7:5/ Neural Aspects of Vision 133 



rates of firing of m and p, or else color sensation would (on this model) 

 vary rapidly with light intensity. For specific color sensations, these 

 ratios must be compared with standards. For large patches of light of 

 the same color, these standards must exist within the nervous system. 



To reconcile the model with Land's experiments, it is necessary to 

 assume that with small randomly distributed patches of colors, the 

 nervous system computes an average value for each ratio and then 

 compares the ratios to this average rather than to absolute standards. 

 Teleologically, this would be desirable because it permits one to dis- 

 tinguish colors independently of the exact spectrum of the lighting used — 

 clouds in the sky, and so on. The model of Figure 2, with this added 

 assumption, predicts correctly that two broad bands of light, illuminating 

 two slides, should be able to produce all possible color sensations. Two 

 narrower bands cannot excite as many different values for the ratios of 

 the responses of the m and p cells to the response of the s cells, and hence 

 cannot simulate all colors. 



This interpretation emphasizes that the model of Figure 2 uses three 

 types of receptors and is thereby a tricolor model. However, the data 

 from these three are analyzed as one absolute value, used for acuity and 

 brightness sensations, and as two ratios used for color sensations. One 

 might well ask if the added assumption is valid. Although more experi- 

 ments are necessary to answer this question, it may be noted that at any 

 rate the assumption of average standards instead of absolute ones is, 

 a priori, no more unreasonable than the possibility of neural sharpening. 

 (As recently as 1950 the latter was considered unlikely.) 



Another question one might raise is whether the nervous system uses 

 the same standards for the mjs and p/s ratios throughout the visual field. 

 Land's experiments show that people identify colors correctly with three 

 different pairs of broad bands of light in three different parts of the 

 visual field. Teleologically, this also is desirable because it allows one 

 to recognize colors as the same, some of which are in direct sunlight and 

 others in shade. 



Whether the model of Figure 2 continues to prove useful, or needs to 

 be drastically revised, the experiments described in this section indicate 

 that the nervous system carries out many complex, computer-like actions. 

 As with most actions of this nature, the exact neural mechanisms are not 

 well understood even though the evidence for their occurrence is very 

 strong. 



5. Cortical Representation 



The complex series of synapses of the visual pathway through the 



