SECTIONAL TRANSACTIONS.— J. 607 



deficiencies. The errors which cause financial loss to printing firms are usually 

 minute misjudgments of hue and saturation. The more elaborate spectrometer 

 investigations of individuals demand a skilled technique, are fatiguing to the subject, 

 are dilticult to interpret in terms of norms, and take too long to be an economic 

 proposition as selection tests for industrial purposes. 



Further, existing methods are not flexible. A printer works under varying 

 conditions of illuminatiop, and in lights of continuous varying spectral composition. 

 His colour judgments may be made under every disadvantage of adaptation, simul- 

 taneous or successive contrast affecting the surfaces he is comparing, or fatigue. 

 The problem was to devise a simple portable test, interesting and involving a situation 

 similar to the ordinary colour matching operation, and with such a test to investigate 

 individual differences in the abilit}' to discriminate small differences of saturation and 

 of hue, and in the ability to match colours. Further, to find the effect on these 

 fimctions of — 



1. Variations of the absolute and relative illumination of test surface and back- 

 ground. 



2. Background colour (contrasting or similar). 



3. Education in colour matching. 



Other allied problems are absolute memory for colour, and the connection between 

 ability to discriminate a difference of colour and ability to specify what the difference 

 is. A printer must know how to correct the colour of his print. 



Various methods were considered and rejected, either on the ground that they 

 did not lend themselves to the investigation of these problems or else that the necessity 

 for a psychophysical method involving tedium and loss of interest to the subject would 

 tend to vitiate results. 



The final test took the form of sets of small coloured discs, prepared by a technique 

 ■which ensured that a step-by-step transition in saturation occurred through the set, 

 even though individual steps were too small for certain discrimination. Three sets 

 are being tried, blue, yellow, red, and each set consists of two duplicate series of 

 sixteen samples. 



In the test the subject first sees a single series scattered at random on a table in 

 a dark room, under a standard and uniformly illuminated field of approximately 

 ■daylight quality. The series is set on a neutral background, but any coloured back- 

 ground can be used. The first test is to arrange the series in order of saturation. 



In a second test the subject is given the duplicate series and is asked to match each 

 member with its duplicate in the first series, which is again arranged at random. 

 The samples are numbered one to sixteen, and the results are scored by taking the 

 sum of all departures from true order in the first test and from true matching in the 

 second. A high score indicates low discriminative ability. 



The validity of the series is assured by analysing the placings of the discs. If 

 in a table of placings or of matches by the accumulated subjects the mode of the 

 frequency distribution of the placings of each disc is in its proper place (j.e. if sample 4 

 is placed most often in place 4, or matched most often with No. 4 of series 2) then 

 correctness of gradation can be assumed. 



Though the tests have not yet been carried out on enough subjects to generalise 

 xesults are encouraging. Out of the first ten tested, two individuals have been found 

 to be far superior to the rest in every variation of the tests. Tests are still proceeding 

 ■and hue-gradation series are being constructed. 



Miss E. M. Yates. — Experimental Work upon Transfer of Training. 



1. Problem of Transfer. 



The doctrine of general training or formal discipline states that the effect of 

 training in any specific form of mental activity may be transferred to any other activity 

 of the same form, although dealing with different material. It may be expressed 

 more simply : ' How far does the training of any mental function improve other 

 mental functions ? ' 



2. General form of experiments on Transfer. 



The essential order of experiments may be summarised thus : (a) test, (6) training 

 in similar but not identical material, then (c) further test. There are two groups of 

 subjects, the first, known as the trained group, is given ' a,' ' b ' and ' c,' the second is 

 given ' a ' and ' c,' but not ' b,' (the training between the two tests). 



