© Biologiezentrum Linz/Austria; download unter www.biologiezentrum.at Polyphagy in True Bugs: A case study of Leptoglossus phyllopus (L.) (Hemiptera, Heteroptera, Coreidae)1 P.L. MITCHELL Abstract: The polyphagous species Leptoglossus phyllopus(L.) was examined with respect to host plant preference, tissue feeding specificity, seasonal dispersal among host plants, and life history. Mark-recap- ture, census, and rearing experiments demonstrated that this species exhibits true polyphagy, in that in- dividual bugs feed on plants from more than one family. Developmental parameters such as growth and survivorship did not differ among plants from several families, but did vary significantly with quality of host (e.g., wild vs. cultivated). Stadium duration, however, varied among wild host plant species in la- boratory experiments. Specialization on reproductive plant parts, coupled with sequential polyphagy and dispersal among a variety of seasonal host plants, allows multiple generations per year. Modes of fee- ding and preferred target tissues among coreids are discussed. Key words: leaffooted bug,Leptoglossus phyllopus, polyphagy, stylet sheath, target tissue. Introduction spp.), for example, employ a macerate-and- flush process, whereas an osmotic pump For phytophagous insects with piercing- mechanism is associated with coreids (MILES sucking mouthparts, feeding selectivity op- & TAYLOR 1994). However, some pentato- erates on two levels: preferred target tissue mids and lygaeids shift between salivary and host plant species. Tissue preference is sheath formation and lacerate-and-flush considered to be a conservative evolution- feeding (MILES1972). ary character (TONKYN & WHITCOMB 1987); for example, specialization on xylem, Preference for a particular plant devel- phloem, or parenchyma is characteristic of opmental stage or structure (e.g., growing subfamilies or families in Auchenorrhyncha shoots, buds, ripe seeds), may also be associ- (PORT 1978). Phytophagous Heteroptera ated with specific groups. Pyrrhocoroidea have been extensively analyzed with respect feed predominantly on mature seeds, tingids to several factors associated with target tis- prefer mature leaves, while phytophagous sue preference: mouthpart morphology, sali- mirids concentrate on flowers, buds, and vary chemistry, preference for plant parts or new foliage (SCHUH & SLATER 1995). In structures, and mode of feeding (MILES some cases, however, such trends only be- 1972; COBBEN1978; MILES& TAYLOR1994; come evident at the tribal or even generic HORI 2000). The enzyme components of level. SCHAEFER & MITCHELL (1983) noted heteropteran saliva are considered to be a that an affinity for vegetative or reproduc- family-level character (MILES, cited in HORI tive structures is characteristic of tribes 2000), although induction of proteolytic en- within the Coreidae. Among Australian zymes has been demonstrated in a zoophy- coreids, KUMAR (1966) recognized two dif- tophagous mirid (ZENG & COHEN 2001). ferent types of feeding – exclusive sap feed- The mode of feeding may also be associated ing is found in some genera whereas others with particular taxa. Mirids (e.g., Lygus feed on both sap and fruit. However, such Denisia 19, zugleich Kataloge 1Dedicated affectionately to Ernst Heiss, my delightful colleague on the Executive Committee of the International der OÖ. Landesmuseen Heteropterists’ Society and heteropterist extraordinaire. Neue Serie 50 (2006), 1117–1134 1117 © Biologiezentrum Linz/Austria; download unter www.biologiezentrum.at diversity of feeding was considered by phytophagous groups, he proposed that sap- COBBEN (1978) to indicate a “phase of un- feeding originated from polyphagous seed- balanced equilibrium.“ feeding, with the former mode of feeding tending toward a more restricted host range Even fewer generalizations may be made (COBBEN 1978). In Hemiptera in general, regarding host-plant specificity. Curiously, tissue-feeding specialization has been associ- the question of diet breadth in Heteroptera ated with host plant specificity (BERNAYS& has not been explored with the intense fer- CHAPMAN 1993), but this observation is vor devoted to insect-plant interactions in based primarily on the British fauna, and is Lepidoptera, Orthoptera, and other orders biased toward Sternorrhyncha and Auchen- of chewing insects. COBBEN (1978) charac- orrhyncha. Clearly, more information on terized the Tingidae as monophagous while polyphagy in Heteroptera is needed. suggesting that “polyphagy prevails“ among the phytophagous Miridae; but he provided The objective of this research was to no comparable analysis for the Pentatomo- conduct a detailed study of polyphagous morpha. Nearly 30 years have passed since feeding behavior in a single coreid species, the publication of COBBEN’s (1978) monu- the leaffooted bug Leptoglossus phyllopus mental study of mouthparts and feeding (L.). Host range and dispersal, seasonal pat- strategies, but our understanding of het- terns of host plant use, preferred feeding eropteran-plant relationships remains limit- site, and target tissue were examined in or- ed. Neither facultative zoophagy nor the ex- der to characterize the factors influencing treme polyphagy associated with some het- host plant selection. Leptoglossus phyllopus eropterans is properly understood (ZENG & damages legumes, tomatoes, tree nuts, and COHEN 2001). WHEELER (2001) observes citrus in the southern and eastern United that despite an extensive literature on crop States (MITCHELL2000), and is considered a injury, there are “large gaps in our knowl- minor economic pest. The primarily edge of mirid-plant interactions.“ Informa- Neotropical genus Leptoglossus GUERIN was tion for some heteropteran groups is lacking, selected for study because it includes species detailed ecological studies of bugs outside of exhibiting widely disparate feeding strate- agroecosystems are uncommon, and adult gies ranging from strict monophagy to ex- host plant records are not always reliable in- treme polyphagy, as well as a number of dicators of breeding hosts (SCHAEFER & species of economic importance. Several MITCHELL1983). Leptoglossusspecies are presently expanding Polyphagy in Heteroptera is of great in- their ranges – and potential for economic terest from both an economic and an evolu- damage – geographically. The highly tionary standpoint. Many of the major agri- polyphagous L. zonatus (DALLAS), a pest of cultural pests (e.g., Nezara viridula(L.), Rip- many crops in Mexico and South America, tortus clavatus (THUNBERG), Leptoglossus has recently invaded the southeastern Unit- zonatus (DALLAS), Lygus lineolaris (PALISOT ed States to Florida (BUSS et al. 2005) and DE BEAUVOIS)) have a wide range of host has become a pest of Satsuma oranges in plants (SCHAEFER& PANIZZI2000). Howev- Louisiana, displacing L. phyllopus (HENNEet er, it is important to distinguish between al. 2003). Leptoglossus occidentalis HEIDE- generalist species composed of more special- MANN, once restricted to conifers in western ized populations and polyphagy at the indi- North America, has spread rapidly eastward vidual level (i.e., diet mixing) (FOX& MOR- across the continent to Ontario and New ROW 1981; BERNAYS & MINKENBERG 1997). England (RIDGE-O’CONNOR 2001). Acci- The evolution of omnivory has been shown dentally introduced into Italy in 1999 (TAY- to correlate with polyphagy in Heteroptera LORet al. 2001), this species has now spread (EUBANKS et al. 2003), and these authors to Central Europe (RABITSCH & HEISS proposed that polyphagous, seed-feeding 2005). Thus, information on the feeding be- herbivores may be more likely to expand the havior of L. phyllopusmay be applicable not diet to consume animal prey. COBBEN(1978, only in the southeastern United States, but 1979) considered polyphagous carnivory to to congeneric species of economic impor- be plesiomorphic within Heteroptera. For tance elsewhere. 1118 © Biologiezentrum Linz/Austria; download unter www.biologiezentrum.at Materials and Methods Both L. phyllopus and another coreid, Eu- thochtha galeator(F.), were monitored. Num- Study site:Research was conducted pri- ber of adults, nymphs of all instars, and cop- marily at Brackenridge Field Laboratory ulations were counted. Location of bugs on (BFL), operated by The University of Texas the plant was noted in greater detail than in at Austin. This field station is located inside earlier censuses, differentiating between the city limits of Austin, Texas, along the bud, stem, and leaf primordia as well as be- shores of the Colorado River, and comprises tween mature plant organs. The 1979 cen- 32 hectares of grassland, cedar-post oak sus covered the entire reproductive life span scrub and semi-deciduous woodland. of the host plant Cirsium texanum BUCKL., Four study plots, each measuring 92.9 from bolting through seed set. m2, were established at BFL to investigate A plant census was conducted monthly seasonal activity of L. phyllopus and other in each study site in conjunction with the coreids. All plots were located in either ear- bug censuses in 1977 and 1978. Each plot ly successional or cultivated garden areas. was subdivided into 40 equal quadrats, and a Plot 1, on the river terrace, was dominated wire square measuring 930 cm2 was tossed by Johnsongrass, Sorghum halepense (L.) twice into each quadrat. This quadrat PERS.,and various composite species. Plot 2, method was designed to eliminate the un- bordering an artificial pond, was composed equal representation of marginal areas fre- primarily of grasses and composites inter- quently associated with sampling by random spersed with prickly pear (Opuntia spp.) and toss (SOUTHWOOD1966). After each throw, mesquite (Prosopis glandulosa TORR.). The presence or absence of plant species within remaining study plots were placed in the the wire sampler was noted, in order to com- cultivated garden area to ensure an adequate pute frequency of occurrence from a total of water supply during summer drought. Plot 3, 80 samples per study plot. Additionally, formerly under cultivation, had returned to counts were taken of number of stems, buds, native early successional vegetation two flowers, developing fruits, and mature fruits years before the commencement of research. of host plants within the confines of the Plot 4, the agricultural study plot, was plant- wire square. All counts were made visually ed each spring with tomato (Lycopersicon es- in the field, and no plant material was re- culentum P. MILL.), giant sunflower (He- moved from the study areas. All Leptoglossus lianthus annuus L.), and okra (Hibiscus host plants were identified to species. With [=Abelmoschus] esculentus L.). A sprinkler the exception of grasses, non-hosts were system provided water for both garden plots identified at least to genus. Within the twice weekly, and fertilizer was applied to Poaceae only S. halepensewas distinguished; cultivated plants according to standard other grasses were lumped in a single cate- agronomic practices. gory. Plant family designations follow JUDD et al. (2002); species and describer names Field census: From July to November 1977, and February to November 1978, L. follow RADFORDet al. (1968) and CORRELL phyllopusindividuals were visually counted in & JOHNSTON (1970). In 1979, only host plants were surveyed. Finer distinctions each study plot at weekly intervals. All cen- were used in the census of plant structural suses were taken within two hours of sunset, and reproductive organs, to correspond with a time of maximal copulation and feeding ac- the more detailed bug distribution studies. tivity. Field work was discontinued during the winter (mid-November through mid- For statistical analysis, plants were cate- February), as no bugs were active during this gorized as reproductive or non-reproductive, period. For each individual observed, the fol- and bug counts (log bug density for each 10 lowing data were collected: location on plant non-zero census date) were correlated with (leaf, stem, bud, etc.), plant species and re- density of reproductive plants for C. tex- productive condition, instar (nymphs), and anum in Plot 1 and Heterotheca latifolia copulatory activity (adults). BUCKL. in Plots 2 and 3. From March to June 1979, a modified Mark-recapture: Mark-release-recap- bug census was conducted weekly in Plot 1. ture studies were carried out at BFL from 1119 © Biologiezentrum Linz/Austria; download unter www.biologiezentrum.at Fig. 1: Leptoglossus phyllopus adults; June 1976 through November 1977. Adult bugs to emerge. The effect of prominent clockwise from upper left: aggregated on L. phyllopus were individually marked by white dorsal markings on bugs in the field is Cirsiumsp., March, Brazos Co., Texas painting the pronotum with typewriter cor- unknown, but this technique does not sig- (Photo: W. O. Ree Jr.); in copula on Gaura parviflora, July, Burleson Co., Texas (Photo: rection fluid (Liquid Paper™) and number- nificantly alter survivorship or copulatory A. Calixto); aggregated with marked bug ing with india ink (Fig. 1). A total of 1,070 activity of individuals under laboratory con- on Helianthus annuus, August, Travis Co., individuals was marked during the course of ditions (Mitchell unpubl. data). Mark-re- Texas; feeding on Solidagosp., October, Brazos Co., Texas (Photo: W. O. Ree Jr.); the study. Bugs were collected in the field by lease-recapture studies were designed to ex- feeding on G. parviflora, July, dropping net bags over aggregations on host amine dispersal, host plant shifts, and length Burleson Co., Texas (Photo: A. Calixto). plants, and were returned to the laboratory of residency, rather than population size. for marking. With rare exceptions, each Consequently, initial marking was a discon- batch was completed and released on the tinuous process, although study areas were date of capture. Bags of marked individuals checked daily or twice daily for recaptures. were attached to a branch or stem of the Thirteen subsites (host plant patches) were original host plant and opened, allowing delineated in the mark-release-recapture 1120 © Biologiezentrum Linz/Austria; download unter www.biologiezentrum.at study. Size of subsites was variable, being a Colony maintenance: Leptoglossus phyl- function of host plant density in a given lopus colonies were reared at 26 ±2°C, and area, and ranged from a single Callicarpa a photoperiod of 16:8 [L:D], and were fed americanaL. bush to a 30 m2stand of H. lat- whole green beans, Phaseolus vulgarisL. and ifolia. Distance between subsites was meas- shelled organically grown sunflower (H. an- ured from center to center of each patch. nuus) seed. Adults were housed in screen Analysis of dispersal therefore did not ac- and Plexiglas™cages (25 x 25 x 25 cm) pro- count for movement within the larger sub- vided with wooden applicator sticks for sites, and total movement may have been oviposition. Nymphs were maintained in slightly underestimated. Each subsite was smaller cages constructed from 300 ml plas- characterized by one dominant host plant tic drinking cups (cup cages) and provided species. with a constant source of water through cot- ton dental wicks. The following data were collected for Rearing experiments: Conducted in each marked individual: sex, subsite, time of both laboratory and field, these experiments collection, host plant species, absence of examined the ability of fifth-instar L. phyllo- legs, presence (and position) of macrotype pusto develop on a variety of hosts from dif- tachinid eggs on exoskeleton, degree of scle- ferent plant families at different stages of de- rotization (teneral vs. mature) and wing velopment. Three variables were used to wear condition (tattered vs. whole). Begin- measure host plant suitability: duration of ning in August 1976, bug length was meas- the fifth instar, growth increment during ured with calipers to the nearest 0.5 mm. At this period, and survivorship. Experimental each subsequent recapture original observa- bugs were derived from field collected par- tions were rechecked, with the exception of ents, and were reared in the laboratory. length measurement. Further information Rearing colony cages were checked nightly, collected for resighted individuals included at growth chamber sunset. All newly molted location on host plant and activity (e.g., sit- fifth instars were collected at this time, iso- ting, walking, probing, sucking, copulating). lated in screen-covered cups overnight, and Data from the pilot study (June to Sep- supplied with a green bean for moisture. tember 1976) and overwintering periods Distribution of nymphs among treatments (October to February of both years) were was completed within 24 hours. Assignment analyzed separately from the major mark-re- to treatment was randomized within the set lease-recapture project in 1977. Marking of of host plants in the appropriate develop- the overwintering generation proved unsuc- mental condition at a given point in time. cessful; no bugs marked in late fall of either Each nymph was measured with calipers year were ever sighted again the following to the nearest 0.5 mm before placement on spring. Pilot project data were analyzed by the experimental host plant. Cages were hand to determine host plant and subsite checked daily at sunset, and newly molted shifts by individual bugs. adults were re-measured. Growth increment The mark-release-recapture project in was calculated as the difference between 1977 involved a total of 543 individuals and these two length measurements. Instar dura- three complete generations of L. phyllopus. tion was defined as days elapsed between the fifth instar and adult molts. Survivorship, Bugs were separated into four generational assessed only in laboratory experiments, was categories (overwintering, spring, summer, calculated as number of surviving nymphs fall) on the basis of wing wear and scleroti- per cage. zation. Data were processed using the LONGPOP program (WATT et al. 1977, In the laboratory, detached fruits or seed 1979). This program applies the capture-re- heads of a single plant species were offered capture methods of JOLLY(1965) for calcula- as a food source. Each experimental cup tion of survivorship, and also tabulates dis- cage contained five bugs, individually persal and other population parameters. Sex marked with enamel dots on the tibial folia- ratios were tested using χ2 goodness-of-fit tion. Laboratory experiments involved six (ZAR1999). treatments, consisting of laboratory diet 1121 © Biologiezentrum Linz/Austria; download unter www.biologiezentrum.at control and five host plants: Callicarpa amer- Field data from mark-recapture studies icana L. var. lactea (Lamiaceae), Campsis were also analyzed to substantiate the exper- radicans (L.) SEEM. (Bignoniaceae), Gaura imental results for fifth instar growth incre- parviflora DOUGL. (Onagraceae), and ment. Adult length of L. phyllopus marked Solanum eleagnifolium CAV. (Solanaceae). as teneral individuals was compared among Each treatment consisted of four cages for a various host plants. Adult length of males total of 20 nymphs per treatment. and females was analyzed separately for these data, because sex ratios were not In field experiments, bugs were caged equivalent across treatments. individually on live host plants (n = 25). Tulle sleeves, measuring 45 by 25 cm, were Measurement data from both field and lowered around plant stems or branches, laboratory experiments were analyzed by t and fastened at the base with fine wire. Veg- tests or by one way analyses of variance etative portions of the plant, flower buds, or (ANOVA) followed by Fisher’s LSD. Loga- one or more developing seed heads (de- rithmic (base-10) transformation was used pending on the growth form of the plant) prior to ANOVA if needed to normalize da- were contained within each sleeve cage. Af- ta or correct for unequal variance (MINITAB ter addition of a nymph, each sleeve was 2000). Count data were analyzed by non- wired together at the top, enclosing both parametric tests, including Friedman two- bug and plant. Plants were maintained in way ANOVA, Kruskal-Wallis (MINITAB the bud stage by clipping flowers within the 2000), and non-parametric rank analog of cage as they opened, or moving the entire Tukey’s test (CONOVER1971). cage to a new plant. Field experiments con- Laboratory choice tests:Choice exper- centrated mainly on species in the family iments were conducted in small cup cages. Asteraceae. Treatments included Baccharis Naive second instars were used in all tests. neglecta BRITT. (Astereae), C. texanum Field observations identified this instar as (Cynareae), Gutierrezia texana (DC) TORR. the first critical period of host plant selec- & GRAY (Astereae), H. annuus (He- tion, because the previous instar is a non- liantheae), H. latifolia (Astereae), Solidago feeding stage, and oviposition site does not altissima L. (Astereae), and G. parviflora necessarily reflect host plant preference (Onagraceae). The last four species were (MITCHELL & MITCHELL 1986). Nymphs tested in all three developmental stages; used in experiments were derived from field- other hosts listed were tested only in repro- collected adults, but reared from the egg ductive condition. stage in the laboratory. Prior to placement Field experiments were conducted from in testing cages, bugs received only green September to November of 1977, and May bean (a water source), and were not exposed through July of 1978, according to the phe- to laboratory diet. nology of host plants studied. Separating the Since early instar nymphs are normally influence of climatic variation from genuine gregarious, bugs were tested in groups of five treatment effects was therefore critical. A per cage. Preference between plant species series of controls was established, in order to was tested by offering seed heads of H. lati- monitor the combined effects of tempera- folia (Asteraceae) and G. parviflora (Ona- ture and daylength and also laboratory stock graceae) as alternative food choices; prefer- quality. In the laboratory control treatment, ence for plant parts was tested using buds bugs received standard rearing diet (beans, and immature seed heads of C. texanum. For sunflower seeds and distilled water), and each comparison, 36 replicates were con- were maintained in the growth chamber at ducted, using a total of 180 second instars. 16:8 [L:D] photoperiod conditions. Addi- Each testing cage contained one cut stalk tionally, tulle cages supplied with laboratory (ca. 8 cm) of each plant alternative, insert- diet were set up on the roof of the Zoology ed in separate water vials. Fresh plant cut- building on the University of Texas campus. tings for each experiment were collected One set of these “roof controls“ was con- daily at the field station. Following addition ducted in fall of 1977, and one in summer of of nymphs, cages were observed at 4-h inter- 1978. vals during the 16-h growth chamber day, 1122 © Biologiezentrum Linz/Austria; download unter www.biologiezentrum.at and again at sunrise of the following day. Stylet tracks of L. phyllopuswere exam- Stylet insertion was considered to represent ined on the wild host C. texanum, and also definitive choice. A single observation, that on cultivated green cherry tomato (L. escu- which involved the greatest number of feed- lentum), and green bean (P. vulgaris). In ad- ing individuals, was selected from each ex- dition, feeding by the coreid E. galeator on perimental data set. A sign test (ZAR1999) C. texanum was investigated, as these two was performed on the resulting values. species coexist on thistle in the spring. Histology: Plant tissue exposed to het- Voucher specimens: Specimens of the eropteran feeding was sectioned and exam- dipteran parasitoid Trichopoda pennipesF. are ined under a light microscope. Field collect- deposited in the California Academy of Sci- ed plant material was supplemented by ences, Department of Entomology. Coreid caging bugs on potted plants and cuttings in specimens are in the insect collection at the laboratory. Sites of observed stylet inser- Texas A&M University, Department of En- tion were marked on the plant surface with tomology. an indelible pen. Marked plant tissue was cut with a single edge razor blade, allowing Results a margin of ca. 2 mm. around the point of insertion, and fixed in either Carnoys solu- Field census. Although a diverse array tion (0.5 hr) or formalin-acetic acid-alcohol of plants from seven families served as L. (FAA) (48 hrs). Material fixed in Carnoys phyllopus breeding hosts at BFL (Table 1), was dehydrated through a series of ethyl al- plant species were not used in direct propor- cohols to xylene, while FAA treated tissue tion to frequency of occurrence. One species was dehydrated with tertiary butyl alcohol, in each study plot was always colonized to a following the method of JOHANSEN (1940). greater extent than predicted on the basis of Although both procedures were satisfactory, abundance alone, and others (e.g., grasses) that of Johansen proved more consistently were never used as breeding hosts at all. reliable. Dehydrated tissue was embedded in Temporal patterns of bug distribution on Paraplast™and sectioned at 15 microns on various wild and cultivated host plants are a rotary microtome. Serial sections obtained shown in Figure 2. Small vertical arrows in- in this manner were rehydrated and stained dicate the beginning of plant reproductive with a 1 % solution of safranin in 50 % activity. On thistle, tomato, sunflower and ethanol, and counterstained with fast green goldenrod, bug density increased dramati- (0.2 % in 50 % ethanol). Best results were cally following the transition from vegeta- obtained using a staining schedule of 2 h in tive to reproductive growth. Colonization safranin solution, followed by 1 min in fast green and 2 min in each of two 95 % alco- was concurrent with fruiting in the latter hol baths. Table 1: Nymphal host plants of L. phyllopus, Travis County, Texas, 1977-1979. Asterisk After staining and subsequent dehydra- indicates observations at locations other than Brackenridge Field Station. tion of sections, coverslips were mounted Plant family Plant species Common name with Permount™. Bug salivary sheaths stain Asteraceae Achillea millefoliumL. yarrow bright pink with safranin. Xylem stains red Baccharis neglectaBRITT. desert willow Cirsium texanumBUCKL. Texas thistle in contrast to the fast green absorbed by Gutierrezia texana(DC) TORR. & GRAY broomweed parenchyma and other non-lignified plant Helianthus annuusL. sunflower tissues. Both stylet tracks and damaged Heterotheca latifoliaBUCKL. camphor weed Solidago altissimaL. goldenrod plant tissue were easily observed using this Bignoniaceae Campsis radicans(L.) SEEM. trumpet creeper* staining technique. As the plane of section- Cucurbitaceae Cucurbita pepoL. squash ing only rarely coincided with the path of Lamiaceae Callicarpa americana L. French mulberry insertion, seriality of sections was critical in Malvaceae Hibiscus (=Abelmoschus)esculentusL. okra determining the termination point of the Onagraceae Gaura parvifloraDOUGL. lizard-tail stylet tracks. Rare sections bearing com- Orobanchaceae Agalinis heterophylla(NUTT.) SMALLEXBRITT. prairie false foxglove plete, parallel-cut salivary sheaths were pho- Agalinis strictifolia(BENTH.) PENNELL stiffleaf false foxglove tographed at 100x with a Nikon Eclipse Scrophulariaceae Verbascum thapsus L. woolly mullein* E600 microscope equipped with an Olympus Solanaceae Lycopersicon esculentumMILLER tomato DP10 digital camera. Solanum eleagnifoliumCAV. silver-leaf nightshade 1123 © Biologiezentrum Linz/Austria; download unter www.biologiezentrum.at hatched prior to flowering only in 1979, and buds were used temporarily as a food re- source until developing seed heads became available. Preference among plant reproductive structures is best documented by the de- tailed data for C. texanumin spring of 1979 (Fig. 3). Weekly plant censuses and biweek- ly bug counts covered the period from ap- pearance of the first bud primordia in April until seed set in early June. For each census date, numerical tallies were converted to percentages, to facilitate comparisons be- tween plant organ availability and bug uti- lization. Percent bug utilization was calcu- lated as number of bugs on a given plant structure (e.g. bud, flower, etc.) divided by total bugs observed on reproductive organs of that host plant species. Plant percentage Fig. 2: Temporal patterns of L. phyllopus three species. Individuals dispersed between distribution on various cultivated and wild host plants with overlapping life cycles, as values were determined in a similar manner. host plants based on 1978 census counts, Preference was indicated when bug percent- one host reached senescence and the next Travis Co., Texas. age use exceeded plant percentage availabil- species began fruiting. Okra and H. latifolia ity. Total daily bug counts during the census become reproductive several months before ranged from 10 to 122; plant parts counted bug invasion, but were never colonized dur- in each census ranged from 40 at first ap- ing the vegetative stage. For a given plant pearance of primordia to 302 at peak species, bug density corresponded closely growth. with abundance of plants in reproductive condition. The number of reproductive Early in the season, bugs concentrated plants was significantly correlated with the on primordia (Fig. 3a), areas of the plant ex- log of bug density in the study area for both hibiting rapid growth. As true buds ap- C. texanum(r = 0.927, n = 15, P < 0.01) and peared (Fig. 3b), abrupt switching behavior H. latifolia(r = 0.692, n = 44, P < 0.01); oth- was evident, leading to disproportionately er plant species were not numerous enough lower use of the still common primordia. for statistical analysis. Identical switching behavior accompanied the first appearance of developing seed Leaffooted bugs were active from early heads (Fig. 3d), which continued as the pre- spring to late fall, and overwintered as ferred feeding site for the remainder of the adults. Spring emergence varied from mid- season. Flowers (Fig. 3c), occurring during February to late March, depending on the the same time period as immature seed harshness of the winter. For 1976-1978, heads, were rarely used. Ripening of seed good correspondence was apparent between heads in mid-June was not accompanied by bug emergence and the condition of C. tex- an increased proportion of bugs on mature anum, the first breeding host. In early heads (Fig. 3e). spring, while thistles remained in the rosette stage, an occasional leaffooted bug could be Similar census data for 1978 (not pre- observed probing on B. neglecta, a woody, sented) also indicated an overwhelming late fall host.However, large aggregations of preference for immature fruit and develop- emerging adults did not appear in spring un- ing seed heads, and concomitant avoidance til thistles commenced bolting and pro- of open flowers and mature fruits and seeds; duced the first central bud primordia. Copu- this was observed for C. texanum, C. ameri- lation began 1-3 weeks after emergence cana, H. latifolia, and S. altissimaconsistent- from overwintering, and production of ly and for H. annuusin early season (June). nymphs coincided with flowering and seed Mature seed heads of the last species were development of C. texanum. Nymphs preferred in July, and all seed heads were 1124 © Biologiezentrum Linz/Austria; download unter www.biologiezentrum.at used in proportion to abundance at the end of the season in August. A comparison of L. phyllopus distribu- tion with that of a less common coreid, E. galeator, on the same host plant shows a clear difference in utilization of plant parts of C. texanum(Fig. 4). Significantly higher numbers of L. phyllopusare found on repro- ductive structures (buds, flowers, and seed heads) compared with leaves or stems (T = 8.22, df = 2, P = 0.016), whereas signifi- cantly more E. galeator are found on stems (includes branches, petioles, and peduncles) than on leaves or reproductive parts (T = 9.39, df = 2, P < 0.01). Mark-recapture. Analysis of 1977 cap- ture-recapture data provided information on dispersal, longevity, copulatory activity, sex ratio, predation, and parasitism of adults, in addition to the primary objective of docu- menting individual host plant shifts. For males, 116 individuals were recaptured out of 300 marked (38.7 %); in total, 299 recap- ture events resulted in an average recapture frequency of 2.58. Of 243 females marked, 86 (35.4 %) were recaptured in 153 recap- ture events, for an average recapture fre- quency of 1.78. Females showed a greater tendency to disperse, although distance trav- eled was equivalent between the sexes; 31 fe- male dispersants (36 %) moved an average of 42.31 m, whereas for the 33 male dispersants (28.5 %), the mean distance moved by dis- persing individuals was 44.01 m. The majority of individuals recaptured (71.5 % of males, 63.9 % of females) re- mained in a single patch of host plants. Most dispersants were recorded moving be- tween separate patches of the same host plant species. However, multiple host shifts Marked bugs were also observed in cop- were documented. Teneral adults marked on Fig. 3: Floral and fruiting stages of Cirsium ula. Among male L. phyllopus, eight bugs C. texanum were subsequently recaptured texanum, Travis Co. Texas, April-June 1979, while feeding on H. annuus(n = 4) and L. were observed to be mating on two separate showing frequency of occurrence of each occasions. One marked individual engaged structure and distribution of L. phyllopus; esculentum(7). One mature, fully sclerotized squares connected by solid lines indicate in four copulations on different days. For fe- adult marked on G. parviflora was recap- percent occurrence of plant structure; tured feeding on H. annuus. Mature individ- males, three individuals were seen mating circles connected by dotted lines indicate uals marked on H. annuus were observed twice, and two engaged in copulations at percent bug distribution on a structure. feeding on H. latifolia(2), L. esculentum(1), least three times during the course of the and Zea mays L. (3). Marked individuals study. Since subsites were checked only from H. latifolia fed upon G. parviflora (1), briefly each day (ca. 10 minutes per site), and H. annuus (4), while those from L. es- these observations represent only a small culentum dispersed to feed upon H. annuus fraction of total copulatory activity. Caged (8). The maximal number of host shifts ob- bugs in the laboratory mated nearly every served was four per marked adult. evening. No individuals marked as teneral 1125 © Biologiezentrum Linz/Austria; download unter www.biologiezentrum.at adults were observed in copula on C. tex- what more frequently than legs with simple anum, but four of the 11 dispersants sighted tibiae, this trend may indicate only that again on sunflower and tomato mated at predators preferentially approach from the least once on the adult host plant. After rear, or that the expanded leaf-feet present a bugs reached adulthood, average residence larger target for attack. Larger sample sizes time on the nymphal host plant (C. tex- would be required to verify a predator escape anum) was 3.8 +2.1 days for males (n = 18), function for the tibial expansions. and 3.5 +2.1 days for females (n = 11). A major natural enemy of L. phyllopusat Field longevity (e.g., residence rate) was BFL is the tachinid fly Trichopoda pennipesF. slightly higher for males, and the longest in- This parasitoid species (actually a complex terval between initial marking and final re- of cryptic species or biotypes) is an impor- capture was 51 days, as opposed to 42 for fe- tant natural enemy of the economic pest N. males. The recapture duration decay plot viridula (JONES 1988), and leaffooted bugs over time indicated a Type II survivorship serve as an alternative host in many crops in pattern for both sexes. For males, the regres- the southeastern United States (TILLMAN sion line may be expressed as y = 4.55 - 2006). Other coreid genera attacked include 0.0878x, where x = days in residence and y Acanthocephala, Anasa, Archimerus, Chelin- = ln number of individuals (F = 1728.3; df = idea, Euthochtha, Leptoglossus, and Narnia 1, 49; P < 0.005). For females, y = 3.94 - (ARNAUD 1978). At BFL, T. pennipes has 0.0985x (F = 803.1; df = 1,40; P < 0.005). been reared from Acanthocephala declivis SAY, A. femorata (F.), A. terminalis (DAL- Only the sex ratio of the overwintering LAS), Chelinidea vittiger UHLER, and Lep- Fig. 4: Distribution of Leptoglossus generation on C. texanum was significantly phyllopus(A) and Euthochtha galeator(B) toglossus oppositus (SAY)as well asL. phyllo- biased toward males (s.r. = 0.602, n = 103, on leaves, stems, and reproductive parts of pus(Mitchellunpublished). the host plant Cirsium texanum, Travis Co., χ2= 4.28, df = 1, P < 0.05). Census data for Texas, April-May, 1979. Median value, first 1978 corroborated this result (s.r. = 0.597, n Intensity of parasitism was estimated as and third quartiles, and upper and lower = 613, χ2= 23.1, df = 1, P < 0.01). For five the proportion of adult bugs bearing a ranges indicated by box and whisker plot. weeks following the initial colonization of macrotype egg (or eggs) on the exoskeleton. the first spring host, males predominated; af- Mark-recapture data indicated that high ter this time and throughout subsequent levels of parasitism were attained only in the generations, the ratio of males to females overwintering and first spring generations, did not differ significantly from 1:1. with 46 % and 27 % egg-bearing individu- als, respectively. Leptoglossus phyllopusis the The large hind leaf-feet of bugs in the first hemipteran species to reach appreciably coreine tribe Anisoscelini detach easily, and high densities in the spring, whereas alter- may operate as a predator escape mecha- nate hosts (e.g., A. femorata, Anasa tristis nism. The relative extent of predation in (DEGEER), various pentatomid species) be- different populations may be estimated by come common on garden crops somewhat frequency of leg loss, in much the same way later in the season. According to both mark- that tail break frequency is used to compare recapture and census data, <10 % of L. phyl- predation rates among lizard populations lopusin summer and fall populations bear ta- (e.g., PIANKA 1970). Leg loss data for indi- chinid eggs. viduals collected during the capture-recap- ture study showed that only the overwinter- Oviposition site on the bug exoskeleton ing generation on C. texanum exhibited a varies, but most eggs are placed on the head high rate of leg loss: 65 % of individuals re- of L. phyllopus. Of 142 eggs examined, dis- mained intact (n = 100). In all subsequent tribution on the body was as follows: head, generations, the percentage of undamaged 72.5 %; pronotum, 13.4 %; elsewhere on individuals was considerably higher, ranging thorax, 9.9 %; abdominal sternites, 2.1 %; from 87.0-90.3 % (n ≥ 93). Summation of abdominal tergites, 0.0 %; tibial foliation, hind, mid, and foreleg loss frequency over 0.7 %; wing, 1.4 %. The presence of a ta- the entire season yields 55 individuals lack- chinid egg on the exoskeleton does not ap- ing at least one hind leg, 41 lacking a mid- pear to deter an ovipositing T. pennipes. Oc- dle leg and 13 with a detached foreleg. Al- currence of supernumerary T. pennipeseggs, though the foliated hind legs are lost some- although infrequent in females, was com- 1126