The Mating Behaviour of a Swarm Forming Insect.

This essay supersedes and expands on earlier attempts to understand swarm based mating behaviour of chironomid midges (McLachlan, 2012a, 2014a, 2018; McLachlan and Neems, 1995; swarm based mating systems 2012b). Chironomids (Chironomidae), are not the dreaded scottish midge (Ceratopogonidae). I focus on the chironomid phenotype, and mostly on behaviour, i.e. response by movement to environmental stimuli. Behaviour is taken to include phenotype limited behaviours as in the size limited behaviours of male midges described below. Unravelling the complexities has proved tricky. There appear to be multiple mating behaviours within a single species but for years I had been trying to distil my observations down to a single mating behaviour. This was a fruitless activity. Over lunch in the senior common room my friend and colleague Max Hammerton asked why it had to be only one behaviour. This led to the thought that it could perhaps involve many alternative behaviours necessary to achieve mating under different environmental conditions. How stupid of me not to have thought of this earlier. Even then I could not see how multiple behaviours would work within a single species. Mary Jane West Eberhard (West-Eberhard, 2003), has the answer. In a inspirational book she shows how alternative phenotypes governed by environmental triggered switch points are widespread among both animals and plants. I hope to show that such a condition dependent approach overcomes many of the obstacles to understanding chironomid mating behaviour. Because I have more data on it than other species, I adopt Paratrichocladius rufiventris, as the main test species for the chironomidae in general but I draw in findings from other species from time to time, e.g. (McLachlan and Allen, 1987; McLachlan and Neems, 1989).

My interest in chironomid midges stems from studies of the larval stages which often dominate freshwater habitats (McLachlan and Ladle, 2009), p.136. Specifically it is those chironomid larvae inhabiting ephemeral waters such as the margins of lakes, rain pools, and upland rivers where   adaptation to the unpredictability of the habitat has attracted my attention (McLachlan, 1974a, 1974b; A. J.  McLachlan, 2014; McLachlan and Ladle, 2001; Walentowicz and McLachlan, 1980). By contrast to work on the aquatic larvae, the adult phase of the life cycle is an aspect of chironomid biology outside the ken of mainstream freshwater ecology. A study of adults requires a fundamental shift from feeding biology to mating biology (Andersson, 1994). Indeed the biology of mating introduces an entirely different world. To start at the beginning - like all sexually reproducing species, to produce a fresh generation of aquatic larvae an adult chironomid must find an individual of a very specific kind. Not only must it be of the right species but it must be of the opposite sex. And finding the appropriate species and sex is not the end of the matter as choice among mates is a fitness priority. Adult chironomids are active in the evenings during early spring and can be seen in huge swarms over landmarks such as tree tops (Fig. 1).

  Fig.1. A swarm of male midges (Chironomus plumosus), over a tree top on a spring evening. From (McLachlan, 2010).  

 Swarms are composed almost entirely of males, usually of a single species, swarming to attract patrolling females (McLachlan and Neems, 1995). A question immediately arises – how do males find each other to form swarms in the first place? Reflect for a moment. After emerging from water, these tiny flies will be scattered by wind and their own flight over many miles. This is like attempting to find another human in the Sahara desert. The solution has been for males to aggregate over a landmark. Here sight appears to be the primary sense, but the sound of high speed wing beats like the whine of mosquito help too. The elaborate antennae of the male (Fig.2.), is the sense organ responsible for detecting wing beat sound. This simplifies the task for patrolling females because the swarms emit a conspicuous auditory signal.

Fig.2. Antennae of a typical chironomid midge showing male antennae (b) and female antennae (d). (Modified after (Freeman, 1955), with permission).

So we have reached the point where male swarms, all of the same species, are present over a landmark. Now to bring the females into the picture. I once succeeded in a striking demonstration of the role of the wing beat sound of a female entering a swarm of males. Borrowing a set of tuning forks from our Physics department, I struck each fork in turn and introduced it into a swarm. Nothing happened until one fork in particular elicited a striking response. The whole male swarm of many thousand males immediately aggregated closely around the fork. I take this particular tuning fork to represent a female. So there is no doubt about the role of female sound. It is worth recording that the same observation involving male mosquitoes aggregating around a buzzing light was made by Hiram Maxim, the inventor of the machine gun, more that 140 years ago (Roth, Roth, & Eisner, 1966). But there is more to it than the response to sound among males and between males and females. When  a female finds and enters an appropriate swarm she searches – not just for wing beat sound of males but for harmonics that match the sound of her own wing beats (McLachlan, 2012b). The use of harmonics is a sophisticated adaptation indeed. Even then there can be mistakes. For example, homosexual pairing is by no means uncommon (McLachlan, 2011; Sales et al., 2017).

As I have already hinted, selection theory predicts that individuals are highly selective in choice of the ‘best’ possible mate among those available. What is meant by best mate is an intriguing question and it is here that research on the mating behaviour of animals has been directed over the last few decades (Thornhill and Alcock, 1983). We have already seen that individuals, both male and female in a swarm, are searching for harmonics of wing beat sound. So, mate choice depends neither  on the female alone as is normally the case, nor does it depend on the male as is sometimes the case, but rather may be mutual with both sexes collaborating in choice. Chironomid swarms are lek base mating systems better known among birds such as grouse (Davies, Krebs, and West, 2012). Leks involve male animals coming together in a definite place such as a patch of ground or treetop year after year. In a lek, males provide females with nothing but sperm - no resources such as shelter or food are exchanges for sex. After mating females leave the swarm to oviposit in nearby water. Successful males by contrast, can potentially return to the swarm and achieve further matings. With chironomids, it has proved difficult to identify the ‘mating system’, essentially comprising of the alternatives - multiple male matings (polygyny) or multiple matings by females (polyandry). I have never been able to distinguish between the two for chironomids but J. A. Downes (Downes, 1969), in a clever manipulation of Chironomus plumosus, swarms found polygamy. The Downes findings appear to be contradict those of Goran Arnqvist (Arnqvist, 1998). Thus this important aspect seems not yet to be resolved but may be explained, as is much else, by environmental switches between alternative tactics within a single strategy (see below).

There is more: size is not a continuous variable among males. Instead size frequency distribution shows three discrete size classes caalled α, β and γ, with α containing the largest males and γ the smallest (A. J. McLachlan, 2014b)(Fig.3).

Fig.3. Approximate relative sizes of α and γ male chironomids. The β males are intermediate. Scale c. 6x.

 Discontinuous size classes such as these carry with them quite separate phenotype limited mating tactics. The idea of phenotype limited tactics is not original but follows the findings of Shuster and Wade in a marine isopod (Shuster and Wade, 1991). The first hint of the existence to size classes among chironomids was obtained by taking net sweeps samples from swarms and separately from the grass under the swarm. These samples showed the size of male in the swarm and in the grass to be quite different with distinctly smaller males in the grass (McLachlan and Neems, 1989). Interestingly, we found that females attracted to the swarm rest in the grass before entering the swarm and must encounter the γ males there. The possibility exists therefore, that the smallest males never enter the swarm but mate on the ground. By contrast, swarm samples reveal β males to predominate. These, it appears, depend on aerobatics and endurance to obtain a mate (Crompton, Thomason, and McLachlan, 2003), cued by harmonics. Finally, α males, also to be found in the swarm reveal something unexpected (McLachlan, MacLeod, and Neems, 2018). At least to the human observer, (α) males look very like a common predator of the swarm – males of the empid fly Empis stercorea.  Male empids enter the swarm to capture a midge as a nuptial gift for their own females. The strong resemblance between α males and empids led to the idea that mimicry may be involved. What I mean is that α males may gain in fitness by mimicking their predator and thereby decrease the chance of a fatal encounter - and at the same time increase the time spent in the swarm and the increased probability of mating. An hypothesis of mimicry could be strengthened by measuring wing beat sound in both predator and prey. Regrettably I have not done this but when it is done I predict that the wing beat sound of large chironomids and empids would be closely similar. So there are three size determined mating tactics among males. Fitness payoff to the three can be calculated with tactics played off against each other in an ESS (Maynard Smith, 1982).The three distinct tactics might be determined genetically with two alleles in a simple Mendelian model. Assuming simple dominance the two alleles in the game become AA; aa; and the heterozygote Aa. The F1 generation would occur in the Hardy - Weinberg equilibrium with gene frequencies of 1AA; 2Aa; 1aa. Of course these genes would only be expressed in males. In principle it should be easy to verify a gene determined hypotheses by counting the proportions of the three morphs is a species population. In practice this may prove very difficult, if not impossible. Furthermore, size limited male behaviours could well be environmentally induced, in which case fitness payoff to the three tactics need not be equal and proportions in the population cannot be predicted. My colleague John Lazarus has shown how three size limited phenotypes might have evolved under selective pressures from the environment (i.e. condition determined body size), as apposed to genetically determined phenotypes, in a rock – scissor – paper game built on the principles of an ESS (A. J. McLachlan, 2014a).

I come now to the second example of environmental switch points, i.e. the effect of predators and of parasites. The body size effects discussed above, were shown in the male. By contrast, it is the female that shows an adaptive response to parasites and to predators.

Predators

A general problem for animals displaying to attract mates is that others are out there listening or looking for signals. For chironomids, a common predator of mating swarms is the empid fly Empis stercorea, (McLachlan, Ladle, and Crompton, 2003). Their presence acts as an environmental switch to quite a different mating system where harmonics for wing beat sound is no longer important. Instead the mating system becomes one driven by coercion, with the male pursuing reluctant females who mistake them for a predator. So the role of coercion need not be abandoned after all (McLachlan, A. J., 2012b).

Parasites

Two parasites commonly infect P. rufiventris during mating; the mite Unionicola  ypsilophora and the nematode worm Gastromermis rosea (Fig.4).

Fig. 4. Midges infected by parasites. Male with two mites (a). Female with about 7 mites (b).  Female  (c), with a worm (g), bursting out of the abdomen, in the act of  returning  to water to close its life cycle. Scale line 0.5mm.Modified after (McLachlan, 2012a).

Male chironomids typically carry one or two mites placed mid-ventral between thorax and abdomen. Females by contrast, bear up to 23 mites on the ventral surface of the thorax (McLachlan, 1999). Samples of mating pairs reveal infected females to predominate in these pairs, indicating significantly improved mating success for mite infected females. This is a counterintuitive finding. There are at least two hypotheses to account for such a strange outcome. First, since mating depends on aerobatics, females baring mites can be expected to be compromised in flight due to biomechanical drag effects – and indeed there is good evidence that this is so (McLachlan et al., 2003). Second, mites may be revealing a hitherto hidden sexual preference, (Ryan, 2017), pp148 – 167, in this case for  the colour red. Hidden sexual preferences are an intriguing but often overlooked possibility and here imply the existence in midges of colour vision for red. Colour vision for red in a fly is theoretically unexpected (McLachlan, 2009). In the case of P. rufiventris, as the species name indicates (rufiventris), this species has a red tinge to the ventral surface which may act as a sexual display ornament, enhanced by the presence of these bright red mites (Fig 5.). There is yet another aspect of the mite mediated sexual selection. During mating mites transfer from male to female – so mites might be considered a sexual transmitted disease.

Fig.5. Female P.rufiventris with several bright red mites on the ventral surface of abdomen.

As both parasite and host gain by the association, the relationship between mite and host should properly be considered a case of mutualism rather than parasitism. I say this because the host benefits by an improved mating success (fitness). At the same time after mating, the female midge departs to water to lay her eggs, providing the mite with the opportunity to leave the female midge and return to an aquatic phase in water. There appears to be nothing gained by mites on the male except the chance of transferring to a female. The relationship between mites and midges nicely illustrates my central theme; condition dependent switches between alternative phenotype. Mating is no simple matter.

Turning to the second parasite, the worm: here is an intriguing twist to the sexual selection story. Unlike mites, worms are not harmless but render any infected female sterile. Thus an infected female and any male mating with her achieve zero fitness. But as females rarely carry both worm and mite, by mating with a mite baring female the male avoids the cost of a complete loss of fitness. So midges are at the centre of a complex, mutual benefit co-adapted mite – midge – worm interaction (McLachlan, 2006). Such close reciprocal interactions between ‘host’ and ‘parasite’ calls to mind the extended phenotype concept of Richard Dawkins (Dawkins, 1982), p21.

Summary   

This essay attempts to bring together all that is known about the lek based mating system of the chironomid midge. The attempt turns on the realisation that behaviour of a single species is highly plastic and can change adaptively depending on the environment. I have been able to identify three condition determined switch points between ‘normal’ mating behaviours and adaptive phenotypic changes in response to environmental effects. Environmental effects include the response to parasites and to predators in the female midge and to body size in the male midge. The answer has not come easily. I took the chironomid midge Paratrichocladius rufiventris, a common non-biting midge, as a case study, not because it is an especially appropriate choice. Indeed, progress in unravelling mating behaviours of animals has depended largely on the careful choice of study subject against which to test hypotheses. Studies of the dung fly (Scatophaga), Parker (Parker, 1984), and of great tits by Lack (Lack, 1966), illustrate the point. The Sydney Brenner of Caenorhabditis elegans fame expresses the point..."choosing the right organism to work on ....was as important as choosing the right problem to work on (Brenner, 2019). Chironomids have proved a rather difficult subject, but a challenge to me following years studying the larval stages. I offer a largely theoretical framework to promote understanding of the adaptive functioning of the behaviour of what I take to be a typical chironomid midge. I point up ideas that are still in need of testing and how some might be tested. The main point to emerge, I hope, is the importance of  bringing adaptive reasoning more fully into an ecological framework. 

References

Andersson, M. (1994). Sexual Selection. Princeton: Princeton University Press.

Arnqvist, G. (1998). Comparative Evidence for the Evolution of Genitalia by Sexual Selection. Nature, 393, 784-786.
Brenner, S.  (2019). A lucky accident. Wits Review, 41. 38-41.

Crompton, B., Thomason, J., and McLachlan, A. J. (2003). Mating in a viscous universe: the race is to the agile, not to the swift. Proceedings of the Royal Society, London (B). 270, 1991-1995.

Davies, N. B., Krebs, J. R., and West, S. A. (2012). An Introduction to Behavioural Ecology. (4 ed.). Oxford: Wiley-Blackwell.

Dawkins, R. (1982). The Extended Phenotype. (1999 edition). Oxford: Oxford University Press

Downes, J. A. (1969). The swarming and mating flight of Diptera. Annual Review of Entomology 14, 171-297.

Freeman, P. (1955). A study of the Chironomidae (Diptera) of Africa South of the Sahara. . The Bulletin of the British Museum (Natural History), 4, 1-69.

Lack, D. (1966). Population Studies of Birds. Oxford: Clarendon Press.

Maynard Smith, J. (1982). Evolution and the Theory of Games. Cambridge, UK: Cambridge University Press.

McLachlan, A. J. (1974a). Development of Some Lake Ecosystems in Tropical Africa, with Special Refrence to the Invertebrates. Biological Reviews, 49, 365-397.

McLachlan, A. J. (1974b). Recovery of the Mud Substrate and its Associated Fauna Following a Dry Phase in a Tropical Lake. . Limnology and Oceanography, 10, 74 - 83.

McLachlan, A. J. (1999). Parasites promote mating success: the case of a midge and a mite. Animal Behaviour, 57, 1199-1205.

McLachlan, A. J. (2006). You are looking mitey fine: parasites as direct indicators of fitness in the mating system of a host species. . Ethology Ecology & Evolution 18, 233-239.

McLachlan, A. J. (2009). Do Flies See Red?, http://www.google.co.uk/atholmclachlan.blogspot.co.uk.

McLachlan, A. J. (2010). Fluctuating Asymmetry in Flies, What Does it  Mean? Symmetry, 2, 1099-1107

McLachlan, A. J. (2011). Homosexual Pairing within a Swarm-Based Mating System: The Case of the Chironomid Midge. Psyche, ID 854820, 5 pages.

McLachlan, A. J. (2012a). Phenotypic plasticity and adaptation in a holometabolous insect, the chironomid midge. ISRN Zoology 2012, 8 pages.

McLachlan, A. J. (2012b). SWARM BASED MATING SYSTEMS. 13/03/2012. http://www.atholmclachlan.blogspot.com/.

McLachlan, A. J. (2014). Blaxter Lough, http://www.google.co.uk/atholmclachlan.blogspot.co.uk.

McLachlan, A. J. (2014a). Phenotype limited male mating tactics among some non-biting midges. (pp. http://www.co.uk/atholmclachlan.blogspot.co.uk): Google.

McLachlan, A. J. (2014b). Phenotype Limited Mating Behaviour., http://www.google.co.uk/atholmclachlan.blogspot.co.uk.

McLachlan, A. J. (2018). http://www.google.co.uk/atholmclachlan.blogspot.co.uk.

McLachlan, A. J., and Allen, D. F. (1987). Male mating success in Diptera: advantages of small size . Oikos, 48, 11-14.

McLachlan, A. J., and Ladle, R. (2001). Life in the puddle: behavioural and life-cycle adaptations in the Diptera of tropical rain pools. Biological Reviews 76, 377-388.

McLachlan, A. J., and Ladle, R. (2009). The evolutionary ecology of detritus feeding in the larvae of freshwater Diptera. Biological Reviews, 84, 133-141.

McLachlan, A. J., Ladle, R., and Crompton, B. (2003). Predator-prey interactions on the wing: aerobatics and body size among dance flies and midges. Animal Behaviour, 66, 911-915.

McLachlan, A. J., MacLeod, K. J., and Neems, R. M. (2018). SSD revised. Sexual Size Dimorphism in the chironomid midge: A Sheep in Wolf's Clothing. , http://www.google.co.uk/atholmclachlan.blogspot.co.uk.

McLachlan, A. J., and Neems, R. M. (1989). An Alternative Mating System in Small Male Insects. . Ecological Entomology, 14, 85-91.

McLachlan, A. J., and Neems, R. M. (1995). Swarm based mating systems. In S. R. Leather & J. Hardie (Eds.), Insect Reproduction. Neww Yourk:CRC  Press
 McLachlan, A.J. 13/03/2012. http://google.co.uk/atholmclachlan.blogspot.co.uk.                                  Parker, G. A. (1984). Evolutionarily stable strategies. In J. R. Krebs & N. B. Davies (Eds.), Behavioural Ecology: An Evolutionary Approach. Oxford: Blackwell.

Roth, M., Roth, L. M., and Eisner, T. E. (1966). The allure of the female mosquito. Natural History, 75, 27.

Ryan, M. J. (2017). A Taste for the Beautiful. Princeton, New Jersey.: Princeton University Press.

Sales, K., Trent, G., Gardner, J., Lumley, A. J., Vasudeva, R., Michalczyk, L., et al. (2017). Experimental Evolution with an Insect Model reveals that Male Homosexual Behaviour ocours due to Inaccurate Mate Choice. . Animal Behaviour, 139, 51- 59.

Shuster, S. M., and Wade, M. J. (1991). Equal mating success among male reproductive strategies in a marine isopod. Nature, 350, 608-610.

Thornhill, R., and Alcock, J. (1983). The Evolution of Insect Mating Systems. London: Harvard University Press.

Walentowicz, A. T., and McLachlan, A. J. (Eds.). (1980). Chironomids and Particles: a field experiment with peat in an upland stream. Oxford: Permagon Press.

West-Eberhard, M. J. (2003). Developmental Plasticity and Evolution. Oxford: Oxford University  Press.

Sexual Size Dimorphism

Sexual size dimorphism in the chironomid midge: A sheep in wolf's clothing?

Athol J. McLachlan^a,*, Kirsty J. MacLeod^b, Rachel M. Neems^a.

^a School of Biology, University of Newcastle, Newcastle upon TyneNE1 7RU

^b Zoology Department, University of Cambridge, Downing Street, Cambridge XB2 3EJ.

ABSTRACT

We measure wing length as an indicator of body size for males and females in mating pairs of two ubiquitous species of chironomid midge, Microtendipes definis (Edwards) and Chironomus anthracinusZetterstedt, in the wild. These measurements show that the ratio of male to female size , sexual size dimorphism (SSD), scales hypoallometrically with overall size in the mating pair. This observation is consistent with the predictions of SSD theory and leads to testable hypotheses concerning the mating strategies of the chironomid midge. The measurement of SSD reveal a hitherto hidden relationship between the sexes in a mating pair, and opens the door to a better understanding of the midge mating system; explicitly a strong correlation is revealed between SSD and female size which leads to the suggestion that mimicry is central to male mating success in these midges. 

Key words: mimicry, predation, Rensch's rule,  sexual selection, size dimorphism.

 *correspondence: A. J. McLachlan, Larach Mhor, Torloisk, Isle of Mull, Argyll, Scotland, PA74 6NH. tel, 01688500103. E-mail address@ a.j.mclachlan@virgin.net.

INTRODUCTION

     The mating system of the chironomid midge is based on the mating swarm (McLachlan and Neems, 1995), the details of which are full of unsolved questions such as the nature of mate choice and predator avoidance (McLachlan, 2012). What is known is that the swarm functions as a lek where males form conspicuous assemblages on the wing which attract patrolling females. Females entering the swarm are captured by males to form pairs which descend to the ground where they can in turn be captured by us to make wing length measurements (McLachlan, 1999). Concurrently, empids are attracted to the swarm which they enter to attempt the capture of a midge (McLachlan, Ladle, & Crompton, 2003). The empid we studies, Empis tesselata,  is  a predacious fly which appears to make heavy use of the midge swarm as a source of  nuptial gifts for his own mate. Since males vastly outnumber females in a midge swarm the captured midge is likely to be a male. Hence, the predator/prey relationship is predominantly between male empids and male midges. For this reason we concentrate on males in this study. As so often, body size is important in this mating system. (McLachlan et al., 2003), and by extension, to predation as well. Both mating pressure (sexual selection) and predation pressure (natural selection) are likely to influence selection on body size and there may be trade-offs between sexual and natural selection.

      Among male midges three phenotypes have been recognised; ά, β and γ males, each appearing to adopt different size determined mating tactics (McLachlan, 2012). For example, both their appearance and wing-beat sound are quit distinct to each size determined phenotype (McLachlan et al., 2003), and size may thus have a strong bearing, not only on mating but on predation as well (Crompton, Thomason, & McLachlan, 2003).  In the first instance we do not set out to test specific hypotheses but rather we here explore what light can be shed on this mating system through a study of relative size of males and females in mating pairs. In this study we measure patterns of Sexual Size Dimorphism (SSD) in the midge mating system. Previous work (R. M.  Neems, Lazarus, & McLachlan, 1992; R. M. Neems, Lazarus, & McLachlan, 1998), has compared size frequency distributions in mating systems, but has not focused explicitly on the relative size of males and females. Our study reveals new testable hypotheses specifies in the Discussion, and lays foundation for future work. SSD is widespread among animals, both invertebrate and vertebrate, and has attracted much attention (Fairbairn, Blanckenhorn, & Szekely, 2007). SSD theory gives prominence to body size effects which are fundamental to the understanding of mating behaviour (M.  Andersson, 1994). Rensch’s rule (Maynard Smith, 1977), is important  to the study of SSD and states that SSD decreases with size when females are the larger sex but it increases in size when males are the larger sex. This holds both between and within species (M. Andersson & Wallander, 2004). Why SSD should be so widespread suggests adaptive advantages in the mating stakes based on the size relationship between the sexes. This is the observation which we here examine for the specific case of the chironomid midge.

     This paper is not a study of SSD or Rensch's rule per se which are both fully explored by Fairbairn and colleagues (Fairbairn et al., 2007). Our assumption is that the two common midge species available for study, Microtendipes difinis and Chironomus anthracinus, have a mating system typical of the chironomid midges. There is much conjecture in this paper. Our hope is that it provides sufficient material to encourage further work.

METHODS

     We obtained three samples of M. difinis from different mating swarms on different evenings, yielding a total of 203 mating pairs. Samples were collected at Washington Wildfowl Park in Northumberland (UK) in late 1980. One sample of C. anthracinus was also obtained, yielding a total of 50 mating pairs. Wings were removed from flies and measured under a stereo microscope at 10 magnifications. Wing length was taken between the anal notch and the distal edge of the wing, excluding the hair fringe (McLachlan, 1986). To test hypotheses about SSD, we use wing length as a proxy for body size, as it is a readily obtained measure, and is closely correlated with other measures of body size (McLachlan & Allen, 1987). Wing length measurements are hereafter referred to as body size.

     For each species, we tested the bias and extent of sexual size dimorphism by comparing male and female wing lengths using Students t tests and Mann Whitney U-tests, depending on estimates of variable normality. Correlation between male and female wing lengths was explored using Pearson’s product moment and Spearman’s rank tests, also depending on normality. Means are reported with standard deviations.

     To assess the allometric relationship between male and female wing length, and to determine if these species conform to Rensch’s rule (that SSD increases with the size of the male in a mating pair if males are larger than females, or vice versa), we used Model II major axis regression. This allowed us to ascertain the allometric exponent β of the power law, y = α × x^β that describes an allometric relationship: β is the slope of the log-log regression. If females are larger than males, and β<1 and="" are="" body="" decreases="" females="" hypoallometry="" if="" larger="" males="" or="" size="" ssd="" than="" with="">1, SSD increases with body size (hyperallometry). In these instances, Rensch’s rule would be supported. To carry out these regression analyses, we used the package “lmodel2” in R (Legendre 2015) to estimate β, and the package “smatr” (Warton et al. 2012) to test the null hypothesis that the slope is not different from 1. Male and female wing lengths were log-transformed before analysis. Probabilities are two-tailed, and 95% confidence intervals are reported. Note that ^β  here is not the same of  β  males discussed elsewhere.

There was a notable male wing length outlier in the M. dificilis sample (1mm smaller than the mean male wing length). We present results for the sample both with and without the outlier.

RESULTS

In both species, male wing length was significantly smaller than that of females (Table 1; M. difinis Mann Whitney test U₂₀₃= 16664.5, P < 0.001; C. anthracinus Mann Whitney test U₄₉= 710.5,, P < 0.001). In M. difinis this effect is maintained when a small male outlier is removed (Students  t test t_365.8 = -3.37, P < 0.001). According to Rensch's rule, we should therefore see a hypoallometic relationship between male and female body size, (SSD should decrease as body size increases, given that females are the larger sex).

Male and female wing length correlate positively in M. difinis(Spearman’s rho = 0.19, S₂₀₂ = 1144792, P < 0.01). The effect is maintained with the male outlier removed (Pearson’s correlation = 0.19, t₂₀₁ = 2.83, P < 0.01). By contrast, in C. anthracinus, male and female wing length do not correlate (Spearman’s rho =  0.10, S₄₉ = 18657.81, P = 0.47).

Table 1. Mean male and female wing lengths for i) M. difinis and ii) C. anthracinus. Results for i) are presented with and without a small male outlier. P values are reported for tests investigating the significance of the difference between male and female wing length.

Species	Male wing length mm ( ± s.d.)	Female wing length mm (± s.d.)	Difference (P)
M. difinis	4.39 ± 0.15	4.45 ± 0.18	< 0.001
M. difinis (without outlier)	4.39 ± 0.13	4.45 ± 0.18	< 0.001
C. anthracinus	5.20 ± 0.17	5.35 ± 0.31	< 0.001

Male and female wing length scale hypoallometrically in C. anthracinus (β = -0.001; Fig.1), and M. difinis with (β = 0.36), and without the outlier (β = 0.27) (Table 2). In all cases, the slopes of these relationships are significantly different from 1 (Table 2). Because females were significantly larger than males in both species, these hypoallometric relationships confirm that both C. anthracinus and M. difinis conform to Rensch’s rule.

Table 2. Results of major axis regressions between male and female wing lengths in C. anthracinus and M. difinis (the latter with, and without, a major outlier), reporting reduced major axis slopes (β), R²of the regressions, and 95% confidence intervals of the slope estimates. P values reported denote the significance of the difference between β and 1.

Species	R2	β	95% C.I.	P
C. anthracinus	0.00	-0.001	-0.23 - 0.23	<0 .001="" p="">
M. difinis	0.02	0.36	0.01 – 0.81	<0 .05="" p="">
M. difinis (without outlier)	0.04	0.27	0.09 – 0.48	<0 .001="" p="">

Fig.1 Male and female body size scale hypometrically in C. anthracinus (a), and M. difinis, with (b) and without (c) an outlier. (Note that in both M. difinis datasets, there is a significant correlation between male and female body size). The major axis regression slopes showing the correlations between male and female body size are shown in red; 95% confidence intervals are shown in grey.

 DISCUSSION

Wing length is used as a surrogate for body size in this study of sexual size dimorphism (SSD), in chironomid midges. Such a measure is a long established devise in insects in general (Fairbairn et al., 2007), p6,  and  in chironomids and related Nematocera in particular (Maiga, Dabire, Lehmann, Tripet, & Diabate, 2012; McLachlan & Allen, 1987; Packer & S., 1989; Takamura, 1999; Xue & Ali, 1994; Yuval, Wekesa, & Washino, 1993). This holds for both sexes (Packer & S., 1989; Xue & Ali, 1994). Size differences between the sexes are almost universal in animals, and   follows from the different selective pressures to which each sex is subject (Greenwood & Adams, 1987). Among chironomid midges males are selected for agility and females for load baring (McLachlan, 1986). The interpretation of Renscsh's rule depends entirely on these differences. To quote (Fairbairn et al., 2007), p9...."SSD primarily reflects the adaptation of each sex to its distinct reproductive role".

Sexual size dimorphism (SSD) is measured here as the difference between size in each sex in a mating pair.  This is a rigorous approach because it eliminates that non-mating part of a species population which does no contribute to the next generation. This is the approach adopted by (Blanckenhorn, 2007). Making individual midges responsible for obtaining a sample of a species population avoids human based sampling errors.

     A comparison of the ratio of male over female size (SSD) in the place of size distribution data for each sex separately, as previously studied (McLachlan & Neems, 1995), reveals something new. That is, a strong correlation appears between SSD and female size. Such a finding supports Rensch's rule and supports the suggestion that there is some sort of change in mating tactic as male size in mating pairs increases (M. Andersson & Wallander, 2004). The change in male mating tactic moves from aerobatic agility for small males (γ males) (Crompton et al., 2003), to something else. We hypothesise that the large males (α males), depend not on aerobatics but rather on the imitating of their own empid predator? The suggestion rests on two observations. First, that there is a size related change in the ratio of males to females. Second, that there is a strong similarity in both appearance (to the human observer) and flight characteristics of alpha males and empids (McLachlan et al., 2003). We find this an arresting hypothesis but it awaits testing.                                       

Our study suggests the possibility of mimicry between midge and empid, specifically the mimicry of a male animal of its predator to achieve a mating, hence the suggestion that the α male should be regarded as a sheep in wolf's clothing. This suggestion begs an adaptive explanation i.e. what does the α male gain by such mimicry? Research on this mating system suggests that mating success for α males can be expected to depend on the length of time he can remain active in the swarm. Since the time spent is curtailed if he falls victim to an empid and since the empid is less likely to attack a fellow empid there are clear fitness gains for the mimic. We recently stumbled upon this insight in the beautiful experiments of (Greene, Orsak, & Whitman, 1987). They discovered a close convergence in appearance and behaviour of a fly and a spider suggesting that the prey is imitating its own predator to escape capture. They point out that: "..."the mimicry of a predator behaviour by its own prey had never before been recorded". We are pleased to acknowledge their priority and regard their aphorism as independent corroboration of the same thing within the mating world of the chironomid midge. Since the work of Greene et al., other examples have come to light, notable the experimentally verified mimicry of the hawk by the cuckoo (Davies, 2008). The mimicry of a male animal of its predator as a means of acquiring a mate would appear an unusual case of sexual selection which opens up conjecture over its evolution. We do not pursue the matter further here. Note that if this is indeed an example of mimicry, it is mimicry hinging on biomechanics, including the adaptive convergence of predator and prey on wing beat sound. Note too that any relationship between size of males and females in mating pairs becomes evident only when size is first converted to SSD. That is the value of SSD in this study of the pairing behaviour in the sexes. 

     We wish to emphasise that this paper is not about mating per se except as a consequence of predation. In other words, there may be a pre-existing sensor bias (Ryan, 2017), on wing beat sound that has been co-opted to the ultimate adaptive role in mimicry.

     The idea of convergence of flight behaviour between predator and prey is testable. A combination of emerging imaging technologies (Crompton et al., 2003; McLachlan, 2010, 2011; McLachlan et al., 2003; McLachlan, Pike, & Thomason, 2008), and of models of predator and prey along the lines of the inspirational work on the empid Ramphomyia longicauda (Funk & Tallamy, 2000) might be rewarding.  Even simple counts of the composition of mating pairs emerging from a swarm might serve the purpose.

We end with a note of caution. We should not try too hard to force ridged mating tactics on a population. Tactics may vary for a variety of ecological reasons as dramatically demonstrated by Thornhill and Alcock, p 281-277, for Panorpa scorpion flies (Thornhill & Alcock, 1983).

Acknowledgements

Michael Cant provided helpful comments on an early draft.  Research was supported by SERC studentship no. 87305549 to R.M.N.

References

Andersson, M. (1994). Sexual Selection. Princeton: Princeton University Press.

Andersson, M., and Wallander, J. (2004). Relative Size in the Mating Game. Nature, 431, 139-141.

(Blackenhorn, 2007). Case studies of the differential equilibrium hypothesis of sexual seize dimorphism in two dung fly species. In D. J. Fairbairn, W. U.

Crompton, B., Thomason, J., & McLachlan, A. J. (2003). Mating in a viscous universe: the race is to the agile, not to the swift. Proceedings of the Royal Society, London (B). 270, 1991-1995.

Fairbairn, D. J., Blanckenhorn, W. U., & Szekely, T. (Eds.). (2007). Sex, Size and Gender roles: Evolutionary Studies of Sexual Size Dimorphism. Oxford: Oxford University Press.

Funk, D. H., & Tallamy, D. W. (2000). Courtship Role Reversal and Deceptive Signals in the long-tailed dance fly Ramphomyia longicauda. Animal Behaviour, 59, 411-421.

Green, E., Orsak, L. J., & Whitman, D. W. (1987). A tephritid fly mimics the territorial displays of its jumping spider predators. Science, 236, 310-312.

Greenwood, P. J., & Adams, J. (1987). The Ecology of Sex. London: Edward Arnold.

Maiga, H., Dabire, R. K., Lehmann, T., Tripet, F., & Diabate, A. (2012). Variations in energy reservs and role of body size in the mating system of Anopeles gambiae. Journal of vector ecology, 37, 289-297.

Maynard Smith, J. (1977). Parental Investment: a prospective analysis. Animal Behaviour, 25, 1-9.

McLachlan, A. J. (1986). Sexual dimorphism in midges: strategies for flight in the rain-pool dweller Chironomus imicola (Diptera: Chironomidae). Journal of Animal Ecology, 55, 261-267.

McLachlan, A. J. (1999). Parisites promote mating success: the case of a midge and a mite. Animal Behaviour, 57, 1199-1205.

McLachlan, A. J. (2010). Fluctuating Asymmetry in Flies, What Does it  Mean? Symmetry, 2, 1099-1107

McLachlan, A. J. (2011). Homosexual Pairing within a Swarm-Based Mating System: The Case of the Chironomid Midge. Psyche, ID 854820, 5 pages.

McLachlan, A. J. (2012). Phenotypic plasticity and adaptation in a holometabolous insect, the chironomid midge. ISRN Zoology                2012, 8 pages.

McLachlan, A. J., & Allen, D. F. (1987). Male mating success in Diptera: advantages of small size . Oikos, 48, 11-14.

McLachlan, A. J., Ladle, R., & Crompton, B. (2003). Predator-prey interactions on the wing: aerobatics and body size among dance flies and midges. Animal Behaviour, 66, 911-915.

McLachlan, A. J., & Neems, R. M. (1995). Swarm based mating systems. In S. R. Leather & J. Hardie (Eds.), Insect Reproduction. New York: CRC Press.

McLachlan, A. J., Pike, T. W., & Thomason, J. C. (2008). Another kind of symmetry: are there adaptive benefits to the arrangement of mites on an insect host? Ethology Ecology & Evolution, 20, 257-270.

Neems, R. M., Lazarus, J., & McLachlan, A. J. (1992). Swarming behavoiur in male chironomid midges: a cost-benefit analysis. Behavioural Ecology, 3, 295-290.

Neems, R. M., Lazarus, J., & McLachlan, A. J. (1998). Lifetime reproductive success in a swarming midge: trade-offs and stabilizing selection for male body size. Behavioural Ecology, 9, 279-286.

Packer, M. J., & S., C. P. (1989). Size variation and reproductive success of female Aedes punctor (Diptera: Culicidae). Ecological Entomology., 14, 297-309.

Takamura, K. (1999). Wing length and asymmetry of male Tokunagayusrika akamusi chironomid midges using alternative mating tactics. Behavioural Ecology, 10, 498-503.

Thornhill, R., & Alcock, J. (1983). The Evolution of Insect Mating Systems. London: Harvard University Press.

Xue, R. D., & Ali, A. (1994). Oviposition and body size of a pestiferous midge, Chironomus crassicaudatus (Diptera: Chironomidae). Journal of the mosquito control association., 10, 29-34.

Yuval, B., Wekesa, J. W., & Washino, R. K. (1993). Effect of body size on swarming behaviour and mating success of male Anopheles freeobnni (Diptera: Culicidae). . Journal of Insect Behaviour, 6, 333.

Eating a Lake. How insect larvae transform lake mud to their needs.#

# This title was inspired by Olive Morton (2007), 'Eating the sun'.

One of Charles Darwin’s greatest contributions to our understanding of the world was his appreciation of the huge changes wrought on the environment by tiny inconspicuous animals (Darwin, 1881). I refer to the burrowing activities of the humble earth worm, each individual making a miniscule contribution to a dramatic overall effect on their soil habitat. The loosening of the soil sometimes leads to human buildings sinking slowly into the ground over the years. Tiny changes with large cumulative effects are not confined to earthworms. Other well known examples are termites and moles.

I wish to consider the role, not of soil dwellers, but of inhabitants of the mud of lakes. I will attempt to show that mud dwellers slowly consume the bottom of a lake and by so doing profoundly change the nature of their habitat. In the words of Lawton and Jones (1995), these organisms are ecosystem engineers. These aquatic ‘earthworms’ are typically dominated by insect larvae rather than annelids, particularly larvae of the chironomid midge (Armitage, Cranston, and Pinder, 1995; McLachlan and McLachlan, 1976). But it is not easy to demonstrate their effect. This is because most lakes are thousands or millions of years old and we arrive to collect samples at an unknown point in the lake’s history. What is needed is a situation which permits access to a lake before mud dwelling animals can start to reshape it. Just such an opportunity is provided regularly by a common type of giant swanp lake, such as Lake Chilwa in Malawi (Africa) (Fig. 1.). Such lakes typically occupy thousands of km² when full but dry up completely and refills every few years.

Fig. 1. Map of Chilwa showing open water with depth contours and surrounding swamp (k). After (McLachlan, 1974).

With great good fortune I was appointed to a lectureship at Chancellor College (University of Malawi) for the period 1965 to 1970. Chancellor College is conveniently close to Chilwa and when I arrived was nearly completely dry but refilled too during my tenure. What an excellent opportunity to study the whole process of drying and recolonization and the accompanying ecosystem changes. At that time none of us knew what to expect during a refilling of such a large lake, but Professor Margaret Kalk, head of the zoology department, set up the Lake Chilwa Coordinated Research Project at just the right time. The appearance of Chilwa during a typical dry and refilling cycle are shown in Figs. 2 and 3.

Fig. 2.   Chilwa full after 6 weeks with fishing in full swing and with Typha swamp in the background (photograph by G. Lenton).

Fig. 3. Chilwa drying with Brian Moss inspecting the dry lake scattered with the shells of the snail Lanistes ovum (see also Fig 9.3 in (McLachlan, 1979) 

Lake Chilwa covers some 2000 km², half of which is impenetrable Typha swamp. It is the open water habitat which concerns me here. To start at the point of drying – at this time the dry mud surface was scattered with the shells of the mollusc Lanistes ovum (Fig. 3). The presence of these shells infers that before drying the mud supported a thriving population of this snail. For various reasons which I will not go into here, I believe these molluscs inhabited the mud of the open lake rather than being migrants from the Typha swamp. In any case, their presence on the mud surface in Fig.3. is sufficient to suggest that the mud before drying was firm enough to support these large snails.

During refilling the dry lake bottom is stirred and eroded by wave action leading to a fine precipitate of mud several cm deep. Informal trials in the laboratory showed that this loose material was incapable of supporting even the larvae of the common midges measuring a up to nearly 2 cm in length and presumable thousands of times lighter than the golf ball sized Lanistes. Mature larvae of the chironomid Chironomus transvaalensis introduced to a beaker of water with precipitated mud from the newly flooded lake, simply sank through the mud surface and kept going to the bottom of the beaker.  Associated characteristics of the fine material were described by me in 1974 (McLachlan, 1974). It can reasonable be concluded that the same effect explains the complete absence of both chironomid larvae and Lanistes from the open lake mud just after refilling. However, larvae of chironomids but not Lanistes were abundant wherever there was any solid substrate such as that provided by the submerged leaves of typha and other aquatic vegetation at the edge of open water - but there only (Fig. 4.).

Fig. 4.  Distribution of animals just after refilling in February 1969. The area of each circle is proportional to the biomass of the faun. Hatched area – C. transvaalensis. Unshaded part of the circles - other fauna. Shading of the open water can be ignored for the present purposes. After (McLachlan, 1974). The swamp is not shown.

So, at some point between refilling and drying there appears to have been a change in the physical structure of the mud surface which then became able to support a dense fauna of chironomid larvae but still no Lanistes. This leads me to the hypothesis that it is the activity of the larvae of chironomids that converts the original almost liquid mud to a structurally firm one. I propose that it is specifically the feeding activities of chironomid larvae that are responsible. Feeding results in the production of hard faecal pellets, each larva producing a substantial 0.2mg of pellets day^-1 (McLachlan and McLachlan, 1976). The larvae of chironomids are thus an extraordinary lake-bed processing engine. So it is probably the accumulation of pellets in the mud that is ultimately responsible for the changes in the mud habitat. I imagine the process starting at the swamp margin and progressing to the lake centre. When the centre is reached the process is complete - provided the process is not interrupted by a dry phase. 

I got close to this realisation in the late seventies. “As pointed out by us (McLachlan and McLachlan, 1976), chironomid larvae have the ability to change the prevailing particle size in their environment. This is done by converting the fine material taken in the search for food into aggregates many time greater in diameter, by the formation of faecal pellets, these pellets are ideally suited to tube building operations and, being bound together with silk, persist for a considerable time, perhaps years, before eventually disintegrating. Given time, a larval populations can eventually process a sediment such as that on the Lake Chilwa bed, to create a habitat more acceptable to themselves”, (McLachlan, 1979) - and it should be added, to other animals such as Lanistes as well. To be clear, I have simplified the situation on Lake Chilwa by putting aside the important role of water chemistry and the role of the silk chironomid larvae use in construction the tubes they inhabit. Such details can be found in the publications listed below.

To summarise, I suggest that, by consuming the young lake bottom chironomid larvae provide the proximate explanation for the mud becoming habitable to other animals. Faecal pellets provide the ultimate cause. In brief chironomid larvae, by consuming the lake bottom, turn it into faecal pellets.

I am not suggesting that the consequences of chironomid feeding are some kind of community level adaptation. Explicitly not so. These effects are by-product effects which may benefit other speies but are clearly not an adaptation evolved to help others (Williams, 1966, p247)(McLachlan, 2011, p545).

The time has come to test my hypothesis. Passing soil through screens of various mesh sizes is a standard method for characterising the physical nature of soil in terms of particle size composition. The same method is applicable to the mud of lakes of course. By screening mud through an appropriate series of sieves, at regular time intervals immediately after filling starts and for as many years as necessary thereafter, the particle size distribution of the mud habitat can be recorded. Specifically, the role of faecal pellets can be determined because as trials showed, mud passes through a 105µ mesh sieve leaving virtually nothing but faecal pellets behind (McLachlan and McLachlan, 1976). The sieve method thus gives neat and clear results. I predict an increase in the proportion of pellets over the years leading eventual to a stable mud substrate. The character of faecal pellets found in the field will benefit from a separate programme to confirm their resistance under various conditions, see also (Joyce, Warren, and Wotton, 2007).

Processing the lake bed by consuming it is a vivid example on a grand scale of what has become known as niche construction (Odling-Smee, Laland, and Feldman, 2003), i.e. the restructuring of habitats by the activities of the inhabitants. In an historical context, lakes such as Chilwa provide a wonderful opportunity for the prepared mind. By resetting the ecological clock such lakes can reveal how the ecosystem gets started as well as the full course of succession following from this - something hidden in most lakes. It seems to me that it is this opportunity that lead my colleague and friend Brian Moss from his studies of Chilwa in the 1960s, to a novel and compelling view of the nature of the lake ecosystem, a concept set to succeed and replace the traditional view (Moss, 2015).

Re-reading this post in January 2021 suggested an alternative introduction to me. I offer it below:

I am interested here in animal engineering reviewed by Hansell (Hansell, M. H. 1984). His concern is with the building activities of animals leading to the construction of termitaria, the dwelling tubes of caddis fly nymphs. earthworms in the soil and many, many others. Hansell's could have included, in his  fascination review, the tube building activities of the larvae of the chironomid midges which, in terms of the variety of structures produced, far outstrip that of  the cadddis flies (Imada, Y. In press).

 

references

Armitage, P., Cranston, P. S., and Pinder, L. C. V. (1995). The Chironomidae. The biology and ecology of non -biting midges. . London: Chapman & Hall.

Darwin, C. (1881). The Formation of Vegetable Mould, through the Action of Worms, with Observations on Their Habits. . London: Murray.

Joyce, P., Warren, L. L., and Wotton, R. (2007). Faecal pellets in streams: their binding, breakdown and utilization. Freshwater Biology.

Hansell, M. H. (1984). Animal Architecture and Building Behaviour. Longman, London and New. 

Imada, Y. (In Press). Diversity of Underwater chironomid tube structures. Zookeys, 47834.
Lawton, J. H. & Jones, C. G. (1995).Linking species and ecosystems: organisms as ecosystem engineers. In: Jones CG, Lawton JH (eds). Linking species and ecosystems. Chapman & Hall, London.

McLachlan, A. J. (1974). Recovery of the Mud Substrate and its Associated Fauna Following a Dry Phase in a Tropical Lake. . Limnology and Oceanography, 10, 74 - 83.

McLachlan, A. J. (1979). Decline and Recovery of the Benthic Invertebrate communities. In M. Kalk, A. J. McLachlan and C. Howard-Williams (Eds.), Lake Chilwa. Studies of change in a Tropical Ecosystem. London: W. Junk. Publishers.
McLachlan, A. J., and McLachlan, S. M. (1976). Development of the mud habitat during the filling of two new lakes. Freshwater Biology, 6, 59 - 67.
McLachlan, A. J. and Ladle, R. J. (2011). Barriers to adaptive reasoning in community ecology. Biologival Reviews. 86, 543 - 548.
Morton, O. (2007). Eating the Sun. How Plants Power the Planet. Fourth Estate. London.

Moss, B. (2015). Mammals, freshwater reference states and the mitigation of climate change. Freshwater Biology, 60, 1964 - 1976.

Odling-Smee, F. J., Laland, K. N., and Feldman, M. W. (2003). Niche Construction. The Neglected Process in Evolution. Princeton: Princeton University Press.
Williams, G. (1966). Adaptation and Natural Selection. Princepton Univerity Prtess. Princeton, New Jersey.