December 14, 1917] 



SCIENCE 



589 



the Eoyal Society, 186 A, p. 343. These curves 

 include sis types besides the normal curve, but 

 the book makes no reference to the five addi- 

 tional types of curves recently published by 

 Pearson.^ The book presents a treatment of 

 the correlation of two systems of variates. The 

 treatment is, in general, clear, and should 

 serve a useful purpose in making better known 

 to persons who are applying these methods to 

 data the nature of some of the limitations that 

 underlie the interpretations of correlation co- 

 efficients. However, the reviewer has one 

 criticism to offer. On p. 155, using r for the 

 correlation coefficient, we are told that " it 

 seems doubtful whether any serious meaning 

 can be attached to values of r which are less 

 than .5." It seems to the reviewer that this 

 statement should be modified. To be sure, the 

 statement would hold if the correlation coeffi- 

 cient r were calculated from such a small num- 

 ber of observations that the probable error of r 

 is not particularly small compared to r. But 

 when the conditions under which the formula 

 for probable error of r is derived are well satis- 

 fied, r may be much smaller than 0.5 and have 

 decided significance if derived from large 

 enough number of observations to make its 

 probable error small in comparison to the value 

 of r. 



A useful chapter is devoted to harmonic 

 analysis from the standpoint of least squares, 

 including an interesting section on a practical 

 method of investigating periodicities. The 

 last chapter deals with the periodogram, in- 

 cluding a treatment of hidden periodicities. 

 H. L. KiETZ 



Universitt of Illinois 



SPECIAL ARTICLES 



THE PRODUCTION OF GASEOUS IONS AND 



THEIR RECOMBINATION 



Gaseous ionization has played a large part in 

 recent advances in both physics and chemistry. 

 In the ordinary college- and high-school 

 courses given in these subjects little, if any, 

 attempt is made, however, to demonstrate 

 methods of producing gaseous ions or of meas- 

 uring their recombination or diffusion con- 



iPhil Trans., 216 A, p. 429. 



stants. Practically no laboratory work along 

 these lines by elementary students is attempted. 

 This may be explained in part by the fact that 

 most investigators in this field of research have 

 made use of the electrometer, an instrument 

 well adapted for demonstration purposes but 

 inappropriate for use by the inexperienced stu- 

 dent. An electroscope of very simple design 

 has, however, proved entirely satisfactory in 

 place of the more cumbersome and possibly less 

 sensitive electrometer. 



Some elementary experiments are suggested 

 in the first part of this paper using apparatus 

 involving little or no expense and which may 

 be assembled by any high-school student. This 

 is followed by a description of some results ob- 

 tained in verification of the law governing the 

 recombination of the ions of a gas. 



PART I 



The type of electroscope used is shown in 

 Fig. 1. 



Fig. 1. 

 A brass rod passes through a sulphur plug 

 into the hollow cylindrical chamber (C) 12 

 cm. high and of 4 cm. radius. On this rod is 

 mounted a flat brass strip which supports the 

 gold leaf. The top of this mounting projects 

 through a large opening in the square metal 

 box surrounding the gold leaf to permit the 

 electroscope to be charged by removing the 

 metal cap (.D). {B) is a brass tube approxi- 

 mately 2 meters long the radius of which will 

 depend upon the laboratory facilities for pro- 

 viding a suitable current of gas. If air ioniza- 

 tion is to be studied and compressed air is not 

 available, a suction pump attached to a water 

 faucet will provide a convenient velocity for 

 carrying ionized air through (5) if its radius 

 is of approximately 3 cm. diameter. 



