October 8, 1891] 



NATURE 



541 



coutume de definir le rendement d'un foyer de lumiere par le 

 rapport de I'energie situee dans la partie visible du spectre a 

 I'energie totale rayonnee par le foyer. Sans insister sur ce que 

 cette definition a de defectueux (je traiterai prochainement 

 cette question dans la Revue generale des Sciences), je rappel- 

 lerai qu'on niesure d'ordinaire le rendement en recevant 

 successivement sur un radiometre quelconque (pile de Melloni, 

 bolometre, radiomicrometre de Boys) la radiation totale du 

 foyer, et la radiation qui a traverse une certaine epaisseur d'eau ; 

 on admet que les radiations obscures ont ete retenues, et on fait 

 le quotient de ces deux quantites. Aucun physicien, je suppose, 

 ne croit que I'absorption par I'eau commence a I'endroit precis 

 oil cesse la vision, et devient immediatement totale, mais on 

 pense en general que le resultat ainsi obtenu est assez approche. 



Or nous pouvons determiner directement le rendement photo- 

 genique d'une source en mesurant la superficie des courbes 

 d'energie rayonnante visible et invisible. En partant des 

 nombres de M. Langley, on trouve ainsi, pour le rendement 

 d'une lampe a gaz une valeur comprise entre I et 2 pour cent. 

 D'autre part, les recherches de M. Knut Angstrom ont montre 

 que I'absorption par I'eau est presque nuUe pour A = i/x, et n'est 

 totale qu'a partir de A = 2ju environ. Une couche epaisse d'eau 

 laisse passer pres de lo pour cent de I'energie rayonnante 

 invisible. La methode ordinaire donnerait done, pour le 

 rendement d'une lampe a gaz, ii a 12 pour cent, c'est a dire 

 une quantite six fois trop forte. 



Je ne quitterai pas ce sujet sans faire remarquer le singulier 

 usage en vertu duquel la puissance de la radiation solaire 

 est rapportee a la minute, tandis que toutes les puissances 

 possibles — cheval, horse-power, watt, ainsi que toutes les radia- 

 tions— sont exprimees par rapport a la seconde. II serait temps 

 de faire disparaitre cette anomalie. 



Ch. Ed. Guillaume. 



Pavilion de Breteuil, Sevres, France, 

 25 septembre, 1891. 



Weather Cycles. 



With reference to this most interesting question, may I be 

 allowed to call attention to the following figures ? Having had 

 to consult Dr. Rutty's " Natural History of Dublin," 1772, vol. 

 ii., I casually found on p. 353 of that volume, in his remarkable 

 detailed registry of the weather in Dublin for a long series of 

 years, the following remark : " It has been remarked that the 

 following years were memorable for great frosts in England, 

 viz, 1638, 1661, 1684, 1708, 1716, 1739." Now the intervals 

 between these dates are 23, 23, 24, 8, 23. He further remarks, 

 on p, 368 : — "It is to be observed that whereas since the great 

 frost of 1739, until the latter end of the present summer, 1744, 

 we had generally an unusual prevalence of dry weather, in 

 autumn our usual wet weather returned." It may be remarked 

 that the interval of 23 years is about double the sun-spot period, 

 and furthermore that the years mentioned by Rutty correspond 

 roughly with years of sun-spot minima or maxima as given in 

 "Wolf's Catalc^ue, mentioned by Guillemin in his work " Le 

 Ciel" (1877), P- 104- This correspondence would appear as 

 follows : — 



J. p. O'Reilly. 

 Royal College of Science for Ireland, Stephen's Green, 

 Dublin, September 25. 



NO. II 45, VOL. 44] 



Occurrence of the Ringed Snake in the Sea. 



The readiness with which the British snake ( Tropidonotus 

 natrix) will enter fresh water is well known. Its occurrence in 

 the sea seems anomalous, and therefore I venture to submit the 

 following details. 



The specimen in question was seen on September 7, from a 

 small boat on the east coast of the Isle of Wight, while about 

 a thousand yards distant from the shore, and about midway 

 between Shanklin and Luccombe Chines. When first seen it 

 was swimming straight out to sea — viz. in an easterly direction. 

 The sea was calm and a strong current was flowing from the 

 south, so that the creature was swimming across the current. 

 At first it took no notice of the boat, but as the boat was rowed 

 towards it, it changed its course and swam directly away from 

 the boat. It was soon captured, and found to be uninjured and 

 in good condition. Upon dissection it proved to be a male ; 

 the entire alimentary canal was absolutely empty. The internal 

 organs were free from disease or other abnormality. It measured 

 33 inches in length. It is most probable that this snake entered 

 the sea about a mile from where it was obtained, as the beach 

 is bounded by almost perpendicular cliffs, some 300 feet high, at 

 that place. J. Cowper. 



A Rare Phenomenon, 



Mr. Wilson's letter in your issue of September 24 (p. 494), 

 recalls what I myself saw on the same evening. On Friday, the 

 nth, I was returning with a friend to town after a day's ramble 

 in Epping Forest. We caught the 8.36 p.m. train at Epping, 

 which is due at Woodford at 8.59, and was, I think, only a few 

 minutes late. Just as the train was nearing Woodford Station, 

 my friend and myself simultaneously noticed a luminous band, 

 such as that observed by Mr. Wilson, and extending from the 

 horizon almost to the zenith. Our first unreflecting thought was 

 to refer it to the revolving light at the Naval Exhibition, only it 

 did not revolve, and the direction was quite wrong. The fact that 

 both of us thought of this is indicative of the appearance which 

 the luminous beam bore. The night was clear and starlit, and I 

 observed that the point in the horizon from which the beam rose 

 was almost under the Great Bear, but a little to the left as I 

 fact d it. We saw it only for a minute or two before it was 

 hidden from us by the shed of Woodford Station, in which 

 station we stayed for what seemed a long while. When we got 

 into the open country again, the phenomenon had disappeared. I 

 may add, that my own eye being unfortunately defective for red, I 

 asked my companion if he noticed any red tinge in the light, and 

 he answered that it seemed quite white. 



Burlington House. Herbert Rix. 



The narrow luminous band described in Nature, September 

 24 (p. 494) was seen here on Friday, the nth inst., between 

 8.30 and 9 p.m., at the same time at which it was seen by Mr. 

 Wilson in the county Westmealh, but about twenty-two hours 

 later than it was seen by Prof. Copeland in Aberdeenshire. It 

 passed close south of Cassiopeia, and nearly through the zenith. 

 Half an hour later it had drifted 8° or 10" southward, and had 

 becom? very faint. 



There can be little doubt that the very rapidly moving 

 "comet" seen by Mr. Eddie at Grahamstown, South Africa, 

 on October 27, 1890, was a phenomenon of this kind. 



J. L. E. Dreyer. 



The Observatory, Armagh, September 28. 



The Heights of Auroras. 

 The rare part of the phenomena described by your corre- 

 spondents is the extreme narrowness of the auroral arches seen 

 on the loth in the north of Scotland, and on the nth at Ryde. 

 I take all the other descriptions on the nth to refer to one arch 

 — a different one from that seen at Ryde ; and it was a much 

 wider one, and therefore less unusual, its width having been 

 about 5° as seen here. Your correspondents do not give its 

 width, except that, as seen from Nottingham, it was evidently 

 very broad, and is not stated to have been an arch at all, though 

 I should suppose it was one. The observation at Nottingham 

 Forest, compared with those further north, gives a good oppor- 

 tunity/or ascertaining the he^ht of the top of the aurora ; but, 

 as Mr. A. Marshall has not given the altitude of the base of the 

 aurora as seen from Nottingham, there are no materials for cal- 



