June i, 191 i] 



NATURE 



459 



for the week was above the average over the whole area 

 erf the British Islands, the greatest excess being 3-6° in 

 the Midland counties, and the least i-i° in the north of 

 Scotland. The absolute temperature ranged from 73° in 

 the east of England to 33° in the east of Scotland. 



A NEW edition (the seventh) of the handy " Hints to 

 Meteorological Observers," prepared under the direction of 

 the Council of the Royal Meteorological Society by Mr. 

 W. Marriott, has been received. The present edition has 

 been revised and enlarged ; the explanations and illustra- 

 tions of ordinary and self-recording instruments are very 

 satisfactory. We are glad to find a considerable addition 

 to the very useful glossary of meteorological terms, in- 

 cluding those most recently introduced. We think this 

 might be still further improved by more additions, and 

 occasionally by a Uttle fuller explanation. We notice here 

 and there a slight departure from the explanations usually 

 given, e.g. the order of the colours of the corona. We can 

 only repeat the opinion before expressed, that the work 

 takes a high place among the best of such handbooks 

 published in any country. 



An interesting article on the weather in the seventeenth 

 century, by Mr. W. Sedgwick, is published in Symons's 

 Meteorological Magazine for May, containing extracts 

 relating to the spring (March-May) between 1658 and 1705 

 from the diaries of John Evelyn, F.R.S., and Samuel 

 Pepys, F.R.S. The author proposes in this and subse- 

 quent numbers to give an opportunity of considering 

 whether these extracts show that any marked change in 

 the climate of London has occurred since that period. 

 Statements are often made that our climate has under- 

 gone considerable changes in comparatively modern times. 

 On the other hand, well-known investigators of the pre- 

 sent day have shown that any apparent changes either in 

 temperature or rainfall can be accounted for by the differ- 

 ence in the instruments and their exposure. Although 

 these instruments were known before the close of the 

 seventeenth century, there were but few in existence ; they 

 cannot have been used regularly, if at all, by Evelyn or 

 Pepys, and the tendency in the case of non-instrumental 

 observations would be to record abnormal rather than 

 normal conditions. Another important consideration 

 pointed out by the author is the change from the Julian 

 to the Gregorian Calendar, which was made in England 

 in 1752. With reference to the popular belief about the 

 old-fashioned Christmas, in several years during the last 

 decade there have been considerable falls of snow after 

 Christmas which would have occurred before Christmas 

 if the Julian Calendar had been still in force. 



The Journal of the Royal Statistical Society for May 

 <ontains an imjwrtant paper, by Mr. E. C. Snow, on a 

 new method of estimating post-censal populations, i.e. the 

 populations of different districts of a country in the years 

 following a census. The estimation of such populations 

 often offers considerable difficulties, especially in districts 

 of a rapidly changing character in the neighbourhood of 

 large towns, and the method at present in official use — 

 ased on the assumption of the approximate maintenance 

 of the rate of change during the preceding intercensal 

 decade — may lead to very serious errors. For example, 

 the birth- and death-rates in Salford in 1890, based on 

 the estimated population, were calculated at 288 and 22-4 

 respectively, but when the results of the census taken in 

 the following year were made known, these figures were 

 altered to 356 and 27-6. Mr. Snow suggests that definite 

 indices of a change of population, such as changes in the 

 number of births, deaths, marriages, or houses, should be 



NO. 2170, VOL. 86] 



used as the basis of the estimate, that regression equa- 

 tions should be formed by the method of correlation 

 between the change in population of a district and the 

 changes in these several indices during a completed inter- 

 censal decade, and that these regression equations should 

 be applied to the following decade. Trial of the method 

 on several groups of districts of diverse characters showed 

 that it would lead to greatly increased accuracy. 



Investigations of the hitherto almost unknown ultra- 

 violet spectrum — the Schumann region — are of special 

 interest, because the conditions attending the production of 

 these extremely short wave-length radiations are obviously 

 of a different character from those attending the production 

 of the more familiar spectrum. In this research Mr. 

 Theodore Lyman has taken a great part, and in the March 

 number of The Astrophysical Journal (vol. xxxiii.. No. 2, 

 p. 98) he publishes results obtained from an investigation 

 of the nature of the radiation from oxygen, hydrogen, 

 nitrogen, helium, and argon in the region more refrangible 

 than A. 1900. No lines of helium and oxygen have yet 

 been discovered in this region ; if they exist they are too 

 faint to be disclosed by the present methods. By varying 

 the conditions of discharge in the vacuum tube, two spectra 

 of nitrogen were revealed, one of faint bands with heads 

 on the more refrangible edges, the other consisting of two 

 pairs of sharp lines at \\ 1492-8, 14948, 17427, and 

 I745-3- No lines were seen in the " red spectrum " of 

 argon, but a considerable number, about forty between 

 ^ 1333-7 and A. i886-i, exist in the " blue spectrum." 

 Repeating the experiments made by Schumann, Mr. 

 Lyman has, as yet, been unable to obtain the primary 

 spectrum of hydrogen which Schumann suspected. An 

 interesting spectrum, apparently intimately associated with 

 hydrogen, apf>ears in the region \ 1650-A. 1450, and con- 

 sists of five groups, each group containing five lines. 

 Argon containing a trace of hydrogen at a pressure of a 

 or 3 mm. shows this spectrum well if aluminium electrodes 

 and no capacity are employed. Nitrogen, oxygen, and 

 helium containing a trace of hydrogen do not show the 

 groups, and if other electrodes are used considerable 

 enfeeblement occurs. With pure hydrogen these groups 

 always appear — with the other lines — and they disappear 

 if the last trace of hydrogen is removed from the argon 

 mixture. If their origin is an impurity in the hydrogen, 

 it must be of a fundamental character, for the groups 

 appear in all the hydrogen used by Lyman and by Schu- 

 mann for many years ; they may be a new spectrum of 

 hydrogen. For a description of the apparatus and 

 methods employed in the research the reader is referred 

 to the original paper. 



The April number of Le Radium contains a memoir by 

 M. L. Dunoyer, of the laboratory of Madame Curie, on 

 the production of a material radiation of purely thermal 

 origin. A tube of hard glass about a centimetre in 

 diameter and about 30 centimetres long is joined at its 

 upper end to a wider tube, the length of which varies from 

 2 to 13 centimetres in different cases. A side tube leads 

 from the enlarged head to a Gaede pump. The lower end 

 of the tube is covered inside with a film of metallic sodium 

 obtained by distillation in vacuo from a heated tube origin- 

 ally communicating with the experimental tube, but sealed 

 off when the deposit has been formed. When the lower 

 part of the tube is now heated, so as to vaporise the 

 sodium and drive it into the upper part of the tube, it is 

 found that if diaphragms with small openings of the order 

 2 or 3 millimetres diameter are placed in the tube above 

 the sodium deposit, the molecules of vapour are driven 

 through the openings with such velocity that they form a 



