December 28, 1900. ] 



SCIENCE. 



1001 



by the same system of pipes, and the drain- 

 age of the sinks is simple and not liable to 

 get out of order. The drain pipes connect 

 with four-inch delivery pipes on each side 

 of the room, by sanitary T's, and these dis- 

 charge into soil pipes in the corners. All 

 the drainage is thus taken from the build- 

 ing bj' four pipes provided with traps, with 

 an additional sewer pipe, of course, to drain 

 the lavatories. 



The plan of the building also provides for 

 a system of high pressure steam pipes from 

 ,the university engineering shops, for blast 

 and vacuum pipes for each room, and for 

 distilled water to be prepared in the attic 

 by boiling water with the high pressure 

 steam. The distilled water is then con- 

 veyed to the different laboratories by means 

 of block tin pipes. 



There has been expended upon the build- 

 ing the sum of $55,000, leaving some of the 

 less important rooms unfinished, and the 

 furnishings in others incomplete. It is es- 

 timated that when the building is com- 

 pletely furnished, as the plans provide, the 

 total cost will be about $80,000. 



In the construction of this laboratory no 

 great originality is claimed, but the effort 

 has been made to combine the best features 

 of several of our most modern buildings, as 

 far as this could be done at moderate ex- 

 pense. So far as tested the arrangements 

 for heating and ventilation, perhaps the 

 most important points in laboratory con- 

 struction, which have some novel features, 

 seem to be very effective. It is believed 

 that greater utility can with difficulty be 

 secured anywhere at the same cost. 



E. H. S. Bailey. 



SCIENTIFIC BOOKS. 

 A Treatise on the Theory of Screws. By SlE 

 Robert Stawell Ball, LL.D., F.R.S., 

 Lowndean Professor of Astronomy and 

 Geometry in the University of Cambridge. 

 Cambridge, The University Press ; New 



York, The Macmillan Company. 1900. Pp. 



xix + 544, quarto. 



Ball's famous work was first given to the 

 world in 1876 ; later (1889), in a German treatise 

 edited by Gravelius with Ball's cooperation and 

 additions by the editor. Both of these having 

 become inadequate, the present monumental 

 publication, containing a systematic presenta- 

 tion of the present state of knowledge on the 

 subject, was undertaken and completed by the 

 original author. The theory of screws in relation 

 to rigid dynamics begins, on the one hand, with 

 the kinematic theorem of Chasles, that any 

 displacement of a rigid body may be reached by 

 a translation along a definite line called the 

 central axis, and a rotation around it ; and on 

 the other hand with the dynamic theorem of 

 Poinsot, that any number of forces or of torques 

 actuating a rigid body in any way may be re- 

 duced to a single force and a single couple (col- 

 lectively a wrench), with the axis of the latter 

 parallel to the direction of the former. The 

 reasoning thence is naturally along the lines of 

 modern geometry or of quaternions, for a screw 

 is a linear magnitude with a definite unit called 

 pitch (advance per radian) associated with it. 

 A twist thus bears the same relation to a rigid 

 body that a vector does to a point. Hence the 

 reader wishing to derive full advantage from 

 Ball's great treatise mvist be familiar with the 

 modern treatment of geometry. A good account 

 of Ball's theory is given in Schell's ' Theorie der 

 Bewegung und der Krafte' (Vol. II., Chapter 

 VIII.), as well as in Routh's ' Treatise on Ana- 

 lytical Statics.' However, such is the lucidity 

 of Ball's style, that the reader who knows only 

 the ordinary dynamic methods will find the 

 book accessible somewhere in almost all parts 

 except those specially devoted to higher geom- 

 etry. 



The chapters follow an orderly development : 

 After the fundamental principles are laid down 

 in the first five chapters, equilibrium, inertia, 

 potential, harmonic motion are successively 

 discussed in the four chapters following. There- 

 after the six orders of freedom are treated 

 consecutively in nine chapters. The eight re- 

 maining chapters deal with the higher develop- 

 ment of the subject in ordinary as well as in non- 

 euclidean space. The generality of the methods 



