286 Intelligence and Miscellaneous Articles. 



large spark, 5 metres or longer ; and with this experiment we can 

 account in part for the enormous length of lightning-flashes, assu- 

 ming that the filings represent the minute droplets of water 

 suspended in the atmosphere. Along the surfaces of water (in my 

 case placed in as many large glass troughs arranged in series) we 

 obtain a discharge more than a metre in length, and of equal 

 length if discharged through a large gas-flame. 



The following is the new experiment which has given occasion 

 to this Note. 



A platinum wire, 3| metres in length (or even a little longer) 

 and -giy of a millim. in diameter, being fused by the discharge, is 

 instantly changed into a beautiful corona ol incandescent globules ; 

 but if we take a shorter path in the same wire, for instance a 

 metre and a half, we observe the following curious phenomenon. 

 The moment the discharge takes place, a white spark a metre and a 

 half in length is observed in the place occupied by the wire, which 

 is rectilinear if the wire is straight, but follows the shape of the 

 wire if bent. Of course no trace is observed in the wire behind 

 the spark ; there is only produced from this long spark a little 

 smoke with a characteristic odour. 



With iron, brass, or gold wire, with a thin and very narrow ribbon 

 of steel, or of magnesium, or tinfoil, an analogous phenomenon is 

 observed. The spark becomes yellow with iron and with gold, and 

 green with copper. With these metals the smoke of the discharge 

 is more dense and abundant, but does not produce the penetrating 

 odour which platinum does. 



The formation of this spark may be explained as follows. 

 The first portions of the discharge are sufiicient to convert the 

 wire into the state of vapour; the remainder of the discharge 

 then finds a column of metallic vapour at a high temperature which 

 offers an easy path. It forms instantaneously, as it were, a 

 Geissler's tube, the sides of which are formed of the surrounding 

 cold air, full of rarefied gas, because at a high temperature. 



To test this explanation, I devised the following experiment. 

 Above about the middle of the wire A B (which is stretched between 

 two stouter ones) I place a conductor with a knob C, which is con- 

 nected with the stout wire to which A is attached. If the above 

 explanation is correct, this is what should happen. The moment the 

 discharge takes place, this should commence by traversing the wire 

 AB and volatilizing it, provided the knob is at a suitable distance 

 from the wire ; but then, instead of forming a spark from A to B 

 through the metallic vapour, it should form one simply from C to B. 

 Then the M'ire should evaporate altogether, but the large spark appear 

 only on the right of C. 



Having frequently made the experiment, I have observed that it 

 succeeds completely as prevised, and that the moment the discharge 

 passes all the wire evaporates, and only a spark passes from C to B. 

 — Bull. Acad, dei Lincei, Dec. 16, 1888. 



