642 



SCIENCE 



[N. S. Vol. XLVII. No. 1226 



absolute values, the coefficients of variability 

 and of correlation as relative values. It seems 

 to the reviewer that the latter, rather than the 

 former, are absolute values; for they are the 

 ones that are independent of the units of 

 measurement used. The point is not one that 

 is likely to cause confusion, as it is at all 

 times clear what the authors mean. 



In the chapter on species hybridization the 

 " reaction system " idea is discussed at some 

 len^h. According to this hypothesis, which 

 has been developed by Goodspeed and Clausen 

 on the basis of their species crosses with 

 tobacco plants, the whole group of genetic 

 materials of a species (the " reaction system " 

 of that species) may behave as a single imit, 

 or nearly so, in influencing dominance and 

 viability in hybrids. The term " reaction sys- 

 tem" seems to the reviewer to be rather tm- 

 fortunate, as it is commonly applied to any 

 system that is imdergoing change, organic or 

 otherwise. There can be no question of the 

 very great interest of the facts that have been 

 discovered by Goodspeed and Clausen, but in 

 the opinion of the reviewer more detailed evi- 

 dence is needed, especially as regards the 

 cytological behavior of the Fj^ hybrids, and 

 the genetic behavior of the plants produced 

 by back crossing the Fj^ to both parent species, 

 before the conception can be adopted as more 

 than an interesting suggestion. 



In parts 2 and 3 a large amoimt of data 

 bearing on the genetics of domestic animals 

 and plants has been brought together, and has 

 been presented in a thoroughly scientific man- 

 ner. This makes these sections useful also 

 to the non-agricultural geneticist. To the 

 practical breeder these sections should be in- 

 valuable, not alone because of the genetic data 

 they contain, but also because of the discus- 

 sions of methods of securing and recording in- 

 formation, and of the practical application of 

 genetic knowledge. The chapter on beliefs of 

 practical breeders is especially noteworthy; it 

 gives in concise and convincing form the evi- 

 dence against telegony, maternal impressions, 

 and similar notions sometimes held by 

 breeders. 



The subject of eugenics is treated only very 



briefly and incidentally, and even then with a 

 word of warning as to the reliability of the 

 conclusions reached. No attempt is made to 

 take advantage of the great popular interest 

 in eugenics by exaggerating the importance 

 and significance of the results that have been 

 reported. 



The authors are to be congratulated on a 

 book that is well printed, well illustrated, well 

 written, and that contains a surprisingly large 

 amount of material that is conveniently ar- 

 ranged and adequately presented. 



A. H. Sturtevant 



SPECIAL ARTICLES 

 A NEW GRAPHICAL METHOD FOR COMPARING 

 PERFORMANCE WITH PROGRAM OR EX- 

 PECTATION 



The graphic method is generally recognized 

 as a most important means to interpret facts 

 to administrative executives as well as to the 

 public. For the sake of simple comparison 

 the well-known bar diagram (besides many 

 other diagrams) is, of course, frequently used, 

 but the latter, especially in the case of a va- 

 riable delivery or production against time and 

 a iixed quantity (requirements or expected 

 production), loses its value. This character- 

 istic rigidity of the bar diagram permits analy- 

 sis of only one particular instant of the situ- 

 ation, with no reference to the past or future. 

 For instance, if the total output of flour of a 

 certain milling division starting from January 

 1, 1918, up to, let us say February 9, 1918, is 

 expected to be 1,314,000 barrels, and up to 



so. OOP 6hU - /OO^ ^. 



!. 14% . I 

 ourrt/r up ro J^f/ s. f& 

 BS7.oaol,6U. • /■f% 



Fig. 1. 



January 5, 227,000 barrels of flour have been 

 manufactured, the situation expressed by a bar 

 diagram (Fig. 1), using the expected output 

 as a 100 per cent, basis, would be : 



