Journal of Chemical Ecology, Vol. 17, No. 11, 1991

CHEMISTRY vis-d-vis MATERNALISM IN LACE BUGS (HETEROPTERA: TINGIDAE): ALARM PHEROMONES AND EXUDATE DEFENSE IN Corythucha AND Gargaphia SPECIES

JEFFREY JAMES

R. A L D R I C H , 1'* J O H N

W.

E. O L I V E R , 1 a n d W I L L I A M

NEAL,

Jr., 2

R. LUSBY 3

ilnsect Chemical Ecology Laboratory, USDA-ARS, Bldg 467 2Florist and Nursery Crops Laboratory, USDA-ARS, Bldg 470 31nsect Neurobiology and Hormone Laboratory, USDA-ARS, Bldg 467 Beltsville, Maryland 20705 (Received March 20, 1991; accepted July 23, 1991) Abstract--The hawthorn lace bug, Corythucha cydoniae, and the eggplant lace bug, Gargaphia solani, possess alarm pheromones that are produced in dorsal abdominal glands (DAGs). When G. solani nymphs are grasped, they emit secretion from both DAGs; the posterior DAG secretion alone elicits alarm, but the anterior DAG secretion may hasten the response. In C. cydoniae, the response is due to a synergism between the anterior and posterior DAG secretions, and nymphs am apparently unable to voluntarily release their DAG secretions; both DAGs must be ruptured for the pheromone to escape. The alarm pheromones are interspeeifically active in patterns matching the intraspecific activities. Compounds identified from tingid DAG secretions that are involved in the alarm messages are: (E)-2-hexenal, (E)-4-oxo2-hexenal, acetaldehyde, geraniol, and linalool. A new natural product of unknown function (designated nerolidol aldehyde) was identified from the anterior DAG secretions of both species. Key Words--Hemiptera, Tingidae, Corythucha, Gargaphia, altruism, semiochemical, defensive behavior, maternalism, allomone, abdominal glands, (E)-2-hexenal, (E)-4-oxo-2-hexenal, acetaldehyde, geraniol, linalool.

INTRODUCTION

Of the 1800 species of tingids worldwide, relatively few are gall-makers or myrmecophiles; the vast majority feed gregariously on mesophylt underneath * To whom correspondence should be addressed. 2307 0098-0331/91/1100-2307506.50/0 9 1991 Plenum Publishing Corporation

2308

ALDRICH ET AL.

leaves (Drake and Ruhoff, 1965). Exposed, sedentary lace bugs seem likely to have specialized defenses against predators and parasitoids. Indeed, two adaptations by which tingids blunt their potential attackers have now been extensively studied. Females of the eggplant lace bug, Gargaphia solani, and other Gargaphia species, assiduously guard their eggs and nymphs (Fink, 1915). Maternalism does effectively deter most predators, although a "cheating" counterstrategy has evolved in G. solani whereby some females furtively " d u m p " their eggs among those of a guarding female (Tallamy and Horton, 1990, and references therein). However, maternal care is not widespread among tingids, at least not to the extent observed for Gargaphia spp. (Faeth, 1989), but nymphs of species from many genera have secretory setae bearing fluid droplets scattered over their bodies (Livingston, 1978). In Stephanitis and Corythucha the droplets contain compounds (probably in aqueous solution) having an oxygenated six-member ring with a long side chain (acetogenins) (Oliver et al., 1990, and references therein). A biological role for these acetogenins has not been firmly established, yet it seems likely that the droplets fulfill a defensive role. Perhaps the acetogenins simply act as soaps to allow the droplets to wet small aggressors (Aldrich, 1988a). Kearns and Yamamoto (1981) reported that aggregated nymphs of G. solani quickly become alarmed when a conspecific nymph is crushed nearby. Furthermore, they described similar alarm behavior toward crushed nymphs for the hawthorn lacebug, Corythucha cydoniae, as well as two other Corythucha spp., and found cross-activity between G. solani and all three Corythucha spp. To our knowledge, alarm pheromones in the Tingidae have not been explored further. Therefore, the present study of the alarm pheromone systems of G. solani and C. cydoniae was undertaken.

METHODS AND MATERIALS

Insects. The hawthorn lace bugs, C. cydoniae, used in the study were from a colony reared on cuttings of Washington hawthorn, Crataegus phaenopyrum, at 26 ___ 1 ~ with a 16 : 8 hr light-dark photoperiod. The initial behavioral and chemical work with G. solani used nymphs collected in the field from horsenettle, Solanum carolinense, near the Beltsville laboratory during late summer, 1989. Later experiments with G. solani were performed using individuals from a laboratory colony maintained on greenhouse-grown eggplant, Solanum melongena, under insectary conditions as above. Scanning Electron Microscopy. Lace bug exuviae were sputter coated with gold-palladium in a Technics Hummer 5 instrument and examined with a Hitachi S-530 scanning electron microscope.

LACE BUG PHEROMONE

2309

Extractions. Immature tingids possess two dorsal abdominal glands (DAGs) attached to the 3-4 and 4-5 intersegmental membranes (Cobben, 1978), and the contents of the glands are shed at each molt with the exuviae (Aldrich, 1988b). For G. solani, the combined contents of the DAGs could be satisfactorily extracted by cutting out a small piece of cuticle containing the gland reservoirs under a dissecting microscope from fresh exuviae and extracting several such reservoir pairs with 25-100/zl CH2C12 or CS2. 1-Nonanol was used as an intemal standard in one extract of 20 G. solani fifth-instar exuviae ( < 1day-old) for quantitation of DAG volatiles. DAG samples from G. solani were also prepared by pressing small pieces of filter paper down on the abdominal tergum of live nymphs and then extracting the papers in ca. 100 #1 of CH2C12. For C. cydoniae, these methods were unsatisfactory because the acetogenins from the setal droplets (Lusby et al., 1989) overwhelmingly contaminated the samples. The following sampling technique was devised to avoid this problem. Fresh exuviae were removed from leaves and placed ventral-side down on double-stick tape on a slide under the dissecting microscope. Using microscissors, a cut was made laterally along each side of the tergum; this flap of cuticle was peeled backwards with forceps and stuck to the tape, thus exposing the inner surface of the two DAG reservoirs. A capillary tube drawn to a very fine point was inserted in a Drummond Microcap (Broomall, Pennsylvania) holder and fastened in a micromanipulator next to the microscope stage. When the reservoir of a DAG was pierced with the micropipet, the secretion was captured by capillary action, and could then be expelled into solvent or used directly for bioassay. This method was subsequently used to separately sample the DAGs of G. solani. Chemical Analysis. Samples were analyzed by gas chromatography (GC) on a bonded methyl silicone column (0.25 mm film, 14 m • 0.25 mm ID; DB1, J&W Scientific, Folsom, Califomia) using a Varian 3700 GC with helium as carrier (40 cm/sec), and a temperature program from 45~ for 2 min to 230~ at 15~ Electron impact mass spectra (EI-MS) were obtained at 70 eV using a Finnigan 4510 GC-MS, equipped with a 30-m DB-1 column. Volatiles from exuviae of C. cydoniae were derivatized with o-benzylhydroxylamine after the method of Ollett et al. (1986). Approximately 50 exuviae were crushed in a 250-/~1 gastight syringe equipped with a GTS Valve (Hamilton Company, Reno, Nevada) in the off position, and then the volatiles were bubbled through the derivatizing solution by opening the valve and slowly depressing the syringe plunger. The o-benzyloximes were also analyzed by chemical ionization mass spectrometry (CI-MS) using NH3 and CH4 as reagent gases. One underivatized sample of C. cydoniae exuviae crushed in a gastight syringe as described above was analyzed by GC-MS. The sample was injected under pressure using the GTS valve with the GC oven turned off and cooled to ca.

2310

ALDRICH ET AL.

0~ with Dry Ice in order to obtain EI-MS of very volatile compounds normally inseparable from organic solvents. A head-space sample of acetaldehyde was analyzed in the same fashion. The unknown compound eluting at a retention time (RT) of 11.0 min (Figures 2 and 3) was isolated from an exuvial extract of G. solani by applying the concentrated CH2C12 extract to a Baker 3-ml silica gel extraction cartridge, and eluting successively with hexane (5 ml), 10% ethyl acetate in hexane (5 ml), and 25 % ethyl acetate in hexane (3 ml). The unknown reproducibly eluted in the last fraction. A small sample thus prepared was hydrogenated for 20 rain (PtO2, EtOH, 1 atm). For NMR, ca. 165 mg of Gargaphia exuviae were processed as described, and the unknown was further purified by preparative GC (15 m • 0.53 mm ID, DB-1). A [1H]NMR spectrum for this compound was recorded in D6-benzene at 300 MHz on a General Electric QE-300 instrument with TMS as an internal standard. Standards. The following compounds were obtained commercially: acetaldehyde and 1-nonanol (Aldrich Chemical, Milwaukee, Wisconsin); geraniol and nerol (Bedoukian Research, Danbury, Connecticut); and linalool (Givaudan, Clifton, New Jersey). (E)-4-Oxo-2-hexenal was synthesized according to Ward and VanDorp (1969). Nerolidol aldehyde [(E,E)-2,6,10-trimethyl-10hydroxy-2,6,11-dodecatrienal] was synthesized after the method of Demole and Enggist (1973). Bioassays. Bioassays were begun with C. cydoniae using a procedure similar to that of Keams and Yamamoto (1981). A feeding aggregation of > 10 fifth-instar nymphs was chosen for observation, and a conspecific nymph from another aggregation (normally on a different cutting) was sacrificed for the test. A leg was grasped with forceps and the nymph was held 5-10 mm above the test aggregation for ca. 1 min; then the entire body of the test nymph was positioned between the prongs of the forceps, crushed, and immediately held over the test aggregation. A response was considered positive if most of the nymphs walked away within i min. Fifth-instar nymphs in the process of ecdysis (evident in aggregations by their unmelanized cuticle) were also bioassayed as above, after which the fresh exuviae were tested for alarm induction as follows: First the exuviae were lifted by an appendage and offered to the test aggregation for 1 min; then the DAG reservoirs of the exuviae were ruptured with forceps under a microscope and the macerated exuviae were again held over the same test aggregation. Alarm pheromone tests were performed with the hawthorn lace bug by positioning a leaf aggregation of C. cydoniae fifth-instar nymphs about 1 cm in front of the exit port from the thermal conductivity detector of the GC equipped with a DB-1 megabore column. For GC-etIluent testing, several aggregations ( > 200 individuals) of C. cydoniae would be quick-frozen by placing the leaves

LACE BUG PHEROMONE

2311

on which they were aggregated onto Dry Ice. The frozen nymphs were brushed off the leaves into a 10-ml gastight syringe barrel, having a Teflon membrane (0.5-ram pore size; Millipore Corp., Bedford, Massachusetts) in the bottom, and equipped with a GTS valve. The barrel of the syringe was flushed with a stream of N2, and the plunger was inserted as far as possible without crushing the nymphs. Then, with the GTS valve in the off position, the nymphs were crushed and the volatiles injected into the GC under pressure. Control experiments were performed in an analogous manner with an empty syringe. Responses were judged positive if most of the aggregated nymphs were dispersed by the effluent. For bioassay of the DAG secretions (test I), the contents of the anterior and/or posterior gland exuvial reservoirs were extracted in a micropipet as previously described. A leaf with a nymphal aggregation to be tested ( > 10 thirdto fifth-instar nymphs) was placed on the microscope stage with the illumination lowered so as not to warm the insects. A second micromanipulator was arranged on the opposite side of the microscope stage with a pair of forceps holding a small triangular piece of filter paper. The DAG secretion(s) to be tested were transferred to the point by bringing the pipet in contact with the paper and then the impregnated paper was quickly lowered to within about 0.5 cm of the aggregation. As part of this series, solutions (3/~g//zl CH2C12) of geraniol, linalool, (E)-2-hexenal, (E)-4-oxo-2-hexenal, and combinations thereof, were tested by wetting the point tip with test solution and proceeding as for the gland extracts~ Absence of a response was scored a ( - ) , a moderate response was (+), and complete dispersal was a (+). In an effort to fully mimic the natural response of the hawthorn lace bug with synthetic standards, a second bioassay series (test II) was performed using only C. cydoniae nymphs. The day before the testing session, five fifth-instar nymphs (or fourth instars if necessary) were placed on single hawthorn leaves with petioles in water vials; 25-50 such preparations were set up for each session. Test solutions of acetaldehyde, geraniol, linalool, nerol, (E)-2-hexenal, (E)-4-oxo-2-hexenal, and combinations thereof, were prepared in CH2C12 at a range of concentrations, plus CH2C12 controls (see Table 2 below). Responses were scored by a "blind" observer with no prior knowledge of the material being tested. An aggregation to be tested was positioned by one person (J.W.N.) on a microscope stage with reduced illumination, while a second person (J.R.A.) applied 1 #1 (Drummond Microcap) of test solution to a paper point on the end of an insect pin. The loaded filter paper was held about 0.5 cm above the aggregation and the response of the majority of nymphs was scored as follows: 0 = no response, 1 = agitation and jiggling in place, 2 = wandering, 3 = slowly moving away (after a delay of > 20 sec), and 4 = quickly moving away (within 20 see).

2312

ALDRICH ET AL.

RESULTS Gargaphia solani nymphs do not have setal droplets (Figure 1A). In C. cydoniae nymphs, setal droplets are abundant (Figure 1B), and after ecdysis or when droplets are brushed off due to overcrowding, they are regenerated. In G. solani the openings of the anterior and posterior DAGs are evident in SEMs, but in C. cydoniae the external openings of the DAGs are not apparent. Preliminary experiments (J.W.N.) with the hawthorn lace bug confirmed that a potent alarm pheromone is released from crushed nymphs. Feeding aggregations of nymphs usually respond within 10-15 sec by removing their stylets from the leaf, jiggling and bobbing, and rapidly dispersing. If a C. cydoniae nymph is grasped without being crushed, aggregated nymphs are not alarmed. Occasionally, aggregated C. cydoniae nymphs failed to respond to crushed nymphs. Microscopic examination (J.R.A.) indicated that the alarm pheromone was released only if the abdomen was crushed, suggesting that the pheromone

FIG. 1. (A) Gargaphia solani fifth-instar nymph with setae lacking droplets and, (B) Corythucha cydoniae fifth-instar replete wi~ setae bearing droplets (SEMs: 80 •

LACE BUG PHEROMONE

2313

FIG. 1. Continued

is contained in the DAGs. Exuviae with intact DAGs did not elicit alarm, but when the exuvial DAG reservoirs were raptured the full range of alarm behavior was elicited. When the abdomen of the newly ecdysed fifth instar (ca. 10 min after ecdysis) was crushed and the nymph was held near an aggregation, no alarm was induced. This experiment was repeated many times with the same result. In contrast to C. cydoniae, it is not necessary to crash a nymph of G. solani in order to elicit alarm in aggregated conspecific nymphs. Merely grasping a nymph firmly by a leg is usually sufficient to elicit clear-cut alarm among clustered G. solani nymphs. Pinched G. solani nymphs that induced alarm had a distinctly terpenoid odor (J.R.A., personal observation). Techniques for directly sampling DAG secretions were first perfected using G. solani fiflh-instar nymphs. Extracts of exuviae with intact reservoirs exhibit four prominent peaks by GC, but late-eluting compounds (characteristic of tingids having setal droplets) are absent (Figure 2; bottom trace, compounds 14). A fifth-instar G. solani nymph produces about 8 ng of 1, 19 ng of 2, 34 ng

2314

ALDRICH ET AL.

~_

Gargaphia so/ani Dorsal Abdominal

Gland (DAG) Secretions ~ o /l~2AnteriorDAG 1

(E)-2-hexenal o 2

aor DAG

(~-4-o•

Anterior+ 4 PosteriorDAG 3

geraniol

4

nerolidol aldehyde

~

~

6

8

Minutes

FIG. 2. Dorsal abdominal gland (DAG) secretions of Gargaphia solani fifth-instar nymphs: anterior DAG (one exuvial reservoir; splitless GC), posterior DAG (one exuvial reservoir; splitless GC), and anterior plus posterior DAGs (60 exuviae; 1 : 50 GC split).

of 3, and 100 ng of 4, based on quantitation of an extract of 20 exuviae. Extracts of filter paper pressed onto the tergum of G. solani nymphs produced GC traces (not shown) matching those of exuvial DAG reservoirs, verifying that these nymphs are able to voluntarily release secretion from both DAGs. GC analysis of secretion isolated from the anterior DAG reservoir of exuviae showed that this secretion contains compounds 1, 2, and 4 (Figure 2, top), whereas secretion isolated from the posterior DAG exuvial reservoir contains compound 3 (Figure 2, middle). For C. cydoniae, separate extraction and GC analysis of the anterior and posterior DAG reservoirs also showed that the two glands produce qualitatively different secretions (Figure 3). Compound 4 is the predominant component in the anterior DAG secretion of the hawthorn lace bug, but compound 2 was not detected and compound 1 was barely detectable. In the posterior DAG secretion of C. cydoniae, compound 3 occurs with another component (5) not found in G. solani nymphs.

LACE BUG PHEROMONE

2315

CorythucacydoniaeDorsal Abdominal Gland (DAG) Secretions Anterior DAG

1 4

(E)-2-hexenal

~OH 3

1

/ t

geraniol

~OH

4 nerolidol aldehyde

Posterior DAG

Minules

linalool

FIG. 3. Dorsal abdominal gland (DAG) secretions of Corythucha cydoniae fifth-instar nymphs: anterior DAG (10 exuvial reservoirs; splitless GC) and posterior DAG (two exuvial reservoirs; splitless GC). The EI-MS of compounds 1, 2, and 5 matched those for the following compounds commonly encountered in heteropteran exocrine secretions (Aldrich, 1988a,b): (E)-2-hexenal, (E)-4-oxo-2-hexenal, and linalool, respectively. The EI-MS of compound 3 gave a near-perfect computer-matched fit to the MS of geraniol. The identities of these four compounds were verified by coinjection with authentic compounds. A match for the EI-MS of compound 4 was not found in the computerized mass spectral library, but the pattern of lower mass (m/z < 140 ainu) ions resembled that of nerolidol. The CI-MS with NH3 indicated a molecular weight of 236, and ND 3 CI-MS showed that the molecule contains one exchangeable hydrogen. Hydrogenation produced a compound with a molecular weight of 244 and two exchangeable hydrogens, suggesting the initial presence of three C = C bonds and one C = O bond, the latter possibly as an aldehyde. The [lH]NMR (C6D6) spectrum was as follows: 6 1.10 (s), 1.46 (s), 1.64 (s), 1.84 (t, J = 7.2), 2.03 (m), 4.96 (d, J = 11.7), 5.11 (t, J = 7), 5.19 (d, J = 16.8), 5.74 (dd, J = 18 and 11), 5.89 (t, J = 7), and 9.31 (s). The singlet at 6 = 9.31 clearly indicates an aldehyde, and both the mass spectral and NMR data are

2316

ALDRICH ET AL.

consistent with an oxygenated nerolidol, most likely with one of the distal methyls oxidized to --CHO. A search of the literature revealed that Demole and Enggist (1973) had synthesized (E,E)- and (E,Z)-2,6,10-trimethyl-10-hydroxy2,6,11-dodecatrienal. Although different solvents precluded a direct comparison of our NMR data with those reported, the E,E isomer seemed most consistent with our data. Spectral comparisons and GC coinjection experiments of the synthetic standard of (E,E)-2,6,10-trimethyl-10-hydroxy-2,6,11-dodecatrienal with component 4 (designated nerolidol aldehyde) confirmed this identification for the natural product. Bioassays testing the response of C. cydoniae nymphs to volatiles from crushed conspecifics eluting from the GC were performed before determining that the DAGs produce the alarm pheromone. Nymphs at the effluent were alarmed within 30-40 sec after injection, indicating that one or more extremely volatile compounds were involved. Once it was established that exuvial DAG reservoirs contain the alarm pheromone, the low temperature GC-MS experiment to identify the very volatile compound(s) was undertaken using C. cydoniae exuviae. A small peak occurred just after a scan time of 150 sec at ca. 0~ and the El-MS of this peak gave a high computer fit for the MS of acetaldehyde (Figure 4A). GC-MS of a head-space sample (ca. 90 ng) of authentic acetaldehyde under similar GC conditions (temperature was difficult to exactly duplicate) produced a peak having nearly the same retention time and an EI-MS virtually identical to that for the exuvial compound (Figure 4B). CI-MS of C. cydoniae exuvial volatiles derivatized with o-benzylhydroxylamine confirmed the presence of the benzyloxime of acetaldehyde (CH3CH-----NOCH2C6H5; mol wt 149); NH3 CI-MS m/z (%): 167 ([M+NH4] +, 100) and 184 ([M+(NH3)2+H] +, 55); CH 4 CI-MS m/z (%): 91 (100) and 150 ([M+H] +,

75). The first intraspecific bioassays using DAG secretions showed that alarm behavior in C. cydoniae occurs only if nymphs are simultaneously exposed to anterior and posterior DAG secretions, whereas G. solani nymphs are alarmed by exposure to posterior DAG secretion alone (Table 1). Inclusion of the anterior DAG secretion with the posterior secretion caused G. solani nymphs to respond sooner. When C. cydoniae nymphs were tested using secretions from G. solani exuviae, the results were the same; neither secretion alone had significant activity, while the combined secretions were highly alarming. The result of the one reciprocal interspecific test performed was also consistent with the intraspecific tests; posterior DAG secretion from C. cydoniae elicited alarm in G. solani nymphs. Bioassays with synthetic compounds 1, 2, 3, and 5 showed that for G. solani nymphs geraniol (3) by itself released typical alarm behavior (Table 1). Geraniol with (E)-2-hexenal (1) and/or (E)-4-oxo-2-hexenal (2) may have slightly hastened the response of Gargaphia nymphs, but this effect was not precisely assessed. Linalool (5) by itself or with 1 or 2 was inactive for G.

LACE BUG PHEROMONE

2317

Corythuca cydoniaeExuvlae

A

29

Air 44

L

L~

,[,

!

I

I

Acetaldehyde head-space 29

B Air

44

15

9

i

50

i

100 150 200 Scan Time in Seconds

250

FIG. 4. Low-temperature gas chromatography-mass spectrometry (GC-MS) of (A) Corythucha cydoniae exuviae and (B) acetaldehyde head-space (reconstructed ion chromatograms and EI-MS).

solani. For C. cydoniae nymphs, the bioassay results using synthetics were less clear than for G. solani. Although (E)-4-oxo-2-hexenal (2) was not detected in the DAGs of C. cydoniae nymphs, this compound alone showed moderate alarm-inducing activity, and combinations 2 + 3, 2 + 5, and I + 2 + 3 yielded

2318

ALDRICH ET AL.

TABLE 1. ALARM PHEROMONE TEST I USING DORSAL ABDOMINALGLAND (DAG) SECRETION AND SYNTHETIC COMPOUNDSa

Response (# replicates) Test material

G. solani

C. cydoniae

G. solani DAG secretion Anterior Posterior Anterior and posterior C. cydoniae DAG secretion Anterior Posterior Anterior and posterior Geraniol (3) Linaloolb (5) (E)-2-Hexenal (1) 4-Oxo-(E)-2-hexenai C(2)

- (3) + (8) + (1)

- (3) - (2) + (4)

NT d + (2) NT + (3) - (1) - (2) - (2)

+ +

1 + 3 1 + 5b

+ (I) - (1)

+ (1)

2 + 3c

+_ (2)

+ (1)

2 + 5 cb

- (1)

+ (1)

+ (5)

+ (3)

2 + 1+3

~

(9) (9) (7) (3) (2) (2) (2) + (1)

~No response = - ; moderate response = +; positive response = +. Secretion collected from exuviae (_< 1 day after-ecdysis); synthetics total 3 #g/#g CH2C12. Test material applied to filter paper point and held over aggregations (>_ 10 third-fifth instars) (details in text). bNot found in G. solani DAGs. CNot found in C. cydoniae DAGs. dNot tested.

the highest activities. Despite the positive response rating for the latter three c o m b i n a t i o n s , it m u s t be p o i n t e d out that this c h e m i c a l l y elicited alarm b e h a v i o r did not fully m i m i c the natural response o f C. cydoniae n y m p h s . T h e second bioassay series with C. cydoniae n y m p h s used a more precise scoring system and included two additional c o m p o u n d s (acetaldehyde and nerol) (Table 2). The results are consistent in that ( E ) - 4 - o x o - 2 - h e x e n a l (2) was more active than ( E ) - 2 - h e x e n a l (1) a n d the m o n o t e r p e n e s . The c o m b i n a t i o n o f 2 + 3 was quite active (score = 3.33 + 0.49), whereas c o m b i n i n g nerol, which is a geometrical i s o m e r o f geraniol (3), with 2 greatly reduced activity. The role o f acetaldehyde in the alarm p h e r o m o n e r e m a i n s questionable; the two highest c o n c e n t r a t i o n s tested i n d u c e d responses interpreted as alarm, b u t the observer s o m e t i m e s noted the responses " s e e m e d d i f f e r e n t . " A n attempt to demonstrate a s y n e r g i s m b e t w e e n acetaldehyde and c o m p o u n d s 1 - 3 (concentration = 0.0065

0

Acetaldehyde (CH3CHO)

0.73 -t- 0.16 (33)

0.125 0.13 __. 0.09 (15)

0.25 0.50 + 0.27 (14)

0.50

1.67 5:0.33 (3) 2.25 5:0.67 (4) 1.00 5:0,58 (3)

0.625

1.00 + 0.21 (4)

1.58 5:0.42 (12) 1.73 + 0.23 (15)

1.00 + 0.49 (15) 0.33 5:0.12 (t5)

0.0065 M

2.33 + 0.88 (3) 1.00 5:0.41 (4) 2.33 + 0.88 (3)

0.33 5:0.33 (3)

3.17 5:0.54 (6)

1.25

2.50

0.89 5: 0.20(9) 0.67 5:0.67 (3) 0.33 + 0.33 (3) 1.67 5:0.33 (3) 2.25 5:0.25 (8) 1.00 5:0.31 (7) 2,00 _+ 0.58 (4) 0.67 + 0.33 (3) 3.33 + 0.49 (6) 1.33 + 0.33 (3) 1.33 _+ 0.88 (3)

"Responses were scored from 0 (no response) through 4 (full response) as described in the text. Scores listed are X 5: SEM(N ). bResponses to CH2CI2 alone = 0.60 5:0.13 (40), 'Units are # g / # l CH2C12 except for one solution tested on a molar basis (0.0065 M). dNot found in C. cydoniae or G. solani DAGs. ~Occurs in G. solani DAGs, but not detected in C. cydoniae DAGs,

1 + 2" + 3 + CH3CHO

2" + nerola

2" + 5

2" + 3

1 + nerola

1 +5

1 + 3

(E)-4-Oxo-2-Hexenal" (2)

(E)-2-Hexenal (1)

Neml a

Liaalool (5)

Geraniol (3)

0.0125

Compound(s) h

Concentration"

TABLE 2. ALARM RESPONSES OF Corythuca cydoniae NYMPHS TO SYNTHETIC COMPOUNDS"

3.00 _+ 0.58 (3)

12.5

2320

ALDRICHET AL.

M) failed to show any appreciable synergism. None of the responses recorded in this bioassay series consistently equaled a natural response. Nerolidol aldehyde (4) was synthesized after both bioassay series were completed, but when 4 was tested (2.5/zg//zl) against C. cydoniae nymphs it failed to elicit alarm behavior either alone or with 3 and/or 5.

DISCUSSION In G. solani, matemalism appears to substitute for the droplet chemical defense. A potent alarm pheromone is a second line of defense for Gargaphia, as well as for other tingids, and our discovery that tingid alarm pheromones emanate from the DAGs adds to the list of heteropterans having alarm pheromone activity associated with the secretions from these glands (Blum, 1985). Gargaphia solani nymphs clearly are able to release DAG secretion. Notwithstanding the discussion of Keams and Yamamoto (1981), the alarm pheromone system of the eggplant lace bug does not resemble that of various treehoppers (Homoptera: Membracidae), where the body wall of a nymph must be ruptured for the alarm pheromone to be released (Nault et al., 1974; Wood, 1976). The Gargaphia alarm-aggregation system seems much more like that of cotton stainer bugs (Pyrrhocoridae: Dysdercus spp.), where the posterior DAG secretion of nymphs triggers alarm and the two anterior DAGs produce an aggregation pheromone (Farine, 1987; Calam and Youdeowei, 1968). A diol corresponding to the tingid sesquiterpenoid is known from plants (Bohlman and Zdero, 1980), but nerolidol aldehyde is a new natural product. Whether this compound is part of an aggregation pheromone in Gargaphia remains to be determined. While the alarm pheromone system of Gargaphia is not exceptional for Heteroptera, the system as it appears to function in Corythuchais extraordinary. Evidence that the alarm message in Corythucha is due to synergism between the two DAG secretions is, we believe, compelling. Yet all indications are that the nymphs are unable to voluntarily emit their DAG secretions; the glands must be physically ruptured for the pheromone to escape, a system truly analogous to that of treehoppers. Furthermore, many other tingid lineages having setal droplets might possess alarm pheromone systems as in Corythucha. This is apparently true for nymphs of the azalea lace bug Stephanitispyrioides and the rhododendron lace bug S. rhododendri (J.W.N., personal observation). The occurrence of sealed DAGs is not unprecedented in the Heteroptera; adults and nymphs of some mirids in the Bryocorinae specializing on toxic plants reportedly have one DAG with no external opening (Aryeetey and Kumar, 1973). When immatures of these mirids are disturbed, the DAG is vibrated, and drops of fluid simultaneously exude from setae over their bodies. It is pos-

LACE BUG PHEROMONE

2321

sible that host-plant alkaloids are sequestered in the DAG and delivered when needed via the blood or an epidermal syncitium, as in milkweed bugs (Lygaeidae: Oncopeltus spp.) (Scudder et al., 1986), to setae or cuticular weak points (Aldrich, 1988a). However, none of the compounds identified in Corythucha DAG secretions have been detected in their setal droplets (Lusby et al., 1989). Many questions remain unanswered in the present study, not the least of which is: what role, if any, does acetaldehyde play in the alarm behavior of Corythucha? For that matter, is acetaldehyde present in Gargaphia, and why do crushed adults cause delayed alarm among their nymphs (Kearns and Yamamoto, 1981)? (A single effort to directly observe acetaldehyde from exuviae of G. solani by low-temperature GC was inconclusive, and the metathoracic scent gland secretion has not been analyzed for any tingid adults.) Perhaps the most interesting question to emerge from this study is: has the alarm pheromone system of some nonmaternalistic tingids become totally altruistic?4 Acknowledgments--We wish to thank Rose Haldenmann for tireless efforts maintaining tingid colonies and preparing for bioassays, and Rolland Waters for the NMR spectrum. We are also grateful to Kim Kal for helping to collect Gargaphia and prepare extracts. Ken Wilzer helped collect tingids and provided invaluable computer assistance in the preparation of the manuscript, and Dawn Harrison performed many of the mass spectral analyses.

REFERENCES ALDRICH, J.R. 1988a. Chemical ecology of the Heteroptera. Annu. Rev. Entomol. 33:211-238. ALDRICH, J.R. 1988b. Chemistry and biological activity of pentatomoid sex pheromones, pp. 417431, in H.G. Cutler (ed.). Biologically Active Natural Products: Potential Use in Agriculture. ACS Symposium Series No. 380, Washington, D.C. ARYEETEY, E.A., and KUMAR,R. 1973. Structure and function of the dorsal abdominal gland and defence mechanism in cocoa-capsids (Miridae: Heteroptera). J. Entomol. Ser. A 47:181-189. BLUM, M.S. 1985. Alarm pheromones, pp. 193-224, in Comprehensive Insect Physiology, Biochemistry and Pharmacology, Vol. 9. Behaviour. G.A. Kerkut and L.I. Gilbert (eds.). Pergamon Press, Oxford, U.K. BOHLMAN,F., and ZDERO, C. 1980. Neue Sesquiterpenlactone und Nerolidol-derivate aus UrsiniaArten. Phytochemistry 19:587-591. CALAM, D.H., and YOUD~OWEI, A. 1968. Identification and functions of secretion from the posterior scent gland of fifth instar larva of the bug Dysdercus intermedius. J. Insect Physiol. 14:1147-1158. COBDEN, R.H. 1978. Evolutionary Trends in Heteroptera. Part II. Mouthpart Structures and Feeding Strategies. Veenman & Zonen, Wageningen, The Netherlands. 407 pp. 4Nymphs of the olive lace bug, Froggattia olivinia, collected in Queensland, Australia, since the submission of this manuscript were observed microscopically (J.R.A.) to possess short setae with conspicuous droplets over the entire body and to be able to excrete material from their DAGs. Mild alarm behavior was elicited from crushed nymphs and exuviae, and GC-MS analysis of an exuvial extract indicated the major DAG components are 2-phenethyl and benzyl esters of isovaleric and isobutyric acids.

2322

ALDRICH ET AL.

DEMOLE, E., and ENC~XST,P. 1973. Applications synth6tiques de la cyclisation d'alcools tertiares ~/-ethyl6niques en a-bmmotetmhydrofurannes sons l'action du N-bromosuccinimide. III. Synth~se du trimethyl-2,6,10-hydroxy-10-dodecatriene-2,6,11-al (trans/trans et trans/cis), hydroxy-aldehyde apparente aus sinensals. Helv. Chim. Acta 56:2053-2056. DRAKE, C.J., and RUNOFF,F.A. 1965. Lacebugs of the World, A Catalogue (Hemiptera: Tingidae). U.S. National Museum Bulletin 243. FARINE, J.-P. 1987. The exocrine glands of Dysdercus cingulatus (Heteroptem, Pyrrhocoridae): Morphology and function of nymphal glands. J. Morphol. 194:195-207. FAEa'n, S.H. 1989. Maternal care in a lace bug, Corythucha hewitti (Hemiptera: Tingidae). Psyche 96:101-110. FINK, D.E. 1915. The eggplant lace-bug. Bulletin U.S. Department of Agriculture 239, pp. 1-7. KEARNS,R.S., and YAMAMOTO,R.T. 1981. Maternal behavior and alarm response in the eggplant lace bug, Gargaphia solani Heidemann (Tingidae: Heteroptera). Psyche 88:215-230. LWINOSTONE,D. 1978. On the body outgrowths and the phenomenon of "sweating" in the nymphal instars of Tingidae (Hemiptera: Heteroptera). J. Nat. Hist. 12:377-394. LusBg, W.R., OLIVER,J.E., NEAL, J.W., JR., and HEATh, R.R. 1989. Acylcyclohexanediones from setal exudate of hawthorn lace bug nymph Corythucha cydoniae (Hemiptera: Tingidae). J. Chem. Ecol. 15:2369-2378. NAULT, L.R., WOOD, T.K., and GoFF, A.M. 1974. Treehopper (Membracidae) alarm pheromones. Nature 249:387-388. OLIVER, J.E., LtJSBY, W.R., and NEAL, J.W., JR. 1990. Exocrine secretions of the andromeda lace bug Stephanitis takeyai (Hemiptera: Tingidae). J. Chem. Ecol. 16: 2243-2252. OLLETT, D.G., ATTYGALLE,A.B., and MORaAN, E.D. 1986. Microchemical method for determining formaldehyde, lower carbonyl compounds and alkylidene end groups in the nanogram range using the Keele micro-reactor. J. Chromatogr. 367:207-212. SCUDDER,G.G.E., MOORE,L.V., and ISMAN,M.B. 1986. Sequestration of cardenolides in Oncopeltusfasciatus: Morphological and physiological adaptations. J. Chem. Ecol. 12:1171-1187. TALLAMY,D.W., and HORTON, L.A. 1990. Costs and benefits of the egg-dumping alternative in Gargaphia lace bugs (Hemiptera: Tingidae). Anita. Behav. 39:352-359. WARD, J.P., and VANDORP, D.A. 1969. A stereospecific synthesis of 4-oxo-2-trans-hexenal. Recueil 88:989-993. WOOD, T.K. 1976. Alarm behavior of brooding female Umbonia crassicornis (Homoptera: Membracidae). Ann. Entomol. Soc. Amr. 69:340-344.