Behavioral Ecology Advance Access originally published online on August 29, 2006
Behavioral Ecology 2006 17(6):992-997; doi:10.1093/beheco/arl040
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Fitness effects of parasite-mediated spatial heterogeneity within a swarm
Theoretical Ecology Group, Division of Environmental and Evolutionary Biology, Graham Kerr Building, University of Glasgow, Glasgow G12 8QQ, UK
Address correspondence to T.W. Pike. E-mail: t.pike{at}bio.gla.ac.uk.
Received 15 February 2006; revised 6 July 2006; accepted 14 July 2006.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
The fitness consequences associated with the position an individual adopts within a dynamic group are not well understood. I investigated mate acquisition by male chironomid midges using a simple swarming model and empirically collected data on midge aerobatic ability. Previous work has suggested that the aerobatic ability of a male is an important predictor of his reproductive success, although there is contrary (and counterintuitive) evidence that infection with ectoparasitic mites increases a male's chance of mating, despite having negative effects on flight speed. The model used here suggests that a male's location within the swarm, brought about passively through interindividual differences in flight speed, may explain these contradictory results. Specifically, slower flying males (including those burdened with mites) adopted positions nearer the center of the swarm, whereas faster males tended to occupy the periphery. This in turn affected their access to females because any mechanism that brought females nearer the swarm's center before capture (including high female flight speed and selective pairing by either males or females) significantly increased the relative reproductive success of both larger and parasitized males, with the benefits of parasitism peaking at around 4 mites per host. There may be selective pressure on hosts and parasites to maintain this optimal mite density because both are likely to benefit from the relationship: hosts enjoy an increased reproductive success, whereas only through host copulation can mites transfer to a female midge and return to water (their next life-history stage) during host oviposition.
Key words: aggregation behavior, Chironomus plumosus, spatial sorting, swarming.
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
Although the spatial structuring of individuals within a group (such as a flock of birds, shoal of fish, or swarm of insects) can have important ecological and evolutionary consequences (Okubo 1980
; Krause and Ruxton 2002), few previous studies have considered that the costs and benefits of living in a group are dependent on the spatial positions individuals adopt (reviewed in Krause 1994; see also Parrish 1993; Barber and Huntingford 1996; Bumann et al. 1997; Ward et al. 2002). Because the members of any group are not intrinsically identical, spatial differences are likely to result from individual variation in, for example, body size or energetic state (Pitcher et al. 1985; Parrish 1989; DeBlois and Rose 1996; Krause et al. 1996). For instance, Krause (1993) has demonstrated that starved roach (Rutilus rutilus) tend to adopt positions near the front of the shoal more often than those that are satiated. Thornhill's (1980) work on lovebug (Plecia nearctica) mating swarms suggests that there is distinct vertical stratification with respect to body size, giving larger individuals toward the bottom of the swarm first access to females, and there is evidence that parasitized fish occupy peripheral positions within a shoal of unparasitized conspecifics (e.g., Krause and Godin 1994). In addition, a recent theoretical model (Couzin et al. 2002) suggests that an individual's location within a group may be affected by factors such as its speed, with slower individuals adopting positions toward the group's center. However, the fitness consequences associated with position in a dynamic group have rarely been considered and indeed very little is known about the movement and spatial structuring of individuals within 3-dimensional aggregations due to their complexity and the difficulty of obtaining suitable data.
Adult male chironomid midges (Diptera: Chironomidae) form large, conspicuous mating swarms each evening in late spring (reviewed by McLachlan and Neems 1995), usually over distinctive landmarks that act as the catalyst for swarm formation (Downes 1969). These aggregations attract patrolling females that enter the swarm and emerge again after a short time coupled with a male that has presumably outcompeted rival males to pair with her (McLachlan and Neems 1995). Males successful at pairing are smaller on average than unmated males within the swarm in several chironomid species (for a review, see McLachlan and Neems 1995), which has led to the suggestion that the increased aerobatic ability demonstrated for small males (Crompton et al. 2003) gives them an advantage in competition over access to females. However, counterintuitively, males infested with ectoparasitic mites (predominantly the hydracharinid mite Unionicola ypsilophora) also appear to enjoy increased mating success (McLachlan 1999), despite the fact that relative to the size of their host these mites are large (see Figure 1 in McLachlan 1999) and so would be expected to impair flight performance in the male and thus reduce his access to females. There should, however, be strong selection pressure on these mites to increase the mating success of their host males. Because they appear unable to distinguish between male and female hosts at the time of attachment (McLachlan et al. 1999; Edwards and Smith 2003), only by transferring to a female midge during copulation will they have the opportunity to return to an aquatic environment (during host oviposition) and hence complete their lifecycle (Di Sabatino et al. 2000). There is not yet any satisfactory mechanism that can explain the increased mating success of parasitized males, although it is possible that some facets of aerobatic ability are in fact enhanced in parasitized individuals (AJ McLachlan, TW Pike, and JC Thomason, unpublished data). However, if the spatial position males adopt in the swarm is constrained by some correlate of parasitism, such as reduced flight speed (Couzin et al. 2002; McLachlan et al. 2003), then this may affect their access to females.
|
In this paper, I use the example of swarming chironomid midges to investigate whether spatial heterogeneity can result from interindividual differences in flight speed and how this might impact on male and female mating success. Specifically, I employed a simple swarming model, which used parameter values based on empirical data collected from parasitized and unparasitized midges, to investigate the hypothesis that the location of individuals within a group (in this case males in a mating swarm) can affect their access to females. The findings provide useful insights into the effects of spatial heterogeneity on group living.
|
METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
Measuring the flight speed of live midges
Male and female Chironomus plumosus midges were collected from Washington Wildfowl Park by inverting vials over individuals resting against windows during the early evening. These midges were kept overnight in individual glass vials supplied with water and raisin juice dried on filter paper as food (details in Neems et al. 1990
). Twenty-two males, which varied in their degree of mite infestation (from 0 to 6 mites) but had no other visible ectoparasites, and 10 unparasitized females were flown to assess the maximum flight speed they could achieve (for full details of the video technology, calibration methods, and computation of flight speed, see Crompton et al. 2003). Briefly, individual midges were flown in a periscopic chamber (50 x 50 and 180 mm high) and filmed with a digital high-speed video camera (Roper Fast-cam 10 K fitted with an Olympus 8–80 mm zoom lens) running at 250 frames per second. This frame rate permitted me to capture the first 50 frames (0.2 s) of the takeoff flight. A mirror fixed at 45° to the top rear edge of the chamber enabled simultaneous plan (to give x and y coordinates) and elevation (to give x and z coordinates) views to be recorded on each frame, hence permitting 3-dimensional analysis using a single camera (Pitcher 1975). Coordinates (x, y, and z) were corrected for perspective error (Crompton et al. 2003) and used to calculate the maximum speed (meters per second) achieved during each flight. Data presented here represent the maximum speed achieved by each individual over 5 flights.
Modeling swarming behavior
To simulate swarming in male midges, I have adopted the model developed by Couzin et al. (2002), in which simple attraction and repulsion rules can generate characteristic swarming behavior. This reference should be consulted for full mathematical details of the model. Briefly, swarms consisted of 100 individual males (i = 1, 2, ..., 100) with position vectors ci and unit direction vectors vi simulated in continuous 3-dimensional space. Time was partitioned into discrete time steps t with a regular spacing 
(because the response latency of midges is unknown, was set to 0.1 s [see Couzin et al. 2002]) during which individuals assessed the position of n neighbors. The primary rule in the model was to move away from neighbors located within a "zone of repulsion" (modeled as a sphere centered on the individual with radius rr); this rule always took precedence over subsequent behaviors. Secondarily, when individuals were not performing an avoidance maneuver, they were attracted toward neighbors within their "zone of attraction" (a sphere centered on the individual with radius ra, where ra > rr). The desired direction of movement resulting from attractance or avoidance behavior was defined by the vector di. I set rr at 1 cm and ra at 10 cm, which generated realistically sized swarms. Each movement was subject to error (with a standard deviation [SD], 
, of 5 [Couzin et al. 2002]). Because midges are likely to respond to auditory rather than visual cues during swarming and mating (Fyodorova and Azovsky 2003), it was assumed they could detect others in any direction (i.e., the field of perception, 
[see Couzin et al. 2002], was set at 360°).
Each male in the model was characterized by 3 parameters: turning rate (
i), number of ectoparasitic mites (mi), and flight speed (si). Turning rate (in degrees per second) was drawn at random from a Gaussian distribution with mean 50.2 and SD 11.8 (from empirical data generated by Crompton et al. 2003). Males were parasitized with a probability of 0.04 (the average proportion parasitized in naturally occurring swarms; McLachlan 1999), and, if parasitized, the number of mites, rounded to the nearest integer, was drawn at random from a chi-square distribution with 1 degree of freedom (which approximates the frequency distribution of mites occurring naturally; AJ McLachlan, TW Pike, and JC Thomason, unpublished data); otherwise mi was set to 0. Speed (in meters per second) is a function of mite load and so was taken at random from a Gaussian distribution with mean 0.705 – 0.0782mi and SD 0.19 (see Figure 1).
The starting positions of males within the swarm were determined randomly around the origin (x = y = z = 0) so that each male was within at least one other's zone of attraction. In nature, swarms attempt to maintain an approximately stable elevation and horizontal location over a distinctive landmark (McLachlan and Neems 1995). To model this, after performing avoidance and attraction maneuvers, an individual's desired direction vector di was multiplied by –1 to maintain the swarm's center around the origin (this still allowed a degree of horizontal and vertical movement that is characteristic of natural swarms). Each swarm was flown for 100 time steps (equivalent to 10 s of real time), allowing them to stabilize before analysis of male positions and the introduction of a female.
Females j were characterized by a single parameter, speed (sj), which was drawn at random from a Gaussian distribution with a mean (0.78 ms–1) and SD (0.20), determined from empirical data (see Results). Assuming that females are parasitized at a similar level to males (probably only around 4% [McLachlan 1999] given the apparent inability of mites to distinguish between male and female hosts at the time of attachment [McLachlan et al. 1999; Edwards and Smith 2003]), and adjusting flight speeds accordingly has minimal impact on the outcome of the model, and so it is assumed in all simulations presented here that females are unparasitized. Each female was given a random starting position vector cj (outside of any males' zone of attraction) and a unit direction vector vj directed toward the center of the swarm at time step t = 0 (see Discussion for a consideration of the consequences of relaxing this assumption). After 100 time steps, she flew in a straight line until capture (a behavior that has been observed in natural swarms; Kon 1984). A variation on this straight flight path was also explored, where a female actively avoided capture by adjusting her desired direction vector dj away from males within her zone of repulsion, while still attempting to fly toward the swarm's center. Males responded to a female within their zone of repulsion by adjusting their desired direction toward her at the expense of both repulsion and attraction maneuvers toward other males. A male was considered successful at capturing a female if both ci and cj were within 0.1 cm of each other. After capture, the simulation was stopped and a new swarm generated. Altogether 1000 such swarms were generated and used for analysis.
Data analysis
The center of each swarm was calculated by finding the mean position vector of each male and its edge by averaging the position vector of the 5 outermost males (those with the greatest distance from the swarm's center). Distances are presented as proportional from the center of the swarm. Where the number of data points was very large (>1000) or data were categorical, they have been summarized and analyzed using Pearson's correlations. All other data met the assumptions of normality and were analyzed using general linear models or t-tests.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
Empirical findings
Unparasitized male midges achieved average maximum flight speeds of 0.74 ± 0.19 ms–1 (mean ± SD) (which is very close to the average maximum speeds recorded by Crompton et al. 2003
), and there was a significant negative relationship between mite number and the maximum flight speed (r2 = 0.49, F6,15 = 3.35, P = 0.022; Figure 1). Unparasitized female midges from the same population had an average maximum flight speed (±SD) of 0.78 ± 0.20 ms–1.
Results of the model
Flight speed
The model revealed marked spatial heterogeneity within the swarm as a result of individual differences in flight speed: there was a positive relationship between male flight speed (si) and distance from the center of the swarm (r = 0.92, P = 0.003; Figure 2a), with faster males tending to occupy the periphery of the swarm and slower males being found nearer the center. In addition, faster flying females penetrated deeper into the swarm before capture (r2 = 0.02, F1,998 = 17.69, P < 0.001; Figure 2b), and so there was a negative relationship between male and female speed in mating pairs (r2 = 0.01, F1,998 = 5.46, P = 0.021; Figure 2c), with faster males tending to catch slower females and slower males catching faster females. There was no significant relationship between a male's turning rate (i) and his spatial position within the swarm (P > 0.3).
|
Parasitism
Because the degree of parasitism impacts on male flight speed, there was a negative relationship between mite load and average distance from the center of the swarm (r = –0.97, P < 0.001; Figure 3a), with parasitized males swarming nearer the center. Consequently, parasitized males tended to capture faster females that entered further into the swarm (mean ± SD flight speed of females captured by parasitized males: 0.89 ± 0.16 and unparasitized males: 0.76 ± 0.19; t997 = 4.09, P < 0.001), resulting in a positive relationship between female flight speed and her probability of being captured by a parasitized male (linear regression weighted by the sample size used to calculate each proportion: F1,4 = 23.11, P = 0.009; Figure 3b). Overall, the proportion of females captured by a parasitized male was 0.036, which is not significantly different from the proportion expected if each male in a swarm, parasitized or otherwise, had an equal probability of mating (0.04; binomial test: P = 0.522). However, slower females were almost never captured by parasitized males, whereas faster females were more likely to mate with infected males than by chance (Figure 3b): if the data are analyzed in terms of female speed, then the proportion of "fast" females (those flying faster than the average speed of 0.78 ms–1) captured by parasitized males was significantly higher than the proportion of "slow" females (those slower than 0.78 ms–1) (0.057 vs. 0.017; 2-sample binomial test: z = 3.31, P = 0.001). Moreover, there was a significant tendency for parasitized males to capture fast females more frequently than by chance (binomial test vs. 0.04: P = 0.047), whereas with slow females infected males did significantly worse than chance (P = 0.004).
|
The model's assumption that females fly in a straight line until capture is simplistic but demonstrates that, given the passive adoption of spatial structure within a swarm, any mechanism that brings a female closer to the swarm's center before capture will increase her chances of mating with a parasitized male. For example, when the model was rerun so that females actively avoided capture by all males while still attempting to reach the center of the swarm, then the mean flight speed of successful males was significantly less than the average flight speed in the swarm (i.e., slower males had a greater mating success: 1-sample t-test, 0.72 ± 0.01 vs. 0.74 ms–1, t = 3.54, n = 1000, P < 0.001) and the total proportion of females captured by parasitized males was significantly greater than chance (binomial test vs. 0.04: n = 1000, P = 0.023). Similarly, larger and parasitized males enjoyed a significantly elevated mating success if faster flying males (faster than 0.74 ms–1) ignored faster (and hence smaller) females (faster than 0.78 ms–1), allowing them to get closer to the swarm's center before capture (larger males: 0.71 ± 0.01 vs. 0.74 ms–1, t = 4.99, n = 1000, P < 0.001; parasitized males: n = 1000, P = 0.002). The biological significance of these mechanisms is discussed below.
When the original model was rerun so that the number of mites infesting each parasitized male was drawn at random from a uniform distribution in the interval (0, 7) and only captures by parasitized males were recorded, I found that females flying faster than average (i.e., faster than 0.78 ms–1) tended to be captured by highly parasitized (slower flying) males (mean ± SD number of mites: 4.41 ± 1.79, n = 100), whereas females flying slower than average tended to be caught by faster males with less parasites (mean ± SD: 3.86 ± 1.89 mites, n = 100) (Figure 4). Although these 2 distributions differed significantly (t198 = 2.11, P = 0.036), the mean of both distributions and the combined mean (4.14) are approximately 4, suggesting that this may represent the optimal number of mites needed to increase male mating success regardless of female flight speed.
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
The results of the model presented here suggest that spatial heterogeneity of males in midge mating swarms, brought about through individual differences in flight speed, can affect mating success through access to females. The observed effect of a male's flight speed on his position in the swarm was predicted by the original model and results of Couzin et al. (2002)
because during flight the angle of turn changes as a function of speed, even though all individuals had a fixed turning rate per unit time in the simulation. Thus, slow-flying individuals, if turning in a similar direction for multiple time steps, will have a tighter turning angle than faster individuals. As a result, faster males tended to swarm closer to the periphery of the swarm, whereas slower males (including those burdened with ectoparasitic mites) tended to swarm nearer the center. This simple model demonstrates that, given the passive adoption of spatial structure within a swarm, any behavior that brings a female closer to the swarm's center before capture will increase her chances of mating with a parasitized male. Indeed, the proportion of fast females (those flying faster than 0.78 ms–1) captured by parasitized males was both significantly greater than the proportion of slow females caught by infected males and significantly greater than chance, suggesting that slower flying (and hence parasitized) males enjoy an increase in mating success in situations where females are predominantly fast flyers. Similarly, when only parasitized males are considered, we find that fast females tended to be captured by highly parasitized (slower flying) males, whereas slow females were predominantly caught by faster males with fewer parasites.
In natural swarms, it is possible that larger (slower flying; Crompton et al. 2003) males may be preferred if body size is positively related to the quantity of sperm produced—and subsequent fertilization success—as in other dipteran species (e.g., Simmons and Parker 1992; Otrenen 1994) or indicates other desirable, heritable qualities such as foraging ability. Females may thus actively attempt to reach these larger males at the swarm's center. Indeed, adapting the model so that females actively avoid capture by all males while attempting to reach the center of the swarm allowed slower flying and parasitized males to significantly increase their mating success. Alternatively, small males may tend to ignore smaller (faster flying) females, which are often less fecund (Fox and Czesak 2000; Bonduriansky 2001); again, incorporating this mechanism into the model increased the reproductive success of parasitized males beyond that expected by chance. Although this may mean that slow-flying males only have the chance to mate with females with low fecundity, from the mite's point of view, this doesn't matter as long as mating (and hence the opportunity for host transfer) occurs. Any of these situations would take females into the zone of not only larger but also parasitized males before capture, as predicted by variations to the basic model outlined here, and so could explain the increased mating success of parasitized males observed in natural swarms (McLachlan 1999).
However, if it were adaptive to fly slowly in natural swarms, we would expect this strategy to be adopted by all males, not just large males and those constrained to do so by their parasitic burden, although with the obvious corollary that spatial structuring in the swarm would be lost. Consequently, in order to maintain such heterogeneity in natural swarms, despite the potential reproductive benefits of swarming near the center, the costs associated with adopting a central position would have to outweigh the benefits. Although the potential costs are not known for this system, it is possible that the costs of flying slowly are too much for most males to bear or individuals near the center of the swarm may suffer an increased chance of predation. There are many aerial predators for which chironomids probably form a common and easily available prey, including dance flies (Diptera: Empididae), and McLachlan et al. (2003) have estimated that large (slower flying) C. plumosus males are approximately 900 times more likely to be caught by dance flies than small males in the same swarms, although whether capture took place near the swarm's center is not known. Generally predation tends to be greater at the edge of groups than at the center (reviewed in Krause 1994), as it may be in midge swarms predated on by avian or mammalian predators, but because dance flies fly faster than their midge prey on average (McLachlan et al. 2003), it might be predicted that they will predominantly catch the slower flying individuals near the swarm's center. Alternatively, males may exhibit a mating preference for larger (and hence slower flying) females if size signals fecundity (a phenomenon observed in many insect species; reviewed in Fox and Czesak 2000; Bonduriansky 2001) and hence adopt peripheral positions in order to intercept these preferred mates. It is also possible that females that fly fast enough to penetrate into the swarm's center are in fact rarer than my empirical data suggest, and so parasitized males only occasionally enjoy increased reproductive success and not frequently enough for a slow-flying strategy to be selected. Any of these scenarios could maintain swarm heterogeneity while still increasing the reproductive success of parasitized males in some situation, although they remain to be empirically tested.
Even though there was a significant difference between the mean number of mites on males capturing fast and slow females, in both cases, the mean mite density approximated 4 suggesting that this level of parasitism may represent the optimal number of mites needed to increase male mating success (and hence a mite's chance of transferring to the female and being transported back to water to continue the next phase of its lifecycle), regardless of female flight speed. This has implications for the mutualism between parasite and host, whereby parasites benefit by increasing the mating success of their male hosts and males obviously benefit by being parasitized; but only up to a certain mite density after which both mites and midges incur the costs of low host mating success (an interesting example of the Allee effect). There may thus be competition between mites to attain (or maintain) optimal infestation densities and also the potential for host manipulation of mite number; for example, midges have been observed "grooming" their mite-infested ventral region (B Crompton, personal communication), although whether this is used to actively manipulate mite number is not known. In this model, an apparently optimal mite number probably occurred because highly parasitized, slower flying individuals tended to be "outcompeted" by faster individuals with fewer mites when maneuvering to capture a female and so could also apply to wild midges, although this remains to be tested.
Obviously, the model used here is simplified and does not take into account physical competition between males or include further constraints of parasitism, such as fatigue, which may affect how long infested males can swarm for. It also doesn't take into account female parasitization, although including probable natural rates of infestation in the model (only around 4%; McLachlan 1999) had very little effect on the findings presented here because of the small effect it had on the distribution of female flight speeds. More critical to the model is the initial direction vector of female (i.e., orientated toward the center of the swarm). If this parameter is relaxed so that females fly toward the swarm in a random direction, then the probability that a fast female is caught by a parasitized male becomes no different to chance (P > 0.05). However, in natural swarms, females do appear to fly toward the center of the swarm mass, rather than the edges (Downes 1969; Kon 1984; TW Pike, personal observation), and so this is likely to be a valid assumption. In contrast, this assumption was not important in determining the high mating success of parasitized males under either of the additional scenarios explored.
The model presented here suggests that the passive adoption of different spatial positions within a dynamically moving group could have significant fitness consequences across a range of vertebrate and invertebrate grouping systems, including many large fish shoals and insect aggregations where heterogeneity is likely to affect access to food, mates, and predators (e.g., Parrish 1993; Bumann et al. 1997) in 3 dimensions. However, because of the difficulties of studying these effects empirically, they remain poorly understood. Simple models such as this one could allow the formation of testable predictions in such systems to help better understand the mechanisms leading to and fitness consequences of interindividual differences in group position.
|
ACKNOWLEDGEMENTS |
|---|
I would like to thank Athol McLachlan, Andrew Jackson, and 3 anonymous referees for useful comments that helped greatly improve the manuscript and Ben Crompton and Jeremy Thomason for help with the high-speed photography and analysis of the empirical data.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
Barber I and Huntingford FA. (1996) Parasite infection alters schooling behaviour: deviant positioning of helminth-infected minnows in conspecific groups. Proc R Soc Lond Ser B 263:1095–102.
Bonduriansky R. (2001) The evolution of male mate choice in insects: a synthesis of ideas and evidence. Biol Rev 76:305–39.[Medline]
Bumann D, Krause J, Rubenstein D. (1997) Mortality risks of spatial positions in animal groups: the danger of being in the front. Behaviour 134:1063–76.
Couzin ID, Krause J, James R, Ruxton GD, Franks NR. (2002) Collective memory and spatial sorting in animal groups. J Theor Biol 218:1–11.[CrossRef][ISI][Medline]
Crompton B, Thomason JC, McLachlan AJ. (2003) Mating in a viscous universe: the race is to the agile, not to the swift. Proc R Soc Lond B 270:1991–5.[Medline]
DeBlois EM and Rose GA. (1996) Cross-shoal variability in the feeding habits of migrating Atlantic cod (Gadus morhua). Oecologia 108:192–6.[CrossRef]
Di Sabatino A, Gerecke R, Martin P. (2000) The biology and ecology of lotic water mites (Hydrachnidia). Freshw Biol 44:47–62.[CrossRef]
Downes JA. (1969) The swarming and mating flight of Diptera. Annu Rev Entomol 14:271–98.[CrossRef][ISI]
Edwards DD and Smith HG. (2003) Host sex preferences and transmission success by the water mite Unionicola foili (Acari: Unionicolidae) parasitic on the midge Chironomus tentans (Diptera: Chironomidae). J Parasitol 89:681–5.[CrossRef][Medline]
Fox CW and Czesak ME. (2000) Evolutionary ecology of progeny size in arthropods. Ann Rev Entomol 45:341–69.[CrossRef][ISI][Medline]
Fyodorova MV and Azovsky AI. (2003) Interactions between swarming Chironomus annularius (Diptera: Chironomidae) males: role of acoustic behaviour. J Insect Behav 16:295–306.[CrossRef]
Kon M. (1984) Swarming and mating of Chironomus yoshimatsui (Diptera: Chironomidae): seasonal change in the timing of swarming and mating. J Ethol 2:37–45.[CrossRef]
Krause J. (1993) The relationship between foraging and shoal position in a mixed shoal of roach (Rutilus rutilus) and chub (Leuciscus cephalus): a field study. Oecologia 93:356–9.[CrossRef]
Krause J. (1994) Differential fitness returns to spatial positions in groups. Biol Rev 69:187–206.[Medline]
Krause J and Godin J-GJ. (1994) The influence of parasitism on the shoaling behaviour of the killifish (Fundulus diaphanus). Can J Zool 72:1775–9.
Krause J, Godwin J-GJ, Brown D. (1996) Phenotypic variability within and between fish shoals. Ecology 77:1586–91.[CrossRef]
Krause J and Ruxton GD. (2002) Living in groups(Oxford University Press, Oxford).
McLachlan AJ. (1999) Parasites promote mating success: the case of a midge and a mite. Anim Behav 57:1199–205.[CrossRef][ISI][Medline]
McLachlan AJ, Ladle R, Bleay C. (1999) Is infestation the result of adaptive choice behaviour by the parasite? A study of mites and midges. Anim Behav 58:615–20.[CrossRef][ISI][Medline]
McLachlan AJ, Ladle R, Crompton B. (2003) Predator-prey interactions on the wing: aerobatics and body size among dance flies and midges. Anim Behav 66:911–5.[CrossRef]
McLachlan AJ and Neems RM. (1995) Swarm-based mating systems. In Leather SR and Hardie J (Eds.). Insect reproduction(CRC Press, New York) pp. 199–214.
Neems RM, McLachlan AJ, Chambers R. (1990) Body size and lifetime mating success of male midges (Diptera: Chironomidae). Anim Behav 40:648–52.[CrossRef]
Okubo A. (1980) Diffusion and ecological problems: mathematical models; lecture notes in biomathematics(Springer-Verlag, New York) Volume 10:.
Otrenen M. (1994) Fertilization success in the fly Drymyza anilis (Dryomyzidae): effects of male size and the mating situation. Behav Ecol Sociobiol 35:33–8.[CrossRef]
Parrish JK. (1989) Layering with depth in a heterospecific fish aggregation. Environ Biol Fishes 26:79–86.
Parrish JK. (1993) Comparison of the hunting behavior of four piscine predators attacking schooling prey. Ethology 95:233–46.
Pitcher TJ. (1975) A periscope method for determining the three-dimensional positions of fish in schools. J Fish Res Board Can 32:1533–8.
Pitcher TJ, Magurran AE, Edwards JI. (1985) Schooling herring and mackerel choose neighbours of similar size. Mar Biol 86:319–22.[CrossRef]
Simmons LW and Parker GA. (1992) Individual variation in sperm competition success of yellow dung flies, Scathophaga stercoraria. Evolution 46:366–75.[CrossRef]
Thornhill R. (1980) Sexual selection within mating swarms of the lovebug Plecia nearctica (Diptera: Bibioidae). Anim Behav 28:405–512.[CrossRef]
Ward AJW, Hoare DJ, Couzin ID, Broom M, Krause J. (2002) The effects of parasitism and body length on positioning within wild fish shoals. J Anim Ecol 71:10–4.[CrossRef]
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||