Search tips
Search criteria 


Logo of nihpaAbout Author manuscriptsSubmit a manuscriptHHS Public Access; Author Manuscript; Accepted for publication in peer reviewed journal;
Anim Behav. Author manuscript; available in PMC 2017 April 1.
Published in final edited form as:
Anim Behav. 2016 April 1; 114: 203–210.
doi:  10.1016/j.anbehav.2016.01.028
PMCID: PMC4834894

Effects of sodium puddling on male mating success, courtship and flight in a swallowtail butterfly


In many Lepidoptera species usually only males puddle for sodium. Two explanations have been offered for this: (1) neuromuscular activity: males need increased sodium for flight because they are more active flyers than females; and (2) direct benefits: sodium is a type of direct benefit provided by males to females via ejaculate during mating. Surprisingly, there is little direct experimental evidence for either of these. In this study, we examined both explanations using the pipevine swallowtail butterfly, Battus philenor L. If sodium increases neuromuscular activity, males consuming sodium should be better fliers than males without sodium. If males collect sodium for nuptial gifts that benefit their mates, males consuming sodium may have greater mating success than males without sodium. In that case, females then need an honest cue/signal of the quality of male-provided direct benefits that they can assess before mating. If sodium affects male courtship flight by increasing neuromuscular activity, how a male courts could serve as such a premating cue/signal of male benefit quality. Therefore, sodium may benefit males in terms of obtaining mates by increasing their neuromuscular activity. In this study we found that males that consumed sodium courted more vigorously and had greater mating success than males that consumed water. In addition, the courtship displays of males consuming sodium were significantly different from those of males consuming water, providing a possible honest cue/signal of male benefit quality that females can assess. Interestingly, we did not find evidence that sodium consumption affects male flight outside of courtship. That only aspects of male flight related to mating were affected by sodium, while aspects of general flight were not, is consistent with the idea that sodium may benefit males in terms of obtaining mates via effects on neuromuscular activity.

Keywords: direct benefit, flight performance, Lepidoptera, mate choice, nuptial gift, puddling, sexual selection, sodium

Mud-puddling behaviour, seen in many species of Lepidoptera as well as in some other insects (reviewed in Molleman, 2010), refers to individuals feeding on soil, excrement, carrion, etc. to obtain micronutrients such as sodium and nitrogen. In many Lepidoptera species, usually only males puddle (Boggs & Jackson, 1991). Previous studies have found that sodium is most commonly sought during puddling (e.g. Arms et al., 1974; Boggs & Dau, 2004; Smedley & Eisner, 1995), although some species seek nitrogenous compounds (e.g. Beck et al., 1999; Boggs & Dau, 2004). To date, there have been two main explanations for why, in many species of Lepidoptera, usually only males puddle for sodium (reviewed in Molleman, 2010). First, it has been suggested that sodium may be a type of male-provided direct benefit to females, provided to them via male ejaculate during mating. Second, it has been suggested that sodium may be needed more by males because they are the more active flyers, and sodium may promote neuromuscular function (Arms et al., 1974). Both explanations are only weakly supported.

In choosing mates, females should be selected to mate with males that provide them with high fitness benefits (Andersson, 1994). These benefits can take different forms, and can be direct, such as material resources or parental care that increase the fitness of the female or her offspring (e.g. Price et al., 1993; South & Lewis, 2011; Vahed, 1998), or indirect, where males provide their offspring with alleles that increase viability of offspring or increase the attractiveness of male offspring (e.g. Fisher, 1930; Lande, 1981; Zahavi, 1975). For the evolution of adaptive female choice under either scenario, males must vary in benefit quality and females must reliably assess the benefit quality that males can provide prior to mating. For direct benefits to favour adaptive female choice, there needs to be a mechanism that keeps males from cheating (i.e. not providing the benefit to a female mate; Wagner 2011).

Some evidence consistent with the first explanation, hereafter referred to as the direct-benefits explanation, for why males puddle more often than females, has been reported among the Lepidoptera. Nutrients provided in ejaculates by males during mating have been found in eggs (Boggs & Gilbert, 1979): males transfer sodium to females during mating, and eggs of females mated to males that have puddled on sodium solution have higher sodium levels (Smedley & Eisner, 1996). In addition, in the few species where both sexes puddle for sodium, male spermatophores have been shown to contain little sodium (Molleman et al., 2005). Sodium may be a desired benefit by females for several reasons. First, because sodium is important to the function of the digestive, excretory and neuromuscular systems of insects (reviewed in Molleman, 2010), females may benefit by receiving it during mating. Second, adult butterflies are sodium limited (Arms et al., 1974; Smedley & Eisner, 1996). Third, a previous study found that, when fed sodium, previously mated males showed increased spermatophore size, mass of accessory gland substances and number of sperm relative to virgin males (Niihara & Watanabe, 2009).

There are two conspicuous gaps in our knowledge with regard to the direct-benefits explanation for male puddling behaviour. First, only two studies to our knowledge have examined the effects of male-provided sodium on females or their offspring, and these studies found weak and/or no effects: one study found that, under drought conditions, egg mortality was marginally lower for females mated to males that puddled on sodium solution (Pivnick & McNeil, 1987), and a second study found no effect of sodium on female egg production or fertility (Molleman et al., 2004). If sodium puddling is a male-provided direct benefit, then females mated to males that have puddled on sodium solution should be expected to gain fitness benefits. Second, while there is evidence that sodium gained by males during puddling is transferred to females during mating and may end up in eggs (Smedley & Eisner, 1996), and some evidence that sodium uptake improves mating success (Pivnick & McNeil, 1987), there is no evidence to explain how sodium uptake might make males more attractive. In this study, we set about addressing this second gap.

There is also some evidence for the second explanation, hereafter referred to as the neuromuscular activity explanation, for why usually only males puddle for sodium. Comparing 124 species in 41 genera in the family Riodinidae, Hall and Willmott (2000) found that species with relatively higher wing loading (indicative of the amount of weight that an individual can carry per unit wing area) puddled more than species with lower wing loading. However, Molleman et al. (2005) found no correlation between puddling and wing loading in a group of 98 species of African fruit-feeding butterflies from three subfamilies, Charaxinae, Nymphalianae and Satyrinae, from the family Nymphalidae. Lastly, sodium deprivation as larvae decreased adult flight speed in Helicoverpa armigera (Xiao et al., 2010), and increasing the concentration of sodium in larval diet resulted in increased flight muscle mass of male cabbage white butterflies, Pieris rapae (Snell-Rood et al., 2014). However, there have been no direct tests of the effects of adult sodium consumption on flight.

It is worth noting that the direct-benefits and neuromuscular activity explanations are not mutually exclusive. Males of many Lepidoptera species perform intricate aerial courtship displays (Rutowski et al., 2010). It is conceivable that flight performance during courtship is improved with sodium consumption by males, and may serve as an honest signal to females of the quality of the direct benefit provided by the male.

In this study, we examined both explanations for why males puddle. Specifically, we addressed the following questions in the pipevine swallowtail, Battus philenor L. (1) Do males with access to sodium have greater mating success than males that are not provided sodium? (2) Does sodium affect their courtship flight behaviour? (3) Does sodium affect male flight outside courtship? We predicted that, (a) if males with access to more sodium are providing females with better-quality direct benefits and if females prefer these males, they should have higher mating success than males without access to sodium; (b) if males advertise the quality of the direct benefits they can provide via improved courtship, then sodium-treated and water-treated males should differ in aspects of their courtship flight; and (c) if males puddle for sodium because it improves their flight, as implied by Arms et al. (1974), then sodium-treated males should fly faster, fly longer, and/or be able to generate more lift than water-treated males.


Study Species

Our study species was the pipevine swallowtail butterfly, B. philenor, a large black papilionid occurring widely through North and Central America. We conducted our experiments in southern Arizona, where this butterfly is common in most years between March and October. In this region the larval host plant is Watson's Dutchman's pipe, Aristolochia watsonii.

Individuals of this species have wing spans of 7–13 cm, and the undersides of hindwings are brightly coloured, with orange submarginal spots on iridescent blue. In addition, males have bright iridescent blue or blue-green coloration on the upper surface of the hindwings. Previous research has shown that ablation of this iridescence on males decreases their mating success (Rutowski & Rajyaguru, 2013). Males court females by flying in loops around them; they approach from behind, fly beneath the female, then fly up in front of her, and drop back behind her to restart the aerial manoeuvre (Rutowski et al., 1989; C. Mitra, personal observation). This pattern of flight probably acts to advertise the male's iridescent blue dorsal hindwing coloration to females.

Mud puddling for sodium by males is an excellent behaviour to study, especially in B. philenor, for four main reasons. First, quantity of sodium is a simple benefit to control, as we can manipulate male access to sodium while controlling for other environmental variables. Therefore, any differences between sodium-treated males and water-treated (control) males must be due to differences in sodium availability. Second, in B. philenor, as in many puddling species, sodium is transferred to females via a spermatophore during mating (Mitra, Papaj, & Davidowitz, n.d.); therefore, if females prefer males that have consumed sodium, there must be a cue/signal that females assess before mating that is correlated with sodium consumption status. Third, previous work in this species has shown that there is substantial variation in the size of the spermatophores that male B. philenor transfer to females (mean ± SE = 6.5 ± 2.02 mg, N = 75, as reported in Rajyaguru et al., 2013), suggesting that direct benefit quality may vary considerably among males. Lastly, most B. philenor females only mate once or twice in their lifetime (Burns, 1968; Rutowski et al., 1989), suggesting that if male sodium increases the fitness of his offspring, natural selection on males may decrease the likelihood that deception may evolve in males of this species.

General Methods

Individuals used in experiments were wild B. philenor collected in our Aristolochia fimbriata garden plots in Tucson, Arizona. We grew A. fimbriata (as they grow larger and faster than the native A. watsonii) in plots at the University of Arizona main campus, and the University of Arizona Campus Agricultural Center located 5 km north of the main campus. Wild B. philenor females found and laid eggs on the plants at both plots. We allowed these eggs to hatch and the larvae to feed freely on the plants until late fifth instar. Larvae were then brought to the laboratory and placed individually into 0.47-litre plastic cups, each of which contained sprigs of A. fimbriata placed in a water pick. Larvae were maintained at approximately 23 °C and 50% RH, in a 12:12 h light:dark cycle. The tops of these cups were covered with fibreglass window screen held in place by rubber bands. The larvae were provided with food ad libitum until they pupated inside the cups. After the pupae completed sclerotization (2 days post pupation), we removed them, and attached them with tape onto plastic medium Pla-House habitat lids (0.28 × 0.18 m, KollerCraft, Fenton, MO, U.S.A.), which were placed inside a 0.61 × 0.61 × 0.91 m nylon-mesh emergence cage (Bioquip, Rancho Dominguez, CA, U.S.A., catalogue no. 1466CV).

Cages were monitored daily for emerging adults. Newly emerged animals were removed from the emergence cage, separated by sex and placed into 0.30 × 0.30 × 0.30 m nylon-mesh cages (Bioquip 1466AV). One day after emergence (adult age 1), the date of emergence was written on the ventral side of one of their hindwings with a Sharpie extra-fine point metallic gold paint marker. All adult butterflies were hand-fed ad libitum with 20% volume-to-mass sugar solution (made with deionized water and commercially available table sugar) once per day. Adult butterflies were maintained at approximately 23 °C and 50% RH, in an LD 12:12 h cycle.

All males used in tests were mated prior to treatment in order to deplete their sodium stores gained from larval feeding. For these initial matings, we placed males and females into a 1.83 × 1.83 × 1.83 m nylon-screen test cage (Bioquip 1406B) that was erected indoors, lit by four 500 W halogen work-lights, suspended 0.3 m above the top of the cage. With the lights on, temperatures within the cage varied from 26–29 °C near the top, to 24–25 °C near the floor. We allowed 30 min for mating to occur, our experience being that if butterflies were going to mate on a given day, it happened within 30 min (C. Mitra, personal observation). Butterflies were allowed up to 3 days to mate and ca. 90% mated within that period. Mating pairs were carefully removed as soon as they formed, and placed in a 0.61 × 0.61 × 0.91 m nylon-mesh cage (Bioquip 1466CV) to minimize harassment from unmated individuals. The day after mating, males and females were separated into sex-specific cages, and males were marked with a test-specific individual ID using a Sharpie as described above. Females in these initial matings were 2–6 days old.

Test males were not fed for at least 12 h prior to receiving a sodium or water treatment on the day after their initial mating. We randomly assigned males to either the sodium treatment or control group. To treat males, we first placed them in a glassine envelope (Bioquip 1130B) and determined their initial mass to 0.0001 g (Sartorius Analytic Balance A210P). Then we fed them either ad libitum distilled water (control) or ad libitum 0.1 M NaCl (treatment), by lowering their proboscis into the solution and allowing them to drink. To assess how much saline solution or distilled water the males consumed, we reweighed the males after treatment. They were then allowed to drink ad libitum 20% volume-to-mass sugar solution, which they did readily. For a subset of males, we reweighed the males after they drank sugar solution to ensure that there was no difference in the amount of sugar water consumed post treatment between water-treated and sodium-treated males. We found no difference in the amount of sugar solution consumed (water-treated: mean ± SE = 0.06 ± 0.01 g, N = 12; sodium-treated: mean ± SE = 0.05 ± 0.01 g, N = 12; P = 0.514), and there was no significant correlation between the amount of water or sodium solution that an individual consumed and the amount of sugar solution it consumed (Pearson correlation: r22 = -0.09, P = 0.679). In all experiments, males were tested for treatment effects (effect of sodium solution consumption versus water) the day after being treated. As the animals we used in these experiments were wild caught in Tucson, and as we minimally manipulated them during trials, most were re-released immediately after trials. Any that were injured during trials were euthanized by placing them in an envelope and putting them into a freezer for at least 24 h.

We controlled for male age and body mass in all the analyses below, but dropped them as covariates when their effects were not significant. Unless stated otherwise, there was no statistically significant difference in male age or body mass between males of the two treatment groups.


At the start of a trial, one sodium-treated male and one water-treated male were placed with an unmated female 2–5 days old in the test cage (Bioquip 1406B). We recorded (1) the time the trial was started, (2) the time that mating occurred and (3) which male mated. If the female did not mate with one of the males within 30 min, she was replaced with a new female. After all pairs of males had been tested for the day, the unmated males from each pair were placed back in the test cage with unmated females and allowed 30 min to mate. A randomly selected subset of mated pairs were observed to note when they separated as a measure of mating duration. Test males were 4–7 days old on the day they were tested.

We used a logistic regression to assess the difference in the number of sodium-treated and water-treated males that mated when in competition for one unmated female, with age and body mass as covariates. Likewise, we used a logistic regression to assess the difference between the total number of sodium-treated and water-treated males that mated when provided unmated females, with age and body mass as covariates. Logistic regressions were used instead of chi-square tests so that we could examine effects of covariates. The effect of male treatment (sodium versus water) on mating duration (the time that the mated pair remained attached) was also analysed with an ANCOVA with age and body mass as covariates.

Courtship duration and vigour

In the courtship trials, one treated male (sodium or water) was placed in the test cage next to two unmated females 2–5 days old in the test cage. We recorded (1) the time the trial started, (2) the time that mating occurred and (3) the amount of time the male spent flying (between the start of the trial and when mating occurred) as a rough measure of courtship effort. We note that this is a rough measure of courtship effort because our measure also included the time the male spent flying in general. However, since we observed test males to be interacting with the females for the majority of the time they were in flight before mating, we are confident that this was a good, if rough, measure of courting effort. Test males were 4–6 days old on the day they were tested.

We defined courtship duration as the total time that a male spent flying before mating, and courtship vigour as the proportion of time that a male spent flying before mating (courtship duration divided by mating latency). Courtship duration and courtship vigour as measured here are likely dependent on both variation in male effort and female receptivity. However, because we were only interested in differences between treatments, we considered our inability to separate effects of male effort from female receptivity less important. Courtship duration data were log transformed for normality, and the effect of male treatment (sodium versus water) was analysed with an ANCOVA, with age and body mass as covariates. As courtship vigour data were not normally distributed, the effect of male treatment (sodium versus water) on courtship vigour was tested using the nonparametric Kruskal–Wallis test. Use of this test did not allow us to test effects of covariates.

Courtship flight displays

To examine variation in courtship flight, we utilized a large nylon-mesh cage (8 × 8 × 3 m) placed in full sun at the University of Arizona Campus Agricultural Center. Trials were started by feeding 20% volume-to-mass sugar solution to one test male and one female on a tray placed at the centre of the cage. After feeding, the butterflies would fly up and around the cage until they were within 5 cm of each other. At this point we started the stopwatch. A hand-held Sony Handycam HDR-CX150 recorded (at 60 frames/s) subsequent male flight until mating, or until 10 min had elapsed. If the pair had not mated in the 10 min, the test female was removed, a new test female was introduced, and the trial restarted. Each male was provided a maximum of two females, and if they did not mate, they were dropped from the experiment. We tested a total of 63 males, of which five never mated (four sodium-treated males and one water-treated male) and were dropped from the trial, and of which three mated with the second female (one sodium-treated and two water-treated males). A total of 58 successful mating trials (27 sodium-treated and 31 water-treated males) were obtained. Because the males that did not mate also did not actively court females, measuring the same courtship behavior on them as on the successful males was not possible. Experiments were run between 0730 and 1000 hours, between 5 September and 3 October 2014. We noted ambient temperature and relative humidity at the start of each trial.

The videos for these trials were analysed for the time it took a male to complete an aerial courtship manoeuvre (described in Rutowski et al., 1989; Rutowski et al., 2010), where a male displays the blue dorsal iridescent area on his hindwings to a female by flying behind her, swooping below her, and flying up in front of her. We measured the time it took each tested male to complete three such circles around the female, and analysed the repeatability of circling time using the intraclass correlation coefficient (Hayes & Jenkins, 1997). Circling time was highly repeatable within individuals (ICC = 0.40, N = 58, k = 3), with 40% of the variation among males being attributed to individual differences (see Bell et al., 2009 for meta-analysis of repeatability scores for behavioural measures). As the butterflies flew freely around the cage (1) the distance between the video camera and courting pair and (2) the angle the pair was at from the video camera were in constant flux. This made it impossible to measure accurately the size of the swoop from recordings.

Mean circling time for each male was square-root transformed for normality, and the effect of male treatment (sodium versus water) on circling time was analysed with an ANCOVA, controlling for ambient temperature and relative humidity, as well as male age and body mass.

Flight distance and speed

To measure flight distance and speed, we used a flight mill with the following specifications. The mill had a horizontal arm made of 3.97 mm diameter aluminium tubing (K-S Engineering, Chicago, IL, U.S.A.), with a centred pivot rod of 2.37 mm (McMaster-Carr, Elmhurst, IL, U.S.A., catalogue no. 3009A256). The radius of the arm was 31.85 cm, making distance travelled during each revolution exactly 2 m. The central pivot was held by two mini, high-precision, double-shielded, flanged stainless-steel ball bearings (2.38 mm, McMaster-Carr, 57155K28) at both ends of a 2.5 mm brass tube. The ends of this brass tube were cut on a lathe to ensure that both ball bearings were exactly aligned, and the tube was held vertical with a chemistry stand and a three-pronged clamp. An infrared transmitter-receiver (Monarch Instrument, Amherst, NH, U.S.A., catalogue no. DC1250-U01) mounted under the mill arm registered every half rotation of the mill (= 1 m), and flight data (number of rotations and time) were logged into a CF card and later exported to MS Excel. The mill was placed on the floor of the 1.83 × 1.83 × 1.83 m nylon-screen test cage (Bioquip 1406B) described above, during trials.

Test males were clipped to the flight mill via a tether that was attached to the dorsal side of their thorax using a small drop of Loctite Ultra Gel Control Super Glue™. These males were not fed for at least 12 h prior to receiving their sodium or water treatment (as described above). We attached the tethers to the males immediately after treatments were completed, and the glue was allowed to set for at least 12 h before trials. The tethers were made of straight galvanized steel wire with flattened loops at both ends (24 gauge, 50 mm in length). We used rigid tethers that did not require the test animals to generate lift during flight because the added mass of a flexible tether capable of restricting the flight of the test butterflies to a single direction (around the pivot) was too great for them to maintain normal horizontal flight. As we were interested in relative differences in flight capability between male treatments, as opposed to absolute measures of the flight distance/speed the butterflies could achieve, we deemed hard tethers an acceptable alternative. Previous studies have found that tethered animals generate less mechanical power than those in free flight (Riley et al., 1997; Taylor et al., 2010); however, unless there is an interaction between experimental treatment (here sodium versus water) and the effects of tethering on flight, which seems unlikely, then any relative differences in flight measures between treatments should reflect real effects of the treatments.

Immediately prior to testing, the test male was allowed to drink ad libitum 20% volume-to-mass sugar solution. His harness was then clipped to the flight mill via a 30 mm alligator clip attached to the end of the horizontal mill arm and he was allowed to fly undisturbed. The trial ended when the male either came to a stop for 10 s, or when 2 h had elapsed. Only 6 of 95 males tested flew for the entire 2 h. Test males were 4–6 days old on the day they were tested. The tethers were carefully removed from males after trials were completed, and most males were able to fly normally afterwards.

Flight data were collected on all tested males. After an initial analysis of the data we found that some males would not fly on the tether. We therefore set a minimum flight distance of 25 m for an individual to be included in the analysis. Ninety of the 95 tested males flew 25+ m. From the logged data, we calculated several flight measures including total distance flown, average speed and maximum speed for each tested male.

Distance data were square-root transformed for normality, and the effect of male treatment (sodium versus water) on flight distance was analysed with ANCOVA, with age and body mass as covariates. Maximum speed was calculated by dividing the entire flight into 10 s periods, measuring speed over these periods, and recording the greatest speed that a male attained. The effect of male treatment (sodium versus water) on maximum speed was analysed with an ANCOVA, with age and body mass as covariates.

Average speed was measured in two ways: (1) over the entire flight distance and (2) over 500 m, between distances of 100 m and 600 m, in order to attain an average speed over a standardized distance for all test males. This methodology also allowed for males requiring time (first 100 m) to acclimate on the flight mill after being handled. For this measure, any male that had not flown at least 600 m was removed from the analysis. The effect of male treatment (sodium versus water) on our two measures of average speed were analysed with ANCOVA, with age and body mass as covariates.

Lift generated during flight

As the tethered flight mill design described above did not allow us to measure the effect of sodium on the ability of males to generate lift, we tested whether sodium affected lift generated during flight by using methods modified from Dillon and Dudley (2004). We created a series of bead strings, with small glass beads and fine sewing thread, ranging from a string with one bead to one with seven beads. The beads were individually attached to a wire earring base (Jo-Ann Stores, Hudson, OH, U.S.A., item no. 2736866).

Before testing, males had a small piece of fine sewing thread tied in a loop around their body between their thorax and abdomen. This thread was in turn tied to a small, metal, open jump ring (4 mm, Jo-Ann Stores, item no. 11372216) approximately 50 mm from the male's body (Fig. 1a) to which the wire earring base of the bead strings could be easily attached during tests. Immediately prior to testing, the test male was allowed to drink ad libitum 20% volume-to-mass sugar solution. The string with three beads was then attached to the jump ring on the butterfly, using the wire clasp. The butterfly was then released approximately 1 m above the ground inside the 1.83 × 1.83 × 1.83 m nylon-screen test cage (described above) and we observed the male's flight across the cage. In order to be counted as a success the male had to flap his wings and gain height during flight. If the male succeeded, the string with three beads was removed, the one with four beads was attached, and the male was retested. If the male failed, the string with three beads was removed, and the one with two beads was attached, and the male was retested. As preliminary data suggested that the mass of three beads was the average mass that butterflies could carry, we started trials with a string of three beads and added/subtracted beads to assess how much mass the male could successfully lift. All males tested were able to fly successfully with the string with one bead, whereas no males tested were able to fly successfully with more than six beads.

Figure 1
(a) A test male in the lift trial, with a small piece of fine sewing thread tied in a loop around its body, between the thorax and the abdomen, which was then tied to a small, metal, open jump ring. The string with one bead is shown here attached to the ...

Each progressive string had an additional bead that contributed 0.075 g to the final mass (Pearson correlation: r4 = 1.00, P < 0.001; Fig. 1b). As the number of beads on the string was perfectly correlated with total string mass (mass of beads + string + clasp), we used mass of the heaviest string the male carried successfully plus his body mass as our measure of the maximum lift that a male could generate. After trials were complete, we carefully cut the thread tied around the male's thorax and abdomen with thin-tipped scissors. Most males were able to fly normally afterwards.

The effect of male treatment (sodium versus water) on lift was analysed with an ANCOVA, with male wing size and age as covariates. Wing size was measured as the length of the vein on the anterior edge of the discal cell of one forewing of each male, to the nearest 0.1 mm. The length of this vein correlates well with wing area (measured by taking a digital photo of the wing lying flat next to a ruler, and using ImageJ software,; Pearson correlation: r8 = 0.82, N = 10, P = 0.003), and wing area is a good measure of the relative size of butterflies (Hall & Willmott, 2000). Despite males being assigned to treatments randomly, wing size of sodium-treated males was found to be larger than that of water-treated males (sodium-treated = 2.66 ± 0.03 cm; water = 2.56 ± 0.03 cm; P = 0.015).



All males were mated once prior to being treated and tested. During tests, when two males, one sodium-treated and the other water-treated, were placed with one unmated female, there was no difference in which type of male mated first (logistic regression: χ21 = 2.06, N = 79 pairs, P = 0.152; Table 1). However, looking at the total number of sodium-treated and water-treated males that mated successfully when provided with more unmated females, sodium-treated males were significantly more likely to mate than water-treated males (logistic regression: χ21 = 7.81, N = 85 sodium-treated, 89 water-treated, P = 0.005; Table 1). There was no effect of sodium treatment on mating duration (ANOVA: F1,57 = 0.324, N = 33 sodium-treated, 26 water-treated, P = 0.572; Table 1).

Table 1
Mating trials

Courtship Duration and Vigour

There was a tendency for salt-treated males to spend more time courting females before mating than did water-treated males, but the difference was not statistically significant (ANOVA: F1,45 = 3.238, N = 23 sodium-treated, 24 water-treated, P = 0.079; Fig. 2a). In terms of courtship vigour (the proportion of time males spent courting), sodium-treated males courted females with more vigour than did water-treated males (Kruskal–Wallis test: H1 = 5.381, N = 23 sodium-treated, 24 water-treated, P = 0.020; Fig 2b).

Figure 2
(a) Differences in the time that males of each treatment group spent courting before they mated. Cell means and SEs of the log-transformed data are shown; N = 23 sodium-treated, 24 water-treated. (b) Box plot of differences in the proportion of time that ...

Courtship Flight Displays

Sodium-treated males took longer to complete courtship circles around females than did water-treated males (ANCOVA: F1,55 = 4.716, N = 27 sodium-treated, 31 water-treated, P = 0.034; Fig. 3). Ambient temperature was a significant covariate, and circling time decreased as temperature increased (F1,55 = 10.265, P = 0.002). However, there was no difference in ambient temperature (sodium-treated = 30.4 ± 0.6 oC; water-treated = 29.7 ± 0.5 oC; F1,56 = 0.710, P = 0.403) between the trials of sodium-treated and water-treated males.

Figure 3
Differences in mean circling time during courtship for males in each treatment group. Cell means and SEs of the square-root transformed data are shown; N = 27 sodium-treated, 31 water-treated. Different letters above bars designate statistically significant ...

Flight Distance and Speed

There was no effect of sodium treatment on flight distance (ANOVA: F1,88 = 0.458, N = 43 sodium-treated, 47 water-treated, P = 0.500; Table 2), maximum speed attained (F1,88 = 0.763, P = 0.385; Table 2), or average speed over the entire time spent flying (F1,88 = 1.672, P = 0.199; Table 2). Likewise, standardized average flight speed (100 to 600 m) was not different between the two treatments (F1,60 = 0.502, N = 29 sodium-treated, 33 water-treated, P = 0.481; Table 2).

Table 2
Flight trials

Lift Generated During Flight

There was no effect of sodium treatment on how much mass a male could lift during flight (ANCOVA: F1,97 = 0.585, N = 48 sodium-treated, 52 water-treated, P = 0.446; Table 2). Male wing size was a significant covariate (F1,97 = 15.475, P < 0.001), with larger males (i.e. those with larger wings) being able to lift more mass.


Increased male-provided direct benefits to mates, and increased neuromuscular activity in males, are the two main explanations put forth in the literature for why usually only males puddle for sodium in many species of Lepidoptera. Here we examined these two explanations in the context of flight, comparing courtship flight to flight outside courtship.

The direct-benefits explanation suggests that by collecting sodium males have stores available that they can transfer to females as part of a nuptial gift (Pivnick & McNeil, 1987; Smedley & Eisner, 1996; Watanabe & Kamikubo, 2005). Under this hypothesis, sodium would constitute a type of direct benefit, increasing the fitness of the males' mates and/or offspring. Only one study has examined effects of sodium consumption on male mating success (Pivnick & McNeil, 1987). It found no difference in the mean total number of matings by sodium-treated and control males, or in the proportions of sodium-treated versus control males that remated post treatment. However, they did find evidence that more sodium-treated males remated the day after their first mating and treatment than did control males.

In our study, we found that whereas sodium consumption did not affect mating success of males when in direct competition with a water-treated male, a larger proportion of sodium-treated males mated when placed with virgin females than did water-treated males. In this regard, sodium consumption appears to be increasing male mating success in this species. Their improved success may be due to their ability to coerce females to mate, if mating success depends mostly on males. Alternatively, this increase in male mating success may be due to differences in female preferences for males in different treatment groups, if mating success depends mostly on females. In this system, the complexity of courtship and mating behaviour of both males and females that occur in flight makes it difficult to distinguish between these two possibilities. However, in light of both the complexity of the male courtship behaviour and the extent of male ornamentation, it seems likely that female preferences are important (Andersson, 1994). In addition, vigorously courting males are sometimes unable to get matings with virgin females that later mate successfully with other males (C. Mitra, personal observation), suggesting that mating probability is not dependent solely on male coercion. Therefore, male courtship flight is likely important for female choice and male mating success.

The nonmutually exclusive neuromuscular activity explanation suggests that males collect sodium because they are more active flyers than females and need the extra sodium for higher neuromuscular performance (Arms et al., 1974). To our knowledge, our study is the first to directly examine the effect of sodium consumption by adult males on flight performance. Outside of the context of mating, we found no effects of sodium on flight speed, duration, distance flown or lift generated during flight. The only context in which sodium appears to affect flight is during courtship. We found that the sodium-treated males spent a significantly larger proportion of their time in flight before mating, and that they took longer to complete a courtship aerial manoeuvre. An aerial swoop of longer duration may allow the male to display his iridescent blue hindwings to the female longer. Because we found that sodium-treated males had higher mating success than water-treated males, we suggest that the increased courting vigour and circling time may make sodium-treated males more attractive to females. However, future female-choice experiments are needed to explicitly test whether increased courtship vigour and circling time increase male mating success.

In short, while sodium puddling does affect flight performance, it only appears to do so in the context of courtship. This result brings together the two explanations, neuromuscular activity and direct benefits, for why males puddle for sodium more often than females. We suggest that sodium consumption increases male neuromuscular activity, allowing them to court females more vigorously. The quality of male courtship flight then serves as a signal of male ability to provide sodium as a nuptial gift to potential mates. If males that have low body sodium are unable to produce the same signal quality due to lower neuromuscular activity, this may be an honest signal of male quality.

The next two questions we need to ask are whether males in this species transfer sodium to females during mating, and if so, how that affects the fitness of females and their offspring. In this study, we found that sodium consumption affects both male courtship flight and male mating success. If sodium consumption makes male courtship more successful, and therefore increases male mating success, then males may be selected to transfer relatively little sodium to females during mating in order to increase chances of future matings. Alternatively, males may be selected to spend additional time and energy foraging for sodium between matings. Indeed, Niihara and Watanabe (2009) found that in Papilio xuthus, virgin males produced larger spermatophores, with both more accessory products and more sperm, than did males mating a second time. But, feeding on saline solution between the two matings restored ejaculate size. However, as most females in this species mate only once or twice in their lifetimes, then if male sodium is indeed a direct benefit that increases a male's fitness by increasing the number and/or quality of the offspring his mate produces, males should be selected to provide sodium as part of a nuptial gift.

  • We examined two explanations for why mostly male Lepidoptera puddle for salt.
  • We asked if salt improves flight performance of males.
  • We found that salt affects flight, but only in the context of courtship.
  • We found that salt increases male mating success.


This research was primarily supported by the National Institutes of Health (NIH K12 GM000708) through the Postdoctoral Excellence in Research and Teaching (PERT) Program of the University of Arizona's Center for Insect Science (CIS). Additional funding was provided by grants from the National Science Foundation (IOS-0921280 to D.P., and IOS-1053318 to G.D.). We thank the editors and anonymous referees who helped improve our study. We also thank J. Bronstein, R. Rutowski, W. E. Wagner Jr. and the Davidowitz and Papaj labs for feedback along the way. Lastly, we thank the undergraduate research assistants from the University of Arizona and Pima Community College, who helped care for our larvae and butterflies, and run trials.


Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.


  • Andersson M. Sexual selection. Princeton, NJ: Princeton University Press; 1994.
  • Arms K, Feeny P, Lederhouse RC. Sodium: stimulus for puddling behavior by tiger swallowtail butterflies, Papilio glaucus. Science. 1974;185:372–374. [PubMed]
  • Beck J, Muhlenberg E, Fiedler K. Mud-puddling behavior in tropical butterflies: in search of proteins or minerals? Oecologia. 1999;119:140–148.
  • Bell AM, Hankison SJ, Laskowski KL. The repeatability of behaviour: a meta-analysis. Animal Behaviour. 2009;77:771–783. [PMC free article] [PubMed]
  • Boggs CL, Dau B. Resource specialization in puddling Lepidoptera. Environmental Entomology. 2004;33:1020–1024.
  • Boggs CL, Gilbert LE. Male contribution to egg-production in butterflies: evidence for transfer of nutrients at mating. Science. 1979;206:83–84. [PubMed]
  • Boggs CL, Jackson LA. Mud puddling by butterflies is not a simple matter. Ecological Entomology. 1991;16:123–127.
  • Burns JM. Mating frequency in natural population of skippers and butterflies as determined by spermatophore counts. Proceedings of the National Academy of Sciences of the United States of America. 1968;61:852–859. [PubMed]
  • Dillon ME, Dudley R. Allometry of maximum vertical force production during hovering flight of Neotropical orchid bees (Apidae: Euglossini) Journal of Experimental Biology. 2004;207:417–425. [PubMed]
  • Fisher RA. The genetical theory of natural selection. Oxford, U.K: Clarendon Press; 1930.
  • Hall JPW, Willmott KR. Patterns of feeding behaviour in adult male riodinid butterflies and their relationship to morphology and ecology. Biological Journal of the Linnean Society. 2000;69:1–23.
  • Hayes JP, Jenkins SH. Individual variation in mammals. Journal of Mammalogy. 1997;78:274–293.
  • Lande R. Models of speciation by sexual selection on polygenic traits. Proceedings of the National Academy of Sciences of the United States of America. 1981;78:3721–3725. [PubMed]
  • Mitra C, Papaj D, Davidowitz G. Passing the salt: sodium puddling in Battus philenor. (n.d.) Manuscript in preparation.
  • Molleman F. Puddling: from natural history to understanding how it affects fitness. Entomologia Experimentalis et Applicata. 2010;134:107–113.
  • Molleman F, Grunsven RHA, Liefting M, Zwaan BJ, Brakefield PM. Is male puddling behaviour of tropical butterflies targeted at sodium for nuptial gifts or activity? Biological Journal of the Linnean Society. 2005;86:345–361.
  • Molleman F, Zwaan BJ, Brakefield PM. The effect of male sodium diet and mating history on female reproduction in the puddling squinting bush brown Bicyclus anynana (Lepidoptera) Behavioral Ecology and Sociobiology. 2004;56:404–411.
  • Niihara S, Watanabe M. Effects of saline intake until the second mating on the ejaculates and sperm number of male swallowtail butterfly, Papilio xuthus Linnaeus. Japanese Journal of Applied Entomology and Zoology. 2009;53:91–96.
  • Pivnick KA, McNeil JN. Puddling in butterflies: sodium affects reproductive success in Thymelicus lineola. Physiological Entomology. 1987;12:461–472.
  • Price T, Schluter D, Heckman NE. Sexual selection when the female directly benefits. Biological Journal of the Linnean Society. 1993;48:187–211.
  • Rajyaguru PK, Pegram KV, Kingston ACN, Rutowski RL. Male wing color properties predict the size of nuptial gifts given during mating in the pipevine swallowtail butterfly (Battus philenor) Naturwissenschaften. 2013;100:507–513. [PubMed]
  • Riley JR, Downham MCA, Cooter RJ. Comparison of the performance of Cicadulina leafhoppers on flight mills with that to be expected in free flight. Entomologia Experimentalis et Applicata. 1997;83:317–322.
  • Rutowski RL, Alcock J, Carey M. Hilltopping in the pipevine swallowtail butterfly (Battus philenor) Ethology. 1989;82:244–254.
  • Rutowski RL, Nahm AC, Macedonia JM. Iridescent hindwing patches in the pipevine swallowtail: differences in dorsal and ventral surfaces relate to signal function and context. Functional Ecology. 2010;24:767–775.
  • Rutowski RL, Rajyaguru PK. Male-specific iridescent coloration in the pipevine swallowtail (Battus philenor) is used in mate choice by females but not sexual discrimination by males. Journal of Insect Behavior. 2013;26:200–211.
  • Smedley SR, Eisner T. Sodium uptake by puddling in a moth. Science. 1995;270:1816–1818. [PubMed]
  • Smedley SR, Eisner T. Sodium: a male moth's gift to its offspring. Proceedings of the National Academy of Sciences of the United States of America. 1996;93:809–813. [PubMed]
  • Snell-Rood EC, Espeset A, Boser CJ, White WA, Smykalski R. Anthropogenic changes in sodium affect neural and muscle development in butterflies. Proceedings of the National Academy of Sciences of the United States of America. 2014;111:10221–10226. [PubMed]
  • South A, Lewis SM. The influence of male ejaculate quantity on female fitness: a meta-analysis. Biological Reviews of the Cambridge Philosophical Society. 2011;86:299–309. [PubMed]
  • Taylor RAJ, Bauer LS, Poland TM, Windell KN. Flight performance of Agrilus planipennis (Coleoptera: Buprestidae) on a flight mill and in free flight. Journal of Insect Behavior. 2010;23:128–148.
  • Vahed K. The function of nuptial feeding in insects: review of empirical studies. Biological Reviews of the Cambridge Philosophical Society. 1998;73:43–78.
  • Wagner WE. Direct benefits and the evolution of female mating preferences: conceptual problems, potential solutions, and a field cricket. Advances in the Study of Behavior. 2011;43:273–319.
  • Watanabe M, Kamikubo M. Effects of saline intake on spermatophore and sperm ejaculation in the male swallowtail butterfly Papilio xuthus (Lepidoptera: Papilionidae) Entomological Science. 2005;8:161–166.
  • Xiao K, Shen K, Zhong JF, Li GQ. Effects of dietary sodium on performance, flight and compensation strategies in the cotton bollworm, Helicoverpa armigera (Hubner) (Lepidoptera: Noctuidae) Frontiers in Zoology. 2010;7:11. [PMC free article] [PubMed]
  • Zahavi A. Mate selection: selection for a handicap. Journal of Theoretical Biology. 1975;53:205–214. [PubMed]