790 



NATURE 



and henc<' an upwcllinK' "f deeper water due to a fall 



in It.iKMii'trir prfS'ii'.rr should cause n rise in specific 



II) tin: 



I Laccaiiivc Sc;t Uiiimg the 



I rhniary 1023. It is now 



1, f liarometrlc pres.siire and 



8) \ith each other. (In this 



anti !M v\\\. J, ;is i nave no records made on lM>ard 



ship, I have taken the average barometric pressure 

 registered in ralcnttn, as given in the Barometer 

 Manual, luiu.) At this season of the year the rain- 

 fall is sHght, and, owing to evaporation, the surface 

 layer tends to have a higher specific gravity than 

 water underlying it. In consequence an upwelling of 

 wall r owing to lowered barometric pressure, produces 

 a tall m specific gravity. 



1 uiall\ , in Fig. 3, I have given the results obtained 



Fig. 3. — Average specific gravity of surface water and average baromttric 

 pressure off the west coast of India in May 1923. 



off the west coast of India in May 1923. The south- 

 west monsoon had already set in, and there had been 

 a considerable amount of rainfall in the region under 

 investigation since the middle of April. In conse- 

 quence the surface layer of water had again become 

 diluted, and we now get a return to the condition 

 foiiiul in the month of October, i.e. a rise of barometric 

 pre.ssure causing a fall in salinity, and a reduction in 

 the pressure being accompanied In- a rise due t6 the 

 upwelling of denser water from below. 



NO. 2822, VOL. 112] 



app. 

 hav< 



limi; 



[December i. 1923 



records taken during the 



Uiing from the results of 



should expect to find a 



icrnmpanieMl by a rise of 



?o the effect 



most levels. 



it>U' to publish 



it, .is the matter 



sKjc-rable interest, I 



'\ publishing a pre- 



TJif Indian Must 



Speitial .Scries in tliv ()\\^cn (>i<)uj). 



^p( < till in of oxygen there occur two types 



(y; .„....,, triplet series, and ^'"' s^'-^^alled " singlet " 

 series. The terms associat' the former type 



are designated by Fowler (' i , jh Series in Line 

 Spectra ') as ms, mp, and md, while those associate<l 

 with the latter type are designated as mS, mP, and 

 mD. 



One of us (Hopfield) has recently found a numlx-r 

 of new oxygen lines occurring in the ultra- vi<. 

 (Nature, September 22, p. 437). These appear 

 fourteen triplets, and for their representation o; 

 fifteen different lenns art- n-cl''!. l-"ourteen 

 these are previously known terms of oxygen 



(i5 to 7S, 2D to 7D, ami iie fifteenth is a 



new triplet term, of larger trecjuency value than any 

 previously known oxygen term. Each of the fourteen 

 triplets represents, on the Bohr theor-- *'- '-ansition 

 from some known singlet energy !■ ae new 



triplet level. The nomenclature jusi k'>^" '"iS been 

 used in all prcMous < oinnuinications and is identical 

 with that of I'oulcr, a point apparently not made 

 clear, since the editor of Nature added a note to 

 the letter of September 11. saving that the .S. P, 

 and D terms did not :as 



of the same designation 



The fourteen oxygen terms include only 5' lor s) 

 and D terms, and this leads to the conclusion that 

 the new energy level corresponds to a P (or />) term. 

 Whether it is a P or ^ term is not at once evident. 

 Since the known mp terms of oxygen are triple, 

 while the mP terms are single, the suggestion was 

 made by one of us (Birge) that the new level be 

 designated o/j,2.t I" ^^^t. the spectral diagrams 

 devised by Brackett and Birge [Physical Review, 21, 

 710, 1923, and Jour. Optical Society of America, 

 now in the press) predict this level at precisely the 

 position found. But it appears, as mentioned in 

 the letter of September 22, that in the op-is triplet 

 the opi-is component is detinitely lacking, and a 

 similar phenomenon was later found in the case of 

 the corresponding new sulphur triplets which are 

 discussed in the same communication Since the is 

 portion of the op^-is designation is known to be 

 correct, and since the known is-nip series of oxygen 

 and of sulphur consist alwavs ot triplets, the theory 

 of inner quant numbers leads inevitably to the 

 conclusion that op is » / the correct designation of 

 the new triple level. Hence tiie designation has been 

 changed to oPj,). An assignment of inner quant 

 numbers to the various terms, according to Sommer- 

 feld's method, then leads immediately to the prohibi- 

 tion of the oPi-15 component, by the Selective 

 Principle. The inner quant numbers in the case of 

 the triple oP term run in a direction opposite to 

 the usjial, in keeping with the " inverted " character 



