ON THE KEW OBSERVATORY. 139 



ously) they would not collapse for five or ten minutes ; and after these 

 supporting wires had been charged, the straws continued slowly to increase 

 in divergence during an hour or more sometimes*. This was proved by 

 comparing them with natural undried straws. 



I therefore tried many experiments upon straws gilded in various ways, 

 but even these did not appear to afford such complete freedom from the 

 above-mentioned defect as was required. 



Passing over many details (tedious but not instructive perhaps), I will now 

 describe shortly the apparatus, &c. which I call my registering (or night) 

 electrometers, the results of many trials. 



Three receivers, 5\ inches high and 4 inches diameter, were fitted air-tight 

 to ground brass plates at their bases and necks. In these tlie electrometers, 

 supported as before, could be charged by means of moveable and insulated 

 wires, without interfering with the air-tightness of the receivers, and they 

 contained a rather larger quantity of the chloride. 



In lieu of the straw electrometers recourse was had to a modification of ray 

 old instruments of fine wires f, very accurately straightened, and in order to 

 prevent as much as possible dissipation, without materially increasing their 

 weight, minute globules of gum-arabic were applied at their extremities, 

 whilst they were electrified for the occasion. 



A scale which could be read in terms of the standard voltaic electro- 

 meter was thus prepared : a slip of ivory was properly cut (to the radius of 

 the wires) and fixed at one extremity of a ruler one foot long ; the other end 

 of the ruler carried a sight-piece, like C (fig. 5) ; this ruler was held in the 

 hand, and the scale-end made to touch the receivers when used. The gradu- 

 ation was easily effected (not in exactly equal divisions of course) by mark- 

 ing on the scale (before engraving) the degrees of divergence of the wires, 

 as seen through the sight, which corresponded with the divergences of the 

 ordinary standard electrometer, placed in good conducting contact with these 

 wire electrometers. 



In order to compare these registers with each other and with the standard, 

 the moveable insulated wires and the standard were placed in contact with 

 an insulated horizontal wire, so that they might be all charged simultaneously ; 

 then their contact with the horizontal wire was suddenly broken, and at the 

 same moment the contact of the moveable wires with the electrometric wires. 



The following Table on the next page exhibits a specimen of the perform- 

 ance of these registers, called C, D, and E. 



If a quarter of a degree of this scale be added for every hour which may 

 elapse betwen the time at which any one of these registers was charged, and 

 the time at which it is read, up to the 45th degree, we may perhaps be tole- 

 rably sure of knowing what the charge was within something less than a 

 tenth of a degree (and this is a quantity which cannot be appreciated by any 

 observation of a voltaic electrometer). 



After the 45th degree (upwards) the loss per hour begins to increase in a 

 much more rapid rate, and after the 90th uncertainty prevails, because spirt- 

 ings "spruzamenti" begin, as Volta found in his electrometers. 



However, the nightly charges of our conductor (after 10 p.m.) in serene 

 weather seldom exceed 45 degrees. 



New experiments must be made on this subject. In the mean time we 

 apply these instruments to the purpose intended, and hope to improve our 

 journal thereby. The particular mode of application and a more detailed 



* We have observed the same land of effect (in much'smaller degree) in the electrometers 

 (exposed to the open air) in the observatory in very dry weather, 

 t Vide Descriptions of an Electrical Telegraph, &c.,"l823, p. 33. 



