142 BULLETIN 623, U. S. DEPARTMENT OF AGRICULTURE. 
have been used almost universally for this purpose. This practice 
has led to the propagation of a continually increasing proportion of 
trees of those strains producing the largest amount of sucker growth. 
Inasmuch as such trees are usually light bearers and produce inferior 
fruits, this practice has been unfortunate and is the direct cause of the 
presence of the large proportion of unproductive trees found in many 
orchards. In cases where fruit wood has been used for propagation, 
such bud wood has been cut without any fruits attaqhed, so that in 
many instances buds from sporting limbs have been used, resulting 
in the development of diverse strains of trees, as shown by the presence 
of such trees in the established orchards. 
The well-known condition of frequent mixture of varieties in 
deciduous orchards, resulting from the use of bud wood cut from young 
nursery trees, where buds from different varieties are likely to be 
mixed or misplaced during propagation, caused the California 
citrus nurserymen largely to avoid this practice. For this reason 
only a limited amount of nursery growth has been used for propa- 
gation, and consequently but little mixture of varieties has been 
found in Washington Navel orange orchards. 
The presence in established orchards of trees of diverse strains is 
the direct result of propagating either from individual trees of these 
strains or from limb variations in trees of the Washington strain. 
Because this condition of bud variation and strain diversity was 
practically unknown until established through these investigations, fhe 
nurserymen who used bud wood accidentally cut from the sporting 
branches of trees of the Washington strain (see figs. 7 and 8) or even 
bud wood from trees of inferior strains are not to blame for the propa- 
gation of these diverse and frequently undesirable strains. However, 
now that the condition of bud variability in the trees of this variety 
has been established beyond question, any propagator who does 
not utilize this information and practice the improved methods for 
securing bud wood from select trees of the best strains will not be 
able to excuse the poor results of his propagations on the ground 
of the lack of definite information on this subject. 
THE ISOLATION OF STRAINS THROUGH BUD SELECTION. 
Fruit-bearing bud wood has been selected from limb variations 
occurring in trees of the Washington or other strains, and in several 
hundred cases where the growth from these buds has fruited every 
selection has come true. In the beginning of this work, the con- 
spicuous bud variations were top-worked in bearing trees in order 
to get thern in fruiting quickly. Later, as opportunity has per- 
mitted, the bud variations have been propagated on nursery stocks. 
Naturally, the results of these progeny tests require considerable 
time in order to secure reliable performance-record data. In so 
