70 



PRINTED MATERIALS, MAPS, CHARTS 



with such characteristics of the vibration as 

 frequency and pattern. 



Interactions 



As indicated by some of the results above, 

 it was found in the Tufts experiments that 

 type size, iUumination, and vibration amph- 

 tude could be separately varied over con- 

 siderable ranges without seriously affecting 

 performance. But when the factors were 

 varied together over the same ranges, un- 

 favorable values of the three in combination 

 produced marked impairment (17, 19). 



Fatigue 



The reading of numerals imposes demands 

 on attention, and probably on eye move- 

 ments, that are different from those made by 

 the reading of verbal text. It might be of 

 some importance to know whether numeral 

 reading is more fatiguing. Available evi- 

 dence is indirect. In the Tufts experiments 

 the schedule was designed to minimize fa- 

 tigue, but some indication of impairment 

 during a working period w^as obtained (un- 

 published). Bitterman (5) in a study of 

 transfer of decrement from one ocular task 

 to another, apparently had no difficulty in 

 setting up the conditions of a number check- 

 ing test to produce a definite decrement in 30 

 minutes. On the basis of present knowl- 

 edge, therefore, the indication that there are 

 only slight fatigue effects in reading verbal 

 text cannot safely be extended to numerals. 



Tentative generalizations suggested by the 

 work on numerals might be restated and 

 amphfied as follows : 



1 . So far as the material permits, tenden- 

 cies to establish integrative habits and to 

 compensate for unfavorable conditions will 

 operate in the reading of numerals just as 

 they do in the reading of verbal material. 



2. With numerals as with verbal text, 

 variables will interact to produce not easily 

 predictable results. 



3. Differences between the reading of nu- 

 merals and of verbal text include : the nature 

 of the psychological task, the relative im- 



portance of the different types of cues, and 

 probably the ocular mechanics and resulting 

 fatigue. 



4. Because of greater dependence on geo- 

 metrical details of form, the reading of nu- 

 merals should be more amenable to analysis 

 in terms of visual receptor theory. 



Tables, Graphs, Scales, and Formulae 



Washburne (55) studied various graphical 

 devices and concluded that pictographs are 

 suitable for very simple material, bar graphs 

 for somewhat more complex static compari- 

 sons, line graphs for dynamic concepts, and 

 tabulated round numbers for specific 

 amounts. Croxton and Stein (20) found 

 that bar figures can be compared most accu- 

 rately by the reader, circles and squares 

 next, and cubes least accurately. Graham 

 (24) has also analyzed illusory tendencies in 

 judging various graphic representations. 



Tables, graphs, and scales^ were compared 

 for speed and accuracy in use in a group of 

 Air Force studies by Carter (11, 12) and by 

 Connell (14). It was concluded that: 



1. Tables are preferable if the steps are 

 small enough to make interpolation unneces- 

 sary. 



2. Scales are slightly better than graphs if 

 no interpolation is required. 



3. When interpolation is required, scales 

 and graphs are about equally effective and 

 are superior to tables. 



4. The frequency of coordinate lines on a 

 graph, and whether the entry is on the x ov y 

 axis, make little difference. 



The reading of mathematical formulae was 

 studied by Tinker (45, 46, 47). He found, 

 among other things, that it is characterized 

 by "a tendency toward analytical reading 

 and a breakdown of ordinary reading 

 habits," and that formula contour can be an 

 aid to perception. 



The use of these various media enlists a 

 battery of functions beyond the simple dis- 



^ "Scales" as here used means two scales repre- 

 senting the related variables plotted on opposite 

 sides of the same base line. 



