52 
MendeV s Laics o] Inheritance. 
followed out in a more complex case. This may be considered 
to be more or less typical of those which will be met with in 
practice. 
In the first case the characters already existing in the 
parents are transferred in their full intensity to the fixable forms 
among the descendants of the hybrids. Where intermediates 
occur, these are incapable of breeding true to type, and they 
always throw off the types which produce them originally. 
Recessives breed true from the moment of their appearance 
among the offspring of the hybrids ; but, on the average, only 
one in three of the individuals showing the dominant feature 
breeds true : the stable forms can only be distinguished by 
raising a further generation from them. 
Where more than one pair of characters are present in the 
parents, new forms are produced by the recombination of these 
characters into fresh groupings. This is illustrated in the 
series in Fig. 3. One parent was a club wheat with a beard, 
the other had lax ears and no beard (Fig. 3, PP). There are 
thus two sets of units to be considered, namely, beards and no 
beards and lax and dense ear-shapes. 
On crossing these varieties together, the hybrid obtained 
was beardless and intermediate with regard to the shape of the 
ear (Fig. 3, F i). In spite of the fact that more than one 
pair of characters is present, the results are the same as before, 
the beardless condition being dominant and the meeting 
of the lax and dense giving intermediate. The offspring of 
the hybrid consisted of the following types : — 
Bearded lax, beardless lax, bearded intermediates, beardless 
intermediates, bearded dense and beardless dense (Fig. 3, i-6). 
These occurred in definite proportions, the lax, intermediate, 
and dense being present in the ratio of 1 : 2 : 1, whilst on the 
average there were three beardless to one bearded of each of 
these types. 
No other types beyond these were produced, no entirely 
fresh features like, for instance, roughness of chaff making 
an appearance. 
The data already obtained from an examination of the 
previous cases will show how the behaviour of each of these 
six types may be predicted when a further generation is raised 
from them, and incidentally it will show how the breeder will 
have to treat such cases in practice. 
Beginning with the intermediate types our previous 
experience would indicate that they would break in the next 
generation into lax and dense (Fig. 1). On sowing, it is found 
that both the bearded and the beardless forms break in this 
way, so that it is policy to throw out all these intermediates 
at once, as fixity of type should be essential in all breeds. 
