Chapter Five

Mate choice in Papilio dardanus

Figure 5-1 A hippocoonides female rejecting a male by landing with wings open, and fluttering them

Behavioural observations
Experiment 5-1
Experiment 5-2
Experiment 5-3
Experiment 5-4

"I find that a male lepidopteran has almost always beaten me to any given female"
Burns (1968)


Cook et al. (1994) studied male mate preferences in the tibullus race of Papilio dardanus found on Pemba, Tanzania. This population of the race tibullus has a tail-less male-like female (andromorph) called trimeni, the common black and white morph hippocoonides, which mimics Amauris niavius (Trimen, 1869), and both of these morphs can appear with orange on the hind-wings which are known as yellow and white lamborni, or prototrophonius and trophonius, respectively (Cook et al. 1994). See Appendix 1 for further details of the races and morphs, and Chapter 3 for further details about this population in particular. Cook et al. pinned out pairs of dead female morphs (and also males and Amauris niavius specimens) and recorded the number of 'approaches' by males (defined as when a male came within 30cm of a pinned out female) to each morph, the amount of time each spent within 30cm of the morph, and whether or not the male attempted to copulate with the female. They also repeated the experiment using only singly presented morphs to avoid possible interaction effects resulting from presenting two females simultaneously. Finally, they followed live females, and recorded the rate of male approaches whilst the different morphs were flying.

The first two experiments gave the same results: the males approached the black and white hippocoonides morph significantly more than trimeni or lamborni, and trimeni females were approached significantly more often than lamborni or old, faded, trimeni females. The difference between the number of approaches to trimeni and lamborni females versus pinned-out males was insignificant. Faded males were approached significantly less frequently than any of the other specimens, and Amauris niavius (the model for hippocoonides) was approached significantly less often than hippocoonides. Males frequently attempted to copulate with the pinned-out females, but never attempted to copulate with males or Amauris niavius. See Appendix 5 for further details on these results. The differences in interaction times that the males spent with each morph followed the same pattern. When Cook et al. followed flying females, they found that hippocoonides females were approached significantly more than trimeni females. The presence of a male preference would have very interesting consequences for the evolution of the species and the morphs, and would affect the balance of morphs in different populations.

Assuming that the polymorphism is balanced, there should be no selective advantage for males which show a preference for a particular morph as each morph must on average leave the same number of offspring. Therefore the results raise two questions. Firstly, how does the population remain in balance? This question is addressed in Chapter 7, but to do that it is necessary to answer the second problem - what is the basis of the observed male preference - and for this there are many possible explanations. Firstly, the black and white females could have been more conspicuous to the males simply due to their coloration contrasting more with the background. To test this, Bateson et al. (unpubl.) pinned hippocoonides and trimeni females on 'cryptic' or 'noncryptic' backgrounds (either plain card or card covered in leaf litter). They found that the preference for hippocoonides remained even when it was made apparently less conspicuous than the trimeni morph. However, the results of Chapter 3 show that the white patches on hippocoonides would be 'true white' (reflectant over their entire visual range) to the butterflies, and therefore they are likely to stand out against the non-UV reflective leaf litter. This is because they would differ both by a difference in hue (since Experiment 3-6 indicated that the butterflies possess red receptors which will not be stimulated by the green background) and brightness (Swihart, 1971 demonstrated that butterflies do detect differences in brightness). The trimeni morph, however, may well have the same degree of brightness as the background, and also their hue will appear shifted more towards green to the butterflies (if they possess true colour vision) due to the UV reflectance, whilst the non-UV reflectant background will remain a greenish coloration.

Another explanation could be that the males perceive the hippocoonides females as being more obvious due to some perceptual bias for their coloration in general. For this case to be demonstrated, it would have to be shown that there was no other basis for the preference. In Chapter 4, Papilio dardanus was shown to have an innate preference for blue flowers over red and yellow. It would therefore be interesting to test whether or not this preference was also shown in mate choice (as is the case in Argynnis paphia - see Chapter 2). However, in pilot tests, males would not react to dead or model females, even when made to rotate on a 'merry-go-round' apparatus, and spin on their axis to give a 'fluttering' effect as was done by Magnus (1958) for Argynnis paphia. Pilot tests in which hippocoonides females were artificially coloured using permanent markers also proved unsuccessful, as the females were never mated, even when they initiated vigorous courtship with the males. It is possible that this was due to the marker pen removing scent-scales from the wings of the females which would normally be brought into contact with the male during courtship and elicit a mating reaction. It was therefore not possible to investigate the presence of a 'hidden preference' for artificial blue morphs in the flight cage (although the fact that males reacted to dead pinned-out females in the field experiments by Cook et al. indicates that it might be possible to perform this test in the field).

A further possible explanation for the apparent bias for hippocoonides involves learning. Males have to be able to recognise females of their own species. Cook et al.'s experiments show that males of Papilio dardanus can do this by sight, although possibly only at close range, given that no males attempted to copulate with a pinned out male or female of the wrong species even though they were approached (the dead specimens would probably not have much residue of pheromones on them). It would be difficult for males of a polymorphic (and mimetic) race to recognise females of their own species by sight innately, and it is possible that males might learn to recognise their own species after their first mating. This would be similar to the learnt flower preferences in butterflies discussed in Chapter 4. Virgin males may approach females at random, but after one mating they may show a preference for females of the same morph as their first mate. Such a preference would explain the observed bias towards the most common morph, and can be tested experimentally using males with a known history of sexual encounters. So, the males may appear to show a preference for the commonest morph (whichever morph that is) simply because their past experiences are most likely to have been with this morph. Should the morph frequencies change, so would the males' apparent preference.

A final explanation for the popularity of the most common morph is that there is a genetic basis to male choice. Males carrying the genes for a morph may also carry the genes for preference for that morph (similar to that which appears to be present in Danaus chrysippus; Smith, 1984). If a gene controlling morph preference were to exist it would be expected to be in linkage disequilibrium with that for the morph coloration itself due to assortative mating between the carriers of the gene for coloration and that for the preference. If male choice were to be controlled by a gene for a morph preference, then the males would be expected to show the same preference all their life (although this may only be a preference, and they may mate with other morphs, though less frequently), and may provide similar data to the case for learning. The theory could easily be tested by determining the genetics of males with a certain preference. It is also possible that there is a gene for a morph preference which is not associated the gene for the morph itself, and to test this males could be allowed to mate only with one particular morph first, and then their subsequent choices studied.

In addition, there may be an element of female choice. The preferences of the males could be evolutionarily irrelevant if the females often refuse matings and their choice governs the mating patterns in the species (as again appears to be the case in Danaus chrysippus; Smith, 1984). Butterflies often display female-limited polymorphism but never male-limited polymorphism. This has classically been explained in terms of female choice (Hespenheide, 1975; Turner, 1978; Silberglied, 1984). Females can refuse males, and therefore they exert a selection pressure on the males to conform to a monomorphic state. In the case of Papilio dardanus, the fluorescent yellow pigment, papiliochrome II, found in the males fades in sunlight to a non-fluorescent orange, and the wings also lose their UV reflectance (see Chapter 3 for further details). This would allow females to use an age-cue when selecting a mate. In addition, females may judge a male's fitness by the extent of his wing damage. A preference for younger males, judged by wing damage, has been found in Colias eurytheme (Rutowski, 1985). This is likely to be due to the fact that male butterflies have been found to produce a much larger spermatophore on their first than on any subsequent mating (Svńrd & Wicklund, 1986). Thus female selection for lack of wing damage could be an explanation for the tails found in many swallowtail species. These break off very easily, and thus could be an honest signal of quality - females may prefer symmetrical males with two tails over those with one tail. It is not clear whether females would thus prefer symmetrical males with no tails over males with 1 tail, and this could also be tested using choice tests. Symmetry itself may also be preferred, either as an indicator of quality (Parsons, 1990), or as the product of another perceptual bias (Johnstone, 1994; Enquist & Arak, 1994).


The aims of this chapter are:

To determine whether male Papilio dardanus have any initial preferences in their selection of a female morph as a mate (as found in their overall mate choices by Cook et al. 1994).

To investigate whether there is any correlation between the initial choice of morph by a male and his subsequent choices, and (if there is) to try to determine whether this is due to learning or a genetic effect.

To determine whether female Papilio dardanus show any preference for males with both swallowtails intact, or for symmetrical males.

The results from this chapter should suggest how the males' preference, as observed by Cook et al. (1994), might come about - whether it is a genetic effect, the result of learning, or whether it is likely to be due to a perceptual bias.

Behavioural Observations

All behavioural experiments were carried out in a flight cage measuring 3.5m x 1.5m x 2.4m in a greenhouse at 30-40o Centigrade. In the cage were placed tall plants (including a Citrus) - see Appendix 3 for further details. Without these plants, and at lower temperatures, the butterflies rarely flew. At higher temperatures, mortality in the butterflies increased.

Butterflies usually took about an hour after being placed in the flight cage before flying around, and females usually refused matings until they had acclimatised for 3 or 4 hours. Males and females often flew at each other as well as members of the opposite sex. Landed males and females were rarely approached, but where they were the approaching butterfly (whether male or female) would fly repeatedly at the landed individual from below, touching the upper surfaces of the landed butterfly's wings with its own. When females approached males (or females) in the air, they flew up at the undersides of the butterfly repeatedly, brushing the upper surfaces of their wings against the lower surfaces of those of the male. Males did the same when approaching females, and if the female accepted, the pair would land together. If accepting, the female curved her abdomen under slightly at the tip, and the males would land next to her, curving his abdomen under to mate with the female. As soon as he had clasped her, the male would drop motionless, to be carried in flight by the female if the pair should be disturbed. Often the female would land and curve her abdomen, but the male would appear to lose sight of her once she had stopped flying, and occasionally he would even land next to her, and perhaps curve his abdomen under, but would not proceed any further with copulation if he did not manage to mate with her immediately. Females would sometimes chase and land with other females, and occasionally one would land on the other as the males do, but none actually attempted copulation as males do. If the mating were successful, the pair remained in copula for about one hour before separating. Females which rejected males would either change their wingbeat pattern to a fast, fluttering flight, at which the males would fly off and leave them, or land and flap their wings vigorously to prevent approach. If pestered consistently by several males, a female would, rarely, drop to the ground and remain still. The sight of a pair in mid-air often attracted other males, and usually the males would be distracted by and pursue each other, ignoring the female.

Males appeared uninterested in females on the day that they emerged, although females would sometimes mate on their first day despite being generally less active. On the second day, both males and females were ready to mate, and on the third day both pestered each other vigorously. After about 4 days, however, the females started to lay eggs (whether they had been mated or not). When doing this they adopted a very fluttering flight, and laid up to 70 or so eggs on the underside of Citrus or Choisya leaves.

Experiment 5-1 Initial morph preference of male Papilio dardanus


Cook et al. (1994) reported morph preferences in wild males. This preference could be innate (due to a genetic or perceptual bias) or learned, so this initial experiment was simply to allow na´ve males to mate with a female of their choice having had no previous experience of females and being presented with even numbers of each morph. Matings of Papilio dardanus in suitable flight cages have been reported (Stride, 1958), so such mate choice experiments could be carried out in a suitable heated greenhouse. If a preference for hippocoonides was found in these na´ve males then it could be assumed that the preference was not learned, but was in fact an innate preference - either a genetically determined preference for a particular morph or due to a perceptual bias for the coloration in general. If no preference was found then it is possible either that the behaviour of the wild males is due to experience or that the behaviour in the flight cage is not the same as that displayed in the field due to factors such as the size of the cage or a difference in the relative conspicuousness of the morphs.


The aim of this experiment is to determine whether or not na´ve male Papilio dardanus show a morph preference when choosing mates.

Method and Materials

Since individual Papilio dardanus from race tibullus, as studied by Cook et al. (1994) could not be obtained in England in the quantities necessary for experimentation, the geographically neighbouring race, meseres, was used (see Appendix 1 for details of the races). The butterflies were sent as pupae weekly, and for the major part of the experimental work (June 1997- September 1998) 20-30 pupae were received each week from a total batch of about 400 per week from the supplier in Moshi, Tanzania. This meant that the individuals were as near unrelated as was practically possible. Unfortunately the andromorph, trimeni, which was present in the Pemba population of race tibullus, used by Cook et al., was not present in this race. Instead a mimetic morph, cenea, was relatively common, and was used as the third option in the mate choice experiments (see Appendix 1 for further details of the morphs).

Males and females were kept in separate nets when they emerged. Virgin males (up to 11 at any one time) were placed in the flight cage together with equal numbers of virgin females of each morph. The numbers of females in the net were kept as high as possible (whilst maintaining equal numbers of each morph) so as to increase the chances of a meeting between the sexes and also decrease the possibility of individual differences amongst the females affecting the results. There were generally between 1 and 3 females of each morph present. The ages of the females were kept as close as possible (and certainly there was no more than a 4 day difference between their emergence dates). Where all three morphs of female were not available, experiments were done with males having only a choice of females from two morphs.

Once a mating had occurred between two butterflies, the pair was placed in a separate net, and the numbers of different female morphs made equal again. The two mated individuals were marked on the undersides of the wings with a permanent marker so that they would not be mixed up with unmated individuals. The female morph that was chosen by each male was noted. The mated females were then placed in laying cages with sprigs of Choisya ternata (the Mexican Orange Flower), and the males kept for further experimentation (see Experiment 5-2).


Males with a choice of all three morphs:

Hippocoonides chosenCenea chosenLamborni chosen
Table 5-1 The morph choices of na´ve male Papilio dardanus in a three-way choice test

The results in Table 5-1 indicate that the males show no significant preferences amongst the morphs (p=0.504, (2 = 1.37, n=19).

Males with a choice of hippocoonides and cenea:

Hippocoonides chosenCenea chosen
Table 5-2 The morph choices of na´ve male Papilio dardanus given a choice of hippocoonides and cenea females

The males appear not to show a significant preference for either hippocoonides or cenea in a pairwise comparison (p=0.252, binomial test, n=20).

Males with a choice of cenea and lamborni:

Cenea chosenLamborni chosen
Table 5-3 The morph choices of na´ve male Papilio dardanus given a choice of lamborni and cenea females

Males with a choice of lamborni and hippocoonides:

Lamborni chosenHippocoonides chosen
Table 5-4 The morph choices of na´ve male Papilio dardanus given a choice of lamborni and hippocoonides females

The data in Table 5-3 and Table 5-4 do not show any obvious preferences, but form data sets too small for statistical analysis.


The results show that the males had no significant preference between the different morphs. This is very different from the results obtained in the field by Cook et al. (1994). There are several possible explanations for this difference:

Firstly, these results were obtained in a small flight cage, and it is possible that the behaviour shown in this cage is not representative of the behaviour of the species in the wild. This could be because there is not a significant difference between the conspicuousness of the three morphs in the relatively small, black mesh net, or that the density of butterflies is simply unrealistically high (although this latter explanation seems less likely as they did not encounter each other particularly often, and mating was relatively rare).

Secondly, it is possible that since only matings are being recorded in this experiment, and not approaches by males to females (as was recorded in Cook et al.'s experiments), then two different phenomena are being compared. As far as the dynamics of the morph populations are concerned, number of matings are all that matter, but it could still be that the males have a preference for hippocoonides which is over-ridden by the females themselves, who determine whether or not mating occurs.

Thirdly, Cook et al. were studying experienced, wild, males who may well have mated previously. So it is possible that these males have mated at random on their first encounter with a female (which is most likely to have been a hippocoonides since it is most frequent in the study population) and have learned from that mating to recognise hippocoonides females as mates, thus causing the preference for hippocoonides observed in the field experiments. Experiment 5-2 aims to investigate this hypothesis by testing whether or not male Papilio dardanus are influenced by previous experience.

These results also do not include male preferences for the andromorphic females. It is therefore not possible to test the hypothesis that these females are either particularly attractive to males (as suggested by Vane-Wright, 1984, in the "pseudosexual selection" hypothesis) or are not attractive to males (as suggested by Cook et al. 1994). Male behaviour towards other males in the flight cage (as described in the "behavioural observations" section of this chapter) suggests that they may approach male-like females more often than others, as they often followed other males rather than mimetic females. This issue is discussed further in Chapter 7.

Experiment 5-2 Second morph choice of male Papilio dardanus


As discussed in the introduction to this chapter, the males' preference for the most common morph, as demonstrated by Cook et al. (1994), could be due to a genetic or learnt preference. An influence of experience has been found in many vertebrate and invertebrate species - in some cases mating experience itself appears to increase a male's chance of mating (Cook, 1995; Pomiankowski, 1990; Clutton-Brock et al., 1989; Burley & Moran, 1979), or non-sexual interactions with other individuals (often parents or siblings) influence the future choice of mate (see Table 5-5). In others an individual's past sexual experience actually seems to affect their choice of mate (see summary in Table 5-6).

rodentssee review by D'Udine & Alleva, 1983
birdsAberle et al., 1963
Cooke & Davies, 1983
Immelmann, 1985
Vos, 1995
fishHaskins & Haskins, 1950
Ferno & Sj÷lander, 1973
Kop & Heuts, 1973
Sj÷lander & Ferno, 1973
Barlow et al. 1990
Breden et al., 1995
Table 5-5 Summary of works relating to the influence of non-sexual encounters on mate choice

Experiment 5-1 appears to indicate that the preference might be learnt, as the na´ve males seem to be choosing at random, rather than showing the preferences illustrated by Cook et al. If the preference is simply due to the hippocoonides morph being more likely to be noticed by the males, then it would be expected that the morphs would be preferred in the same proportions as in Experiment 5-1, but that there would not be a significant correlation between an individual's choice of first and second mates. However, if there is indeed a genetic or learnt preference, there should be an increased probability of males choosing the same morph on their second mating as they did on their first.

SpeciesCommon namePreferenceReference
Taeniopygia guttatazebra finchfemales increased preference for males with high display rate if exposed previously to low display rate and vice versaCollins, 1995
Ficedula hypoleucapied flycatcherfemales may return to and mate with previously visited malesDale et al., 1995
Cottus bairdimottled sculpinfemales choose male which is at least as large as previously encounteredBrown, 1981
Gasterosteus aculeatussticklebackfemales choose male which is more intensely red than previously encounteredBakker & Millinski, 1991
Table 5-6 Summary of effects of past sexual experience found on mate choice in different species


This experiment aims to determine whether or not males are more likely to choose the same morph on their second mating as they did on their first than if they were selecting at random.

Method and Materials

Experiments were carried out in the same net as described in Experiment 5-1 (see Appendix 3 for further details).

Males which had already made one recorded mating (in Experiment 5-1) were placed singly in the flight cage together with equal numbers of virgin females of each morph which was available to the male for his first choice. The numbers of females in the net were kept as high as possible (whilst maintaining equal numbers of each morph) so as to increase the chances of a meeting between the sexes and also decrease the possibility of individual differences amongst the females affecting the results (generally between 1 and 3 individuals of each morph were present). The ages of the females were kept as close as possible (and there was never more than a 4 day difference in their ages).

Once a mating had occurred between two butterflies, the pair was placed in a separate net, and the numbers of female morphs made even again. The female morph that was chosen by each individual male was noted, and the pair were marked on the underside of their wings with a permanent marker to ensure that they could be recognised. The females were then placed in laying cages with sprigs of Choisya ternata (the Mexican Orange Flower).


 Chose same morphChose different morph
Choice of 3 morphs62
Choice of 2 morphs27
Table 5-7 The morph choices of experienced male Papilio dardanus given a choice of either two or three morphs each time

Using a simple binomial probability test, the data from the choice of three morphs shows that the males choose the same morph on their second mating significantly more often that they choose either of the other two (p=0.0197, binomial test, n=8). The data from the choice of two morphs shows no significant preference for either the same morph or the other (p=0.090, binomial test, n=9), and so more data would be required to demonstrate any preference.


From the data from males who had a choice of all three morphs, it appears that males do show a pattern in their mating preferences and that they choose the same morph on their second mating as on their first significantly more often than choosing at random. This means that learning from a first mating could be the explanation for the difference seen between the field data of Cook et al. (1994) from experienced males and the mate choice data recorded in Experiment 5-1. In Chapter 4 it was shown that individual Papilio dardanus showed 90% constancy to the colour of the first flower from which they fed, indicating that they could learn from one experience. It appears that this may be a further demonstration of such a capability. The data from males who had a choice of only two morphs is less clear - it even appears to show a (not significant) bias for the morph opposite to that chosen in the first experiment. More data would be required on this experiment to determine whether or not the results for the two-way choice are indeed different from those for the three-way choice. It was unfortunately not possible to gather more data on either of these experiments due to the restraints of time and supply of butterflies. Therefore the significant results of the three-way choice were accepted, although this result may not stand when further data is gathered.

It is still possible that the consistent choice of one morph by males is genetically determined. In fact, this could even be supported by the difference between the two-way and three-way choices in this experiment as males with a choice of only two morphs may not be able to demonstrate their innate morph preference, and thus may be showing different behaviour from those which have a choice of all three morphs and are therefore always able to choose their preferred morph. This is investigated in Experiment 5-3.

Experiment 5-3 Genetics of males showing morph preferences


It is possible that the morph preferences demonstrated by Cook et al. (1994) have a genetic component, with the 'gene for preference' in linkage disequilibrium with the genes for the pattern formation (as it is likely to be, due to matings between carriers of the morph gene and those carrying the gene for the preference). This assortative mating could explain why most males preferred females carrying the most common morph gene. It could also explain the morph constancy demonstrated in Experiment 5-2 (and the possible anomaly of the two-way choice data being different from the three-way choice data in Experiment 5-2), although it is rendered less likely by the results of Experiment 5-1, where males appeared to be choosing mates at random despite the fact that the hippocoonides morph was most common, indicating that the recessive hippocoonides gene was the most common gene.

Since the genetics of the morphs are well understood (Clarke & Sheppard 1959, 1960a, 1960b, 1962 and see Appendix 1) it is possible in a lot of cases to calculate the genes carried by the males if the morph of the mother is known (for example, in all cases where the mother is hippocoonides, the bottom recessive). Thus, by allowing males to choose freely, and then noting the morphs present in their offspring it should be possible to assess the presence of any genetic effect associated with the morph genes they carry.


The aim of this experiment is to determine whether or not males carrying the genes for a particular morph also show a preference for that particular morph when choosing a mate.

Method and Materials

Females which had been chosen as mates by males given a choice of all three morphs in the free-choice experiment, Experiment 5-1, were left to lay in cages containing Choisya ternata. Any resulting offspring were kept in separate fine-mesh cages, and the female morphs which emerged were noted. This allowed the pattern genes carried by the male to be calculated in most cases.

In addition, Ford (1936) collated the results of broods laid by wild-caught females. These can give a clue to behaviour in the wild. Since all the matings were the result of free choice in the wild, this data can also be used to assess the possibility of a genetic influence on the morph choices of males.


1) Results from flight cage matings: Only five broods were successfully raised to emergence separately. The results of these broods are shown in Table 5-8.

male choicesfemale morph female gene 1female gene 2offspringmale gene 1male gene 2
2 x hippocoonideshippocoonideshh14 x lamborniTT (?)
2 x lambornilamborniT?2 x lamborni??
1 x hippocoonideshippocoonideshh5 x hippocoonides, 6 x ceneaCh
1 x hippocoonideshippocoonideshh1 x hippocoonidesh?
1 x lamborni, 1 x hippocoonideslamborniT?8 x lamborni??
Table 5-8 The results of broods raised after males had been given free choices of all three morphs. The first column shows the morphs the father of the brood chose in choice experiments, the second gives the morph of the mother of the brood. The third and fourth columns give the morph genes carried by the mother (where it is possible to determine them), the fifth column lists the female offspring of the brood, and hence the morph genes carried by the father can sometimes be determined (the last two columns)

The last two columns of Table 5-8 indicate that this experiment is not very informative. In only one case are both the male genes known for certain, although in the first brood shown in Table 5-8 both genes are probably the same given that the brood only contained lamborni females. If the father of the first brood did not carry a hippocoonides gene, then the choices cannot be strictly guided by a gene linked to those for morph as the male chose hippocoonides over lamborni twice in the choice experiments. The second and fifth broods in the table are completely uninformative as it is not possible to determine either of the genes carried by the male. The father of the third brood must have carried both a hippocoonides gene and a cenea gene, and that of the fourth brood at least one hippocoonides gene, and both males chose a hippocoonides female in a choice test. However, there is simply not enough data to analyse.

2) For the results from Ford (1936), see Appendix 4. This data is very difficult to analyse, as so many of the second genes are not known (due to small broods, or dominance of other genes). It is possible, however, to compare the proportions of the genes calculated to be present in the males who mated with each female morph with the proportions of the genes in the populations from which they came (also recorded by Ford). This is done in Appendix 4, and the two samples which were large enough to allow a statistical analysis showed that the gene frequencies of the mated males do not differ significantly from those in the overall population. One further test, which would allow an analysis of a greater proportion of the data, is to compare the ratio of homozygotes to heterozygotes in the wild population. Since the genotypes of the females are known completely in many cases, the proportion of homozygotes represented by them can be compared with that predicted by random mating. This is also done in Appendix 4. For race polytrophus, the predicted numbers of homozygotes versus heterozygotes is 21.7 versus 27.4, whilst the observed numbers are 16 and 33. Thus the observed ratio is not significantly different from that predicted by random mating (binomial test; p=0.919, n=29). For race cenea, the predicted numbers of homozygotes and heterozygotes are 7.5 and 6.5 respectively, and the observed values are 5 and 9. A binomial test on these results indicates that they do not significantly deviate from those expected if mating was at random (binomial test; p=0.152, n=14). It can therefore be said that the ratio does not indicate that males preferentially mate with females carrying the same genes (as indicated by their morph) as themselves. The slight bias towards heterozygotes in the observed data as compared with the expected values may have two contributing factors. Firstly, biases in the collecting of the females (for example, not catching as many hippocoonides females as others), would bias the ratio of homozygotes (e.g. hippocoonides) to heterozygotes. Secondly, the cases where the genotype of the female is not known (and therefore the female is left out of the analysis) are more likely to be cases where the female was homozygote (and therefore the second gene was 'masked' by the first) than when it was heterozygote (when both genes may be become obvious from the morphs apparent in the resulting brood).


From the evidence of the random mate choice in Experiment 5-1 and the calculation of the parentage of the broods reported by Ford (1936) it appears very unlikely that there is a genetic component to the males' mate choice which is associated with the genes for the patterning itself. It therefore appears, if the results of the three-way choice in Experiment 5-2 are accepted, that the males do indeed learn from their first mating and subsequently choose the same morph with a higher probability. This has implications for the population dynamics of the morphs, which is investigated in Chapter 7.

Experiment 5-4 Female preferences for symmetry and swallowtails


The presence of swallowtails on the hindwings of many of the species in the genus Papilio has not been explained. In Papilio dardanus, the mimetic females have on the whole completely lost their swallowtails in order to mimic their un-tailed model species (a few tailed mimics have been reported from Kenya (Ford, 1936), and the rare mimetic females in race antinorii from Ethiopia are tailed (Trimen, 1869)), but the maintenance of the tails in the males and the non-mimetic females of races antinorii, meriones and humbloti implies that there is some advantage to having tails. One explanation for the presence of tails would be that they provide an aerodynamic advantage, but little work has been done on butterfly flight (Chai & Srygley, 1990) and so this has yet to be studied. It is also possible, since the tails are extremely delicate and easily broken off, that they are used in signalling. The presence of strong female choice in sexual selection is often cited as a reason for the abundance of female-limited and absence of male-limited polymorphism in butterflies (Belt, 1874; Turner, 1978; Silberglied, 1984; Hespenheide, 1975). Female choice has only rarely been studied however, and it has been shown that the females appear to choose younger males, possibly on the basis of the extent of their wing damage (Rutowski, 1985).

Females have been shown to prefer symmetry when choosing mates in some species (swallows - M°ller, 1992; humans - Thornhill & Gangestad, 1993; zebra finches - Swaddle & Cuthill, 1994), and symmetry has been shown to be correlated with mating success is others (peacocks - Manning & Hartley, 1991; scorpionflies - Thornhill, 1992; earwigs - Radesńter & Halldˇrsdˇttir, 1993; damselflies - Harvey & Walsh, 1993; midges - McLachlan & Cant, 1995; field crickets - Simmons, 1995; dung flies - Allen & Simmons, 1995; cerambycid beetles - M°ller & Zamora-Mu˝oz, 1997). Some studies have shown conflicting evidence, or no symmetry preference (paradise whydahs - Oakes & Barnard, 1994; red-winged blackbirds - Dufour & Weatherhead, 1998; fruit flies - Markow & Ricker, 1992), but a review and analysis by M°ller & Thornhill (1998) concluded that there was an overall correlation between symmetry and female preference. There is much debate as to whether this represents a perceptual bias (Johnstone, 1994; Enquist & Arak, 1994) or an adaptive choice to avoid 'fluctuating asymmetry' (FA, defined as deviations from perfect symmetry which are normally distributed around a mean of zero: Van Valen, 1962), which can indicate a degree of poor adaptation to the environment (see Parsons, 1990, for a review of the evidence for this). Some insects have been shown to have a preference for symmetrical flowers when feeding (M°ller & Sorci, 1998 although Giurfa et al., 1996, did not find this in bees), which is difficult to explain as a selected trait as there is currently no reason to believe that asymmetrical flowers should provide a lesser reward for nectar feeders. In swallowtail butterflies a female preference for symmetry per se might lead to females preferring males either with no tails at all or both tails intact, whereas a preference for a young, healthy male should lead to females preferring only males with both tails present over males with either one or both missing. If females are selecting their mates on the basis of their tail condition, it should therefore be possible to distinguish between these two mechanisms.


The aim of this experiment is to determine whether females preferentially mate with males with either both swallowtails intact, or symmetrical wings (either with or without swallowtails).

Method and Materials

Because of the comparative rarity of matings obtained in the flight cage, matings obtained in Experiment 5-1 were used as the basis for this experiment. When matings occurred in Experiment 5-1, where virgin males were placed in the flight cage with virgin females and allowed to mate (see Experiment 5-1 for further details), the wing condition of all the males in the cage as well as the male with which the female mated were assessed and recorded (the wing conditions of the males were not artificially altered in any way). The males were recorded as either being '2-tailed' (where both swallowtails on the males' wings were intact), '1-tailed' (where only one swallowtail was intact) or '0-tailed' (where neither were present). In addition, some experiments were carried out in exactly the same manner when equal numbers of each female morph were not available, which were not part of Experiment 5-1. In these cases the numbers of each condition of male were kept as even as possible, although the fact that the males broke their tails during the experimental time meant that at the time of mating (when their condition was assessed) they were not necessarily evenly distributed amongst the conditions.


female numbermale chosen2-tailed males1-tailed males0-tailed males
Table 5-9 The results of females choosing males with differing numbers of swallowtails on their hindwings

The uneven numbers in each experiment, and the fact that in some cases one class had no representatives, make an analysis of the data more difficult than usual. However, it is possible to analyse it on the basis of males making a binomial choice. This was done using a short computer program written in C (see Appendix 6) which calculated the probabilities of all the possible outcomes of the experiment, and summed these for all the combinations in which each possible observable result could have been obtained (since the order in which the data points were gained is irrelevant). The probability that exactly these results were gained by chance is thus the number of possible ways in which the results could have been obtained over the number of possible results, and the p-value can be calculated as the probability that this result, or a more extreme one, could have been obtained by chance.

This method gave a p-value of 0.64 for females preferring symmetrical males (either 2-tailed, or 0-tailed) over non-symmetrical males (1-tailed), and p=0.44 for females preferring perfect, 2-tailed males over any other class of male. Thus females do not appear to have a significant preference for symmetrical or intact males.


Although the statistical method used is not powerful (and therefore it is possible that a trend in the data exists which is not detected), the high p-values indicate that there is unlikely to be any sizeable trend in the data. It therefore appears that females are not mating preferentially with males on the basis of the condition of their swallowtails, or choosing symmetrical males. It thus seems likely that the swallowtails found on the hind-wings of the males are not sexually selected ornaments, and are more likely to be used either as a defence (perhaps as a means of escaping bird attacks) or are present for their aerodynamic effect.

These results do not mean that females are not selecting males on the basis of other characteristics, however. As Rutowski (1985) found, some female butterflies preferentially mate with younger males, who are likely to be more fertile (Svńrd & Wicklund, 1986). In the case of Papilio dardanus, the fading of the males to a non-UV reflectant orange (Cook et al., 1994, and see Chapter 3) could act as a very good cue of age. The colour and patterning of Papilio dardanus differs remarkably from the closely related species, Papilio phorcas and Papilio constantinus (Clarke et al., 1991; Vane-Wright & Smith 1991), and it is possible that this change has been due to a sexual selection pressure by the females. It would be very interesting to test the hypothesis that females prefer unfaded males over faded males which appeared to be older. However, to do this it would be necessary to fade males artificially (they do not fade within their own lifespan in normal English weather), and this proved impossible. Experiment 3-7 indicated that the fading was due to wavelengths between 400-475nm, and it was not possible to find a lamp with high enough power at these wavelengths to cause fast artificial fading in live butterflies, and therefore it was not possible to test this hypothesis.


The data from Experiments 5-1, 5-2, and 5-3 suggest that male Papilio dardanus mate at random when na´ve, but subsequently have an increased probability of mating with a female of the same morph with which they already have experience. Experiment 5-2 suggested that the probability of a male choosing the same morph on his second mating was around 0.75, but as this figure is based on only 8 recorded second matings (when the male was given the choice of all three morphs) it can only be taken as a rough estimate. Since Cook et al. (1994) carried out mate choice experiments on wild males, some of these are likely to have been experienced, and to have mated previously. By comparing the results of Experiments 5-1 and 5-2 with the results of Cook et al. it might be possible to see to what extent the two sets of data are in accordance, and possibly estimate such unknown parameters as the number of times which a male mates in the wild (which will be valuable data for the construction of a mathematical model of the population in Chapter 7).

The choice tests of Cook et al. were all performed as pairwise presentations. Either the data from these individual experiments will be additive (i.e. the probabilities of males choosing A over B and B over C in two pairwise tests will predict the probability of males choosing A over C in a further pairwise test), or the different pairings will affect the preferences of the males in a more complicated way (so that the probability of males choosing A over B and B over C will not predict the probability of males choosing A over C). By analysing the data from Cook et al. it is possible to assess which of these two is the case, and it appears that the data can be taken to be additive, reflecting the overall preferences of the males (see Appendix 5 for the analysis). The preferences of the males (again see Appendix 5 for the calculations) are similar to the percentage of each female morph reported by Cook et al. from the area in which they were working (see Table 5-10).

Morph% in population% preference by males
Table 5-10 Summary of the proportions of the morphs and the males' preferences as reported by Cook et al. (1994) from Pemba

If it is assumed (from the results of Experiment 5-1) that na´ve males in the population mate at random, then it is possible to calculate from each of the percentage preferences recorded by Cook et al. the proportion of males which would be expected to have mated once already, using the probability that they choose the same morph on their second mating recorded in Experiment 5-2 (see Appendix 5 for these calculations).

Such a speculative analysis of the data from Cook et al. gives the unlikely result that around 97% of the males had already mated. This is a very high figure, and it seems very possible that the probability of males choosing the same morph twice, calculated from the results of Experiment 5-2, is inaccurate as it is based only on 8 data points. However, it is possible to calculate both this and the proportion of males who had already mated solely from the data of Cook et al. as three equations can be formed (one for each morph) and solved simultaneously.

By forming these three equations and plotting a surface of the error between the various predicted preferences for each morph (as calculated by changing both the proportion of males who had previously mated and the probability of them choosing the same morph twice in the equations) and the actual preferences calculated from Cook et al. previously it is possible to represent the error between the predicted and measured preferences as both the rate of mating twice and the probability of mating with the same morph twice vary. This is done in Appendix 5, and the minimum point on the surface (reflecting the best fit of the model to the observed values) is when both the proportion of the males who had mated twice and the probability that a male chooses the same morph on both matings is 0.84.

This high proportion for the probability of choosing the same morph twice is credible, as the value of 0.75 observed in Experiment 5-2 is based only on 8 data points and therefore has a very high margin of error. However, such a high proportion of second mating seems very unlikely to be accurate, and therefore one of several things may be occurring. Firstly, it is possible that the males on the island of Pemba do not mate at random on their first mating, as the males in Experiment 5-1 did, although it would be difficult to explain such a racial difference. Secondly it is possible that the results of Cook et al. do not simply reflect the underlying preferences of the males, and are affected by the pairwise method employed. Thirdly, the experimental design could be affected by the fact that the subjects are self-selecting. This could mean that only mated males, which have already had experience in recognising a mate, are attracted to the experimental apparatus. This would also explain why na´ve males tested in the experimental flight cage did not respond to pinned-out dead females or models.


I would like to thank Phil Taylor for help with the flight cage, and John Calvert of Stratford Butterfly Farm for supplying the butterflies.


Aberle, D.F., Bronfenbrenner, J., Hess, E.H., Miller, D.R., Schneider, D.M. & Spuhler, J.N. 1963. The incest taboo and the mating patterns of animals. American Anthropologist 65, 153-165.

Allen, G.R. & Simmons, L.W. 1996. Coercive mating, fluctuating asymmetry and male mating success in the dung fly Sepsis cynipsea. Anim. Behav. 52, 737-741.

Bakker, T.M. & Milinski, M. 1991. Sequential female choice and the previous male effect in sticklebacks. Behav. Ecol. Sociobiol. 29, 205-210.

Barlow, G.W., Francis, R.C., Baumgartner, J.V. 1990. Do the colors of parents, companions and self influence assortative mating in the polychromatic Midas Cichlid. Anim. Behav. 40, 713-722.

Belt, T. 1874. The naturalist in Nicaragua. Murray, London.

Breden, F., Novinger, D., & Schubert, A. 1995. The effect of experience on mate choice in the Trinidad guppy Poecilia reticulata. Environmental Biology of Fishes 42, 323-328.

Brown, L. 1981. Patterns of female choice in mottled sculpins (Cottidae, teleleostei). Anim. Behav. 29, 375-382.

Burley, N & Moran, N. 1979. The significance of age and reproductive experience in the mate preferences of feral pigeons, Columba livia. Anim. Behav. 27, 686-698.

Burns, J.M. 1968. Mating frequency in natural populations of skippers and butterflies as determined by spermatophore counts. Proc. Nat. Acad. Sci. 61, 852-859.

Chai, P. & Srygley, R.B. 1990. Predation and the flight, morphology and temperature of Neotropical rain-forest butterflies. Am. Nat. 135, 748-765.

Collins, S.A. 1995. The effect of recent experience on female choice in zebra finches. Anim. Behav. 49, 479-486.

Cook, D.F. 1992. The effect of male size on receptivity in female Lucilia cuprina (Diptera: Calliphoridae). J. Insect Behav. 5, 365-374.

Cook, S.E., Vernon, J.G., Bateson, M., Guilford, T. 1994. Mate choice in the polymorphic African swallowtail butterfly, Papilio dardanus: male-like females may avoid sexual harassment. Anim. Behav. 47: 389-397.

Cooke, F. & Davies, J.C. 1983. Assortative mating, mate choice and reproductive fitness in Snow Geese. In Mate Choice (ed. Bateson, P.), Cambridge University Press, Cambridge. p279-298.

Clarke, C.A., Gordon, I.J., Vane-Wright, R.I., Smith, C.R. 1991. Phylogenetic relationships of three African swallowtail butterflies, Papilio dardanus, P. phorcas and P. constantinus: new data from hybrids (Lepidoptera: Papilionidae). Systematic Entomology 16, 257-273.

Clarke, C.A. & Sheppard, P.M. 1959. The genetics of Papilio dardanus, Brown. I race cenea from South Africa. Genetics 44: 1347-1358.

Clarke, C.A. & Sheppard, P.M. 1960a. The genetics of Papilio dardanus, Brown. II races dardanus, polytrophus, meseres and tibullus. Genetics 45: 439-457.

Clarke, C.A. & Sheppard, P.M. 1960b. The genetics of Papilio dardanus, Brown. III race antinorii from Abyssinia and race meriones from Madagascar. Genetics 45: 683-698.

Clarke, C.A. & Sheppard, P.M. 1962. The genetics of Papilio dardanus, Brown. IV data on race ochracea, race flavicornis, and further information on races polytrophus and dardanus. Genetics 47: 909-920.

Clarke, C.A. & Sheppard, P.M. 1963. Interactions between major genes and polygenes in the determination of mimetic patterns of Papilio dardanus. Evolution 17, 404-413.

Clarke, C.A., Gordon, I.J., Vane-Wright, R.I., Smith, C.R. 1991. Phylogenetic relationships of three African swallowtail butterflies, Papilio dardanus, P. phorcas and P. constantinus: new data from hybrids (Lepidoptera: Papilionidae). Systematic Entomology 16, 257-273.

Clutton-Brock, T.H., Hiraiwa-Hasegawa, M., & Robertson, A. 1989. Mate choice on fallow deer leks. Nature 340, 463-465.

Cook, D.F. 1995. Influence of previus mating experience on future mating success in male Lucilia cuprina (Diptera, Calliphoridae). J. Insect Behav. 8, 207-217.

Cook, S.E., Vernon, J.G., Bateson, M., Guilford, T. 1994. Mate choice in the polymorphic African swallowtail butterfly, Papilio dardanus: male-like females may avoid sexual harassment. Anim. Behav. 47: 389-397.

Dale, S., Amundsen, T., Lifjeld, J.T. & Slagsvold, T. 1990. Mate sampling behaviours of female pied flycatchers - evidence for active mate choice. Behav. Ecol. Sociobiol. 27, 87-91.

Dawkins, M.S. & Guilford, T. 1995. An exaggerated preference for simple neural network models of signal evolution? Proc. R. Soc. Lond. B. 261, 357-360.

D'Udine, B. & Alleva, E. 1983. Early experience and sexual preferences in rodents. In Mate Choice (ed Bateson, P.) Cambridge University Press, Cambridge. p311-327.

Dufour, K.W. & Weatherhead, P.J. 1998. Reproductive consequences of bilateral asymmetry for individual male red-winged blackbirds. Behav. Ecol. 9, 232-242.

Enquist, M. & Arak, A. 1994. Symmetry, beauty, and evolution. Nature 372, 169-172.

Ferno, A. & Sj÷lander, S. 1973. Some imprinting experiments on sexual preferences for colour variants in the platyfish (Xiphophorus maculatus). Z. Tierpsychol. 33, 417-423.

Ford, E.B. 1936. The genetics of Papilio dardanus Brown (Lep.) Trans. R. Ent. Soc. London 85, 435-465.

Giurfa, M., Eichmann, B. & Menzel, R. 1996. Symmetry perception in an insect. Nature 382, 458-461.

Harvey, I.F. & Walsh, K.J. 1993. Fluctuating asymmetry and lifetime mating success are correlated in males of the damselfly Coenagrion puella (Odonata, Coenagrionidae). Ecol. Entomol. 18 (3), 198-202.

Haskins, C.P. & Haskins, E.F. 1950. Factors governing sexual selection as an isolating mechanism in the poeciliid fish Lebistes reticulatus. Proc. Nat. Acad. Sci. USA 36, 464-476.

Hespenheide, H.A. 1975. Reverse sex-limited mimicry in a beetle. Evolution 29, 780-783.

Immelmann, K. 1985. Sexual imprinting in zebra finches: mechanisms and biological significance. Proc. Int. Ornithol. Congr. 18, 156-172.

Johnstone, R.A. 1994. Female preference for symmetrical males as a by-product of selection for mate recognition. Nature 372, 172-175.

Kop, P.A.M. & Heuts, B.A. 1973. An experiment on sibling imprinting in the jewel fish (Hemichromis bimaculatus). Rev. Comp. Animal 7, 63-76.

Liley, N.R. 1966. Ethological isolating mechanisms in four sympatric species of poeciliid fishes. Behav. Suppl. 13, 1-197.

Magnus, D.B.E. 1958. Experimentelle Untersuchungen zur Bionomie und Ethologie des Kaisermantels. Z. Tierpsychol. 15, 397-426.

McLachlan, A. & Cant, M. 1995. Small males are more symmetrical - mating success in the midge Chironomus plumosus. Anim. Behav. 50, 841-881.

Manning, J.T. & Hartley, M.A. 1991. Symmetry and ornamentation are correlated in the peacock's train. Anim. Behav. 42, 1020-1021.

Markow, T.A. & Ricker, J.P. 1992. Male size, developmental stability, and mating success in natural populations of three Drosophila species. Heredity 69, 122-127.

M°ller, A.P. 1992. Female swallow preference for symmetrical male ornaments. Nature 357, 238-240.

M°ller, A.P. & Sorci, G. 1998. Insect preference for symmetrical artificial flowers. Oecologia 114, 37-42.

M°ller, A.P. & Thornhill, R. 1998. Bilateral symmetry and sexual selection: a meta-analysis. Am. Nat. 151, 174-192.

M°ller, A.P. & Zamora-Mu˝oz, C. 1997. Antennal asymmetry and sexual selection in a cerambycid beetle. Anim. Behav. 54, 1509-1515.

Oakes, E.J. & Barnard, P. 1994. Fluctuating asymmetry and mate choice in paradise whydahs - Vidua paradisaea - an experimental manipulation. Anim. Behav. 48, 937-943.

Parsons, P.A. 1990. Fluctuating asymmetry - an epigenetic measure of stress. Biol. Rev. 65, 131-145.

Pomiankowski, A. 1990. How to find the top male. Nature 347, 616-617.

Radesńter, T. & Halldˇrsdˇttir, H. 1993. Fluctuating asymmetry and forceps size in earwigs, Forficula auricularia. Anim. Behav. 45, 626-628.

Rutowski, R.L. 1985. Evidence for mate choice in a sulphur butterfly (Colias eurytheme). Z. Tierpsychol. 70, 103-114.

Simmons, L.W. 1995. Correlates of male quality in the field cricket Gryllus campestris L.: age, size, and symmetry determine pairing success in field populations. Behav. Ecol. 6, 376-81.

Silberglied, R.E. 1984. Visual communication and sexual selection among butterflies. In The biology of butterflies (ed. Vane-Wright, R.I. & Ackery, P.R.). Acadmeic Press, London. p207-223.

Sjolander, S. & Ferno, A. 1973. Sexual imprinting on another species in a cichlid fish, Haplochromis burtoni. Rev. Comport. Anim. 7, 77-81.

Smith, D.A.S. 1984. Mate selection in butterflies: competition, coyness, choice and chauvenism. In The Biology of Butterflies (ed. Vane Wright, R.I. & Ackery, P.R.) Princeton University Press, USA, 225-244.

Stride, G.O. 1958. Further studies on the courtship behaviour of African mimetic butterflies. Anim. Behav. 6, 224-230.

Svńrd, L. & Wicklund, C. 1986. Different ejaculate delivery strategies in first versus subsequent matings in the swallowtail butterfly, Papilio machaon L. Behav. Ecol. Sociobiol. 18, 325-330.

Swaddle, J.P. & Cuthill, I.C. 1994. Selection for symmetrical males by female zebra finches. Nature 367, 165-6

Swihart, C.A. 1971. Colour discrimination by the butterfly, Heliconius charitonius L.. Anim Behav. 19, 156-164.

Thornhill, R. 1992. Fluctuating asymmetry and mating success in the Japanese scorpionfly, Panorpa japonica. Anim. Behav. 44, 867-879.

Thornhill, R. & Gangestad, S.W. 1993. Human facial beauty: averageness, symmetry, and parasite resistance. Human Nature 4, 237-269.

Trimen, R. 1869. On some remarkable mimetic analogies among African butterflies. Transactions of the Linnean Society of London 26, 497-522.

Turner, J.R.G. 1978. Why male butterflies are non-mimetic: natural selection, sexual selection, group selection, modification and sieving. Biol. J. Linn. Soc. 10, 385-432.

Vane-Wright, R.I. & Smith, C.R. 1991. Phylogenetic relationships of three African swallowtail butterflies, Papilio dardanus, P. phorcas, and P. constantinus: a cladistic analysis (Lepidoptera: Papilionidae). Systematic Entomology 16: 275-291.

Van Valen, L. 1962. A study of fluctuating asymmetry. Evolution 36, 768-776.

Vos, D.R. 1995. The role of sexual imprinting for sex recognition in zebra finches - a difference between males and females. Anim. Behav. 50, 645-653.

Please cite this thesis as:
Freeman, ALJ; 1998; D.Phil thesis, Oxford University.
E-mail to Alexandra Freeman
Back to Table of Contents
Back to Home Page