260 'Journal of Comparative Neurology and Psychology. 



law apparatus. It is a discriiniiiatioii box iii which the two 

 boxes which have heretofore been referred to as white and black are 

 illuminated by standardized incandescent lamps. There are no card- 

 boards and difference in illumination, as desired, is obtained by 

 shifting the position of the source of light for one of the boxes. 

 This apparatus permits easy and fairly accurate measurements of 

 the absolute and relative illumination of the two boxes, and in this 

 respect it is more satisfactory than the cardboard method. Its chief 

 disadvantage is that it compels experimentation in a dark-room or 

 at least with artificial illumination of the boxes. 



In the Weber's law apj^aratus two pairs of dancers were trained 

 systematically until they had been given almost a thousand tests. 

 The individuals represent the age limits of the plasticity experi- 

 ments. The old ones, Nos. 170 and 95, were ten and twelve months, 

 respectively; the young ones, Xos. 294 and 293, were one month 

 old. Instead of a single series of ten tests per day, all these indi- 

 viduals were given two such series each day. 



To start with, all the mice possessed perfectly formed habits of 

 choosing the white box, in the old white-black discrimination appa- 

 ratus. Experiments in the Weber's law apparatus were begun with 

 the two boxes illuminated the one by 80 hefners, the other by 20 

 hefners. The difference in luminosity in this case may be stated as 

 three-fourths, since the latter value is only one-fourth the former. 

 I have found it convenient to keep one of these values constant 

 throughout a training experiment and to vary the latter as need 

 dictated. The fixed value, which may then be known as the standard, 

 is indicated in the table by the abbreviation S. The other value, 

 which may be known as the variahle, is indicated by the abbrevia- 

 tion Y. 



A habit was considered perfect in this experiment when a dancer 

 succeeded in choosing without error in two successive series. As soon 

 as ability to. discriminate a certain degTce of difference in luminosity 

 had been acquired, the amount of the difference was reduced and 

 the training continued under the more difficult condition of discrim- 

 ination. We may now examine the results of this experiment as they 

 appear in Table 12. 



