FOR DETACHED OBSERVATORIES. 523 



For the determination of the hygrometric condition of the air those 

 instruments alone can be used which depend on the expansion and con- 

 traction of certain animal substances, and on the communication of these 

 movements to an index. There will still remain to be provided for, the 

 registration of the direction of the wind by means of the wind-vane, the 

 velocity of the wind by means of Robinson's anemometer, and finally the 

 amount of rainfall. All of these may be simultaneously recorded by the 

 arrangement I am about to describe. 



I. 



DETACHED OBSERVATORIES WITHOUT TRANSMISSION OF RECORDS. 



The slight interest which meteorologists feel in observations the re- 

 sults of which they do not see before the end of a certain period is, with- 

 out doubt, the reason why observatories of this kind have not been long 

 in use. Their erection would, in fact, present no difficulty, and they 

 would have the advantage that a number could be kept up with a small 

 force of assistants — at least if they were not too widely separated. Each 

 would require to be visited but about once in every ten or fifteen days, 

 in order to wind the clock, change the blanks, and inspect the instru- 

 ments. They could be located anywhere, but, preferably, away from 

 habitations ; the first cost of erection would be very slight, and the only 

 precautions necessary would be to protect the instruments from injury by 

 rain and wind or malicious persons. The observations — automatically 

 and continuously recorded according to an established system — would 

 offer to the meteorologist a guarantee of accuracy which no human 

 records possess, since to the latter may always be objected u humanum 

 est err a re.'''' 



The following arrangement appears to me completely to satisfy the 

 proposed conditions : 



An aneroid barometer, a metallic thermometer, and a hair hygrometer 

 are placed side by side on a board so that the pivots of their indexes 

 are in a straight line. These indexes are of equal length and move in the 

 same plane, but so that the greatest sweep of each will not describe an 

 arc of over 120° at most. A little way above these indexes is placed a 

 sheet of paper — kept in uniform motion by clockwork — on which there 

 have been previously drawn lines marking out spaces corresponding to 

 the arc described by the tip of each index. A pencil attached to this 

 tip will trace a curve on the paper which will indicate at the same time 

 the successive readings of the instruments and the corresponding time. 

 But the friction of a pencil on the paper would be too great for such 

 delicate instruments, and would necessarily falsify the indications. It 

 would be infinitely better to have recourse to the method generally 

 adopted at present in physical laboratories for the registration of delicate 

 motions, which consists in the use of glazed white paper previously black- 

 ened above a resinous flame, on which a feather, attached to the tip of 



