ECOLOGY AND POPULATION BIOLOGY Shoot Characteristics Affect Oviposition Preference of the Willow Spittlebug Aphrophora pectoralis (Homoptera: )

1 2 AKIYOSHI NOZAWA AND TAKAYUKI OHGUSHI

Ann. Entomol. Soc. Am. 95(5): 552Ð557 (2002) ABSTRACT The relationships of shoot abscission, shoot toughness, and egg parasitism with shoot lengthwere investigated in relation to oviposition preference by thewillow spittlebug Aphrophora pectoralis Matsumura on Salix miyabeana, S. sachalinensis, and S. integra. Spittlebugs strongly preferred shoots Ͼ20 cm long even though this size class was a small proportion of all shoots. Oviposition increased with shoot length up to 40 cm but leveled off, or even decreased, on longer shoots. Shoot abscission rate decreased and toughness increased with increasing shoot length, but egg parasitism was not affected. Spittlebug oviposition preference is therefore likely to be affected by shoot abscission and shoot toughness.

KEY WORDS Salix species, shoot abscission, shoot toughness, egg parasitism, xylem sucker

OVIPOSITION PREFERENCE IS FUNDAMENTAL to understand- tree during winter. Because A. pectoralis eggs over- ing herbivore population dynamics (Ohgushi 1992, winter inside shoots, shoot abscission results in egg Price et al. 1995) and host plant selection (Thompson death. Spittlebug nymphs of A. pectoralis penetrate and Pellmyr 1991, Mayhew 1997). The patterns of shoots with their stylets to feed, and feeding sites are herbivore attack on host plants have been explained restricted by stem tissue hardness in the spittlebug by plant traits (e.g., age, growthrate, genotypes, re- Philaenus spumarius L. (Hoffman and McEvoy 1986, source quality, and chemical compounds) (Tahvanainen McEvoy 1986). Therefore, shoot abscission and tough- et al. 1985, Price et al. 1987a, Brodbeck et al. 1990, Woods ness may have signiÞcant roles in determining the et al. 1996, Craig et al. 1999, Dorn et al. 2001). oviposition preference for long shoots. Recently, Price (1991) proposed the plant vigor Natural enemies are another important factor de- hypothesis that herbivorous attack vigorous termining oviposition preference of herbivorous in- plants or plant modules more frequently than slower sects (enemy-free space hypothesis, see Thompson growing plants or plant modules. Many studies testing 1988). For example, the mistletoe butterßy Ogyris the plant vigor hypothesis have focused on galling amaryllis Hewitson has an oviposition preference for insects (Price et al. 1995, Price et al. 1997). For ex- plants of the genus Amyema tended by ants, even ample, the galling sawßy Euura mucronata Hartig ex- though the plants are low quality for larvae (Atsatt hibited a strong oviposition preference for longer, 1981). This preference apparently evolved because more vigorous shoots on which larval survival was high the presence of ants on the plants decreases egg par- due to high food quality in rapidly growing shoots asitism by a parasitoid, Centrodora sp. Although two (Price et al. 1987b). However, little attention has been egg parasitoids of the family Aphelinidae and Mymari- given to other herbivorous insects. dae were found in A. pectoralis (Craig and Ohgushi The spittlebug Aphrophora pectoralis Matsumura is 2002), the effects of egg parasitism on oviposition a xylem-sucking uses Salix species as host plants preference have not been investigated. (Komatsu 1997). Craig and Ohgushi (2002) demon- The purpose of this study was to determine the strated that ovipositing females of this species have a effects of shoot toughness, shoot abscission, and egg strong preference for long shoots and that oviposition parasitism on oviposition preference for long shoots in preference is correlated withlarval survival. These A. pectoralis. We investigated the relationship be- results suggest that larval survival is an important fac- tween shoot length and proportion of shoots in which tor in determining oviposition preference of the spittlebug oviposition occurred, the effects of shoot spittlebug. However, the question of what plant traits length on shoot abscission and shoot toughness, and determine the oviposition preference for long shoots the relationship between shoot length and rate of egg remains unresolved. Willow shoots often drop off a parasitism.

1 Institute of Low Temperature Science, Hokkaido University, Sap- Materials and Methods poro, 060Ð0819 Japan, e-mail: [email protected] 2 Center for Ecological Research, Kyoto University, Otsu, 520Ð2113 Study Site. The study site was located on a bank of Japan. the Ishikari River in Hokkaido, northern Japan (43Њ

0013-8746/02/0552Ð0557$02.00/0 ᭧ 2002 Entomological Society of America September 2002 NOZAWA: AND OHGUSHI:OVIPOSITION PREFERENCE OF A. pectoralis 553

11Ј,N,141Њ 24Ј, E) that has been disturbed by dyke recorded the pressure in Newtons (N) required for building. Around the study area, six willow species, S. complete penetration in the apical region. Three mea- miyabeana Seemen, S. sachalinensis Fr. Schm., S. in- surements were made per shoot, and the average value tegra Thunb., S. hultenii Floderus, S. subfragilis was used as an index of shoot toughness. Anders., and S. pet-susu Kimura, occur sympatrically Egg Parasitism. In the study area, one egg parasitoid (Ishihara et al. 1999). Our study included three of of the family Aphelinidae and one of the Mymaridae these: S. miyabeana, S. sachalinensis, and S. integra, all were found (Craig and Ohgushi 2002). In early April of which were abundant at the study site. 1999, we randomly collected 30 shoots (1 yr old) These three species of willows are widely distrib- bearing eggs from S. miyabeana, 34 from S. sachalin- uted in Hokkaido and are dioecious, deciduous trees ensis, and 33 from S. integra. These shoots were kept often occurring on riverbanks. S. miyabeana grows up individually in a small plastic case withmoderate hu- to7m,S. sachalinensis up to 15 m, and S. integra up to midity. We recorded the number of emerged parasi- 1.5 m and in Ishikari ßower from mid-April to early toids and nymphs until late July. In early August, we May. After ßowering, budbreak occurs in May, and dissected these shoots and counted immature parasit- leaf expansion and shoot elongation continue until ized and nonparasitized eggs. It is possible to distin- mid-July. Most leaves become senescent by late Oc- guishbetween theparasitized and nonparasitized tober when new buds appear in leaf axils. eggs, because parasitized eggs are black and brown in Insect Life Cycle. The spittlebug A. pectoralis is a color and nonparasitized eggs are white. We calcu- xylem-sucking insect that utilizes Salicaceae as host lated the rate of egg parasitism as the proportion of plants (Komatsu 1997). It has a univoltine life cycle parasitized eggs to all eggs observed. and overwinters in the egg stage. Overwintered eggs Statistical Procedures. Chi-square goodness-of-Þt begin to hatch in early May, and nymphs produce analysis was used to determine whether distribution of spittle in which they live and feed. Adults begin to spittlebug oviposition in the Þeld was random with emerge in late June. Mating and oviposition occur respect to shoot length (Zar 1999). The Spearman from early August to mid-October. Females lay egg rank correlation was used to determine the relation- masses (Ϸ10 eggs per mass) inside the apex of current- ship between shoot abscission rate and shoot length. year growing shoots. The apical region of the shoots, Shoot toughness and egg parasitism among shoot within which generally oviposition occurs, die within lengthclasses were compared using a Kruskal-Wallis a week because of mechanical damage by the femaleÕs test followed by nonparametric Tukey-type multiple ovipositor. comparisons for each of the three willow species (Zar In this study, we counted each oviposition scar as an 1999). Shoots were assigned to Þve classes according egg mass. Although we did not compare the number to length. When compared among willow species, data of eggs per egg mass among three willow species, no from all shoot length classes were pooled before anal- signiÞcant difference was found in the length of ovi- ysis. position scars (mm) (S. miyabeana: 3.66 Ϯ 0.07 [mean Ϯ SE], n ϭ 40; S. sachalinensis: 3.57 Ϯ 0.06, n ϭ Results 40; S. integra: 3.67 Ϯ 0.06, n ϭ 40; Kruskal-Wallis test; df ϭ 2, H ϭ 1.1, P ϭ 0.59). The number of eggs was Distribution Pattern of Spittlebug Oviposition. In highly correlated with the length of oviposition scars three willow species, 80% of shoots were Ͻ20 cm (Fig. (Nozawa and Ohgushi 2002). This suggests that there 1). Nevertheless spittlebug females strongly preferred was no difference in the number of eggs per egg mass ovipositing in long (Ͼ20 cm) shoots (S. miyabeana: among three willow species. ␹2 ϭ 19.3, df ϭ 2, P Ͻ 0.001; S. sachalinensis. ␹2 ϭ 91.2, Measurement of Shoot Characteristics. In early May df ϭ 3, P Ͻ 0.001; S. integra: ␹2 ϭ 98.9, df ϭ 3, P Ͻ 1998, we randomly selected 117 shoots (1 yr old) from 0.001). The proportion of shoots with spittlebug eggs S. miyabeana, 107 from S. sachalinensis, and 96 from S. increased withshootlengthup to 40 cm. However, the integra. In early September 1998, we measured the incidence of eggs decreased again on Ͼ40 cm shoots length of all current-year shoots on the selected one- in S. miyabeana and leveled off in S. sachalinensis and year old shoots. In late September 1998, we recorded S. integra. Shoot Abscission. Shoot abscission rate de- the number of current-year shoots with oviposition on clined withincreasing shootlengthin threewillow ϭϪ ϭ each of the selected 1-yr-old shoots. To determine species (S. miyabeana: rS 0.93, P 0.05; S. sacha- ϭϪ ϭ ϭϪ ϭ abscission rate of the current-year shoots during win- linensis: rS 0.93, P 0.05; S. integra: rS 0.93, P ter, we recorded the number of the current-year 0.05; Fig. 2). More than 60% of Ͻ10 cm shoots dropped shoots in late April 1999. The abscission rate was ob- off during winter, while no more than 30 cm shoots tained as the proportion of current-year shoots that did. had dropped off during winter. Shoot Toughness. We found signiÞcant differences To determine shoot toughness, we selected 35 cur- in shoot toughness among shoot length classes in the rent-year shoots from each willow species in early three willow species (S. miyabeana, H ϭ 6.95, df ϭ 2, August 1999 when oviposition was occurring. We P ϭ 0.03; S. sachalinensis, H ϭ 7.4, df ϭ 2, P ϭ 0.02; S. chose Ͼ20 cm shoots on which most of oviposition integra, H ϭ 11.29, df ϭ 2, P ϭ 0.004; Fig. 3). Shoot scars were found. Shoot toughness was measured by toughness of three willow species tends to increase texture analyzer TA-XT2 (Stable Micro Systems, Lon- with increasing shoot length. Shoot toughness (N) in don) witha needle penetrometer (model P/2N). We S. sachalinensis was greater than S. miyabeana and S. 554 ANNALS OF THE ENTOMOLOGICAL SOCIETY OF AMERICA Vol. 95, no. 5

Fig. 1. Percentage of current-year shoots (open circle) and current-year shoots with oviposition (solid circle) in each Fig. 2. Percentage of shoot abscission during winter in shoot lengh class (S. miyabeana, n ϭ 648, 19 for total shoots eachshootlengthclass. and oviposited shoots, respectively; S. sachalinensis, n ϭ 659, 40; S. integra, n ϭ 992, 35). willow species. Although 80% of shoots were Ͻ20 cm, the percentage of shoots that received oviposition was integra (S. miyabeana, 4.56 Ϯ 0.31 [mean Ϯ SE[, n ϭ highest in 30Ð40 cm shoots, which were only 6% of all 35; S. sachalinensis, 6.54 Ϯ 0.42, n ϭ 35; S. integra, 5.69 Ϯ shoots. However, oviposition rate leveled off or even 0.30, n ϭ 35; H ϭ 14.29, df ϭ 2, P ϭ 0.0008). decreased in the longest (Ͼ40 cm) shoots. Shoot ab- Egg Parasitism. In S. miyabeana and S. sachalinensis, scission rate decreased consistently as shoot length there were no signiÞcant differences in egg parasitism increased in all three willow species. In contrast, shoot among shoot length classes (S. miyabeana, H ϭ 3.52, df toughness increased as shoot length increased. Egg ϭ 4, P ϭ 0.48; S. sachalinensis, H ϭ 6.87, df ϭ 4, P ϭ 0.14; parasitism was not related to shoot length in three Fig. 4.). In S. integra, egg parasitism in 20Ð40 cm shoots willows, though the percentage of parasitized eggs in was signiÞcantly higher than those in other shoot 20Ð40 cm shoots of S. integra was signiÞcantly higher lengthclasses ( H ϭ 10.10, df ϭ 4, P ϭ 0.04; Fig. 4.). No than those in other shoot length classes. signiÞcant differences were found in egg parasitism This study illustrated that oviposition preference of among willow species (S. miyabeana, 17.1 Ϯ 1.8 the spittlebug is most likely affected by shoot abscis- [mean Ϯ SE], n ϭ 30; S. sachalinensis, 14.9 Ϯ 2.0, n ϭ sion. It is adaptive for females to avoid laying in short 34; S. integra, 20.2 Ϯ 2.9, n ϭ 33; H ϭ 14.9 Ϯ 2.0, df ϭ shoots because of the high abscission rate of short 2, P ϭ 0.34). shoots. As A. pectoralis eggs overwinter in the shoots, when females oviposit in short shoots, eggs are sub- jected to heavy morality due to shoot abscission. This Discussion effect of shoot abscission on oviposition preference We found that females of A. pectoralis exhibited a has also been shown in the galling sawßy Euura lasio- strong oviposition preference for long shoots in three lepis Smith, which has an oviposition preference for September 2002 NOZAWA: AND OHGUSHI:OVIPOSITION PREFERENCE OF A. pectoralis 555

Fig. 3. Shoot toughness at apical region in each shoot ϭ lengthclass. Mean and SE are presented ( n 35 for each Fig. 4. Egg parasitism by the two parasitic wasps in each willow species). Mean values withdifferent letters are sig- shoot length class. Mean and SE are presented (S. miyabeana, niÞcantly different at 5% level in eachwillow species. n ϭ 30; S. sachalinensis, n ϭ 34; S. integra, n ϭ 33). Mean values withdifferent letters are signiÞcantly different at 5% level in eachwillow species. long shoots (Craig et al. 1989). Oviposition preference of the sawßy for long shoots reduces the risk of larval mortality due to shoot abscission. Similarly, oviposit- Similarly, some studies have documented that ing females of the leafminer Stilbosis quadricustatella chemical compounds also have an important impact Chambers chose individual trees of the oak Quercus on the oviposition preference of herbivorous insects geminata Small withlow rates of leaf abscission (Stil- (Zucker 1982, Rowell-Rahier 1984, Tahvanainen et al. ing et al. 1991). 1985, Kolehmainen et al. 1995). For example, a shoot In this study, we did not investigate the possibility galling sawßy Euura amerinae Linnaeus, which has an of nutritional differences among shoot length classes. oviposition preference for long shoots, recognizes However, nitrogen concentration may have an impor- shoot size by the concentration of phenolic glucoside, tant effect on the larval performances of herbivorous whichincreaseswithshootlength(Kolehmainenet al. insects, suchas survival rate, growthrate, and fecun- 1995). Also, a leaf-galling aphid Pemphigus betae dity (Scriber and Slansky 1981). In particular, xylem Doane that prefers larger leaf for oviposition may sap has low concentration in nitrogen compared with recognize leaf size by phenolic concentrations that other plant tissues (Mattson 1980), and some xylem decrease withleaf size (Zucker 1982). feeders use plant quality as a cue to plant choice Although some galling insects have an oviposition (Prestidge and McNeill 1983, Brodbeck et al. 1990, preference for the longest (most vigorous) shoots Thompson 1994, Rossi et al. 1996). For example, the (Price et al. 1995, Price et al. 1997), A. pectoralis did not oligophagous xylem-feeding leafhopper Carneo- prefer the longest shoots. This may be explained by the cephala floridana Ball prefers the salt-marsh cord grass, fact that shoot toughness increases with shoot length. Spartina alterniflora Loiseleur-Deslongchamps, with Toughness of plant tissues often has negative effects higher nitrogen content (Rossi and Strong 1991). on herbivorous insects (Hoffman and McEvoy 1986, 556 ANNALS OF THE ENTOMOLOGICAL SOCIETY OF AMERICA Vol. 95, no. 5

Craig et al. 1990, Scheirs et al. 2001). For example, Japanese Ministry of Education, Science, Sports, and Culture feeding sites of the meadow spittlebug Philaenus spu- Grant-in-Aid for Creative Basic Researchto T.O. marius nymphs were restricted to the apexes of the stems of pearly everlasting Anaphalis margaritacea Benth& Hook. Because stem toughnessincreases with References Cited distance from the apical region, the nymphs could not Atsatt, P. R. 1981. Ant-dependent food plant selection by penetrate the lower parts of stems with their stylets the mistletoe butterßy Ogyris amaryllis (Lycaenidae). (Hoffman and McEvoy 1986). Oecologia 48: 60Ð63. In S. miyabeana, oviposition preference declined Brodbeck, B. V., R. F. Mizell, W. J. French, P. C. Andersen, dramatically in Ͼ40 cm shoots with increased shoot and J. H. Aldrich. 1990. Amino acids as determinants of toughness. However, despite the increased toughness host preference for the xylem feeding leafhopper, Ho- of S. sachalinensis compared with the other two wil- malodisca coagulata (Homoptera: Cicadellidae). Oecolo- Ͼ gia 83: 338Ð345. low species, oviposition preference for 40 cm shoots Craig, T. P., and T. Ohgushi. 2002. Preference and perfor- did not decline. Egg mass density on S. sachalinensis mance are correlated in the spittlebug, Aphrophora pec- was signiÞcantly higher than on S. miyabeana or S. toralis (Homoptera: Cercopoidea) on four species of wil- integra (Craig and Ohgushi 2002), suggesting that lows. Ecol. Entomol. (in press). there are unknown beneÞts (e.g., higher resource Craig, T. P., J. K. Itami, and P. W. Price. 1989. A strong quality, escaping from natural enemies) from ovipos- relationship between oviposition preference and larval iting in S. sachalinensis shoots that may partially com- performance in a shoot-galling sawßy. Ecology 70: 1691Ð pensate for the negative effect of shoot toughness of 1699. the longer shoots. Craig, T. P., J. K. Itami, and P. W. Price. 1990. The window of vulnerability of a shoot-galling sawßy to attack by a Natural enemies can also affect oviposition prefer- parasitoid. Ecology 71: 1471Ð1482. ence in herbivorous insects (Atsatt 1981, Thompson Craig, T. P., W. G. Abrahamson, J. K. Itami, and J. D. Horner. 1988, Ohsaki and Sato 1994, Yamaga and Ohgushi 1999. Oviposition preference and offspring performance 1999). For example, Ohsaki and Sato (1994) studied a of Eurosta solidaginis on genotypes of Solidago altissima. butterßy Pieris napi japonica Shirozu that oviposits on Oikos 86: 119Ð128. Arabis plants withlow quality for larvae performance Dorn, N. J., G. Cronin, and D. M. Lodge. 2001. Feeding in terms of survival rate, development time, and pupal peferences and performance of an aquatic lepidopteran mass. Since these plants grow mainly under other on macrophytes: plant hosts as food and habitat. Oeco- weeds throughout the growing season, parasitoids logia 128: 406Ð415. Hoffman, G. D., and P. B. McEvoy. 1986. Mechanical lim- have difÞculty in Þnding the butterßy larvae on the itations on feeding by meadow spittlebugs Philaenus spu- host plants. Thus, the butterßy may oviposit on these marius (Homoptera: Cercopidae) on wild and cultivated plants to avoid attack by specialist parasitoids, despite host plants. Ecol. Entomol. 11: 415Ð426. the poor food quality available to larvae. Ishihara, M., T. Hayashi, and T. Ohgushi. 1999. Life cycle of However, the two egg parasitoids of the spittlebug the willow leaf beetle, Plagiodera versicolora (Coleoptera: were unlikely to affect oviposition preference, be- Chrysomelidae) in Ishikari (Hokkaido, Japan). Entomol. cause egg parasitism was not a function of shoot size. Sci. 2: 57Ð60. Although in S. integra egg parasitism was signiÞcantly Kimberling, D. N., E. R. Scott, and P. W. Price. 1990. Testing higher in the 20Ð40 cm class compared with other a new hypothesis: plant vigor and phylloxera distribution on wild grape in Arizona. Oecologia 84: 1- 8. shoot length classes, ovipositing females did not avoid Kolehmainen, J., K. Roininen, R. Julkunen-Tiitto, and J. these shoots. Therefore, egg parasitism may not affect Tahvanainen. 1995. Importance of phenolic glucosides oviposition preference on S. integra or the two other in host selection of shoot galling sawßy. J. Chem. Ecol. 20: willow species examined. Why egg parasitism was 2455Ð2466. higher in 20Ð40 cm shoots remains unclear. Komatsu, T. 1997. A revision of the froghopper genus Previous studies have documented that some her- Aphrophora Germar (Homoptera, Cercopoidea, Aphro- bivorous insects (especially galling insects) have an phoridae) from Japan. Part 2. Jpn. J. Entomol. 65: 369Ð383. oviposition preference for vigorous plants and plant Mattson, W. J., Jr. 1980. Herbivory in relation to plant ni- modules (Price 1991). This study indicates that A. trogen content. Annu. Rev. Ecol. Syst. 11: 119Ð161. Mayhew, P. J. 1997. Adaptive patterns of host-plant selec- pectoralis also has an oviposition preference for long tion by phytophagous insects. Oikos 79: 417Ð428. shoots and the preference is affected by shoot abscis- McEvoy, P. B. 1986. Niche partitioning in spittlebugs (Ho- sion and shoot toughness. However, A. pectoralis did moptera: Cercopidae) sharing shelters on host plants. not prefer the longest (most vigorous) shoots, and Ecology 67: 465Ð478. sucha pattern hasnot reported in galling insects. Nozawa, A., and T. Ohgushi. 2002. Life-history and ovipo- Therefore, this result suggests that we should inves- sition preference of the willow spittlebug Aphrophora tigate whether the plant vigor hypothesis explains pectoralis (Homoptera: Aphrophoridae). Entomol Sci (in patterns of oviposition preference in other guilds of press). insect herbivores (Kimberling et al. 1990). Ohsaki, N., and Y. Sato. 1994. Food plant choice of Pieris butterßies as a trade-off between parasitoid avoidance and quality of plants. Ecology 75: 59Ð68. Acknowledgments Ohgushi, T. 1992. Resource limitation on insect herbivore populations, pp. 199Ð241. In M. Hunter, T. Ohgushi, and We thank Tim Craig and Andrew Davis for valuable com- P. W. Price (eds.), Effects of resource distribution on ments on the manuscript. Financial support was provided by -plant interactions. Academic, San Diego, CA. September 2002 NOZAWA: AND OHGUSHI:OVIPOSITION PREFERENCE OF A. pectoralis 557

Prestidge, R. A., and S. McNeill. 1983. - Scriber, J. M., and F. Jr. Slansky. 1981. The nutritional ecol- host plant interactions: leafhoppers and grasses. Ecol. ogy of immature insects. Annu. Rev. Entomol. 26: 183Ð Entomol. 8: 331Ð339. 211. Price, P. W., H. Roininen, and J. Tahvanainen. 1987a. Plant Stiling, P., D. Simberloff, and B. V. Brodbeck. 1991. Varia- age and attack by the bud galler, Euura mucronata. Oe- tion in rates of leaf abscission between plants may affect cologia 73: 334Ð337. the distribution patterns of sessile insects. Oecologia 88: Price, P. W., H. Roininen, and J. Tahvanainen. 1987b. Why 367Ð370. does the bud-galling sawßy, Eurra mucronata, attack long Tahvanainen, J., R. Julkunen-Tiito, and J. Kettunen. 1985. shoots? Oecologia 74: 1Ð6. Phenolic glycosides govern the food selection pattern of Price, P. W. 1991. The plant vigor hypothesis and herbivore willow feeding leaf beetles. Oecologia 67: 52Ð56. attack. Oikos 62: 244Ð251. Thompson, J. N. 1988. Evolutionary ecology of the relation- Price, P. W., T. P. Craig, and H. Roininen. 1995. Working ship between oviposition preference and performance of toward theory on galling sawßy population dynamics, pp. offspring in phytophagous insects. Entomol. Exp. Appl. 321Ð338. In N. Cappuccino and P. W. Price (eds.), Pop- 47: 3Ð14. ulation dynamics: new approaches and synthesis. Aca- Thompson, J. N., and O. Pellmyr. 1991. Evolution of ovipo- demic, San Diego, CA. sition behavior and host preference in lepidoptera. Annu. Price, P. W., H. Roininen, and J. Tahvanainen. 1997. Willow Rev. Entomol. 36: 65Ð89. tree shoot module length and the attack and survival Thompson, V. 1994. Spittlebug indicators of nitrogen-Þxing pattern of a shoot-galling sawßy, Euura atra (Hymenop- plants. Ecol.Entomol. 19: 391Ð398. tera: Tenthredinidae). Entomol. Fenn. 8: 113Ð119. Woods, J. O., T. G. Carr, P. W. Price, L. E. Stevens, and N. S. Rossi, A. M., and D. R. Strong. 1991. Effects of Host-plant Cobb. 1996. Growthof coyote willow and theattack and nitrogen on the preference and performance of labora- survival of a mid-rib galling sawßy, Euura sp. Oecologia tory populations of Carneocephala floridana (Homoptera: 108: 714Ð722. Cicadellidae). Environ. Entomol. 20: 1349Ð1355. Yamaga, Y., and T. Ohgushi. 1999. Preference-performance Rossi, A. M., B. V. Brodbeck, and D. R. Strong. 1996. Re- linkage in a herbivorous lady beetle: consequences of sponse of xylem-feeding leafhopper to host plant species variability of natural enemies. Oecologia 119: 183Ð190. and plant quality. J. Chem. Ecol. 22: 653Ð671. Zar, J. H. 1999. Biostatistical analysis. Prentice-Hall, Engle- Rowell-Rahier, M. 1984. The food plant preferences of Ph- wood Cliffs, NJ. ratora vitellinae (Coleptera: Chrysomelinae). Oecologia Zucker, W. V. 1982. How aphids choose leaves: the roles of 64: 375Ð380. phenolics in host selection by a galling aphid. Ecology 63: Scheirs, J., L. De Bruyn, and R. Verhagen. 2001. Nutritional 972Ð981. beneÞts of the leaf- mining behaviour of two grass miners: a test of the selective feeding hypothesis. Ecol. Entomol. Received for publication 5 November 2001; accepted 26 April 26: 509Ð516. 2002.