XIII, A, 2 Wright and Heise: The Radium Content of Water 51 



regions near the coast of England. He obtained a mean of 1.0, 

 with a maximum range of 0.2 to 1.6. Contrary to Joly's ex- 

 perience Satterly found that his first determination on any given 

 sample was always higher than succeeding tests and concluded 

 that the most probable result was the mean after the first reading 

 had been eliminated. 



Lloyd,** in 1915, made three determinations on a sample from 

 the Gulf of Mexico, his mean result being 1.7. Like Joly he 

 also found that the first reading was slightly lower than suc- 

 ceeding ones and consequently omitted it in the determination 

 of his mean value. 



On a voyage across the Atlantic from Spain to Chile Knoche '■' 

 made several determinations by what is commonly called the 

 shaking method. The water was collected from the surface in 

 buckets and tested immediately for the emanation content, an 

 Engler and Sieveking electroscope being used. Unfortunately 

 his results are expressed in maches. As a mean of twelve deter- 

 minations on the Atlantic he obtained 0.12 mache. Joly, in a 

 summary of Knoche's work, attempts to express Knoche's results 

 in terms of the radium content per liter in grams X 10'^ and 

 calculates that 0.12 mache would be equivalent to 17 X 10'^^ 

 grams radium per liter, or, expressed in the unit used throughout 

 this discussion, the mean radium content found by Knoche for 

 the Atlantic would be 17. The value for Knoche's mean as 

 given by Joly is probably much too low. The only satisfactory 

 way of converting from the one unit to the other is to make 

 a direct calibration of the particular instrument by introducing 

 a known quantity of radium emanation. For the electroscope 

 with attached ionization chamber that we used in most of our 

 determinations on the radium content of waters, one mache 

 equals 285 X 10'^^ grams radium per liter, and on this basis 

 0.12 mache would be equivalent to 34.6 X 10"^- grams. The 

 conversion factor is dependent, however, on the constants of 

 the particular instrument and varies rapidly with variation in 

 the capacity. The factor that we obtained for our instrument 

 is lower than that given for most instruments, which inclines 

 us to the belief that the mean value of Knoche's results for the 

 Atlantic Ocean, expressed in grams radium, is much higher than 

 that given by Joly. In estimating the value of Knoche's results 



' The radium content of water from the Gulf of Mexico, Am. Joum. Sci. 

 (1915), 189, 580-582. 



' Einige Bestimmungen der aktiven Emanation des Meerwassers auf dem 

 Atlantischen Ocean, Phys. Zeitschr. (1909), 10, 157-158. 



