THESlS MATCHIGAN ‘m “3 1293 \ni1\\\uil71'\“\\igli‘Mm This is to certify that the thesis entitled ROLE OF ALLIUM VISUAL AND CHEMICAL STIMULI IN THE OVIPOSITION BEHAVIOR OF DELIA ANTIQUA (MEIGEN) presented by Marion Olney Harris has been accepted towards fulfillment of the requirements for M.S Entomology degree in Major professor Date 20 May 1982 0-7639 LEERARY Michigan gtate University OVERDUE FINES: 25¢ per day per item RETURNING LIBRARY MATERIALS: Place in book return to remve charge from circulatton records é— // 74,37 ROLE OF ALLIUM VISUAL AND CHEMICAL STIMULI IN THE OVIPOSITION BEHAVIOR OF DELIA ANTIQUA (MEIGEN) by Marion Olney Harris A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of MASTER OF SCIENCE Department of Entomology 1982 ABSTRACT ROLE OF ALLIUM VISUAL AND CHEMICAL STIMULI IN THE OVIPOSITION BEHAVIOR OF DELIA ANTIQUA (MEIGEN) by Marion Olney Harris Stimuli emanating from onion stems play an important role in releasing Delia antiqua oviposition behavior. Surro- gate stems were used to assess the role of non-chemical stimuli from stems. In the presence and absence of chemical stimuli, flies laid more eggs on yellow stems, thereby indi- cating a role for stem hue in oviposition behavior. When visual, structural and chemical stimuli were presented separ- ately and in various combinations, combinations of all three stimuli had a synergistic effect on oviposition. While non-propyl disulfides of garlic and propyl disul- fides of onion elicit virtually identical EAG responses, their effects on oviposition are dissimilar. At concentrations where dipropyl disulfide is stimulatory, diallyl disulfide does not increase or decrease oviposition on a chemically neutral visual stimulus. Response to garlic chemicals is, however, synergized by the same factors which synergize onion chemicals. Larval feeding studies indicated that oviposition preferences reflect the suitability of the host for larval survival. To my grandmother and grandfather, Marion and Stoddard Stevens ii ACKNOWLEDGEMENTS My sincere thanks go to my major professor, Jim Miller, for his munificent and unstinting enthusiasm, for allowing me to pursue some of my wilder ideas, and for letting me learn from my own mistakes. I also am deeply indebted to Ring Cardé, who jumped ship as I was finishing my Master's Degree, but whose comments and advice were critical to my understanding of insect behavior. His incisive mind will be sorely missed. My committee members, Edward Grafius, Fumio Matsumura and Guy Bush also gave much useful advice. My thanks also go to the proletariat of the lab. Joan Harlin put in countless hours doing work of the most tedious sort, listened to my endless complaints, and kept me amused during gargantuan field experiments. John Behm converted scribbled diagrams of experimental materials into objects of art. Beatrice Gloria helped on all fronts, both skilled and unskilled, and did so with a competence and cheerfulness that made life in the lab far more pleasant. iii TABLE OF CONTENTS LIST OF TABLES LIST OF FIGURES. GENERAL INTRODUCTION . CHAPTER 1. CHAPTER 2. Synergism of Visual and Chemical Stimuli in the Oviposition Behavior of Delia antiqua . Introduction . Materials and Methods. Results. . . . . . . Stems vs. bulbs . . Surrogates vs. onion stems. Range of colors . Neutrals vs. yellow stems . Relative importance of the three stimuli. Discussion . Influence of Non—Propyl Radicals in Host-Range Determination of Delia antiqua Introduction . Materials and Methods. Growth and survival of larvae on Allium spp. . Electroantennograms (EAG' s) Oviposition tests Results. Growth and survival of. larvae on Allium spp. Electroantennograms (EAG' s) Oviposition tests . iv vi vii 34 35 4O 4O 41 43 45 45 45 48 Discussion. CONCLUSIONS. REFERENCES CITED . 53 57 59 LIST OF TABLES TABLE 1. Relative importance of above- -ground vs. below- -ground stimuli for oviposition— by the onion fly . TABLE 2. Relative effectiveness of onion stems gs. surrogate stems both with and without onion odors in stimulating oviposition by the onion fly. . . . . . . . . TABLE 3. Relative importance of visual/structural stimuli and onion odor for oviposition by the onion fly . TABLE 4. Relative importance of visual/structural stimuli and onion- (dipropyl disulfide) and bacteria-produced (ethyl acetate) volatiles for onion fly oviposition. TABLE 5. Amounts and ratios of sulfur radicals in Allium spp. varying in attractiveness to Delia antiqua (Meigen) TABLE 6. Volatile components of crushed onion (Allium cepa) and garlic (Allium sativum). TABLE 7. Growth and survival of Delia antiqua larvae reared on onion and garlic. TABLE 8. Stimulation of Delia anti ua (Meigen) oviposition by suscept1 e and non- susceptible Alliums and their primary volatile components. . . . . TABLE 9. Relative effectiveness of larval-infested and non-infested onions and garlic in stimulating oviposition by Delia antigua (Meigen) when visual and structural cues were held constant . TABLE 10. Relative effectiveness of various combin- ations of microbially produced volatiles with alk(en)yl disulfides in stimulating oviposition by Delia antiqua (Meigen) when visual and structural cues were held constant . vi FIGURE 1. FIGURE 2. FIGURE 3. FIGURE 4. FIGURE 5. LIST OF FIGURES Influence of stem hue (color) on oviposi- tion of Delia antiqua (Meigen) given various chemical stimuli. Data from 4 replicates are expressed as mean % eggs laid/treatment . Spectral reflectance curves of colored papers used in onion fly oviposition experiments. Reflectance spectra were recorded on a Bausch and Lomb Spectronic 505 . Spectral reflectance curves of neutral and colored papers used in onion fly oviposition experiments. Reflectance spectra were recorded on a Bausch and Lomb Spectronic 505 . Influence of stem brightness on ovipo- sition of Delia anti ua (Meigen) given either n-d1propy su fide and ethyl acetate or chopped onion as chemical stimuli. Data from 3 replicates are expressed as mean % eggs laid/treatment . Normalized EAG responses of male and female Delia antiqua (Meigen) antenna vs. dosage of alk(en)yl disulfides Iaaded in peanut oil. . . . vii 19 21 23 24 47 GENERAL INTRODUCTION The onion fly, Delia antiqua (Meigen), is a muscoid fly in the family Anthomyiidae. The members of this family, which consists of over 550 North American species, are quite diverse in their feeding habits, some being dung breeders, others phy- tophagous, and a small number, aquatic and predaceous (Borror and DeLong 1971). The best known members of the Anthomyiidae are, of course, the black sheep of the family, pest species which attack numerous commercial crops and cause considerable damage. With few exceptions, these pest species fall into the genus leig (revised from Hylemya by Hennig in 1974) and vary in their feeding habits. Some, such as the seedcorn fly, 2. platura (Meigen), are polyphagous while others are special- ists on groups of related plants, such as Q. antiqua on the Allium species and the cabbage fly, 2, brassicae (Wiedemann), on the Cruciferae. Virtually nothing is known about 9. antigua's existence outside of home gardens and commercial Allium fields. Inasmuch as it is an underground feeder, this lack of information may be due to its nonapparency; however, in the Netherlands where the sterile insect control technique is being used, an effort was made to determine the significance of wild hosts in main- taining_p. antigua populations. Still, no wild hosts were found (Loosjes 1976). .5. vineale, which is perhaps the most 1 common of the wild Allium species, was apparently not visited by onion flies and was not a suitable host, inasmuch as larvae placed on bulbs in the lab could not penetrate the outer scales of the bulb and therefore perished (Wolff 1973). Loosjes (1976) has reviewed the literature on the biology of D. antiqua and its association with the onion, é, 932a. While remarkably little is known about 2, antiqua pre- oviposition behavior, there is some indication that g, ggpa's influence begins well before oviposition occurs. When flies were released outside of onion fields, Loosjes (1976) found that whereas females recaptured in onion fields had been mated within 5 to 9 days after emergence, mating was signifi- cantly delayed in females recaptured outside of onion fields. The observation was confirmed by Ticheler (1971) who worked extensively with lab cultures of D. antiqua. Whether onions primarily affect the behavior of the fly or its physiology, and subsequently its behavior, is not known. Host plant odors might affect the physiology of 2. £351- 323 in several ways. Since 2. antiqua females will only mate when the ovaries are well developed (Ticheler 1970 as cited by Loosjes 1976), it is possible that in the absence of host plant odors, ovaries do not develop and that females therefore do not become receptive to males. It is also possible that the absence of odors interferes with mating by inhibiting the release of female sex pheromone. Stimulation of pheromone release by plant odors has been reported to occur in several phytophagous insects (Riddiford and Williams 1967, Renwick and Vite 1972, Rahn 1969). While there is no reported evi- dence that pheromones play a role in Q. antiqua mating behavior (which is rarely observed and not well understood), cuticular hydrocarbons have been implicated as sex pheromones in many Diptera, including those in the closely related Muscidae (Howard and Blomquist 1982). filling fields might also affect the pre-oviposition behavior of D. antiqua by serving as aggregation and mating sites for males and females. Loosjes (1976) found that 70% of the D. antiqua immigrants in an onion field arrived before the average age of mating, and that during the period when mating normally occurs, migration of males and females was reduced. Within onion fields, both male and female flies are highly responsive to synthetic onion volatiles (Vernon et a1. 1981) and chopped onion (Dindonis and Miller 1980b). The use of host plants as mating sites has been well documented in the family Tephritidae (Prokopy 1977). There is one other generally ignored, but extremely im- portant pre-oviposition behavior. Before laying eggs, females must locate and feed on protein sources (McLeod 1964). Where females find these exogenous sources of protein is a mystery. Flies have occasionally been observed feeding on flowers, grasses, bird excrement and rotting onions (Loosjes 1976, Din- donis and Miller 1980b, Kastner 1929) and accumulate in large numbers in traps baited with enzymatic yeast hydrolysate (Miller and Haarer 1981). Given the paucity of concentrated protein sources in nature, it might be expected that if Q. antiqua were to possess a finely tuned and chemically mediated host-finding behavior, it would occur here. Much work has been done on the role of secondary plant chemicals in the oviposition behavior of Q, antiqua. In 1968, Matsumoto and Thorsteinson demonstrated that dipropyl disul- fide and propanethiol, two major volatile components of onion, released oviposition behavior when applied to moist sand. Since then a number of alkyl sulfides and thiols, as well as their analogues have been assayed for activity as oviposition stimulants. The results of these studies prompted two groups (Vernon et a1. 1978, Ishikawa et a1. 1978) to independently conclude that oviposition stimulants must contain a sulfur atom possessing two unshared electron pairs and bonded to a saturated hydrocarbon chain three to five carbons long. In addition to the stimulants emitted by onions, several micro- bially produced synergists of the alkyl sulfides have been discovered, including ethyl acetate and tetramethyl pyrazine (Ikeshoji et a1. 1980). In much of the early work done on insect-plant relation- ships, it was implicitly assumed that the presence of a few key stimulants (Dethier 1954) or the absence of deterrents (Jermy 1958, Thorsteinson 1960) would indicate to the herbi- vore whether or not the site was suitable for oviposition or feeding. However, recent work has indicated that the chemical information used during host finding is considerably more diverse, and that insects respond to a summation of inputs from receptors sensitive to a range of plant constituents (Dethier 1971). Furthermore, Dethier (1982) has pointed out that if an insect is exposed to a stimulus and does not overt- ly "behave", the absence of a particular behavior does not negate the possibility that the stimulus has been perceived and that the information is being used in some manner. This expanded view of insect perceptual capabilities allows one to ask further questions about the types and diver- sity of information used by Q. antiqua when finding and accept- ing a host plant. Instead of using only the so-called "stimulatory" chemicals, might not Q. antigua be sensitive to the non-stimulatory volatile disulfides as well and base host acceptance on the ratio of positive to negative factors emitted by the plant? Furthermore, might not Q. antigua use other sensory modalities to supplement information provided by chem- ical stimuli? Onions differ from surrounding vegetation not only in chemical characteristics, but also in their structural and visual characteristics. Thus, while 2. antiqua may be capable of finding submerged unsprouted onion bulbs for ovi— position (Carruthers 1979), the redundancy which results from the use of several sensory modalities might facilitate rapid and accurate location of larval food resources. Visual and structural stimuli have been shown to be of consequence in the oviposition behavior of other Dgli§_species (Sbmme and Rygg 1972, Zohren 1968). In the following two chapters, the role of chemical and visual stimuli in the oviposition behavior of Q. antiqua is explored. In Chapter 1, I have examined oviposition behavior to determine whether visual and structural stimuli emitted by the host plant significantly modulate responses to secondary plant chemicals. In Chapter 2, I have addressed the question of whether 9. antigua is capable of perceiving volatile di- sulfides which do not fit the specifications for stimulatory disulfides, and whether the non-stimulatory disulfides con- stitute positive, negative or neutral oviposition stimuli. CHAPTER 1 Synergism of Visual and Chemical Stimuli in the Oviposition Behavior of Delia antiqua INTRODUCTION The oviposition preferences of the onion fly, D. antiqua, generally follow plant taxonomic lines, with Allium species being preferred among plant genera, and the onion, 5, ggpa, being preferred among the Allium species. Within g, 9223, plants in certain developmental or physiological states are more commonly oviposited on than others. Within an onion field, large numbers of eggs are found on onions which are rotting (Loosjes 1976), nematode and larval-infested (Loosjes 1976, WOrkman 1958, Armstrong 1924) or diseased (Crete 1972). In the spring, volunteer onions and onions which have been sown early receive a disproportionate number of eggs (Gray 1924, Sleesman 1934, Treherne and Ruhmann 1922). The proximate behaviors leading to increased oviposition on these particular onions and the stimuli which elicit these behaviors are not well understood. Some of the behaviors are undoubtably mediated by secondary plant chemicals. Loosjes (1976) found that dispersal patterns of onion flies from emer- gence sites to neighboring fields could be explained by random movement. Once flies had entered a field, their rate of move- ment slowed 2.5 fold, suggesting that flies accumulate in onion fields because of inverse klino- or orthokinetic rather than anemotactic responses to onion odors emanating from fields. Within fields, onion flies have been observed orienting anemo- tactically when stimulated by volatiles from chopped onion 8 (Dindonis and Miller 1980b, Kastner 1929); however, in both cases it was not entirely clear whether the onion odors were signalling the presence of an oviposition, feeding or mating site. Interpretation of the biological significance of results from field experiments comparing the number of flies caught in traps baited with various sulfur and non-sulfur containing chemicals (Loosjes 1976, Vernon et a1. 1981, Din- donis and Miller 1980ab, 1981ab) is also difficult, being confounded by our lack of understanding of the mating and feeding behavior of adult flies. Matsumoto and Thorsteinson (1968) did report that more females landed on screens and in- serted their ovipositors through mesh above dishes containing dipropyl disulfide and propanethiol. Thus, the sulfur- containing chemicals can mediate both pre- and post-alighting behaviors. Other workers have compared 2, antiqua oviposition pref- erences without determining whether the presented stimuli exerted their effects on pre- or post-alighting behaviors. Hough et a1. (1981) showed that onion flies laid many more eggs on nonsterile seedlings than on sterile seedlings. They suggested, based on the work of King and Coley-Smith (1969), that bacteria in the soil convert sulfoxides, given off by onion roots, into the volatile sulfides which elicit ovipo- sition. Based on extensive work analyzing volatiles from fresh onions and oviposition experiments on onion volatiles and their analogues, Ishikawa et a1. (1978) and Vernon et a1. 10 (1978) hypothesized that chemicals eliciting oviposition must contain a single sulfur atom having two unshared pairs of electrons, and bonded to a saturated hydrocarbon chain three to five carbons long. Ikeshoji et a1. (1980), noting increased oviposition on larval-infested onions in the field, investigated the chemical basis for this preference and show- ed that microbially produced volatiles, ethyl acetate and tetramethyl pyrazine, significantly increased oviposition when presented with dipropyl disulfide. However, traps bait- ed with these oviposition synergists in combination with dipropyl disulfide did not increase trap catch relative to dipropyl disulfide (Harris et al. unpublished). The two synergists may therefore play a role in post-alighting be— haviors either by causing more females to remain and oviposit on a treatment, or by causing single females to deposit more eggs than they would on treatments containing dipropyl disul- fide alone. Since none of the identified oviposition stim- ulants elicit a response equivalent to an onion slice (Pierce et a1. 1978), it might be suggested that oviposition in Q. antiqua is mediated by a complex volatile mixture rather than any single "key" volatile. The importance of complex odors in insect-plant interactions has been noted in other behav- ioral studies (Tichenor and Seigler 1980, Staedler and Hanson 1978) and is supported by electrophysiological studies (Staedler 1976, Dethier 1980). While most of the behavioral work on D. antiqua has ll concentrated on chemical stimuli, there are indications that visual and structural stimuli play a role in releasing ovipo- sition behavior. Gray (1924) found that cull onions planted so that leaves branched below the soil surface received more eggs than cull onions having a pronounced neck and foliage branching well above the soil surface. Though it is possible that chemical stimuli emanating from the differently planted culls differed, it is more likely that flies were responding to differences in visual and structural stimuli. Onion flies also seem to prefer plants which are neither extremely small (Workman 1958) nor plants which are healthy and fully devel- oped (Perron 1972). Once again, this preference could be mediated by chemicals rather than plant size; however, Hough (1981) found that females laid more eggs on taller objects (toothpicks and aquarium reed) even without the presence of chemical stimuli. From experiments with onions in mono- and mixed cultures, Muller (1969) suggested that, while long- range orientation by Q. antigua is mediated by olfactory stimuli, close-range orientation is mediated by plant shape and brightness and might be hindered by concealing onion plants among other plants. The present work addresses the role of structural and visual stimuli in the oviposition behavior of the onion fly. While no attempt was made to determine whether the stimuli being examined affected pre- or post-alighting behaviors, I have established the importance of above-ground stimuli and 12 describe some of the visual characteristics of the onion plant which are important in eliciting oviposition behavior. MATERIALS AND METHODS Small sprouted onions (variety Abundance), having bulbs 2 to 3 cm in diameter and stems 5 to 7 cm long, were cut into two parts, stem and bulb. The bulb was submerged in a cup 8 cm.in diameter and 5 cm deep, filled with 200 m1 of washed silica sand and 50 m1 of distilled water. The stem was placed in a second cup so that it rose vertically from the sand and presented a visual stimulus similar to sprouted cull onions in the field. A third cup, containing only sand and water, served as a control. The surface of every dish of sand used here and in the following experiments was imprinted with 20 holes, 5 mm deep and 2 mm.wide, which are thought to facilitate oviposition (G. Ritcey, University of Guelph). All treatments containing chopped onion were prepared as follows. A layer of sand (15 m1), 5‘ml of chopped onion and a top layer of 80 ml of sand were sequentially added to plastic dishes (8 x 8 x 2.5 cm). Dishes moistened with 20 m1 of dis- tilled water were placed in a white styrofoam frame (9 x 9 x 3.5 cm). To stabilize sand moisture, the Styrofoam frame was placed on top of another plastic dish containing 25 ml of dis- tilled water. Contact between the sand in the upper dish and the water reservoir was maintained by a filter paper strip 13 (7 x 1 cm) inserted through a 2 cm slit in the bottom of the upper dish. Treatments containing synthetic chemicals were prepared similarly. Dipropyl disulfide (Eastman Kodak,Rochester, NY, 98% pure by GLC) and ethyl acetate (Aldrich Chemical Co. Inc., Milwaukee, Wis., 99% pure by GLC) were released from Beem polyethylene embedding capsules (Pelco Electron Micro- scopy Supplies, Ted Pella Co., Tustin, CA.) submerged under 1 cm of sand in the center of the oviposition dish. Eight cm sections, cut from 5 to 8 mm diameter onion stems, were used as above-ground stimuli in experiments com— paring oviposition on real gs. surrogate stems. Surrogate stems consisted of 5 mnlID PyrexTM glass tubes cut in 8 cm sections and heat sealed to a tapered end. The open end of the tube was sealed with a cork during experiments. When called for, 7 x 0.5 cm papers were placed inside glass tubes to provide color or brightness stimuli. Red, orange, yellow, yellow-green and blue silkscreened papers (Geller Artists, New York, NY) and neutral scale (gray) papers (N 0.5 to N 9.5, Munsell Color, Baltimore, Maryland) were printed on one' side. Hence, two sheets of each color were glued back to back and then cut to size. Reflectance spectra (400 to 700 nm) of all papers used as visual stimuli in these experiments were re- corded on a Bausch and Lomb 505 spectrophotometer using a polychromatic detector and magnesium carbonate as the standard. In each oviposition dish two 6.5 cm tall stems, either onion 4 or surrogate, were placed 5 cm apart on the diagonal. 14 In experiments using onion as a chemical stimulus, treat- ments were placed in cages immediately upon preparation; however, in experiments using synthetic chemicals, treatments were first allowed to age for 8 hours. The design of all ex- periments was randomized complete block per cage per sampling time. Rather than completing all blocks of a given experiment and then moving on to another experiment, one block of each experiment was completed in random order and this cycle repeated over time. This precaution was taken to avoid training effects possible with flies (Roitberg and Propoky, unpublished) re- peatedly exposed to the same treatments. During these exper- iments (July through October, 1981), cages containing several hundred flies were placed either outdoors or in the green- house, depending on the weather. Parental stock for the cul- ture of flies used in the experiments was obtained from Dr. Freeman McEwen, University of Guelph, Ontario. When used in these experiments, flies were 8 to 10 generations removed from the field. Food was retained in cages during all exper- iments and consisted of honey, water and the artificial diet of Ticheler (1971). Treatments were generally removed after 24 hours; however, since very few eggs were laid during the first 2 days of the experiment with the range of colors pre— sented without chemical stimuli, these treatments were kept in cages for 3 to 4 days. Eggs were collected from the sand by flotation, counted, and recorded as a percent of the total number of eggs laid in the block. Sand, onion, corks, and capsules containing chemicals were discarded, and surrogate 15 stems were washed, rinsed with acetone and placed in a drying oven for 24 hours before being reused. Data generated by these experiments did not fulfill the assumption of normality either with or without transformations. Therefore, results were analyzed using the nonparametric equi- valent of the F test, Kruskal-Wallis one way analysis of var- iance by ranks (Siegel 1956). RESULTS Stems vs. Bulbs Initial experiments comparing oviposition on stems vs. bulbs included an intact onion plant (i.e. stem and bulb to— gether); however, as almost all eggs (ca. 80 to 90%) were laid on this treatment, we removed it subsequently, to force females to oviposit either on onion stems or submerged bulbs. When given only these treatments and a control, flies laid far more eggs around the stems (Table 1). While the results of this experiment did not indicate whether the stem's chem- ical or visual stimuli played the major role in eliciting oviposition, work by Pierce et al. (1978) discounted the pos— sibility that the stem's chemical stimuli were solely respon- sible. Pentane extracts of onion bulbs elicited 16 times more oviposition than did extracts of either onion stems of leaves. Surrogates vs. Onion Stems The importance of the interaction between the stem's chemical and visual/structural cues was investigated by fash- ioning a surrogate stem and comparing oviposition around 16 TABLE 1. Relative importance of above-ground vs. below-ground stimuli for oviposition by the onion—fly . Treatment Mean2 % eggs laid/treatment3 bulb submerged in sand 5.3 b stem standing in sand 94.7 a sand control 0.0 c Three replicates Means followed by the same letter are not significantly different at p s 0.05 (Kruskal—Wallis one way analysis of variance by ranks). 3 Total eggs = 786. 17 it with oviposition around an onion stem of similar shape and size, both with and without chopped onion as a below- ground chemical stimulus. While there was great variability in egg counts from block to block (Table 2), there was no significant difference between onion stems and surrogate stems as long as both were presented along with chopped onion in the sand. When no chopped onion was placed beneath the sand, onion stems (which provided their own chemical stimuli) re- ceived significantly more eggs than did surrogate stems. It cannot be assumed from these results that olfactory or gust- atory stimuli from onion stems do not play a role in ovipos- ition behavior of D. antiqua. Color stimuli from onion stems and surrogate stems were not identical inasmuch as surrogates contained yellow rather than green papers. Since yellow is more stimulatory than green to many herbivores, it may have somehow compensated for the lack of appropriate chemical stimuli in the surrogate stem. Range of Colors Surrogate stems made it possible to manipulate color stimuli while holding odor and structural stimuli constant. The question of whether vertical objects of all colors were equally stimulatory was addressed by placing red, orange, yellow, green, blue, black and white papers in glass tubes and assaying them alongside clear tubes and dishes containing no tubes. When the range of colors was presented along with chopped onion beneath the sand (Figure 1c) significantly more eggs were laid in treatments having yellow stems. Green was 18 TABLE 2. Relative effectiveness of onion stems gs. surrogate stems both with and without onion odors in stimu- lating oviposition by the onion flyl. Mean2 % eggs laid/treatment3 Below-ground stimulus Above-ground Sand + stimulus Sand Onion onion stem 16.9 b 43.3 a surrogate stem 1.0 c 36.7 a no stem 0.0 d 2.1 c 1 Five replicates 2 Means followed by the same letter are not significantly different at p s 0.05 (Kruskal-Wallis one way analysis of variance by ranks). 3 Total eggs = 6185. (a) (b) (C) ((1) FIGURE 1. l9 n-DIPROPYL onsumoe ,0_ AND ETHYL ACETATE i venom! EGGS N (I l . fimfiifliu 101 AL EGGS 5016 so ,. n-DIPROPYL DISULFIDE 25'- X PERCENI EGGS TOIAL EGGS . ‘669 8 so- CHOPPED ONION 8 v 5 g 25- m e rlnrgl IgIrllflmlrnI[£14u_ YOTAL Ems 2060 50.. NO CHEMICAL STIMULUS PE RCENY £665 25 i r‘nr‘wIIIJ—Irfiri‘lrhr—Ha 701 AL 5335 199‘ Influence of stem hue (color) on oviposition of Delia anti ua (Meigen) given various chemical stimuli (noted above). (R=red, O=orange, Y=yellow, G=green, B=b1ue, BK=black, W=white, C=c1ear, NS=no stems). Data from 4 replicates are expressed as mean % eggs laid/treatment. 20 less stimulatory than yellow, but received significantly more eggs than the remainder of the treatments. When either dipropyl disulfide or dipropyl disulfide and ethyl acetate were used, results were very similar (Figure 1a and lb). When the range of colors was run with no chemical stimulus (i.e. plain sand), essentially no oviposition occurred during the first two days of the experiment; however, on the third and fourth days, when the threshold for release of oviposition behavior had probably been reduced, flies laid significantly more eggs on yellow stems (Figure 1d). In this experiment, oviposition on green stems did not differ statistically from the remain- ing stems, and the variability of oviposition response to all treatments was much greater than in previous experiments. Neutrals vs. Yellow Stems Wavelength discrimination (color vision) is not the only explanation for the oviposition preferences heretofore seen. Spectral reflectance of the colored papers used in the exper- iments (Figure 2) show that while all colors have a character- istic pattern of reflectance at certain wavelengths (hue), they can also be distinguished by how much light they reflect (brightness or value). 0f the colors presented, yellow does not reflect nearly as much light as does white, but reflects far more than does red, orange, green, blue or black. It is possible, therefore, that the onion fly is not distinguishing between hues but is stimulated to oviposit by gray vertical objects which reflect as much light as the yellow used in our experiments.' This hypothesis was tested by comparing 21 w 4_—,,,/’/// 2 ( YELLOW S m ORANGE .4 u. :2 P BLUE RED 2 ul 3 w GREEN a. . BLACK 400 I I 700 WAVELENGTH (nm) FIGURE 2. Spectral reflectance curves of colored papers used in onion fly oviposition experiments. Reflectance spectra were recorded on a Bausch and Lomb Spectronic 505. 22 oviposition in response to yellow, orange, and a series of gray papers (Figure 3), using dipropyl disulfide and ethyl acetate in the first experiment and chopped onion in the se- cond experiment as chemical stimuli. Flies laid significantly more eggs (Figure 4) on yellow stems. Oviposition on the grays, which corresponded to yellow in terms of brightness, did not differ significantly from the darker and lighter grays, in- dicating that the increased amount of oviposition on yellow stems was due to a preference for a particular hue rather than brightness. Relative Importance of the Three Stimuli I tested the relative importance of color, structural and chemical cues both separately and in various combinations. In one experiment, chopped onion was used as the chemical stimulus and six different treatments were presented: sand alone, sand with chopped onion (chemical alone), clear stems in sand (structural alone), clear stems in sand with onion (structural and chemical), yellow stems in sand (structural and visual), and yellow stems in sand with onion (visual, structural and chemical). Combining the three stimuli had a synergistic, rather than additive, effect on oviposition (Table 3). Few eggs were laid on treatments containing only chemical, structural or visual plus structural stimuli. How- ever, the addition of a clear glass tube to a treatment con- taining only moist sand and chopped onion caused a significant increase in the number of eggs laid (2.8% + 15.8%). This suggested that the onion stem elicits oviposition not only 23 90, WHITE YELLOW N8 w . 3 E U g 50* ORANGE I” c N7 5 ‘ 8 1 I 2 N6 1/’ \ d / l m ,,_1 ' / / N3 400 1 ' 700 WAVELENGTH (nm) FIGURE 3. Spectral reflectance curves of neutral and colored papers used in onion fly oviposition experiments. Reflectance spectra were recorded on a Bausch and Lomb Spectronic 505. 25 X PERCENT EGGS 50 25 X PERCENT EGGS FIGURE 4. 24 n-DIPROPYL DISULFIDE AND — ETHYL ACETATE Y TOTAL EGGS = 1987 CHOPPED ONION Y O fi%%fifi%%fififlflgfi TOTAL EGGS = 5514 Influence of stem brightness on oviposition of Delia anti ua (Meigen) given either n-dipropyl disulfide and ethyl acetate or chopped onion as chemical stimuli. (BK=black, N2 — N8 = neutral grays of values 2 to 8, Munsell color notation, W=white, O=orange, Y=yellow, C=c1ear, NS=no stems). Data from 3 replicates are expressed as mean % of eggs laid/treatment. 25 TABLE 3. Relative importance of visual/structural stimuli and onion odor for oviposition by the onion f1y1. Mean2 % eggs laid/treatment3 Chemical stimulus Visual/structural stimulus nothing chopped onion no stems 0.0 d 2.8 c clear glass stems 0.4 dc 15.8 b yellow glass stems 2.8 dc 78.1 a 1 Four replicates 2 Means followed by the same letter are not significantly different at p s 0.05 (Kruskal-Wallis one way analysis of variance by ranks). 3 Total eggs = 2832. 26 because it reflects light of stimulatory wavelength but also because it presents a structure which plays a role in pre- oviposition behaviors. Preliminary observations in the lab and in the field indicate that flies almost invariably spend several minutes walking up and down onion and surrogate stems before probing the soil or sand with their ovipositors and eventually laying eggs. Such a stem walk has also been ob- served during the pre-oviposition behavior of Q. brassicae (Zohren 1968). An alternative, perhaps more conservative, explanation for increased oviposition on glass stems should also be considered. Glass stems are not entirely transparent: specular reflection consisting of all wavelengths of light occurs at the outer surface of the glass (Snodderly 1979). This reflection could impart visually discernable qualities to an apparently transparent object. If this were the case, oviposition on glass stems might have occurred because of visual rather than structural characteristics. A comparison of the relative importance of color, struc- ture and chemical stimuli was also run using dipropyl disulfide, the major volatile component of onion, and ethyl acetate, a microbially produced synergist of oviposition. Again, the combination of color, structural and chemical stimuli elicited the most oviposition (Table 4). The best oviposition stimuli reported in the literature, dipropyl disulfide and ethyl ace- tate (Ikeshoji et a1. 1980) performed poorly without the addi- tion of visual and structural cues. As reported by Ikeshoji et a1. (1980), ethyl acetate did not have an effect on its 27 TABLE 4. Relative importance of visual/structural stimuli and onion-(dipropyl disulfide) and bacteria-produced (ethyl acetate) volatiles for onion fly oviposition . Mean2 % eggs laid/treatment3 Chemical stimulus4 Visual/structural PrzSz stimulus “we ETAC Pl'232 + ETAC no stems 0.0 d 0.1 d 0.1 d 0.1 d clear stems 0.4 d 1.2 d 1.2 d 7.1 c yellow stems 9.2 c 9.9 c 22.6 b 47.4 a Three replicates Means followed by the same letter are not significantly different at p S 0.05 (Kruskal-Wallis one way analysis of variance by ranks). 3 Total eggs = 3597. 4 ETAC = ethyl acetate, PrZS2 = dipropyl disulfide. 28 own but did synergize dipropyl disulfide, causing more than twice as much oviposition in treatments with yellow stems, and over 5 times as much oviposition with glass stems. The addition of structural cues (glass stems) to the best chemical stimulus again caused a significant increase in oviposition (0.7% + 7.1%). DISCUSSION Stimuli emanating from onion stems play an important role in releasing oviposition by D. antiqua. The quality (hue) rather than the quantity (brightness) of light reflected by the stem influences oviposition responses. This contradicts the work of Muller (1969) who suggested that close-range orien- tation in D. antigua is mediated by plant shape and bright- ness rather than hue; however, Mfiller's conclusions concerning the importance of brightness for oviposition were based on an experiment measuring landing frequency of D. antiqua flies on colored disks mounted horizontally and baited with onion juice. The observed landing response may have been unrelated to oviposition behavior or may have been only one of a series of behaviors resulting in oviposition. Q. antiqua preference for yellow stems was expressed in the presence of several different chemical stimuli and also when no chemical stimulus was placed in the sand. Although color perception is not well understood it has been suggested that the preference for yellow (rather than green) shown by many herbivores (Kennedy, Booth and Kershaw 1961, Mbericke 1969, Prokopy and Boller 1971) occurs because yellow reflects 29 more light than green in the wavelength range (540 to 560 nm) where green foliage reflects maximally. Thus, yellow might present a "supernormal foliage-type stimulus" (Prokopy 1972). Preliminary observations comparing 2. antiqua behavior on dif- ferent color stems indicate that yellow stems elicited more oviposition behavior because of effects on both pre- and post- alighting behaviors. Treatments combining chemical and visual/structural char- acteristics have a synergistic, rather than additive, effect on oviposition. The mechanism behind this synergistic effect is not known but could be due to the two stimuli having either a sequential or simultaneous effect on oviposition behavior. The fly might, for instance, respond anemotactically or klino- or orthokinetically to olfactory stimuli, and upon reaching the oviposition site, encounter visual/structural stimuli which, in the presence of chemotactile stimuli, cause the fly to oviposit. 0n the other hand, the presence of olfactory stimuli might cause the fly to "pay attention to" visual stimuli. Vaidya (1969) found that the lemon butterfly, Papilio demoleus, did not respond to isolated citrus odors or colors; however, when an odor source was placed near colored papers, butter- flies responded to colors with a characteristic drumming re- sponse and did not "search" for the source of the odor. Be- havioral observations of individual onion flies on treatments presenting a single stimulus or various combinations of stimuli would be very useful, and might explain how the various stim- uli mediate oviposition behavior. 30 The system used by Q. antiqua to select an oviposition site is therefore fairly flexible: flies appear to discrim- inate between chemical stimuli in the absence of visual/struc- tural stimuli (Vernon et a1. 1978, Ikeshoji et al. 1980, Ishikawa et a1. 1978), and between plant color and size in the absence of chemical stimuli (our work and Hough 1981). This flexibility could be highly adaptive, as the visual, structural and chemical stimuli emitted by A. ggpa_vary with- in fields and change drastically throughout the growing season. In Michigan, flies in the first generation encounter seedling onions and a few sprouted cull onions, whereas second gener- ation flies generally encounter onions with tall, thick fol- iage and large bulbs. By the time the third generation has emerged and mated, onion fields may have been harvested. At this time, the only onions available for oviposition are those left behind by the harvester. These cull onions appear in a variety of states, some being sprouted and others sub- merged beneath the soil surface and therefore presenting no visual stimuli. Volatile composition also changes qualita- tively and quantitatively throughout the growing season and when plants are damaged (Saghir et a1. 1965, Ikeshoji et a1. 1980). Although Q. antigua oviposition preference does gener- ally follow plant taxonomic lines, with the Allium species being preferred among plant genera and A, gepa_being pre- ferred among the Alliums, flies do occasionally ignore taxon- omy and oviposit on non-Allium plants such as corn (Workman 31 1958) and non-A. cepa plants such as A. ampeloprasum L. (Muller 1969), A. fistulosum L. (Perron 1972) and A. sativum L. (Chap- ter 2). Rausher et a1. (1981) have suggested that some insects do not order oviposition preferences along taxonomic lines but instead discriminate among all plants by using a few charac- teristics that cross plant taxonomic lines. The onion fly's preferences could therefore appear to follow taxonomic lines inasmuch as Allium species have a unique secondary plant chem- istry and vary chemically within the genus (Saghir et a1. 1964, 1966); however, preferences for certain visual/struc- tural stimuli would cross plant taxonomic lines, and might cause the fly to oviposit on plants other than A, 3333. Inasmuch as visual and structural stimuli supplement the information conveyed by chemical stimuli, the use of sev- eral sensory modalities during host finding should be adap- tive. Although the onion fly may be able to locate a submerged, non-sprouted cull onion solely by the use of chemical stimuli, visual stimuli would give an accurate indication of where an onion is located. Such accuracy may be critical to the sur- vival of the first instar larvae which move slowly (Gray 1924) and are subject to desiccation in dry soils (Workman 1958). Visual and structural stimuli might also provide information about the size of the resource available for larval develop- ment, and thereby serve as an indicator of how many eggs may safely be laid on a particular resource. In addition, it is conceivable that the spectral reflectance characteristics of the stem convey information about the physiological state 32 of the onion (whether it is stressed by nutrient or moisture deficiencies or by disease) and therefore, give some indica- tion of the plant's chemical and physical defenses. The practical significance of this information is three- fold. Firstly, the observation that the onion fly has a flex- ible system for locating oviposition sites should indicate to plant breeders that efforts to impart resistance to onions should be concentrated on toxic rather than deterrent charac- teristics of the plant. While some flies may leave a field when confronted with "deterrent" or less chemically stimula- tory plants, others may remain in the field, and accept plants which are visually but not chemically stimulatory. Secondly, incorporating visual stimuli into the cone trap which is pre- sently being used in Michigan for monitoring onion fly popula- tions (Miller unpublished) might improve the ability of this trap to detect low populations. Finally, an understanding of host finding behavior may allow the development of cultural methods for control of 2, antigua. Such methods might utilize trap crops consisting of visually "conspicuous" sprouted cull onions which would be more preferred for oviposition than commercial onions (Gray 1924, Armstrong 1924). Trap crops might also serve to concentrate Q. antiqua eggs in one area, and thus, make more effective releases of the egg predator Aleochara bilineata (Gyllo) (E. Groden, personal communica- tion). 2. antiqua oviposition might also be reduced if come mercial onions were made less visually "conspicuous". This might be achieved by the use of cover crops which could be 33 planted either during the growing season or after harvest of onion fields. CHAPTER 2 Influence of Non-Propyl Radicals in Host-Range Determination of Delia antiqua 34 INTRODUCTION Like many other herbivores, host range in the onion fly is determined by the female's propensity to oviposit on cer- tain plants rather than limited larval feeding habits (Loosjes 1976). The stimuli responsible for the selective oviposition behavior of the female are not entirely understood. In Chap- ter 1, it was shown that visual and structural characteristics of host plants may play an important role. However, given the genus Allium, one is confronted with plant species which are remarkably similar in both color and morphological charac- teristics. The secondary chemistry of the Allium species is also quite similar, being dominated by large amounts of alkyl cysteine sulfoxides and y-glutamyl peptides which, upon cell disruption, yielda series of alk(en)y1 substituted mono-, di-, and trisulfides, as well as thiols (Whitaker 1976). Yet, in spite of the apparent similarities, D. antiqua con- centrates oviposition on Allium cepa, though some less pre- ferred species, such as A. fistulosum, are equally, if not better, suited for larval development (Perron and Jasmin 1963). Allium species may be separated on the basis of more subtle differences in secondary chemistry. Proportions of major alk(en)yl radicals differ significantly among Allium species: A. ggp§_contains primarily propyl radicals attached to mono-, di-, and trisulfides, while A. sativum contains primarily 2ww mum mama .mwmmn unwwm3_£mmnm co .Ammmav EmscmsB cam cmfimmum mo sump no ummmm N N l I I .uxmu CH fim>ww maoflumenmmm Go mommm AmammcH>.< cam as>flumm .¢v ummma ou.Ammmo .wuomuuum umoE Eoum poxcmh mum mowooam BSHHH< H Humo.o m.o u: o.H N.o .A Es>wumm .« owaumw I cowco andm.o o.n No.0 m.n w.m .A mamm¢a> .¢ paws I aowco H” m.H nu o.H o.m q.oa .A ESmoHsumHm .< wcwnocsn H” m.H :: m.o m.m m.q .A 85mmumoammmm .fl xmmH 7 I ,3 H” ©.H nu N.o w.o o.H .uuos Showsoamomm .4 uoHHmnm H” m.s -- m.o m.o m.s .A mass .« sosso mHmoHumH Humoum Icocuahmoum owumu amammoumum Hmcmmoumua axsuwa ammoum mmfiommm BbHHH< mama aoEEoo «Aam\~sov maHeauomuuum fig wcwhum> .amm EdHHH< aw mamofiumu namHSm mo mowumn paw mud508¢ .m mqm¢8 38 radicals are not well correlated with different degrees of attractiveness. A. fistulosum, which enjoys a certain degree of resistance to D. antigua due to lower egg deposition (Perron et a1. 1958), produced two times more volatile propyl radicals than does onion, A. 3223. Dethier (1980) has suggested that herbivore specialists show discriminatory behavior not because of "peripheral in- competence" but rather because they have evolved receptor types (including deterrent receptors) which are sensitive to a multitude of secondary plant chemicals. This suggests, that D. antiqua might be sensitive not only to those sulfides which are most stimulatory, but also to those which have no apparent effect on oviposition behavior when tested in isola- tion. 2. antiqua might therefore respond to the summation of both stimulatory (propyl) and non-stimulatory (non-propyl) inputs. Indeed, comparisons of ratios of propyl to non-propyl (methyl, 1-propenyl, 2-propenyl) radicals show a better cor- relation with ovipositional attractiveness: whereas A. EEEE has a ratio of 7:1, the ratio switches to 9:1 in A, fistulosum, and to 0.15:1 in A, sativum (Table 5). This would seem to indicate that the non-propyl disulfides play some role in oviposition behavior, be it as deterrents or simply as less stimulatory disulfides. This chapter addresses several questions concerning the nature and diversity of chemical information used by D, antiqua when locating and accepting a host plant. Two Allium species, A. cepa and A. sativum.were used to investigate these .AommH .memuH53.Boumv mommm pom: Eoum mucmcogEoo mHHumHo> Hmuou mo mwmucmoumm H Ase/Ems .3 Nss-mm Nmm-- Hv Hv Hv oHHusw Emma .$ HV Nmuo NmHIN quum Nmmu¢q Gowao o, mUHmHsmHu muHMHsmHu 3 8333s 15% 8333s :83 03383 HwHHMHu HhHHm HxnumEHu HSSUmE HumohmHu mmHommm EDHHH< maznomzoo MHHH H.AE:>Humm ESHHH .o MHmHumm ESHHHH>HDm Hogan on whom Hm>H>HSm Hm>HMH mmm9m2ma .oHHumw cam GoHGo co umummu mm>umH manucm mHHmn mo Hm>H>Hsm paw nusouu .m mHmHuomamou m ppm N .H uamEHummxm pom «mom paw Hum .mmmm u wwwo Hmuoe 1V .Amxamn up mocmwnm> mo mHmszcm hm3 mao wHHHm3umemaqu no.0 w e um unmuwmmHu mHuamonHame uoc mum mumuumH mEmm mnu ma umaoHHom mGEbHoo :HSuH3 mammz m .ssHmHssss HAHHsHs u NmNHs .mssmsssss Hssossss u NmNss N .ucmEHHmmxm some mo moumoHHmmu Hsom H Annamv n m.n us- In- Houucoo HNmNH QHumnuchm + mnHan ummmosu mucmHm mHonz msHQEHum Hobo m .sxm N .sxm H .sxm N .HmquGOQEoo mHHumHo> humEHum HHmzu was mmmmmwfl oHnHummomSwnaoc unm mHnHumoomsm kn nowuHmomH>o Acomezv mstucm mHHmQ mo GOHumHSEHum .m mHm