OCR Text |
Show 51 that these changes are not due to defective graduation, since I have selected observations where each cistern- barometer reads nearly the same at two or more of the campa given in each series. Nor were the adjusting- Bcrews of the aneroids touched, as shown y the records of these instruments. In fact, many of the aneroids are so constructed that these screws cannot be reached without the trouble of unscrewing and removing a plate upon the back of the instrument, which the ordinary observer is not likely to take without positive orders to that effect. The observers were young gentlemen of more than average education, and the results are undoubtedly as good as may be expected from observers who have been instructed in their duties and conscientiously perform them. My object is to show in this discussion what is actually attained in practice with proper care, and not to select observations with a view to call attention to undue accuracy or to the want of it. - r_- From these observations, it can be seen at a glance that it would be inadvisable ~ to rely upon these instruments for any length of time for absolute altitudes without checking their indices by comparisons with those of the mercurial barometer, since the results would not only be affected by the ordinary sources of error due barometric determinations, but also those due changes in the zero- points of the aneroids themselves. Besides this source of error, aneroids are not always perfectly graduated, though the best English aneroids are well and carefully marked; out, from the nature of the case, only a few points of the scale are in the first place accurately determined, and the intermediate divisions are measured off, and afterward, if required, the instruments are tested under the air- pump through the entire scale. This should always be done, and a table of scale- errors accompany each instrument. In reducing our work, I have found it impracticable to correct for scale- errors, or even to determine their amount. Comparisons have been made at many points varying in altitude between 4,500 and 14,000 feet; but these comparisons are worthless for determining errors of scale, since it is in most oases ( save where the index- error of the instrument remains constant) impossible to separate these errors from changes in the position of the zero- point. La individual and rare oases, it may be practicable to determine these errors by comparison with a good cistern at various altitudes; but in order that this method be successful in general, it is necessary to pass through the range of altitude in a short time, as during the ascent of a mountain, and to make the comparisons very frequently, and in any case to transport the aneroids with the greatest care, avoiding all jolts and jars, as we would with a chronometer or other delicate piece of mechanism. This, at least, is the conclusion to which I have come after the examination of the records of some eighteen aneroids during an entire season of four months, where a rigid system of comparison at various altitudes was enforced. During the past field- season, wherever hourly observations were taken, all aneroids which could be assembled were compared at each hour with the cisterns^ which gave aeries of comparisons extending over several days and through quite wide ranges of temperature, when the instruments were undisturbed, and therefore their index- errors presumably constant. I have examined all these comparisons, and, as far as they furnish data, nave determined, as well as I could, the effects of temperature upon many aneroids, in practice. The instruments were made by Pike, Ewing, and Casella, principally by the latter; of pocket size; graduated to 0.05 inch, but susceptible of being read to hundredths of an inch by estimating the smaller subdivisions. I have first taken the difference between the indications of each aneroid and the corresponding readings of the barometer at 32°, to eliminate the movement in the barometric column itself, and then grouped the temperatures and the corresponding differences according t<> the ascending scale of temperature, and taken the mean of each group. The result has been an ascending scale of temperatures and a corresponding scale of aneroid differences, which I have plotted, together with the curve, Bhowing the effect of temperature upon the height of the mercurial column at the altitude in which these comparisons were made. In a great number of cases I have detected no well- marked law of variation, especially among the aneroids made by Casella. Certainly, in these cases the want of compensation for ordinary ranges of temperature affects the readings by a quantity less than the errors of observation. Others show their want of compensation even in a short series of comparisons, where errors of observation are not sufficiently eliminated to give smooth curves. Of these, some are over- compensated, or the movement produced by heat is the inverse of that produced in the mercurial barometer; others are under- compensated, or the movement is in the same direction as the temperature. It is probable that no aneroid is perfectly compensated; but it is certain that in a great number, perhaps in the majority of instruments in use by us, this desideratum is so nearly approximated that for all intents and purposes it is practically attained for ordinary ranges of temperature. But, before we accept an aneroid as reliable in this respect, we should first satisfy ourselves by examination that we have not an - exceptionally poor one. I give herewith the plotted curves relating to two of the aneroids examined, plotted upon a very large scale, greatly exaggerating the aneroid curve, which, in the two oases given, is not of great amplitude. Of all the aneroids examined, but one shows a greater range of variation than J& e two given upon . this |