Journal (if Chemical Ecology, Vol. 9, No. 12, 1983

CHEMISTRY AND FUNCTION OF MANDIBULAR GLAND PRODUCTS OF BEES OF THE GENUS Exoneura (HYMENOPTERA, ANTHOPHORIDAE) 1

JAMES

H. C A N E 2 a n d C H A R L E S

D. M I C H E N E R

Department of Entomology University o[ Kansas, Lawrence, Kansas 66045

(Received September 27, 1982; revised February 14, 1982) Abstract--Female Exoneura richardsoni, E. bicolor, and E. bicincta (Hymenoptera: Anthophoridae) release a pungent, staining liquid from their mandibular glands upon disturbance. This secretion is primarily composed of ethyl dodecanoate, with lesser amounts of homologous ethyl and methyl esters, salicylaldehyde, and 1,4-benzoquinone. The secretion elicits vigorous grooming when topically applied to antennae of Formica ants. The shared, unique combination of mandibular gland lipids of these three Exoneura species supports their monophyletic classification, while the presence of salicylaldehyde may associate Exoneura (Allodapini) with Pith#is (Ceratinini). Key Words--Exoneura, Allodapini, Hymenoptera, bees, defense, mandibular glands, ethyl decanoate, salicylaldehyde, 1,4-benzoquinone, ester, aldehyde, quinone. INTRODUCTION It has long been suspected t h a t the A u s t r a l i a n genus E x o n e u r a a n d p r o b a b l y o t h e r a l l o d a p i n e bees secrete m a t e r i a l s f r o m the m o u t h region that repel ants a n d p e r h a p s o t h e r p r e d a t o r s . W h e n held between the fingers, a female E x o n e u r a often e x u d e s a r a t h e r copious, b r o w n , yellowish, or r e d d i s h liquid o n t o the skin. This liquid stains white p a p e r or one's skin y e l l o w i s h - b r o w n , and has an o d o r suggestive o f that of c a r a b i d o r t e n e b r i o n i d beetles ( C . D . M . ) ~Contribution number 1842 from the Department of Entomology, University of Kansas, Lawrence, Kansas 66045, "Present address: Division of Entomology and Parasitology, University of California, Berkeley, California 94720. 1525 00984)33 I 83 1200-1525$03,00 0 -r 1983 Plenum Publishing Corporation

1526

CANE AND M ICHENER

with overtones of raisins (J.H.C.). [The sweetish odor may result from an admixture of regurgitated crop contents but ethyl esters such as ethyl decanoate in the secretion may account for it (Amerine and Roessler, 1976; Heath and Pharin, 1978).] Michener (1965) wrote that the females "emit this material either at the nest or, if captured or greatly disturbed, away from the nest. On one occasion [he] found females of E. bicolor so c o m m o n . . . visiting flowers . . . that when the bushes bearing the flowers were beaten with a net, the odor could be perceived several feet away. The effectiveness of adult females in defending their nests is shown by a series of nests of E. variabilis set up for observation in Brisbane. Within a day those containing no adults were robbed of all immature stages by ants. Those containing adults mostly survived for many days." Because of the observations reported above, one of us (C.D.M.) suggested that the liquid from the mandibular gland serves to repel predators such as ants and that it probably contains quinones, well-known repellants produced by other arthropods (Eisner and Meinwald, 1966) but not previously recognized a m o n g glandular products of bees. [A similar function, however, has been shown for other mandibular gland products of some bees (Cane, 1982; Wheeler et al., 1977).] The present paper documents the' mandibular gland lipid chemistry and its defensive function for three species of Exoneura.

METHODS AND MATERIALS To properly relate this paper to previous work on Exoneura biology, the following notes on the specific names seem necessary: (1) Exoneura richardsoni Rayment is probably the same as E. variabilis Rayment, the name used in a study of a Queensland population (Michener, 1965). However, since richardsoni was described from the Dandenong range, the source of our material, we feel that until the group is studied taxonomically, it is appropriate to use that name even though it will probably ultimately be synonymized under variabilis. (2) Exoneura bicolor Smith is used in a traditional and possibly correct sense for a form similar to E. harnulata Cockerell but with the upper hooks of the yellow clypeal mark of the female usually reduced or absent, the clypeal yellow area of the male much reduced, and the abdominal pubescence of the male all black. E. bicolor, however, was described from West Australia and Tasmania, probably from specimens belonging to two species, and the proper application of the name remains in doubt. (3) Exoneura bicincta Rayment is similar to and likely to be a synonym of E. angophorae Cockerell. Our males appear to agree with the male type of bicincta, but are also similar to males of E. angophorae in the Snow Entomological Museum, University of Kansas. Voucher specimens of all

MANDIBULAR GLAND OF E x o n e u r a

1527

three species are placed in the Snow Entomological Museum and in the Australian National Insect Collection, Canberra. Nests of Exoneura bicolor, bicincta, and richardsoni in dead blackberry (Rubus) stems were collected in the Dandenong Ranges, Victoria, Australia, in May 1982, by Mr. Michael Schwarz of Monash University. Immature forms were absent, males were scarce, and the bees were evidently in overwintering quarters. C.D.M. transferred the bees to glass tubes 5 mm in inside diameter (plugged with pipe cleaners), one for each nest. The tubes were refrigerated (about 4~ when practical, and carried by air to Kansas in a plastic bag containing a little moist cotton for humidity. Every specimen arrived alive and was active when not refrigerated. Tests with ants were made in Lawrence, Kansas, in July, using female Exoneura bicolor only. Workers of Crematogaster lineolatus were in a rather weak trail moving up and down a post. Workers of Formica pallidefulva were placed individually in 4-dram vials for topical treatments and observations. Each treatment consisted of touching the middle part of one flagellum a few times with a needle or closed fine forceps bearing the test material. For the chemical studies, made in June, females that had been continuously chilled for at least four days were killed on Dry Ice; then 15-35 glands (depending on the species) were removed with their attached mandibles and placed in chilled methylene chloride. The chemical analysis was made on the day of dissection. All spectral analyses were made with a Girdel gas chromatograph (GC) with split Ros injection, coupled to a Ribermag R-10-10 quadrupole mass spectrometer and P D P 8/A computer in the Department of Chemistry, University of Kansas. The GC oven was programed at 5~ steps from 70~ to 250~ followed by a 10-min isothermal period at 250~ Components were separated using a 12-m, 0.20-mm inside diameter methyl silicone wall coated (WCOT) fused silica capillary column (Hewlett-Packard). Total ion chromatograms were compiled from 1.2-sec scans from 40 to 500 amu at 70 eV, in EI mode. All spectra were taken from chromatogram peak maxima with background subtraction. Components of glandular extracts were positively identified by a combination- of matching retention times (_+2% range) and comparison of our spectra with published spectra (quinone) and authentic samples (esters). Individual injections contained between one half and one glandular equivalent of extracted secretion.

RESULTS

Defensive Function of Mandibular Gland Products. Preliminary observations were made with the ant Crematogaster lineolatus. Freshly macerated heads of females of Exoneura bicolor applied to the ant trail on a post

1528

CANE AND MICHENER

were n o t attractive; the trail was interrupted, but within 3-5 m i n (three trials) the ants r e s u m e d m o v e m e n t a r o u n d the macerated material, which was ignored. Freshly macerated thoraces or a b d o m e n s , however, were immediately attractive; ants in the trail stopped a n d a p p a r e n t l y fed from the material (three trials for each b o d y part). Likewise, five macerated heads inside a glass tube 5 m m in diameter and 5 cm long placed a m o n g ants at the bott o m of the post were not t o u c h e d in an hour, a l t h o u g h at least one ant wandered a b o u t inside the tube. Macerated thoraces a n d a b d o m e n s of two bees inside a n o t h e r tube were removed by the ants d u r i n g the first half hour. These crude experiments suggest either that heads do not c o n t a i n attractive materials present in thoraces a n d a b d o m e n s or that heads c o n t a i n a repellant s u b s t a n c e absent f r o m thoraces a n d a b d o m e n s . T o get less equivocal i n f o r m a t i o n o n possible repellance, a larger ant, Formica pallidefulva, was used for topical application of materials. Secretion of the m a n d i b u l a r glands, t a k e n o n a needle or closed fine forceps directly from the m o u t h area of live Exoneura, was applied (in invisible a m o u n t s ) to one ant a n t e n n a l flagellum. As controls, there is the untreated flagellum of the same a n t a n d other ants similarly treated but with water or raw egg white to see whether sticky material might elicit the same response as the Exoneura secretion. As Table 1 shows, the cleaning response to the Exoneura secretion was intense a n d lengthy. The a n t e n n a was repeatedly (up to 46 times per minute) d r a w n t h r o u g h the strigilis on the foreleg for up to 15 min. Moreover, after every two or three a n t e n n a l cleaning m o v e m e n t s , some individuals cleaned the foreleg involved by d r a w i n g it t h r o u g h the m o u t h p a r t s . No such cleaning of the other foreleg was observed. It was as t h o u g h an irritant affected not only the a n t e n n a but the leg used to clean the a n t e n n a . Egg white often elicited short bouts of cleaning. The m o v e m e n t s were

TABLE 1. ANTENNALCLEANING BY Formica pallidefulva AFTERMATERIALFROM MANDIBULAR GLANDS OF FEMALESOF Exoneura bicolor WAS PLACED ON ONE FLAGELLUM, WITH EGG WHITE AND WATER FOR COMPARISON

Exoneura secretion Duration of r e s p o n s e (rain)~' No. of cleaning movements per min/'

2,5,12,12,15 .g = 37 (20-46, N - 15)

Untreated antenna

Egg white

Water

0,1,1,2,3

0,0,0, l, l

0(N = 15)

X = 8 (0-12, N = 10)

)7 = 5 (0-8, N = 10)

3T= 1 (0-5. N = 35)

~'Based on five different ants for each treatment. The numbers given are minutes after treatment until cleaning of the treated antenna either ceased or decreased to the low level at which the untreated antenna was cleaned. 1'Based on counts for full minutes.

MANDIBULAR GLAND OF Exoneura

1529

TABLE 2. LIPID COMPONENTS OF MANDIBULAR GLAND PRODUCTS OF Exoneura

Proportional integreated peak areas from GC (%)~ Compound

E. richardsoni

E. bicolor

E. bicincta

0.2 0.5 5.5 4.6 0.1 100.0 0.7 1.5 34.4 0.5 4.5 20-30

2.5 0.2 3.2 2.4 10.0 100.0 0.7 0.9 31.6 0.1 -30-40

1.5 0.6 1.0 1.2 3.2 100.0 0.8 0.3 23.1 1.2 1.8 30-40

p-Benzoquinone (1,4-benzoquinone) Salicylaldehyde (2-hydroxybenzaldehyde) Ethyl decanoate Methyl dodecanoate Unknown/' Ethyl dodecanoate Dodecanoic acid Methyl tetradecanoate Ethyl tetradecanoate Tetradecanoic acid Ethyl hexadecanoate Approx. quantity' of lipids gland (#g)

"Calculated as percent of largest peak (ethyl dodecanoate). J'Spectrum resembling a dimethyl naphthol. 'Calculated from GC integrated peak area of ethyl stearate standard. often slow, as t h o u g h considerable force was required to draw the sticky a n t e n n a t h r o u g h the strigilis. It is not clear that water elicited any cleaning movements. Chemistry o f Mandibular Gland Products. The secretion is red in dissected gland sacs; when exposed it is yellowish, gradually browning, presumably as a result of an u n k n o w n oxidation. Table 2 lists the lipid c o m p o n e n t s in the glands of three species of Exoneura. DISCUSSION Evidence is strong that the mandibular gland products in Exoneura serve for defense against ants and perhaps other natural enemies. The mixture of materials in these glands is similar in three sympatric species that are not particularly closely related to one another. P r o b a b l y the most recent comm o n ancestor of these species possessed the defensive mixture, which has been retained with little modification in the descendant species. The mixture may, of course, have other functions also. The m a n d i b u l a r gland p r o d u c t of Exoneura contains a quinone as well as salicylaldehyde and various ethyl and methyl esters of aliphatic acids. Quinones have not previously been recorded as exocrine products of bees, a l t h o u g h they are well k n o w n in defense secretions of other a r t h r o p o d s

1530

CANE AND MICHENER

(Eisner and Mainwald, 1966). It is noteworthy that the mandibular glands do not contain the mevalonic derivatives (e.g., citral, geranyl acetate, etc.) that are present in these glands of Ceratina (including Pithitis) (Hefetz et al., 1979; Wheeler et al., 1977). The differences in chemistry of gland products between the allodapine bee Exoneura and Ceratina support the suggestion that the allopadines might well be placed in a separate tribe, Allodapini, rather than within the Ceratinini as has been the recent custom. However, among all bees so far examined, Pith#is and Exoneura are the only ones producing salicylaldehyde in addition to some of the ethyl esters. If production of these compounds is synapomorphic, Pithitis and Exoneura may have a common ancestor not shared by Ceratina, a conclusion not supported by known morphological features. The large amouint of ethyl dodecanoate is interesting; traces of the same compound are found in mandibular glands of males of an unrelated anthophorid, Centris adani (Vinson et al., 1982). Of course the mandibular gland product is not the only defense by Exoneura at nest entrances. The bee at the entrance will bite at a disturbing object (ant or fiber held by an observer); then if the disturbance continues she will turn and firmly block the constricted nest entrance with the flattened dorsoapical part of the abdomen. Thus the construction of the nest entrance, as well as the behavior and secretions of the guard, contribute to nest defense. At least some guards of nests set up in Brisbane and heavily persecuted by Pheidole megacephala (?) ants seemed to have drops of brownish fluid that looked like the mandibular gland product around the edges of the abdominal area used to plug the nest entrance. Thus it is likely that Exoneura, like Pithitis, rubs the mandibular gland secretion over its body when disturbed (Hefetz et al., 1979). North American ants were used to test the defense secretion of an Australian bee for practical reasons. However, the ant species Pheidole rnegacephala, noted by C.D.M. attacking Exoneura nests in Brisbane, is an introduced species there. Henry Hacker (personal communication, 1957, to C.D.M.) remarked on the great reduction in native bee populations (especially Exoneura) in southeastern Queensland, compared to his experience early in the century. Perhaps the introduction of Pheidole is a factor in this reduction. Even when the guards of Exoneura are able to keep ants out of the nest for many days, the colony is likely to succumb because the Pheidole are persistent and, whenever a bee is entering or leaving, an ant is likely to get in. Presumably defense against native ants is more effective. Acknowledgments--We wish to thank Mr. Michael Schwarz of Monash University who found and collected the Exoneura nests and kindly gave them to us. Drs. Robert W. Taylor, Ian Naumann, and M a x J. Whitten of the Commonwealth Scientific and Industrial Research Organization, Canberra, arranged for C.D.M.'s trip tO Australia in 1982. J.H.C.'s part in this study (primarily the chemical-analyses) was supported by grants from the Danforth Foundation, the Hungerford Memorial Fund (University of Kansas), and the

MANDIBULAR GLAND OF E x o n e u r a

1531

National Science Foundation (Dissertation Improvement Grant DEB 80-23091). Assistance was also provided from NSF grants DEB 77-23035 and BNS 82-00651 (C.D. Michener, principal investigator). The GC-MS was purchased under USPHS grants GM-27430 and RR-5606; use of this intrument was possible thanks to the interest and courtesy of Dr. Robert G. Carlson of the Department of Chemistry, University of Kansas.

REFERENCES AMERINE, M.A., and ROESSLER, E.B. 1976. Wines: Their Sensory Evaluation. W.H. Freeman and Co., San Francisco. 230 pp. CANE, J.H. 1982. Evolution of the lipid exocrine secretions of bees. PhD thesis, University of Kansas, iv + 141 pp + Appendices I-4. EISNER, T., and MEINWALD, J. 1966. Defensive secretions of arthropods. Science 153:13411350. HEATH, H.B., and PHaRM, B. 1978. Flavor Technology: Profiles, Products and Applications. Avi Publ. Co., Westport, Connecticut. 542 pp. HEFETZ, A., BATRA,S.W.T., and BLUM, M.S. 1979. Chemistry of the mandibular gland secretion of the Indian bee Pithitis smaragdula. J. Chem. Ecol. 5:753-758. MICHENER, C.D. 1965. The life cycle and social organization of bees of the genus Exoneura and their parasite, lnquilina. Univ. Kans. Sci. Bull. 46:317-358. VINSON, S.B., WILLIAMS,H.J., FRANKIE,G.W., BLUM, M.S., and COVILLE, R.E. 1982. Mandibular glands of male Centris adani (Hymenoptera: Anthophoridae): Their morphology, chemical constituents, and function in scent marking and territorial behavior. J. Chem. Ecol. 8:319-328. WHEELER, J.W., BLUM, M.S., DALY, H.V., KISLOW, C.J., and BRAND,J.M. 1977. Chemistry of mandibular gland secretions of small carpenter bees (Ceratina spp.). Ann. Entomol. Soc. A m . 70:635-636.