SECTIONAL TRANSACTIONS.— A. 313 



JOHANNESBURG. 



Wednesday, July 31. 



Rt. Hon. Lord Rayleigh, F.R.S. — The Fluorescence of Mercury Vapour. 



This paper deals vnth fluorescence excited by frequencies equal to or less than 

 that of the resonance line X 2537. 



It is known that unexcited mercury vapoiir shows an absorption band beginning 

 very near the resonance line, and extending towards the red as the density increases. 

 In 4 inches of denser vapour it can even be extended as far as X 3750, or more than 

 1200 A. 



It is found that green fluorescence can be observed with excitation at any part 

 of this band, from the resonance line to at any rate X 3450. In all probability only 

 technical difficulties hinder our going further still. 



The fluorescence excited consists of four distinct features : — 



(a) The resonance line X 2537. 



(6) The narrow band 2540, near the resonance line. 



(c) The broad structureless emission band X 3130 to X 3650 not known in 

 absorption. 



(d) The structureless visual band. 



When excitation is applied as far towards the red as practicable, the band c is 

 emitted entire, in spite of the fact that excitation is only applied at the middle of it. 

 Stokes' law is notably violated, the ' anti-Stokes ' efiect extending 230A° towards 

 shorter waves. 



As regards features a and 6, so far as the available evidence goes these are only 

 excited when the excitation is quite near the resonance line. They can, however, 

 be excited by frequencies definitely less than that of the resonance line. 



When excitation is applied by a continuous source broader than the resonance 

 line, it is possible to distinguish a discontinuity in the visual fluorescence. If the 

 pressure of the fluorescent vapour is a few centimeters, the emission due to the core 

 of the resonance line (core effect) extends only a fraction of a millimeter from the 

 wall of the vessel. As we proceed inwards from the wall of the vessel there is a dis- 

 continuity of intensity, the emission becoming suddenly less, and stretching for 

 several centimeters without much further loss of intensity. This latter is due to 

 absorption outside the core of the resonance line and I call it the wing effect. But 

 it probably extends to the limit of absorption (X3750, say), no further discontinuity 

 having been so far recognised. The breadth of the core of the line was determined 

 from this effect as 0-1 A°. It is considered probable that the core effect represents 

 atomic absorption and the wing effect molecular absorption. The same discontinuity 

 of intensity appears by photography through filters which transmit the spectral 

 features a and 6 while suppressing c and d. 



Both the core effect and the wing effect, as observed visually, move down stream 

 from the point of excitation when the vapour is in rapid motion. The visual wing 

 effect certainly, and the visual core efiect probably, show a time interval between 

 excitation and fluorescence. 



The photographic fluorescence (features a and b blended) lasts a much shorter 

 time than the visual. It is possible so to adjust the velocity of the vapour stream that 

 the visual effect extends several centimeters down stream, while the photographic 

 effect remains sensibly stationary. This applies both to the wing and the core efiect. 



But, what is remarkable and unexpected, both the photographic wing effect and 

 the photographic core effect can be observed to move conspicuously down stream in 

 a sufficiently rapid blast. 



Special interest attaches to the photographic core effect at low pressures. Under 

 these conditions there is no visual effect, and the photographic effect is apparently 

 limited to the resonance line. Thus the band features b, c and d are all absent, and 

 we have only the resonance line a. In this case, therefore, we have all the character- 

 istics of what is called resonance radiation, discovered by R. W. Wood in 1912. But, 

 contrary to received views, it is here found to last long enough to be propagated 

 some centimeters down the vapour stream from the point of excitation. This implies 

 a duration of the order of 10~< seconds, yet other methods of investigation indicate 

 that the atom only remains in the excited state for 10~' seconds. The reconciliation 

 of their discrepancy has not yet been found. 



