698 



SCIENCE 



[N. S. Vol. XL. No. 1034 



Pearson's tables for statisticians and 

 biometricians 



When one is told that the advance of sci- 

 ence is in a high degree dependent upon im- 

 provements in technique, one naturally thinks 

 of the astronomical and physical instruments 

 of precision, the calorimeters of the chemist's 

 equipment, the microtomes and microscopes 

 of the general biological laboratory, and of the 

 pure culture and surgical technique of the 

 clinic. With such magnificent examples of 

 instrumental facilities for research, it is easy 

 to forget the large debt of modem science to 

 mathematical methods of description and an- 

 alysis. Even if one limits oneself to the cases 

 in which the mathematical tools have taken 

 the most workable form — that of tables of final 

 constants for given value of observation or 

 tables to facilitate the calculation of such 

 constants — the debt is enormous. Who can 

 estimate the service to applied science of the 

 engineer's pocket books of formulae and tables ? 

 or the value to pure science of the convenient 

 volumes of logs and trigonometric functions? 

 or, to be both specific and modern, of such 

 volumes as the " Physikalish-chemische Ta- 

 bellen " of Landolt and Bornstein, the tables 

 for physicists and chemists of Castell-Evans, 

 and the " Annual Tables of Constants and 

 ISTumerical Data" published by the Commis- 

 sion of the International Congresses of Ap- 

 plied Chemistry? 



The most recent advance of this kind is 

 marked by the publication of a series of tables 

 for the use of statisticians and biometricians. 



With the foundation of the Biometric School 

 of Biology, there were available only the gen- 

 eral aids to calculation — tables of logarithms 

 and trigonometric functions. Barlow's tables'- 

 and Crelle's.^ 



All of these are still to some extent useful, 

 though the improvement of calculating ma- 

 chines has rendered them less indispensable. 



1 Barlow 's tables of squares, cubes, square roots, 

 cube roots and reciprocals of all numbers up to 

 10,000. liondon, Spon. 



2 Dr. A. L. Crelle 's calculating tables giving the 

 products of every two members from 1 to 1000 and 

 their application to the multiplication and division 

 of all numbers above one thousand. Revised by 

 C. Bremilver. New York, Steekert. 



The tables' before us, carefully designed as 

 they are to meet the needs of a special group 

 of students, are in a very different class. To 

 workers in the difficult field of higher statistics 

 such aids are invaluable. Their calculation 

 and publication was, therefore, as inevitable 

 as the steady progress of a method which 

 brings within the grip of mathematical analy- 

 sis the highly variable data of biological ob- 

 servation. The immediate cause for congrat- 

 ulation is, therefore, not that the tables have 

 been done but that they have been done so well. 



In the original prospectus of Biometrika, 

 the editors promised to provide " numerical 

 tables tending to reduce the labor of statistical 

 arithmetic." Since 1901, when the first of 

 these tables was published in Biometrika, the 

 responsible editor has had but one end in view, 

 the publication, as funds would permit, of as 

 full a series of tables as possible. 



A detailed list of the tables which have re- 

 sulted from the grim perseverance in this 

 determination for the past fifteen years is 

 superfluous. Such fundamental series of con- 

 stants as Sheppard's " Tables of the Prob- 

 ability Integral," Elderton's Tables of 

 Values of P for Pearson's x^ Test of Good- 

 ness of Fit of Theory to Observation, 

 Everitt's " Tables of the Tetrachoric Func- 

 tions," Ehind's " Criteria for Frequency Ts^pes 

 and Probable Errors of Frequency Constants," 

 and such convenient aids to calculation as 

 Miss Gibson's values of X^ and X. and Soper's 

 1 — r^ to lighten the labor of the calculation 

 of the probable errors are only sample titles of 

 the fifty-five sets calculated by Bell, Duffell, 

 Elderton, Everitt, Gibson, Greenwood, Heron, 

 Lee, Pearson, Ehind, Sheppard, Soper, " Stu- 

 dent " and Whitaker, which cover with great 

 completeness the whole field of statistical 

 description and analysis. 



The convenient volume in which these are 

 now brought together contains something over 

 75 pages of explanation and illustration and 



3 ' ' Tables for Statisticians and Biometricians. ' ' 

 Edited by Karl Pearson, F.E.S. Issued with as- 

 sistance from the grant made by the Worshipful 

 Company of Drapers to the Biometric Laboratory, 

 University College, London. Cambridge Univer- 

 sity Press, 1914. 



