
864 THE AMERICAN NATURALIST. [Vor. XXXIX. 
be by means of a branch originating on the twig below the 
gall. At first it appears as if the crook in the stem (see Fig. 2) 
might provide for a direct growth by means of a branch from 
the convex side of the stem. Dr. P. Speiser claims that a bend 
in the stem is of advantage in this way in the gall Dichelomyia 
rosaria H. Lw., since the gall is turned aside, and thus permits 
a more or less direct growth from a bud below it. An examina- 
tion of any willow bearing these crooked galls (S. graphaloides) 
shows very plainly that the benefit for continued growth is acci- 
dental rather than habitual. To be of service it would be nec- 
essary that a bud be located just at the beginning of the crook 
in the stem. In one lot of 57 galls, 32 had buds so located and 
25 were without bud in a favorable position. By watching the 
willows for further information, I found that very frequently the 
favorably placed bud either does not develop or sends off a 
branch at an angle from the parent twig. And just as fre- 
quently, when there is no bud so favorably located, a bud far- 
ther down, or one immediately beneath the gall, is the one which 
sends out the most vigorous branch, whose direction is deter- 
mined by its surroundings. 
THE GALL MAKERS. 
The life history of the gall gnat is very simple. In the spring 
the larva which has spent the winter in its cell in the gall trans- 
forms into the pupa and that, a little later, into the gall gnat 
itself. The gnats soon deposit their eggs in apical buds of wil- 
lows. From the eggs larve are hatched and the cycle begins 
again. Walsh speaks of observing larv as late as April 23, and 
pupæ from early in April until late in May. From galls gathered 
and examined on the following dates I found larvze 
gnats: October, November 12, 
ruary, about March 20, and the 
of the gall 
December 26, January 17, Feb- 
! i first and second weeks of April. 
In March and April, many larvee were almost ready for trans- 
forming into pupæ but only a few pupæ were found. Gnats had 
emerged from all the galls examined June 3. I kept some of 
the galls gathered in December and on subsequent dates, to 
watch for transformations. From those of December 26, ine or 
