CHEMICAL DEFENSE AND SOUND PRODUCTION IN AUSTRALIAN TENEBRIONID BEETLES ( ADELIUAI SPP . )

Girdled eucalyptus trees, felled to make room for grazing pasture, are a common sight in rural Australia. In the dead and decaying stumps that are strewn through the countryside, many insects flourish. Two of these, the congeneric tenebrionid beetles ddelium percatum and .4. pustulosum (subfamily Adeliinae) possess interesting defense mechanisms that we here describe. Both have eversible abdominal glands such as are commonly found in Tenebrionidae (Roth, I945), but some features of the chemistry and biology of the glands are anomalous. Moreover, .4. pustulosum has a stridulatory apparatus that may function for acoustical reinforcement of the. chemical defense.

effected by first immersing them in cold (--95C) liquid Freon (while held in forceps to cause eversion) and then transferring them to a tissue freeze-dryer for desiccation.
Instruments for chemical analyses included conventional gas chro- matographs and an LKB-9ooo gas chromatograph-mass spectrometer.
Sound recordings were made at 23-25C with a condenser micro- phone from a Holgate Ultrasonic Receiver MK.V [frequency re- sponse -+ 2 db from to oo kHz; sensitivity at preamplifier out- put 268 mv (rms)//,bar at 6o kHz (determined by substitution) ], an Ithaco Amplifier Model 225 [variable amplification range from o to 80 db; flat frequency response (within 3 db) in range of o.5 Hz to oo kHz] and a Lockheed Magnetic Tape Recorder/Repro- ducer Mo.del 419 [recording speed set at 30 inches/sec; flat fre- quency response (within 3 db) in range from 0.2 to OO kHz; input sensitivity of 1.4 volt (I.O volt rms) for +-4o% deviation].Sound intensity was measured with a Brfiell and Kjaer Precision Sound Level Meter Type 2203 and a Brfiell and Kjar Condenser Micro- phone Cartridge Type 445.The intensity measurements were limited to audio frequencies due to limitations in the frequency response of the Briiell and Kjaer Microphone [flat frequency re- sponse (within 2 db) in range from 2.6 Hz to I8.5 kHz].
Instruments used for displaying and recording temporal patterns of the sound included a Tektronix storage oscilloscope (Model 4IO3N) with differential amplifier (Model 5A22N) and Polaroid camera, and a Tektronix dual beam oscilloscope (Model 5o2A) with a Grass Kymograph camera (Model C4N).Timing marks were generated on the oscilloscope with a Hewlett-Packard function gen- erator (Model 33IOA).
Sound spectograms were made with a Kay Elemetrics Sonograph (Model 7o29A).Low frequency background noise was filtered as needed during playback of the recordings with a Krohn-Hite filter (Model 355o).
CHEMISTRY OF THE SECRETION It was clear from the outset that the secretion of both species was quinonoid in nature.The beetles responded typically to rough hand- ling by protruding their glands, and the oily golden-brown fluid that visibly coated the everted structures had the unmistakable pene- trating odor of benzoquinones, and the characteristic quinonoid property of tanning human skin.Our (ingertips were always darkly mottled ater handling a quantity of the beetles. Secretion for analysis was obtained by seizing and squeezing in- dividual beetles in forceps and wiping the fluid from the everted glands with small pieces of (ilter paper.The chemical procedure and the findings were the same for both species.Extraction of the papers with methylene chloride, followed by gas chromatography of the extract (1% OV-on Gas Chrom.Q, 4' ", 50 to 24oC), revealed the presence of three major components.Analysis by gas chromatography-mass spectrometry showed two of the components to have mass spectra identical to the published spectra (Budzikiewicz   et al., I967) of 2-methyl-I,4-benzoquinone and 2,3-dimethyl-I,4- benzoquinone (Fig. a, b).The third component had a mass spec- trum matching that of I-pentadecene (Stenhagen et al., 1969)   (Fig. c).Definitive confirmation of the position of the double bond in this compound was obtained by ozonolysis.A small sample of the third component was collected from the gas chromatograph and dissolved in methylene chloride, and the cooled solution (-7oC) was subjected to a slow bubbled stream of ozone in oxygen until the blue color persisted.Addition of a small crystal of triphenylphos- phine, followed by gas chromatography-mass spectrometry, showed the higher molecular weight product to have a mass spectrum identi- cal to that of tetradecanal (Stenhagen et al., 1969).
One interesting point in connection with these analyses was the observation of a fourth component in the initial sample of secretion obtained from both species.This component was guessed to be 1,2- dichloropentadecane on the basis of its mass spectrum, and the iden- tification was confirmed by direct comparison of the isolated material with an authentic sample prepared by addition of chlorine to I-penta- decene.However, since additional milkings taken from the same Iga Psyche [March beetles four months later into carbon disulfide, rather than methylene chloride failed to show this component, it is apparent that the diha- lide is an artifact derived from the olefin.STRUCTURE AND DEFENSIVE OPERATION OF THE GLANDS Although the two beetles differ slightly in size and appearance-- d. percatum is less rotund and longer (I.5-.9 cm) than A. pustu- losum (I.3-I.5 cm)--their glandular apparatuses are essentially identical.No dissections of fresh beetles were made, but judging from what could be discerned from specimens somewhat inadequately preserved in ethanol, the glands are anatomically similar in both species and probably homologous to eversible glands such as have been described for other tenebrionids (Lengerken, I925 Roth, I945  Tschinkel, 1972).Such glands are basically cuticular sacs, ordinarily withheld in inverted condition within the body cavity, and presumably everted by blood pressure.What is unusual about the d delium glands is that they extrude to such great lengths.For example, gland length in a specimen of d. percatum .8cm long that was freeze-dried with its glands everted, was recorded at 0.8 cm.Relative to body size, the glands of these beetles must rate among the largest eversible glands known for insects.
Examination of a glandular sac of d. percatum, treated with warm aqueous potassium hydroxide to remove all soft cellular parts, re- vealed a conventional membranous cuticular lining, with typical cuticular organdies attached (Fig. 8).Such organelles are a diag- nostic feature of many insect gland cells (references in Eisner,97o), and the ones of d delium resemble closely those described from cell type z in the defensive gland of the tenebrionid Eleodes longicollis (Eisner et al., 1964).The organelles were distributed along the full length of the glandular sac, indicating that the secretory tissue is not restricted to a limited portion of the gland.
Gentle handling of ddelium usually induced no gland eversion at all.Only when the animals were more forcibly stimulated, as when they were grasped by the body or an appendage and squeezed, did  taken.Figs.3-4.ddelium percatum responding to pinching of individual hindlegs by everting the gland of the side stimu- lated.Figs. 5-6.d. percatum responding to pinching of left midlegs.In Fig.
the gland has been maximally everted and has come into contact with the leg stimulated.In Fig. 6 the gland, on eversion, discharged part of its content as an anteriorly-directed spray (reference bar---5mm).

I94
Psyche [March   eversion occur with some consistency.Reluctance to employ the glands in these animals is undoubtedly related to their possession o a particularly tough integument, which may in itsel provide an eective tqrst line o defense.
In order to observe more precisely the method o employ o the glands, individual d. percatum were affixed to wire tethers with wax, and stimulated by pinching their legs with orceps.ing elicited more complete and usually sustained eversion, and the extruded gland commonly extended outward t:ar enough to contact the seized leg or orceps (Figs.3-5).The target was usually visibly wetted by the g|and.The beetles can apparently exercise some con- trol over the direction o extrusion.Thus, when distal parts o a hindleg were stimulated, the gland usually extended outward at a greater angle relative to body axis than when more proximal parts o the leg were grasped.The gland is apparently aimed by flexion at its base rather than along its length, as evidenced by the act that its shape is relatively xed irrespective o its angle o protrusion.
When gland eversion occurred abruptly rather than gradually, as was generally the case in response to vigorous pinching, the sudden extension o the gland sometimes caused secretion to be ejected t:orcibly as a spray.(Figs. 6, 7).Ejections ranged rom to 5 cm, but in one instance droplets were shot to as ar as 25 cm rom the beetle.Forelegs and such other parts on the ront o a beetle as are ordinarily inaccessible to the glands were inevitably doused by such ejections.Only beetles with replete or nearly replete glands tended to spray.However, animals with partly depleted glands could still effectively administer secretion to their ront end.They did this by relaying secretion rom one pair o legs to another, ollowing initial wetting o the hindlegs by the glands.Even the antennae and the head itsel were eventually wetted with secretion i they had been 1974] Eisner et al.
Tenebrionid Beetles 195 Fig. 7. Response of Atdelium #ercatum to pinching of its right midleg.The animal everted its right gland abruptly, causing secretion to be sprayed forward.The droplets of secretion are shown as black spots on the indi- cator paper that served as substrate (filter paper impregnated with acidified aqueous solution of potassium iodide and starch).Far-ranging discharges as shown here occurred only when beetles had fully replete glands.Fig. 8. Cuticular organelles from secretory cells of gland of At.Iercatum, isolated by treatment of gland with warm 10% aqueous potassium hydroxide (Nomarski interference contrast; reference bar 30#).Fig. 9. Scanning electronmicrograph of hairs on exposed surface of everted gland of At. lercatum (reference bar 5#).
Psyche [March seized in forceps.The initial contact of hindlegs and glands appeared at times to be deliberate rather than fortuitous.Thus, when glands were insufficiently everted to effect automatic contact with the hind- legs, the legs sometimes reached back toward the glands and brushed against them.Wiping of midlegs against hindlegs, and of forelegs against midlegs, quickly followed.In cases where the beetles had been placed on indicator paper, the legs left dark markings on the substrate as evidence of their contamination.Some experiments comparable to the preceding were also done with A. pustulosum, with essentially similar results.However, forcible gland eversion with spray ejection was never seen in this species.
Scanning electronmicroscopic examination of the surface of everted glands o.f A. percatum revealed a covering of slender curved hairs (Fig. 9).Whether these are spaced closely enough to act as a mat- ting remains uncertain.But it is conceivable that they function in this capacity, providing perhaps for the maintenance of uni/orm secretory wetness over the surface of the glands when they are everted and for spontaneous re-spreading of secretion over areas wiped clean by defensive action of the glands.Dispensation of secretion onto target surfaces might also be facilitated by the hairs.

TRIDULATORY ]ECHANISM
When seized or pinched, /. pustulosum produced a distinctly audible sound.The response took precedence over gland eversion, and could persist for seconds if the disturbance was sustained.Two serrate ridges on the last (visible) abdominal tergite (Figs.IO-I 3) are responsible for engendering the sound.By rhythmic downward deflection of the abdominal tip--a motion depicted in Figs.IO and II-the beetles scrape the ridges back and forth across the sharp overhanging elytral margins, thereby inducing a repetitive bi-syllabic chirp.Intensified stimulation eventually caused gland extrusion, which silenced the sound, since during gland eversion abdominal deflection is apparently precluded, z/.percatum lacks the ridges and is soundless.
Stridulatory mechanisms are common in insects, and have been repeatedly noted.In the leaf-cutting ant, .4ttacehpalotes, which pro- duces sound by scraping a file in much the same manner as does ldelium, acoustical output and sound generation have been master- fully analyzed (Markl, 1968).Since this analysis is essentially applicable to ,4delium, we will here describe only the properties of showing the abdominal tip in the undeflected (Fig. 10) and deflected (Fig. 11) positions that mark the limits of the stridulatory motion.The arrow points toward the left serrated ridge (reference bar--1 mm).
The elytral margin against which the ridge is scraped extends across the top of Fig. 12 (reference bar Fig. 13---.05mm).Only two live emales and one male o d. pustulosum were avail- able for sound studies.The animals were affixed to wire tethers glued to their pronotum with wax, and induced to stridulate by pinching their legs individually with orceps.Sound recordings were made with the beetles held at cm from the microphone.An oscillo- gram of a typical chirp sequence is shown in Fig. 4A.The bi- syllabic composition of each chirp, indicative of the bi-directional movement of the ridges across the elytral margins, is clearly evident.
Measurements made from oscillograms, of chirp repetition rates, chirp and syllable durations, and duration of inter-chirp and inter-syllable intervals, are summarized in Table , together with measurements of tooth counts and tooth spacings made from the serrate ridges.These parameters are evidently remarkably constant for any one beetle, although they differ somewhat between beetles.The Table does not give an indication of the gradual decline in the chirp repe- tition rate that characteristically took place with time following onset of pinching of a given leg (e.g. in a female, the rate declined from 2.:5 to .63 chirps/see over a period of 4.5 seconds).The decline occurred by a lengthening of the inter-chirp interval, rather than by prolongation of the syllables or inter-syllabic interval of the chirp itself.Accelerated chirping (or gland eversion) always occurred when pinching was shifted to a new leg or increased in intensity.
As was to be expected by analogy with the stridulatory chirps of d tta, and shown clearly by the oscillograms (Fig. 4B, C), the syllables are essentially trains of sound pulses, generated by impact of the elytral margins with the sequence of teeth on the ridges.
Figs. 4D and I4E show oscillograms of individual pulses from a first and second syllable respectively.
Spectrograph analysis (Fig. 5) of the chirps revealed a frequency composition ranging broadly from below kHz to 6o kHz.The frenquency distribution pattern was essentially similar in the first and second syllables, and differed little in spectrographs made with broad- band (Fig. 5 top) and narrow-band (Fig. 5 bottom) i]lters.Since broad-band filtration essentially analyzed the pulses one at a time, rather than in groups o three to rive as in narrow-band (iltration, it ollows that the pulses themselves contain essentially the ull range of the beetle's sound frequencies. Apparatus for measuring sound intensities became available only at:ter the beetles had died.Intensities therefore had to be measured Psyche [March Fig. 15.Sound spectrograms of one chirp of ddelium pustulosum (the chirp is that lettered BC in Fig. 14/, and shown expanded in Fig. 14B  and C).In each spectrogram the first syllable is on left and second syllable on right.Spectrograms were made from recorded sounds replayed at speed.Top spectrogram: wide band analyzing filter (effective bandwidth 2400 Hz, with effective time response of 0.375 msee) bottom spectrogram: narrow band analyzing filter (effective bandwidth 360 Hz, with effective time response of 2.75 msee).Beside the top spectrogram are shown energy distribution profiles, corresponding to time transects lettered on the spec- trogram itself.
indirectly, rom the tape-recorded sounds o the beetles.This was done by monitoring the playback o the recordings with a microphone pickup and oscilloscope (the same microphone as used previously or recording o the chirps), and adusting the intensity o the playback until the induced oscilloscope signal matched that initially elicited by live beetles held cm rorn the microphone.The intensity o the adjusted playback was then measured with the Sound Level Meter positioned at the site of the microphone, and found to be 65 db (relative to 0.0002 dynes/cm2).
Detailed analysis o the oscillograms showed that the beetles ap- parently do not scrape both ridges with equal consistency during stridulation.In act, stridulation appears to involve predominant scraping of only one ridge during downward deflection of the ab- dominal tip (first syllable), and of both ridges with somewhat unequal force during the upward movement of the tip (second syl- lable).This conclusion rests on the observation that the syllables contain, in addition to a consistent major pulse (which in total num- ber per syllable corresponds to the number of teeth per length of file estimated to be scraped during a syllable) 1, a subsidiary pulse of varying lesser amplitude, which appears sporadically in the /]rst syl- lable, and fairly consistently in the second syllable.These lesser pulses, when they appear, result in a doubling of the pulse repetition rate.Moreover they bear a shifting rather than fixed temporal rela- tionship to the primary pulses, suggesting that they are not subsidiary acoustical concomitants of single tooth scrapings, but a consequence of scraping of a second set of teeth of somewhat unmatched spacing relative to the first.The near absence of secondary pulses in the first syllable suggests that the beetle might not press the abdominal tip as_ vigorously against the elytral margins during downward abdominal deflection as during the return motion.

DISCUSSION
Defensive glands that produce quinones are widespread among arthropods.They have been reported not only from other Tenebrionidae (references in Jacobson, 966; Roth and Eisner, 962; Weatherston, I967; Weatherston and Percy, 97o), but also.from certain termites (Moore, 968), earwigs (Eisner 'From photographs such as Figs. 10 and 11, which depict the motion of the abdominal tip during chirping, it was estimated that this motion has a maximum amplitude of ca.2/3 the length of a serrated ridge.This cor- responds to about 130 teeth, a figure in line with the counted maxima of major pulses in the syllables.have defensive glands (e.g., Tentyriinae and Asidinae).But the glands are frequently present, and when they are, they are either exclusively abdominal as in Adelium, or supplemented by a second pair of glands in the prothorax (Roth,945; Tschinkel, I969).
Hydrocarbons, both saturated and unsaturated, are known from arthropod secretions, and although -pentadecene as such has not been reported, shorter chain homologs o the substance are present in Tenebrionidae (Hurst et al., 1964).We made no effort to evaluate the defensive effectiveness of the secretion of Adelium.However, judging from the proven deterrency of other quinonoid mixtures to predators (references in Eisner, 97o,  972; Eisner and /[einwald t966), there can be no doubt that the glands of Adelium are protective in function.The presence o the alkene may be of more than incidental significance.It has been argued (Blum and Crain, I96), and in one case proven (Eisner  et al., t96I), that lipoidal additives promote spreading and penetra- tion of defensive secretions on target.Defensive glands o.f arthropods vary greatly in structure and oper- ation, and several characteristic types have been recognized (Eisner,97o).The glands of Adelium appear to combine features of two of these types.Basically, by virtue of their extrusibility, they clasify as conventional eversible glands, but since they occasionally eject their Table 1.ddelium pustulosura" Temporal characteristics of the chirps and morphological features of the stridulatory ridges.Temporal measure- ments were made from oscillograms and are given as mean + standard error; sample size (N) is given in parenthesis.Morphological measure- ments were made from microscopic preparations of the last abdominal tergite of the beetles; tooth spacing was calculated by dividing ridge length by tooth number.(The tergite of 9 no.was damaged in preparation, hence the omission of measurements from one ridge.)204 Psyche [March contents, they also quality as spraying glands.We know of no other tenebrionid glands that operate in this way.Eversible glands are common, but they are generally shorter than those of Adelium and do not eject spray on extrusion.Spraying does occur, but only from non-eversible glands, as in Eleodes (Eisner, 966).The peculiar habit of Adelium o'f relaying secretion by way of its appendages is not unique.Other tenebrionids are also said to. use their legs for administration of secretion (Tschinkel, 1972), and comparable be- havior has been reported for opilionids (Eisner et al., I97Ib) and Hemiptera (Remold,196:).
Many insects produce chirps or other "disturbance" calls when they are molested.Cerambycid beetles, for example, very generally show this property, as do other beetles, cockroaches, Hemiptera, adult and immature Lepidoptera, and members of most other groups, including non-insectan arthropods (references in Alexander, 1967; Haskell, 96 Roth and Hartman, I967; Tuxen, I967).The sounds are generally believed to be defensive', but it is by no means clear how precisely they act in this capacity.Convincing evidence has been advanced to account for the function of the calls in arctiid moths.These insects are chemically protected (Bisset et al., 96o) and hence presumably distasteful, and they emit sounds, rich in ultra- sound, in response to the echolocating chirps of bats (Blest et al.,  1963).Bats turn away from the calls, presumably because they have learned on the basis of previous experience with the moths that a prospective meal that "protests" audibly is distasteful (Dunning and Roeder, 1965; Roeder, 1967).The sounds, then, act as acoustical aposematic signals, operative in the dark just as visual aposematic adornments supposedly operate in the light.Defensive calls generally, to the extent that they are produced by chemically or otherwise pro- tected insects and induced by encounters with predators, may function in this way.They might certainly do so in Adelium t)ustulosum, a furtive animal most likely to meet its enemies in darkness.But the calls need not only serve for acoustical reinforcement of a concomi- tant defense.They could also be intrinsically deterrent to predators, perhaps only to some, or they might function socially as do "alarm" sounds in termites and ants (Howse, I964; Markl, I967, 969;  Markl and Fuchs, 97:)to alert conspecifics to states of emer- gency.But these possibilities, like so many other functional sug- gestions that have been advanced for disturbance calls, remain to be proven.One wonders what predators might be affected by the chirps of Adelium, although criteria for compiling a list of enemies are lack- ing.It seems clear, however, that the chirps are o} sufficiently broad frequency composition to be at least potentially audible to virtually any acoustically sensitive predator.

SUMMARY
The black flightless tenebrionid beetles Adelium percatum and A. pustulosum from Australia possess a pair ot eversible abdominal glands that secrete a mixture of 2-methyl-,4-benzoquinone, 2,3- dimethyl-x,4-benzoquinone, and -pentadecene.The tubular glands are unusually large, and when everted may be nearly half as long as the beetles themselves.In A. percatum, the replete glands may, on extrusion, spray their contents forcibly, to a recorded maximum dis- tance of 25 cm.Defensive administration of secretion is also effected by the beetles' legs. A. pustulosum, unlike A. percatum, produces an audible "disturbance sound" when molested.The sound is engendered by the scraping of two serrated ridges on the last abdominal tergum across the elytral margins.The morphology o the stridulatory ap- paratus, and the acoustical properties of the sound, which contains frequencies ranging trom below kHz to 60 kHz, are presented.It is argued that disturbance sounds, when produced by chemically or otherwise protected animals, act as acoustical aposematic signals, operative primarily in darkness.

Figure 2 .
Figure2.Pasture land near Gudgenby, A. C. T., Australia, where both species of ddelium were taken.Figs.3-4.ddelium percatum responding to pinching of individual hindlegs by everting the gland of the side stimu- lated.Figs.5-6.d. percatum responding to pinching of left midlegs.In Figs. 10-11.Dorsolateral view of rear of body of Adelium pustulosum, s sound, without detailed consideration o the physical basis of sound production.
The beetles were sometimes placed over /]lter paper impregnated with an acidiiqed aqueous solution o potassium iodide and starch, a mixture that dis- colors darkly on exposure to benzoquinones (Roth, I943) and pro- vides a good visual means or detecting secretory emissions.The results were identical with all nine specimens tested.Stimulation o a single leg generally caused extrusion o only the gland o the corre- sponding side (Figs.3-6).Simultaneous stimulation o two legs caused bilateral eversion.The extent and rapidity o. extrusion varied, and seemed t) depend on the intensity o stimulation.Jk brie squeeze o a hindleg or midleg, or example, usually induced only a partial eversion o the ipsilateral gland.Protracted or more vigorous squeez- and Blumberg, I959 Schildknecht and Weis, 96o), cockroaches (Roth and Stay, I958), grasshoppers (Eisner et al., 97a), carabid beetles (Anes- hansley et al., 969; references in Moore and Wallbank, 968; Schildknecht et al., 968), staphylinid beetles (Blurn et al., 97; Brand et al., 973), millipedes (references in Jacobson, 966 Roth and Eisner, I962; Weatherston, 967; Weatherston and Percy, I97o) and opilionids (Fieser and Ardao, 956; Eisner et al., 97b).As is often the case with arthropod defenses, they exist as close counterparts in plants: glandular hairs that secrete a benzoquinone have been described for Primula obconica Schildknecht et al., I967).Contrary to claim (Schildknecht et al., 964), not all Tenebrionidae