March 19, 1903 J 



NA TURE 



467 



of frostv days was nineteen in the winter of 1883-4, ar >d 

 there were fewer than thirty-five days with frost in the 

 winters of 1845-6, 1S50-1, 1858-9, 1862-3, 1865-6; 1876-7, 

 1SS1-2, 1883-4, 1S95-6 and 1897-8. In the five out of the 

 six months already elapsed of the present winter there 

 have been twenty-one days with frost, and as yet the 

 screened thermometer has not fallen below 23°-6. The 

 winter (six months) with the lowest mean temperature 

 at Greenwich is 1844-5, when the mean was 38°*8, and 

 the winters with the mean below 40 were 1844-5, 

 1854-5, 1878-9, 1885-6, 18S7-8 and 1S90-1. The winter 

 with the highest mean temperature was 1898-9, when 

 the mean for the six months was 45°'4, and the mean 

 for each of the six months, with the exception of 

 March, was above the" average. The winters with 

 the mean temperature above 44 were 1845-6, 1847-8, 

 184S-9, 1862-3, 1876-7, 1883-4, 1897-8 and 1898-9. The 

 mean for the five out of the six winter months already 

 elapsed (1902-3) is 44°*6, so that it is most highly 

 probable that the present winter will rank as one of 

 the foremost for its general mildness. 



Limiting the winter to a period of three months — 

 December, January and February — the coldest winter 

 in the last sixty years at Greenwich is 1890-1, with a 

 mean temperature of 34°'3, or 5 below the average, 

 and during this period the mean temperature for 

 December was 29° - 9, January 33°'9 and February 39°'i. 

 The second coldest winter was 1846-7, with a mean 

 of 34°'5, or 4°'S below the average. In 1894-5 me 

 mean temperature for the winter was 35°'i, or 4°'2 

 below the average, and February, 1895, with a mean 

 temperature of 29°'3, was the coldest of any winter 

 month since 184 1, with the exception of 29° - 2 in Feb- 

 ruary, 1855, and during this month (February, 1895) 

 frost occurred in the screen on twentv-three nights, 

 the lowest shade temperature being 6°'9, and for six 

 consecutive days the thermometer did not once rise 

 above the freezing point. The warmest winter (three 

 months) was 1868-9, with a mean of 44°'4, or 5°'i above 

 the average. The second warmest winter was 1876-7, 

 with a mean of 43°'7, which is 4°'4 above the average. 

 The winters (December to February) with the mean 

 3° or more below the average are 1844-5, '846-7, 1854-5, 

 1864-5, 1S78-9, 1879-80, 1885-6, 1890-1, 1894-5. "The 

 winters with the mean 3 or more above the average 

 are 1845-6, 1848-9, 1862-3, 1865-6, 1868-9, 1876-7, 

 1898-9, and the present winter, 1902-3, when the mean 

 temperature was 3°-i above the average. The mean 

 temperature of February, 1903, was 45°'3, which is the 

 warmest since 1869, and it was 16 warmer than 1855 

 and 1895. February had been cold for the previous 

 three years, and it seemed probable that it would have 

 been so this year, but experience has proved otherwise. 

 At the commencement of the present winter, it was 

 pretty confidently believed by many meteorologists 

 that the winter would be severe, but such a belief has 

 proved a most complete failure. It is, however, hoped 

 that in the somewhat near future long period weather 

 forecasts may be attempted. At present the forecaster 

 is bound to admit his utter inability to form anything 

 like an accurate estimate of our coming - weather in 

 England for more than twenty-four hours in advance, 

 except when we are experiencing pronounced cyclonic or 

 anticyclonic conditions, when we may with fair safety 

 venture an opinion for, say, a week. The absolutely 

 lowest winter temperatures at Greenwich (below io°) 

 are 4°'o January 9, 1841 ; 7°'7 February 12, 1845; 8 0- o 

 December 25, i860; 6°'6 January 5, 1S67; 9°-S De- 

 cember 25, 1870; and 6° - 9 February 8, 1895. The abso- 

 lutely highest temperatures in each of the three winter 

 months are December, 1848, 62° - 4; January, 1843, 

 57°o; February, 1846, 62°- 3 ; 1868, 6i°7; 1869, 6i°-6; 

 1878, 6o '5; 1891, 62°-i; 1899, 63°-g. The winter of 



NO. 1742, VOL. 67] 



1885-6 was severe and very prolonged, and it is 

 apparently the only winter with skating on the waters 

 around London in each of the four months December 

 to March. 



The average rainfall at Greenwich for the winter six 

 months for the last sixty years is 11-82 inches, and the 

 winters with the aggregate rainfall in excess of 14 

 inches are 1852-3, 1865-6, 186S-9, 1872-3, 1876-7, 1880-1, 

 1S82-3, 1896-7, 1899-1900. The wettest winter of the 

 whole series was 1876-7, with an aggregate rainfall 

 of 18-72 inches. The driest winters, with a rainfall 

 below 8 inches, were 185S-9, 1873-4, 1S79-80, 1890-1 

 and 1897-8. The driest winter was 1879-80, with a 

 total rainfall of 5-54 inches. The aggregate rainfall 

 •for five out of the six winter months of the winter 

 1902-3 is 7-3 inches, which is 4-5 inches less than the 

 average for the six months during the last sixty years, 

 and it is exceedingly improbable that the whole winter 

 (October to March) will prove to be wet. Only two 

 years have been wet at Greenwich out of the last four- 

 teen years, but seven of the last fourteen winters have 

 been wet, and ten of the last fourteen Decembers have 

 been wet. The tail end of the present winter is proving 

 very stormy, and for their destructive character the 

 recent gales, as shown by the publications of the 

 Meteorological Office, seem likely to prove as generally 

 disastrous as any experienced for a long time past. 



Chas. Harding. 



M' 



A UNIQUE VARIABLE STAR. 

 ESSRS. MULLER and Kempf, of the Astro- 

 physical Observatory of Potsdam, have recently 

 announced the discovery of a variable star of so short 

 a period that it must take a unique position among 

 this class of phenomena. Up to this time the variables 

 which went through a complete cycle of their light 

 phases in the shortest time were those two stars situ- 

 ated in the rich star cluster <a Centauri ; these bodies 

 completed their periods in 7h. 11*4111.- and 7h. 42-8m. 

 Another variable running these rather close is that of 

 S. Antli-e, the period of which is 7h. 46*8111. The new 

 variable is, however, of a much shorter period than any 

 of these, nearly one-half as short, occupying only four 

 hours and a few seconds to complete its light changes. 



The discoverers of this variable had their attention 

 first brought to this object in their work on the photo- 

 metric survey, in which it was noticed that there was 

 a great difference between two measures of this 

 star (B.D. + 56°i400, R.A. = 9I1. 36m. 44s., Decl. + 

 56 24'-6, 1900) that exceeded the usual error of observ- 

 ation. A closer examination of the star itself was then 

 undertaken, and a series of observations extending 

 over the year 1902, and part of this year, was made. 

 The account of this research, recently published (Site. 

 Ber. der K. Preuss. Ak. der Wiss., February 5, 1903, 

 vii.), gives the details of the observations and the con- 

 clusions arrived at. 



The diagram accompanying the paper shows that 

 the light-changes at an epoch of minimum vary very 

 quickly, the curve being quite pointed at these times. 

 From minimum to maximum the light changes at 

 rather a slower rate than from maximum to minimum, 

 and at about maximum the star apparently changes its 

 magnitude very little, so that the exact epoch of the 

 maximum is not so easy to determine as that of the 

 minimum. During these changes the magnitude 

 varies from 8-58 to 7*9, and the length of the period, as 

 at present determined, is 4h. om. I2*8s., with an error, 

 as stated, of probably about 0-55. For computing the 

 times of minima the following equation is given : — 



Min. = igo3 January 14, 4h. 32m. (Greenwich mean time; 

 + 4h. om. i2'8s. E. 



