Journal of Chemical Ecology, Vol. 15, No. 6, 1989
REDUCTION IN DIET BREADTH RESULTS IN SEQUESTRATION OF PLANT CHEMICALS AND INCREASES EFFICACY OF CHEMICAL DEFENSE IN A GENERALIST GRASSHOPPER
CLIVE G . J O N E S , 1 D O U G L A S COMPTON,1,
5 PETER
W.
WHITMAN, 2 STEVE J.
J. S I L K , 3 a n d M U R R A Y
S. B L U M 4
1Institute of Ecosystem Studies The New York Botanical Garden, Mary Flagler Cary Arboretum Box AB, Millbrook, New York 12545 2Department of Biology Illinois State University Normal, Illinois 61671 3pheromone Research Group New Brunswick Research and Productivity Council P.O. Box 6000, Fredericton, New Brunswick, E3B 5141, Canada 4Department of Entomology University of Georgia Athens, Georgia 30602 (Received June 1, 1988; accepted August 30, 1988)
Abstract--The lubber grasshopper, Romalea guttata, is a generalist feeding on a broad diet of many herbaceous plant species and has a metathoracic defensive secretion normally containing phenolics and quinones synthesized by the insect. When insects were reared on a restricted diet of wild onion, they sequestered sulfur volatiles from the plant into their defensive secretions. These compounds were not detected by gas chromatography-mass spectroscopy in secretions of insects on an artificial diet or a natural, generalist diet of 26 plants that included wild onion as a component, nor were they present in secretions from field-collected insects. Defensive secretions of insects reared on wild onion were significantly more deterrent, by as much as an order of magnitude, to two species of ant predators than secretions from insects on either of the other two diets, despite a reduction in the concentration of autogenous defensive chemicals in secretions of insects on the onion 5 Present address: University of Washington School of Medicine, Health Sciences Center, A-320, Seattle, Washington 98195. 1811 00984)331/89/0600-1811506.00/0 9 1989 Plenum Publishing Corporation
1812
JONES ET AL.
diet. Sequestration of plant chemicals that increased defensive efficacy occurred when diet breadth was reduced. We suggest that this occurs because under conditions of specialization, plant secondarymetabolites are more likely to be ingested and bioaccumulated in sufficient concentrations to have biological activity against predators. What we define as casual bioaccumulation of bioactive plant chemicals followingdietary specialization may lead to evolution of sequestered defense syndromesin insects, and this process may not necessarily require specific adaptation to or coevolution with a toxic host plant. Key Words--Lubber grasshopper, Romalea guttata (microptera), Orthoptera, Acrididae, onion, generalist diet, artificial diet, sulfur volatiles, ant predation, casual bioaccumulation, specialization, evolution of defenses.
INTRODUCTION
Many herbivorous insect species use chemical defenses against predators (Duffey, 1976; Blum, 1981; Pasteels et al., 1983a). In some cases, autogenous defensive chemicals are made by the insect from dietary precursors. Alternatively, chemicals from the diet may be sequestered into the body tissues, fluids or glands with or without modification by the insect. Examples of sequestered defenses include cardenolides in the monarch butterfly (Brower, 1984), milkweed bug (Duffey and Scudder, 1972), and pyrogomorphid grasshoppers (Rothschild and Parsons, 1962); salicylaldehyde in some chrysomelid beetles (Pasteels et al., 1983b); and cardenolides and pyrollizidine alkaloids in some arctiid moths (Rothschild and Aplin, 1971). The ecological factors that initiate or constrain evolution of the sequestered defense syndrome are of considerable interest but poorly understood. It has been suggested that sequestered defenses initially evolved as a special insect adaptation for avoiding toxicity of specific chemicals in the host plant, by placing these chemicals out of harm's way in tissues or fluids. Storage of these compounds then subsequently conferred a selective advantage against predators (Brower and Brower, 1964; Rothschild, 1973). An alternative possibility is that the sequestered defense syndrome may arise as a consequence of what we term initial casual sequestration or bioaccumulation of ingested chemicals in body tissues (Duffey, 1976, 1980). The chemicals may not necessarily be toxic to the insect, and specific adaptation to a host plant may not be necessary. Bioaccumulated compounds might then deter predators if they were sequestered in sufficient concentrations. In this article we present evidence that two key characteristics of a sequestered defense syndrome--incorporation of plant chemicals into defensive secretions, and a concomitant increase in defensive efficacy against predators--can arise in an insect when there is no evidence for specific adaptation to a plant species, when the plants are not toxic to the insect, and in
DIET BREADTH AND C H E M I C A L DEFENSE IN A GRASSHOPPER
1813
a manner best described as casual sequestration. Furthermore, we show that diet breadth is the major ecological variable that determines whether or not sequestration of at least one class of plant chemicals occurs in sufficient amounts to have effects on predators. The lubber grasshopper, Romalea guttata (Houttuyn) [Orthoptera; Romaleidae = R. microptera (Beauvois)] (Kevan, 1980), is a large, flightless, generalist grasshopper with a local host range of over 80 species of herbaceous and woody plants. Many of these plants will be eaten within the lifetime of an individual insect (C.G. Jones, D.W. Whitman, M.S. Blum, unpublished data). Adult Romalea have an impressive defensive armory, including warning coloration, a pink wing-flashing display and sharp spines on the legs. Most notable is the odoriferous, frothy, glandular secretion produced with an audible hiss from modified metathoracic spiracles (Eisner et al., 1971). This secretion is deterrent to ants and other predators (Eisner et al., 1971; D.W. Whitman, C.G. Jones, M.S. Blum, unpublished data). The major components of the secretion of wild insects are various phenolics and quinones, including hydroquinone, catechol, p-benzoquinone, phenol, guaiacol, and 4-methoxybenzaldehyde (Eisner et al., 1971; Jones et al., 1986, 1987). These compounds are presumably autogenous because they are found in the secretions of insects reared on an artificial diet that does not contain any of these chemicals (Jones et al., t987). The autogenous defensive components are markedly affected by diet breadth. Restricting Romalea to a single host plant or artificial diet results in a reduction in the number and concentration of compounds, and thus alters the composition of the secretion, compared to lubbers reared on a generalist diet or collected in the field (Jones et al., 1987). Eisner et al. (1971) showed that Romalea collected from an area sprayed with the herbicide 2,4-D had 2,5-dichlorophenol in the defensive secretion. We theretbre wondered if Romalea was capable of sequestering natural plant products into the defensive secretion. In our studies on the effects of diet breadth on autogenous defenses (Jones et al., 1987), we noticed that the secretions of insects reared on a single host plant diet of onion appeared to have an onionlike odor. Here, we evaluated the possibility that sequestration of onion compounds might have occurred. We reared Romalea from first instar to adult on three different diets: (1) a restricted single plant species diet of wild onion, Allium canadense L., that we have observed is one of a number of favored food plants of Romalea (C.G. Jones, D.W. Whitman, M.S. Blum, unpublished data); (2)a "natural" generalist diet of plant species used by R. guttata in the field, which included wild onion as a component species; and (3) a restricted artificial diet known to lack any of the insects' autogenous phenolics and quinones and lacking any plant secondary metabolites. We compared the chemistry of secretions from insects reared on the different diets. We then bioassayed the secretions against a known ant predator of Romalea, the red imported fire ant, Solenopsis invicta
1814
JONES ET AL.
Buren (Formicidae: Myrmicinae), and a less aggressive ant species, Tapinoma melanocephalum F. (Formicidae: Dolichoderinae). Our findings are relevant to the ecology and evolution of chemical defense in insects.
METHODS AND MATERIALS
Rearing of Insects. Lubber grasshoppers were reared under the same conditions as used by Jones et al. (1987). The first of the three diets consisted of freshly cut leaves of Allium canadense in water. The second diet consisted of 26 species from 15 families of shrubs, herbs, forbs and grasses known to be eaten by Romalea in the field (C.G. Jones, D.W. Whitman, M.S. Blum, unpublished data). Plants were collected daily and the petioles or stems placed in water. The species list was as in Jones et al. (1987) and included wild onion, Allium canadense. The third diet was an artificial diet (Dadd, 1960) containing cellulose, sucrose, dextrin, mineral salts, cholesterol, linoleic acid, casein, peptone, egg albumin, ascorbic acid, and B vitamins and was fed as a water-moistened cake. Ten males and 10 females were also collected from the field for chemical analysis of their secretions. Secretion Collection. Secretions were collected from adults at the eighth day after the terminal molt, by gently squeezing the thorax while applying a 10-/xl microcapillary to the metathoracic spiracule orifices. A single milking removes all the secretion produced since the last molt. Molting insects discard all previous secretions. The volume collected (range 1-8 tzl) was recorded and the secretion sealed in the microcapillary and stored at - 10~ Secretions from insects reared in the same batch on the different diets were randomly allocated for: (1) separately pooled secretions of males and females for mass-spectral analysis, (2) secretions from individual females and pooled secretions of females for quantitative analysis of sequestered compounds, (3) secretions from individual females for quantitative analysis of autogenous compounds, and (4) pooled secretions of females for bioassays. In addition, separately pooled secretions of wild male and female insects were subjected to mass-spectral analysis. Qualitative Analysis of Secretions. Compounds were identified from splitless injection (10 psi) of undiluted secretion on a Finnigan 4021 gas chromatograph-mass spectrometer (GC-MS) using a 30-m x 0.3-mm-ID DB-5 fused silica bonded phase column (J&W), programmed at 50~ for 1.5 min, then 100-250~ at 15~ The GC-MS was operated in electron impact mode (70 eV, 1400 V, 10 -7 amp/V; scanning 45-350 amu, 1.5 sec/scan) with acquisition commencing at sample injection. Compounds were identified by using the molecular ions and characteristic mass fragments, by comparison with library spectra (r fit), and by GC-MS of authentic standards. Pooled secretions col-
DIET BREADTH AND C H E M I C A L DEFENSE IN A GRASSHOPPER
1815
lected from 10 individual males and females reared on each of the three diets and 10 males and females collected in the field were analyzed by GC-MS. Quantitative Analysis of Sequestered Compounds. Diethyl ether extracts of secretions were analyzed by subambient, capillary gas chromatography (GC) (Varian 6000), using a 30-m • 0.32-mm-ID fused silica bonded phase column (J&W, DB-5), helium carrier (2 ml/min) splitless injection (200~ purge at 0.7 min), and flame ionization detection (300~ 1 • 10 ~2 sensitivity). Initial column temperature was 20~ ramped to 40~ at 10~ then 40~ to 200~ and held at 200~ for 2 min. Secretions from eight individual females and the pooled secretions of three females reared on the onion diet were each extracted with 10/~1 redistilled diethyl ether and immediately analyzed by GC. The 2.5-tA aliquots were injected, and exact volumes of ether recovery and reproducibility were determined by addition of dl-alpha pinene to the ether as an internal standard, prior to extraction (325 t~mol/liter). We determined the concentration of eight sulfur-containing volatiles as the mean of three replicate injections: 1-propanethiol (minimum detectable quantity per microliter, 14 pg), methylthiirane (5 pg), methyldisulfide (21 pg), isopropylsulfide (11 pg), 2,5dimethylthiophene (30 pg), methylpropyldisulfide (13 pg), isopropyldisulfide (28 pg), propyldisulfide (25 pg). Retention times were accurate to _+1% ( P = 0.05); peak heights with tangent baseline correction were accurate to + 10% (P = 0.05), based on analysis of standards and replicate injections of pooled samples, and were measured on a Varian 402 integrator. Retention times and concentrations of compounds were determined by calibration against authentic standards and confirmed by using coinjection and independent injection. Standards were commercially available, except for methylpropyldisulfide, which was synthesized by the method of Carson and Wong (1959). Combined concentrations were determined from the sum of the concentrations of each compound. Quantitative Analysis of Autogenous Compounds. Secretions were analyzed for six phenolics and quinones (hydroquinone, p-benzoquinone, phenol, catechol, guiaicol, 4-methoxybenzaldehyde) using tertiary gradient reversephase high-performance liquid chromatography (HPLC) on a C1s bonded-phase silica column with a water-methanol-l% H3PO4 convex elution gradient, according to the method of Jones et al. (1986, 1987). Secretions from 10 individual females reared on each diet were diluted to 1% in 18 mohm deionized water, and triplicate 10-t~l aliquots were analyzed. Combined concentrations of phenolics and quinones were determined from the sum of the concentrations of each compound. Bioassay of Secretions. In preliminary binary choice experiments, we determined that ants preferred pure sucrose solutions to those containing wild Romalea defensive secretions (D.W. Whitman, C.G. Jones, M.S. Blum,
1816
JONESET AL.
unpublished data). We therefore decided that relative deterrency of the three different secretions was best determined by simultaneously comparing the secretions against one another, when they were made up in sucrose solution. Five laboratory colonies of Solenopsis invicta, varying in size from 1000 to 4000 individuals, were each presented with a 45 x 75-mm glass slide containing three separate 50-/H droplets of 4% sucrose. Each separate droplet contained 1 /A of Romalea metathoracic secretion pooled from individual females reared on either the natural, the onion, or the artificial diet. Ants foraged down a common trail to this comparative bait station. The common foraging trail p r e v e n t e d recruitment to one bait from affecting the total number of foragers approaching the bait station and reduced any biases due to the deterrency of a given bait. The number of ants feeding on each droplet was recorded every 4.5 min, up to the time that the first of the sucrose baits with secretion was completely consumed. Trials were repeated three times for each colony, with the order of the three droplets on the slide being changed for each trial. It was determined that the number of ants visiting the station initially increased, stabilized by 10 min, and remained more-or-less constant for over 30 min. Therefore the number of ants at each bait at 40 min was recorded for each colony and trial. These data were analyzed using ANOVA followed by multiple comparisons (least-significant-difference tests). Two laboratory colonies of Tapinoma melanocephalum (colony size no more than 3000 individuals) were tested using the above procedures. Ten percent sucrose solution containing secretions was used because this species of ant did not recruit sufficient numbers to 2% pure sucrose solution to permit adequate evaluation of deterrency. Trials were repeated twice. In addition, pairwise comparisons of artificial diet secretion (1/H) + 10% sucrose versus natural diet secretion + sucrose, and natural diet + sucrose versus onion diet + sucrose were also carried out using the glass slide technique. Three lab colonies were used with five repeated trials each. Again, the number of ants visiting the baits at 20 min was recorded. Data in these trials were analyzed as above.
RESULTS
Sequestration of Plant Compounds. The secretions of insects reared on the wild onion diet lacked the characteristic color and odor of secretions from insects on the natural diet and had a marked onionlike odor. GC-MS of the volatile fraction of pooled secretions from males and females showed the presence of a large number of sulfur-containing compounds (C.G. Jones and P.J. Silk, unpublished data). Eight compounds were unambiguously identified (isopropylsulfide, methylpropyldisulfide, isopropyldisulfide, methyldisulfide, propyldisulfide, propanethiol, methylthiirane, and 2,5-dimethylthiophene) and were
DIET BREADTH AND C H E M I C A L DEFENSE IN A GRASSHOPPER
1817
identical to those previously isolated and identified from onion (Carson and Wong, 1961; Brodnitz et al., 1969; Boelens et al., 1971). The total concentrations of these eight onion sulfur volatiles that were generally the most abundant in the secretions were quantified in the defensive secretions of eight adult female individuals reared on the onion diet. Mean concentration of all eight of these compounds combined was 0.5 ng//zl (Figure la). None of these sulfur volatiles, nor any other sulfur-containing compounds, were detected in GC-MS analyses of pooled secretions from male or female insects reared on the natural or artificial diet (pooled from 10 of each sex on each diet), or from secretions pooled from 10 individuals of each sex of field-collected insects occurring in an area where a diversity of plant species grow, including wild onion. In a separate rearing experiment, sulfur volatiles were detectable by odor and GC-MS in defensive secretions of insects within five days of transfer from artificial to onion diets. Auto~genous Compounds. Secretions from insects raised on either onion or the artificial diets contained significantly lower concentrations of all autogenous phenolics and quinones than secretions from insects on the generalist, natural diet. Combined phenolic and quinone concentrations were depressed fivefold on the artificial diet and 37-fold on the onion diet (Figure lb). The concentrations of autogenous compounds in secretions of insects on onion diets were also significantly lower than concentrations in artificial diet secretions (t tests, P < 0.0001). These new data agree with our previous findings for autogenous compounds (Jones et al., 1987). Efficacy of Defensive Secretions. To determine whether changes in the secretion chemistry affected biological activity, we bioassayed secretions against two ant species. The imported fire ant, Solenopsis invicta Buren, is a very aggressive predator of insects (Lofgren et al., 1975). This ant species did not evolve in the same habitat as Romalea guttata, but we have observed predation by Solenopsis on Romalea in the field (D.W. Whitman, C.G. Jones, M.S. Blum, unpublished data). Tapinoma melanocephalum F. is a less aggressive, omnivorous scavenger sympatric with Romalea in Florida (Creighton, 1950). We estimated the relative deterrency of the defensive secretions. With Solenopsis invicta, significantly more ants visited and fed at the artificial diet secretion bait compared to the natural diet secretion bait, and significantly more ants visited and consumed both of these baits compared to the onion diet secretion bait (Figure lc shows means + 1 SE). Overall effect of treatments was significant: ANOVA: F = 16.42, df = 44, P = 0.0001. All means in multiple comparisons were significantly different from each other, LSD t tests, cr = 0.05.) Approximately 63, 30, and 7 % of all foraging ants visited and fed at the artificial, natural, and onion diet secretion baits, respectively. The less aggressive ant, Tapinoma, showed even greater differences between treatments. In trials where all three secretion baits were presented
1818
JONES ET AL.
0.6-
0.5 {33 E tO
0.4
0.3
0 tO 0
0.2 0.1
Natural Diet
Artificial Diet
Onion Diet
b 7000" :~
6000-
C 5000= C .9 4000-
E aooo0 2000" O 1000ONatural Diet
Artificial Diet
Onion Diet
131211109876-
e~
E Z
Natural Diet
Artificial Diet
Onion Diet
D I E T B R E A D T H A ND C H E M I C A L D E F E N S E IN A G R A S S H O P P E R
18 t 9
together, the proportion o f ants visiting and feeding at the baits was approximately 70 (artificial diet), 29 (natural diet), and 1% (onion diet) (mean _+ 1 SE number o f ants at bait at 20 min for two colonies and two trials was: artificial 13.3 + 3.4; natural 5.0 + 1.6; onion 0. A N O V A : F = 9.55, df = 11, p = 0.006. All pairs o f means in multiple comparisons were significantly different from one another, LSD t tests, c~ = 0.05). In comparisons with Tapinoma, where pairs o f secretions were tested together, the number of ants at artificial diet secretion baits at 40 min (mean + 1 SE = 14.9 _+ 1.7) was significantly greater than the number at natural diet secretion baits (mean + 1 SE = 7.0 + 1.4; A N O V A : F = 8.43, df = 49, p = 0.007). Numbers at 40 min at natural diet secretion baits (mean +_ 1 SE = 45.1 _+ 4.4) were significantly greater than numbers at onion diet secretion baits (mean + 1 SE = 0.3 _+ 0.16; A N O V A : F = 35.49, d f - - 49, p = 0.0001).
DISCUSSION Diet ,clearly affects the quality and quantity o f the chemical components of the defensive secretion of Romalea and the efficacy of the defensive secretion against these predators. Sequestration of measurable quantities o f sulfur compounds occurred only when Romalea specialized on the onion diet. The natural diet also included wild onion as a component species, and yet secretions from insects reared on this diet did not contain any detectable sulfur-containing compounds, nor were these compounds found in field-collected insects that fed on a large number o f plant species. Since these compounds were not present in the artificial diet, their absence from secretions o f insects specializing on this diet was to be expected. In addition, specialization on either artificial or onion diets again caused a reduction in the concentrations o f autogenous phenolics and quinones, as has been previously reported (Jones et al., 1987). The response o f ants to baits containing these secretions is consistent with the chemical differences resulting from the presence of sequestered compounds and the depression of autogenous compounds. Sequestration o f sulfur corn.( FIG. t. (a) Mean combined concentrations of eight sulfur-containing volatiles (ng//~l secretion, +_ 1 SE) in defensive secretions of individual adult female R. gutmta (N = 8), reared on an onion diet. Pooled secretions of females (N = 10) on the other diets did not contain detectable levels of sulfur volatiles. (b) Mean combined concentrations of six autogenous phenolics and quinones (ng/~l secretion, + 1 SE) in defensive secretions of adult female R. guttata (N = 10), reared on three diets. (c) Mean number of Solenopsis invicta ants (_+ 1 SE, five colonies, three trials each) visiting and feeding at 20 min on sucrose baits containing defensive secretions of R. guttata females reared on three diets.
1820
JONES ET AL.
pounds in onion diet secretions is associated with an increase in defensive efficacy compared to natural diet secretions, despite depression of the concentrations of autogenous phenolics and quinones. Reduction of autogenous phenolic and quinone concentrations in artificial diet secretions is associated with a reduction in the efficacy of defense compared to natural diet secretions. The deterrent activity of components sequestered on the restricted onion diet appears to be much greater than the activity of the autogenous compounds. The combined sequestered sulfur compounds occurred at concentrations about 104 lower in onion diet secretions than combined concentrations of phenolics and quinones in the natural diet, and yet the onion diet secretion bait was visited and fed on by Solenopsis at about one quarter the rate of the natural diet secretion bait. Dietary specialization in R. guttata therefore led to rapid and simultaneous changes in two attributes important to its chemical defense: acquisition of compounds directly from the diet and a reduction in autogenous compounds. This led to a concomitant increase in defensive efficacy. Experimental specialization in Romalea produces results strikingly similar to many specialist insect herbivores where chemical defense against predators depends upon sequestration of host-plant chemicals (Brower, 1984; Rothschild, 1973; Roeske et al., 1976; Duffey, 1976, 1980; Blum, 1981). Why should specialization result in sequestration in Romalea ? We propose that the degree of sequestration of plant chemicals in this insect is primarily a function of the concentration of compounds ingested in the diet. When a generalist diet is consumed, compounds from an individual plant species may be too dilute to be detectable or biologically active against predators in comparison to the autogenous compounds. Conversely, specialization results in the ingestion of sufficiently high concentrations of compounds from one plant species such that these compounds may be sequestered in detectable levels that have biological activity. Interestingly, data from Eisner and coworkers (1971) may be relevant to our contention. They reported 2,5-dichlorophenol in the defensive secretion of R. guttata collected in herbicide-sprayed areas and suggested that this compound was derived from the herbicide 2,4-D. 2,5-Dichlorophenol was not present in secretions of insects from unsprayed areas. Although in this case Romalea was consuming a generalist diet, it may have been, in effect, specializing on a halogenated phenol diet because all of the plants were sprayed with herbicide. Our results may have important implications for understanding the evolution of plant-derived chemical defense in insects. What we term casual bioaccumulation of plant chemicals active against predators may be an important process that initiates selection toward the plant chemical-based defenses found in many insects. Bioaccumulation of natural and synthetic chemicals is a widespread phenomenon (Duffey, 1980). A large number of plants contain high levels of biologically active compounds, and a general relationship between
DIET BREADTH AND CHEMICAL DEFENSE IN A GRASSHOPPER
1821
concentration of such compounds in the diet and subsequently in the body does exist for many organisms (Duffey, 1980). It follows that situations in which a high concentration of compounds are ingested may be more likely to lead to concentrations of plant compounds in herbivorous insects that are active against predators. Our data on R. guttata suggest that the ecological equivalent of this dose relationship is diet breadth. Generalist consumers may be less likely to sequester chemicals in bioactive concentrations than their specialist counterparts because the generalist diet dilutes the concentrations of specific compounds ingested from each plant. It is possible that initiation of the evolution of sequestered defenses may arise when an insect feeds on a single host plant or limited numbers of species because casual sequestration of chemicals occurs in sufficient amounts for these compounds to be active against predators. If so, evolution of sequestered defenses may not necessarily initially require a high degree of adaptation or coevolution with a toxic host plant, as has been previously suggested. Acknowledgments--We thank C. O'Connor, T. Hess, B. Price, S. Tan for technical assistance; J.S. Coleman, P. Cusack, S.S. Duffey, T. Eisner, J. Kolasa, A.C. Lewis, G.E. Likens, G.G. Parker, J. Pasteels, S. Pickett, and M. Shachak for critical comment. Contribution to the program of the Institute of Ecosystem Studies, The New York Botanical Garden. Funding was provided by the Mary Flagler Cary Trust and NSF (DEB-8117943) to C.G.J. and NSF (DEB8117999) to M.S.B.
REFERENCES BOELENS, M., DE VALOIS, P.J., WOBBEN, M.J., and VAN DER GEN, A. 1971. Volatile flavor compounds from onion. J. Agric. Food Chem. 19:984-991. BLUM, M.S. 1981. Chemical Defenses of Arthropods. Academic Press, New York. 562 pp. BRODNITZ, M.H., POLLOCK, C.L., and VALLON, P.P. 1969. Flavor components of onion oil. J. Agric. Food Chem. i7:760-763. BROWER, L.P. 1984. Chemical defence in butterflies. Symp. R. Entomol. Soc. London 11:109134. BROWER, L.P., and BROWER, J.V.Z. 1964. Birds, butterflies and plant poisons: A study in ecological chemistry. Zoologica 49:137-159. CARSON, J.F., and WONG, F.F. 1959. Separation of aliphatic disulfides and trisulfides by gasliquid partition chromatography. J. Org. Chem. 24:175-179. CARSON, J.F., and WONG, F.F. 1961. The volatile flavor components of onions. J. Agric. Food Chem. 9:140-143. CRE1GHTOr~, W.S. 1950. The ants of North America. Bull. Mus. Comp. Zool. Harv. Univ. 104:8585. DADD, R.H. 1960. The nutritional requirements of locusts 1. Development of synthetic diets and lipid requirements. J. Insect Physiol. 4:319-347. DUFFEr, S.S. 1976. Arthropod allomones: chemical effronteries and antagonists, Proc. Intl. Congs. Entomol. 15:323-394. DUFFEY, S.S. 1980. Sequestration of plant natural products by insects. Annu. Rev. Entomol. 25:447477. DUFFEu S.S., and SCUDDER,G.G.E. 1972. Cardiac glycosides in North American Asclepiadaceae,
1822
JONES ET AL.
a chemical basis for unpalatability in brightly coloured Hemiptera and Coleoptera. J. Insect Physiol. 18:63-78. EISNER, T., HENDRY, L.B., PEAKALL, D.B., and MEINWALD, J. 1971. 2,5-Dichlorophenol (from ingested herbicide?) in defensive secretion of grasshopper. Science 172:277-278. JONES, C.G., HESS, T.A., WmTMAN, D.W., SILK, P.J., and BLOM, M.S. 1986. Idiosyncratic variation in chemical defenses between individual generalist grasshoppers. J. Chem. Ecol. 12:749-761. JONES, C.G., HESS, T.A., WHITMAN, D.W., SILK, P.J., and BLUM, M.S. 1987. Effects of diet breadth on autogenous chemical defense of a generalist grasshopper. J. Chem. Ecol. 13:283297. KEVAN, D.K.M. 1980. Romalea guttata (Houttuyn), name change for well-known "eastern lubber grasshopper" (Orthoptera: Romaleidae). Entomol. News 91:139-140. LOFGREN, C.F., BANKS,W.A., and CLANCY, B.M. 1975. Biology and control of imported fire ant. Annu. Rev. Entomol. 20:1-30. PASTEELS, J.M., GREGOIRE, J.-C., and ROWELL-RAHIER, M. 1983a. The chemical ecology of defense in arthropods. Annu. Rev. Entomol. 28:263-289. PASTEELS, J.M., ROWELL-RAHIER, M., BRAEKMAN,J.C., and DUPONT, A. 1983b. Salicin from host plant as precursor of salicylaldehyde in defensive secretion of chrysomeline larvae. J. Physiol. Entomol. 8:307-314. ROESKE, C.N., SEmER, J.N., BROWER, L.P., and MOFFITT, C.M. 1976. Milkweed cardenolides and their comparative processing by monarch butterflies (Danaus plexippus L.) Recent Adv. Phytochem. 10:93-167. ROTHSCHILD, M. 1973. Secondary plant substances and warning colouration in insects. Symp. R. Entomol. Soc. London 6:59-83. ROTHSCHILD, M., and APLIN, R.T. 1971. Toxins in tiger moths (Arctiidae: Lepidoptera). Pestic. Chem. 3:177-182. ROTHSCHILD, M., and PARSONS, J. 1962. Pharmacology of the poison gland of the locust Poekilocerus bufonius Klug. Proc. R. Entomol. Soc. London C27:21-22.
REDUCTION IN DIET BREADTH RESULTS IN SEQUESTRATION OF PLANT CHEMICALS AND INCREASES EFFICACY OF CHEMICAL DEFENSE IN A GENERALIST GRASSHOPPER
CLIVE G . J O N E S , 1 D O U G L A S COMPTON,1,
5 PETER
W.
WHITMAN, 2 STEVE J.
J. S I L K , 3 a n d M U R R A Y
S. B L U M 4
1Institute of Ecosystem Studies The New York Botanical Garden, Mary Flagler Cary Arboretum Box AB, Millbrook, New York 12545 2Department of Biology Illinois State University Normal, Illinois 61671 3pheromone Research Group New Brunswick Research and Productivity Council P.O. Box 6000, Fredericton, New Brunswick, E3B 5141, Canada 4Department of Entomology University of Georgia Athens, Georgia 30602 (Received June 1, 1988; accepted August 30, 1988)
Abstract--The lubber grasshopper, Romalea guttata, is a generalist feeding on a broad diet of many herbaceous plant species and has a metathoracic defensive secretion normally containing phenolics and quinones synthesized by the insect. When insects were reared on a restricted diet of wild onion, they sequestered sulfur volatiles from the plant into their defensive secretions. These compounds were not detected by gas chromatography-mass spectroscopy in secretions of insects on an artificial diet or a natural, generalist diet of 26 plants that included wild onion as a component, nor were they present in secretions from field-collected insects. Defensive secretions of insects reared on wild onion were significantly more deterrent, by as much as an order of magnitude, to two species of ant predators than secretions from insects on either of the other two diets, despite a reduction in the concentration of autogenous defensive chemicals in secretions of insects on the onion 5 Present address: University of Washington School of Medicine, Health Sciences Center, A-320, Seattle, Washington 98195. 1811 00984)331/89/0600-1811506.00/0 9 1989 Plenum Publishing Corporation
1812
JONES ET AL.
diet. Sequestration of plant chemicals that increased defensive efficacy occurred when diet breadth was reduced. We suggest that this occurs because under conditions of specialization, plant secondarymetabolites are more likely to be ingested and bioaccumulated in sufficient concentrations to have biological activity against predators. What we define as casual bioaccumulation of bioactive plant chemicals followingdietary specialization may lead to evolution of sequestered defense syndromesin insects, and this process may not necessarily require specific adaptation to or coevolution with a toxic host plant. Key Words--Lubber grasshopper, Romalea guttata (microptera), Orthoptera, Acrididae, onion, generalist diet, artificial diet, sulfur volatiles, ant predation, casual bioaccumulation, specialization, evolution of defenses.
INTRODUCTION
Many herbivorous insect species use chemical defenses against predators (Duffey, 1976; Blum, 1981; Pasteels et al., 1983a). In some cases, autogenous defensive chemicals are made by the insect from dietary precursors. Alternatively, chemicals from the diet may be sequestered into the body tissues, fluids or glands with or without modification by the insect. Examples of sequestered defenses include cardenolides in the monarch butterfly (Brower, 1984), milkweed bug (Duffey and Scudder, 1972), and pyrogomorphid grasshoppers (Rothschild and Parsons, 1962); salicylaldehyde in some chrysomelid beetles (Pasteels et al., 1983b); and cardenolides and pyrollizidine alkaloids in some arctiid moths (Rothschild and Aplin, 1971). The ecological factors that initiate or constrain evolution of the sequestered defense syndrome are of considerable interest but poorly understood. It has been suggested that sequestered defenses initially evolved as a special insect adaptation for avoiding toxicity of specific chemicals in the host plant, by placing these chemicals out of harm's way in tissues or fluids. Storage of these compounds then subsequently conferred a selective advantage against predators (Brower and Brower, 1964; Rothschild, 1973). An alternative possibility is that the sequestered defense syndrome may arise as a consequence of what we term initial casual sequestration or bioaccumulation of ingested chemicals in body tissues (Duffey, 1976, 1980). The chemicals may not necessarily be toxic to the insect, and specific adaptation to a host plant may not be necessary. Bioaccumulated compounds might then deter predators if they were sequestered in sufficient concentrations. In this article we present evidence that two key characteristics of a sequestered defense syndrome--incorporation of plant chemicals into defensive secretions, and a concomitant increase in defensive efficacy against predators--can arise in an insect when there is no evidence for specific adaptation to a plant species, when the plants are not toxic to the insect, and in
DIET BREADTH AND C H E M I C A L DEFENSE IN A GRASSHOPPER
1813
a manner best described as casual sequestration. Furthermore, we show that diet breadth is the major ecological variable that determines whether or not sequestration of at least one class of plant chemicals occurs in sufficient amounts to have effects on predators. The lubber grasshopper, Romalea guttata (Houttuyn) [Orthoptera; Romaleidae = R. microptera (Beauvois)] (Kevan, 1980), is a large, flightless, generalist grasshopper with a local host range of over 80 species of herbaceous and woody plants. Many of these plants will be eaten within the lifetime of an individual insect (C.G. Jones, D.W. Whitman, M.S. Blum, unpublished data). Adult Romalea have an impressive defensive armory, including warning coloration, a pink wing-flashing display and sharp spines on the legs. Most notable is the odoriferous, frothy, glandular secretion produced with an audible hiss from modified metathoracic spiracles (Eisner et al., 1971). This secretion is deterrent to ants and other predators (Eisner et al., 1971; D.W. Whitman, C.G. Jones, M.S. Blum, unpublished data). The major components of the secretion of wild insects are various phenolics and quinones, including hydroquinone, catechol, p-benzoquinone, phenol, guaiacol, and 4-methoxybenzaldehyde (Eisner et al., 1971; Jones et al., 1986, 1987). These compounds are presumably autogenous because they are found in the secretions of insects reared on an artificial diet that does not contain any of these chemicals (Jones et al., t987). The autogenous defensive components are markedly affected by diet breadth. Restricting Romalea to a single host plant or artificial diet results in a reduction in the number and concentration of compounds, and thus alters the composition of the secretion, compared to lubbers reared on a generalist diet or collected in the field (Jones et al., 1987). Eisner et al. (1971) showed that Romalea collected from an area sprayed with the herbicide 2,4-D had 2,5-dichlorophenol in the defensive secretion. We theretbre wondered if Romalea was capable of sequestering natural plant products into the defensive secretion. In our studies on the effects of diet breadth on autogenous defenses (Jones et al., 1987), we noticed that the secretions of insects reared on a single host plant diet of onion appeared to have an onionlike odor. Here, we evaluated the possibility that sequestration of onion compounds might have occurred. We reared Romalea from first instar to adult on three different diets: (1) a restricted single plant species diet of wild onion, Allium canadense L., that we have observed is one of a number of favored food plants of Romalea (C.G. Jones, D.W. Whitman, M.S. Blum, unpublished data); (2)a "natural" generalist diet of plant species used by R. guttata in the field, which included wild onion as a component species; and (3) a restricted artificial diet known to lack any of the insects' autogenous phenolics and quinones and lacking any plant secondary metabolites. We compared the chemistry of secretions from insects reared on the different diets. We then bioassayed the secretions against a known ant predator of Romalea, the red imported fire ant, Solenopsis invicta
1814
JONES ET AL.
Buren (Formicidae: Myrmicinae), and a less aggressive ant species, Tapinoma melanocephalum F. (Formicidae: Dolichoderinae). Our findings are relevant to the ecology and evolution of chemical defense in insects.
METHODS AND MATERIALS
Rearing of Insects. Lubber grasshoppers were reared under the same conditions as used by Jones et al. (1987). The first of the three diets consisted of freshly cut leaves of Allium canadense in water. The second diet consisted of 26 species from 15 families of shrubs, herbs, forbs and grasses known to be eaten by Romalea in the field (C.G. Jones, D.W. Whitman, M.S. Blum, unpublished data). Plants were collected daily and the petioles or stems placed in water. The species list was as in Jones et al. (1987) and included wild onion, Allium canadense. The third diet was an artificial diet (Dadd, 1960) containing cellulose, sucrose, dextrin, mineral salts, cholesterol, linoleic acid, casein, peptone, egg albumin, ascorbic acid, and B vitamins and was fed as a water-moistened cake. Ten males and 10 females were also collected from the field for chemical analysis of their secretions. Secretion Collection. Secretions were collected from adults at the eighth day after the terminal molt, by gently squeezing the thorax while applying a 10-/xl microcapillary to the metathoracic spiracule orifices. A single milking removes all the secretion produced since the last molt. Molting insects discard all previous secretions. The volume collected (range 1-8 tzl) was recorded and the secretion sealed in the microcapillary and stored at - 10~ Secretions from insects reared in the same batch on the different diets were randomly allocated for: (1) separately pooled secretions of males and females for mass-spectral analysis, (2) secretions from individual females and pooled secretions of females for quantitative analysis of sequestered compounds, (3) secretions from individual females for quantitative analysis of autogenous compounds, and (4) pooled secretions of females for bioassays. In addition, separately pooled secretions of wild male and female insects were subjected to mass-spectral analysis. Qualitative Analysis of Secretions. Compounds were identified from splitless injection (10 psi) of undiluted secretion on a Finnigan 4021 gas chromatograph-mass spectrometer (GC-MS) using a 30-m x 0.3-mm-ID DB-5 fused silica bonded phase column (J&W), programmed at 50~ for 1.5 min, then 100-250~ at 15~ The GC-MS was operated in electron impact mode (70 eV, 1400 V, 10 -7 amp/V; scanning 45-350 amu, 1.5 sec/scan) with acquisition commencing at sample injection. Compounds were identified by using the molecular ions and characteristic mass fragments, by comparison with library spectra (r fit), and by GC-MS of authentic standards. Pooled secretions col-
DIET BREADTH AND C H E M I C A L DEFENSE IN A GRASSHOPPER
1815
lected from 10 individual males and females reared on each of the three diets and 10 males and females collected in the field were analyzed by GC-MS. Quantitative Analysis of Sequestered Compounds. Diethyl ether extracts of secretions were analyzed by subambient, capillary gas chromatography (GC) (Varian 6000), using a 30-m • 0.32-mm-ID fused silica bonded phase column (J&W, DB-5), helium carrier (2 ml/min) splitless injection (200~ purge at 0.7 min), and flame ionization detection (300~ 1 • 10 ~2 sensitivity). Initial column temperature was 20~ ramped to 40~ at 10~ then 40~ to 200~ and held at 200~ for 2 min. Secretions from eight individual females and the pooled secretions of three females reared on the onion diet were each extracted with 10/~1 redistilled diethyl ether and immediately analyzed by GC. The 2.5-tA aliquots were injected, and exact volumes of ether recovery and reproducibility were determined by addition of dl-alpha pinene to the ether as an internal standard, prior to extraction (325 t~mol/liter). We determined the concentration of eight sulfur-containing volatiles as the mean of three replicate injections: 1-propanethiol (minimum detectable quantity per microliter, 14 pg), methylthiirane (5 pg), methyldisulfide (21 pg), isopropylsulfide (11 pg), 2,5dimethylthiophene (30 pg), methylpropyldisulfide (13 pg), isopropyldisulfide (28 pg), propyldisulfide (25 pg). Retention times were accurate to _+1% ( P = 0.05); peak heights with tangent baseline correction were accurate to + 10% (P = 0.05), based on analysis of standards and replicate injections of pooled samples, and were measured on a Varian 402 integrator. Retention times and concentrations of compounds were determined by calibration against authentic standards and confirmed by using coinjection and independent injection. Standards were commercially available, except for methylpropyldisulfide, which was synthesized by the method of Carson and Wong (1959). Combined concentrations were determined from the sum of the concentrations of each compound. Quantitative Analysis of Autogenous Compounds. Secretions were analyzed for six phenolics and quinones (hydroquinone, p-benzoquinone, phenol, catechol, guiaicol, 4-methoxybenzaldehyde) using tertiary gradient reversephase high-performance liquid chromatography (HPLC) on a C1s bonded-phase silica column with a water-methanol-l% H3PO4 convex elution gradient, according to the method of Jones et al. (1986, 1987). Secretions from 10 individual females reared on each diet were diluted to 1% in 18 mohm deionized water, and triplicate 10-t~l aliquots were analyzed. Combined concentrations of phenolics and quinones were determined from the sum of the concentrations of each compound. Bioassay of Secretions. In preliminary binary choice experiments, we determined that ants preferred pure sucrose solutions to those containing wild Romalea defensive secretions (D.W. Whitman, C.G. Jones, M.S. Blum,
1816
JONESET AL.
unpublished data). We therefore decided that relative deterrency of the three different secretions was best determined by simultaneously comparing the secretions against one another, when they were made up in sucrose solution. Five laboratory colonies of Solenopsis invicta, varying in size from 1000 to 4000 individuals, were each presented with a 45 x 75-mm glass slide containing three separate 50-/H droplets of 4% sucrose. Each separate droplet contained 1 /A of Romalea metathoracic secretion pooled from individual females reared on either the natural, the onion, or the artificial diet. Ants foraged down a common trail to this comparative bait station. The common foraging trail p r e v e n t e d recruitment to one bait from affecting the total number of foragers approaching the bait station and reduced any biases due to the deterrency of a given bait. The number of ants feeding on each droplet was recorded every 4.5 min, up to the time that the first of the sucrose baits with secretion was completely consumed. Trials were repeated three times for each colony, with the order of the three droplets on the slide being changed for each trial. It was determined that the number of ants visiting the station initially increased, stabilized by 10 min, and remained more-or-less constant for over 30 min. Therefore the number of ants at each bait at 40 min was recorded for each colony and trial. These data were analyzed using ANOVA followed by multiple comparisons (least-significant-difference tests). Two laboratory colonies of Tapinoma melanocephalum (colony size no more than 3000 individuals) were tested using the above procedures. Ten percent sucrose solution containing secretions was used because this species of ant did not recruit sufficient numbers to 2% pure sucrose solution to permit adequate evaluation of deterrency. Trials were repeated twice. In addition, pairwise comparisons of artificial diet secretion (1/H) + 10% sucrose versus natural diet secretion + sucrose, and natural diet + sucrose versus onion diet + sucrose were also carried out using the glass slide technique. Three lab colonies were used with five repeated trials each. Again, the number of ants visiting the baits at 20 min was recorded. Data in these trials were analyzed as above.
RESULTS
Sequestration of Plant Compounds. The secretions of insects reared on the wild onion diet lacked the characteristic color and odor of secretions from insects on the natural diet and had a marked onionlike odor. GC-MS of the volatile fraction of pooled secretions from males and females showed the presence of a large number of sulfur-containing compounds (C.G. Jones and P.J. Silk, unpublished data). Eight compounds were unambiguously identified (isopropylsulfide, methylpropyldisulfide, isopropyldisulfide, methyldisulfide, propyldisulfide, propanethiol, methylthiirane, and 2,5-dimethylthiophene) and were
DIET BREADTH AND C H E M I C A L DEFENSE IN A GRASSHOPPER
1817
identical to those previously isolated and identified from onion (Carson and Wong, 1961; Brodnitz et al., 1969; Boelens et al., 1971). The total concentrations of these eight onion sulfur volatiles that were generally the most abundant in the secretions were quantified in the defensive secretions of eight adult female individuals reared on the onion diet. Mean concentration of all eight of these compounds combined was 0.5 ng//zl (Figure la). None of these sulfur volatiles, nor any other sulfur-containing compounds, were detected in GC-MS analyses of pooled secretions from male or female insects reared on the natural or artificial diet (pooled from 10 of each sex on each diet), or from secretions pooled from 10 individuals of each sex of field-collected insects occurring in an area where a diversity of plant species grow, including wild onion. In a separate rearing experiment, sulfur volatiles were detectable by odor and GC-MS in defensive secretions of insects within five days of transfer from artificial to onion diets. Auto~genous Compounds. Secretions from insects raised on either onion or the artificial diets contained significantly lower concentrations of all autogenous phenolics and quinones than secretions from insects on the generalist, natural diet. Combined phenolic and quinone concentrations were depressed fivefold on the artificial diet and 37-fold on the onion diet (Figure lb). The concentrations of autogenous compounds in secretions of insects on onion diets were also significantly lower than concentrations in artificial diet secretions (t tests, P < 0.0001). These new data agree with our previous findings for autogenous compounds (Jones et al., 1987). Efficacy of Defensive Secretions. To determine whether changes in the secretion chemistry affected biological activity, we bioassayed secretions against two ant species. The imported fire ant, Solenopsis invicta Buren, is a very aggressive predator of insects (Lofgren et al., 1975). This ant species did not evolve in the same habitat as Romalea guttata, but we have observed predation by Solenopsis on Romalea in the field (D.W. Whitman, C.G. Jones, M.S. Blum, unpublished data). Tapinoma melanocephalum F. is a less aggressive, omnivorous scavenger sympatric with Romalea in Florida (Creighton, 1950). We estimated the relative deterrency of the defensive secretions. With Solenopsis invicta, significantly more ants visited and fed at the artificial diet secretion bait compared to the natural diet secretion bait, and significantly more ants visited and consumed both of these baits compared to the onion diet secretion bait (Figure lc shows means + 1 SE). Overall effect of treatments was significant: ANOVA: F = 16.42, df = 44, P = 0.0001. All means in multiple comparisons were significantly different from each other, LSD t tests, cr = 0.05.) Approximately 63, 30, and 7 % of all foraging ants visited and fed at the artificial, natural, and onion diet secretion baits, respectively. The less aggressive ant, Tapinoma, showed even greater differences between treatments. In trials where all three secretion baits were presented
1818
JONES ET AL.
0.6-
0.5 {33 E tO
0.4
0.3
0 tO 0
0.2 0.1
Natural Diet
Artificial Diet
Onion Diet
b 7000" :~
6000-
C 5000= C .9 4000-
E aooo0 2000" O 1000ONatural Diet
Artificial Diet
Onion Diet
131211109876-
e~
E Z
Natural Diet
Artificial Diet
Onion Diet
D I E T B R E A D T H A ND C H E M I C A L D E F E N S E IN A G R A S S H O P P E R
18 t 9
together, the proportion o f ants visiting and feeding at the baits was approximately 70 (artificial diet), 29 (natural diet), and 1% (onion diet) (mean _+ 1 SE number o f ants at bait at 20 min for two colonies and two trials was: artificial 13.3 + 3.4; natural 5.0 + 1.6; onion 0. A N O V A : F = 9.55, df = 11, p = 0.006. All pairs o f means in multiple comparisons were significantly different from one another, LSD t tests, c~ = 0.05). In comparisons with Tapinoma, where pairs o f secretions were tested together, the number of ants at artificial diet secretion baits at 40 min (mean + 1 SE = 14.9 _+ 1.7) was significantly greater than the number at natural diet secretion baits (mean + 1 SE = 7.0 + 1.4; A N O V A : F = 8.43, df = 49, p = 0.007). Numbers at 40 min at natural diet secretion baits (mean +_ 1 SE = 45.1 _+ 4.4) were significantly greater than numbers at onion diet secretion baits (mean + 1 SE = 0.3 _+ 0.16; A N O V A : F = 35.49, d f - - 49, p = 0.0001).
DISCUSSION Diet ,clearly affects the quality and quantity o f the chemical components of the defensive secretion of Romalea and the efficacy of the defensive secretion against these predators. Sequestration of measurable quantities o f sulfur compounds occurred only when Romalea specialized on the onion diet. The natural diet also included wild onion as a component species, and yet secretions from insects reared on this diet did not contain any detectable sulfur-containing compounds, nor were these compounds found in field-collected insects that fed on a large number o f plant species. Since these compounds were not present in the artificial diet, their absence from secretions o f insects specializing on this diet was to be expected. In addition, specialization on either artificial or onion diets again caused a reduction in the concentrations o f autogenous phenolics and quinones, as has been previously reported (Jones et al., 1987). The response o f ants to baits containing these secretions is consistent with the chemical differences resulting from the presence of sequestered compounds and the depression of autogenous compounds. Sequestration o f sulfur corn.( FIG. t. (a) Mean combined concentrations of eight sulfur-containing volatiles (ng//~l secretion, +_ 1 SE) in defensive secretions of individual adult female R. gutmta (N = 8), reared on an onion diet. Pooled secretions of females (N = 10) on the other diets did not contain detectable levels of sulfur volatiles. (b) Mean combined concentrations of six autogenous phenolics and quinones (ng/~l secretion, + 1 SE) in defensive secretions of adult female R. guttata (N = 10), reared on three diets. (c) Mean number of Solenopsis invicta ants (_+ 1 SE, five colonies, three trials each) visiting and feeding at 20 min on sucrose baits containing defensive secretions of R. guttata females reared on three diets.
1820
JONES ET AL.
pounds in onion diet secretions is associated with an increase in defensive efficacy compared to natural diet secretions, despite depression of the concentrations of autogenous phenolics and quinones. Reduction of autogenous phenolic and quinone concentrations in artificial diet secretions is associated with a reduction in the efficacy of defense compared to natural diet secretions. The deterrent activity of components sequestered on the restricted onion diet appears to be much greater than the activity of the autogenous compounds. The combined sequestered sulfur compounds occurred at concentrations about 104 lower in onion diet secretions than combined concentrations of phenolics and quinones in the natural diet, and yet the onion diet secretion bait was visited and fed on by Solenopsis at about one quarter the rate of the natural diet secretion bait. Dietary specialization in R. guttata therefore led to rapid and simultaneous changes in two attributes important to its chemical defense: acquisition of compounds directly from the diet and a reduction in autogenous compounds. This led to a concomitant increase in defensive efficacy. Experimental specialization in Romalea produces results strikingly similar to many specialist insect herbivores where chemical defense against predators depends upon sequestration of host-plant chemicals (Brower, 1984; Rothschild, 1973; Roeske et al., 1976; Duffey, 1976, 1980; Blum, 1981). Why should specialization result in sequestration in Romalea ? We propose that the degree of sequestration of plant chemicals in this insect is primarily a function of the concentration of compounds ingested in the diet. When a generalist diet is consumed, compounds from an individual plant species may be too dilute to be detectable or biologically active against predators in comparison to the autogenous compounds. Conversely, specialization results in the ingestion of sufficiently high concentrations of compounds from one plant species such that these compounds may be sequestered in detectable levels that have biological activity. Interestingly, data from Eisner and coworkers (1971) may be relevant to our contention. They reported 2,5-dichlorophenol in the defensive secretion of R. guttata collected in herbicide-sprayed areas and suggested that this compound was derived from the herbicide 2,4-D. 2,5-Dichlorophenol was not present in secretions of insects from unsprayed areas. Although in this case Romalea was consuming a generalist diet, it may have been, in effect, specializing on a halogenated phenol diet because all of the plants were sprayed with herbicide. Our results may have important implications for understanding the evolution of plant-derived chemical defense in insects. What we term casual bioaccumulation of plant chemicals active against predators may be an important process that initiates selection toward the plant chemical-based defenses found in many insects. Bioaccumulation of natural and synthetic chemicals is a widespread phenomenon (Duffey, 1980). A large number of plants contain high levels of biologically active compounds, and a general relationship between
DIET BREADTH AND CHEMICAL DEFENSE IN A GRASSHOPPER
1821
concentration of such compounds in the diet and subsequently in the body does exist for many organisms (Duffey, 1980). It follows that situations in which a high concentration of compounds are ingested may be more likely to lead to concentrations of plant compounds in herbivorous insects that are active against predators. Our data on R. guttata suggest that the ecological equivalent of this dose relationship is diet breadth. Generalist consumers may be less likely to sequester chemicals in bioactive concentrations than their specialist counterparts because the generalist diet dilutes the concentrations of specific compounds ingested from each plant. It is possible that initiation of the evolution of sequestered defenses may arise when an insect feeds on a single host plant or limited numbers of species because casual sequestration of chemicals occurs in sufficient amounts for these compounds to be active against predators. If so, evolution of sequestered defenses may not necessarily initially require a high degree of adaptation or coevolution with a toxic host plant, as has been previously suggested. Acknowledgments--We thank C. O'Connor, T. Hess, B. Price, S. Tan for technical assistance; J.S. Coleman, P. Cusack, S.S. Duffey, T. Eisner, J. Kolasa, A.C. Lewis, G.E. Likens, G.G. Parker, J. Pasteels, S. Pickett, and M. Shachak for critical comment. Contribution to the program of the Institute of Ecosystem Studies, The New York Botanical Garden. Funding was provided by the Mary Flagler Cary Trust and NSF (DEB-8117943) to C.G.J. and NSF (DEB8117999) to M.S.B.
REFERENCES BOELENS, M., DE VALOIS, P.J., WOBBEN, M.J., and VAN DER GEN, A. 1971. Volatile flavor compounds from onion. J. Agric. Food Chem. 19:984-991. BLUM, M.S. 1981. Chemical Defenses of Arthropods. Academic Press, New York. 562 pp. BRODNITZ, M.H., POLLOCK, C.L., and VALLON, P.P. 1969. Flavor components of onion oil. J. Agric. Food Chem. i7:760-763. BROWER, L.P. 1984. Chemical defence in butterflies. Symp. R. Entomol. Soc. London 11:109134. BROWER, L.P., and BROWER, J.V.Z. 1964. Birds, butterflies and plant poisons: A study in ecological chemistry. Zoologica 49:137-159. CARSON, J.F., and WONG, F.F. 1959. Separation of aliphatic disulfides and trisulfides by gasliquid partition chromatography. J. Org. Chem. 24:175-179. CARSON, J.F., and WONG, F.F. 1961. The volatile flavor components of onions. J. Agric. Food Chem. 9:140-143. CRE1GHTOr~, W.S. 1950. The ants of North America. Bull. Mus. Comp. Zool. Harv. Univ. 104:8585. DADD, R.H. 1960. The nutritional requirements of locusts 1. Development of synthetic diets and lipid requirements. J. Insect Physiol. 4:319-347. DUFFEr, S.S. 1976. Arthropod allomones: chemical effronteries and antagonists, Proc. Intl. Congs. Entomol. 15:323-394. DUFFEY, S.S. 1980. Sequestration of plant natural products by insects. Annu. Rev. Entomol. 25:447477. DUFFEu S.S., and SCUDDER,G.G.E. 1972. Cardiac glycosides in North American Asclepiadaceae,
1822
JONES ET AL.
a chemical basis for unpalatability in brightly coloured Hemiptera and Coleoptera. J. Insect Physiol. 18:63-78. EISNER, T., HENDRY, L.B., PEAKALL, D.B., and MEINWALD, J. 1971. 2,5-Dichlorophenol (from ingested herbicide?) in defensive secretion of grasshopper. Science 172:277-278. JONES, C.G., HESS, T.A., WmTMAN, D.W., SILK, P.J., and BLOM, M.S. 1986. Idiosyncratic variation in chemical defenses between individual generalist grasshoppers. J. Chem. Ecol. 12:749-761. JONES, C.G., HESS, T.A., WHITMAN, D.W., SILK, P.J., and BLUM, M.S. 1987. Effects of diet breadth on autogenous chemical defense of a generalist grasshopper. J. Chem. Ecol. 13:283297. KEVAN, D.K.M. 1980. Romalea guttata (Houttuyn), name change for well-known "eastern lubber grasshopper" (Orthoptera: Romaleidae). Entomol. News 91:139-140. LOFGREN, C.F., BANKS,W.A., and CLANCY, B.M. 1975. Biology and control of imported fire ant. Annu. Rev. Entomol. 20:1-30. PASTEELS, J.M., GREGOIRE, J.-C., and ROWELL-RAHIER, M. 1983a. The chemical ecology of defense in arthropods. Annu. Rev. Entomol. 28:263-289. PASTEELS, J.M., ROWELL-RAHIER, M., BRAEKMAN,J.C., and DUPONT, A. 1983b. Salicin from host plant as precursor of salicylaldehyde in defensive secretion of chrysomeline larvae. J. Physiol. Entomol. 8:307-314. ROESKE, C.N., SEmER, J.N., BROWER, L.P., and MOFFITT, C.M. 1976. Milkweed cardenolides and their comparative processing by monarch butterflies (Danaus plexippus L.) Recent Adv. Phytochem. 10:93-167. ROTHSCHILD, M. 1973. Secondary plant substances and warning colouration in insects. Symp. R. Entomol. Soc. London 6:59-83. ROTHSCHILD, M., and APLIN, R.T. 1971. Toxins in tiger moths (Arctiidae: Lepidoptera). Pestic. Chem. 3:177-182. ROTHSCHILD, M., and PARSONS, J. 1962. Pharmacology of the poison gland of the locust Poekilocerus bufonius Klug. Proc. R. Entomol. Soc. London C27:21-22.
