4 1 6 



NATURE 



[September 28, 1911 



rhi qu 1 ■. bi asked, Could the remarkable 



tliis summer (so hot and dry) bi i seen in 



any measun r I consider there was reason to expect a 

 very small rainfall (al Greenwich) ; and I would submit the 

 igh dot diagi am in suppoi oi this view. 



It is got thus : the rainfall of spring plus summer 

 -August) in each year, 1S41-1910, is noted, and 

 these values are added in groups of five (1841-5, 1842-6, 

 &c.). Then each sum is compared, by the dot method, 

 with the difference between it and the fifth after, a plus 

 value ni' aning that the later sum is the higher, while a 

 minus value means it is lower. (Each dot signifies one 

 sum by the horizontal scale, and the difference by the 

 vertical.) 



Before this summer we had got to comparing the sum 

 for 1903 (i.e. the group 1901 to 1905), which is 684 inches, 

 with the difference between it and the sum for 1908 (i.e. 

 1906-10), which is 6 inches (68-4-62.4). Next we have the 

 sum for 1904, which is 653 inches, and the value for 1909 

 has to be ascertained (i.e. the group 1907-11). 



Find 65-3 in the horizontal scale (see arrowhead), and 

 consider where the new dot is likely to go. (The encircled 

 dot shows its actual position.) It will be seen that, 

 yei the vinns have exceeded (say) 62 inches, the fifth 

 sum after has been a lower value hitherto. Let us lessen 

 the sum 65-3 by 1 inch, and see what we get. We have 

 thus 643 inches as (say) an extreme upper limit for the 

 group 1007-11. Now, the sum for 1007-10 was known, 

 viz. 54.1; ; and 64-3 — 54-5 = 9-8 for March-August, 1911. 

 The spring had 5-2 inches, so that we might look for 

 rainfall not ovei 1 inches. (The average is 

 6-7 inches.) The actual amount appears to have bei n 



3-7 inches. The difference of the sums for 1004 and 1 



is —i-9 inches; and the previous distribution of dots, 

 indeed, points to a greater diminution than 1 inch. 



Vu B MacDowall. 



Limits of Explosibility in Gaseous Mixtures. 

 In making [3 on the explosion ol 



by means of an incandescent wire, I 1 ed much 



limits ol explosibility than those usually given. The 

 following table shows the numbers obtained: — 



plain 



candescent 



Marsh gas 2-51024 5 to 13 



Coal gas 4 to 2S 6 to 29 



Hydrogen ... 3 to 75 S to 72 



In eai 1 ,1 with air. The explosion 



was carried out in a glass tube of about 10 



2187, VOL. 87 



having a stop-cock at each end, and a men 



sma11 attached to the middle of the tube. ' When the 



explosion occurs the mercury moves more 1 



to the force of the explosion, but a distim 



meat, apart from that due merely to the 1 



le the heating ol the in ra bi 

 seen with the proportions 

 given abovi . The mai sh gas was 

 prepared from zinc methyl, as 

 aluminium carbide, on treating with 

 water, was found to give 33 per cent, 

 hydrogen. 



I am not clear as to why the limits 

 should be wider than those usually 

 obtained; possibly it is due to cata- 

 lytic action of the platinum wire, or 

 possibly the apparatus is more sensi- 

 tive than that usually employed, 

 ["he matter seems of interest in con- 

 nection with colliery explosions. 

 Using the same apparatus, I have 

 obtained explosions with coal-dust 

 and air, and with lycopodium powder 

 and air, no other gas being present. 

 E. P. Perm an. 

 1 nivi rsitv Colli ge, Cardiff. 



Working Hypotheses v. Collection 

 of Bare Facts. 

 Perhaps it is desirabl 



that my review of Prof. Schu 

 book published in Nature ol Sep 

 tember 21 was written and 

 to you before Prof. H. H. Turner 

 had delivered his address to the British Associatioi 

 president of Section A. Consequently, nothing in my 

 review is a reply to, or has any reference to, Prof. Tui 

 excellent address. It would be discourteous to critii is< 

 important pronouncement of a leader in science in 



back-handed way. My statement and his are p 



not really in opposition, though they to some e: 

 emphasise opposite types of investigating activity. 



Oliver I 



Use of Windfurnaces in Smelting. 



It may interest your correspondent, Mr. George Turner 

 (p. 381), to be reminded that wind-furnaces (furnaces with- 

 out any blast but that of the wind) were used, at all 1 

 for lead smelting, much less than 900 years ago. Until 

 some time in thi seventeenth century the Derbyshire lead- 

 smelters did not employ an artificial blast. They, like their 

 predecessors in Roman times, built their furna 

 tops of hills and facing the quarter of the prevalent winds, 

 Dr. Percy in his "Metallurgy of Lead" gh 

 from Bishop Watson's "Chemical Essays" and from an 

 earlier writer, Joshua Childrey, describing such wind- 

 furnaces. 11. I 



Royal s e oi Vrts, W.C., September 24. 



Meteor-showers. 



Besides the ordinary display of Orionids that 1 

 in. 11 the middle of October, a considerable anion 

 meteoric activity may be looked for about the beginning 

 oi the month. Indeed, the intensity of I meteor- 



1- promises to be considerably greatei than thai ol 

 any of the subsequent ones in October. The Following 

 computed particulars of two important meteor-showers that 



mi dui befon Octol 



Epoch September 20, 3I1. (G.M.T.), second order of 

 magnitude. Principal maximum Septem 25m. 



Secondary maxima September 29, i6h., and Octob 

 16I1. 25m. 



Epoch Octobet 5, 6h. 30m., third ord 

 Principal maxin um Octobei .;. 22h. 10m. Secon 

 maxim. 1 Octobei , 2I1. 45m-, and October 4, nh. 

 Vfter these there will be meteoric qui' • 1 



12. l"ii- R. Hi 



B Igrave Villas, Rathmines, Dublin, Septemb 



