[114] REPORT 4, UNITED STATES ENTOMOLOGICAL COMMISSION. 
tree of Heaven ") fell from it paralyzed, and soon died. So, when the caterpillars at- 
tempted to cross my fence, I placed in their way, at short intervals, branches of Ail- 
anthus leaves, and killed immense numbers of them, effectually protecting my yard 
and garden. I have to suggest the expediency cf trying this native poison, so abund- 
ant and easily accessible, on the cotton-worm. I have found the common larkspur 
an effective poison on insects. Would it not answer as well for the Cotton Worm ? 
Note 55 (p. 184). — The universal belief that these two species of dog fennel are never 
attacked by any insect is without any foundation. We found a small Longicorn 
borer (larva of Mecas inornata) boring in the stem ; an unnamed species of Baris bores 
in the root, while the flower-heads are badly infested by several species of Bracliy- 
tarsus. 
Note 56 (p. 288). — The results of these experiments were published in Bulletin 3, 
Division of Entomology, United States Department of Agriculture. The machine 
had been perfected to a large extent without accurate held test of its practical work- 
ing, and in order to learn whether any improvements could be made in its several 
parts, or what faults it possessed as a working machine, as soon as news came that 
the worms were at work around Selma, Dr. Barnard was sent down with instructions 
to make the proper experiments. His report, which follows, would seein to show 
that considerable modification in the details, especially of attachment, is necessary. 
Future experiments may lead to the abandonment of the attempt to spray cotton from 
the ground up, on account of tile irregularity of the rows in the average cotton-field, 
and the adoption of lateral or oblique spraying from nozzles that do not drag entirely 
on the ground, but hang some inches above it. While the crookedness in the rows in 
ordinary cotton fields is an obstacle to the use of any complicated machinery, yet it 
is a mistake to suppose that cotton is planted so irregularly in all parts of the South. 
While it holds particularly true in the easily washed and hilly country, character- 
istic of the larger parts of the cotton-growing sections of the Carolinas, Georgia, and 
Alabama, it is by no means to the same extent true in the richer cotton-producing sec- 
tions of the South, as in the larger part of Mississippi and Texas, and the canebrake 
portions of Alabama. Hence the machine in question, as illustrated in the report, 
will prove more satisfactory in these sections, the elastic fork giving sufficient play to 
accommodate the nozzles to whatever slight irregularities in width are found in well- 
planted fields. 
The Cotton Worm machine described in the annual report for 1881-82, and now 
subjected to field tests, is shown to be suited only for cotton so planted that the rows 
are spaced apart very equally, since it lacks adaptability to the usual great differences 
of interspaces between the rows. Unfortunately, nothing very closely approaching 
ideal straightness of rows or equality of width between them can be detected in the 
South, even in such fields as are said to be " planted perfectly true." In the more 
evenly disposed cotton, stiff fork apparatus, made light and snorter, to supply only 
four rows at each drive, and hung loosely upon hooks instead of eyes, without the 
ratchet lever elevator, and capable of being easily slid by hand to the left or right, 
as infringement on row crooks from time to time required, proved susceptible of use 
with due watchfulness ; but the eight-row machine was too heavy to be thus shifted 
by hand, and being stiff-backed with rigid descending pipes, no eight consecutive 
rows could be found regular enough to be callipered for much distance by this device. 
The inflexibility also prevented conformability of the apparatus to inequalities of the 
ground, an elevation straining hard on one descending pipe, lifting the others from 
the ground, &c, and the light, flexile, jointed nozzle-arms, being borne upon severely 
by the stiff pipe system, soon became impaired, whereas they had formerly and have 
since worked well on the yielding stem-pipes of the adjustable machines which were 
tested at the time of the Atlanta exposition, as well as in these last experiments. 
For under-spraying, this old-fashioned, stiff, cross-pipe system is shown to be wrong, 
as originally foreseen, unless some power can be brought to bear to enforce a system 
of greater straightness and equality in planting cotton. A considerable amount of 
the irregularity in rows has been attributed to the " constitutional perversity and 
crookedness of the nigger," appearing from the bad execution of his instructions. 
But even if this could be corrected it is not the matter of vital importance, for the 
planter himself, as well as the field-hand, is guided by a natural principle which will 
always control and stand against any contrary theoretical or mechanical rule. Ac- 
cording to " the strength of the ground," the size of plant it will produce, will the 
rows run wider or narrower in any particular ''cut" or part of a "cat." This 
