36 
Proceedings of the Eoyal Society of Edinburgh. [Sess. 
It is then seen, with the average subject, that the COg-proportions are 
higher, especially at and beyond the crest-load, and that the crest-load 
itself is greater than when normal air was breathed. Oxygen-want, which 
is the main factor limiting the duration and intensity of physical exertion, 
is staved off by breathing enriched air. 
By experimenting on a large number of healthy men ranging in type 
from the athlete in perfect training to the ultra-sedentary person, it was 
found that the higher the degree of fitness the less the A and B curves 
diverged, and, indeed, that when exceptionally fit men were tested, the 
resulting graphs were similar to those of fig. 2, where the curves for all 
practical purposes are coincident to the crest and only show divergence at 
FT. LBS. PER MIN. 
0 3000 6000 9000 12000 15000 
FT. LBS. PER MIN. 
the over-loads. It was also discovered that if the same man were kept 
under observation for several months and were tested at different states 
of health, or at intervals during a course of physical training, the B curve 
would remain constant (within the limits of experimental error), but the 
A curve would vary in form and position according to the state of health 
or of training. Fitness, which may be defined as the efficiency of oxy- 
genation of brain, heart, and muscles during exercise, is therefore inversely 
as the extent of divergence of the two curves, and can be evaluated. 
By drawing upon the graph the horizontal line ac (fig. I) at the level 
of the expired-C02"percentage at rest, and then measuring the crest 
ordinates ah and cd, the fitness factor can be expressed as ah-^cd. Thus, 
fig. I, which records the data for a sedentary man, gives his fitness as 
46 per cent, while the curves of fig. 2, which are those of an athletic 
sergeant-instructor in physical drill, show his fitness to have been 100 
per cent. 
If stamina be defined as the power of dealing with sustained exertion. 
