SECTIONAL TRANSACTIONS.— J. 883 



(d) An advertisement that is of little value when presented in black and white 

 may be made highly efficient b3'^ the use of colour. 



Dr. J. Drever. — A Hue- Discrimination Spectrometer. 



This instrument is a spectrometer which has been devised mainly for the determina- 

 tion of colour thresholds. It gives a uniform colour field, one half of which can be 

 varied by means of deviating prisms, while the other half remains constant. The 

 two fields can also be varied in brightness by means of sliding neutral glass wedges. 

 In addition, there is an attachment for mixing colours, and in particular for giving 

 the Rayleigh equation by mixing Lithium red and Thallium green so as to produce 

 Sodium yellow. 



Dr. Mary Collins. — Individual Differences in Sensitivity to Red and Green. 



The investigation to be reported was carried out with Dr. Drever's Hue-Discrimina- 

 tion Spectrometer. The first part is concerned wth the determination of thresholds ; 

 so far only those for lithium red and thallium green have been determined. It has 

 been found there are considerable individual diSerences. 



By means of a special arrangement, already described in Dr. Drever's paper, the 

 mixture in the Rayleigh equation was determined for the same subjects. They were 

 asked to equate a mixture of lithium red and thallium green to sodium yellow. This 

 was also carried out with .spectral colours on a colour miser. The object of this 

 experiment was to see if there was any correlation between the threshold results and 

 variation in the Rayleigh equation, that is, whether a big deviation from the normal 

 in the Rayleigh equation was accompanied by threshold values in red and green 

 larger than the normal threshold values. 



Mr. G. C. Grindley. — Experiments on the Functions of Rods and Cones in 

 Vision. 



Mr. S. J. F. Philpott. — Apparent Periodicities in the Work Curve. 



It is not suggested that the curves to be shown are periodic in a strictly 

 mathematical sense. A better description is wavelike or oscillatory, and it is quite 

 possible that this is all we should expect normal work curves to be. 



Apparent waves can be made manifest to the eye by appropriate methods of 

 smoothing. Most of those first met with in this research seemed to be geometric. 

 Thus, one little group of curves has crests at the 45th, 90th, 180th, 360th . . . seconds 

 from zero time. These values are in G.P. with log increment -3010. Other log 

 increments have been measured, ranging in value from -0160 to about '6000, these 

 limits being determined by the nature of the curves used. 



Occasionally, it appeared as though there were also waves of arithmetical nature, 

 but most of the early curves seem to indicate that the geometric waves are dominant. 

 On the other hand, work now in progress with more powerful methods of analysis, 

 e.g. the method of serial correlations, shows that arithmetical periods or durations 

 can be inferred much more frequently than one first supposed. The work is not yet 

 complete, but it seems that curves based on one or two experiments will usually 

 appear more geometric than arithmetic (there are cases where the converse happens), 

 but that as more and more experiments are pooled, arithmetic durations become 

 more and more obvious. The weaker curves are very variable, showing some 

 possible geometric durations. Grand total curves tend to resemble one another, 

 i.e. they approach a standard form. It is interesting therefore to note that, so far, 

 the arithmetical durations (whether seen in weak or strong curves) all seem to have 

 a duration lying between 55 and 60 seconds or simple fractions thereof (approxi- 

 mately 30, 20 or 15 seconds). The evidence strongly suggests that poohng results 

 in an averaging of geometric elements with consequent clarification of the arithmetio 

 element or elements. 



The durations or periods of geometric waves are apparently neither general nor 

 specific in the accepted sense. One cannot be sure when, or with what task or with 

 what individual any particular length of wave will appear. We can conclude therefore 

 that the geometric waves are due to some influence extrinsic to the task or the 

 permanent characteristics of the subject. What those influences are one cannot 



