66| 



PRINTED MATERIALS, MAPS, CHARTS 



increases, performance scores improve at a 

 decreasing rate. On the question of a criti- 

 cal level the evidence is conflicting. If we 

 minimize the controversial data on bhnk rate 

 and heart rate, the remaining findings, 

 though not in good agreement \Yith. each 

 other, suggest a value somewhat above the 

 three foot-candles for short reading periods — 

 perhaps 10 foot-candles.- If we include a 

 common-sense safety margin of 10 foot- 

 candles, practical recommendations based 

 only on those studies in which the subjects 

 actually did normal reading under good con- 

 ditions would come to 13 foot-candles for 

 brief, and 20 foot-candles for long, reading 

 periods. 



An extreme illumination situation is met 

 in the use of faint red light to preserve dark 

 adaptation. Reading of books and maps is 

 alleged not to be difficult under this condition 

 (Keil, 28) but quantitative information is 

 meager. 



Fatigue 



Subjective impressions, and analogy with 

 other forms of work, support the hypothesis 

 that sustained reading is accompanied by 

 significant amounts of fatigue. However, 

 the phenomenon has proved pecuharly elu- 

 sive under experimental attack. The mono- 

 graph by McFarland, Holway, and Hurvich 

 (31) reported an extensive exploratory study 

 of the mechanisms of visual fatigue which 

 failed to find grounds for a consistent general 

 theory. Several other studies mentioned 

 above in the discussion of problems of illumi- 

 nation found that reading was accompanied 

 by changes in certain functions (4; 30, pp. 

 107, 360; 31, 50), but usually under only the 

 lower illuminations or in terms of the meas- 

 ures most subject to dispute. The most 

 systematic positive findings are those re- 

 ported by Hoffman (26), who took electrical 

 records of the eye movements of 30 subjects 

 as they read continuously for four hours 

 under 10.5 foot-candles. From an analysis 

 of the records, indices were computed for 

 blinks, fixations, regressions, lines read 



(speed), fixations per line, regressions per 

 line, and variability of fixations. By the 

 end of the four-hour reading period, all of 

 the indices showed significant changes, except 

 regressions, which fluctuated irregularly. 

 Carmichael and Dearborn (10),^ on the other 

 hand, in a very similar experiment in which 

 more extensive data were gathered, obtained 

 negative results. High school and college 

 students read book print for six-hour periods 

 under 16 foot-candles. The entire procedure 

 was repeated with microfilm, results for which 

 are discussed below. For the book reading, 

 neither eye movement records analyzed for 

 the same indices as in the Hoffman experi- 

 ment, nor comprehension measures obtained 

 on the average every 25 pages, showed re- 

 liable changes over the six hours. Acuity, 

 measured before and after, showed no 

 changes that could be clearly assigned to the 

 effect of reading. Heart and brain- wave 

 records, not fully analyzed, showed no obvi- 

 ous relation to the reading task. Subjective 

 impressions gathered at the end of reading 

 included a number of reports of discomfort, 

 but these the authors interpreted as related 

 rather to the restrictions in posture than to 

 the reading itself. It appears that any phys- 

 iological changes produced by the large 

 amount of ocular work involved in prolonged 

 reading can readily be compensated for by 

 the subjects, and the degree of discomfort 

 does not seem to be excessive. 



The foregoing comments on typography, 

 illumination, and fatigue in ordinary reading, 

 taken as a whole, permit the conclusion that 

 optimal conditions are within the range of 

 current facilities and practices. New de- 

 partures are not indicated, with the possible 

 exception of higher illumination, the need 

 for which is under debate. From the stand- 

 point of methodology, certain considerations 

 relevant to the evaluation of other visual 

 situations can be formulated. 



First, since variables tend to interact, com- 



3 This reference contains a bibliography of 409 

 titles. 



