iVA TURE 



[September 26, 1S95 



The data pre\ious to 1841 are, I suppose, less reliable ; but I 

 may add these two cases of August rainfall under average : 



Sun-spot max. 1S30 

 ■837 



Rainfall of August 1831, 2'I4 in. 

 1S3S, 0-93 ,, 



By way of showing that in other parts of the countr)- there 

 has been, in recent years at least, a similar variation, I add three 

 similarly smoothed curves of .\ugust rainfall for Haverfordwest, 

 Llandudno, and Boston (Lincoln) resjwctively (1, r/, c). The 

 data, however, do not extend back further than 1 866. 



The case of Greenwich may be presented as follows : — Take 

 each maximum sun-sp<'>t year, and a year on either side, and 

 tabulate the .-Vugust rainfall in each of these. Indicate by 

 the letters d (for dry) and ii' (for wet) whether this rainfall has 

 been below or above the average. Then we have : — 



Mxximum. 

 1S47, 1S4S. 1S49 



1S59. i860. 1 86 1 

 1869, 18-0. 1 87 1 

 IS82, ISS3. 1884 



1-95 4-25 045 



I -13 368 057 



I -21 2'02 0'86 



I '16 071 067 



d -Ji d 



d w d 



d d d 



d d d 



Here we find ten cases of a dr)- .\ugust out of twelve. 

 twelve \-alucs give an average of I "55 inches. 



Now do the same with minimum sun-spot years : — 



Those 



Here we find eleven cases of a wet .\ugust out of fifteen. 

 Those fifteen cases give an average of 2*84 inches. 



It would Ik: interesting to know to what extent such relations 

 suiisist elsewhere, and perhaps some of your readers may be 

 dis|xiscd to investig.ate the matter. A. B. M. 



Alteration in the Colours of Flowers by Cyanide 

 Fumes. 



It is well known that the yellows of some insects are turned 



to red liy the fumes from potassium cyanide : but I have not, 



after some inquir)'. t)een able to obtain any literature describing 



the effects of such fumes upon the colours of flowers. The 



reactions I have observed are very curious, and while it seems 



improlxible that they are hitherto wholly unknown, it may not 



lie amiss to direct attention to them. .\ few lumps of the 



cyanide are placed in a corked tul>e, covered with a little cotton, 



and the flowers are placeil on the cotton. It is prolwbly 



I1L-. L>.ary th.1t the day should be hot, or the tube slightly 



1. The pink flowers of CUomc inUf;rifolia and Moiiarda 



• a turn to a brilliant green-blue, and finally become pale 



> . A purple-red I'erliena becomes bright blue, then pale 



The purple flowers of Solauuiii slnagnifoliiim go 



' ' inrl then yellow. The while petals of Argeiiume 



irn yellow — the natural colour of ./. m:xi<ana. The 



'!! il.iwcrs of Mdituliii Hilda turn a <leeper yellow. 



'■IIS art^cnteiis, var., turn pale yellow. White 



'1'iwcrs turn yellow. The scarlet flowers of 



■ turn pale dull pink, resembling some- 



■)f the same. .\ny of your re.idcrs will 



' " '>iii.iin siiniiar results with the flowers growing in their 



y. T. D. A. Ci)i KKKBl.l.. 



'.-■^ Cnices, New Mexico, '' ^ N ^'ptcmber 3. 



THE CONSTITUENTS OF THE CAS IN 

 CLEVEITE. 



\\ 'F, have investit;atcd the spectrum of the gas dis- 

 ' ' covered in the mineral cleveite by Ramsay, and 

 have fountl it to be most regular. It consists of six series 

 oflin- •' ■ • of the lines in each series decreasing 

 with ' '--langths. .Similar scries of lines 



have 1... ■. ., ; m many spectra. The first series 



was discovered by Dr. Muggins m the ultra-violet spectra 

 of a number of stars. It proved to belong to hydrogen, 

 and to be the continuation of the four strong hydrogen 



NO. 1352, VOL. 52] 



lines in the visible part of the spectrum. Johnstone 

 Stoney had already shown that three of the wave-lenyths 

 of the visible hydrogen lines were most accurately pro- 

 portional to the \alues 95, 43, 98, when Balmcr 

 discovered that these \alues were gi\en by the formula 



for m = 3, 4, 6, and that the other wave-lengths of the 

 series were proportional to the values obtained by sub- 

 stituting^ for m the other entire numbers yre.iter than 

 three. The series has now been followed from nt = 3 to 

 III = 20, the lines growing weaker and weaker to the more 

 refrangible side, and approaching each other closer and 

 closer. The formula shows that they approach a definite 

 limit for large values of ///. This is seen more clearly 

 wOien we consider wave-numbers instead of wave-lengths, 

 which according to the formula would be proportional to 



Many series of lines similar to the hydrogen series 

 were discovered by Liveing and Dewar. Tliey have 

 called them harmonic scries, and have compared them to 

 the series of over-tones of a vibrating body. Tliey have 

 been further studied by Rydberg and by Kayscr and 

 Runge. We cannot here enter into any detailed account. 

 We only want to explain so much as to make the con- 

 clusions understood which we have drawn from the 

 spectrum of the gas in cleveite. The wave-lengths X of 

 the lines belonging to the same series arc always approxi- 

 mately connected by a formula somewhat similar to 

 Balmer's 



l/K = A - B/m- - C/m*. 



A determines the end of the series towards which the 

 lines approach for high values of /«, but docs not influence 

 the difference of wave-numbers of any two lines. IS has 

 nearly the same value for all the series observed, and C 

 may be said to determine the spread of the series, cor- 

 responding intervals between the wave-numbers being 

 larger for larger \alues of C. .■\s H is approxim;itcIy 

 know n, two wave-lengths of a series suffice to determine 

 the constants .A. and C, and thus to calculate approxi- 

 mately the wave-lengths of the other lines. It was by 

 this means that we succeeded in disentangling the 

 spectrum of the gas in cleveite, and showing its 

 regularity. 



In the spectrum of many elements two series have been 

 observed for which .A has the same value, so that they both 

 approach to the same limit. In all these cases the scries 

 for which C has the smaller value, that is to say wliicli 

 has the smaller spread, is the stronger of the two. In the 

 spectrum of the gas in cleveite wc have two instances of 

 the same occurrence. One of the two pairs of series, the 

 one to which the strong yellow double line belongs, 

 consists throughout of double lines whose wave-numbers 

 seem to have the same difTercncc, while the lines of the 

 other pair of series appear to be all single. Lithium is 

 an instance of a pair of series of single lines approaching 

 to the same limit. Hut there are also many instances of 

 two series of double lines of ecpial difference of wave- 

 numbers ending at the same place as sodium, potassium, 

 aluminium, &c. There are .ilso cases where the members 

 of each series consist of triplets of the same difiercnce of 

 wave-numbers as in the spectrum of magnesium, calcium, 

 strontium, zinc, cadmium, mercury. Hut there is no 

 instance of an element whose spectrum contains two pairs 

 of series ending at the same place. This suggested to us 

 the idea that the two pairs of series belonged to different 

 elements. One of the two pairs being by far the stronger, 

 we assume that the stronger one of the two remaming 

 scries belongs to the same element as the stronger pair. 

 We thus get two spectra consisting of three series each, 



