January 23, 1919] 



NATURE 



403 



value "l 1 'S for 85 Pegasi, bul adopted 

 i - o; .ind on p. 233 he also points out ihai in 

 Sirius we have a system in which is an even 

 .1 disparit) between mass and luminosity in 

 the inn components, and he adopts Boss's corn- 

 ed value. Incidentally, this weakens one of 

 the Facts ol observation made use "I in discussing 

 the- "origin of the binary stars." Future progress 

 is dependent on departures from our preconceived 

 ideas. 



In chaps. \. and xi. the author is mon 

 in dealing with Ins material a large percentage 

 being his own contribution during the last twenty 

 years and we are given n number of most 

 interesting tables and results. Here are a few : 



\i leas) one- in ever) eighteen stars, on the 

 average, in the northern half of the sky, which 

 arc as bright .is o/o B.I), magnitude, is a close 

 double star visible with the 36-in. refractor. (2) 

 I'lie percentages ol double stars by magnitude 



. lasses arc : 



7'6 to 80 percentage 6 S 

 S-. „ 8-5 ., ' 5-3 



S - 6 „ 90 ., 41 



Mag. to 6*5 percentage n't 

 '■6 ,, 7-0 ,. 7-9 



r • .. 7-5 ., 7-2 



l) Visual doubles are relatively more numerous 

 in the Milky Way than elsewhere in the sky. (4) 

 Visual binaries as brighl as 05 magnitude are in 



-s amongst class {, stars, and least in K and 

 M. (5) Visual and spectroscopic binaries of every 

 spectral class increase in numbers as the Milky 

 Way is approached. (6) Spectroscopic binaries as 

 brig-ht as s'.s magnitude arc far the most numerous 

 amongst stars of spectral type B. 



To these are added, in chap. xi. the following 

 from points brought out in the previous chapters: 

 17) The considerable percentage of multiple sys- 

 tems. (8) Close correlation between period and 

 ellipticity. (9) Period and spectral type. (10) 

 Relative brightness and relative masses of the 

 two components. (11) Relatively great mass of 

 a binary compared with the sun. (12) Spectro- 

 scopic binaries of class P>, on the average, are 

 three times as massive as those of later types. 



All these and other minor points are discussed 

 in connection with their bearing on the "origin 

 of the binary stars." Of the three theories: (1) 

 Capture, (2) fission, (3) independent nuclei, the 

 author, having no alternative theory of his 

 own, favours (2). The book contains a mass 

 of interesting data well discussed. The physicist 

 as will as the astronomer will find it a real 

 treasury. It must also appeal to the wider circle 

 of our readers. 



////■; SCIENCE OF IRON-FOUNDING. 

 ('tist Iron in the Light of Recent Research. By 



Dr. W. II. Hatfield. Second edition, revised 

 and enlarged. Pp. xvii + 292. (London: 

 Charles Griffin and Co., Ltd., 1918.) Price 



12S. 6d. net. 



' I '"HIS important work on the metallurgy of 



*■ iron and steel has been enriched, in its 



d edition, to the extent of some forty-six 



S of new matter, including thirty-nine fine 



xo. 2569, vol. 102] 



reproductions ol micro-structures. The chapter 



I lie Heal Treatment of Cast Iron " has beeSn 



led SO that the annealing of grey cast iron is 



treated separately. The new matter includes 



Prof. Carpenter's valuable contribution on "The 



I licet of Working Temperatures on Parts of In- 



l-combustion Engines," the author's report 

 to the Ministry of .Munitions on "The Present 

 Position ol the Malleable Casting Industry in this 

 Country"- -an illogical bul welcome inclusion — 

 I in Influence of Sulphur in the Presence >>i 

 Silicon" (1 hap. \\.), and "The Limits of Phos- 

 phorus in Malleable Castings " (chap. xvi.). 



A careful perusal of this work suggests that 

 nut much research, recent or ancient, has escaped 

 notice, and everything worthy of note on the 

 theoretical or the quasi-practical side of cast iron 

 is included within the covers of the book. The 

 difficulty is to locate and isolate in a concise 

 form any particular information required. The 

 impression is created that the author has been 

 more or less overwhelmed by the mass of data 

 collected. This results in a sense of uncertainty 

 and a tendency to confusion in the mind of the 

 reader. The author, who is undoubtedly an expert 

 in his subject, would be well advised in any 

 future edition to set out clearly a summary of 

 each chapter, even more fully than has been done 

 in the two new chapters, which in this respect are 

 fairly well equipped. 



The portions dealing with heat treatment, 

 especially that which treats of malleable cast iron, 

 bear the true impress of authority, and when this 

 idea is conveyed by the rest of the book it should 

 become the standard work on the metallurgy of 

 cast iron as distinct from iron-founding. All 

 such works should bear the hall-mark of a con- 

 vincing personality. 



The technically trained practical man will find 

 great help in fitting himself either to meet speci- 

 fications or to account for failures. Graduates 

 and others seeking promising fields of research 

 will find them in plenty, whilst those who delight 

 in public controversy, provided they have easy 

 access to a scientific library (so that the all-too- 

 numerous references may be turned up), will be 

 in clover. In the hope that this book may become 

 the standard classic, we suggest that careful 

 attention should be paid to what may appear to 

 be matters of detail. For example, several graphs 

 do not show clearly the increments which con- 

 stitute the co-ordinates, and analyses occasionally 

 arc far from complete. 



It is to be regretted that the chapter on 

 mechanical properties is included in its present 

 form. This should be entirely re-written. 

 Even the pressure ot war-time duties does not 

 excuse the use of such terms as "breaking strain 

 in tons per square inch" and "compressive 

 strain." Admitting- that cast iron is almost devoid 

 of ductility, "breaking load" is not the correct 

 expression, \gain, in dealing with the transverse 

 test, too little attention is paid to the relative 

 value of the maximum stress under the two 



rnised standard conditions, and there is no 



