( 824 ) 
Then the tube c was connected with the Gaede-pump through two 
U-shaped tubes placed one behind the other. 
The first of these tubes was placed in a vessel with solid carbonic 
acid and aleohol to condense all the water-vapour, the second being 
immerged in a vessel with liquid air to solidify the last traces of 
phosphorus so that a high vacuum could be obtained. 
Now first of all the apparatus was exhausted, and the flask D 
was. carefully heated with tbe flame to distill off the water. 
When nothing of the water was to be detected any more we 
continued the boiling in vacuo for another half hour to be sure 
that also the water dissolved in the phosphorus was entirely removed. 
After the connection of the apparatus with the pump had been 
broken by sealing it at 4, three fourths of the phosphorus was dis- 
tilled over into 4, in which a liquid was obtained, which at first 
opalized somewhat, but became perfectly clear and colourless afterwards. - 
Though this phosphorus appeared to be as pure or purer than the 
purest product obtained by other methods, it was not yet pure 
enough for our purpose, as experiment showed that this phosphorus 
could not be made to melt sharply at one temperature by any means 
whatever. 
To reach a still higher degree of purity the phosphorus in B was 
almost entirely melted, and then made to erystallize again by slow 
cooling ; when three quarters of the mass had solidified, the remaining 
liquid was conveyed to C by tilting of the apparatus, and solidified 
there by strong local cooling of the supercooled liquid (by means of 
solid earbonie acid and aicohol). 
This manipulation was repeated a great many times, in which the 
mass remaining in B did not only get continually a higher melting- 
pou:t, but also became coarser and more perfectly crystalline. 
When through this repeated partial crystallisation and removal of 
the liquid four fifths of the quantity originally present in 4 was 
conveyed into C, we filled the vessel A with the phosphorus which 
had remained in 5 by melting it in B, and by then making it flow 
into A by tilting the apparatus. 
By then sealing off at a we broke the connection of the melting- 
point-vessel with the other part of the apparatus, and the experi- 
ment could begin. | 
The perfectly colourless coarsely crystalline phosphorus obtained in 
this way pow appeared te be so pure, that on immersion of the 
melting puint-vessel in a thermostat the temperature of which varied 
slowly, a melting-range was found smaller than 0.02°, so that we may 
say that we have determined the unary melting-temperature of the 
