]\gt ; ' On the Efficiency of Muscular Work .
and one only , which makes , viz. , when less than the angle of total reflection for the yellow and 24 ' less than for the blue ) .
At the position so defined there will be a band with borders only faintly coloured .
On either side of this baIxl.the order in which the colours appear will be reversed .
On the Efficiency of Work .
By M. GREENWOOD , Captain R.A.M.C. ( T.F. ) , Lister Institute Preventive Medicine .
( Communicated by Prof. Leonard Hill , F.RS .
Received January 18 , 1918 .
) In a paper communicated to the Royal Society in 1913 , * Prof. J. S. Macdonald published a series of observations upon the heat production of persons performing certain known ' quantities of work upon a bicycle ergometer .
In that paper , and again in a more recent publication , Prof. onald has outlined certain methods of interpreting his results , which are of much impol.tance ; these I shall discuss in the latter half of this communication , but , before doing so , it will be interesting to examine some purely numerical questions to which Macdonald 's paper gave rise .
In a note on Macdonald 's earlier paper , Messrs. Glazebrook and have published a formula descriptive of Macdonald 's numerical results .
This formula is , ( 1 ) where production in calories , body mass in kilogrammes , work equivalent in calories , , and are constants .
The values of the constants , which were obtained by a graphical process led in the particular case to the equation , ( 2 ) and this equation was found to provide values in very fair agreement with the observed results .
It will be noticed that , when the body weight iS constant , the heat production ' Roy .
Soc. Proc , vol. 87 , p. 96 ( 1914 ) .
ibid. , vol. 89 , p. 394 1917 ) .
Ibid. , vol. 87 , p. 311 ( 1914 ) .
ecause oazebrook a 's .
event , the test is a crude one , so that no claim is made that the regression ( formula is the better of the two ) .
The regression equation does not lead to oduction axceeds teatp dtion iions , fortiori , those oersons resbinga tircumqway other dataany pative vproducti b Table I.\mdash ; Constants Deduced from Data of Glazebrook and Dye 's Table II 1 ( op .
cit. , p. 313 ) .
: Capt. M. Greenwood .
204 Capt. M. Table V Table ibutio Greater the standard deviation , positive Between and negative .
Greater \ldquo ; muscular work .
The particular work he rather than in degree .
In his recent Amar , ' Le Moteur Humain , ' On the Efficiency of Muscular Work .
work of a metal filer in an hours ' day to be 61,600 kilogramme-metres ; that of a smith as rather more than twice as much .
Iaking these quantities as equivalent to a medium and a heavy day 's work , it will be interesting to compute from equation ( 6 ) the requisite energy requirements of a man weighing 67 kgrm .
Taking the round numbers 60,000 and 120,000 rammeo -metres , and reducing to equivalents , we reach 3049 oalories for the medium and 3929 calories for the heavy work .
These figures are not very different from the usual standards .
So far I have merely examined the data from the same standpoint as that of Glazebrook and Dye , viz. , that of finding a formula which shall summarise particular experimental data , and it appears that this comparatively humble task can be readily accomplished , i.e. , we can , with very little arithmetical work , construct interpolation formulae which reproduce the observations sufficiently well for the formulae to be used to approximate to intermediate values within the range of the data .
It is , however , more than doubtful whether such formulae throw light upon the efficiency of the body as energy transformer when this term is strictly used , and I now propose to examine the fundamental The customary method of the efficiency of the human body as a machine , prior to the publication of Macdonald 's paper , was to subtract from the total heat production associated with the performance of a measured amount of external work the heat production of the same subject when " " at rest\ldquo ; ; the heat equivalent of the work divided by this has been termed the " " net efficiency the " " gross efficiency\ldquo ; being the quotient of work by total heat production .
Clearly the " " net efficiency\ldquo ; will vary with the base line chosen , i.e. , whether " " at rest\ldquo ; is taken to be lying on a couch , sitting still on the ergometer , rotating its pedals without load , etc. Benedict and Cathcart* discuss the choice of base line at length in their admirable monograph , and point out the inconvenient diversity of base lines used or proposed .
But even these writers do not remark that the method is fundamentally open to attack .
Thus , if heat transformed be a function of work , the proper measure of extrinsic efficiency would seem to be the reciprocal of the first derivative of with respect to , the limiting value of the incremental change of heat with work .
This measure will differ considerably in value from the ' net efficiency\ldquo ; and may even lead to a different interpretation of the results .
For instance , Benedict and Cathcart note a tendency towards an increase in both gross and nct efficiencies with increasing loads .
With * Benedict and Cathcart , 'Muscular Work : a Metabolic Study , with Special Reference to the Efficiency of the Human Body as Machine , ' Washington , 1913 .
Op. , p. VOL. On the Efficiency of Work .
the whole of the heat production on the rest day from the day 's quota is to rate the efficiency too high .
* Macdonald 's recent paper is an attempt to transfer the inquiry to a different plane altogether .
If I have followed his extremely ingenious reasoning correctly , his position is as follows .
He holds that the efficiency of work performance is a function of body mass , and he found that when the thermal equivalent of the external work performed was divided by the efficiency ( as defined above ) and the quotient subtracted from the total heat production , the residue was , in his experiments , sensibly constant for a given range of work performance .
This residue is defined as the cost of movement , it varies with the velocity of movement and can again be expressed as a function of mass and rate of movement .
The details of Macdonald 's analysis are perhaps open to criticism .
Thus the agreement between his arithmetical calculations and the observed results is not always of such importance as appears on the face of the figures , since he is using formulae with several constants for absolutely few observations .
Thus , he has a theoretical formula for cost of movement ( Q ) in terms of velocity of movement , an expression of the form , where and are constants to be dete1mined from five observations only .
But these criticisms are of subsidiary importance and detract in no way from the suggestiveness of the argument .
Looking at the theory as a whole , the following conclusions appear to follow .
The real efficiency of the muscular machine is indeterminate and may even be unity .
The apparent efficiency of the same individual performing the same kind of work at different rates be represented by the reciprocal of the first derivative with to of , i.e. , is equal to is constant and might be called the indicated efficiency , but ( the cost of movement ) will vary with the rate : of performance , belng a minimum at the most economical rate of movement .
Thus the locus of heat produced as a function of work is a family of parallel straight lines intersecting the axis of heat at different points .
It seemed desirable to test the deduction on the basis of Benedict and 's important data .
To this end the observations upon their professional subject were sorted out into groups for each of which the rate of pedalling fell within narrow limits .
Only two such groups were really suitable for the purpose ; that already given in which the rate of pedalling * This criticism is implicit in Lefevre 's excellent discussion of the efficiency problem ' Chaleur Animale et Bioenerg6tique , ' pp. 924 , etc. , Paris , 1911 ) , which I had not seen until most of this paper had been written .
is to be noted that Chauveau 's formula also includes a velocity , see op .
, pp. 686 et seq. On the Efficiency of Work .
209 Table X.\mdash ; Observations of Table VIII Grouped for Testing Fit .
linear law , but some reservatio1ls must be made .
Thus in each the value of is chiefly due to a single group , while , careful as are the experiments , absolute uniformity of external conditions can never be secured in inquiries of this kind , so that too much stress ought not to be put upon absolutely small deviations .
I am not confident therefore that the result of this test is sufficient to condemn the ] as being inapplicable .
I also think that no significance attaches to the difference between the regression coefficients ( the reciprocals of the efficiencies ) .
The difference is , which is more than three times the probable difference , but the applicability of the normal law of error to such cases may justly be questioned ; that the divergence may be of no practical importance is by the results obtained with the next higher rates of , although these are determined from smaller numbers of observations .
They are shown in Tables XI and XII , and it will be seen that the regression coefficients do not continue to increase , but are well within the range of likely fluctuation about the value deduced from the observations at lower speeds .
Table XI.\mdash ; Heat Production and Work .
Observations of Benedict and Cathcart on M. A. M. pedalling at the rate of 103-107 revolutions per minute ( inclusive ) .
On the Efficiency of Work .
211 The algument can best be expressed symbolically .
Let be the resting heat production in unit of time , the work performed ( in thermal units ) , the real efficiency , i.e. , , the energy needed to develop units of work , and .
Then we have subject to the conditions ( i ) cannot xceed 1 and ( ii ) where is an unassigned fuqction of W. These give and attains its greatest possible value when So that until and thereafter ' hence at first , and subsequently This is , however , inconsistent with the experimental facts , viz. , that when is large , although is approximately hnear in , it is not equal to but to , and arbitrarily assumes to be given by .
Hence we must write our equation in the more general form : subject to ( i ) ( ii ) ( iii ) At and after , where is constant .
From ( ii ) we have } from ( iii ) .
On the Efficiency of .
213 being done nor is energy expended in maintaining potential , which would the case , for instance , when standing erect .
In conclusion , I note certain physiological factors which , coming into play at the commencement of muscular work , can hardly increase in intensity pari passu with the output of work , but increase the loss of heat , and therefore depress the apparent efficiency .
The first is the circulatory readjustment , leading to an enhanced flow of blood through the muscles .
Evidently this partial diversion of the stream through the viscera has an upper limit which may be reached at a comparatively low rate of muscular output .
No measurements suitable for the present inquiry have been made .
A second factor is the loss of heat due to air currents .
In such work as that on a bicycle ergometer , currents of some magnitude will be set up by the rotatory ements of the legs .
Hill , Griffith and Flack* concluded that the total heat loss from the surface of a wet katathermometer was represented where is the velocity of air movement , while and independent of V. The movement factor might help to explain the rise of heat production associated with a rotation of the pedals of the bicycle ergometer by the motor , the subject performing no work , as compared with the 's heat production while sitting still on the machine .
Benedict and Cathcart observed outputs of from calories per minute to ( rates of rotation from 58 to 104 ) under these conditions .
The range for sitting still was from to .
This cannot , ) wever , be the whole explanation .
The rate of heat loss due to this cause must diminish with since , from Hill , Griffith and Flack 's equation we have where is a constant , while Benedict and Cathcart 's data show that the rate .
of increase of heat loss is greater between 86 and 96 revolutions per minute than between 60 and 86 .
The observations are not , however , very numerous , and I hope to be able to submit in a subsequent paper the experimental results dealing with the point .
The results which emerge from the study here published and seem to me interesting ( 1 ) Within fairly wide ranges , simple formulae of linear ression describe the relations subsisting between heat production , body mass and work performance with an accuracy sufficient for such purposes as roughly computing !
: the energetic needs of workers , doing the kind work studied .
As , however , this work is of .
specially simple kind , the type of calculation is 'Phil .
Trans , vol. 207 , pp. 183-220 ( 1916 ) .
V0L .

