some 

 times increased and sometimes diminished it, relative errors, amounting to between five and 

 six hundred feet, would have resulted had it been neglected. 



METHOD OF COMPUTATION. 



After undergoing the four corrections above mentioned, the observations were ready for com 

 putation. The tables of Professor Elias Loomis were used for this purpose, the calculation 

 being somewhat shortened by the method of preparing the observations. These tables are 



