18 BULLETIN 360, U. S. DEPARTMENT OF AGRICULTURE. 
976, from the branch 321. On very old brooms of the western larch 
it is often noticed that the needles begin to turn yellow some time 
before those on the branches of uninfected trees. Exactly the re- 
verse may occur in the case of recently formed brooms, owing to 
the larger amount of newly stored food materials in the swelling 
on the main branch and the branches of the brooms. That the 
broom may be the cause of a great localization of food substances 
is indicated by the fact that in heavily infected Douglas fir and 
larch the last part of the tree to succumb is usually the smaller and 
younger brooms of the tree. Frequently trees of these species are 
noticed with only a single small broom living, the rest of the branches 
being apparently dead; likewise the old and exhausted brooms. The 
increase in the number of needles on the broom due to the multi- 
plication of its branches is usually at the expense of the needle de- 
velopment on the normal parts of the tree. For this reason an 
excess of food materials for the tree as a whole does not take place. 
The foliage beyond the broom becomes thin and, in most cases, 
the end of the branch dies (figs. 12 and 14). The food materials 
are entirely stored and appropriated by the broom itself. The 
phenomenon is analogous to the formation of spiketop of the main 
trunk. | | 
That brooms do not always necessarily mean an increase in foliar 
surface for the host, since we have seen that parts of the branches 
not supporting brooms frequently die, is shown by a comparison of 
the needles of old brooms with those of normal branches either of 
_ the same tree or of uninfected trees. Such a study was made in the 
case of the Douglas fir. It was found that the needles of the brooms 
on the trees studied were uniformly a little less than one-half as long 
as the leaves of the normal branches (PI. IV, fig. 1). Neither were 
they as thick or as broad. By compensation it would be possible to 
determine approximately the actual folar surface of a given broom 
and compare it with that of a given normal branch of the same 
whorl and of the same age. This difference in the size of the needles 
was found to hold good only in the case of old, mature brooms of trees 
which were beginning to be suppressed. Young brooms, especially 
on young trees from 10 to 20 years old, often have abnormally long 
needles on the still upright branches, but this condition is not long 
maintained. Soon these branches begin to droop, the broom be- 
comes denser, the needles disappear from the center outward, and 
they are often sparingly distributed along the stems but more densely 
assembled on the last few years’ growth (fig. 13). With continued 
suppression of the Douglas fir and exhaustion of the broom, a new 
type of branching often appears. The long trailing, weeping-willow- 
like branches cease to elongate and the cortical stroma of the parasite 
is enabled to catch up with the terminal bud and killit. The branch 
