i()Z Mr, Morgan's Oofcrvat'ions and E>:perifiieuts on 



By examining the fliime of a common candle we may ob-< 

 ferve, that its iowefc extremities, or the part In which the black 

 colour of the wick terminates, dlf^jharges the lead heat; and 

 th;)t, as the vertex of the flame is approached, a fucceffive order 

 of jxirts is pafled through, in which the loweft is continually 

 addiiig to the heat of what is juH: above it, till wc come to 

 the top of the flame, near whicii all the heat is collected into a | 

 focus. At tile loweft extremity, however, where the heat is - 

 inconfiderable, a blue colour may be always obferved ; and 

 from this appearance, amongfl: others, it may, I think, be 

 fafely concluded, that the blue rays are fome of thofe which 

 efcape from combuftibles in an early period of their decompo- 

 fition ; and that if the decom^pofition could be examined in a 

 period ftill more early, the colour of their flame w^ould be 

 violet. By an a priori dedu6lion of this kind, I was led to 

 watch the appearances of a candle more attentively ; whence I 

 found that, to the external boundary of a common candle 

 is annexed a filament of light, which, if proper care be 

 taken to prevent the efcape of too much fmoke, will appear^ 

 moll: beautifully coloured with the violet and indigo rays. To 

 the preceding inftance of a common candle many fa6ls may 

 be added, which fpeak a fimilar language. If fulphur or 

 gether is burned, or any of thofe combuftibles whofe . 

 vapour is kindled in a fmall degree of heat, a blue flame 

 will appear, which, if examined by the prifm, will be found 

 to confiil of the violet, the indigo, the blue, and fometlmes a , 

 fmall quantity of the green rays. The beft mode, however, 

 of fhewing the efcape of fome rays l^y that degree of heat, 

 which will not feparate others till increafed, is the following* 

 Give a piece of brown paper a fpherical form, by prefling it 



upon 



