REVIEWS 519 



ENGINEERING 



A Manual of Machine Design. By Frank Castle, M.I.Mech.E. [Pp. x + 352, 

 with 207 figures and illustrations.] (London : Macmillan & Co., 1919. 

 Price 7s. 6d.) 



This book will be hailed with delight by the hundreds of technical teachers 

 throughout the country. It has the distinct merit of being written by a teacher 

 of great technical standing for students who are bound hand and foot by 

 examination results. It is true that the Board of Education has made efforts 

 to break down the examination bogey, but it has not carried its efforts to a 

 logical conclusion. While it abolishes the greater number of its own examinations 

 and encourages committees and teachers to conduct their own affairs, it allows, 

 without at least a public protest, numerous bodies, unions, and associations, to 

 be established, with the one aim and object of conducting examinations. 



When the students of technical institutions are examined under conditions 

 that obtain in our teaching universities the uses of Mr. Castle's book will be 

 practically nil, and teachers will fall back once more on those types of books 

 which point clearly to the value of research, to the value of experience, which 

 make students think, and which, more than anything else, indicate that two 

 designs, totally different, may accomplish a set end and both be economical, 

 efficient, and theoretically and practically sound. 



Such books also indicate that sound design invariably leads to a certain 

 degree of standardisation chiefly owing to limitations imposed and conditions to 

 be fulfilled. 



A student knowing Mr. Castle's book thoroughly would, without any question, 

 pass certain examinations with credit. Therein, as we have said, lies the merit of 

 the book. 



The same student would more than likely prove a distinct trouble in the 

 drawing office. Take a few examples. 



A young designer is asked to design a knuckle joint as given in Example No. 6, 

 p. 71. The rod works out to 178 in. diameter, or i}| in. There is not a 

 suggestion of the young designer looking up makers' sizes to see if i£f in. is 

 a stock size of rod, much less looking up the stock-book of the particular firm 

 concerned. 



Now iff in. is of course rolled, and would, perhaps, be in stock at the makers', 

 but the chances are that ii in. would prove the nearest local size. Not a 

 suggestion of this important point is given in the book. 



Then, again, Example 16, p. 113, is for the design of a rolled steel girder to 

 carry a brick wall over a 20 ft. clear span. 



The design finishes with a flange 6 in. wide. Not a hint is given as to the safe 

 pressure per square foot for brickwork, or as to the wisdom, or otherwise, of 

 carrying a brick wall on a surface 6 in. wide. 



The last example we propose to take is the question of the design of bolts. 

 Here the product of Mr. Castle's book may be really dangerous. Working 

 stresses in the examples given vary between 3,000 lb. and 5,000 lb. per sq. in. 



Not a word of warning is given that for a black f in. diameter bolt on a rough 

 joint (a kind frequently used in practice) the lower value (3,000 lb. per sq. in.) 

 would be much too high — indeed, Mr. Castle does not discriminate between 

 rough and finished joints nor between large and small bolts. 



It is with relief that we think our serious designers have been guided by 

 Unwin and Seaton, and we suggest that in the revision of the next edition the 



