( 467 ) 



temperature was from 10° to 'M)° higiior than a centimeter heluvv 

 the plane of demarcation ; thus in one of the experiments e.g. 1H0° 

 was found for the temperature of the upper laver, and 130° foj' 

 that of the k)wer Layer, l)ut as was said rhis difference can also 

 amount to 10°, tliis depends on the method of experimenting', hi 

 any case, however, a great difference of temperature is found, and 

 if this difference of temperature is maintained, the two hiyers can 

 be made to contijiue at pleasure. 



It is clear that the found diffei'ence of temperature perfect I)- 

 explains the distinctness of the [)henomenon, which seems lo come 

 about chiefl}' in the following ^^■ay. 



The liquid layer which is in contact with the glass wall, is of 

 course cooled most i-apidly. If the temperature of this layer has 

 fallen below^ =b 160 , it differs so much in specific weight from the 

 layers that lie deeper, that it floNvs dowjiward, and collects at the 

 bottom, in which the sulphur of higher temperature, darker colour, 

 and smaller sjtecific weight is forced upwards. 



It nmst, however, be noted here that if the sj)ecifically heavier 

 liquid did not flow downward along the wall, yet the lowest part 

 of the liquid column wcudd get clear first, for even then the cold 

 air which rises upward round the hot tube, would bring about a 

 fall of temperature iji the sulphur in vertical dii-ection, in consequence 

 of which the lowest temperature would ahvays be found at the 

 bottom; but that in this case the phenomenon of the layers does 

 nut make its appearance, could be made clear by the experiment 

 as will appe ir presently. 



Probably the two phenomena take |)lace side by side; the former 

 will [)reponderate when wide tubes are used, the latter in the case 

 of narrow tubes. When the colder liquid flows downward there will 

 be a great chance of mixing, specially in case of wide tubes, and 

 this is the reason llial llie two-layer phenomenon is much less clear 

 with a lube of a diameter of 20 iiiiii. than with narrower tubes 

 both with regard lo dilference of colour and shar[»ness of demarcation. 



It is not very difficult now to account for the fact that stirring 

 down lo 170° prevents ihe occurrence of the two-layer i)heii()inenon. 

 For as the mass had throughout the lemperature of 170° at the end 

 of Ihe slirring, I lie dilference of temperature which arises be! ween 

 Ihe differenl lasers wilhout slirring on further cooling, is not sufficiejit 

 to bring about Ihe phenomeuoii in (piestion. 



With a view lo auginenling Ihe experimental material al our disposal, 

 some more experiments were made, which all without exception con- 

 firmed our view. AVhen using the tube of 16 mm. diameier we placed a 



