October 28, 1909] 



NATURE 



515 



seasonal rainfall, and a general table of the annual 

 rainfall at all stations. There is also a coloured 

 frontispiece map showing the relation of the rainfall 

 of 1908 to the average of 1870-90. The rainfall of 

 Scotland and Ireland, generally, was practically 

 normal, that of England and \\'ales rather more than 

 one-tenth less than the average. Only a small part 

 of England and Wales, but a large part of Scotland 

 and Ireland, had more than the average; parts of the 

 south of Ireland, south-west and east of England and 

 east of Scotland were very dry. The greatest annual 

 amount recorded was 237'3 inches, at Llyn Llydaw 

 (Snowdon) ; the least, 15b inches, at Bourne (Lincoln- 

 shire). Among the changes introduced in this volume 

 may be mentioned (1) that, in discussing monthly 

 rainfall, maps of the actual fall are given side by side 

 with those showing the percentage difference from the 

 normal ; (2) much fuller treatment of the sections 

 relating to heavy falls on rainfall days and in short 

 periods. Although efforts have been made to econo- 

 mise space, and no part is a repetition of any previous 

 issue, the present volume is larger than any of its 

 predecessors, and, we thinh, compares favourably with 

 them. 



La Mesure rapide des Bases geodesiques. By J. 



Rene Benoit and Ch. Ed. Guillaume. Quatri^me 



(Edition. Pp. 228. (Paris : Gauthier-Villars, 1908.) 

 The invention of invar, the nickel-steel alloy with a 

 small or zero coefficient of expansion, in 1897 im- 

 ported new conditions into determinations of length. 

 The great value of this invention for the measure- 

 ment on the ground of the base-lines of a survey 

 was at once apparent, and during the next year the 

 new method was tried by the joint Russo-Swedish 

 expedition in Spitsbergen. The results obtained 

 equalled the most sanguine expectations. Not only 

 was a high limit of precision attained, but the rapidity 

 of the work, as compared with the old methods, was 

 enormously enhanced. It was obvious that the geo- 

 desist had in his hands a new tool which greatly 

 accelerated the most laborious portion of his opera- 

 tions, and at the same time gave him a degree of 

 accuracv at least equal to that hitherto achieved with 

 far more cumbrous apparatus. 



These results came into prominent notice at the 

 International Geodetic Conference held at Paris in 

 1900, and a further and more minute investigation 

 of the whole problem was undertaken by the Comite 

 international des Poids et Mesures. 



At the meeting in 1905 a report was furnished by 

 the present authors, to be expanded into a more com- 

 plete form for presentation to the Geodetic Conference 

 of 1906. The large demand for copies of this report, 

 and the fact that the new method has now been 

 adopted by almost all surveys having any pretensions 

 to execute work of the first order, have induced the 

 authors to put the record of their investigations into 

 a permanent and convenient shape. Of the present 

 liltle volume we can only say that it is one that must 

 be in the hands of every geodesist. It contains, in a 

 succinct form, the general theory of measurements by 

 wires hung freely between supports, a short discus- 

 sion of the physical properties of invar, an account of 

 the testing and standardisation of the wires and of 

 their possible distortions under different conditions of 

 tension, temperature, repeated windings and unwind- 

 ings, and, in general, their stability under the prac- 

 tical conditions of their employment in the field. This 

 is followed by a description of the auxiliary apparatus 

 used for base measurement, all of a very simple char- 

 acter, and a full account of the routine of the field 

 work and of the calculations for the reduction of the 

 measures to the horizontal, including the necessary 

 tables. 



NO. 2087, VOL. Si] 



A summary of the actual results attained in practice 

 shows that a rate of measurement of about 5 kilometres 

 per day can be kept up with a limiting error of he- 

 tween jgjJijQjj and YuoJjj^jj. With special precautions a 

 still higher degree of apparent accuracy can be 

 reached, but such appearance is largely delusive, and 

 in geodetic work would soon disappear in the angular 

 measures. 



We congratulate the authors, not only on their most 

 valuable investigations, but also upon the excellent 

 form in which their conclusions are presented. 



E. H. H. 



Bibliotheca Geographica. Jahreshihlio graphic der 

 gesamten geographischen Literatur. Heraus- 

 gegeben von der Gesellschaft fiir Erdkunde zu 

 Berlin. Bearbeitet von O. Baschin. Band xiv. 

 Jahrgang, 1905. Pp. xvi-i-545. (Berlin: W. H. 

 Kiihl, 1909.) 

 This is one of those publications (and there are not 

 a few) for which British geographers may well be 

 grateful to German. We do not produce many works 

 of this kind — works which can hardlv bring profit 

 sufficient to reward the labour of their compilation. 

 The " Bibliotheca Geographica " is a wonderfully full 

 bibliography of geographical books and papers pub- 

 lished, in all parts of the world, during the vear 1905 

 — it is not to be wondered that the volume bears a 

 date of issue four years later, when the magnitude 

 of the task of tracing such a vast number of publica- 

 tions is considered. General treatises on the various 

 scientific branches of geography are given first ; after 

 these there follows what occupies the bulk of the 

 volume — a bibliography according to topographical 

 divisions. Each topographical division is minutely 

 subdivided according to special subjects, an arrange- 

 ment which partially disarms the criticism that the 

 entry of a publication under its own title or the name 

 of its author seems somewhat arbitrary. 



It is doubtful, however, whether any large biblio- 

 graphv would entirelv escape this criticism. Some 

 difficulty, again, is evidently felt with regard to the 

 entry of individual papers out of collected volumes. 

 Thus, when the report of a research committee of the 

 British Association is entered only under the name 

 of the secretary of the committee, it may be doubted 

 whether this method gives the reader the best chance 

 of finding the reference. On the other hand, cross- 

 references are provided from one subject subdivision 

 to others of a kindred nature, and at the end there 

 is an authors' index, so that one cannot but recognise 

 that the system of the whole work is well-nigh per- 

 fect. Moreover, the entries, so far as can be judged, 

 appear to be admirably accurate. 



The Iiivicia Numhcr Scheme. Bv 1. W. Ladner. 

 (London : George Philip and Son, Ltd., n.d.) Hand- 

 book, price 8d. net; Number Board, price, with 

 plain edges, 6s. 6d. per dozen ; with edges, cloth.i 

 mounted, Ss. 6d. per dozen. 

 This device is a method found useful by a practical 

 schoolmaster of experience in teaching the funda- 

 mental principles of arithmetic by constant reference 

 to the decimal system of notation. The plan utilises 

 not onlv the ears, but the hands and eyes of the chil- 

 dren. Though many original teachers will have 

 developed equally good expedients for rendering their 

 lessons in arithmetic practical, interesting, and in- 

 telligible, the scheme may be recommended to the 

 attention of teachers who have not as yet adopted con- 

 crete aids in their instruction. It is now very gene- 

 rally agreed that children learn best by doing, and 

 Mr.' Ladner's method of teaching will certainly assist 

 the children to arrive at the rules they have to learn 

 from the' results of their own experiments. 



