Vespula germanica Foragers Do Not Scent-Mark Carbohydrate Food Sites

C 2005) Journal of Insect Behavior, Vol. 18, No. 1, January 2005 (
DOI: 10.1007/s10905-005-9344-x

Vespula germanica Foragers Do Not Scent-Mark Carbohydrate Food Sites Jennifer M. Jandt,1 Laura Riel,1 Benjamin Crain,1 and Robert L. Jeanne2,3 Accepted March 16, 2004; revised August 4, 2004

We tested whether foragers of Vespula germanica leave a chemical attractant at a carbohydrate food site. Foragers were allowed to make 50 or 100 visits to a feeder filled with a rich, unscented fructose solution. We then gave them a choice between this and an identical but unvisited feeder. There was no difference between the number of individuals that chose the visited feeder and those that chose the unvisited one. Instead, wasps showed a tendency to choose a feeder based on its relative position on the feeding stand. We conclude that foragers of V. germanica do not scent-mark food sites. Instead, in the absence of food odor, they rely on local visual cues to relocate carbohydrate food sources. KEY WORDS: Vespula germanica; social wasp; attractant; scent-marking; visual cue; foraging behavior.

INTRODUCTION At abundant sources of nectar, such as flowers that can replenish nectar quickly, honey bees and bumble bees leave attractant scent marks as signals that save foragers the time and energy of searching for new 1Department

of Entomology, University of Wisconsin, Madison, Wisconsin. of Entomology and Zoology, University of Wisconsin, Madison, Wisconsin. 3To whom correspondence should be addressed at Department of Entomology, University of Wisconsin, 1630 Linden Drive, Madison, Wisconsin 53706. e-mail: jeanne@ entomology.wisc.edu. 2Departments

19 C 2005 Springer Science+Business Media, Inc. 0892-7553/05/0100-0019/0


20

Jandt, Riel, Crain, and Jeanne

food sources (von Frisch, 1967; Free, 1968; Cameron, 1981; Schmitt and Bertsch, 1990). Because vespine wasps frequently feed at copious food sources, such as carrion or honeydew, that do not become depleted after the first visit or two (Duncan, 1939; Raveret-Richter, 2000), it may be similarly advantageous for them to leave an attractant behind at the food site. Odor and vision provide the primary cues that initially attract foraging wasps to a food source. In some contexts, olfactory cues associated with the food are more important than visual cues in inducing landing (RaveretRichter and Jeanne, 1985; Chang, 1988). Many vespine wasps respond to particular odors when searching for food (Jander, 1998; Landolt, 1998). In other circumstances, visual cues, such as colors (Sharp and James, 1979; Chang, 1988) and the presence of other wasps (D’Adamo et al., 2000), serve as cues indicating new food sources. There is some evidence that Vespula germanica foragers leave attractant chemicals at carbohydrate food sources. Beier (1983) reported that foragers chose a feeder that had been heavily visited for 5 days over a clean one placed next to it, and concluded that this was due to scent-marks left behind by previous foragers. However, it is unknown whether this chemical marker can be perceived after one or a few wasps visited, as would be expected if the marker were a signal. Overmyer (1995) found that V. germanica foragers were more likely to enter a feeder and to do so with decreasing hesitation after the feeder had been visited at least 46 times. She concluded that attractive chemicals left by foragers facilitated landing at the food source. However, it is unclear whether the behavioral changes she observed translate into an increased likelihood of choosing the visited feeder if given a choice. Here we address the question left unresolved by these studies: Can wasp foragers mark a carbohydrate food source with a chemical that acts as a signal to attract conspecifics and/or nestmates? If it were a signal (Seeley, 1989), it seems reasonable to expect the chemical to be applied and to be attractive after one or a few visits by a forager to a rich food source. Alternatively, a scent-mark may act as a cue (Seeley, 1989), left inadvertently by a forager where she has fed. In this case, we could expect that the chemical might not attract others unless it was in a fairly high concentration. Our approach, therefore, was to start by determining whether a feeder that had been visited 50 or 100 times was more attractive in a choice test than a clean one. If so, the second step would be to determine the minimum number of visits required to make a feeder attractive. This is the first direct test of the hypothesis that V. germanica foragers mark rich food sources with an attractive scent.

Vespula germanica Foragers Do Not Scent-Mark Food Sites

21

METHODS Colonies and Field Site Three V. germanica colonies were located behind the brick facing walls of three buildings on the campus of the University of Wisconsin-Madison in Dane County, WI. All colonies were separated by at least 50 m, but were within 200 m of one another. Data were collected between the hours of 08 00 and 17 00 from 27 August to 9 October 2002, when colonies were heavily foraging for carbohydrates (Akre et al., 1980).

Experimental Set-Up Foragers from each nest traveled 15–100 m to a feeding station, where they were tested. At each station, a feeder was placed in the center of a 30 cm × 30 cm platform on a 60-cm-high metal tripod (Fig. 1a). The feeders (Fig. 1b) were constructed from 4.5-cm-diameter round, seamless, tin boxes. A rectangular opening approximately 3 cm × 1 cm was cut into the side of each feeder and a 2.5 cm × 7.5 cm glass microscope slide was placed in the opening to act as a landing platform and ramp leading into the feeder. The feeder was filled with enough 2 M fructose solution to cover the lower end of the slide (Fig. 1c). A square piece of yellow construction paper (11 cm × 11 cm) placed under each feeder served as a visual cue. Because wasps orient upwind to chemical attractants, the openings of the feeders were directed downwind. To avoid contamination from fingerprints or oils, experimenters handled the feeders only while wearing latex gloves.

Forager Training All unmarked wasps that arrived at the feeding tray on their own were considered trained and were captured, cooled, and marked on the thorax using Deco Color paint pens. Each wasp received a color code that corresponded to a unique number, allowing it to be easily identified upon visiting the feeder. When wasps finished feeding, they were observed to fly straight back to their nest. Therefore, we were able to determine to which nest each wasp belonged by taking the bearing of her postfeeding flight. To increase the rate at which foragers discovered and learned the location of the feeder, wasps were caught in vials as they exited the nest, cooled on ice until they became immobile, then placed in groups of four to six

22

Jandt, Riel, Crain, and Jeanne

Fig. 1. Feeding station and feeders, (a) Feeding station set-up. The original position of the test feeder is indicated with dotted lines. Solid lines indicate feeder positions in place for choice tests, (b) Side view and (c) cross sectional view of feeding dish with the protruding slide, acting as a ramp to the solution.

on or next to a feeder containing 2 M fructose solution scented with 0.1% pure vanilla extract. To encourage wasps to feed as they warmed up, the top 14 cm of a 2-L plastic soda bottle was cut off, covered with black plastic, and placed over the wasps to confine them with the feeder. After approximately 10 min, the enclosure was removed and the wasps were allowed to depart. Preparing the Test Feeder To control for intercolony effects, only members of a single colony were allowed to participate in each trial. All other wasps arriving at the

Vespula germanica Foragers Do Not Scent-Mark Food Sites

23

feeder were captured and put on ice until the end of the trial. After at least two individuals were trained from the same nest, a single feeder with a clean slide (the Test Feeder) was filled with unscented 2 M fructose solution and placed in the center of the feeding tray on top of a clean paper square. The feeder top was covered with a round 4-cm-diameter piece of filter paper. Wasps were allowed to make either 50 or 100 visits to the Test Feeder, after which all wasps except one were captured and put on ice. The remaining wasp was subjected to a choice test. Using only one wasp at a time eliminated the influence of local enhancement on the decision of the wasp (D’Adamo et al., 2000).

Choice Test To begin a choice test, the Test Feeder was moved to one side of the tray and a Control Feeder was placed next to it on a clean paper square (Fig. 1a). The Control and Test Feeders were positioned to be approximately 6 cm from each other and equidistant from the center of the tray. The Control Feeder was filled with unscented 2 M fructose solution and provided with a clean slide and filter paper. The only difference between these two feeders was that one had been visited either 50 or 100 times, while the other had not been visited at all. If there was any air movement, the feeders were lined up cross wind at the beginning of the choice test, so that a wasp approaching from downwind had an equal chance of perceiving and encountering both feeders (Fig. 1a). This starting orientation of the feeders on the tray was maintained for the duration of the trial, regardless of changes in wind direction. For each trial, the test wasp was allowed to choose between the Test Feeder and the Control Feeder on 20 consecutive visits. To count as having made a choice, the forager had to land on the slide and walk all the way into the feeder. To control for any bias that the position—right vs. left as seen from downwind—of the previous choice may have had on subsequent choices, we switched the position of the two feeders after each visit. If the Control Feeder was visited, its slide was replaced as soon as the wasp departed, so it remained clean throughout the trial. Filter paper and paper squares were replaced if they became contaminated with fructose solution or otherwise soiled. Choices were not included in the analysis unless feeders had been switched and the Control Feeder had a clean slide before the wasp approached the feeding station. For each trip, the forager’s choice, as well as her arrival and departure times, were recorded. Temperature, humidity, wind speed, and wind direction were also recorded for each test. If the feeding site was disturbed or the weather became unfavorable, the test

24

Jandt, Riel, Crain, and Jeanne

was ended early. Data were not collected during poor weather conditions such as rain or temperatures below 10◦ C.

Data Analysis Twenty-three foragers, 14, 5, and 4 from the three nests, were tested. We used Fisher’s Exact Test to test for the effects of different nests and distances to the feeding station. The first trip was analyzed separately from the rest because it was the first time the wasp was confronted with a choice, and therefore the only na¨ıve choice between a visited and unvisited feeder. A chi-square test, with the correction for continuity (Sokal and Rohlf, 1981), was used to test the null hypothesis that wasps chose feeders randomly on the first trip. The alternative hypothesis was that feeder choice was based on the presence of a scent-mark. Binary logistic regression was used to determine if the first response differed between feeders visited 50 vs. 100 times (Chatterjee et al., 2000). Binary logistic regression was also used to test if the strength of the mark decayed from the time of the last departure during preparation of the Test Feeder until the first choice, and thus, whether that interval affected the first choice. Statistical computer software, SAS and Minitab v.13
, were used to run Fisher’s Exact Tests and regression analyses. On subsequent visits, wasps had a second variable, position, that could potentially influence their decision. The null hypothesis remained the same, but now two alternative hypotheses were tested: (1) feeder choice was based on the presence of a scent-mark vs. (2) choices were based on the position of the feeders. In order to test these hypotheses with a chi-square test, the second choice was analyzed separately. The first alternative hypothesis was tested by comparing the number of wasps that chose the Test and Control Feeders on the second visit. The second hypothesis was tested by comparing the number of wasps that chose the feeder in the same position on the second choice as they did on the first with those that chose the feeder in the other position. This also tested whether a wasp could remember the position of a food source based on only one visit. Fisher’s Exact Test was used to test for an interaction between the presence of a scent-mark and the position of the feeder on the second visit. Because successive choices on trips 2–20 were dependent on one another, we used Markov chain analysis (Lehner, 1996) to test the scent-mark and position hypotheses on all 20 trips (per wasp per trial). Results from all individuals were combined in order to satisfy the requirements of this analysis (Slater, 1973). The probability of choosing the Test Feeder was calculated for each individual, where P = 1 if a wasp chose the Test Feeder, C

Vespula germanica Foragers Do Not Scent-Mark Food Sites

25

P = 0 if a wasp chose the Control Feeder, and P = 0.5 if wasps chose randomly. The probability of returning to the feeder in the same position as the one visited on the previous trip was calculated for each individual, where P = 1 if a wasp chose the feeder in the same position all 20 times, P = 0 if a wasp switched position each time, and P = 0.5 if she chose randomly.

RESULTS Forager Behavior The average time (±SE) required for the Test Feeder to be prepared was 45.75 (±5.56) min for 50 visits and 86.87 (±4.58) min for 100 visits. During feeder preparation, we observed nothing to suggest that foragers were performing any specialized scent-marking behavior at the feeders. No body parts were dragged or daubed on the surface of the slide, nor was there anything unusual about the walking gait. However, slides that had been visited 50 or 100 times had a clear residue on them, visible in glancing light as an irregular smear. In a typical trip to the feeding station, the wasp flew straight to the feeder; 15% of arriving wasps circled the area once or twice before landing, and 92% landed directly on the slide. After landing, the wasp walked into the feeder via the protruding slide. She fed for 30–60 s inside the feeder, walked out on the slide, and took off. Filter paper and paper squares, therefore, rarely needed to be changed during a trial.

First Choice There was a marginally significant colony effect (P = 0.054). In order to obtain an adequate sample size for the remaining analyses, data from all three nests were pooled. Nest was blocked in all regression analyses. The distance to the feeding station did not affect whether the wasp chose the Test or Control Feeder (P = 0.114), nor did the date that wasps were tested (P = 0.302, Odds Ratio = 1.05, CI = 0.96–1.14). Wasps were not more likely to choose the Test over the Control Feeder (Table IA), nor were they more likely to choose the Test Feeder visited 100 times than they were the Test Feeder visited 50 times (P = 0.258, Odds Ratio = 0.98, CI = 0.95–1.01). Therefore, we cannot reject the null hypothesis that foraging wasps chose randomly on their first choice. The average interval between the last departure during preparation of the Test Feeder and a wasp’s first arrival at the choice test was 129.4

26

Jandt, Riel, Crain, and Jeanne Table I. Results of Choice Tests Feeder No. visits A. First choice 50 100 Total B. Second choice 50 100 Total

Test

Control

χ2

P

6 7 13

7 3 10

0.00 0.90 0.17

1.00 0.34 0.68

6 3 9

7 7 14

0.00 0.90 0.70

1.00 0.34 0.40

Note. Data are numbers of wasps coming to each feeder after it had been visited 50 or 100 times.

(±22.5) s. The duration of this interval did not influence whether the wasp chose the Test or Control Feeder (P = 0.968, Odds Ratio = 1.00, CI = 1.00–1.00). Thus, there was no evidence for decay of the scent-mark during this interval.

Second Choice For their second choice, wasps were not more likely to choose the Test over the Control Feeder (Table IB), nor were they more likely to choose the Test Feeder visited 100 times than they were the Test Feeder visited 50 times (P = 0.425, Odds Ratio = 1.01, CI = 0.98–1.05). Therefore, we cannot reject the null hypothesis that wasps chose randomly. However, wasps were more likely to return to the feeder in the same position (right or left) they chose on the first visit than to choose the feeder in the other position (same = 20, different = 3; χ2 = 11.83, P < 0.001). This supports the alternative hypothesis that wasps’ second choice is based on the position of the feeder. There was no evidence for an interaction between the position of the feeder and the presence of a scent-mark (P = 1.00).

All Choices One trial was omitted when analyzing all choices because the feeding station was damaged after the third visit. Markov chain analysis showed that wasps were more likely to switch between the Test and Control Feeder on subsequent visits, instead of continuing to go to the same feeder

Vespula germanica Foragers Do Not Scent-Mark Food Sites

27

Table II. Transition Matrices of Choices for 22 Trials A.

Following choice

Previous choice

Test

Control

Test Control

56 142

150 56

χ2 = 80.12, p < 0.001 B.

Following choice

Previous choice

Left

Right

Left Right

144 57

55 148

χ2 = 80.19, p < 0.001 Note. Including all choices made in each trial (n = 404 visits).

(Table IIA). However, the position of a wasp’s previous choice was a reliable predictor of its following choice (Table IIB). The mean probability of returning to the Test Feeder was 0.503 (±0.024, n = 22; Fig. 2a), and for the feeder in the same position as on the previous trip was 0.733 (±0.042, n = 22; Fig. 2b). DISCUSSION V. germanica foragers did not choose visited feeders over otherwise identical nonvisited ones. This result, combined with the absence of any evident scent-marking behavior at the feeders, indicates that the visible residue left on the glass slides by foragers is not a signal that is attractive to subsequently arriving conspecific foragers (Lloyd, 1983; Seeley, 1989). A true food-site-marking pheromone should be attractive after one or a few visits. Instead, this residue is probably an accumulation of cuticular hydrocarbons inadvertently tracked onto the slides by the tarsi that gradually becomes attractive after large numbers of forager visits (Beier, 1983; Steinmetz et al., 2003). Even if attractive chemicals were beginning to accumulate at the feeder, the position of the feeder was a stronger factor on subsequent visits. Instead of responding to the chemical left at the food site, our wasps tended to choose a feeder at random on their first visit, learn its position, then show a strong bias for that position for the duration of the trial. Thus, in the absence of food odor, they relied on local visual cues for subsequent returns. If wasps had merely been confused because the

28

Jandt, Riel, Crain, and Jeanne

Fig. 2. The distribution of probabilities of (a) choosing the Test Feeder and (b) returning to the same position as the previous visit for each wasp.

visited feeder had been moved, they would not have made choices based so strongly on position, and there would have been an interaction between the two factors. Other species of vespid wasps also make use of local visual landmarks when returning to food sources (Takagi et al., 1980; RaveretRichter and Jeanne, 1985). Also, before taking off from a food source that has not been exhausted, wasps perform an orientation flight, during which they apparently learn the landmarks in the area, especially if they are gathering food in a new location with new landmarks (Raveret-Richter, 2000).

Vespula germanica Foragers Do Not Scent-Mark Food Sites

29

Our results do not support those of Beier (1983), who found that, when given a choice between a previously visited feeder and a clean one, V. germanica foragers chose to enter the visited one. However, in Beier’s study the visited feeder had been entered by hundreds, if not thousands of foragers prior to the test; thus, the accumulation of chemicals may have been heavy enough to give the feeder the familiar odor of the colony. This odor could be perceived by the human nose and overpowered the scent added to the sugar solution. Two additional factors could have affected Beier’s results. First, it is possible that during the 5 days of training to the original feeder, wasps in Beier’s study may have learned its position, since it is not clear whether the original feeder was moved. Second, unlike our wasps, Beier’s wasps were tested as a group instead of individually; therefore, the attractive effect that local enhancement has on foraging wasps (D’Adamo et al., 2000) could have influenced wasp choice. Overmyer (1995) reported changes in the behavior of wasps approaching a single feeder as the number of previous visits increased, particularly after 46 visits. Wasps were more likely to enter heavily visited feeders and did so in less time than if the feeder had not been visited. Although accumulation of cuticular hydrocarbon residues at a food resource may facilitate landing/acceptance of a feeder after it has been visited many times, the results here have shown that even 100 visits is not sufficient to make the feeder more attractive than a clean one. The results of Beier (1983) and Overmyer (1995) suggest that there is a dose response to this phenomenon, similar to the dose response seen in nest-entrance chemical trails (Butler et al., 1969; Jandt et al., in press), but that was not tested here. The slight bias we found toward the visited feeder in the first choice at 100 visits may represent the beginning of a dose response, but it was not significant; in any case, the bias shifted to the clean feeder in the second choice. Unlike yellowjackets, honey bees and bumble bees leave attractive cues at the food site (von Frisch, 1967; Free, 1968; Cameron, 1981; Schmitt and Bertsch, 1990). Honey bees attract newcomers to a food source by releasing a pheromone via the Nasanov gland (von Frisch, 1967; Free, 1968), while bumble bees appear to deposit the attractant as they groom themselves before taking off from the feeder (Cameron, 1981). Since bees typically forage in patches of flowers, an attractant that distinguishes a highly rewarding flower from the rest may decrease search time on return visits (Cameron, 1981). Wasps, on the other hand, generally feed at abundant food sources (e.g., carrion and ripe fruit). Since each has a unique combination of visual (e.g., landmarks or colors) and olfactory cues, an attractant may not provide a significant additional advantage in relocating a rewarding food source.

30

Jandt, Riel, Crain, and Jeanne

It is important to emphasize that the focus of this study was on carbohydrate foraging. Wasp marking behavior may be different at protein or other short-lived food sources, just as wasps react differently to visual cues at carbohydrate vs. protein resources (D’Adamo and Lozada, 2003). ACKNOWLEDGMENTS We thank Lisa Chanbusarakum, Ken Howard, Andy Bouwma, and Reid Maier for help in collecting nests and Lisa Chanbusarakum for help in gathering data. We are grateful to Ken Raffa and Rick Lindroth for their suggestions in planning. We thank Andy Bersch and Kevin Gross for their help with statistical analyses. Cristie Hurd, Andy Bouwma, Ken Howard, and Christine Curry made valuable suggestions in improving the manuscript. Research supported by Hatch grant no. 4433 to RLJ and by the College of Agricultural and Life Sciences, UW-Madison. REFERENCES Akre, R. D., Green, A., MacDonald, J. F., Landolt, P. J., and Davis, H. G. (1980). Yellow jackets of America North of Mexico, U.S. Department of Agriculture, Agriculture Handbook No. 552, 102 p. Beier, W. (1983). Beobachtungen und Experimente Zur Orientierung der Deutschen Wespe (Paravespula germanica) am Futterplatz. Zool. Beitr. 28: 321–348. Butler, C. G., Fletcher, D. J. C., and Watler, D. (1969). Nest-entrance marking with pheromones by the honeybee- Apis mellifera L., and by a wasp, Vespula vulgaris L. Anim. Behav. 17: 142–147. Cameron, S. A. (1981). Chemical signals in bumble bee foraging. Behav. Ecol. Sociobiol. 9: 257–260. Chang, V. (1988). Toxic baiting of the Western Yellowjacket (Hymenoptera: Vespidae) in Hawaii. J. Econ. Ent. 81: 228–235. Chatterjee, S., Hadi, A. S., and Price, B. (2000). Regression Analysis by Example, 3rd edn., Wiley, New York. D’Adamo, P., Corley, J., Sackmann, P., and Lozada, M. (2000). Local enhancement in the wasp Vespula germanica: Are visual cues all that matter? Insectes Soc. 47: 289–291. D’Adamo, P., and Lozada, M. (2003). The importance of location and visual cues during foraging in the German wasp (Vespula germanica F.) (Hymenoptera: Vespidae). N. Z. J. Zool. 30: 171–174. Duncan, C. D. (1939). A Contribution to the Biology of North American Vespine Wasps, Stanford University Publications, Biological Sciences 8: 1–272. Free, J. B. (1968). The conditions under which foraging honeybees expose their Nasanov gland. J. Apic. Res. 7: 139–145. Jander, R. (1998). Olfactory learning of fruit odors in the Eastern yellowjacket, Vespula maculifrons (Hymenoptera: Vespidae). J. Insect Behav. 11: 879–888. Jandt, J. M., Curry, C., Hemauer, S., and Jeanne, R. L. (in press). The accumulation of a chemical cue: Nest-entrance trail in the German yellowjacket, Verpula germanica. Naturwissenschaften. Landolt, P. J. (1998). Chemical attractants for trapping yellowjackets Vespula germanica and Vespula pensylvanica (Hymenoptera: Vespidae). Environ. Entomol. 27: 1229–1234.

Vespula germanica Foragers Do Not Scent-Mark Food Sites

31

Lehner, P. N. (1996). Handbook of Ethological Methods, 2nd edn., Cambridge University Press, Cambridge, UK. Lloyd, J. E. (1983). Bioluminescence and communication in insects. Ann. Rev. Entomol. 28: 131–160. Overmyer, S. L. (1995). Scent-mediated recruitment and food-site marking in the German yellowjacket, Vespula germanica (Hymenoptera: Vespidae), Master’s Thesis, University of Wisconsin-Madison. Raveret-Richter, M. (2000). Social wasp (Hymenoptera: Vespidae) foraging behavior. Ann. Rev. Entomol. 45: 121–150. Raveret-Richter, M. A., and Jeanne, R. L. (1985). Predatory behavior of Polybia sericea (Olivier), a tropical social wasp (Hymenoptera: Vespidae). Behav. Ecol. Sociobiol. 16: 165–170. Schmitt, U., and Bertsch, A. (1990). Do foraging bumblebees scent-mark food sources and does it matter? Oecologia 82: 137–144. Seeley, T. D. (1989). The honey bee colony as a superorganism. Am. Sci. 77: 546–553. Sharp, J. L., and James, J. (1979). Color preference of Vespula squamosa. Environ. Entomol. 8: 708–710. Slater, P. J. B. (1973). Describing sequences of behavior. In Bateson, P. P. G., and Klopfer, P. H. (eds.), Perspectives in Ethology, Vol. 1, Plenum, New York, pp. 131–153. Sokal, R. R., and Rohlf, F. J. (1981). Biometry, 2nd edn., W. H. Freeman, New York. Steinmetz, I., Schmolz, E., and Ruther, J. (2003). Cuticular lipids as trail pheromone in a social wasp. Proc. R. Soc. Lond. 270: 385–391. Takagi, M., Hirose, Y., and Yamasaki, M. (1980). Prey-location learning in Polistes jadwigae Dalla Torre (Hymenoptera, Vespidae): Field experiments on orientation. Kontyu 48: 53– 58. von Frisch, K. (1967). The Dance Language and Orientation of Bees, The Belknap Press of Harvard University Press, Cambridge, MA.