'1896-97.] SPATIAL THRESHOLDS OF COLOIRS. 235 



higher than those of Aubert as the ratio lo :6 might indicate. There 

 was in fact no such regularity. 



However, there was in operation another influence which would 

 materially modify the case and have a tendency to produce a compara- 

 tive increase in the characteristic thresholds for experimentation. Such 

 was the fact that Aubert and Von Wittich both permitted their 

 observers to be aware by the very method of their procedure what 

 colours to expect. The effect of such a mental preparation, undoubtedly, 

 as is well known, would be to lower the point at which the colour 

 could be known as such by counterbalancing that difficulty of decision 

 among the various possible colours which placed our observer's judg- 

 ment in suspense through a transitional region of no inconsiderable size, 

 until the colour became so definite in its characteristics as to permit no 

 further hesitation in judging it to be one colour tone as against all 

 others. Such taken in connection with the obscure influences which we 

 have noticed in the employment by the different observers of different 

 representative colour pigments which are quite incomparable, and the 

 varying play of contrast influences through the different investigations, 

 .vill account for the main discrepancies between Von Wittich's and 

 Aubert's and our experiments made under largely the same conditions. 



So much for reference to our more carelessly executed experi- 

 ments which were made solely for the sake of comparison with the 

 results of Aubert and Von Wittich. We will pass on to a brief consid- 

 eration of some of the results of less inaccurate experiment methods, 

 wherein we secured fairly pure spectral colours and confined the con- 

 trast influence within definite channels and to fixed degrees equally dis- 

 tributed over the entire series of colours under investigation. We 

 cannot, of course, give any lengthy discussion of the results in this 

 preliminary paper, but must simply indicate their main trend, referring 

 for fuller information regarding them to their subsequent presentation 

 with tables and diagrams, whereby their significance can be more easily 

 explained and more readily apprehended. 



On black ground we obtained some rather interesting results. The 

 light or achromatic thresholds show a quite decided tendency to be 

 lowest in the region around the blue and blue-green with a slight 

 tendency to be highest in the red. On the other hand, the chromatic 

 thresholds though not with perfect unanimity show a tendency to be 

 lowest in the red and highest at the other end of the spectrum. This 

 fact is not contrary to what we might be naturally led to expect, 

 because r^^is a colour which emerges as light, no sooner than it is visible 



