6 90 SCIENCE PROGRESS 



A part of this discrepancy is obviously due to the transitional stage through 

 which the book is passing in this edition, the old philosophy and the new 

 experiment not having been brought entirely together into their community of 

 interest. We are reminded of an eminent physiologist who, having sat through 

 a surfeit of papers in which one vital process after another had been represented 

 by smooth log-curves, said wearily, " It is very beautiful, and one believes it while 

 actually listening, but it can'/ be as simple as that." The fact is that we resent 

 having things made easy for us. Again, we biologists are overshadowed and 

 overawed by the giants of the past ; we cannot write on heredity without citing 

 Darwin, Lamarck, Weissman, and Mendel, and the mere mention of their names 

 induces sub-conscious prejudices. One could wish that some insignificant David 

 of biology might attempt a book on heredity which would carefully avoid making 

 a single reference to any previous investigator, and would take the facts for 

 granted, stringing the whole together into a logical sequence of interpretations, 

 beginning at the physico-chemical basis of physiology in order to examine the 

 phenomena of fluctuation, and passing through segregation and mutation to end in 

 ecology. Such a book would be anathema to the multitude, but it would give 

 to the earnest student at least one clear pathway through the wilderness. And 

 it would lay down definitions which are badly needed, especially in the New 

 World. In the present volume we find such elasticity of nomenclature that 

 while "variation" includes the fluctuation in the form of successive leaves on a 

 mulberry branch, the rigid term " mutation " is extended to include strange 

 plants which crop up in a hybrid population. This way madness lies. 



At the risk of seeming ungenerous we would emphasise the looseness of 

 nomenclature in the present volume, but remembering that it sins in good 

 company. How, for example, can the act of rogueing, even when done "on 

 hands and knees," " push a variety to greater excellence " ? the variety itself is 

 unaffected (p. 263). Such an expression as " have a definite ideal " is not 

 equivalent in significance to a conception of unit-characters, as the authors 

 assert it to be (p. 202) ; and the use of the phrase " an antipathy to crossing " 

 does not explain anything. In some places this looseness is indubitably mere 

 carelessness, as in the legend to Fig. 63, where two forms of cabbage are 

 described as " egg-shaped " and " oval " respectively ; such identity may be a 

 commercial distinction, but it is out of place in a book of this standing. 



The detailed presentment of methods for determining statistical constants 

 is a useful feature, but there seems to be a general tendency in America to 

 overrate the value of statistical expression, at the expense of graphic presentment. 

 Statistical methods are invaluable tools, and also provide a means of condensing 

 data for publication ; it is well that the student should know of them, but he 

 should be shown very carefully that in too many cases they bundle together 

 imperfectly understood data^which are not homogeneous. For ordinary work, 

 and for the greater part of biological research, data should be kept in graphic 

 form until nothing more can be made of them by inspection and dissection ; 

 smoothed curves, averages, and statistical constants should always be utilised, 

 but rarely regarded as anything but compromises. A brilliant mathematician laid 

 down the law on the subject to the present reviewer in the injunction " Never 

 smooth any curve." Also, computation takes a deal of time. 



A feature of the book which gives it a place in the reference library also is 

 the bibliography, over sixty pages long, of plant-breeding references from 1905 

 • 1912. The value of the publications included in it is necessarily variable, 

 but very few of importance seem to be omitted. 



