STATE PLANI' BOARDL)
ET-163 October 1940
United States Department of Agriculture
Bureau of Entomology and Plant Quarantine
A MECHANICAL TRAP FOR THE SAMPLING OF AERIAL INSECT POPULATIONS,
By Joseph C. Chamberlin and F. R. Lawson, Division of Truck Crop
and Garden Insect Investigations 2
The trap described herein was developed in its original form
during July 1935 in the course of investigations on the flight
habits and dispersal of the beet leafhopper (Eutettix tenellus
(Bak.)). It was devised to meet a need for quantitatively sampling
the numbers of leafhoppers actually in flight. The trap has been
employed principally in studies of the flight habits and dispersal
of the beet leafhopper and the pea weevil (Bruchus pisorum (L.)).
The essential principle involved in the construction of all
models of this trap is that power is so applied to one or more
rigid insect nets as to cause them to rotate in a fixed horizontal
plane at a speed sufficient to capture and concentrate in a remov-
able collecting bag any insects which might be intercepted.
A trap using essentially this same principle was independ-
ently developed by Williams and Milne3 and Davies4 and described
under date of Deceober 1935. The Williams and Milne trap used
A brief, popular description (apparently a news reporter's
account) of this trap appeared under the heading "Patents, Proc-
esses, Inventions" on page 96 of the August 1939 issue of Science
Digest, Vol. 6, No. 2. See also: Cody, Chas. E. Rotary Trap Aids
in Collecting Field Data on Insects. Market Growers Journal
(Louisville, Ky.) 65 (12): 512, illus. (Dec. 15, 1939.)
2 The writers are indebted to George T. York, of the Division
of Truck Crop and Garden Insect Investigations, who helped design
and build the trap shown in figure 5, and to Alfred A. Barney, of
the same division, for suggestions and help with the mechanical
details of several models.
3 Williams, C. B., and Milne, P. A. A Mechanical Insect
Trap. Bul. Ent. Res. 26 (4): 543-552, illus. 1935.
4 Davies, W. Maldwyn. A Water-Power Mechanical Insect Trap.
Bul. Ent. Res. 26 (4): 553-558, illus. 1935.
electric fans mounted in the nets for propulsion; the Davies adap-
tation of the same trap used a water wheel for power in much the
same fashion that we employed an electric or gasoline motor. Wil-
liams and Milne and Davies used open-mesh cloth for nets, while we
used 16-mesh wire screen, which we found much more efficient as
well as more durable. Both the above papers mention difficulties
caused by insects remaining in the net. We have not had this
trouble, all insects caught being blown into the small collecting
bag. This may be due to the speed of net rotation. Williams and
Milne used a speed of only four and one-half and Davie- used five
revolutions per minute. The speeds employed by us ranged from 50
to 67 r. p. m. The length of the net arms was about the same for
the two types of traps. The comparatively high speed adopted by
us also prevented insects from dodging the moving nets and from
"rawliing or flying out of the net after being caught. Neither
Williams and Milne nor Davies mention any such difficulty, but,
as they were working with aphids, it would not be expected that
this would be a factor. Those authors had trouble keeping the speed
of rotation constant in variable winds. No difference in speed
could be noted in our traps in winds from 0 to 20 miles per hour
when the net speed was set for a minimum of 55 to 60 r. p. m. The
estimated cost of the Williams and Milne trap is about the same as
9f the one here described.
A trap built upon the plan here presented will efficiently
sample aerial insect populations, with the possible exception of
certain swiftly flying Odonata, Diptera, etc. It offers a quanti-
tative method of determining the density of insects at different
elevations and at different periods of the diurnal and annual cycles
relative to such factors as time, temperature, humidity, light
intensity, and so on. Accurate records can be obtained for periods
as brief as 5 minutes or less or for as much longer as desired. If
the interval between the taking of collections is not too long, the
insects caught are generally not seriously damaged and are for the
most part in suitable condition for preservation as museum spec-
An excellent illustration of the type of data which can be
obtain',d '.'ith this machilie is contained in a paper by Barnes,
Fiy'er, and KaloosLian,s who used one of the machines built by the
present writes. Each net of this trap (net opening, 2 square
feet in area; diameter of circle described by net, 12 feet) will
theoretically collect the insects from approximately 248,000 cubic
feet of air per hour at a speed of 55 r. p. m.
s Barnes, Dwight F., Fisher, Charles K., and Kaloostian,
George H. Flight Habits of the Raisin Moth and Other Insects as
Indicated by the Use of a Rotary Net. Jour. Econ. Ent. 32 (6):
859-863, illus. 1939.
The cost of the machine, exclusive of the motor, will range
from $50 to $70 if built by hired labor. The machine can be built
by any good automobile mechanic or machinist.
The writers have collectively and individually been respon-
sible for the building of six different traps. All were alike in
essential features. Probably the best trap, mechanically, ,'.as one
of those constructed by the senior author at Corvallis, Oregon, for
use in studies of pea weevil flight. This model will be described
in some detail, with references to other models. It is obvious that
great variation in structural detail is possible within the broad
limits imposed by the required performance.
The essential mechanical features of the trap are illustrated
in the accompanying illustrations (figs. 1 to 7).
The net assembly comprises one or more nets, together with
supporting radial arms and adjustments. Where only one net is
employed, it must be suitably counterbalanced. Each net is mounted
at the en'2l of a braced, 6-foot arm, so that in rotation the nets de-
scribe a circle 12 feet in diameter. The entire net aseiiembly may
be either heavy bolted wooden construction, as shown in figure 1,
or more lightly made of 1-inch welded angle iron, as in the model
shown in figure 2.
Since the speed of the motor is much greater than that of the
net, and the axis of rotation of the motor is in a horizontal plane
while that of the nets is in a vertical plane, it is necessary that
there be an intervening set of gears. In the traps illustrated, a
cutdown rear axle of a model T Ford car, mounted L fashion on the
net support, serves this purpose. Any automobile rear axle should
be equally adaptable. One axle is cut off just beyoc',' the differ-
ential housing, and a plate is then welded across the casing to
make it oil-tight. The hub of one of the wheels of the :&ea axle
assembly is then employed for mounting the net arms, while the
universal transmission joint of the drive shaft serves the same
purpose in the trap as in the original automobile. Because of the
light load, a light crankcase motor oil replaces the usual heavy
lubricant in the differential. The whole unit is very durable and
The machine is driven by a transmission shaft extending from
the net drive shaft to the motor and clutch assembly, which is
mounted beyond the sweep of the rotating nets. A flexible connec-
tion should be used to connect the motor or drive shaft and the
The motor and clutch assembly (figs. 3 and 4) comprises a
small gasoline or electric motor. We have employed 1 to 1 horse-
power gasoline motors and to I horsepower electric motors with
complete satisfaction. Where electric power is available its use
is preferable, since electric motors require less attention and are
more reliable than gasoline motors. If a single-phase electric
motor is employed, the use of a clutch is unnecessary, as the trap
can be started and stopped by turning a switch. With split-phase
motors the clutch must be used to avoid throwing too heavy a start-
ing load on the motor. One net is about the maximum load for a
-horsepower motor. We have used a --horsepower motor to pull two
nets. The power required depends somewhat upon the efficiency of
the gears employed.
If the automobile transmission is used, one or more reduction
pulleys or gears between the motor and clutch are nr.cessary in order
to achieve the required net speed. The exact nature and size of
these reduction pulleys or gears will vary with the normal motor
speed and the desired speed of rotation. Where belt transmission
is used, slippage must be allowed for if a particular net speed is
Two distinct but equally satisfactory motor and clutch
assemblies are shown in figures 3 and 4. Others equally satis-
factory from a mechanical standpoint can no doubt be improvised by
any good mechanic. The use of a small manufactured clutch assembly,
such as is used in machine and woodworking shops (fig. 3), is prob-
ably the most satisfactory arrangement in the long run, although the
homemade assembly shown in figure 4 has proved fairly satisfactory.
In this later assembly the machine is started gradually and stopped
by tightening and loosening the belt between the motor and clutch
assembly by means of a screw made from an automobile jack.
In the first model of the trap which was built a worm gear
from a wnshing machine provided the gear reduction. This was
satisfactory as long as the load was limited to one small net. In
some models the reduction in speed from the motor to the trap was
obtained by a series of belts and countershafts. This arrangement
is cheap and is g-asier for amateur mechanics to build, but is not
so satisfactory as a reliable set of gears. The belts must be pro-
tected from the rain.
One of the belt-driven models is shown in figure 5. This
particular model was designed to study the height of flight of the
beet leafhopper. The frame was of wood, bolted together and held
in place by guy wires. The height of the nets was 21, 15, and 32
feet above the ground. The center shaft was made of 1-inch iron
pipe hung from a large thrust bearing at the top to prevent kinking.
Thu details of the net construction are shown in figure 6.
An adjustable method of mounting the nets on the net arms is shown
in figure 8. The framework of the net is of '-inch iron rod and
is welded throughout. The net proper is of 16-mesh wire screen.
Two types of collecting bags have been used. One is shown
in detail in figure 9. The fine copper-guaze bottom prevents
serious injury or destruction of the trapped insects. If all-cloth
(muslin or similar material) bags are employed, the use of a small
wire frame to keep the bag permanently distended minimizes.injury
to the trapped insects. If such a frame is not used, strong eddies
in combination with a marked whipping of the bag badly damage the
At the end of each collecting period the small collecting
bags are removed and immediately replaced by new bags. The insects
collected in the small bags are killed by dropping the bag into a
large-mouthed killing jar and are later sorted and counted.
The dimensions of the net, the length of the net arm, and
the speed of rotation are optional. The only essential points to
be considered are that the taper of the net be sufficient to prevent
the escape of small insects and that the speed be great enough to
prevent large insects from dodging the net or escaping after being
caught. The writers have considered that a large-meshed screen with
a long taper was more efficient than a fine screen with a short
taper. The long-tapered nets direct even very small insects into
the collecting bag without loss or serious injury, whereas a net
with a short taper would result in loss through the mesh or crushing
and sticking on the screen. The long-tapered net with wider mesh
allows more air to pass through than a short-tapered net with narrow
mesh, and thus increases efficiency.
A net speed of 20 to 25 miles per hour is adequate to retain
most insects. Even at this speed, however, strong insects may
escape, not by flying out of the net but by crawling out of it. The
loss from this source is probably unimportant for short runs, al-
though with runs as long as an hour or more it might be serious with
some species. On the other hand, with aphids or similar passive
flyers a speed as low as 4 to 5 r. p. m. would be satisfactory, as
proved by Williams and Milne and by Davies (loc. cit.). The attain-
ment of a given net speed is obviously dependent upon the number
of revolutions per minute of the net arms and the length of the
net arms. The smaller the circle described by the nets, the faster
the speed of rotation required to provide a given net speed. The
writers preferred a relatively large circle of revolution in order
to avoid such things as "flicker" which might possibly prove attrac-
tive or repellent to certain insects. With 6-foot net arms, such
as were used in the traps illustrated, a net speed of 55 r. p. m. is
equivalent to 23 to 24 miles per hour.
If it is desirable to sample insect aerial populations over
an extended area rather than in one spot, a net like the ones used
on the rotary traps can be mounted on an automobile.6 When the car
is in motion, insects are collected in exactly the same manner as
by the rotary trap. Such a mount is shown in figure 8. It is
important that the net be mounted at least 1 foot away from the
fender or hood of the car to avoid strong eddies which tend to
force insects through the mesh of the screen. The writers have
used car speeds up to 45 miles per hour. This apparatus has proved
very useful under certain circumstances and can be built very
6 The use of a net on an automobile for the collection of
insects in flight was apparently first employed by A. Bonnet in
1911. (Recherches sur les causes des variations de la faunule
entomologique aerienne. Compt. Rend. Acad. Sci. (Paris) 152: 336-
339. 1911.) See also: McClure, H. Elliott. Insect Aerial Popu-
lations. Ann. Ent. Soc. Amer. 31 (4): 504-513, illus. 1938.
.- ^. ,
1 *' '
-.. W i e *"-"- ""
.'' : *7 -. ^ "f l
Figure l.--General view of rotiry trap.
(Photograph by K. W. Gray.)
Corvallis, model A.
- -' ...e ~
Figure 2.--General view of rot..ry trap. Corvallis, model B.
Note the pyraMnid'Ll frame of welded angle iron instead of
the bolted wooden construction of model A.
Figure 3.-Motor and clutch assembly of Corvallis trap, model A. Note
connecting rod with flexible connection, extending from drive pulley
to net drive shaft. The bendix spring shown in this figure is not
especially durable and was later replaced by a tubular iron sleeve
which was united to each shaft by a pin. The clutch is a manufactured
unit used in small machine and woodworking shops. The motor is a 1-
horsepower, air-cooled gasoline engine. (Photograph by K. W. Gray.)
Figure 4.-Motor and clutch assembly used on Corvallis trap, model B.
The trap is started or stopped by tightening or loosening the belt
by means of the specially mounted automobile jack, which tilts the
countershaft. This is as effective as the clutch arrangement shown
in figure 3, but causes greater wear on the belt.
Figure 5.-Model of the trap used to study height of insect flight. The
top net is 32 feet from the ground. (Photograph by Dwight F. Barnes.)
:. if. - -ii ia & ;::? = "':"
Figure 6.--Diagram showing the dimensions and construction of the collecting net. The framework
is of 1/4-inch iron rod and welded throughout, except for the 12-inch supporting arm, which is
of 1/2-inch rod or pipe. A, Front view. B, Side view. The net covering (not shown) is of
16-mesh screen wire with all seams soldered.
* ~ ~
Figure 7.-Mounting of the collecting net. The net is fastened to
the net arm by an iron rod, which runs in a vertical slot at the
tip of the net arm and is adjus.t'.ble as to height within certain
limits. It is held in place by a set scre%. A wire brace extends
from the large and i .Zn .,i of the net to the net, ern at the
point of the brace attii.n I-, to prevent the net being displaced
from its proper position by centrifugal force. (Photograph by
K. W. Gray.)
Figure 8,.-Collecting net, of the type used on the rotary trap,
mounted on an automobile for use in sampling aerial popula-
tions over extended areas.
,- w awi
Figure 9.-Di&gram showing construction of the collecting bag. The
length is 7 inches, the diameter 3-1/4 inches. The material is
muslin, with copper gauze bottom. A, Diagrammatic cross section
showing method of sewing in the drawstring and the screen bottom.
B, The finished bag.
UNIVERSITY OF FLORIDA
I 1i HII II it111111 I II 11 I N11111111 I I I U l I I I NBIW
3 1262 09239 5226