Journal of Chemical Ecology, Vol. 22, No. 5, 1996

CHEMISTRY OF METAPLEURAL GLAND SECRETIONS OF THREE ATTINE ANTS, Atta sexdens rubropilosa, Atta cephalotes, AND Acromyrmex octospinosus (HYMENOPTERA: FORMICIDAE)

RUTH

R. D O N A S C I M E N T O , I E R I C S C H O E T E R S , 2

E. D A V I D

M O R G A N , l'* J O H A N

B I L L E N , 2 and

D A V I D J. S T R A D L I N G 3

IDepartment of Chemisto,. Keele Universi O' Staffordshire ST5 5BG, England "-Zoological Institute, University of Leuven Naamsestraat 59, B-3000, Leuven, Belgium 3Department of Biology, Universi D, of Exeter £~'eter EX4 4PS, England (Received May 15, 1995; accepted January 5, 1996)

Abstract--The chemical composition of the secretions of the metapleural glands of workers and soldiers of two Atta species, Atta sexdens rubropilosa and A. cephalotes, and workers of Acromyrmex octospinosus, has been studied. As indicated by infrared spectrometry and confirmed by the ninhydrin test. the secretions contain chiefly proteins. Of the volatile acidic portion, which is present as ionized salts, phenytacetic acid is the major component in workers and soldiers of A. s. rubropilosa and A. cephalotes. Both Atta species also contain 3-hydroxydecanoic acid and its homolog as minor components together with indoleacetic acid. While there are qualitative similarities in the acidic composition in the secretions of A. s. rubropitosa and A. cephalores, they differ quantitatively, The secretion of Acromyrmex octospinosus contains 3-hydroxydecanoic and indoleacetic acids, but lacks phenylacetic acid. The bactericidal and fungicidal actions of the three major substances have been confirmed, Key Words--Atta ,~'xdens rubropilosa, Atta cephalotes, Acrotn~,rmex octospinosus, metapleural glands, phenylacetic acid, Formicidae.

*To whom correspondence should be addressed.

987 0098-n331/Y(~/05(~l-09875()9.5(}/0 c , 1 ~ 6 Plenum Publishing C~rp~)nlliun

988

DO N A S C I M E N T O ET AL.

INTRODUCTION

The metapleural glands of ants are located in the metathorax, where they open on either side through a permanently open duct adjacent to the hindleg coxae. The duct is intemally connected with a large sclerotized bulla or reservoir, into which open the numerous secretory cells by means of their accompanying duct cells (Hrtldobler and EngeI-Siegel, 1984; Schoeters and Billen, 1993). Some functions, such as production of nestmate recognition substances, territorialmarking pheromones, protection against microorganisms (Maschwitz et al., 1970), and pollen germination inhibition (Sanderson and Wright, 1989) have been attributed to the secretion from these glands. Work carried out some 20 years ago on the secretion from two myrmicine ants, using large amounts of insect material, identified phenylacetic acid (1) and indoleacetic acid (3) as components of the secretion of Atta sexdens and A. s. rubropilosa (Maschwitz et al., 1970; Schildknecht and Koob, 1970). Schildknecht (1976), using the metathoracic extracts of 40 ants, later identified three further components in the secretion of A. s. rubropilosa, namely 3-hydroxyhexanoic, 3-hydroxyoctanoic and 3-hydroxydecanoic acids. No further work was recorded on metapleural glands until recently, when three further reports appeared. Attygalle et al. (1989) showed that the myrmicine species Crernatogaster deforrnis has hypertrophied metapleural glands that contain 3-propyl and 3-pentylphenol, 5-propyl- and 5-pentylresorcinol and a related coumarin. Beattie et al. (1994) extended the claim for antibacterial action of the metapleural secretion to the Australian ant Myrmecia gulosa. Knapp et al. (1994) at the same conference, however, drew attention to the antifungal activity of the mandibular gland secretion of attine ants, while Jaff6 et al. (1994) described the metapleural secretion of a large number of ant species as being composed largely of fatty acids. In the leaf-cutting species Atta taevigata, Acromyrmex subterraneus, and A. landolti, the major component was palmitic acid. They reviewed the claim for antimicrobial activity (Jaff6 et al., 1994). Independently, we have carried out studies to confirm the work of Schildknecht (1976), which up to this time had not been reexamined in the intervening two decades. Our objective was to work only with pure metapleural secretion and to avoid any contamination with other tissue or secretions. We have shown that the metapleural secretions of three attine species contain acidic compounds, as described by Schildknecht (1976), and the composition varies with species. The secretion also contains proteins. Microbiological tests confirm the activity of the secretion and synthetic substances against some bacteria and fungi,

METAPLEURAL GLAND SECRETIONS OF ANTS

989

METHODS AND MATERIALS

Colonies of Atta sexdens rubropilosa Forel, A. cephalotes, (L.) and Acromyrmex octospinosus (Reich.), from Vicosa and other locations in Brazil and Trinidad were maintained in laboratories at Leuven, W/irzburg, or Exeter. Secretions from the metapleural glands of workers and soldiers of the two Atta species and workers of A. octospinosus were collected in fine glass capillaries (50-70 ~m diameter) as described below. These capillaries were placed in large soft glass capillaries (1.8 +__ 0.2 mm x 20 + 5 mm), which had been first sealed at one end (Morgan, 1990). The larger capillaries were then completely sealed and transported to Keele for analysis. To collect secretion, a system was devised that required a 10-~1 syringe mounted in an expanded polystyrene block. The 50 to 70-p~m capillary was drawn from a larger capillary in a flame, and the larger end was attached to the microtiter syringe needle with a silicone tubing connection. The capillary was held steady in another small polystyrene block. The wall of the bulla of the metapleural gland was pierced with a fine insect-mounting pin, the ant was then held so the fine tip of the capillary are inserted into the hole in the bulla (Figure 1) and the secretion withdrawn into the capillary by gently raising the plunger of the syringe. With practice, this could be done routinely for a number of ants, collecting secretion from the paired metapleural glands in one capillary. The infrared spectrum of the secretion of A. sexdens rubropilosa major workers was recorded with a FTIR Brucker microscope attached to a Brucker IFS 48 Fourier Transform infrared spectrometer. For the ninhydrin test, the section from one metapleural gland of each species was placed on filter paper and sprayed with 0.1% ninhydrin reagent (Merck) and heated to 110°C. Approximately equal quantities of hemolymph from each species were tested similarly. The chemical analyses of these secretions were carried out by gas chromatography-mass spectrometry (GC-MS) with a solid-sampling technique (Morgan and Wadhams (1972). Preliminary analyses were made by directly crushing the sealed capillaries in the heated injection zone of the gas chromatograph and separating the components on a polar phase capillary column (polyethylene glycol, 12 m x 0.20 mm ID x 0.25-#m film thickness). This method did not give very satisfactory results for the very small quantities of acidic secretion. We had to modify the technique by opening each capillary at the tip, converting the acids to their methyl esters by adding a solution of diazomethane in ether (50 #1), and blowing off the ether with a stream of nitrogen, then resealing the capillary and using the solid-sampling method. These methyl esters were then analyzed on a column with a nonpolar phase (dimethylsiloxane, 12 m x 0.22 mm ID × 0.25-~m film thickness). The oven was programmed from 30°C at

990

o o NASCIMENTO ET AL.

Flo. 1. Method for removing metapleural gland secretion from an ant. After perforating the reservoir cuticle with an insect pin, the capillary was inserted into the opening. The ant was held with fine forceps. Arrow indicates secretion transport inside the capillary.

8°C/min to 250°C and helium was used as the carrier gas at 1 ml/min. The split vent was closed before crashing the sample and reopened 30 sec later. The chemical substances separated were identified by interpretation of the mass spectra of their methyl esters and confirmed by comparison of retention time and mass spectrum with synthetic standards. The analyses were carried on at least two colonies of each species, using at least 10 individual specimens for each colony caste. The methyl esters of phenylacetic and indoleacetic acids were prepared by esterification of their corresponding acids with diazomethane, and the methyl esters of the 3-hydroxy acids were obtained via a Reformatsky reaction. These compounds were used as external standards for quantification of the components, and for microbial assays. Preparation of Methyl 3-Hydroxy Esters. Methyl 3-hydroxy esters were prepared from a mixture of methyl bromoacetate (2 ml, 20 mmol, Fluka) and either hexanal, octanal, or decanal (20 mmol, Aldrich) dissolved in a mixture of dry toluene and ether (4: 1). The solution was added dropwise to zinc powder

METAPLEURAL GLAND SECRETIONS OF ANTS

991

(2 g, 20 mmol) and the mixture was refluxed for 1.5 hr, giving the zinc alkoxide of the hydroxy-ester. This was treated with a solution of sulfuric acid (10%), followed by aqueous sodium carbonate (10%). The product was extracted with diethyl ether (50 ml) and the solvent layer was dried (CaSO4), filtered, and the solvent removed by rotatory evaporation to yield the methyl esters of 3-hydroxyhexanoic, -octanoic, -decanoic and -dodecanoic acids, respectively. The yield in each reaction was approximately 60%. Microbiological Tests. To test tot the inhibition of spore germination by metapleural gland secretion of A. s. rubropitosa the tbllowing fungi were used: Botrytis cinerea, Pyricularia or3,zae, Trichoderma viride, and Venturia inaequalis. The spores were stored in a refrigerator and subjected to heat-shock immediately before use to induce germination. Spore suspension (10 p-l) was transferred into sterile potato dextrose bouillon (2.4%) with glucose (1%) (10 ml, diluted 1:2) containing tetracycline (final concentration 0.05%). An aliquot of this suspension (80 p.l) was placed in a microliter plate either with (experimental) or without (control) metapleural gland secretion ( 1-2 p-t) and maintained at 25°C for 33 hr. The samples were observed by light scattering in a spectrophotometer at 595 nm. For antibacterial tests, the following species were used: Bacillus cereus (6923, numbers in the Difco catalogue), Bacillus .sphaericus (7134), Pseudomonasfluorescens (1794), and Pseudomonas putida (2257). The bacterial were grown on Muelder-Hinton agar (Difco). To inoculated Petri dishes were added either metapleural gland secretion (1 p-l) or an aqueous suspension of the synthetic acids in distilled water (1 : 1) (1 p-l).

RESULTS The milky white secretion from the metapleural glands can be collected in a fine glass capillary as reported by Schildknecht (1976), but can be collected more easily by first piercing the bulla (Figure 1) and then inserting the capillary into the orifice. We first examined a droplet of the secretion of A. s. rubropilosa with the aid of a microscope and a Fourier transform infrared spectrometer. The spectrum showed very broad OH and possibly NH absorption with a sharp maximum at 3287 cm-~ and a secondary maximum at 3086 cm -~. There was no carboxylic acid C ~ O absorption near 1700 cm ~ . There were two strong bands at 1655 cm ~ and 1626 cm -~, which we attributed to amide I and II, and a broad absorption at 1549 cm-~, which can be attributed to ionized carboxylate ( C O 0 - ) (Figure 2). We concluded that the spectrum showed chiefly the characteristics of a concentrated peptide or protein solution. This was confirmed for each species using the ninhydrin test for protein, which gave clear positive tests.

992

DO NASCIMENTO

,,

~

i

i

ET AL.

i

I

~o

4000

3500

30~

2500 Wavenumber

2000

1500

1000

c m -I

FIG. 2. Infrared spectrum of the total metapleural gland secretion obtained from a major worker of Atta sexdens rubropilosa.

For comparison, the same quantity of hemolymph from an ant gave a weaker reaction. Phenylacetic acid (Figure 3, 1) was identified in the secretion of A. s. rubropilosa on a polar phase capillary gas chromatography column. However, converting the free acids into their corresponding methyl esters with diazomethane showed the presence of other acids in the secretions (Figure 4). The second most important constituent was 3-hydroxydecanoic acid. The exact composition varied with the individual, as indicated by the large standard deviations in Table 1. The soldiers showed a greater proportion of phenylacetic acid than the workers, but no more secretion. There were no differences found between colonies. Schildknecht and Koob (1970) found on average 1.4/.tg of phenylacetic acid per insect. We found a mean of 0.54 Izg per major worker, and 0,82 ~g per soldier. The difference may suggest incomplete removal of secretion by our method. In A. cephalotes, phenylacetic acid and indole acetic acid were the major substances (Figure 5). Again there were differences in composition between soldiers and workers, the soldiers having more secretion than we found in the workers. The secretion of A. cephalotes contained traces of the odorous compound skatole (4), which was not detected in A. s. rubropilosa, while the 3-hydroxy acids were undetected in workers.

METAPLEURALGLANDSECRETIONSOF ANTS

~

o

993

~

O

OH

OH

[1]

\

\ H

H

[3]

[4]

F~G. 3. Structures of the compounds found in the metapleural glands: 1: phenylacetic acid; 2: 3-hydroxydecanoic acid; 3: indoteacetic acid; 4: skatole.

A

r

B

3

a

b

~ ~c

10

2

d

15

~ e

2~

~5

FIG. 4. Total ion chromatogram of the volatile acids (as their methyl esters) from the metapleural glands of Atta sexdens rubropilosa. (A) From a major worker: (B) from a soldier caste. Peaks a-e are impurities introduced with the diazomethane solution.

994

DO NASCIMENTO ET A L

TABLE [. CHEMICAL COMPOSITION OF VOLATILES IN METAPLEURAL GLAND SECRETION OF Atta

Compound number in Figure 4

sexdens rubropilosa

12)

WORKERS AND SOLDIERS (N =

Mean proportion (% _+ SD) Conlpound"

I 2 3 4

Phenylacctic acid 3-Hydroxyoctanoic acid 3 - H y d r o x y d e c a m ~acid ic 3-Hydroxydodecanoic acid lndoleacetic acid Mean total amount (#g)

Workers

Soldiers

57.3 _+ 24,5 t 35.9 + 24.9 6.85 ± 5.6 t 0.95 ± 12

86,5 + 13.4 12]) ± 12.6 1,5 ± 3.2 t 0.95 ± 1.4

"The compounds were identified as their methyl esters after derivatizalion with diazomethane.

A

c,~

a

5

I~

15

Time

[min.

20

e

g

B

f

25

h

3~D

)

FIG, 5. Total ion chromatogram obtained after treatment with diazomethane of the volatile part of the metapleural gland secretion of Atta cephalotes. (A) Major workers, (B) soldier, For numbered peaks, see Table 2. Peaks a - f are fatty acid esters ranging from Ct2 to C~8 that were found to be contaminants in the diazomethane solution and peaks g and h are hexadecane and octadecane, respectively.

995

METAPLEURAL GLAND SECRETIONS OF ANTS

The metapleural gland secretions of A. octospinosus were characterized by the presence of 3-hydroxydecanoic acid (2) and indoleacetic acid (3) and by the absence of phenylacetic acid (Figure 6). Again no colony differences were found. In tests on the phytopathogens, Botrytis cinerea showed germination and hyphal growth in both the experimental and control conditions. Pyricularia o ~ z a e showed abundant mycelial growth in both control and experimental conditions, while Trichoderma viride showed strong mycelial growth in the control, but no germination in the experimental conditions. Venturia inaequalis showed only slight reduction of spore germination in the presence of the metapleural gland secretion. For antibacterial tests, the secretion of minor, medium, and major workers of A. s. rubropilosa was placed on an agar plate together with phenylacetic acid, 3-hydroxydecanoic acid, and indoleacetic acid. The metapleural gland secretion showed strong growth inhibition against Bacillus sphaericus and Pseudomonas putida and a weaker effect against B. cereus and P. fluorescens (Figure 7). It showed little effect against E. coti and S. aureus. Phenylacetic acid in the pure state was inhibitory towards all six species, and 3-hydroxydecanoic acid was inhibitory towards all but S. aureus. Indoleacetic acid, which tended to diffuse into the agar and gave larger, more diffuse spots, was strongly inhibitory

IE~ I1nle

13 mtrl.

30

=[4

)

Fie. 6. Ion chromatogram of m/z 103 for the metapleural gland secretion of an Acromyrrnex octospinosus worker, after treatment with diazomethane. Peak a, 3-hydroxydecanoic acid; b, indoleacetic acid (as their methyl esters). The amount of secretion was very small and the peaks did not show up well in a total ion chromatogmm.

996

DO NASCIMENTO ET AL.

,,~

[3

FIG, 7, Results of inhibition of hacterial growth, An emulsion of phenylacetic acid and distilled water, 1:1 (I); 1 t~l metapleural gland secretion ofA. s. rubropilosa major (2); medium (3) and minor (4) workers were placed on the agar plates (MHZ) together with indoleacetic acid and water 1 : 1 (5); and 3-hydroxydecanoic acid and water 1 : 1 (6). (A) Escherichia coli; (B) Staphylococcus aureus, (C) Bacillus cereus; (D) Bacillus sphaericus, (E) Psuedomonas fluorescens, (F) Pseudomonas putida.

METAPLEURALGLANDSECRETIONSOF ANTS

997

toward S. aureus and B. sphaericus, weakly inhibitory towards P. putida and E. coli, and showed no effect against B. cereus and P. fluorescens.

DISCUSSION

The metapleural gland secretion of the three species of attine ant examined here is a milky white substance, and is thus an emulsion, not a solution. The secretion is readily seen through the thin cuticle of minor and medium workers of A. s. rubropilosa, so that individuals with full reservoirs can be identified, but it is not visible through the more opaque cuticle of major workers and soldiers. The infrared spectrum of the secretion of A. s. rubropilosa was not at all like that of a mixture of carboxylic acids. There was no evidence of carboxylic acid C = O absorption of phenylacetic acid (at 1697 c m - ~) or of the 3-hydroxydecanoic acid (at 1730 cm-~). It showed strong OH absorption, typical of associated hydroxy groups, which could be a concentrated aqueous solution. It showed a sharp maximum at 3287 c m - ~, typical of secondary amides, and two very strong bands at 1655 cm -~ and 1626 cm -~, which could be attributed to amide 1 and II bands of peptides or proteins. The broad band at 1549 cm -~ can be correlated with ionized carboxyl. We concluded that the secretion was an aqueous emulsion of protein and the ionized form of the carboxylic acids. The secretion gave a strong, positive ninhydrin test for protein, stronger than that of hemolymph. Schoeters and Billen (1993) have already described the glandular endoplasmic reticulum in the secretory cells of the metapleural gland, which suggests considerable protein-synthesizing ability. The remainder of the study concentrated on the volatile part of the secretion~ Carboxylic acids are not ideal compounds for gas chromatography because they are very polar and small quantities can be lost through adsorption on surlaces in the instrument. Initially, we showed that phenylacetic acid was present in the secretion of A. s. rubropilosa. It gave a clear peak when the whole secretion was chromatographed on a polar column, but no other substances were visible. Thereafter, we converted all carboxylic acids to their less polar methyl esters by treating the secretion on a microscale with diazomethane solution. Minor peaks were then visible, which from their mass spectra and coincidence of retention times were identified as the compounds listed in Tables I and 2. While Schildknecht and Koob (1970), Maschwitz et al. (1970), and Schildknecht (1976) apparently were unaware of the presence of protein in the secretion, we confirmed the report of Schildknecht (1976) that the major volatile and acidic substances in A. s. rubropilosa were phenylacetic acid followed by 3-hydroxydecanoic acid, which we synthesized for comparison. However, where Schildknecht found the C 6, C~, and C~o/3-hydroxyacids, we found the Cs, C~o, and

DO NASC1MENTOET AL.

998

TABLE 2. CHEMICAL COMPOSITION OF VOLATILESIN METAPLEURAL GLAND SECRETION OF Atta cephalotes WORKERS AND SOLDIERS (N = 15) Compound numbers in Figure 5 1

Mean proportion (% ± SD) Compound"

Phenylacetic acid Skatole 2 3-Hydroxydecanoic acid 3 4-Hydroxyphenylacetic acid 4 3-Hydroxydodecanoic acid Unknown 3-hydmxy acid I 5 3-Hydroxytetradecanoic acid 6 3-Hydroxyhexadecanoic acid Unknown 3-hydroxyacid 1I lndoleacetic acid Mean total amount Lag)

Workers

Soldiers

99~3 _+ 1.7 0.t + 0.4

63,2 + 26.1 0.9 + 1.6

t

0.6 +_ 1.7 0,01 + 0.03

6. I 2,7 4.1 3,8 0.1 18.4 0.2

+ + ± ± + + ±

7. I 4.4 4,7 7. I 0~3 9.9 0,2

"The compounds were identified as their methyl esters after derivatization with diazomethane.

CI2 ~-hydroxy acids. The mean c o m p o s i t i o n of the large workers' and soldiers' secretions were different, but because o f the wide individual variations, it is not clear if this difference is important. Microbiological tests s h o w e d that the antibacterial action o f the secretion can be ascribed to the volatile acid portion. H o w e v e r , as pointed out by H611dobler and Engel-Siegel (1984), the presence of an antiseptic effect in the secretion does not necessarily m e a n that the primary function o f the gland is antisepsis. The work o f Van den Broeck (1993) and Knapp (1994) s h o w e d that there are powerful antiseptic effects in the m a n d i b u l a r glands o f attine ants. According to Diehl-Fleig and V a l i m - L a b r e s (1993), isolation of Aspergillus spp., Mucor spp., and Penicillium spp. from Atta sexdens piriventris and Acromyrmex heyeri workers shows that m a n d i b u l a r and metapleural gland secretions are not fungicidal to at least some species. T h e s e results are also supported by the fact that one can find fungus ants with Beauveria bassiana ( e n t o m o p a t h o g e n ) (DiehI-Fleig et al., 1992) and Metarhizium anisopliae. T h u s , it seems that the ants cannot avoid the occurrence o f alien fungi. In preliminary tests, Van den Broeck (1993) found that the Atta garden fungus (Attamyces sp,) did not inhibit g e r m i n a t i o n of Trichoderrna hamatum spores. T. hamatum and T. viride are mycotoxic fungi. We, therefore, concluded there was no natural protection against pathogenic fungi in the fungus garden, and included T. viride in o u r tests. Strong hyphal growth o f T. viride was

METAPLEURAL GLAND SECRETIONS OF ANTS

999

observed in the control, but no growth at all occurred in the experiment. This indicated that the metapleural substances could provide protection for Attamyces from pathogens. Powell and Stradling (1986) have already shown that the metapleural gland secretion had only a minor influence on the growth of Attamyces. The metapleural gland secretion of A. cephalotes workers contains phenylacetic acid, and the secretions from the soldiers contain four/3-hydroxy acids (C10, CI2, CI4, and Cir). In addition, the soldiers contained two unidentified compounds, which after conversion to methylesters, had mass spectra that showed a base peak of m/z 103 and fragmentation patterns similar to unsaturated hydrocarbons. This is characteristic of methyl ~-hydroxyesters. We therefore concluded that these are from unsaturated/3-hydroxy acids. Skatole, not detected in A. s. rubropilosa, was also present in A. cephalotes and is probably responsible for the distinct odor of the secretion. The secretion of Acromyrmex octospinosus workers was more difficult to collect and different from that of the two Atta species. By searching the mass spectra, 3-hydroxydecanoic acid and indoleacetic acid were identified and cornfirmed by selective ion monitoring. A. octospinosus secretion did not contain phenylacetic acid. It is noteworthy that Schildknecht and Koob (t970) also recorded that Acromyrmex subterraneus lacked phenylacetic acid. Indoleacetic acid (a heteroauxin) is well known as a plant growth hormone. It is clear that these leaf-cutting ants possess a metapleural gland secretion that contains, in part, volatile acids that have antibacterial action. They could protect the ant fungus garden from other fungi, which, like Trichoderma viride, produce mycotoxins. The earlier work of Maschwitz et al. (1970) and Schildknecht (1976) are confirmed to that extent. However, the gas chromatography, powerful as it is, is not the solution to all analytical problems, and it cannot detect peptides and proteins. To obtain a solution in an organic solvent and to analyze the volatile components of that solution may be to overlook other substances. Unless the secretion is homogeneous and the components identified can account for the full mass of the secretion, such results must be treated with caution. The metapleural gland secretions of these ants require further investigation before we know their full composition and function. Acknowledgments--We thank B. Hrlldobler for providing some of the ant material.

REFERENCES ATTYGALLE, A. B., SIEGEL, B., VOSTOWSKY, O,, BESTMANN, H. J., and MASCHWITZ, U. 1989. Chemical composition and function of metapleural gland secretion of tile ant, Crematogaster deJbnnis Smith (Hymenoptera: Myrmicinea), J. Chem. Ecol. 15:317-328. BEATT1E, A. J., MACKINTOSH, J. A., CRUSE, A., and VEAL, D. A, 1994. Defense against micro-

1000

DO NASCtMENTO ET AL,

organisms in the ant Myrmecia gulosa and the termite Nasutitermes exitiosus, p. 108, in A, Lenoir, G. Amotd, and M. Lepage eds.). Les Insectes Sociaux. Universit6 Paris Nord, Paris. DIEHL-FLEIG,E., and VALIM-LABRES,M. E. 1993. Fungi isolated from leaf-cutting ants Atta sexdens piriventris and Acromyrmex heyeri (Hymenoptera: Formicidae): Mucor spp. effects on Beauveria bassiana entomopathogen Cienc. Cult. 45:142-144. DtEHL-FLEIG, E., DA SmVA, M. E., VALtM-LABRES, M. E., and SPECHT, A. (1992). Ocorr~ncia natural de Beauveria bassiana (Bals.L Vuill. no Rio Grande do Sul. Acta Biol. Leopoldensia 14:99-104. HOLLDOBLER, B., and ENGEL-SIEGEL,H. 1984. On the metapleurat gland of ants. Psyche 91:201224, JAFFI~.,K., VASQUEZ,C., BRANDWIJK,t . , and CABRERA, A. 1994. Metapteural gland secretions in ants, p. 217, in A. Lenoir, G. Arnold, and M. Lepage (eds.). Les Insectes Sociaux. Universit~ Paris Nord, Paris. KNAPP, J. J., JACKSON,C. W., HOWSE,P. E., and VILELA.E. F, 1994. Mandibular gland secretions of leaf-cutting ants: Role in defense against alien fungi, p. 109, in A. Lenoir, G. Arnold and M. Lepage (eds.). Les Insectes Sociaux. Universit6 Paris Nord, Paris. MASCHWlTZ, U,, KOOB, U , and SCrt~LDKNECHT,H. 1970. Ein Beitrag zur Funktion der Metathorakaldrfise der Ameisen. J, Insect Physiol. 16:387-404. MORGAN, E. D. 1990. Preparation of small-scale samples from insects for chromatography, Anal, Chim. Acta 236:227-235. MORGAN, E. D., and WADHAMS, L. J. 1972. Gas chromatography of volatile compounds in small samples of biological materials. J. Chromatogr. Sci. 10:528-529. POWELL, R. J., and STRADLING,D. J. 1986. Factors affecting the growth ofAttamyces bromatificus, a symbiont of attine ants. Trans. Br. Mycol. Soc. 87:203-213. SANDFRSON,T., and WRIGHT. P. J. 1989. Inhibition of pollen germination by ant secretions. Actes Colloq. his. Soc. 5:25-30. SCHILDKNEeHT,H. 1976. Chemical ecology--a chapter of modem natural products chemistry. Angew. Chem. Int. Ed. Engl. 15:214-222. SCmLDKNECHT,H., and KooB, U. 1970. Plant bioregutators in the metathoracic glands of myrmicine ants, Angew. Chem. Int. Ed. Engl. 9:173. SCHOETERS, E., and BILLEN, J. 1993. Anatomy and fine structure of the metapleural gland in Atta (Hymenoptera: Formicidae). Belg. J. ZooL 123:67-75. VAN DEN BROECK, P. 1993. Onderzoek Naar de lnhibitorische Werking van Klierextracten op Bacterien en Schimmels bij Enkele Mieresoorten. Licentiaatsverhandeling K.U. Leuven. pp. 79.