Observations on the Allison Magneto-Optic Apparatus

Abstract

The Allison magneto-optic apparatus has been subjected to critical tests by the checking of minima reported from other laboratories, by the working of unknowns and by the distribution of scale readings over a given range. About 300 of Allison's minima have been checked. Two sets of three unknowns, prepared in such a way that there was one chance in six of a fortuitous correct solution, have been successfully analyzed. One set in which three salts were chosen out of a list of seven was correctly analyzed after the three had been mixed. The chances were 34 to 1 against a lucky selection of the correct combination. A second set of three out of seven, similarly run, resulted in the identification of only one constituent. The overall average shows 10 out of 12 correctly reported. All unknowns were given to the observer at concentrations of 1 part in ${10}^7$ of water and they were diluted to 1 part in ${10}^{10}$ before subjecting to analysis. A statistical study has been made of the distribution of 1698 scale readings, made over a period of three years by five different observers. The readings were plotted against the number of times they were observed and definite peaks are evident in the curve. 84.2 percent of all readings fall within the limits of experimental error of a peak, and 30.1 percent fall on a peak. This is contrary to the results of Slack and of MacPherson, who found only chance distribution.