116 BULLETIN 624, U. S. DEPARTMENT OF AGRICULTURE. 
be blamed for the unfortunate and undesirable propagations of poor 
strains which have been so generally made. However, now that 
bud variability in trees of this variety has been established defi- 
nitely, any propagator who does not carefully select his bud wood 
from trees of the best strain on the basis of their performance records 
will not be able to excuse the poor results of his propagations on the 
ground of the lack of information on this subject. 
THE ISOLATION OF STRAINS THROUGH BUD SELECTION. 
In the beginning of these investigations fruit-bearing bud wood 
was selected from limb variations in trees of the Valencia and other 
strains and top-worked in bearing trees in order to get them into 
fruiting quickly. About 100 such propagations have come into 
bearing, and as far as can be judged from trees of their age they have 
in every instance reproduced the strains which were propagated. 
Later, as opportunity has permitted, the bud variations have been 
propagated on nursery stocks. These progeny tests will require 
several years to secure reliable performance-record data, but in so 
far as such trees have fruited the results have confirmed those 
secured by top-working bearing trees. 
Enough evidence has been secured in these investigations to war- 
rant the assertion that all of the 12 strains described in this bulletin 
can be isolated through bud selection. It seems probable that other 
strains also can be isolated in the same manner. 
This conclusion does not include the idea that bud variation within 
these strains can be entirely eliminated. On the contrary, the in- 
vestigations have shown that variation will likely continue as long 
as the strains are propagated. What has been demonstrated is that 
variation can be controlled by bud selection to such an extent 
that the individuals of a strain can be brought to a condition of 
practical uniformity as regards crop production and other 
characteristics. 
Bud selection should be based on definite knowledge of the in- 
dividual tree, which can be secured most satisfactorily by means of 
performance records and careful study. A thorough investigation of 
most of the established Valencia orchards in the light of individual- 
tree records will reveal the necessity for greater care in selecting buds 
for use in propagation. 
The bud wood found most satisfactory in these investigations is 
that bearing typical fruits of the strain desired and of the one or two 
periods of growth preceding the production of the fruits. One or 
more fruits of the desired strain should be cut off with each bud 
stick as an indication or label showing the probable character of 
fruit that will be produced by the buds immediately back of the fruit. 
Figure 9 shows a bud stick cut in this manner with typical fruits of 
