Professor Johannseris Experiments in Heredity. 237 
As regards the weight of the beans Professor Johannsen finds 
Off-spring of the 
Minus variants Plus variants 
100.9 98-5 
this result being based on 5494 seeds ; and as regards the length 
breadth ratio (2440 seeds) 
Off-spring of the 
Minus variants Plus variants 
998 100 0 
In short (p. 39) “ the regression is complete, right back to the type 
of the line. The individual character of the mother-bean has no 
influence, nor has that of the grandmother, but the type of the line 
(der Typus der Linie)—of course in conjunction with the whole 
environment in the year concerned —determines the average 
character of the offspring.” While “ in some lines selection 
appears to be effective, in others the result is the reverse of 
what one would expect—on the whole nothing whatever is attained 
by the process of selection within the pure lines.” The writer does 
not, of course, dispute the effective character of selection in an 
ordinary population, but in such a case (summary p. 58), “ the work 
is carried out on impure material: the result depends on the more 
or less complete isolation of those lines whose types deviate in the 
required direction from the mean of the population.And it 
will consequently be easily understood that the operation of selection 
cannot be carried beyond a certain limit—it must cease when the 
purification, the isolation, of the lines which deviate most strongly 
in the required direction is, practically speaking, complete.” Later 
the author adds, to further emphasise the point (p. 64), “ selection 
in a population is only effective in many cases in so far as it picks 
out the representatives of already existing types. These types are 
not formed in succession, e.g. by protection of those individuals that 
exhibit fluctuating variations in the required direction; they are 
simply found and isolated.” These extracts give, I think, a fair idea 
of Professor Johannsen’s position; but quotations removed from 
their context are never quite satisfactory, and I hope the reader 
will refer to the original. 
The researches have been carried out on novel and carefully 
considered lines, and the results obtained are clearly of the 
highest importance for both practice and theory. As regards the 
former, they show conclusively that the breeder should proceed by 
noting the separate lines and not merely the offspring eti masse 
of the individuals showing desired characteristics; by such a 
