February 12, 1920] 



NATURE 



645 



holism of 133 1 Calories. The regression formula jjives 

 i486 Calories! 102 ; the surface rule 1620 Calories ± 130. 



Neither prediction is close to the truth, but that 

 of the regression equation is decidedly superior. On 

 the whole, there is no doubt that the regression 

 method is preferable, and that the tables provided by 

 the authors are of value. In general, however, the 

 difference is not such as to give rise to any apprehen- 

 sions that the use heretofore made of I he surface rule 

 has led to practical inconvenience. Prof. Lusk recently 

 wrote : — " For the study of metabolism processes it is 

 certainly most fortunate that the unit of surface area 

 eliminates the same amount of heat in the normal 

 adult within 10 per cent, of a determined average. 

 The reason is not clear." 



This remark is not invalidated by Drs. Benedict and 

 Harris's results, although thev have provided a some- 

 what superior criterion. It is proper to notice this 

 point, as our authors are a little too prone to lecture 

 their physiological colleagues upon the real meaning 

 of scientific "laws," and, as we think, exaggerate the 

 importance which the surface "law" assumes in the 

 minds of those who have employed it as a convenient 

 working rule. 



Space would have been saved and the statistical 

 results and methods might have been brought to a 

 sharper focus had a number of sententious generalities 

 been omitted. .\n instance occurs on p. 148. Tho 

 authors properly remark that surface must be less 

 variable than mass (when Meeh's formula is used), 

 ;ind imply that it is a direct arithmetical consequence 

 that heat per surface unit is less variable than heat 

 per mass unit. This really needed further investiga- 

 tion. The coefficient of variation of an index is a 

 function not only of the coefficients of variation of 

 its constituents, but also of their correlation; the sub- 

 stitution for one constituent of another less variable 

 constituent might not reduce the variability of the 

 index were the correlation also greatly reduced. In 

 the present case the two correlations are nearly equal, 

 .ind the authors could have actually made their point 

 more securely had they explained the theory of the 

 matter more fully. 



Were physiologists really so ill-acquainted with 

 statistical methods as the authors hint, the remarks 

 on p. 16 with respect to the "probable error" of the 

 mean of a sample of four might be dangerous. But we 

 think that British physiologists are quite rdive to the 

 importance of biometric methods, sufficiently so to 

 congratulate Drs. Benedict and Harris upon the con- 

 clusion of a laborious task which has yielded results 

 of appreciable value to other students of metabolism. 



The second volume describes two series of experi- 

 ments upon squads of college students, originally 

 twelve in each sqiiad. In squad .\, diet providing 

 approxima.tely 2500 Calories and 13 grams of nitrogen 

 was maintained for four months; in sauad B, a still 

 lower diet, yielding about q grams of nitrogen and 

 1500 Calories, was used for three weeks. The former 

 squad lost on the average 12 per cent, of the initial 

 body-weight, and the basal metabolism per square 

 metre per twenty-four hours fell from 1)40 to S17 

 Calories. Squad B lost 65 per cent, of the average 

 body-weight, and the basal metabolism also declined 

 considerably. The urinary excretion of nitrogen in 

 both series was not so variable as might have been 

 expected, and did not decline, pari passu, with the 

 diminished intake. By the end of the experiment on 

 squad A, an average loss of 175 grams of nitrogen 

 (based on nine men continuously observed) was in- j 

 fcrred. J 



In the course of the experiments intervals of un- I 

 controlled feeding occurred, viz. during the Thanks- ! 

 giving celebrations, November 29 to December 2, 1017 ! 



NO. 2624, VOL. 104] 



and during the Christmas holidays, Decemljer 20, 19 17, 

 to January 6, 19 18. On alternate Sundays the diet was 

 usually uncontrolled. A very large number of anthro- 

 pometric, physiological, and psychological measure- 

 ments were made upon these men. In addition to 

 the decline of basal metabolism above noted, both 

 pulse-rate and blood-pressure diminished. Little ob- 

 jective evidence of a decrease of ])hysical efficiency 

 was obtained. The majority of the psychological tests 

 pointed to a diminution of efficiency, or at least to a 

 lack of improvement with practice. The students' 

 college work, on the other hand, did not appear to 

 suffer. 



The value of this research, which •erifies upon a 

 relatively large scale the practicability of diminishing 

 both the rate of metabolism and the quota of 

 body nitrogen by a simple reduction of intake, should 

 be appreciated by scientific clinicians, while man\ 

 issues of physiological interest are raised and lines of 

 advance suggested to the pure physiologist, especially 

 in connection with the study of levels of nitrogenous 

 metabolism. 



.\s a contribution to the science and art of national 

 dietetics, this elaborate — one might almost say over- 

 elaborate — study is not of so much importance. It is 

 shown that a great reduction, both of available energy 

 and protein, may be borne for some months by healthy 

 adults without either immediate breakdown or sign 

 of enduring deterioration; these are facts which in- 

 voluntary experiment on a vast scale had already 

 demonstrated. The correlation between variations of 

 nutrition and of resistance to infection, which 

 "common sense" has postulated and which the vital 

 statistics of Europe seem to substantiate, must engage 

 the attention of future investigators; .it oresent an 

 adequate experimental technique is lacking. We do 

 not think that Dr. Benedict and his associates have 

 obtained any results either invalidating tho general 

 conclusions expressed by the Food (War) Committee 

 of the Roval Societv or filling up lacunae in our know- 

 ledge of the general subject shown by that committee 

 to exist. The experimental and statistical study of 

 national dietetics is still in its infancy. M. G. 



ISOSTATIC COMPENSATION IN THE 

 EARTH'S CRUST. 



T^WO articles on isostasy by the late Prof. Joseph 

 -•■ Barren, which appeared in the American Journal 

 of Science for October, 19 19, contain what may be 

 regarded as his mature views on the subject. The 

 first, entitled "The Nature and Bearings of Isostasy," 

 was a summary of six lectures delivered by Prof. 

 Barrell at Columbia L^niversity in 1916; it gives a 

 general account of the theory of isostatic compensa- 

 tion, and of the methods of investigation which have 

 led to the recognition of the phenomenon. The second 

 article, "The Status of the Theory of Isostasy," was 

 written just before the author's death; it vindicates 

 the theory as to attacks which have been made on it 

 bv MacMillan and others, and describes the various 

 views which are held as to the degree of perfection of 

 isostatic balance existing in Nature. Hay ford's 

 general conclusion is maintained : that surface in- 

 equalities of contour and mass are accompanied bv 

 inverse inequalities of density beneath the surface, 

 so that at a depth of about 120 km. equal areas have 

 equal masses superposed; but a view different from 

 Hayford's is taken as to the exactness of this com- 

 pensation. 



Hayford estimated the greatest departure from com- 

 pensation as being 250 ft. above or below the level for 

 perfect balance over an area of one square degree on 



