METEOROLOGY. 509 



ratus for registration works with perfect correctness, and without injuri- 

 ous reflex action upon the accuracy of the barometer proper; it is sur- 

 prising that many pieces of self-registering apparatus are still used in 

 which this i mportant condition is neglected. 



The probable error of the self- registering Hasler barometer (Wild's), 

 at St. Petersburg, has only lately and after several years of experience 

 been brought down from 0.23 in 1871 to 0.085 in 1878, and this by virtue 

 of many laborious computations and reductions. The probable error of 

 the Schreiber baro-thermograph is still about 0.21. In general the per- 

 formance of Sprung's balance barograph — with sliding rod and sliding 

 weight, and invariable position of barometer tube — has proved thoroughly 

 satisfactory, and proves it to be an exception to the statement of Dr. 

 Schreiber, in December, 1881 (Z. O. G. M., xvi, p. 500), that all other sys- 

 tems of registration, except those depending on the motions of the 

 barometer, are mere methods of interpolation. The insulation of the 

 recording and the measuring portions of the apparatus as accomplished 

 by Sprung has been applied by him to thermometers, rain-gauges, &c. 

 [Was it not first applied to meteorological apparatus by G. W. Hough 

 at Albany, in 1861, whose printing barograph and thermograph still 

 serve as the typical apparatus, though by no means giving such accu- 

 rate results as those of Wild, Schreiber, and Sprung.] (Z. O. G. M., 

 xvn, p. 46.) 



Crova has described a modification of Eegnault's dew-point hygrom- 

 eter, which promises to give highly accurate results. Crova draws the 

 air to be examined into the interior of a highly-polished tube, whose 

 outer surface is cooled by evaporating ether or other processes, and 

 whose temperature can be determined by a thermometer immersed 

 therein. The dew is deposited on the interior of the polished tube, 

 and its appearance and disappearance can be accurately detected. 

 (Z. 0. G. M., xvn, p. 374.) 



Mignon and Kenard describe a condensation hygrometer so arranged 

 that all the vapor present in the air is precipitated and is collected in 

 a special vessel for measuremeut. Fonvielle suggests that this instru- 

 ment is especially adapted to collect for investigation the dust parti- 

 cles swimming in the atmosphere, and that interesting results would 

 be obtained by such apparatus on balloon voyages among and above 

 the clouds. {Z. O. G. M., xvn, p. 375.) 



Stellung and Wild have devised a form of evaporimeter which floats 

 in a river or basin of water, thereby preserving the evaporating water 

 at the same temperature and as nearly as possible under the same con- 

 ditions as the main body of water in the river itself. The quantity of 

 water evaporated is accurately measured to one-twentieth of a millime- 

 ter. When rain falls and fills the evaporating dish above a certain level 

 the surplus is carried into the interior of the apparatus and can subse- 

 quently be poured away ; if rainfall is too heavy, in consequence of 



