The Ectoparasites of the European Badger, Meles meles, and the Behavior of the Host-Specific Flea, Paraceras melis

Transcription

1 Journal of Insect Behavior, Vol. 12, No. 2, 1999 The Ectoparasites of the European Badger, Meles meles, and the Behavior of the Host-Specific Flea, Paraceras melis R. Cox, 1 P. D. Stewart, 1 and D. W. Macdonald 1,2 Accepted August 24, 1998; revised October 27, 1998 The badger-specific flea, Paraceras melis, jumps repeatedly when separated from its host Meles meles; thereafter fleas settled into sheltered positions. After separation from badgers, some 42% of fleas (n = 63) voided their gut contents; this was associated with a significant increase in mean jumping distance. The maximum longevity of fleas away from the host was 89 days, with 50% mortality at around 35 days. Badger lice, Trichodectes melis, survived for up to 3 days postcapture. We conclude that the badger's habit of frequently swapping dens with a mean period of return of 6 days is unlikely to bring about significant mortality of adult fleas but may effectively eradicate lice. Fleas abandoned in "bedding" in a simulated badger sett were mobile, being drawn toward light and moving upward. This response would draw the fleas to the den entrance, which may be a suitable site to intercept returning badgers. The fleas responded to stimuli which might signal the proximity of the host: they jumped toward sources of carbon dioxide and of carbon dioxide in air current directed at the flea. The strongest response was seen when a mixture of stimuli consisting of carbon dioxide, a dark circle of card, and movement were tested; the majority of fleas jumped toward the mixed stimulus. Finally, fleas separated from the host responded to exhaled air by running and jumping; this is in marked contrast to their response to those stimuli when on the host, when fleas run downward and very rarely jump. These contrasting observations find adaptive explanation in the two contexts. KEY WORDS: flea; Paraceras melis; behavior; ectoparasite; host-finding; badger; Meles meles. 1 Wildlife Conservation Research Unit, Department of Zoology, University of Oxford, South Parks Road, Oxford, OX1 3PS, United Kingdom. 2 To whom correspondence should be addressed /99/ $16.00/ Plenum Publishing Corporation

2 246 Cox, Stewart, and Macdonald INTRODUCTION The badger Meles meles is commonly infested with ectoparasites such as fleas, ticks, and lice (Hancox, 1980, 1988). Here we are concerned mainly with the badger and its host-specific flea, Paraceras melis, with incidental observations on the louse Trichodectes melis. Adult fleas are parasitic, drawing resources from the host (Price, 1980), and therefore have the potential to reduce the host's fitness (reviews by Moller, 1990; Lehmann, 1993). Unusually high ectoparasite loads can cause death from anemia and weight loss (Neal, 1977). Ectoparasites can also serve as vectors of a variety of endoparasites (Hancox, 1980; review by Hart, 1990). In the badger's case the blood parasites Trypanosoma pestanai [5% prevalence (Macdonald et al., 1999)] and Theileria meles [70% prevalence (Macdonald et al., 1999)] are candidates for ectoparasite transmission, and bovine tuberculosis, Mycobacterium bovis, is a candidate for mechanical transmission (from contact with the ulcers and dags of tuberculosis badgers). The cumulative effects of several types of macroparasites may be measurable, even if each alone has no detectable cost to fitness. While fleas and lice may become detached from the host by its grooming, the large number of fleas found in badger bedding suggests that they may also detach voluntarily (Neal and Roper, 1991). The seasonal decline in fleas found on captured badgers (unpublished data) may not simply reflect seasonal cycles [since the den represents a relatively stable year-round microclimate (see Neal and Cheeseman, 1996)]; instead, low skin temperatures on the badger while outside the den in winter [a result of fat rather than fur being the primary insulation, Stewart (1997)], do not favor these ectoparasites. Under such conditions fleas may voluntarily detach to remain within the den. Therefore relocation behavior may not be confined to freshly emerged fleas and those detached by host grooming. Our aim was to discover how the flea finds and remains on the badger. Host location methods (review by Rea and Irwin, 1994) may use both chemical and physical sensory input (Pike, 1990). MacInnis (1976) concluded that signals to which parasites respond during active host location may originate from the host (e.g., heat), within the host (by products of physiological processes) or the environment (e.g., air currents). Responses to signals may operate in a number of ways. Direct response will decrease the distance between the host and the parasite. General responses to the environment may bring the ectoparasite into areas with a greater potential for host contact (e.g., den entrances). Random activity of the parasite can increase the probability of contact with the potential host, especially shortly after detachment. Signals can also initiate development or physiological changes that prepare the parasite for the next step (Rea and Irwin, 1994). The badger and flea exhibit behaviors that impede and promote the main-

3 Badger Flea Behavior 247 tenance of mutual contact, respectively. Badgers frequently groom (Neal, 1977), which entails costs including time and energy and is often seen when the badger has left the sett in the evening. Stewart (1997) concluded that fleas attempt to escape the disturbance characteristic of grooming by running in the fur, typically in the direction of fur growth and toward the badger's posterior. The badger counters by rapidly switching grooming sites and often enlisting conspecific help to reach relatively inaccessible regions (shoulders and back). When on the host the fleas seldom respond to disturbance by jumping, even though this represents the most rapid means of escape, perhaps because it carries an increased risk of separation from the host. In this context we ask, What is a flea's immediate reaction to being separated from the host? How does it keep itself within the host's active space and to which host stimuli does the flea respond? GENERAL METHODS Subjects Badgers (Meles meles) were live-trapped during May and July 1997 at Wytham Woods, Oxfordshire, and during June 1997 at North Nibley, Gloucestershire. They were anesthetized by an intramuscular injection of mg/kg ketamine hydrochloride (Vetelar; Parke Davis Veterinary). Fleas were gently picked off anesthetized badgers by hand between 0700 and Individually identified fleas were housed separately in small clear plastic containers with a piece of moist white filter paper (diameter, approximately 3 cm), to maintain a high humidity, and a piece of white muslin (diameter, approximately 3 cm), as a grasping substrate. These were kept moist and changed regularly (usually every 4 days). They were kept at 9DC (the typical temperature of a badger den) (T. Roper, personal communication) under a light:dark cycle of 12:12 h. All fleas used in the experiments were identified as the badger flea Paraceras melis (Smit, 1957). Initial Observations During initial observations fleas removed from the badger jumped repeatedly (n = 50). If allowed to return to a badger, fleas generally ceased jumping and ran through the fur (Stewart, 1997). This study observed that fleas void their stomach contents rapidly after capture, a behavior envidenced by one or more small red stains on the paper substrate of their containers. After half an hour or so of separation fleas were observed to be far less active, often squeezing between the folds of the damp paper lining their containers. Vibration of containers containing quiescent fleas stimulated them to run and recommence jumping. Sustained vibration ceased to cause this jumping response, but it could be

4 248 Cox, Stewart, and Macdonald reinstigated by breathing into the container. Fleas housed communally showed a cascade response to a single active flea dropped into the container, and fleas generally jumped in response to any contact from another flea. When fleas and lice were housed together we found no indication of the lice attempting to attach to the fleas' legs. This response has been deduced from lice found attached to the legs of fleas in taxonomic collections and interpreted to be an example of phoresy (Britt, 1982). The rare instances of lice grasping to the legs of fleas were observed only when there were no other fibrous substrates to which the lice could attach. The fleas responded to a louse's attachment by vigorously kicking. Female fleas laid eggs in the containers, and eggs hatched after approximately 2 weeks. Eggs were quite resistant to water and exposure to light, but fleas, lice, and flea larvae were very prone to entrapment and drowning in the water meniscus. Similarly fleas or flea larvae kept in dry containers died within a few days (lice died rapidly under any conditions away from the host; see experiment 2 below). Flea larvae were successfully fed on dried badger blood and survived for up to 3 weeks, growing slowly. We were not able, however, to rear them through pupation under either cool (9DC) or warm (21DC) conditions, at high or low humidity. These initial observations suggested the following experiments (± values are standard errors, unless indicated otherwise). EXPERIMENT 1: DOES VOIDING OF GUT CONTENTS INCREASE A FLEA'S JUMPING RANGE? Method A total of 63 fleas was carefully picked off badgers and their time of capture recorded. Each flea was placed in the center of a jumping arena of radius 40 cm which was marked with concentric circles, 1 cm apart. The flea was allowed to jump from the center of the arena four times and the mean jumping distance was calculated. Each flea was weighed using a four-figure Oertling microbalance. The fleas were jumped at intervals as close as possible to 0, 2, 5, 10, 20, 30, 60, and 90 min after capture. During these measurements we recorded whenever a flea voided its gut contents and immediately retested its jumping ability. To estimate the mean mass of the fluid voided by a flea, five fleas were kept in separate clear plastic containers without filter paper until voiding occurred. The diameter of the hemispherical fluid drop produced on the side of the container was measured and a mean size of 1.5 mm calculated. Fluid droplets of mean size 1.5 mm were recreated on a glass slide using badger blood (from an anesthetized badger) and weighed. The mean mass of one droplet was calculated to be 0.82 ± 0.02 mg.

5 Badger Flea Behavior 249 Data Analysis Individual variability in jumping ability was controlled for by comparing the mean jumping distance of an individual before and after voiding gut contents using a paired t test, the data being normally distributed. Data from the fleas which were not seen to void their gut contents were analyzed in the same way to elucidate their change in jumping ability over time. A t test was used to compare (a) the body mass and (b) the mean jumping distance, before voiding, of the fleas that were seen to void their gut contents compared to those that were never observed to void their gut contents. Results A total of 27 of 63 fleas was seen to void (at least part of) their gut contents. Twenty-six of 27 of these showed an increase in mean jumping distance after voiding (Fig. 1). Fleas were able to jump significantly farther when their guts were empty compared to full (paired t test, t = 8.04, n = 27, P < 0.001). The mean mass of the fleas that were seen to void their gut contents was mg, thus the mean void represents 3.3 ± 0.008% of a freshly caught flea's mass. The mean time following capture to voiding of the gut contents was 21 min 58 s ± 3 min 56 s. After voiding the mean increase in jumping distance was ± 0.17 cm (range, 0.08 to 3.36 cm). Thirty-six fleas did not void their gut contents. A mean jump distance was calculated from measurements taken before 22 min postcapture (the mean time of voiding) and after 22 min postcapture. Twenty-two of 36 fleas showed an increase in mean jumping distance after 22 min, but this difference was not significant (paired t test, t = 1.29, n = 36, P = 0.21). There was no detectable Fig. 1. Distance of mean jump by fleas before and after voiding of gut contents. Means are shown with standard error bars (except where only one set of jumps was observed).

6 250 Cox, Stewart, and Macdonald significant difference between the mass (t test, t = 0.60, df = 58, P = 0.55) or the mean jumping distance (t test, t = -1.93, df = 47, P = 0.059) of the fleas that were seen to void their gut contents compared to those that did not, which may be a result of the inadequate sensitivity in the balance employed. EXPERIMENT 2: HOW LONG DO FLEAS AND LICE SURVIVE SEPARATION FROM THE BADGERS? Method The fleas from experiment 1 plus a further 18 fleas (total of 81 fleas) were housed separately as described previously. On the day of capture and on each day thereafter, each flea was weighed and jumped four times (to obtain a mean jumping distance), between 1000 and 1500, until the rate of weight loss had slowed sufficiently to warrant weighing and jumping every alternate day. A further 157 fleas were collected in the same way as before. Sixty-seven were individually identified and placed directly into separate containers and 90 were housed in groups of 10. They were all stored at 9DC. These fleas were not weighed or jumped but the rate of mortality was recorded to act as a control for the jumping experiments. One hundred lice were carefully picked off anesthetized badgers by hand. Lice were housed separately in small clear plastic containers as described for the fleas. The rate of mortality was recorded by noting the number of individuals surviving each day postcapture. Data Analysis Pearson's correlation coefficient was used to assess the relationship between mass and jumping distance and number of days separated from the host. For the regular jumping fleas, only data for fleas that had been jumped and weighed for at least 3 days after capture were used. Results Fleas that were kept as controls and were not weighed or jumped at all lived for up to 87 days (kept individually) and up to 89 days (kept in groups of 10 individuals). Figure 2 shows the proportion of dead fleas with increasing time. Mortality was greatest between approximately 20 and 50 days after separation. Figure 3 shows the change in body mass and mean jumping distance after separation from the host of the fleas that were weighed and jumped regularly. The maximum length of survival after capture of these fleas was 63 days. The mean jumping distance of a flea increased during approximately the first 5 days after capture. The maximum distance jumped was cm (mean of four jumps measured in 1 day) by a flea weighing 10 mg, separated from the host for 14

7 Badger Flea Behavior 251 Fig. 2. Mortality of fleas separated from the host which were not weighed or jumped. Percentage of fleas dead each day postcapture from samples housed either individually (total of 67 fleas) or housed in groups of 10 (total of 90 fleas). Method of housing did not appear to affect longevity. days. The jumping distance of the fleas began to decrease only after about days of separation from the host. Of the fleas that were jumped regularly until they died (total 47), 30 showed a decrease in jumping ability prior to death. The mean time that the jumping distance began to decrease was 5.3 ± 0.7 days (range, 1-12 days) before death. Lice survived for a maximum of only 3 days of separation from the host; 39% died after day 1 and 73% had died after day 2. The mean mass of fleas decreased with time of separation from the host. There was no significant relationship between the mean distance jumped and the time of separation from the host. The results tend toward a negative relationship between the mean distance jumped and the mean mass of the flea, although this was not statistically significant at the 5% level and there is much scatter (Table I). Figure 4 shows a trend that the lighter a flea is, the farther it can jump. Although the data were widely scattered, the triangles highlight regions of the graph for which there were no observed jumps and illustrates that only heavy fleas cannot jump very far and only very light fleas jump very short distances (normally prior to death).

8 252 Cox, Stewart, and Macdonald Fig. 3. Change in mean body mass and mean jumping distance after separation of the fleas from the host. Standard error bars are shown. Table I. Correlation of Mean Distance Jumped and Mean Flea Body Mass Over Time Postcapture a Distance of mean jump Mean flea body mass Time (days postcapture) *** Distance of mean jump (*) a N = 63 (data for 63 days of capture). ***P< (*)P = EXPERIMENT 3: TO WHAT STIMULI DO FLEAS RESPOND WHEN ABANDONED UNDER SIMULATED DEN CONDITIONS? Method Ten fleas were placed in some bedding made up of lightly dampened grass, straw, and debris. The bedding and the fleas were placed in a cardboard tube (length, 2 m; diameter, 10 cm) through a small door (12 cm long, 9 cm wide) cut in the center of the side. The tube was kept damp to maintain a high humidity and stored at 9DC. The ends of the tube were covered with clear plastic film to prevent the fleas escaping and to stop any air movement influencing their behavior. The tube was subject to one of the following conditions.

9 Badger Flea Behavior 253 Fig. 4. Relationship between flea body mass and mean jumping distance. Each flea was weighed and jumped regularly from the day of capture until death. Each plot mark represents the mean distance jumped by a flea at that particular mass. The triangles highlight areas where there are no data points. (1) Light/dark One end of the tube was draped with a piece of thick material to make it dark. The tube was laid horizontally. (2) Up/down Both ends of the tube were draped with material to provide uniform darkness. The tube was tilted at an angle of approximately 30 D to the horizontal. The number of fleas visible within 30 cm of each end of the tube was recorded at 1200 daily, for 3 consecutive days. Each experiment was repeated six times with a new set of fleas. Between each experiment the stimulus was changed to the opposite end of the tube to control for external factors. The fleas used had been separated from a badger for between 1 and 7 days. Data Analysis The Wilcoxon paired signed-ranks test was used to compare the number of fleas at each end of the tube on each day of the test and also to analyze the movement of the fleas over the 3 days of observation. Results More fleas were observed at the light end of the tunnel than at the dark end on each day of the experiment. There was a significant change in the number of fleas at each end of the tube between day 1 and day 2 and also between day 1 and day 3 (Table II).

10 254 Cox, Stewart, and Macdonald Table II. Analyses Comparing (a) the Number of Fleas at the Light End of the Tube Compared to the Dark End and (b) the Number of Fleas at Each End of the Tube Over Time a Time Day 1 Day 2 Day 3 Day 1-Day 2 Day 2-Day 3 Day 1-Day 3 Wilcoxon statistic a 21* 21* 21* b 0* 0 0* a For each experiment n = 10. *P < Table III. Analyses of (a) the Number of Fleas at the Raised End of the Tube Compared to the Lower End and (b) the Number of Fleas at Each End of the Tube Over Time a Time Day 1 Day 2 Day 3 Day 1-Day 2 Day 2-Day 3 Day 1-Day 3 Wilcoxon statistic a * b 0 1 0* a For each experiment n = 10. *P < There were more fleas observed at the higher end of the tunnel compared to the lower end only on day 3 of the experiment. A significant difference was also observed in the movement of the fleas between day 1 and day 3 (Table III). EXPERIMENT 4: TO WHAT HOST CUES DO FLEAS RESPOND? Method The jumping arena as described in experiment 2 was marked to divide it into eight numbered, equal-angle segments. The arena was placed on a table surrounded by white sheeting hanging from the ceiling to the floor (an observation port was provided for the researcher). The lighting within the arena was uni-

11 Badger Flea Behavior 255 formly diffused via a skylight covered with white paper. A flea was placed in the center of the arena and a clear plastic container placed upside down over it to prevent it jumping any distance. A piece of wire 1.5m long was attached to the top of the container so that it could be removed by the observer from outside of the enclosed arena. A stimulus was presented in one of the segments, 15 cm from the center of the arena (unless stated otherwise) so that it was within jumping distance of the flea (see Table IV for a description of stimuli). The plastic container in which the flea was enclosed was removed and the segment to which the flea jumped was recorded. The stimulus and control were alternated with each consecutive flea. The procedure was carried out on 100 fleas. In each experiment the observer was positioned outside the arena at the corner between segment 5 and segment 6. The segment into which each different stimulus was placed was varied between experiments to account for any external influencing factors. The temperature and humidity were monitored using a Testo 625 probe and there was no observed difference in either temperature or humidity between each experimental and control procedure. The experiments were carried out on fleas which had been separated from a badger for between 0 and 3 days. Data Analysis Analysis of circular statistics followed Batschelet (1981). The Raleigh z test was used to analyze whether the population from which the sample is taken differs significantly from randomness (i.e., whether there is statistical evidence of one sidedness). A chi-square test was used to test the significant difference between the experimental and the control values for each experimental test. Where expected values were less than 5, data from adjacent segments were merged. We used an improved Bonferroni correction to adjust the significance of the P values in each experiment for the total number of experiments conducted (Haccou and Meelis, 1995). Results The Raleigh z test (Table V) indicates that the fleas were jumping nonrandomly when the stimuli of light, carbon dioxide, and carbon dioxide with breeze and a mixture of stimuli were tested. The Bonferroni procedure excluded the light stimuli from significance. There was a significant difference between the experimental and the control results when carbon dioxide, carbon dioxide with breeze, and a mixture of stimuli was presented to the fleas, indicating that the fleas were responding positively in jumping towards the stimuli. Figure 5 illustrates the results of the experiments where a significant reaction by the fleas was recorded.

12 256 Cox, Stewart, and Macdonald Table IV. Description of Stimuli and Procedure in Each Test for 4 Stimulus tested Light A lamp was placed in segment 4. The temperature was monitored while it was switched on or off; it was found that the temperature remained constant during the short time that the lamp was on or off. Silhouette A white piece of card was positioned upright on segment 4. A lamp was positioned behind the card to cast a silhouette across the segment. Vibration A piece of wire 1.5 m in length was tapped onto the arena board in segment 6 by the observer from outside of the arena. Breeze A small fan was placed 50 cm from the center of the arena in segment 7. The wind speed at the center was measured using an anemometer as a mean of 1.18 m/s. Visual A roughly circular piece of black card of mean diameter 19 cm was placed in segment 3. Carbon dioxide A piece of dry ice (which produces pure carbon dioxide as it melts) of mean mass 1.5 g was placed in segment 4. Carbon dioxide and breeze Method repeated as with carbon dioxide but in segment 7. A fan producing a breeze of speed 1.18 m/s was placed 50 cm from the center of the arena in segment 7. The carbon dioxide produced by the dry ice was directed toward the center of the arena. Movement A circular disk of card of diameter 19 cm with black and white stripes of width 2 cm was taped to the fan. This was placed in segment 4. The fan rotated slowly, although no measurable breeze was produced. Heat A water bottle at 30DC containing 1.5 pt of hot water was placed in segment 1. The body temperature of a badger had been measured previously and was found to be a mean of 32.7D C. Badger hair smell A pile of badger hair of mass 3.0 g was placed in segment 4. Mixture of stimuli al methods for carbon dioxide, movement, and visual stimuli were combined, in segment 6. Lamp was switched off. Lamp was switched off. The piece of wire was moved in the same way without tapping the board. Fan switched off. A piece of white card replaced the black card. A pebble of approximately the same size and color replaced the dry ice. A pebble replaced the dry ice as above; the fan was switched off. The fan was switched off. Cool water replaced the hot water so the bottle temperature was the room temperature of 21.3D C. A pile of washed badger hair of mass 3.0 g replaced the unwashed badger hair methods for carbon dioxide, movement, and visual stimuli were combined.

13 Badger Flea Behavior 257 Table V. Response of Fleas to Various Stimuli: (a) Results of Raleigh's z Test to See If Fleas Were Jumping Randomly; (b) Results of Chi-Square Test Comparing al and Results a Stimulus Light Silhouette Vibration Breeze Visual Carbon dioxide Carbon dioxide and breeze Movement Heat Badger hair Mixture of stimuli (a) Raleigh's z 8.916*** *** *** *** (b) x (6) (6) (7) (5) (5) t (5) t (5) (5) (5) (6) t (4) a For each experiment and control n = 50. Degrees of freedom are shown in parentheses. ***P < t Chi-square values retaining significance after application of a sequential Bonferroni adjustment; P < EXPERIMENT 5: IS THERE A THRESHOLD CONCENTRATION OF CARBON DIOXIDE TO WHICH FLEAS RESPOND? Method Ten opaque plastic 40-ml containers were taped securely to an upturned cardboard box. Each container had a small hole in the side where a rubber tube of 4-mm diameter, and 12-cm length fit securely; this was also taped to the

14 258 Cox, Stewart, and Macdonald Fig. 5. Response of fleas to certain stimuli. Graphs a to c represent three experiments where the results are shown to be significant. Each graph represents an arena divided into eight equal-angle segments. A stimulus was presented in one segment. The direction that a flea jumped from the center was recorded. Total number of fleas = 50. During each experiment a control was carried out at the same time (shown to the right of the experimental graph). (a) Stimulus: carbon dioxide in segment 4 (mean angle of direction of jump D). (36.3 D). (b) Stimuli: Carbon dioxide with breeze from segment 7 (301.5 D). (354.6 D). (c) Stimuli: Mixture of stimuli in segment 6 (235.7 ). (0 D).

15 Badger Flea Behavior 259 cardboard. A 60-ml syringe could be attached to the other end of the tube so that the contents could be expelled into the container. A piece of moist filter paper lined the base of each container to maintain a high humidity. A flea was placed in each container and allowed to acclimate to the surroundings. The lid of the container was balanced on the top to prevent the flea from escaping, while not being so secure that the air from the syringe increased the pressure within the container. The end of the tube in the container was packed with sponge to diffuse the air current when the syringe was depressed. A piece of dry ice was put into a syringe and the nozzle blocked with a piece of plasticine. As the dry ice melted, it produced pure carbon dioxide and the plunger was forced back up the tube. A mixture of approximately 2% carbon dioxide in air could be achieved in the syringe by making up 1.2 ml of the pure carbon dioxide to 60 ml. This method allowed the air mixture to reach room temperature after the dry ice had melted. The syringe was attached to the tube leading from the container and was taped to the box so that when the plunger was depressed the movement did not disturb the behavior of the flea. The syringe was depressed slowly over 20 s and the activity of the flea during a total of 90 s was recorded. The flea was classed as active if it was jumping or running. The observer recorded from a level lower than the container so as not to influence the flea activity. The method was repeated in the next container with a syringe containing atmospheric air. Thus the experiment was carried out alternating with a control test of normal air. The experiment was repeated on 100 fleas. The experiment was then repeated on different fleas with concentrations of 4% carbon dioxide approximately the concentration of exhaled air (2.4 ml of pure carbon dioxide in a 60-ml syringe) and 15% carbon dioxide (9 ml of pure carbon dioxide in a 60-ml syringe). Data Analysis A chi-square test or Fisher's exact test (where expected counts were less than 5) was used to test for a significant difference between the activity of the fleas in air compared to a concentration of 2, 4, or 15% carbon dioxide. Chisquare tests were also used to test for a significant difference between the activity of the fleas in 2% compared to 4% carbon dioxide, as these experiments were carried out on the same day. Results Figure 6 shows the difference in activity of fleas at each concentration of carbon dioxide. Fleas were significantly more active in concentrations of 4 and 15% carbon dioxide compared to air (Table VI). Furthermore, fleas were significantly more active in concentrations of 4% carbon dioxide compared to 2% carbon dioxide (x 2 = , df = 1, P < 0.001).

16 260 Cox, Stewart, and Macdonald Fig. 6. Activity of fleas in response to different concentrations of carbon dioxide. Each experiment compares carbon dioxide concentration with atmospheric air. Table VI. Statistical Analysis of the Activity of Fleas in Response to Different Levels of Carbon Dioxidea Concentration of carbon dioxide (%) x2 value(or Fisher's exact P) -* *** *** afor each experiment n = 40. *P < ***P < DISCUSSION Immediate Response to Separation from the Host When fleas void their gut contents the possible jumping distance increases by ± 1.88% of the mean jump before voiding. The loss of fluid may help the fleas' jumping ability in other ways not investigated here (e.g., height of jump). The prediction that a reduction in the mass of fluid stored in the body will increase the distance jumped is based on equations used to model saltation in frogs (Marsh, 1994). Some frog species have been observed to empty their bladders in response to predators (Buchanan and Taylor, 1996). On average frogs with empty bladders jumped 18.5% farther than frogs with full bladders. Even

17 Badger Flea Behavior 261 though the frogs increase their risk of desiccation, the immediate need to escape from predators is apparently paramount. Similar arguments may apply to the fleas, which must balance relocating the host quickly, against the longer-term risk of desiccation if the gut fluids are lost and the host is not relocated. Fewer than half (27 of 63) of the fleas were seen to void their gut contents, but the other fleas may have done so unobserved while we attempted to catch them on the badger. Another immediate response by the fleas to separation from the host was a high rate of jumping, often after the flea had reached the local peak of the substrate by running. This may be an attempt to relocate the host before it moves further away. After more prolonged separation fleas travel into protected crevices in the substrate, perhaps to seek protection against desiccation. Vibration stimulates them to emerge from these crevices and to recommence running and jumping. Longevity After Separation from the Host Fleas can survive away from the host and without feeding for up to 89 days when kept under conditions of temperature and humidity similar to those encountered in badger dens. The mass of the fleas decreased with the time after separation. The jumping ability of the fleas did not decrease until just prior to death, indicating that they are mobile and able to search for a host long after separation. One cannot assume, however, that, even with food, all the fleas could recover from starvation. Badgers have been observed to switch their sleeping location within the sett regularly, a behavior that has been interpreted as a strategy to reduce ectoparasite burdens (Butler and Roper, 1996). They observed that treating badgers with antiparasite spray reduced the frequency at which they switched from one nest site to another. Badgers typically return to the nest after 6-7 days. Our observations have shown that this may be a sufficient interval to cause mortality of the lice within the den, but not the adult fleas. It is important to note that the laboratory conditions (high humidity and low temperature) were controlled to approximate those encountered in a den (Moore, 1997; see Neal and Cheeseman, 1996); abandoned fleas may die faster outside the den. Fleas are vulnerable in the early stages of development and the duration of the pupal stage is temperature dependent (Rothschild, 1965). A possible advantage of the use of multiple sleeping sites could therefore be that when a nest chamber is vacated, the development of immature fleas is showed or terminated (Butler and Roper, 1996). We observed that captive adult fleas, flea larvae, and lice were susceptible to both desiccation and entrapment or drowning in a water-drop meniscus. Humphries (1968) showed that fleas are highly susceptible to water loss and, if

18 262 Cox, Stewart, and Macdonald kept under drier conditions, showed a more rapid decrease in jumping ability. Fleas, and to a lesser extent lice, are found in large numbers in excavated bedding (Neal, 1986; Hancox, 1988). An effective strategy by the badger against the relatively immobile flea larvae could be the renewal and airing of bedding [badgers drag bedding out of the den and leave it there for a few days before taking it back (Neal and Roper, 1991)]. This will subject it to alternating desiccation and damp. Behavior of Fleas Abandoned in the Nest Fleas left in bedding material within a tube simulating the conditions of a badger's sett abandon it within the mean 6-day period of return of a badger (perhaps to optimize the rate of feeding and hence reproduction). Given that the badger's return interval is within the typical flea mortality period, an alternative strategy would be for the flea to adopt a safe, but potentially less fecund, strategy of "sit and wait." The flea's abandonment response may be another reason badgers relocate sleeping site; by doing so they effectively "redistribute" their fleas among the other occupants of the sett. Fleas free to wander in a length of tubing concentrated at its light end and moved upward. This strategy, if pursued in a real den network, would result in the fleas accumulating around the entrance of the den, where they might intercept badgers. Studies on bird fleas have similarly shown that if infected nest sites are not used, fleas will start searching for a new nest by waiting at the nest entrance or by moving away from the nest in the expectation of jumping onto a passing bird (Bates, 1962; Humphries, 1968; Du Feu, 1992). The fleas in our experiment showed a greater response to light than to the angle of the tube, perhaps indicating that this is a more influential stimulus. Response to Stimuli Fleas that have been separated from the host strongly responded to the stimuli of carbon dioxide, carbon dioxide with a breeze, and the mixture of carbon dioxide, movement, and dark object by jumping toward the source. The results indicate that when attempting to relocate a badger directly, fleas may react most strongly to a combination of host stimuli. During the experiments involving carbon dioxide, the fleas often waved their heads backward and forward immediately before jumping. This implicates uni- or bisensory organs (such as antennae) in a taxic orientating response for the jump. The eyes of fleas are simple in structure and probably enable the insects to distinguish only between light and darkness (Rothschild, 1965). The badger fleas showed no significant response to shadows, dark objects, or movement. These stimuli may not be so important for flea species which live in dark burrows and

19 Badger Flea Behavior 263 on hosts that are active at night. The response to light in experiment 3, however, indicates that the fleas can detect the direction of a light source. s by Bates (1962) with dark cardboard disks showed no evidence that sand martin fleas are attracted to shadows or to dark patches on sand cliffs and similarly concluded that it is probable that these fleas do not detect burrows by vision. Du Feu (1992) noted that bird fleas have a greater willingness to jump onto darker targets. Humphries (1968) recorded head waving behavior (similar to the taxic orientating response we observed during experiments involving carbon dioxide) in the hen flea C. gallinae. When a piece of card was used to vary the amount of light falling on a flea, it invariably responded with a lateral swinging to and fro ("peering") of the head and thorax, which was sometimes interrupted by a shuffling reorientation of the body, prior to jumping. These movements appeared to align the flea more accurately toward the object casting the shadow. Fleas kept in sealed containers were observed to respond to the initial movement of the container by running upward and jumping. They appeared to habituate to this sort of physical disturbance quite rapidly, however, so the tests of directionality in response to vibration alone may be deceptive in fleas which have recently been handled. Fleas are believed to detect air currents with the aid of sensor bristles on the abdomen (Bates, 1962). He reports that sand martin fleas show intense activity for s at the entrance of a deserted burrow (a cliff face) when stimulated by vibration. There was no significant response to the stimuli of heat or badger hair (smell). It may be that these factors play a more important role when the parasite has found a potential host. We might expect the importance of different senses to vary between flea species depending on their host behavior and lifestyle. Response to Carbon Dioxide Fleas respond to moderately elevated carbon dioxide by running and jumping toward the source. The fact that they showed only a very low level of activity at 2% carbon dioxide and a significantly higher level of activity at 4% carbon dioxide may indicate that they can detect only levels above a certain threshold or that they choose to wait until the concentration reaches a certain level before becoming active and attempting to find a badger. It could be argued that fleas react because carbon dioxide is potentially lethal at high concentrations. However a level of 4% in air is approximately that of exhaled air and the results from experiment 5 also show that fleas tend to jump not away from the carbon dioxide but toward it. Fleas may habituate to some stimuli, for example, movement, and so an increased concentration of carbon dioxide may provide a more reliable indicator of the host's proximity under such conditions. In a series of experiments, Stewart (1997) placed fleas in an open-bottomed container on an anesthetized badger and subjected them to approximately 5% carbon dioxide and

20 264 Cox, Stewart, and Macdonald 0 of 20 jumped in response. This indicates how the response to such a stimulus may change depending on other external factors. Initial indications suggested that it is the texture of the substrate (fur) which inhibited jumping. We have considered how the badger flea locates a host by responding to a number of stimuli. There is the possibility of additive effects of stimuli and that the level of one stimulus may dictate the response to others. It must be remembered that under natural conditions spurious stimuli may emanate from a variety of sources. Whether the elicited behaviors improve location rates has not been tested here. Once on the potential host, other stimuli such as odor, taste, warmth, and substrate texture may ultimately determine whether the parasite stays or leaves (Pike, 1990). ACKNOWLEDGMENTS We thank Tom Simmons for practical assistance, Paul Johnson and Steven Freeman for help with statistics and critical review of the manuscript, and Clive Hambler for help with identification of Paraceras melis. This work was carried out in partial fulfillment of the M.Sc. in Applied Animal Behaviour and Animal Welfare at Edinburgh University and was funded by grants from the Peoples Trust for Endangered Species and the BBSRC. REFERENCES Bates, J. K. (1962). Field studies on the behaviour of bird fleas. 1. Behaviour of the adults of three species of bird fleas in the field. Parasilology 52: Batschelet. E. (1981). Circular Statistics in Biology, Academic Press, London. Britt, D. P. (1982). Possible phoretic association between mallophagan and flea ectoparasites of the badger Meles meles L. Ann. Trop. Med. Parasitol. 76(4): Buchanan, B. W., and Taylor. R. C. (1996). Lightening the load: Micturation enhances jumping ability performance of squirrel tree frogs. J. Herpetol. 30: Butler, J. M., and Roper, T. J. (1996). Ectoparasites and sett use in European badgers. Anim. Behav. 52: Du Feu, C. R. (1992). How tits avoid flea infestation at nest sites. Ring. Migrat. 13: Haccou, P., and Meelis, E. (1995). Statistical Analysis of Behavioural Data, Oxford University Press, Oxford. Hancox, M. (1980). Parasite and infectious diseases of the Eurasian badger (Meles meles L.): A review. Mammal Rev. 10(4): Hancox, M. (1988). The nidiculous fauna of badger setts. Entomol. Month. Mag. 124: Hart, B. L. (1990). Behavioural adaptations to pathogens and parasites: Five strategies. Neurosci. Biobehav. Rev. 14: Humphries, D. A. (1968). The host-finding behaviour of the hen flea, Ceratophyllus gallinae (Schrank) (Siphonaptera). Parasitology 58: Lehmann, T. (1993). Ectoparasites: Direct impact on host fitness. Parasitol. Today 9: 1. Macdonald, D. W., Anwar, M., Newman, C., Woodroffe, R., and Johnson, P. (1999). Inter-annual differences in the age-related prevalences of Babesia and Trypanosoma parasites of European badgers (Meles meles). J. Zool (Lond.) 247: MacInnis, A. J. (1976). How parasites find hosts. Some thoughts on the inception of host-parasite

21 Badger Flea Behavior 265 integration. In Kennedy, C. R. (ed.), Ecological Aspects of Parasitology, North-Holland, Amsterdam, pp Marsh, R. L. (1994). Jumping ability in anuran amphibians. In Jones, J. H. (ed.), Advances in Veterinary Science and Comparative Medicine, Vol. 38B. Comparative Vertebrate Exercise Physiology. Phylectic Adaptations, Academic Press, San Diego, CA, pp Moller, G. P. (1990). Parasites and sexual selection: Current status of the Hamilton and Zuk hypothesis. J. Evol. Biol. 3: Moore, J. A. (1997). Internal Environment of Badger (Meles meles) Setts, D.Phil thesis, University of Sussex, Sussex. Neal, E. G. (1977). Badgers, Blandford Press, London. Neal, E. G. (1986). The Natural History of Badgers, Croom Helm, London. Neal, E. G., and Cheeseman, C. L. (1996). Badgers, T & AD Poyser Natural History, London. Neal, E. G., and Roper, T. J. (1991). The environmental impact of badgers (Meles meles L.) and their setts. Symp. Zool. Soc. London 63: Pike, A. W. (1990). Interpreting parasite post location behaviour. Parasitol. Today 6(11): Price, P. W. (1980). Evolutionary Biology of Parasites, Princeton University Press, Princeton, NJ. Rea, J. G., and Irwin, S. W. B. (1994). The ecology of host-finding behaviour and parasite transmission: Past and future perspectives. Parasitology 109: S31-S39. Rothschild, M. (1965). Fleas. Sci. Am. 213(6): Smit, F. G. A. M. (1957). Handbooks for the Identification of British Insects. Vol. I, Part 16. Siphonaptera, Royal Entomological Society of London. Stewart, P. D. (1997). The Social Behaviour of the European Badger Meles meles, D.Phil thesis, University of Oxford, Oxford.