Unformatted Document Excerpt
St. Mary NE
Course Hero has millions of student submitted documents similar to the one
below including study guides, practice problems, reference materials, practice exams, textbook help and tutor support.
Course Hero has millions of student submitted documents similar to the one below including study guides, practice problems, reference materials, practice exams, textbook help and tutor support.
Exp. J. Biol. (1965), 43, 185-192
With 1 plate and 9 text-figures
Printed in Great Britain
THE SIGNAL GENERATED BY AN INSECT IN A
BY D . A. PARRY
Zoological Laboratory, University of Cambridge
(Received 29 January 1965)
It is generally accepted that web-spinning spiders detect and find prey in their
webs through the mechanical signal generated by the prey. There is also a good deal
of evidence that spiders discriminate between different types of signal. Barrows (1915)
caused Epeira sclopetaria to orientate, or move, towards a tuning fork; and found t he
frequency-band of 24-300 cyc./sec. most effective. Meyer (1928) obtained responses
from several species of Agriopidae (orb-web spinners) to a tuning-fork (435 cyc./sec).
Peters (1931), with Epeira diademata, found that the spiders did not respond to a dead
fly placed gently in its web; if, however, theflyarrived in the web with a jerk or if,
once in the web, it was suddenly tapped with a needle or stimulated with vibrating
forceps, then the spider responded. Liesenfeld (1956) also found, with Zygiella
x-notata, that a vibrator suddenly switched on would produce a response, while, if
the amplitude of vibration was slowly raised, no response occurred. Walcott & Van der
Kloot (1959) obtained responses in Achaearanea ( = TheridionV) tepidariorum to a
vibrating phonograph needle over the range 400-700 cyc./sec., while between 700
and 3000 cyc./sec. the spider retreated or dropped from the web. Tretzel (1961)
similarly obtained responses from Coelotes terrestris to sinusoidal vibrations generated
by a loudspeaker movement and drew attention to the ability of females to distinguish
between prey and their own young. Most recently Bays (1962) has conditioned
Araneus diadematus to discriminate between tuning forks vibrating at 262 cyc./sec.
(= c) and 523 cyc./sec. ( = c').
In contrast to the amount of work on the response of spiders to sinusoidal and other
artificial signals, there is little information about the natural signal generated by living
prey. Liesenfeld (1956) analysed cinematograph records of flies moving in the web
of Zygiella and recognized wing vibration, body movement and the resonance of the
insect in the web as components of the movement pattern. Tretzel (1961, see above)
tape-recorded signals from the web of Coelotes produced by prey and by young spiders,
and compared their frequency content. He found that the signals produced by prey
contained higher frequencies, and showed a greater range of intensities, than those
produced by young spiders. But he doubted whether the spider's evident ability t o
discriminate between prey and young did in fact depend on frequency. Walcott (1963)
published spectral analyses of recorded signals produced by bees and flies in the web
of Achaearanea but found the energy distributed over a wide frequency range (mostly
within 100-5000 cyc./sec.)
Thus we have evidence that various spiders discriminate between sinusoidal signals
The signal generated by an insect in a spider's web
of different frequencies, and also evidence that they discriminate between living prey
and artifacts. But there is little evidence that the latter discrimination (which has
obvious selective advantage) is frequency dependent, neither is there any particular
reason to suppose that the slight movement of insects in a web would produce a signal
with some characteristic frequency spectrum. Clearly more information is needed
about natural signals, and this paper contains a study of the vibration produced in
the ' cobweb' of the British house-spider Tegenaria atrica by small cockroaches, which
it readily detects and attacks.
Spiders and webs. This work has been done on the British house-spider Tegenaria
atrica (Koch) which readily spins its horizontal' cobweb' in confinement. For experimental purposes the animals are given rings of wood (7^ in. inside diameter) and they
build their webs across the ring. These can be bolted to a steel plate to which the
transducer is also secured. Fig. 1 shows the layout.
Fig. 2. Transducer response (a) undamped; (b) damped with silicone fluid,
1250 centistokea. Calibration: iooo eye./sec.
Signal source. When a small cockroach is dropped on the web it usually lies still for
a time and then begins to make slight movements with legs and antennae. These
movements are enough to provoke an attack if a spider is present; the spider runs
across the web, in the direction of the insect.
A cockroach will make slight movements of this sort while recovering from CO a
anaesthesia and it is the signal produced under these conditions that has been examined.
Cockroaches approximately 10 mm. long (50 mg.) are anaesthetized and placed in the
centre of the web. Recordings are made when movements begin, and are discontinued
when the insect eventually starts walking across the web as the signal amplitude then
exceeds the linear range of the pre-amplifier.
Transducer. This consists of a in. length of PZT multimorph (Brush-Clevite Co.)
used as a cantilever, and damped with silicone fluid (see Fig. 1). It has a resonant
frequency of just under 2 kc./sec., and progressive damping was obtained with fluids
of up to 12500 centistokes, but not beyond this viscosity.
The response was studied by dropping a ball-bearing on the end of the transducer;
the ball triggering an oscilloscope trace, just before impact, through interruption of a
light beam. Fig. ia shows the response of the undamped transducer; Fig. zb shows
the damped transducer as used in these experiments. The damping factor is very
approximately 0-45 (Q = 1) and the sensitivity 1-75 mV/dyne.
D. A. PARRY
Electronics (see Fig. 3). The transducer is RC-coupled to a Tectronix pre-amplifier
( x 1000, maximum unbalanced input + 5 mV). The output is fed to a Telequipment
oscilloscope and can also be monitored aurally through headphones.
The gain is flat (+ 3 db.) from 2 cyc./sec. to 30 kc./sec, calibration being carried
out with a sinusoidal signal fed into the cathode follower through a capacitance of
840 ju,F. which is equal to the capacitance of the transducer.
Fig. 3. Diagrammatic lay-out.
To avoid distortion due to the mechanical resonance of the transducer the signal
was frequency-limited to 1 kc./sec, using the simple RC filter in the pre-amplifier.
This signal is referred to as the ' broad-band' signal. For further frequency-limiting,
a Krohn-Hite model 315 A variable band-pass/high-pass filter was used; and its output
could be displayed on one trace of the oscilloscope while the broad-band signal was
displayed on the other trace.
1. A typical broad-band signal is shown in Fig. 4, and further examples are given
in Figs. 6 and 7 (upper traces). It is a very irregular wave-form but there are, nevertheless indications of a 60 msec, periodicity (equivalent to a frequency of 15 cyc./sec).
Fig. 4. Typical broad-band wave-form produced by an insect making slight movements in the
web. Calibration: 50 cyc./sec.
This can be shown in the following way to be the period of the transverse oscillation
of the insect in the web. The cockroach, from which the records of Figs. 4 and 6 were
made, was killed and placed in the centre of the web. A small glass bead was dropped
on to it and the resultant signal 5 recordedFig. (a). I t is a heavily damaged oscillation
with a period of 60 msec.
2. Fig. 6 shows (upper trace) a broad-band signal and (lower trace) the same signal
band-limited to 50-500 cyc./sec. and with a relative gain of x 5. An occasional
periodicity of very approximately 10 msec, is evident on the lower trace; and this is
the order of magnitude of the rotational oscillation of the insect in the web. The
The signal generated by an insect in a spider's web
evidence for this is shown in Fig. $(b), which is the signal recorded whan a glass
bead is dropped excentrically on a dead cockroach to emphasize the rotational oscillation. The upper, broad-band, trace shows the transverse oscillation as in Fig. 5 (a).
The lower trace (band-limited to 50-500 cyc./sec.) shows a damped oscillation with
a period of about 10 msec, (equivalent to 100 cyc./sec).
3. Fig. 7 shows (upper trace) a broad-band signal and (lower trace) the same signal
band-limited to 200-500 cyc./sec. and with a relative gain of x 5. T he band-limited
Fig. 5. Response of a dead insect to an impulse (glass bead), (a) broad band. (6) upper trace:
broad band; lower trace: band limited to 50-500 cyc./sec. Calibration: 50 cyc./sec.
Fig. 6. Typical wave-forms produced by an insect making slight movements in the web. Upper
trace: broad-band; lower trace: band-limited to 50-500 cyc./sec. Calibration: 50 cyc./sec.
D. A. PARRY
signal is characterized by sporadic pulses which can be seen to correspond to sudden
discontinuities in the broad-band signal. Similar discontinuities can be seen in Fig. 6,
and one such discontinuity is particularly well shown in Fig. 8. Fig. 7 (4th row) shows
a burst of 'activity' in the narrow-band channel; but note that the occurrence of discontinuities is not obviously related to the intensity of the broad-band signal. This is
well shown in Fig. 7 (5th row) where the broad-band signal amplitude becomes very
large but is only accompanied by a few pulses.
Fig. 7. Typical wave-forms produced by an insect making alight movements in the web. Upper
trace: broad band; lower trace: band-limited to 200-500 eye./sec. Calibration: 50 eye./sec.
Fig. 8. Example of a'fast transient'. Upper trace: broad band; lower trace: band-limited to
200-500 cyc./sec. Calibration: 50 cyc./sec.
The object of this investigation was to get information about the mechanical disturbance set up in the web of Tegenaria by an insect making the sort of movements
which normally evoke the spider's attack response. These movements will bring the
insect's limbs into impulsive contact with threads of the web, and it is not surprising,
therefore, that the resultant wave-form is very irregular. It is dominated by lowfrequency components (in the order of 10 cyc./sec.) which are due to the transverse
The signal generated by an insect in a spider's web
oscillation of the insect in the web. There is also evidence of a higher-frequency component (in the order of 100 cyc./sec.) due to rotational oscillation. But possibly the
most significantand certainly the most unexpectedfeature of the signal is the
sporadic occurrence of discontinuities which will be referred to as 'fast transients'.
The same effect can be obtained (Fig. 9) by moving an insect's leg or antenna, or a
light string, very gently in the web with a micromanipulator. The precise cause of the
fast transients is not known. They could be due to threads of the web being displaced
by the insect and then suddenly snapping back into position. Alternatively, they might
break or, more likely, suddenly become detached from the rest of the web which would
then move into a new equilibrium position. PI. 1 shows that the threads do not
run in straight lines from one edge of the web to the other, but are attached to one
Fig. 9. Insect antenna gently drawn across the web: band limited to 200500 cyc./sec.
Calibration: 50 cyc./sec.
another to form a complex network. This network is under tension, due to the way it
is produced and also the to fact that it is weighed down by the insect. In normal use
it may also be tensioned by the spider itself which has a characteristic way of standing
on the dege of the web (often in a ' lair') and grasping a bundle of threads in the claws
of its front pair of legs. If the insect's movements have the effect of releasing some of
the tension, then the web is acting as an alarm system in which the energy is coming
from the web itself and the insect is merely triggering its release. This could form the
basis of a very sensitive detector.
The possibility that the fast transients form the basis of the spider's recognition of
prey in its web is now being investigated.
1. There is evidence that web-spinning spiders discriminate between prey and
artifacts in their webs, and that the signal involved is a mechanical one. As a contribution to our understanding of the basis of this discrimination, an analysis has been
made of the natural signal generated by an insect in the web of the British house spider
2. T he signal investigated was frequency-limited to 1 kc./sec, this being the upper
limit of the linear response of the specially designed transducer.
3. T he signal has an irregular wave-form with most of the energy lying below
50 cyc./sec. Damped transverse and rotational oscillations of the mass of the spider
in the compliance of the web have been recognized. In addition there are 'fast
transients', most likely due to the sudden release of tension in the web by slight movements of the insect.
4. The possibility that the fast transients form the basis of prey-recognition is being
D. A. PARRY
I am indebted to Dr K. E. Machin for a number of helpful discussions in the course
of thia work and for reading the manuscript. Some of the work was done while I was
a visiting lecturer at Harvard University, Mass., U.S.A.; and I am particularly indebted to Prof. J. H. Walsh for the hospitality I enjoyed while working in the Harvard
BARROWS, W. M. (1915). The reactions of an orb-weaving spider, Epeira sclopetaria Clerck, to rhythmic
vibrations of its web. Biol. Bull. Woods Hole, 39, 316-32.
BAYS, S. M. (1962). A study of the training possibilities of Araneus diademattu (Cl). Experientia, 18,
423-24L1E8ENFBLD, F. J. (1956). Untereuchungen am Netz und tlber den Erschutterungssinn von Zygiella
X-notata (Cl). (Araneidae). Z. vergl. Phytiol. 38, 563-92.
MEYER, E. (1928). Neue Sinnesbiologische Beobachtungen an Spinnen. Z. Morph. Okol. Tiere, i a,
PETERS, H . (1931). Die Fanghandlung der Kreuzspinne (Epeira dxademata L .). Z. vergl. Pkytiol. 15,
TRETZEL, E . (1961). Biologie, Okologie und Brutpflege von Coelotet terreitrit (Wider) (Araneae, Agelenidae). Teil I I : Brutpflege. Z. Morph. Okol. Tiere, 50, 375-542.
WALCOTT, C. (1963). The effect of the web on vibration sensitivity in the spider Achaearanea tepidariorum (Koch). J. Exp. Biol. 40, 595-611.
WALCOTT, C. & VAN DER KLOOT, W. G. (1959). The physiology of the spider vibration receptor. J. Exp.
Zool. 141, 191-244.
EXPLANATION OF PLATE
Part of the web of Tegenaria atrica, illustrating the attachment of threads to
one another, and their varying sizes.
journal of Experimental Biology, Vol. 43, No.
D. A. PARRY
(Facing p. 192)