-
Chemoecology 10:8187 (2000)09377409:00:00008107 $1.500.20
Birkhauser Verlag, Basel, 2000
Chemical defense of an earwig (Doru taeniatum)*Thomas Eisner1,
Carmen Rossini2 and Maria Eisner1
1Department of Neurobiology and Behavior, Cornell University,
Ithaca, NY 14853, USA2Facultad de Qumica, Universidad de la
Republica, Montevideo, Uruguay CP 11624
Summary. The earwig Doru taeniatum (Dermaptera,Forficulidae) has
a pair of defensive glands, opening onthe 4th abdominal tergite,
from which it discharges aspray when disturbed. It aims the
discharges by revolv-ing the abdomen, a maneuver that enables it
simulta-neously to use its pincers in defense. The
secretioncontains two quinones (methyl-1,4-benzoquinone
and2,3-dimethyl-1,4-benzoquinone) present in the glands asa
crystalline mass, together with pentadecane and a(presumably)
aqueous phase. The gland openings areminute, with the result that
virtually no quinone crys-tals are expelled with the spray. Only
the two liquidphases are discharged, together with the ca. 1%
quinonethey carry in solution. Such a solute-economizing dis-charge
mechanism appears to be without parallelamong insect defensive
glands.
Key words. Defensive glands discharge mechanism quinones
hydrocarbon
Introduction
The insects of the order Dermaptera, the so-calledearwigs, are
doubly protected against predation. Thepincers at the end of their
abdomen provide for me-chanical protection, while the dischargeable
glands thatmany of them have in the abdomen (Pawlowsky
1927;Vosseler 1890) provide for chemical protection.
Most thoroughly studied has been the Europeanearwig, Forficula
auricularia (family Forficulidae). Thisearwig has two pairs of
defensive glands, opening onthe posterior margins of the 3rd and
4th abdominaltergites, from which it discharges a secretion
consistingof a mixture of methyl-1,4-benzoquinone and
ethyl-1,4-benzoquinone (Schildknecht & Weiss 1960). It ejects
thesecretion in the form of jets, which it is able to directwith
some accuracy toward parts of the body subjectedto assault. It uses
its glands and pincers in combination.When attacked, it first
revolves the abdomen in such
fashion as to bring the pincers to bear upon the at-tacker. In
so doing it automatically aims the glands, sothat if it then
discharges it inevitably targets the bodysite under assault. Ants
and frogs were shown to bedeterred by these defenses (Eisner
1960).
Recently we had occasion to study another speciesof earwig, Doru
taeniatum (henceforth referred to asDoru) (Fig. 1A). We found this
insect, also a member ofthe Forficulidae, to use its pincers and
aim its dis-charges in much the same manner as F.
auricularia,although we noted it to produce a different
quinonoidmixture than the latter species and to have but twoinstead
of four glands. More interesting, however, wasthe finding that Doru
has a special glandular dischargemechanism by which it economizes
on the amount ofquinone it expels when spraying. The mechanism
issimple and seemingly without parallel among arthropo-dan
defensive glands.
We here describe our findings. We succeeded incollecting only
eighteen Doru. Our data are thereforenot nearly as extensive as we
would have liked.
Materials and methods
The earwig
The Doru stemmed from Lake Placid, Highlands County,
Florida,where they were collected at lights on the grounds of the
ArchboldBiological Station. They were maintained in the laboratory
in smallplastic containers and given freshly cut up mealworms
(larvae ofTenebrio molitor) and water (soaked cotton wad).
Gland morphology
Dissections were done under saline solution. For light
microscopy,preparations were examined fresh (mounted in glycerin)
or afterconventional histological preparation (Bouins fixation;
hematoxylinstaining).
For scanning electronmicroscopy, preparations were
critical-point dried and gold-coated.
Spray emission
We knew from the odor of the secretion that it was quinonoid
andthat we would therefore be able to use the same indicator paper
fordepiction of the glandular discharges that we had used with
otherquinone-spraying arthropods (Eisner 1958, 1960). That paper is
pre-pared by soaking a sheet of filter paper in a freshly-made
acidifiedsolution of potassium iodide and starch. The paper is then
laid out on
Correspondence to : T. Eisner, e-mail: [email protected]* Paper
no. 170 in the series Defense Mechanisms of
Arthropods. Paper 169 is Eisner T and Eisner M, Defensive use
ofa fecal thatch by a beetle larva (Hemisphaerota cyanea). Proc
NatlAcad Sci USA 97:26322636
-
T. Eisner, C. Rossini and M. Eisner CHEMOECOLOGY82
a sheet of glass and blotted off. It changes instantly to an
intenseblue-black where contacted by benzoquinones.
To elicit the discharges, we used the same technique
employedpreviously with F. auricularia (Eisner 1960). Individual
Doru wereaffixed to a metal rod with a droplet of wax placed
dorsally on thethorax, and the rod was adjusted in such manner as
to cause theearwig to be positioned in a normal stance on the
indicator paper.The earwig was then stimulated by pinching parts of
its body withfine (watchmakers) forceps. The ensuing spray
registered instantly asa pattern on the paper.
Chemistry
All chemicals were obtained from Sigma Chemical Co. (St.
Louis,MO). Gas chromatographic analyses were done using a
HewlettPackard (HP) 5890 instrument with a split:splitless injector
and aflame ionization detector. Secretion extracts were introduced
by splitinjection. The HP-ChemStation software program was used to
ac-quire and integrate data. Low-resolution electronionization
massspectra were obtained using a HP Gas Chromatograph coupled to
aHP 5971 Mass Selective Detector. In both cases, the same columnwas
used [25 m0.25 mm fused-silica capillary column coated withHP-5 (5%
phenylmethylsilicone) stationary phase (0.25 mm filmthickness)],
under the following oven temperature condition: 60C for4 min,
increased to 260C at 10C:min.
Results
The glands
The two glands of Doru are situated immediately be-neath the 4th
abdominal tergite and open on the poste-rior margin of that tergite
(Fig. 1B, C). The cuticle oftergite 4 bulges outward conspicuously
on each side toaccommodate the glands within (Fig. 1C).
The glands are capacious (Fig. 2B). A calculation ofgland
volume, based on linear measurements from ascanning
electronmicrograph (Fig. 2C), and certain ge-ometric assumptions
(legend, Fig. 2C), yields a value of0.34 ml per replete gland. The
gland opening is elongateand minute (Fig. 1B, 3A).
The glands are devoid of compressor muscles [ex-amination of a
fresh glycerin mount of a glandular sacin polarized light failed to
reveal presence of envelopingmuscle fibers (Fig. 1E), and so did
surface examinationof a gland by scanning electronmicroscopy (Fig.
2A)].However, a muscle does occur in association with thegland
opening. We assume this muscle, which insertsjust inside the gland
opening on the narrowed, ordinar-ily collapsed exit duct of the
gland, and originates onthe tergite wall, to serve as the opener
muscle thatclears the path for glandular emission. The same
musclehas been noted to occur in F. auricularia (Vosseler1890).
Given the absence of enveloping muscles, weassume gland compression
to be effected by a rise inhemocoelomic pressure, triggered perhaps
by momen-tary telescoping of the abdominal segments. We envi-sion
glandular ejections to occur when such a rise inblood pressure is
coupled with contraction of theopener muscle.
The glands of Doru are integumental structures andas such are
endowed with a thin cuticular lining thatsurvives KOH treatment
(Fig. 2B). Microscopic exami-nation of whole mounts of the gland
wall showedpresence of the large specialized cells that
presumablysecrete the gland contents (Fig. 2DF). These cells
are
Fig. 1 (A) Doru taeniatum (3.3 ). (B) Base of abdomen, in
transmitted light, showing the defensive glands beneath tergum of
4th segment; notethe tiny white circles denoting the gland openings
(preparation treated with KOH, consisting of cuticle alone) (18 ).
(C) Dorsal view of base ofabdomen, showing gland openings at
posterior margin of fourth abdominal tergite (30 ). (D) Inside view
of freshly isolated 4th abdominaltergite, with glands attached; the
glands are conspicuously filled with yellow crystalline quinone (12
). (E) Whole mount of gland in glycerin,flattened out under a cover
slip, photographed in partially polarized light. The yellow mass of
quinone crystals is clearly discernable, as are someof the brown
droplets of the oily phase (presumably pentadecane, bearing
dissolved quinone). Most of the oil has been squeezed from the
gland(some has drifted beyond the field of vision). The aqueous
phase, also having been squeezed out, has become mixed with the
mounting medium.The dark rectangle is a portion of the 4th tergite
to which the gland is attached (25 ). (F) Same as preceding, in
fully polarized light, showinga better resolution of the crystals
of the quinonoid mass (30 )
-
Vol. 10, 2000 Chemical defense of an earwig (Doru taeniatum)
83
Fig. 2 (A) Surface view of gland (Bouins solution-fixed,
critical point-dried); the nodular structures that beset the
surface are the gland cells(90 ). (B) Isolated 4th abdominal
tergite (KOH-treated and consisting of cuticle alone) with glands
attached (one empty, one distended) (30 ).(C) Same distended gland
as in B, indicating the two dimensions, radius (r0.3 mm) and height
(h0.8 mm), used in the calculation ofapproximate gland volume. That
volume was considered to be equal to two half spheres and a
cylinder (4:3 p r3p r2 h0.34 ml). (D) Surfaceview of gland (Bouins
solution, hematoxylin stain) showing large, intensely basophilic
gland cells (360 ). (E, F) Two views of the same glandcell at
different optical planes (940 ); the secretory vesicle is denoted
by arrow in (E); at a plane closer to the cuticular lining of the
gland, theduct that drains the vesicle is discernable (F). (G)
Drainage duct of gland cell (2260 ); the duct has been isolated by
KOH treatment. The insetshows site where a duct opens through the
cuticular lining of the gland into the gland lumen (4275 )
Fig. 3 (A) Close-up view of gland opening(660 ). (B) The
crystalline quinonoid masswithin a gland (partially polarized
light) (78 );the inset shows the gland opening to scale
(dimen-sions of opening taken from A). Bar in B10 mm
large enough to bulge out from the gland wall (Fig.2A). They are
distributed singly. An actual count madeon one gland wall showed
presence of 387 such cells.Each has a central vesicle (Fig. 2E,
arrow), drained bya cuticular duct (Fig. 2F), an arrangement
frequentlyfound in secretory cells of insectan defensive
glands(Noirot & Quennedey 1974). The duct opens directly
into the gland lumen. It survives KOH treatment and isevidently
cuticular (Fig. 2G).
Although chemical analysis was to show the secre-tion to consist
of benzoquinones and a hydrocarbon,compounds that one might expect
to be produced bydifferent cells, no second type of secretory cell
wasfound to be associated with the Doru gland.
-
T. Eisner, C. Rossini and M. Eisner CHEMOECOLOGY84
Spray emission
As is clear from Fig. 4, Doru discharged its secretion asa
coarse spray. It is also clear that Doru exercised somecontrol over
the directionality of the emissions. Thus,stimulation of a front
leg or of the head resulted inforwardly-directed discharges (Figs.
4B and D), whilestimulation of a prong of the pincers resulted in
anejection accurately directed toward that prong (Fig.4C). As is
evident from the broadness of the spraypatterns, the secretion was
not always efficientlytargeted. In Figs. 4A and 4B, for instance,
where thestimulus was to an individual leg, some spray
wasmisdirected in each case to the side opposite of the leg,and in
Fig. 4D, some spray was misdirected to the rear.
Doru aimed its ejections in the same manner as F.auricularia. As
soon as a stimulus was inflicted, itflexed its abdomen and
attempted to bite the offend-ing forceps with its pincers. The
discharges never oc-curred prior to these abdominal adjustments,
but theymay have occurred in some cases while the adjustmentswere
taking place. Secretion ejected prematurely duringthe adjustments
could account for the misdirected por-tion of discharges (the
droplets to the rear of theearwig in Fig. 4D could have been
ejected before theabdomen was fully coiled forward toward the
head).Misdirection of part of the spray to an opposite side(as in
Figs. 4A and B) could be a reflection of the factthat the two
glands, as set in the abdomen, point
divergently and are therefore unable ever to
dischargesynchronously in convergence upon a target. Theglands need
not, of course, always discharge simulta-neously. Sharply focussed
spray deliveries, when theyoccurred (Fig. 4D), could have been due
to unilateralglandular emission.
Only 1 to 3 discharges could be elicited from eachof the 5 Doru
that were stimulated in this fashion.
Chemistry
We had seven individual Doru available for analyticalwork. We
divided these into two groups, which weretreated as follows:
a. (individuals 13). These were killed by freezing,then thawed
out and dissected under saline solution.The glandular sacs were
then excised intact, and ex-tracted with dichloromethane for
analysis.
b. (individuals 47). These were first renderedlethargic by
exposure to 4C, then held in forceps andsqueezed, so a to cause
them to eject secretion. Theeffluent from each was taken up in a
microcapillarytube, held over the gland openings as the earwig
wasbeing stimulated. The animals were then killed by freez-ing and
their glands dissected out. The weighed effluentand the glands were
separately extracted with di-chloromethane for analysis.
Qualitatively the secretion was shown to consist ofthree
volatile components, evidenced by three distinct
Fig. 4 Spray patterns, on indicator paper, of discharges
elicited from individual Doru by stimulation of: the right midleg
(A); the left foreleg (B);the right prong of pincers (C); and the
head (D). The diagram of the earwig is included for positional and
size reference purposes only; the animalsare not shown in the
special stance they adopted (that is, with abdomen coiled and
pincers brought toward the site stimulated) when theydischarged.
The spray patterns were traced from photocopies made of the sheets
of indicator paper on which the patterns were originally
registered(1.2 )
-
Vol. 10, 2000 Chemical defense of an earwig (Doru taeniatum)
85
Fig. 5 A reconstructed ion chro-matogram of volatiles in a
di-chloromethane extract of the defensivesecretion of Doru
(compound abbrevia-tions as in Table 1)
peaks in the gas chromatograms (Fig. 5). These peakswere
identified by GC:MS as corresponding to methyl-1,4-benzoquinone,
2,3-dimethyl-1,4-benzoquinone, andpentadecane. The
characterizations were confirmed bycomparison with authentic
samples.
The quantities in which the three components arestored in the
glands are given in Table 1. For the 4earwigs in group (b) these
quantities were calculated foreach compound by adding the amount
present in theeffluent and the amount present in the glands
aftereffluent emission.
It is clear from these values, that the compoundsare stored in
substantial quantity in the glands. Indeed,upon dissection, the
glands were noted to be ladenwith golden-yellow crystals that we
took to be the
quinones (Fig. 1D). Examination of glycerin mounts offreshly
dissected glands showed the secretion withinto be triphasic. The
quinone crystals, which weredensely packed, made up the solid phase
(Fig. 1E, F).Of the other two phases, both liquid, one was oilyand
consisted mostly of large light-brown droplets(Fig. 1E), while the
other, more plentiful and morefaintly colored, provided the carrier
fluid in whichthe solid and oily phases were suspended. We
presumethe carrier phase to be aqueous, and the oily phase tobe
made up of pentadecane (indeed, earwig no. 6,whose effluent was
noted to be free of oily droplets,proved upon analysis to have no
pentadecane in theeffluent; see Table 2). We attribute the
coloration ofthe two liquid-phases to quinones that they have
Body MassDoru Total per earwig (mg) % of Body Mass
DMQ 15C15C MQDMQno.:sex (mg) MQ
681 F 25.4 0.180.4145380.040.1513123933.92 F
3 M 30.0 46 36 15 0.26 0.054 F* 22.4 55 33 14 0.39 0.065 F*
0.050.2515603430.2
46.5 0.040.07168226 M*0.030.2411286237.07 M*
0.2790.12 0.0690.0232.298.0(X9SD) 1891147916 30917
Table 1 Total quinone andpentadecane content of theglands of 7
individual Doru.Asterisk denotes individuals ofgroup b for which
totals werecalculated as the sum of theamount of chemical in
effluentdischarged by glands, and theamount detected in glands
af-ter emission of effluent (seeTable 2).
MQmethyl-1,4-benzoquinone;
DMQ2,3-dimethyl-1,4-benzoquinone;15Cpentadecane
Total in Effluent (mg)Doru % of EffluentEffluent
15C MQDMQ 15Cno.:sex (mg) MQ DMQ
0.110.484 F 0.37330 0.92 0.662.42 1.005 F 410 2.65 0.591.44
0.566 M 530 1.89 1.061.667 M 260 2.36 1.96 1.60 0.62
0.9290.543839116 1.9590.76 1.2890.56 0.4390.281.4691.03
Table 2 Volume and chemi-cal content of glandular
fluid(effluent) discharged by the 4individual Doru of group
b.Abbreviations as in Table 1
-
T. Eisner, C. Rossini and M. Eisner CHEMOECOLOGY86
taken up by partition from the solid phase. Penta-decane can be
expected to have the greater affinity forquinones (Peschke &
Eisner 1987), hence the darkercolor of the oily phase.
The sum total of quinones and pentadecane in thetwo glands
ranged from 46151 mg (Table 1). Assuminga density of l mg:ml for
these materials, this corre-sponds to 0.050.15 ml, or 722% of the
summedcapacity of the two glands. The aqueous carrier phasecan
therefore be expected to make up in the order of7893% of the gland
contents.
The effluent itself of the glands, when taken up
inmicrocapillary tubes, was noted to be virtually free ofquinone
crystals, and to consist (by visual estimate) ofa few large
droplets of the oily phase suspended withina much larger quantity
of the aqueous phase.
It follows from the preceding that the dischargedsecretion
should have a much lower quinone contentthan the mixture stored in
the gland. Analysis bore thisout (Table 2). The amount of fluid
that we collected aseffluent, which can be expected to provide a
measure ofthe amount of secretion ordinarily voided by the
earwigwhen spraying, ranged from 260530 mg, an equivalent(if one
again assumes a density of 1 mg:ml) of 3878%of the combined
capacity of the two glands. Yet thetotal amount of quinone in the
effluent was only in therange of 1.64.3 mg or, on average, 0.92% of
thesample. Doru evidently discharges a dilute solution ofquinones,
such as one could envision occurring if itejects solvent only (in
other words, quantities of thetwo liquid phases) and holds back on
quinone crystalemission. We suggest that this is precisely what
theearwig does, and that the reason the quinone crystalsare not
expelled is that they are not readily mobilizedand flushed out
during ejection. Physically the crystalsseemed to form a pasty
aggregate. Individually theyseemed on the whole to be too large to
fit through thetiny gland openings (Fig. 3A, B).
A simple calculation yields a figure for the volumeof secretion
that the earwig would need to accommo-date if it stored its entire
quinone supply at the concen-tration prevalent in the effluent: to
be stored at aconcentration of 0.92%, 77 mg of quinone would
require8.4 ml of solvent, the equivalent of 12 times the com-bined
volume of the two glands.
An explanation is needed for why pentadecane wasalso present in
minimal quantities in the effluent. Onepossibility is that the
compound is not readily dislodgedfrom the gland lumen because of
its physical affinity forthe crystalline quinone mass and the
cuticular lining ofthe gland. We noted that when we compressed a
freshlydissected gland mounted in glycerin under a coverslip,the
aqueous phase would be squeezed out first, togetherwith only a
fraction of the oily material.
Discussion
Although much of our argument about the operationof the Doru
glands remains speculative, the basic fea-
tures of the mechanism seem established. The earwigstores a
substantial quantity of the quinones in crys-talline form in the
glands, and on ejection avoidsexpelling these crystals, discharging
instead a dilutesolution of the quinones. The solvent, according
toour argument, is the liquid mix in which the quinonesare
suspended in the glands, a mix consisting of anoily phase
(pentadecane) and an aqueous phase. Theadaptive advantage of such a
mechanism is that theearwig can thereby avoid incurring a shortage
ofquinones. For as long as it is able to replenish thesolvent
fluids when these are lost as a consequence ofdefensive use of the
glands, it is ready for action.It would clearly have been desirable
to determine therate at which these fluids are secreted into the
glands,just as it would have been interesting to establishwhen and
at what rate the quinones are produced.One wonders whether the
gland contents are lost atmolting, and if so, whether quinone
production is ata peak immediately after molting. Unfortunately
noneof our Doru were immatures.
While we carried out no tests with predators,there can be little
doubt that Doru derives defensivebenefit from its glands. Ants must
certainly be de-terred by the secretion, as they typically are by
qui-none-bearing fluids (Blum 1981; Eisner 1958, 1960).At the
concentration of about l% at which Doru ex-pels its quinones, these
compounds (including methyl-1,4-benzoquinone itself) have been
shown to bepotently irritating in a bioassay with
cockroaches(Peschke & Eisner 1987).
Pentadecane may contribute to the defensive effec-tiveness of
the secretion in several ways. Support forthe notion that the
compound serves as solvent forthe quinones is provided by evidence
that it effec-tively takes up quinones by partition (Peschke &
Eis-ner 1987). But pentadecane can also promote spreadof the
secretion over the integument of the enemy,and be deterrent in its
own right. Indeed, penta-decane has been shown to be an effective
surfactant,as well as a repellent to ants and a topical irritant
tocockroaches (Peschke & Eisner 1987).
We know of no other arthropods with a defensiveglandular
discharge mechanism comparable to that ofDoru. One mechanism that
bears some similarity isthat of the opilionid Vonones sayi. This
arachnidstores crystalline benzoquinones in two glands thatopen
anterolaterally on the edge of the carapace.When disturbed, the
animal mixes a small amount ofthese crystals with regurgitated
enteric fluid, and theneffects dosaged delivery of the mixture by
brushing iton the attacker with the tips of its forelegs. Ants
areeffectively repelled. It was estimated that at any onetime a
Vonones stores enough quinone in its glands tomix 55 such potions
(Eisner et al. 1971).
A final point concerned the possibility that F.auricularia
shared some of the adaptive featuresof the defensive apparatus of
Doru and that thismight have escaped notice in previous studies.
Wecollected a number of F. auricularia in Ithaca, NY
-
Vol. 10, 2000 Chemical defense of an earwig (Doru taeniatum)
87
and dissected these and milked them of secretion. Theglands were
found to be replete with brown fluid, anddevoid of both yellow
crystals and an oily phase. Thesecretion was found to consist of
methyl-1,4-benzo-quinone and ethyl-1,4-benzoquinone, as previously
re-ported (Schildknecht & Weiss 1960), and to
lackhydrocarbon.
Acknowledgements
We acknowledge with thanks partial support of thisresearch by
National Institutes of Health grantAI02908 (T. E.) and a fellowship
from the Johnson &Johnson Corporation (C. R.). We are also very
grate-ful to the personnel of the Archbold Biological Sta-tion, and
especially Mark Deyrup, for hospitality andhelp during our stay at
the Station.
References
Blum MS (1981) Chemical Defenses of Arthropods. New
York:Academic Press
Eisner T (1958) Spray mechanism of the cockroach Diploptera
punctata.Science 128:148149
Eisner T (1960) The chemical and mechanical weapons of an
earwig.Psyche 67:6270
Eisner T, Kluge AF, Carrel JE, Meinwald J (1971) Defense
ofphalangid: liquid repellent administered by leg dabbing.
Science173:650652
Noirot C, Quennedey A (1974) Fine structure of insect
epidermalglands. Ann Rev Entomol 19:6180
Pawlowsky EN (1927) Gifttiere. Jena: Gustav FischerPeschke K,
Eisner T (1987) Defensive secretion of a beetle (Blaps
mucronata): physical and chemical determinants of
effectiveness.J Comp Physiol 161:377388
Schildknecht H, Weiss KH (1960) VI. Mitteilung uber
Insektenab-wehrstoffe. Zur Kenntnis des Pyrigialdrusensekretes
vomgemeinen Ohrwurm, Forficula auricularia. Zeitschr
Naturforsch15:755757
Vosseler J (1890) Die Stinkdrusen der Forficuliden. Arch Mikr
Anat36:565578
Received 14 February 2000; accepted 21 February 2000.
.