Fierce Little Creatures

Butterfly-ant mutualism


Text Steve Woodhall Photographs Steve Woodhall and others

Predatory butterflies…​​?!!

Symbiotic and parasitic relationships abound in the world of insects, and butterflies are no exception. Many of their fellow arthropods prey on them or parasitise them, as we saw in previous LE articles. Butterflies (and moths) have a gentle, herbivorous image. People are usually shocked to learn that some of them are out-and-out predators on other insects. This usually occurs in the larval stage, although many butterflies feed on the bodily secretions of aphids, scale insects etc.

The best-known inter-order relationships that butterflies display is with ants. There is a spectrum running all the way from no observable relationship, to the ants being deterred by chemical defences to them possessed by the larvae, to ants protecting the larvae from predators, to the larvae living inside the ants’ nests and preying on them, or even being fed by them, rather like baby cuckoos! So who discovered this?

Where it all started

The first indications that scientists had, that something was going on between butterflies and ants, was in England, in the 19th Century. Butterfly collecting was all the rage at the time – surpassing even the modern-day birding passion. Collectors had realised that perfect cabinet specimens could be obtained by rearing larvae to adulthood – or by collecting pupae in the wild.

Before the end of the century all British butterflies had been reared – except the Large Blue (Lycaena – now Phengarisarion). Larvae could be found up to the third instar – roughly half-grown – on the host plant, Wild Thyme. After that they disappeared. F W Frohawk, the greatest authority at the time, had set out to record the life histories of all British butterflies, but was thwarted by the Large Blue.

It wasn’t until 1906 that he discovered that the fully grown larvae could be found inside the nests of certain ant species. And it took until 1915 for it to be discovered that they were actually feeding on the ant brood. The first carnivorous butterfly had been discovered. It wasn’t until 1953 that EB Purefoy published the life history, and not until the 60’s was it found just how complicated the situation was. It wasn’t certain ant species, it was only one; and the butterfly-ant relationship depended on other factors like herbivore activity, land use, and climatic factors. All of which had changed since the butterfly had been discovered. By 1979 it was extinct.

Large Blue, Phengaris arion near Cirencester in Gloucestershire. England – not African, but the first.

Fortunately, in the meantime enough had been learned about the Large Blue to allow, in 1985, the Scandinavian subspecies to be reintroduced to south-west England. This has been a remarkable success… it has now spread across most of its old range, and I was able to visit one of its haunts in 2016. It was flying in good numbers!

At third instar Large Blue larvae stop feeding on their host plant Wild Thyme, ‘hunch themselves up’ and emit chemicals that fool the ants into thinking they are actually their larvae. This image shows one being carried to the nest by a Red Ant, Myrmica sabuleti (Myrmicinae).

Photo: © Peter Eeles,

This habit of feeding on plants, then switching to ant brood, is called phyto-predacious behaviour. It’s rare, and has only arisen in the genus Phengaris, and the Africa-centred genus Lepidochrysops.

The Ant-Butterfly ‘spectrum’

What’s happening with the Large Blue isn’t mutualism, it’s parasitism. As we saw in a recent LE, parasites exploit other creatures. Mutualism is where a symbiotic, or synergistic, relationship exists. Both parties to it, gain.

Since Purefoy and Frohawk’s time, a lot of research has been done in the ant-butterfly field, leading to the realisation that ant-association occurs across a spectrum. All butterflies in the family Lycaenidae (gossamer-wings is their collective name) – Coppers, Blues, Silverlines etc – have the toolkit to deal with ants in some way. Some use defensive chemistry to defend themselves against ant attacks and are said to be myrmecoxenous.

Others (and there are many) have regular contact with ants, either mutualistic or parasitic, and are said to be myrmecophilous. They mimic the chemicals the ants use to communicate with one another, to control the ants. Much of this research has been carried out on southern African species, because we are very fortunate here. Our area (and Australia) has a preponderance of such butterflies. Compared to the northern hemisphere, or South America, we have a relatively large number of them and they are quite easy to find and study. We have some in Kloof (or to be more accurate, the Ethekwini area). I’m going to share with you some examples.

Myrmecoxeny – ants tolerate, and don’t attack

First of all, an example of what is probably a myrmecoxenous butterfly, the familiar Zebra Blue, Leptotes sp. There are four species in the genus Leptotes whose adults are impossible to tell apart without dissecting their genitalia, but the commonest is Leptotes pirithous, the Common Zebra Blue.

Zebra Blue Leptotes sp. on its preferred food – seed pods of Plumbago auricula.

The sluglike larva, feeding on a high protein food source, is typical of the Lycaenidae. This larva has the extra-thick cuticle (skin) typical of myrmecoxenous species; it gives the ants little to grab onto. If you look closely it is covered with tiny white spots. Each one of those is a specialised gland that exudes chemicals that the ants react to, a perforated cupola organ, or PCO for short.

Scanning electron micrograph of a PCO from a Chrysoritis species.

Image: courtesy Alan Heath

These PCOs are found on the skin of all lycaenid larvae. They may secrete substances to pacify ants that might otherwise attack the soft-bodied larvae. They could represent a key evolutionary development that allowed the ancestral lycaenids to avoid ant attacks by appeasing them, which later evolved into mutualism.

Zebra Blues Leptotes sp. are familiar garden butterflies. Here you can see both sexes getting ready to mate.

Mutualism – ‘give and take’

Mutualistic larvae usually have some kind of ‘reward’ that attracts ants and keeps them near. This often takes the form of a ‘honey gland’ or dorsal nectary organ (DNO) on the seventh abdominal segment, that exudes nutritious secretions. A good example of this is seen in the genus Euchrysops (Smoky Blues).

Half-grown larva of Grey Smoky Blue Euchrysops malathana photographed near Uve Rd in Kloof. It’s feeding on the pods of the Wild Black-eyed Pea, Vigna unguiculata (high protein food source again) and the DNO is clearly visible as a lens-shaped pale patch on the dark back of the larva near its rear (left) end. PCOs are also visible as tiny dark dots all over its body.

Adult male Grey Smoky Blue Euchrysops malathana, perching on a grass stem close to a patch of the larval host plant. The genus Euchrysops is mutualistic with ants, but it is very closely related to a parasitic (actually phyto-predacious) genus, Lepidochrysops, but more about those later…

Mutualism is common in Lycaenidae and there are many examples. The species covered so far are all in the subfamily Polyommatinae – covering a wide range of strategies. Also in that subfamily is the genus Anthene (Ciliate Blues), that has many mutualistic species, with very close relationships with ants. An example is the Spotted Ciliate Blue, Anthene larydas, a common butterfly in our area.

This third instar Spotted Hairtail Anthene larydas larva has exuded a drop of secretion from its ‘honey gland’ or DNO, attracting an attendant ant, Lepsiota sp., probably capensis (Formicinae). The secretion (which is known to contain sugars as well as amino acids) is clearly a benefit to the ants.

The same Spotted Hairtail Anthene larydas larva, extruding its eversible pair of tentacle organs (TOs). These are situated on the eighth abdominal segment and secrete volatile substances that attract and alert ants if a caterpillar is alarmed.

These substances have been shown in some lycaenids to mimic the ants’ alarm pheromones. They excite the ants and by doing so induce them to protect the larva from predacious wasps and flies. In this way the larva benefits from the ants it attracted using the ‘honey gland’ – i.e., true mutualism with both organisms in the relationship benefiting from the other’s presence. Note that in addition to the protection from the ants, the larva is well camouflaged on its host plant, a Vachellia species. It feeds on the growing shoots of this leguminous plant – again, a high protein food source.

These Spotted Hairtail Anthene larydas pupae are formed on the leaves of the host plant. It is not certain whether these pupae are protected from ant attack. Pupae of some species of lycaenid in the subfamilies Aphnaeinae and Polyommatinae are definitely attended by the ants in whose nests they were formed.

Spotted Hairtail Anthene larydas is a familiar, beautiful small butterfly found commonly around Kloof. This is a male.

Another lycaenid subfamily, Theclinae, has many examples of mutualism. One of the most familiar species is the Common Fig-tree Blue, Myrina silenus ficedula.

Common Fig-tree Blue Myrina silenus ficedula lays conspicuous bright white eggs on many species of Fig, as well as the related Mulberry, Morus alba.

Common Fig-tree Blue Myrina silenus ficedula larvae are often attended whilst on their host plant by various ant species, like these Technomyrmex pallipes (Dolichoderinae) ants. This one was in my garden in Gillitts, feeding on Tickey Creeper Ficus pumila. This alien I tolerate because of its colonies of butterflies and ants!

Common Fig-tree Blue is one of our most well-loved small butterflies; males like to choose a prominent twig or branch and perch there, showing off their exquisite blue wings to delight the eye. (He’s actually doing it to scare off other males and entice females, but we won’t mention that…)

Mutualism – ‘moving in’

The subfamily Aphnaeinae is noted for its ornate larvae that often take mutualism to the point where they live inside the ants’ nests. The relationship is still transactionally ‘fair’ – the ants get ‘honeydew’ and the larvae, an ‘enemy-free’ space – shelter and protection from predators. A few have taken it still further and are actually predacious on the ants, or have a ‘cuckoo’ – type relationship with them – the larvae are fed mouth-to-mouth by ants. This is known as ‘trophallaxis’.

Fully grown larva of Natal Silverline (until recently, Natal Bar), Cigaritis natalensis. It feeds on a variety of host plants, including Turkey-berry Canthium inerme, Smooth Tinderwood Clerodendrum glabrum, and Sourplum Ximenia caffra.

Like many Aphnaeinae, Cigaritis natalensis’ larva has a full set of ant-controlling organs. At the rear end (L) it has TOs, a DNO, and the body is covered with PCOs as well as specialised bristles, or ‘setae’, many of which are branched or club-shaped, and are sensitive to touch. The larvae are usually found sheltering inside hollow stems of the host plant, in the company of Cocktail Ants (Crematogaster sp., Myrmicinae). This individual was reared to adulthood from an egg, in the absence of ants – showing that the ants are not essential to its survival.

Male Natal Silverline, Cigaritis natalensis, a very familiar little butterfly. This is the attractive winter form obscura, showing his spectacular iridescent blue lustre.

It is a speciality of the KZN coastal area up into the hills. Males can be seen whirling around hilltops and prominent bushes, perching on twigs and sallying forth to chase away other insects. They will even take on much larger butterflies like Charaxes, or even dragonflies! They are the Jack Russell terriers of the butterfly world, and are often seen in Kloof gardens.

Endemic to South Africa, and mostly mutualistic with at least one parasitic species, the genus Chrysoritis (Coppers and Opals) has about 43 species. Like the related Cigaritis, most of the larvae can be reared in the absence of their host ants, but females will rarely lay on plants devoid of the right Crematogaster ant species. The larvae live inside structures created by the ants, from simple corrals to elaborate ‘carton’ nests. ‘Carton’ is a form of paper made by the ants from chewed up vegetable matter.

Mooi River Opal Chrysoritis lycegenes is found in the KZN Midlands, where its larva feeds on plants such as Diospyros, Searsia, Myrsine spp. and Osteospermum moniliferum (Tick-berry). This larva is being attended by Crematogaster ants in a corral they have built in the bark of the host plant, and is extending its TOs to chemically control the ants.

The closely related Natal Opal Chrysoritis natalensis is found near clumps of its local host plant Osteospermum moniliferum. Most of the colonies are close to the coast, sometimes on the actual beach.

This Natal Opal Chrysoritis natalensis pupa was found in a shelter made of carton between the leaves of the host plant.

Mooi River Opal Chrysoritis lycegenes is found across a wide range of country in the KZN Drakensberg foothills. Their fiery orange colour is conspicuous against the green background of their home.

Natal Opal Chrysoritis natalensis has a golden opalescent glow at the bases of the wing uppersides. It’s worth looking for these beauties whenever you visit the beach – watch the Tick-berry bushes!

Non-vegetarian butterflies – parasitism and entomophagy

Now let’s look at a subfamily of the Lycaenidae that doesn’t touch plant matter at all… the Miletinae. Some examples are easy to find in the Highway area – Woolly Legs (Lachnocnema sp.) and the Basuto Skolly (Thestor basuta basuta).

Here’s an entomophagous (insect-eating) caterpillar – Southern Pied Woolly Legs Lachnocnema laches. It’s feeding on a nymph of a plant hopper (Homoptera) and the lady who photographed it, Allison Sharp of Hoedspruit, assures me that once the larva bites the homopteran, that insect doesn’t move, and allows itself to be eaten. Venom at work… who knows? A Ph.D waiting for some entomologist here! 

This is probably the first (and currently only) photograph of a Southern Pied Woolly Legs larva being fed by trophallaxis from an ant, probably of the genus Pheidole (Myrmicinae) – but the experts are not 100% sure! This is something that occurs in other Lycaenid subfamilies but to date not recorded in Miletinae, and it was photographed in Westville at the home of the Caterpillar Rearing Group stalwart, Suncana Bradley. In fact, until recently these larvae were thought not to be myrmecophilous, but we now know that they shelter in ants’ nests at the base of the plant hosting their homopteran prey.

A female Southern Pied Woolly Legs Lachnocnema laches interacting with Pheidole ants that appear to be ‘farming’ tiny scale insects (too small to be identified!) such insects are known to ‘bribe’ ants by producing sugary secretions.

Butterflies in the subfamilies Miletinae and Poritiinae are very fond of scale insect honeydew, and may be found in numbers wherever they may be.

A male Southern Pied Woolly Legs Lachnocnema laches sucking ‘honeydew’ from scale insects (probably Pulvinaria sp.), in the company of Pheidole ants.

 ‘Woolly Legs’ butterflies are very widespread and common across Africa. Most pesticide-free gardens will have a population of them, and they carry out a similar control mechanism to Ladybugs and Hoverfly larvae – a natural pest control system.

A good reason NOT to spray insecticides. The males are territorial and are often seen whirling around prominent bushes, tiny little turbocharged black atoms, perching for a few seconds on a prominent leaf or twig.

Eggs of Basuto Skolly Thestor basuta basuta. The genus Thestor is probably the lycaenid group with the most extreme behaviour. The white dots are eggs of Basuto Skolly Thestor basuta basuta, laid amongst the young of these Jassid plant lice (Homoptera). The adult Jassids can be seen hanging from the plant – which is a leaf of Hypoxis (African Star-grass) which has been deformed by their sap sucking activities.

Basuto Skolly larvae feed on Homopterans such as the Cottony Cushion Scale Icerya purchasi, at least to the third instar, as in this image by the late John Joannou. Photographs of these are extremely rare, and it’s not known how they complete their larval stages. We know they enter the nests of the Pugnacious Ant, Anaplolepis custodiens, but what they feed on, we don’t know. It could be detritus (the ants’ wastes) or trophallaxis.

Final instar larva of Peninsula Skolly Thestor yildizae. Skollies (Thestor sp.) have radiated across the Karoo and Fynbos areas, splitting up into 27 different species.

Very few Skollies have known life histories; the fully grown larvae are found deep inside ants’ nests where they are difficult to find, let alone photograph. This final instar Peninsula Skolly Thestor yildizae, from Cape Town, was photographed by lycaenid authority Alan Heath, who assisted me greatly with this article. It is being fed by trophallaxis from a Pugnacious Ant, Anaplolepis custodiens – these larvae are the cuckoos of the butterfly world.

Male Basuto Skolly Thestor basuta basuta. Fortunately, you don’t have to go all the way to the Cape provinces to find these fascinating butterflies. Basuto Skollies can be found in the local Sandstone Sourveld Grassland, wherever termite mounds are found.

Pugnacious Ants live in underground nests close to termite mounds, where they prey on the termites. We know that Basuto Skolly larvae live inside Pugnacious Ants’ nests, but to date we don’t know for sure what they do there. Do the ants feed the larvae by trophallaxis on chewed up termites? Or do the larvae survive on the debris in the ants’ nest – dead adults, pupal skins (exuviae) etc, known as detritus?

These Pugnacious Ants Anaplolepis custodiens are preying on alate termites near a Basuto Skolly colony.

Phyto-predacious butterflies in South Africa

There are only 10 or so species of the Eurasian genus of Large Blues, Phengaris, in the world. They are scarce. The genus Lepidochrysops, Giant Cupids, has 136 species of which 50 are found in South Africa. Some of them are very widespread, such as the Patrician Giant Cupid, Lepidochrysops patricia. All whose life history is known, are phyto-predacious – the females lay eggs on (usually) the flowers of the host plants, and the young larvae eat the flowers and ovaries (high protein food sources). As in Phengaris, at the end of the third instar the larvae solicit ant attention and are carried into the ants’ nests, where they spend the rest of the larval stage as parasites on the ants.

This behavioural type has only evolved twice in the Lycaenidae. The latest theories propose, with some supporting evidence, that this is an adaptation from mutualism for survival during the aridification of Africa during the Miocene era – which also drove our ancestors out of the forest, where they had to adapt to an arid grassland existence by walking on foot. The butterflies adapted to spend the dry seasons sheltering inside the ants’ nests, where they became carnivorous and thus had abundant food when vegetation was scarce.

Although Lepidochrysops are widespread in South Africa, they are quite scarce in our local area. Photographs of adults are commonplace, but photos of the early stages are scarce, and those of the larvae interacting with ants, almost impossible to get hold of!

Patrician Giant Cupid Lepidochrysops patricia is found in Krantzkloof Nature Reserve. This is one of the most widespread of the genus. The female lays her eggs on the flowers of Lantana, usually the indigenous L. rugosa (Bird’s Brandy) but they have been known to use the exotic invader, Lantana camara. When they moult to fourth instar, they solicit carriage by Camponotus maculatus, the Spotted Sugar Ant. They spend the rest of their larval stage in the ants’ nest, devouring their brood.

So there we have it – South Africa is one of the best places to study butterfly-ant mutualism, parasitism and predation. Scholars from all over the world are busy working on these – many papers have been published to date, and several are in preparation as I write this. And it’s not only the butterflies that require study. The relationships between, and identity of, our ant species are poorly understood. Several entomologists are currently at work trying to unravel these, and their relationships with butterflies may be key to their understanding.

So next time you see ants crawling all over a plant in your garden, resist the temptation to reach for the can of ‘Doom’. You may be about to exterminate a very special butterfly!


This article could not have been produced without the help and assistance of some of the researchers in the field. Alan Heath, Peter Hawkes, Simon van Noort and Marianne Espeland all provided insights and identification assistance for this article. Photos were provided by Allison Sharp, Alan Heath, Peter Eeles of UK Butterflies, and the late John Joannou.

Steve Woodhall is a butterfly enthusiast and photographer who began watching and collecting butterflies at an early age. He was President of the Lepidopterists’ Society of Africa for eight years, and has contributed to and authored several books, including Field Guide to Butterflies of South Africa and Gardening for Butterflies. His app, Woodhall’s Butterflies of South Africa, is described as the definitive butterfly ID guide for South Africa.