Mans Mechanical Efficiency in Work Performance. 399 



As a consequence the values of E formerly given must be altered from 

 (1) 3-57, (2) 3-65, (3) 3-84, (4) 457, to (1) 3-50, (2) 3"58, (3) 3-88, (4) 449, and 

 the efficiencies prevalent in the four different cases are : (1) . : 28"6, (2) 28'0, 

 (3) 25-8, and (4) 22'3 per cent, respectively. 



Expressed in terms of the " stripped weights " of the different subjects the 

 observed values of E are as follows : — 



(1) Kemp 6457 W '? 06 , 



(2) Eae 6457 W^ 703 , 



(3) Bennett 64-57 W^" 703 , 



(4) Armstrong 6457 W^ 0-706 . 



It is therefore clear that efficiency in woi'k-performance is dominated by 

 the value of the body-weight, that it is greater with the greater weight, and 

 therefore that on this account greater weight is an advantage, diminishing 

 heat production. In the particular case just examined it is found, in fact, 

 that 



E = W°- 705 /646. 



However, it is essential, when making this statement, that some indication 

 should be inserted of my opinion that this expression covers only a particular 

 problem and refers, in unmodified form, only to work upon a particular machine. 

 I shall, therefore, venture to insert in the statement a particular function 

 of the body weight, 1/P (see (7), p. 407), which is of the average value of 

 W°' 208 /4 - 04 in these particular subjects, and has a definitely particular 

 average value in this particular case. Modified by its insertion, the efficiency 

 statement is. therefore, as follows : — 



F = Wt/16 P. 



Note. — It is of interest that there is some slight evidence in these data of the 

 influence of factors of secondary importance. Thus in the data of Group D (' Roy. Soc. 

 Proc.,' loc. cit., p. 109) five separate experiments are recorded on Bennett in which his 

 rates of cycling were respectively 59 - 7, 60 - 0, 60 - l, 59 - 0, and 60'0. Notwithstanding these 

 differences in rate of movement, and the still larger consequences which ensue in the 

 rate of work performance on the brake (see above), the respective heat-productions 

 are recorded as 338, 332, 336, 333, and 338. Thus, more especially attending to the 

 fourth of these experiments, in which the rate of cycling fell to an annoying degree, 

 Bennett's temperature may explain the fact that the heat production did not similarly 

 fall. At the time he maintained that the cycle-counter was at fault, and that as a 

 fact the rate was properly continued, but his record differs on that day from every other 

 day in the following important points : (1) initial temperature 37 - 5° C, replacing average 

 37 "1° C. ; (2) rectal temperature during experiment 37 33° C. instead of 36 - 8° C. ; 

 (3) surface temperature 33'65° C. instead of 30 - 81° C, and consequently (d) difference of 

 level between rectal and surface temperatures 37° C. instead of 60° C. In short — 

 and not as an explanation — on that occasion Bennett suffered from, and complained of, a 

 heavy cold. Somewhat similar pathological interest is to be attached to the experi- 

 ment on Sharrard in the same group. 



