BUD VARIATION IN THE MAESH GEAPEFEUIT. 109 
certain trees he decided to top-work those trees with select buds 
from trees of the Marsh strain. The senior writer was asked to select 
several desirable trees of this strain as sources of bud wood. Five 
trees were selected for this purpose and the bud wood was cut by 
men having had long experience in securing bud wood for propaga- 
tion. After several hours' work one of these men approached the 
senior writer, who was working in the investigational performance- 
record plat near by, and said that they had been unable to obtain 
enough bud wood from the five trees originally selected for this 
purpose. He explained that, in order to secure an adequate supply, 
they had cut some bud wood from a neighboring tree. He was 
asked whether they had examined any of the fruits from this sixth 
tree, and he replied, "No, we never cut any fruits, but the tree looks 
just as good as the other five trees." 
A doubt as to whether or not the additional tree was a desirable one 
led to an inspection of its fruits. The first fruit cut was found to con- 
tain more than 100 seeds. The production of this tree was low and 
the fruits of very inferior quality. Further examination showed that 
this tree was a typical representative of the Seedy strain, the very 
strain, in fact, which the workmen were preparing to eliminate by 
top-working. Inasmuch as the bud wood from the seedless and the 
seedy trees had been mixed together in the bundle of bud sticks, it 
was necessary to secure an entirely new lot of buds from other trees 
of the Marsh strain. It is interesting to note that the 118 trees of 
the Seedy strain top-worked in that orchard at that time have now 
all come into fruiting, and without exception the fruits borne by the 
growth from the buds of the Marsh strain are all commercially seedless 
and have the valuable characteristics of the fruits borne by the trees 
from which the buds were taken. 
It can easily be seen from this experience how trees of the undesir- 
able strains in established orchards of Marsh grapefruit have been 
unintentionally propagated under the old system of cutting bud wood. 
This experience also offers effective evidence of the need of great 
care in bud selection, which can be most accurately accomplished 
through the use of individual-tree performance records and definite 
tree knowledge. 
THE ISOLATION OF STRAINS THROUGH BUD SELECTION. 
The results of the propagation work in these investigations have 
proved beyond a reasonable doubt that the Marsh strain of grape- 
fruit can be established and maintained through the careful selec- 
tion of bud wood from select trees of the Marsh strain. In the 
course of these investigations many propagations have been made 
from select Marsh strain trees. Several commercial orchards have 
been propagated in this way. So far as these trees have come into 
