August 30, 191 7] 



NATURE 



D^D 



VIII.-XIII. 2-2 (07) 20 (0-5) 27 (17) O-S (0-2) 



I.-XIII. 2-3 (0-9) 17 (0-5) 2-4 (1-3) 0-8 (0-2) 



The conclusions of the whole investigation may be 

 summarised as follows (all this relates, of course, to 

 winter temperatures in \V. Europe) :— 



(i) Within each interval of 445 years (759-5-8o30 

 . . . i872o-i9i6-5), the first half is colder than the 

 second. 



[The difference in the amount of temperature-devia- 

 tion has been found on an average 20° per 44 winters; 

 after the year 1383 on an average 26''. 



Exceptions, or apparent exceptions, from this rule, 

 two out of twenty-six cases since 760, none since 1200.] 



(2) Within each interval of 89 years, to begin with 

 the year 759-5 (1827-5), the first half is colder than 

 the second. 



[The difference in the amount of temperature-devia- 

 tion has been found on an average 22° per 89 winters. 



Exceptions from the rule, two out of thirteen cases 

 since 760, besides two doubtful ones; since 11 16 one 

 exception.] 



(3) The chance that the last quarter of an 89-year 

 period (826-25-848-5 . . . i894-25-i9i6-5) contains a 

 smaller number of hard winters than the preceding 

 and following 22-year intervals is o-88. Within the 

 last quarter of an 89-year period the chance that any 

 winter will be severe (or very severe) is less than 

 04 (or 0-007), *-^- l^ss than A(l) of the general chance. 

 In the neighbouring 22-year intervals {e.g. 1872-93 

 and 1916-37) this chance is about three (live) timeh 

 as great. 



(4) Increased and accelerated activity of the solar 

 surface corresponds in general with the winter-cold in 

 Western Europe setting in more forcibly and quickly 

 jdian usual ; inversely, a weakened and retarded activity 

 bf the sun corresponds with winters setting in more 

 Tiildly and in a later part of the period. 



j The forecast for the period 1917-38, derived from 

 j:hese statistics, indicates at least two very cold and one 

 «vere winter; the average winter temperature for 

 hese twent\'-two years being generally below the 

 J^year mean. C. Easton. 



Amsterdam, June, 1917. 



Auroras and Magnetic Storms. 



With reference to yoiir note in Nature of August 16 

 eferring to a magnetic storm on the night of August 

 -10, it may be of interest to learn that an aurora 

 /as seen here that night. It was first seen a few 

 linutes before 10 p.m. (G.M.T.), when it appeared 

 s a glow in the northern sky. Two streamers were 

 1st discernible at first, but they gradually increased 

 t numbers and became clearer, at the same time 



NO. 2496, VOL. 99] 



growing longer and brighter and moving towards the 

 west. The longest reached to the centre of the Great 

 Bear. Small, sharp, and delicate streamers, although 

 not prominent, were distinctly seen in the larger 

 streamers. There was no colouring seen at all, but 

 merely a white glow. By 11. 15 all traces of it had 

 vanished. L. Cave. 



Testing Squadron, Royal Flying Corps, 



Martlesham Heath, Suffolk, August 21. 



An Unusual Rainbow. 



An unusual rainbow display was visible at sea be- 

 tween 6.30 and 7 p.m. on thp evening of August 16. 

 The primary and secondary bows were complete and 

 of exceptional brilliancy. Between these two lay two 

 arcs of a third bow, cutting the primary bow near the 

 horizon and ending in the secondary bow about 20" 

 above the horizon in the manner shown in the accom- 

 panying diagram. 



The blue of this bow w^as towards the primary 

 bow, and the red towards the secondary bow. This 

 third bow cut the primarv bow at an angle of 

 25°-30°. 



Outside the secondary bow were visible two arcs 

 of a fourth bow (less distinct than the others) which 

 cut the secondary bow in much the same way as the 

 third cut the primary. 



Unfortunately, I am unable to give you at present 



the ship's position at the time when the phenomenon 

 was seen. The sun's altitude was about 7° when the 

 bows were most clearly seen. The afternoon was 

 warm and sultr}- and there was practically no wind. 

 A thunderstorm took place at some distance from the 

 ship during the afternoon. 



I shall be glad if any of the readers of Nature can 

 give me an explanation of the phenomenon, which 

 has caused considerable discussion among the officers 

 of the ship. • All.an J. Low. 



August 16. 



An Invasion of Ants. 



Yesterday afternoon (Bank holiday) the weather 

 suddenly became brilliantly sunny and very hot, after 

 some days of gloom with rain and thick east wind 

 atmosphere; and about five o'clock I became aware 

 that apparently every ants' nest in the garden had 

 chosen that precise moment for the emergence of its 

 winged inhabitants. There they were in myriads, 

 swarming out of holes in the drive, gravel paths, flag- 

 stones, the rock-garden, where they had been devas- 

 tating Sempervivum clumps, and all over the lawns. 

 They were nearly all the small red ant, only a few 

 nests of the small black one. 



The tiny winged males niiuch outnumbered the large- 

 bodied winged females, and both were attended by 

 fussily anxious "workers"; by seven o'clock all were 

 gone. Can ants delay their appearance above ground 

 until the onset of suitable hot, dry weather? 



Eleonora Armitage. 



Dadnor, Herefordshire, August 7. 



The appearance in swarms of male and female ants 

 for the nuptial flight is deiscribed by many observers. 



