would be easier when the detection of saturation was difficult, whereas looking for a saturation difference would be easier when both colors are saturated to an obvious degree.
(C) The long recognized difficulties of spatial and temporal separation of compared samples were evidently present in this study where widely separated parts of a chart so often must be compared. Contrast and end-effects due to proximity of samples and terminal positions have already been mentioned.
(D) Finally, the visual estimates of the colors were all made with the samples in the regular positions and patterns of the Munsell system. This condition in itself would serve as a resistance to change. The untrained subject, or even the trained subject in a doubtful case, would tend to be affected or guided by the system (56); therefore, slight adjustments were probably not made and larger adjustments may not have been great enough. Still, this resistance to change might be considered advantageous in the sense of producing conservative estimates of the revisions required.
In the face of all of these difficulties what did the observers actually do? Only the more typical tendencies can be mentioned here. (a) The series of colors, or color differences, to be judged was first surveyed as a whole and the observer essayed a decision with respect to rank order for the attribute in question. Often no progressive graduation was found; and instead of arriving at an order, the observer was able to identify a few extreme colors and a number which seemed about equal with respect to the attribute in question. (b) After some such preliminary survey, the observer might proceed to the assignment of vectors. A common practice was to draw in vectors representing the more extreme displacements first, and then, partly on the basis of those, to fill in the shorter vectors indicating the lesser irregularities. (c) There was the tendency to "carry" a standard subjectively even though some particular perceptual standard had been chosen. Occasionally, the subject might refer to the standard sample for verification, but would often make many comparisons without doing so.
This "absolute" procedure (20, p. 205; 84) and the ranking procedure (20, p. 244; 81) mentioned above were both hit upon by many subjects without instruction, probably because they facilitated the subject's difficult task. Certainly it is easier to decide which of two sensory magnitudes is greater than to estimate how much greater; and it is sometimes easier when making numerous comparisons to carry a subjective standard than to refer back each time to a perceptual one.
Evidently, it would be a mistake to conclude that the ratio method has been applied to the spacing problem in anything like pure form. The actual procedure was so complex that if the ratio method were not pointed out it might pass unrecognized. An advantage of the complicated instructions is that they call for so thorough an examination of the charts that marked irregularities in spacing can scarcely be overlooked.
Observers
Since there were 7 constant-value charts and 20 constant-hue charts used in the survey, and since each of these 27 charts was prepared with the three backgrounds, a total of 81 charts were to be examined, each chart requiring about 100 judgments. The task was exacting, and those observers who returned complete sets of data from observation of the constant-value charts, the constant-hue charts, or both, usually took a year or more to do so. Without substantial contributions of this character, the successful completion of the investigation would have been impossible.
Data were received from forty-one subjects. Complete or nearly complete returns were secured from the following: W. H. Beck, Marion Belknap, B. R. Bellamy, Elizabeth Burris-Meyer, Claire Dimmick, F. L. Dimmick, Dean Farnsworth, Loraine Fawcett, F. A. Geldard, W. C. Granville, Paul Henry, M. M. Jackson, D. B. Judd, K. L. Kelly, D. L. MacAdam, S. M. Newhall, Dorothy Nickerson, R. W. Russell, Walter Scott, L. L. Sloan, W. T. Spry, Samuel Talbot, Irving Taylor, J. Weitz. Partial or supplementary observations which were summarized with the others were furnished by the following: Genevieve Becker, H. P. Gage, S. R. Gilmore, I. H. Godlove, G. W. Haupt, R. S. Hunter, H. E. Hussong, L. A. Jones, E. M. Lowry, Alexander Murray, P. Nutting, M. R. Paul, F. H. Rahr,R. H. Sawyer, Albert Smith, W. B. VanArsdel, K. S. Weaver.
