Oct. 2, 1879] 



NATURE 



545 



as fast as he could, and reached a solid standing-ground, but not 

 before perceiving that the guide, who walked behind him, had 

 sunk up to the middle in the sand, and indeed soon disappeared. 

 All the slope sank with increasing swiftness, swelling out at the 

 bottom and opening out to give forth a stream of mud of at least 

 ten metres in height, which shot like a wild torrent across the 

 plateau, and precipitated itself towards the Val de Bove. Im- 

 mediately all the moving sand, to more than thirty metres in 

 height, was diluted and drawn along, as were also the stones and 

 blocks it contained, and which rolled over each other pell-mell. 

 Fortunately they succeeded in extricating the guide, covered with 

 mud, torn and bruised. 



Seeking the origin of these eruptions of mud, M. de Saussure 

 observes first that they take place in the neighbourhood of the 

 central cone of Etna, and that they all escape from the foot of 

 the cone and the very 1; cliued slopes which support it. They 

 proceed then exclusively from the water stored in the beds of 

 ashes. These causes seem to account for this mass of water. 



1 . The hot vapours. The cone seems almost exclusively com- 

 posed of ashes and movable stones, and during the eraption an 

 enormous mass of vapour must have traversed it. There are, 

 moreover, numerous fumeroles in the upper part. These vapours 

 have been condensed, all the more readily that they have encrusted 

 the mantle of snow which envelops the summit of the mountain. 



2. The melting of this snow, determined less by the fall of 

 ashes on the surface than by the heat which radiates from the 

 interior and the vapours which traverse the porous mass of the 

 cone. 3. The abundant vapours falling from the cloud of 

 vapours projected by the crater. The two first causes are the 

 most important ; the last has most especially given rise to sur- 

 face streams, the traces of which are apparent. The water 

 gradually collecting in excess in the lower parts of the cone, 

 ends by forming these enormous deposits, which, at a given 

 moment, yielding to their weight, cause an eruption to the out- 

 side, turning over and diluting the beds of ashes which oppose 

 its exit. 



In a communication to us M. de Saussure gives an analysis of 

 the little mud-heaps which have projected through the snow. 

 They are composed of a grey powder formed of pozzolana 

 (spongy amorphous silicates), white, grey, and black (augite), 

 mixed with (//) amorphous sulphur (yellow and orange coloured) ; 

 (c) calcium sulphate (hydrated) in numerous well-defined small 

 crystals; (d) sodium and potassium sulphate; (e) calcium and 

 potassium chloride ; (f) perchloride of iron and oxychloride of 

 iron ; (g) indication of copper salts. 



The crystals are all of alabaster, and so wet that' they can 

 only have been formed after the eruption of the mud in drying. 

 There is no crystallised silica nor feldspar. The sulphur in the 

 mud is curious to observe. It must have condensed out of the 

 eruptive vapours which formed the water to make the mud by 

 mixing with old ashes under the snow, and by their expansion 

 driven small quantities of that mud through the snow. This 

 mud, when heated, evaporates sulphur, sulphurous and sulphuric 

 acid, some perchloride of iron, and some hydrochloric acid. 



THE ACTION OF HEAT IN VACUO 

 METALS'" 



ON 



TN the course of my experiments on electric lighting I have 

 developed some striking phenomena arising from the heating 

 of metals by flames and by the electric current, especially wires 

 of platinum, and platinum alloyed with iridium. These experi- 

 ments are in progress. 



The first fact observed was that platinum lost weight when 

 heated in a flame of hydrogen, that the metal coloured the flame 

 green, and that these two results continued until the whole of 

 ihe platinum in contact with the flame had disappeared. A 

 ]ilatinum wire four-thousandths of an inch in diameter, and 

 weighing 306 mgrms., was bunched together and suspended in 

 a hydrogen flame. It lost weight at the rate of a fraction less 

 than I mgrm. per hour as long as it was suspended in the flame. 

 When a platinum wire is .stretched between two clamping posts, 

 and arranged to pass through a hydrogen flame, it is coloured a 

 light green ; but when the temperature of the wire is raised above 

 that of the flame, by jiassing a current through it, the flame is 

 coloured a deep green. To ascertain the diminution in the 

 weight of a platinum wire when heated by the electric current, 

 I placed between two clamping posts a wire five-thousandths of 



' A Paper rend by Mr. T. A. Edison b*forc the Americin Association for 

 the Advancement of Science ; Saratoga Meeting. 



an inch in diameter, and weighing 266 mgrms. This wire, after 

 it was brought to incandescence for twenty minutes by the current, 

 lost I mgrm. The same wire was then raised to incandescence ; 

 for twenty minutes it gave a loss of 3 mgrms. Afterwards it 

 was kept incandescent for one hour and ten minutes, at which 

 time it weighed 258 mgrms. — a total loss of 8 mgrms. Another 

 wire, weighing 343 mgrms., was kept moderately incandescent 

 for nine consecutive hours, after which it weighed 301 mgrms., 

 showing a total loss of 42 mgrms. A platinum wire twenty- 

 thousandths of an inch in diameter was wound in the form of 

 a spiral one-eighth of an inch in diameter and one-half an inch 

 in length. The two ends of the spiral were secured to clamping 

 posts, and the whole apparatus was covered with a glass shade 

 2i inches in diameter and 3 inches high. Upon bringing the 

 spiral to incandescence for twenty minutes that part of the globe 

 in line with the sides of the spiral became slightly darkened ; 

 in five hours the deposit became so thick that the incan- 

 descent spiral could not be seen through the deposit. This 

 film, which was most perfect, consisted of platinum, and I 

 have no doubt but that large plates of glass might be coated 

 economically by placing them on each side of a large sheet 

 of platinum, kept incandescent by the electric current. This 

 loss in weight, together with the deposit upon the glass, pre- 

 sented a very serious obstacle to the use of metallic wires 

 for giving light by incandescence, but this was easily sur- 

 mounted after the cause was ascertained. I coated the wire 

 forming the spiral with the oxide of magnesium, by dusting upon 

 it finely powdered acetate of magnesium : while incandescent 

 the salt was decomposed by the heat, and there remained a 

 strongly adherent coating of the oxide. This spiral so coated 

 was covered with a glass shade, and brought to incandescence 

 for several minutes ; but instead of a deposit of platinum upon 

 the glass, there was a deposit of the oxide of magnesia. From 

 this and other experiments I became convinced that this effect 

 was due to the washing action of the air upon the spiral ; that 

 the loss of weight in and the coloration of the hydrogen flame 

 were also due to the wearing away of the surface of the platina 

 to the attrition produced by the impact of the stream of gases 

 upon the highly incandescent surface, and not to volatilisation, 

 as commonly understood ; and I venture to say, although I have 

 not tried the experiment, that metallic sodium cannot be vola- 

 tilised in high vacua by the heat derived from incandescent 

 platinum ; any effect that may be produced will be due to the 

 \^ashing action of the residual air. After the experiment last 

 described I placed a spiral of platinum in the receiver of a 

 common air-pump, and arranged it in such a manner that the 

 current could pass through it, while the receiver was exhausted. 

 At a pressure of 2 millimetres the spiral was kept at incan- 

 descence for two hours before the deposit was sufficient to 

 become visible. In another experiment, at a higher exhaustion, 

 it required five hours before a deposit became visible. In a 

 sealed glass bulb, exhausted by a Sprengel pump to a point where 

 a quarter of an inch spark from an induction-coil would not pass 

 between points i millimetre apart, was placed a spiral, the 

 connecting wires passing through the glass. This spiral has 

 been kept at the most dazzling incandescence for hours without 

 the slighte^t deposit becoming visible. 



I will now describe other and far more important phenomena 

 observed in my experiments. If a short length of platinum wiu 

 one-thousandth of an inch in diameter be held in the flame of a 

 Bunsen burner, at some part it %vill fuse, and a piece of the wire 

 will be bent at an angle by the action of the globule of melted 

 platinum ; in some cases there are several globules formed 

 simultaneously, and the wire assumes a zigzag shape. With a 

 wire four- thousandths of an inch in diameter this effect does not 

 take place, as the temperature cannot be raised; to equal that of 

 the smaller wire, owing to the increased radiating .surface and 

 mass. After heating if the wire be examined under a microscope, 

 that part of the surface which has been incandescent will be 

 found covered with innumerable cracks. If the wire be placed 

 between clamping posts, and heated to incandescence for twenty 

 minutes, by the passage of an electric current, the cracks will;be 

 so enlarged as to be seen with the naked eye ; the wire, under 

 the microscope, presents a shrunken appearance, and is full of 

 deep cracks. If the current is continued for several hours these 

 effects will so increase that the wire will fall to pieces. This 

 disintegration has been noticed in platina long subjected to 

 tile action of a flame by Prof. John W. Draper. The failure 

 of the process of lighting invented by the French chemist Tessie 

 du Motay, who raised sheets of platinum to incandescence by 



