FOR DETACHED OBSERVATORIES. 533 



ail insulator. The observation would only be incorrect in case the air of 

 the distant point has a higher temperature than that of some point in 

 the course of the rubber tube, as in this ease a part of the vapor would 

 be condensed in the tube, and consequently would not reach the areom- 

 eter. But, in general, this would not be likely to occur, because, since 

 the temperature of the air usually decreases as we ascend, the existence 

 of a warm and moist stratum above a colder one will be of but excep- 

 tional occurrence. 



Before concluding, I must add a word in regard to the case where it is 

 desired to make use of a captive balloon for carrying the observatory to 

 the elevated regions of the atmosphere and supporting it there for a 

 considerable time. The apparatus I propose is perfectly adapted to this 

 purpose, and requires but two conducting wires in the cable which holds 

 the balloon captive. 



It will, of course, be objected that the balloon, losing gas rapidly, could 

 not be kept up for more than two days; but this difficulty appears to me 

 capable of being overcome in a very simple way. The two conducting 

 wires in the cable should be insulated from each other, which might be 

 simply done by inclosing each in an India-rubber tube. Xow, one of these 

 tubes may be made use of to carry a continuous current of gas into the 

 balloon to supply its losses, while through the other the air of the elevated 

 region may be drawn down by an aspirator to the hygrometric areometer. 

 Thus supplied, the balloon would float in the air until the approach of 

 a tempest — foretold by the observations — would render it necessary to 

 draw it down again. 



But there is still another objection, to which so far I can make no 

 satisfactory answer, namely, the impossibility of maintaining the cap- 

 tive balloon at the same height, especially in a region where strong 

 winds prevail. In such a case the record of the barometer is of no 

 service except to give an approximate idea of the elevation of the 

 instruments, and it will be impossible to tell whether the observed 

 barometric variations are due to differences in the height of the balloon 

 or to differences in the atmospheric pressure at the same height. 



The use of captive balloons is always attended with a serious defect, 

 namely, the limited height (3,000 feet at most) to which they can be 

 raised, since, without mentioning other difficulties, the weight of the. 

 cable will soon become so great that the balloon must be very large to 

 support it. But we can obtain meteorological observations from much 

 greater elevations (from 15,000 to 30,000 feet, for example) by means of 

 the detached observatorj' without transmission of records. By its rel- 

 atively moderate price and its light weight this apparatus would, in fact, 

 be well adapted to be suspended from a small balloon allowed to float 

 freely in the atmosphere. Supposing that of ten balloons left to them- 

 selves one or two should fall into the sea or in uninhabited places and 

 rhus be totally lost, the expense of the experiment would still be much 

 less than the ascension of a balloon large enough to carry the observers. 



