Orthoptera: Acrididae): Comparison Between a Specialist Species on a Shrubby Legume and Three Graminivorous Species

PLANT-INSECT INTERACTIONS Decision Making for Food Choice by Grasshoppers (Orthoptera: Acrididae): Comparison Between a Specialist Species on a Shrubby Legume and Three Graminivorous Species

1 1 2 F. PICAUD, E. BONNET, V. GLOAGUEN, AND D. PETIT

Faculte´ des Sciences, E.A. 3176, 123, av. A. Thomas, 87060 Limoges, France

Environ. Entomol. 32(3): 680Ð688 (2003) ABSTRACT Dwarf gorse bush( Ulex minor) heathlands in Limousin, France, are ecological islands often separated by tens of kilometers of grasslands and hedges, where several species of grasshoppers belonging to the genus Chorthippus (Acrididae: Gomphocerinae) coexist. Chorthippus binotatus (Charpentier) feeds only on Ulex minor; nymphs feed exclusively on leaves whereas adults become ßorivorous at the end of the season. The other species studied (C. biguttulus (L.), C. albomarginatus (De Geer), and C. parallelus (Zetterstedt)) are all graminivorous. The importance of sugars, nitrogen content, sparteine (a quinolizidine alkaloid), and plant architecture in food selection was investigated. Chorthippus binotatus is sensitive to sucrose and fructose, consistent with the high sugar content of Ulex minor ßowers. Experiments with grass coated with sparteine showed that this molecule associated with sucrose is a phagostimulant for this grasshopper. Different behavioral responses of graminivorous species are observed with sparteine alone, but never phagostimulation. We compared the response times corresponding to decision-making between the different species toward several components involved in food selection. The food choice toward host plant and sugars is as quick for C. binotatus as for the two graminivorous species (C. parallelus and C. albomarginatus), whereas C. biguttulus is slower and exhibits atypical reactions. Chorthippus binotatus can feed on Poaceae, but withmore time spent, leading to an increasing predation risk. This situation is a limitation toward dispersal between different heathlands (patchy habitats).

KEY WORDS decision making, patchy habitats, alkaloids, Chorthippus, diet shift, graminivory

CHAPMAN (1982) NOTED that in the majority of phyto- Gangwere and Morales-Agacino (1973). All species phagous insect orders, Ͼ50% of the species are oli- are graminivorous withtheexception of Chorthippus gophagous (they feed on one plant family only) or binotatus (Charpentier) (Picaud et al. 1999). This spe- monophagous. The Orthoptera, and particularly Ac- cies is present in France, Spain, and Morocco and ridoidea, stand apart from the other orders: 60% of comprises six subspecies, Þve of which occur in France grasshoppers have been classiÞed as polyphagous, and (Defaut 1999). Chorthippus binotatus binotatus feeds a further 25% are graminivorous. almost exclusively on various species belonging to the In NorthAmerica, at least two Gomphocerinaespe- Genisteae tribe (Fabaceae) (Grasse´ 1929), withsome cies have evolved specialized diets from graminivory regional specialization: Ulex minor Rothin Pe´rigord under conditions in which grasses were often very and Limousin, Ulex europaeus L. and Cytisus scoparius sparse (Otte and Joern 1977). Ligurotettix coquilletti (L.) Link in Charente-Maritime, Genista scorpius (L.) McNeill feeds on Larrea (Zygophyllaceae), Atriplex DC. In the French Pyrenees (Defaut 1999), and Ulex (Chenopodiaceae), and Lycium (Solanaceae) (Chap- parviﬂorus Pourret in Vaucluse (Mossot and Petit man and Joern 1990). Bootettix argentatus Bruner 2000). In Limousin, the dietary preference of C. bino- lives and feeds on the creosote bush Larrea tridentata tatus binotatus during the season is synchronized with (Zygophyllaceae) (White and Chapman 1990). The the life history of Ulex minor: the grasshopper feeds basis for this preference depends partly on the pres- shoots before blooming and ßowers during the sum- ence of the lignan nordihydroguaiaretic acid (NDGA) mer (Picaud et al. 1999, 2002). In the laboratory, the (Chapman et al. 1988). entire life cycle can be completed with Ulex minor as In France, the genus Chorthippus (Gomphoceri- the only food supply. nae) comprises 16 species (Bellmann and Lucquet It is certain that C. binotatus evolved from gramin- 1993). The diet of all species was characterized by ivorous species: the 15 other species in the genus Chorthippus feed on Poaceae, as well as the French 1 Socie´te´ Entomologique du Limousin, Limoges, France. species of the sister genera Omocestus and Stenoboth- 2 Laboratoire CSN, Faculte´ des Sciences de Limoges, France. rus (Bonnet 1996). It is well known that grass-feeding

0046-225X/03/0680Ð0688$04.00/0 ᭧ 2003 Entomological Society of America June 2003 PICAUD ET AL.: DECISION MAKING FOR FOOD CHOICE BY GRASSHOPPERS 681

Fig. 1. Distribution map of Chorthippus binotatus subspecies binotatus in France (hatched parts) and in Limousin Region (open circles), divided in three departments (after Kruseman, 1982; Defaut, 1999). Plant samples for analyses were collected in Ceinturat heathland (black circle). grasshoppers feed mainly on the most abundant basis of food selection in grass feeding and Ulex minor Poaceae species (Bernays and Chapman 1994). Be- specialist grasshoppers. We analyzed sugars and ni- cause the chemical composition of secondary metab- trogen content in the plant species with a special focus olites of Graminae are rather uniform (Bruneton on Ulex minor and Poaceae. As for secondary com- 1999), one can assume that grass-feeding grasshoppers pounds, it is well known that numerous quinolizidine are oligophagous and therefore relatively specialist alkaloids characterize the tribe of Genisteae (Brune- species. ton 1999). Our investigation was limited to sparteine, In Limousin, the heathlands correspond to ecolog- the major alkaloid in the genus Ulex (Picaud et al., ical islands (Brugel et al. 2001). Thus, for species using unpublished data). The inßuence of plant architec- food resources that are limited and patchy, the dis- ture was also tested: ramiÞed in the case of Ulex, tance between these bush formations can be a con- straight in the case of Poaceae. straint for dispersal. A less obvious constraint that may We discuss the Þndings in relation to the beneÞt of be associated with oligophagy is the problem of food quick decision times (Bernays and Bright 2001, quality, including nitrogen and/or carbohydrates lev- Abrams and Schmitz 1999) and the disadvantage of els, and the balance between deterrents and phago- being a specialist on Ulex minor, as the Ulex minor stimulants. heathlands form islands between grasslands and pas- Bernays and Bright (2001) showed that food choice tures. of Schistocerca americana (Drury) between different combinations of two plant species causes interrupted Materials and Methods feeding as a result of inefÞcient decision making. So, feeding on only one plant species saves time and po- Experimental Animals. Chorthippus adults (C. bino- tentially beneÞts grasshoppers. One can ask several tatus, C. biguttulus, C. albomarginatus, and C. paral- questions about the properties linked to feeding upon lelus) were collected in September and October 1993Ð a given plant: (1) in the case of binary choice (Ulex 2001 in heathlands in the Limousin region (45Ð46 minor versus Poaceae), do the insects feed exclusively degrees N. latitude, 1Ð2 degrees E. longitude), France on their host plants; (2) what are the characteristics of (Fig. 1). If we take into account the mean of absolute Ulex minor involved in the diet preference of C. bino- plant cover in seven heathlands where C. binotatus tatus; and (3) is the specialist species on Genisteae was present, the most abundant plant species are: Ulex bushes as efÞcient as each graminivorous species in minor (48%), Calluna vulgaris (34%), Molinia caerulea their food selection? (23%), Erica cinerea (17%), Pteris aquilinum (7%), and In this work, we examined the feeding behavior and Erica tetralix (4%). Insects were placed individually in the food selection of all the Chorthippus species co- glass jars (6 cm diameter ϫ 10 cm height) and kept at existing with Chorthippus binotatus in Limousin heath- room temperature (20 Ϯ 2ЊC), under a L:D 12:12 h lands: C. biguttulus (L.) (subgenus Glyptobothrus, as photoregime and 50% humidity for the duration of the C. binotatus), C. parallelus (Zetterstedt), and C. albo- experiments. Theywere fed daily withfreshplants marginatus (De Geer) (subgenus Chorthippus). It is collected at the Limoges University: Ulex minor for known that the three latter species are graminivorous C. binotatus and a mix of Agrostis tenuis, Holcus lanatus, (Bonnet 1996). We investigated food quality as the and Lolium perenne for the graminivorous species. 682 ENVIRONMENTAL ENTOMOLOGY Vol. 32, no. 3

Every week after the experiments, all living insects were returned to the wild and replaced by newly collected specimens. Tested Plants and Food Selection Experiments. We chose Ulex minor (ßowering time in JulyÐOctober) and the most abundant grass in heathlands (Molinia caerulea) for testing. Another grass species, Lolium perenne, was also tested because of its abundance at the Limoges University. Before each experiment, food Fig. 2. Drawings presented to insects in plant architec- was withheld from insects for 4Ð6 h (Bernays and ture experiments. Chapman 1970). Food choice tests with 16Ð32 individual adults of bothsexes of eachspecies were performed in a Petri avoid consumption of the piece of Þlter paper. The dish(14.2 cm diameter ϫ 2 cm height) isolated in a choice consisted of two pieces of Þlter paper (each green enclosure. A 40 W-lamp above the center of the one measuring 6 ϫ 6 cm, and four folds), placed at dishprovided constant lightand minimized temper- random on opposite edges of the dish, one with one M ature ßuctuations. Eachexperiment was limited to sugar solution (1 ml), and one withdistilled water 15 min to avoid postingestive feed-backs. Thus, deci- (20Ð21 individuals for eachspecies). sions by grasshoppers depended on receptor sensitiv- Decomposition Analysis of Shape of the Plant. In- ity and the physiological state of individuals. The dividual grasshoppers were placed for 10 min in a grasshopper was transferred to the center of a dish in Y-maze withopaque walls (30 ϫ 8 ϫ 11 cm) covered a small glass jar. It was given a choice between two with Plexiglas. The grasshoppers had to choose be- separate sources, eachconsisting of freshplants tween two drawings (8 ϫ 11 cm) placed vertically and (Ϸ1 g), cut with scissors just before the start of the at random ends of the two branches (18 ϫ 8 ϫ 11 cm). bioassay. Three different behavioral responses were Choice between the drawings was recorded when the recorded (Williams 1954, Bernays and Chapman 1970, insect reached the end of the Y-maze and climbed on Uvarov 1977): orientation (the grasshopper walked to the drawing. The time from initiation of movement the plant (item 1), palpated the plant (item 2), and ate until climbing onto the drawing was recorded. Two the plant (item 3). To help understand food choice, we pairs of green patterns on a yellow background, to examined sugars, architecture, and the secondary me- mimic Ulex branches in the environment, were pre- tabolite sparteine. sented: eachindividual (19Ð21 by species) hadto During plant choice experiments, the sum of ori- choose between straight versus branching aspects and entation and palpation times from initiation of move- straight versus crossed aspects in separate bioassays ment until the decision (feeding upon or rejecting it) (Fig. 2). was documented. Concerning the experiments with Alkaloid Sensitivity. Preliminary experiments sugars and the alkaloid, non - volatile chemical com- demonstrated (data not shown) that C. binotatus and pounds, only palpation time before decision was taken C. biguttulus have no preference or rejection behavior into account. for 1% acetic acid versus water, on Þlter paper. Alka- Chemical Analysis of Plants (soluble carbohydrates loid content in Leguminosae vary generally between and nitrogen content). The samples of Ulex minor 0.1 and 3% of plant dry weight (Bruneton 1999). Our ßowers, shoots, and stems were collected from Cein- analyses of different organs of Ulex minor (Picaud et turat heathland (Fig. 1), Þve in September 1997 and al., unpublished data) have shown that sparteine rep- four in July 1999 at the beginning of blooming for a resents between 30Ð48% of nonlinked alkaloids. Dif- series of analyses. Four samples of two species of ferent concentrations (in 1% acetic acid) of sparteine Graminae were collected in grasslands of the Limoges 99% (Sigma Aldrich, Saint-Quentin, France) were University in early July 1999. The separation and iden- tested under the same conditions as described in tiÞcation of soluble sugars (glucose, fructose and su- “sugar sensitivity” trials, witha 1% acetic acid treat- crose) was adapted from Harisson et al. (1997) and ment as a control. Experiments were repeated three Cataldi et al. (1999) and performed by high perfor- times withdifferent sets of 16Ð24 individuals of mance liquid chromatography (HPAEC-PAD). Elu- C. binotatus, tested separately with0.1%, 0.5%, 1% and tion of the samples was carried out in isocratic con- 2% of sparteine. The principal behavioral response ditions by using 0.16 M aqueous NaOH at a ßow rate recorded was the arrestment on the Þlter paper. of1mlϫ minϪ1. Calculations of mono- and disaccha- InterspeciÞc comparisons of behavioral responses ride concentrations were made from dry weight ex- toward sparteine were undertaken witha solution of traction residues, and then expressed relative to initial 1% sparteine in 1% acetic acid on Þlter paper in the sample dry weight. The measurement precision was center of a glass container (5 cm diameter at base, 7 cm tested by duplicating analyses of the samples. high), with a 1% acetic acid treatment as a control. For The nitrogen content was evaluated by Kjehldahl eachspecies, we observed 20Ð24 separated individu- analysis (AOAC 1984). als. For eachindividual, we recorded 10 items during Sugar Sensitivity. The experimental apparatus was a 5-min continuous observation. The behaviors were: the same as that described above, with the exception palpation in the air (Pair); palpation of Þlter paper that the length of the test was reduced to 5 min to (Palp); arrestment on the Þlter paper (Arrest); put- June 2003 PICAUD ET AL.: DECISION MAKING FOR FOOD CHOICE BY GRASSHOPPERS 683

Fig. 3. InterspeciÞc comparison of three behaviors of Chorthippus binotatus (bin), C. biguttulus (big), C. albomarginatus (alb), and C. parallelus (par) in binary food plant experiments. (a) Results (mean Ϯ SEM of differences) from Wilcoxon test on behaviors (orientation: 1, palpation: 2, consumption: 3) for plant choice. Kruskal-Wallis tests of all combined data are noted on eachÞgures. Different letters correspond to signiÞcant differences between species (MannÐWhitney U tests, P Ͻ 0.05). Ϯ (b) Mean reaction times SEM (seconds, Log10-transformed data) to move and palpate the plant. ANOVA of the combined ϭ ϭ ϭ data of all species concerning reaction time for ßowered Ulex minor: F3,27 3.62, P 0.026. ANOVA for Lolium perenne: F3,30 17.16, P Ͻ 0.001. Different letters correspond to signiÞcant differences between species (TukeyÕs test, P Ͻ 0.001). ting the fore legs on the Þlter paper and touching the for interspeciÞc comparisons. In eachcase, TukeyÕs mouth parts (Leg); putting the labrum on the Þlter tests were performed for pair-wise comparisons. paper (Lab); biting and masticating (Bmas); leg with- Frequencies of different items withsparteine trials drawing above the Þlter paper (Wleg); regurgitation and coated grass experiments were compared witha (Reg); leaving Þlter paper (Leave); jumping resulting MannÐWhitney U test. from Þlter paper contact (Jump). We tried to enhance grass consumption by C. bino- Results tatus in adding some characteristic molecules present in Ulex minor: we studied its behavior with Lolium Behavioral Comparison among Species. Chorthip- perenne covered, witha brush,withdifferent combi- pus binotatus and C. biguttulus preferred Ulex minor nations of solution diluted in 1% acetic acid: sucrose ßowers (Fig. 3a) to Lolium perenne (Kruskal-Wallis ϭ ϭ Ͻ one M, sparteine 1%. Lolium perenne covered with1% test: H3 18.166, N 88, P 0.001) and Molinia ϭ ϭ ϭ acetic acid were used as control. We recorded the time caerulea (Kruskal-Wallis test: H3 16.30, N 79, P spent palpitating before decision and the frequency of 0.001). When the Ulex minor branches have not ßow- individuals (16Ð22 individuals for eachexperiment) ered yet, the graminivorous species (C. biguttulus, consuming Lolium perenne. C. albomarginatus, and C. parallelus) preferred Grami- Ͻ Ͻ Ͻ Ͻ Statistical Tests. The individual behaviors for each nae (Kruskal-Wallis test: H3 10.96, 74 N 79, P choice were analyzed with Wilcoxon tests. Differ- 0.015 withbothbinary plants choice). ences given by these tests were used to perform The shortest time (Fig. 3b) between the beginning Kruskal-Wallis tests for interspeciÞc comparisons us- of the experiment and the palpation of Ulex minor in ing SYSTAT ver. 7.0 (SPSS 1997). the case of binary choices was observed for C. bino- ϭ ϭ The reaction times were compared using a multi- tatus (ANOVA: F3,27 3.62, P 0.026). C. albomar- way analysis of variance (ANOVA) because all of the ginatus and C. parallelus show the quickest reaction ϭ Ͻ data were normally distributed. Log-transformed pal- time to Lolium perenne (ANOVA: F3,30 17.16, P pation times were compared using multi-way ANOVA 0.001). 684 ENVIRONMENTAL ENTOMOLOGY Vol. 32, no. 3

Fig. 4. InterspeciÞc comparison of three behavioral items of Chorthippus binotatus (bin), C. biguttulus (big), C. albomarginatus (alb), and C. parallelus (par) in an experiment testing the response to different sugars versus water. (a) Mean Ϯ SEM of differences from Wilcoxon test on behaviors (orientation: 1, palpation: 2, consumption: 3). Kruskal-Wallis test of all combined data are noted on eachÞgures. Different letters correspond to signiÞcant differences between species (MannÐ Whitney U tests, P Ͻ 0.05). (b) Mean palpation time Ϯ SEM (seconds, log-transformed data) of soluble carbohydrate solution (1M). ANOVA of the combined data of all species for each sugar experiment: see text. Different letters correspond to signiÞcant differences between species (TukeyÕs test: P Ͻ 0.05).

The most marked preference for sucrose (Fig. 4a) levels in Ulex minor (Ͼ58 mg ⅐ gϪ1 of extract residue). was observed for C. binotatus (Kruskal-Wallis test: Shoots of Ulex minor also contained fructose (mean Ϯ ϭ ϭ ϭ ϭ Ϯ ⅐ Ϫ1 Ϯ ϭ H3 10.49, N 80, P 0.015). This result is corrob- SEM 3 1.4 mg g ) and sucrose (mean SEM orated by the fact that C. binotatus showed the quickest palpation time (Fig. 4b) before making a decision ϭ ϭ in the presence of sucrose (ANOVA: F3,28 3.91, P 0.019). Curiously, C. biguttulus responded slowly to sucrose. As for fructose, the most complete series of behaviors was performed by C. binotatus and C. par- ϭ ϭ Ͻ allelus (Kruskal-Wallis test: H3 7.56, N 82, P 0.056). Unexpectedly, C. biguttulus had the quickest ϭ ϭ reaction time (Fig. 4b, ANOVA: F3,31 3.53, P 0.026) in spite of its low motivation for fructose (Fig. 4a). C. biguttulus is the only species sensitive to glucose ϭ ϭ ϭ (Kruskal-Wallis test: H3 10.23, N 81, P 0.017), but no difference between decision time was recorded ϭ ϭ (ANOVA: F3,29 0.77, P 0.52). Finally, C. binotatus was the only species attracted Fig. 5. InterspeciÞc comparison of orientation behavior (Fig. 5) by branching aspects (Kruskal-Wallis test: of Chorthippus binotatus (bin), C. biguttulus (big), C. albo- H ϭ 13.41, N ϭ 82, P ϭ 0.004). marginatus (alb), and C. parallelus (par) in a Y-maze. Results 3 (mean Ϯ SEM of the differences) of Wilcoxon test on ori- Soluble Carbohydrates and Nitrogen Contents. entation behaviors an experiment comparing different plant Flowers of Ulex minor had high levels of sucrose architecture. Kruskal-Wallis test of all combined data are Ϯ ϭ Ϯ ϫ Ϫ1 (mean SEM 18.9 0.7 mg g of extract noted on eachÞgures. Different letters correspond to sig- residue) in the autumn but not summer (Table 1). niÞcant differences between species (Mann-Whitney U tests, Glucose and fructose were bothpresent at very high P Ͻ 0.05). June 2003 PICAUD ET AL.: DECISION MAKING FOR FOOD CHOICE BY GRASSHOPPERS 685

Table 1. Determination of soluble carbohydrates by HPAEC-PAD analyses and nitrogen content by Kjehldahl analysis. Sugars and nitrogen (mean ؎ SEM) are expressed in mg.g؊1 for dry weight. Analyses were performed on 500-mg plant tissues of 5 samples of Ulex minor and 4 samples of each grass

Glucose Fructose Sucrose Nitrogen Flowers of Ulex minor (end of July) 58.3 Ϯ 0.9 67.7 Ϯ 6.7 3.5 Ϯ 0.2 5.16 Ϯ 0.5 Flowers of Ulex minor (September) 69.1 Ϯ 0.5 77.4 Ϯ 0.9 18.9 Ϯ 0.7 No data Stems of Ulex minor 5.6 Ϯ 2.1 2.6 Ϯ 0.4 8 Ϯ 0.8 9.6 Ϯ 1.2 Shoots of Ulex minor 3.95 Ϯ 0.8 3 Ϯ 1.4 6.6 Ϯ 0.8 9.8 Ϯ 0.6 Leaves of Agrostis tenuis 7 Ϯ 2.5 8.4 Ϯ 4.1 9.6 Ϯ 0.9 No data Leaves of Holcus lanatus 4.2 Ϯ 1.7 3.6 Ϯ 2 8.7 Ϯ 1 No data

6.6 Ϯ 0.8 mg ⅐ gϪ1), representing Ͻ10% of the quantity tation. It should be noticed that except one individual present in ßowers. The two species of grasses con- of C. albomarginatus, arrestment and leaving Þlter tained sugars in comparable amounts to stems of paper were never associated. C. binotatus differed Ulex minor (Table 1). Ulex minor ßowers were low in from remaining species by three behaviors when nitrogen content (0.5Ð0.6% of dry weight). Nitrogen sparteine is dropped on Þlter paper. C. binotatus pal- made up Ϸ1% of the dry weight of shoots and stems pated the air more than the other species (Kruskal- ϭ ϭ Ͻ of Ulex minor. Wallis test: H3 18.94, N 57, P 0.001). Arrestment Sparteine Sensitivity. At concentrations between on the piece of Þlter paper characterized C. binotatus ϭ ϭ ϭ 0.5 and 2% sparteine, 25% of C. binotatus individuals (Kruskal-Wallis test: H3 16.27, N 57, P 0.001). were arrested by the piece of Þlter paper. For weaker In the same way, C. binotatus was the only species ϭ concentrations, the percentages of responses clearly never leaving Þlter paper (Kruskal-Wallis test: H3 fall (12% of individuals for 0.1% sparteine). Conse- 19.03, N ϭ 57, P Ͻ 0.001). quently, the sparteine concentration used in interspe- Otherwise, regurgitation characterized C. albomar- ciÞc experiments was 1%. ginatus compared withotherspecies (Kruskal-Wallis ϭ ϭ Ͻ Among the 10 behaviors considered, four were of test: H3 21.12, N 57, P 0.001). special interest because of statistical differences be- Within each species, we tested to determine tween species (Fig. 6): air palpation, arrestment on the whether there was a signiÞcant difference between piece of Þlter paper, leaving Þlter paper and regurgi- the frequencies of each behavior in response to acetic

Fig. 6. Frequency of different behaviors in response to 1% sparteine diluted in 1% acetic acid on Þlter paper (shaded boxes) or to 1% acetic acid alone (open boxes). Asterisks indicate signiÞcantly different values from all other species tested (Kruskal-Wallis test, P Յ 0.001 in eachcase). Thecircles indicate signiÞcant differences between frequencies of response to sparteine and the acetic acid control (Mann-Whitney U test, P Ͻ 5%) in eachcase), whenbothfrequencies are not null. In the case of acetic acid alone, the behavior “no-response” is not represented. Pair: palpation of air; Palp: palpation of Þlter paper; Arrest: arrestment on Þlter paper; Leg: placing fore legs on Þlter paper and touchingwithmouthparts; Lab: touching Þlter paper withlabrum; Bmas: biting and masticating; Wleg: withdrawingleg above Þlter paper; Reg: regurgitation; Leave: leaving Þlter paper; Jump: jumping in response to contact withÞlter paper. 686 ENVIRONMENTAL ENTOMOLOGY Vol. 32, no. 3 acid versus sparteine diluted in acetic acid. The responses of C. binotatus are generally stronger when sparteine is present (MannÐWhitney U test, P Ͻ 0.01 in each case), except for biting behavior. The behaviors related to avoidance (leaving Þlter paper, jumping, leg withdrawing, regurgitation) are generally expressed in graminivorous species when sparteine is present. Curiously, C. parallelus is the only graminivorous species palpitating Þlter paper withsparteine at a higher frequency than acetic acid (MannÐWhitney U test, P Ͻ 0.05). Responses of C. binotatus toward Lolium perenne Coated with Different Combinations of Solutions. The time spent on palpation of L. perenne before decision (consumption or rejection) is shown in Fig. 7a. The shortest time, i.e., the quickest decision cor- responded to the combination of sucrose and sparteine. This time differs signiÞcantly from the one recorded with L. perenne coated withacetic acid only ϭ ϭ (ANOVA: F1,26 5.97, P 0.02). Consumption frequency of Lolium perenne by C. binotatus (Fig. 7b) was enhanced by the same combination of sparteine ϩ sucrose (Kruskal-Wallis ϭ ϭ ϭ test: H3 8.24, N 69, P 0.04). The effect of addition of sucrose or sparteine alone did not signiÞcantly differ from the test.

Discussion Analysis of Food Choice. Chorthippus binotatus pre- fers Ulex minor, especially ßowers. Both C. albomarginatus and C. parallelus prefer Graminae to Ulex, whether blooming or not. Our experiments lead us to deÞne three feeding groups. The Þrst one is represented by C. binotatus, which specializes on Ulex mi- nor, and the second one comprises C. albomarginatus and C. parallelus, which are “purely graminivorous.” Fig. 7. Behavior of Chorthippus binotatus in response to The third group consists of C. biguttulus, which be- Lolium perenne covered withdifferent combinations of so- haves like the purely graminivorous species when Ulex lutions diluted in acetic acid 1% (sucrose: 1 M, sparteine: 1%). is not ßowering, but like the specialist species on (a) Mean time Ϯ SEM spent (seconds, log-transformed data) ϭ Ͻ Genisteae when Ulex is ßowering. We can call this in palpation (global ANOVA: F3,52 3.41, P 0.05). Different species “atypically graminivorous.” Thus, both C. bino- letters correspond to signiÞcant differences between treat- Ͻ tatus and C. biguttulus respond to Ulex minor ßowers, ments (TukeyÕs test: P 0.05). (b) Consumption frequency. Combined data of eachcombination are compared with which probably contain one or more attractants. ϭ ϭ Ͻ Kruskal-Wallis test: H3 8.24, N 69, P 0.05). Different These observations are conÞrmed by measuring letters correspond to signiÞcant different frequencies be- time to decision making. The two purely graminivo- tween treatments (Mann-Whitney U tests, P Ͻ 0.05). rous species have the quickest reaction time for Graminae and the longest one for Ulex minor, whereas the specialist species on Genisteae shows the opposite C. parallelus. Chorthippus biguttulus is most sensitive behavior. Curiously, the “atypical graminivorous spe- to glucose. As a result, sucrose and fructose are prob- cies” (C. biguttulus) is slow to make a decision, re- ably the major feeding stimulants, among sugars, on gardless of its choice (Ulex minor or Graminae) and Ulex minor. The sugar analyses show that Ulex minor has the broadest diet, accepting Ulex as well as grasses. ßowers in September had the highest content of sug- This agrees with the Þndings of Bernays (1998, 2001) ars. However, the attraction of C. binotatus to its host that the broader the food types accepted, the slower plant in July is not explained by sugar content. the response (assuming grasses to be rather uniform in Presence of receptors that can detect sugars has secondary chemical compounds). This disadvantage is been demonstrated in numerous insects (Mitchell somewhat balanced by the fact that C. biguttulus has 1981) and especially in grasshoppers (Bernays and a larger habitat niche than either of the other species Chapman 1994): the pyranose site is maximally sen- (Bellmann and Luquet 1993). sitive to sucrose and glucose and the furanose site to Chorthippus binotatus is the species most respon- fructose. Feeding on Ulex minor ßowers is partly ex- sive to sucrose, and it is as responsive to fructose as plained by sensitivity of C. binotatus to sucrose and of June 2003 PICAUD ET AL.: DECISION MAKING FOR FOOD CHOICE BY GRASSHOPPERS 687

C. biguttulus to glucose. Fructose does not play any should be an increase of Þtness because of the pres- role as the “strictly graminivorous” species are not ence of carbohydrates. We hypothesize that the sugars attracted by Ulex minor ßowers. present in ßowers favor the reproductive investment Sparteine alone leads to different behavioral re- that occurs from the end of August until October, i.e., sponses according to the species. Leaving the sub- the main ßowering season. strate containing sparteine is a response characteristic Graminivorous species have no constraint for dis- for graminivorous species. C. albomarginatus regurgi- persal as grasslands and hedges are connected. The tates in response to sparteine. Contrary to the other occurrence of C. binotatus, however, is limited by the species, C. binotatus palpates the air and remains on patchy distribution of Ulex minor heathlands. How- the substrate. Thus, sparteine is at least involved in the ever, we showed that this specialist species can feed on food selection of the insect species enabling it to feed Poaceae when there is no other available resource. on Genisteae. This situation can be compared with the Gue´guen et al. (1975) found Poaceae phytolithes in phagostimulant power of pyrrolizidine alkaloids to C. binotatus faeces in Brittanny. It should be pointed Zonocerus elegans Thunberg (Boppre´ et al. 1984), al- out that when the species specializing on Genisteae though it is a polyphagous and opportunistic species. feeds on Poaceae, its vigilance level probably de- Our study on C. binotatus with Lolium perenne creases, leading to a predation risk (Dukas 1998). coated with different solutions showed that the com- Thus, population migrations between relatively close bination of sparteine and sucrose is probably involved heathlands are possible but certainly infrequent. It in food choice of C. binotatus. This is in agreement would be interesting to know the percentage of indi- with our results that the highest consumption level viduals able to complete their life cycle with Poaceae was recorded in presence of the same combination, as the only food source. Our preliminary experiments demonstrating its role as a phagostimulant. (Picaud et al. 2002) suggest that this percentage is very With respect to the plant architecture, we found low at best. If a few generations could develop on that C. binotatus was attracted to the branchy drawing grass, it would allow the species to colonize distant over the straight one. This can partly explain its pref- heathlands islands. Genetic distances between popu- erence to bushes versus grasses. However, plant ar- lations should be measured to estimate gene ßow to chitecture cannot be the only explanation for the address whether these populations are part of one or feeding behavior of this species, because it is never several metapopulations, and thus the extinction risk attracted to other ßowering bushes in heathlands, e.g., of eachpopulation. Calluna vulgaris, Erica tetralix, and Ulex europaeus. Contrary to the results of Williams (1954) and Bernays and Chapman (1970), graminivorous species were not Acknowledgments signiÞcantly attracted to a given architecture in our We are grateful to Cećile Nouhaut for her help in trans- experiments. These authors used only three-dimen- lation, and A. Bellet, S. Verot, and I. Vannier for their tech- sional lures and never drawings. nical assistance. Benefits and Disadvantages of Being a Specialist on Fabaceae Bushes. Chorthippus binotatus has evolved from graminivorous species. We hypothesize that the References Cited shift could have been the consequence of an evolved Abrams, P. A., and O. J. Schmitz. 1999. The effect of risk of sensibility to sparteine (and/or other quinolizidine mortality on the foraging behaviour of animals faced with alkaloids) in the shoots and not to Genisteae ßowers time and digestive capacity constraints. Evol. Ecol. Res. 1: because (1) young instars of the subspecies C. b. bino- 285Ð301. tatus are never exposed to ßowers, whereas (2) the [AOAC] Association of Official Agricultural Chemist. 1984. adults of subspecies C. b. moralesii feed on bothgrasses OfÞcial Methods of Analysis. A.O.A.C., Washington, DC. and nonßowering Genisteae (Picaud et al. 1999). Bellmann, H., and G. Luquet. 1993. Guide des Sauterelles, However, the diet shift within the genus Chorthippus Grillons et Criquets dÕEurope occidentale. Delachaux et has ecological consequences on nutrient balance, dis- Niestle´, Lausanne. persal, and possibly predation risk. Bernays, E. A. 1998. The value of being a resource specialist: behavioral support for a neural hypothesis. Am. Nat. 151: Grasses growing on richsoils in spring contain as 451Ð464. muchnitrogen as Genisteae shootsand stems. How- Bernays, E. A. 2001. Neural limitation in phytophagous in- ever, nitrogen concentration of grasses decreases with sects: implications for diet breathand evolution of host plant maturity (Chapuis and Lefeuvre 1980). Finally, afÞliation. Annu. Rev. Entomol. 46: 703Ð727. Genisteae in the autumn are richer in nitrogen than Bernays, E. A., and K. L. Bright. 2001. Food choice causes grasses growing on the poor soils of heathlands (Roze´ interrupted feeding in the generalist grasshopper Schis- 1980, Glyphis and Puttick 1989). Consequently, feed- tocerca americana: further evidence for inefÞcient deci- ing on Genisteae represents a gain in terms of nitrogen sion-making. J. Insect Physiol. 47: 63Ð41. supply, at least when Ulex minor shoots and stems are Bernays, E. A., and R. F. Chapman. 1970. Experiments to determine the basis of food selection by Chorthippus consumed. However, when the specialist species shifts parallelus (Zetterstedt) (Orthoptera, Acrididae) in the its diet toward U. minor ßowers, we assume that the Þeld. J. Anim. Ecol. 39: 761Ð776. nitrogen supply falls. If compensatory feeding occurs, Bernays, E. A., and R. F. Chapman. 1994. Host plant selec- C. binotatus should eat twice as much on ßowers as tion by phytophagous insects. Chapman & Hall, New shoots. Nevertheless, the major advantage of this shift York. 688 ENVIRONMENTAL ENTOMOLOGY Vol. 32, no. 3

Bonnet, E. 1996. Phylogeńie et partage des ressources chez et des Cype´races. 2. Application a` lÕe´tude du re´gime les Orthopte`res Acridiens. The`se doctorat, Universite´ de alimentaire des Orthopte`res Acridiens. C.R.A.S., Paris 281 Limoges, Limoges, France. (D): 929Ð932. Boppre´, M., U. Seibt, and W. Wickler. 1984. Pharmacoph- Harisson, J., J. A. Gallagher, and C. J. Pollock. 1997. A simple agy in grasshoppers? Entomol. Exp. Appl. 35: 115Ð117. and rapid method for the analysis of water-soluble car- Brugel, E., L. Brunerye, and A. Vilks. 2001. Plantes et ve´g- bohydrates from small segments of cereal leaf tissue. e´tation en Limousin. Atlas de la ßore vasculaire. Conser- J. Plant Physiol. 151: 654Ð659. vatoire Re´gional des Espaces Naturels, St-Gence, France. Kruseman, G. 1982. Mate´riaux pour la faunistique des Or- Bruneton, J. 1999. Pharmacognosie, Phytochimie, Plantes thopte`res de France. Fasc II. Les acridiens des Museés de me´dicinales. Techniques et Documentation, Paris. Paris et dÕAmsterdam. Instituut voor taxonomische zo- Cataldi T. R., Campa, C., Casella, I. G., and S. A. Bufo. 1999. o¨logie (Zoo¨logish Museum), Amsterdam, The Nether- Determination of maltitol, isomaltitol, and lactitol by lands. high-pH anion-exchange chromatography with pulsed Mitchell, R. 1981. Insect behaviour, resource exploitation, amperometric detection. J. Agric. Food Chem. 47: 157Ð and Þtness. Annu. Rev. Entomol. 26: 373Ð396. 163. Mossot, M., and D. Petit. 2000. Pre´sence de Chorthippus Chapman, R. F. 1982. Chemoreception: the signiÞcance of binotatus binotatus (Charpentier, 1825) dans le Vaucluse receptor numbers. Adv. Insect Physiol. 16: 247Ð356. (Orth., Acrididae). Bull. Soc. Entomol. Fr. 105: 251Ð252. Chapman, R. F., and A. Joern. 1990. Biology of grasshop- Otte, D., and A. Joern. 1977. On feeding pattern in desert pers. Wiley, New York. grasshoppers and the evolution of specialized diets. Proc. Chapman, R. F., E. A. Bernays, and T. Wyatt. 1988. Chem- Natl. Acad. Sci. 128: 89Ð126. ical aspect of host-plant speciÞcity in three Larrea-feed- Picaud, F., C. Nanthieras, I. Coudon, A. Bellet, and D. Petit. ing grasshoppers. J. Chem. Ecol. 14: 561Ð579. 1999. Graminivory abandonment gradation among Chapuis J. L., and J. C. Lefeuvre. 1980. Evolution saison- two subspecies of Chorthippus binotatus (Acrididae, nie`re du re´gime alimentaire du Lapin de Garenne, Oryc- Gomphocerinae): descriptive and experimental ap- tolagus cuniculus (L.), en lande: re´sultats de deux ans proaches. Ann. Soc. Entomol. Fr. 35: 242Ð247. dÕanalyses. Bull. Ecol. 11: 587Ð597. Picaud, F., Gloaguen, V. and D. Petit. 2002. Mechanistic Defaut, B. 1999. Synopsis des Orthopte`res de France. As- aspects of feeding preference in Chorthippus binotatus sociation pour la Caracte´risation et lÕe´tude des Entomo- (Acrididae, Gomphocerinae). J. Insect Behav. 15: 513Ð ceńoses. Be´deilhac, France. 526. Dukas, R. 1998. Constraints on information processing and Roze´, F. 1980. Re´partition et e´volution de lÕazote dans their effects on behaviour, pp. 89Ð128. In R. Dukas [ed.], quelques landes bretonnes. Bull. Ecol. 11: 315Ð321. Cognitive ecology: The evolutionary ecology of informa- SPSS Inc. 1997. SYSTAT ver. 7, The system for statistics. tion processing and decision making. Chicago University SysStat Inc., Evanston, IL. Press, Chicago, IL. Uvarov, B. P. 1977. Grasshoppers and locustsÑA handbook Gangwere, S. K., and E. Morales-Agacino. 1973. Food selec- of general acridology, vol. 2. Centre for Overseas Pest tion and feeding behaviour in iberian Orthopteroidea. Research, London, England. An. Inst. Nacl. Inv. Agr., Serie Pr. Veg. 3: 251Ð337. White, P. R., and R. F. Chapman. 1990. Olfactory sensitivity Glyphis, J. P., and G. M. Puttick. 1989. Phenolics, nutrition of gomphocerine grasshoppers to the odours of host and and insect herbivory in some garrigue and maquis plant non-host plants. Entomol. Exp. Appl. 55: 205Ð212. species. Oecologia (Berl.) 78: 259Ð263. Williams, L. H. 1954. The feeding habits and food prefer- Grasse´, P. P. 1929. Etudes ećologiques et biogeógraphiques ences of Acrididae and the factors which determine them. sur les Orthopte`res français. Bull. Biol. Entomol. Fr. Belg. Trans. Royal Entomol. Soc. Lond. 105: 423Ð454. 63: 489Ð539. Gue´guen, A., A. Chenouvrier, and J. C. Lefeuvre. 1975. Es- sai de de´termination du re´gime alimentaire dÕanimaux Received for publication 1 April 2002; accepted 18 February herbivores a` lÕaide des phytolithes siliceux des Gramineés 2003.