Extended series of observations with this apparatus and with trained subjects have shown that this measure of visual efficiency is remarkably constant from day to day and shows a standard deviation of approximately 4 per cent of the average threshold. The normal series for 1917, previously referred to, contains threshold values for 61 subjects, which range from 38" to 106" on the arc of vision. The average was 68.7", with a standard deviation for the 61 men of 18.5", which is 27 per cent of the average threshold value. It should be noted that good vision was one of the factors operative in the selection of the normal group of men, 25 of whom (somewhat more than 40 per cent) show thresholds of 60" or somewhat less.1 It is something of a surprise to find that in the first normal measurement for this threshold on Squad B, the average of the 10 men is 58.2", that is, 10" lower than the average for the normal group of prospective aviators, the personnel of which was supposedly selected with some reference to keen vision. Naturally this factor of vision was a matter of pure chance in the selection of the personnel for Squad B. These first averages show almost the same range as the normal series of i917, that is, the range for B is 41" to 110". The latter figure (the record for Mac) is not included in the average. If included, it would raise the average somewhat, but not to the level of the average for the normal series of 1917. It is regretted that no normal records are available for Squad A; their first measurement for this threshold was on October 13; this was the first observation after food reduction, which began on October 4. The average for the 10 men on this date is 86.2", and would not be materially different if Spe and Fre were included. According to the only available normals, this seems to be high.

1 In the test-letters on charts used to determine acuteness of vision, the width of one stroke or of the space between two strokes composing a letter is the fundamental width or unit for vision. A determination of 20/20 means that the subject distinguishes the direction of lines the stroke width of which occupies 60" on his arc of vision. Similarly, as recorded on page 175, the width of one dark band or of one light band, or to state it in another way, since the boundary between dark and light is not sharp in our test object (see figure 51), the distance from the center of one dark band to the center of an adjoining light band was taken as the unit width in terms of which the visual efficiency was given. A threshold of 60" signifies that the subject became aware of lines in the test field and could correctly indicate their direction (vertical, horizontal, etc.) when the unit width, from center to center of adjoining dark and light bands equaled 60" on his arc of vision. In a personal communication Captain Percy W. Cobb, M. R. C, states that workers at the Nela Research Laboratory have regularly employed this same unit with these test objects and believe that the results thus expressed are in closest conformity with the regular Snellen units. Dr. Cobb states further that 20/20 is quite common as an ophthalmologic finding and that in persons who have no defect or are well corrected, it is not uncommon to get 20/15 or better, i. e., 45" or less as visual angles.

Table 175. - Squad A - Visual Efficiency And Its Mean Variation

[Values given in seconds on the arc of vision].

Date.

Acuity.

Bro.

Can.1

Kon.

Gar.

Gul.

Mon.

Moy.

Pea.

Pec.

Spe.

Tom.1

Vea.

Fre.

Av.

1917.

Oct. 13..

M.

94

108

..

43

104

91

56

133

86

42

89

58

100

86.2

M.V.

5.3

13.3

..

3.9

9.6

5.3

6.0

21.5

8.4

1.9

11.9

5.1

7.2

9.03

Oct. 27..

M.

47

76

71

46

99

83

74

90

87

41

80

51

..

73.3

M.V.

4.5

231.0

7.6

6.9

9.3

6 5

7.2

10.9

7.4

2.6

11.9

6.8

..

10.24

Nov. 10..

M.

43

91

69

46

103

89

77

101

80

237

83

47

..

76.0

M.V.

4.1

220.0

10.7

4.3

6.3

5.3

7.6

15.3

6.2

2.8

11.9

3.2

..

8.42

Nov. 24..

M.

40

77

73

45

100

76

84

93

81

37

82

45

..

72.3

M.V.

3.0

230.0

5.6

2.6

5.9

8.2

5.0

13.1

5.8

2.5

13.0

4.2

..

9.08

Dec. 8..

M.

41

81

75

41

91

68

86

105

86

36

78

44

..

72.1

M.V.

2.6

226.0

9.0

2.3

8.1

3.4

8.4

19.0

5.5

3.6

12.5

2.7

..

9.05

Dec. 19..

M.

42

68

75

43

88

67

83

68

78

..

76

50

..

66.3

M.V.

3.6

27.3

6.2

3.0

4 1

7.6

6.6

0 3

8 1

..

11.0

6.2

.

8.68

1918.

Jan. 12..

M.

38

75

80

41

87

76

75

85

77

..

76

42

..

67.2

M.V.

2.0

22.0

3.7

3.2

4.7

8 7

6.2

9.6

6.4

..

11.8

5.4

..

8.00

Jan. 26..

M.

36

68

68

42

90

65

68

78

84

..

71

40

..

64.2

M.V.

1.3

27.0

6.0

1.4

5.0

5.5

6.7

8.0

8.4

..

10.1

2.5

..

7.59

Feb. 2..

M.

36

75

69

41

91

74

86

62

78

..

69

39

..

65.1

M.V.

1.4

24.8

6.1

1.8

4.0

3.2

8.4

23.0

6.7

..

6.4

5.5

..

8.52

Low-diet av.

M.

46.3

79.9

72.5

43.1

94.8

76.6

76.6

90.6

81.9

38.6

78.2

46.2

100.0

71.4

M.V.

3.1

24.6

6.9

3.3

6.3

6.0

6.9

14.4

7.0

2.7

11.2

4.6

7.2

8.73

1 Subject commonly wore glasses but not when tested.

2 In the vertical axis Can has a threshold of about 48": it is his pronounced astigmatism, therefore, that accounts for large M. V. of this subject. 3 The subject's father was a visitor in the room when these measurements were taken.

The individual values in tables 175 and 176 are, in each case, the average of from 12 to 20 threshold determinations, distributed about equally between the four axes which were used, as outlined on page 175. The mean variation computed by taking all of an individual's s observations on one evening and considering them as one group, is partly a measure of astigmatism as well as of true variability. The number of observations made at each axis was too small for satisfactory determination of a separate mean or standard deviation. The average mean variation, so-called, is found to be 8.7" for Squad A (see lower right-hand corner of table 175). It is somewhat smaller in the case of Squad B, being 5.3" for their normal measurements and 4.8" for their low-diet averages. (See table 176).

Table 176. - Squad B - Visual Efficiency And Its Mean Variation

[Values given in seconds on the arc of vision].

Date.

Acuity.

Fis.

Har.

How.

Ham.

McM.1

Kim.

Lon.

Mac.2

Sch.

Liv.

Sne.

Tho.

Van.

Wil.

Av.

1917.

Nov. 3..

M.

50

52

41

46

67

..

57

110

..

79

78

68

51

60

58.2

M.V.

2.3

2.9

2.6

3.7

4.2

..

5.8

10.5

..

9.2

12.5

4.4

3.3

6.3

5.30

Nov. 17..

M.

44

84

63

45

62

..

83

97

..

85

45

80

47

72

64.8

M.V.

3.1

8.7

6.8

3.3

5.1

..

9.4

7.7

..

5.0

3.4

7.8

5.8

6.2

5.95

Dec. 15..

M.

..

78

41

41

55

..

72

88

..

84

82

87

42

60

65.2

M.V.

..

9.3

1.7

1.9

5.7

..

5.5

7.5

...

8.4

5.3

3.7

5.2

8.6

5.51

1918.

Jan. 5..

M.

44

81

49

345

64

(4)

..

..

89

91

46

91

43

58

61.0

M.V.

2.8

9.8

3.5

1.6

5.7

..

..

..

6.5

5.3

2.9

3.7

5.3

4.4

4.37

Normal av.

M.

46.0

73.8

48.5

44.3

62.0

..

70.7

98.3

89.0

84.8

62.8

81.5

45.8

62.5

62.3

M.V.

2.7

7.7

3.7

2.6

5.2

..

6.9

8.6

6.5

7.0

6.0

4.9

4.9

6.4

5.28

Jan. 13..

M.

40

90

43

40

..

64

70

..

81

78

70

83

42

67

62.3

M.V.

2.2

4.9

9.7

7.2

..

7.8

6.1

..

4 9

5.5

8.6

7.6

2.5

7.6

6.19

Jan. 19..

M.

41

80

38

36

..

53

74

..

83

89

38

84

40

61

58.1

M.V.

1.5

6.3

1.8

2.8

..

4.4

4.7

..

5.8

6.4

2.9

5.5

1.7

5.2

3.88

Jan. 27..

M.

40

74

39

44

...

54

72

..

86

83

532

79

40

62

56.5

M.V.

1.9

4.0

3.4

7.9

..

4.6

4.5

..

7.0

6.1

1.9

4.6

3.5

4.0

4.18

Low-diet av.

M.

40.3

61.3

40.0

40.0

...

57.0

72.0

..

83.3

83.3

46.7

82.0

40.7

63.3

59.0

M.V.

1.9

5.1

5.0

6.0

..

5.6

5.1

..

5.9

6.0

4.5

5.9

2.6

5.6

4.75

1 Subject commonly wore glasses but not when tested.

2 The right eye was used in the test as with the other subjects, but Mac informs us that he has better vision with his left eye. 3 Ham complained that his eyes were tired from reading on the train while coming to Boston. 4 This was the first session for Kim and this measurement was not given for lack of time. 5 Sne said he could see the fixation dot better in this test than in any previous experiment.

In figure 115 the mean variations (see the two lower curves) are very consistent from experiment to experiment, with a slight increase for Squad B at the time of the first reduction date (January 13). The curve for Squad A is definitely and consistently below that for Squad B, with two slight depressions on October 27 and November 24 and a tendency to smaller variations near the close of the series of experiments, but these fluctuations are certainly not larger than might normally occur.

The average thresholds for the two groups, as expressed in degrees on the arc of vision and shown in the upper curves in figure 115, maintain from October 27 and November 17 about the same relative levels with continuous improvement. The wide and opposed variations in the thresholds shown by the two squads at their first measurements cannot be satisfactorily explained. In the case of Squad B, the rise in the threshold, shown on November 17, is of course partly due to the records for Har, How, and Lon, as shown in table 176. However, the average threshold does not improve on December 15, but is, in fact, slightly higher on this date, for some of the subjects, particularly Sne, show an increase. Considering Squad B's successive averages for November 17, December 15, January 5, and January 13, it seems very probable that the average threshold for November 3, the first time this measurement was taken, is very low for some chance reason which is not revealed. This conclusion seems further justifiable in view of the fact that this threshold for Squad B, as was pointed out earlier, is below the average for the normal group of 1917.

With Squad A the first average threshold is very high, and indeed, by present standards, seems abnormally high. It is of course associated with food reduction, although on this particular date (October 13) the subjects received an average of 1,993 net calories per man, an amount which was 250 calories above the average for the 9 days during the low-diet period on which the subjects of Squad A came to Boston. An examination of table 175 for Squad A would indicate that this high threshold for October 13 is due principally to the subjects Pea, Can, and Bro. The first designated subject was poor in threshold measurements, as will be seen by examining his record for the electrical threshold. Can commonly wore glasses and apparently found some difficulty in adjusting himself to taking the test without them. No explanation can be given for the relatively high threshold of Bro, who later in the series showed very consistent results and also a very low threshold; furthermore his mean variation was usually small, the average being 3.1 seconds. The higher threshold for October 13 may be partly due to the food reduction. The poor result for November 10 corresponds with the lowest average net energy figure for any of the experimental dates (see table 146). These correspondences do not definitely prove that the reduced diet raised the threshold for visual efficiency with Squad A, and as the results for B give no clear substantiation, the findings appear to be negative. The difference in level between the two squads can hardly be assigned as a low-diet effect, although if each squad were several times larger, this might be significant.

Visual efficiency and its mean variation. Solid lines represent Squad A, and broken lines Squad B.

Fig. 115. - Visual efficiency and its mean variation. Solid lines represent Squad A, and broken lines Squad B.