Mallee Bronze Azure

Ogyris subterrestris subterrestris (Field)

Interesting Aspects

These butterflies are rarely seen flying, and when they do fly it is usually below about head height. This butterfly, like all other Ogyris species, has cryptically marked undersides, which makes the butterflies indistinguishable from the rough bark of trees and bushes, or the ground, when they are settled with wings folded.

The butterfly belongs to the compact O. genoveva - O. idmo Species Group of interesting Lycaenid butterflies within the Ogyrini Tribe, having very similar morphology and which are closely associated with the large Camponotus sugar-ants. O. subterrestris, along with O. halmaturia and O. otanes (all found in South Australia) and O. idmo (found in Western Australia) form the isolated O. idmo Species Subgroup within this group, as these four species have very similar adult morphology and early stages that associate only with Camponotus terebrans. O. subterrestris is one of the more specialised species of the sub-group having larvae which are believed to be predatory on the immature stages of these ants, which would allow these species to overcome the adverse effects of the periodic loss of foodplant, particularly from bush-fires, from which the very similar Ogyris otanes suffers.

It is presumed these butterflies have evolved from O. otanes type butterflies. Larvae of the latter will often strip their foodplants bare, and as they are incapable of moving to another nearby foodplant bush, those larvae not big enough to pupate, will perish. It is a small step for these larvae to either feed on other larvae within the colony (carnivorous feeding), or feed on the immature stages of the ants (myrmecoxeny) or on coccids that also associate with the ants. Lycaenid larvae often eat their own kind when their food source becomes insufficient to support all the larvae, and there are many Lycaenid species already known to practice myrmecoxeny. However, although it was probably a successful specialisation historically for Ogyris butterflies, it has not been very successful with the coming of European style habitation, as even though the ants are widespread and common, O. subterrestris (and Ogyris halmaturia the other myrmecoxenous butterfly in the group occuring in South Australia) are more threatened than O. otanes and have not adjusted very well to agricultural and urban development.

The Camponotus ant nests with which this butterfly is associated with, are usually situated at the base of tall mallee eucalypts (of various species), but will also utilise other trees and lower shrubs. There is usually more than one nest entrance around the base of the larger trees. The ants are nocturnal, and are large and ferocious (but without a sting). Adult butterflies emerge in the late morning after the ant activity above ground ceases. Newly emerged butterflies are covered in a loose fluffy down which may help protect them from ant attacks. They crawl up and out of the ant nest entrance, along the tree trunk, and after reaching about a metre from the ground, will then stop to expand their wings. This entire emerging process can take from 15 - 30 minutes before flight takes place.

Males usually congregate in open lek areas separate from the brood areas, and mostly wait settled with closed wings on or near ground level. Newly emerged females fly out to the males to mate, then return to the brood areas where they remain, separated from the males. As the ants are nocturnal, very little ant activity occurs above ground during the midday heat. Consequently, the female does most of the egg laying at this time, which she does by seeking out the ants nests, and usually those that already contain butterfly brood. She usually lands near the base of the trees containing the ant nest, and then walks down to the ant nest entrance to check for the presence of ant activity, and if all clear she will proceed to lay eggs. These are usually laid on the tree trunk, or even inside the entrance to the ant nest, and she does this by backing into the entrance upside down as far as she thinks it is safe to go without provoking an attack by the guard ants. She gains further ground by extending her abdomen considerably before proceeding to lay her egg cluster. More timid females will lay above the ground level entrance to about 5 cm, but in some instances up to 30 cm from the entrance. The eggs are also sometimes laid on bark or leaf debris, or even stones on the ground near the ant nest entrance. The further the eggs are placed away from the ants' nest entrance the more they will be subject to predation and parasitoids.

This butterfly has only recently been scientifically described, prior to which it was confused with the similar, but larger Ogyris halmaturia. It is unlikely that the two butterflies would occur together, due to their different ecological requirements. The two are difficult to tell apart. There are subtle morphological differences, but adult O. halmaturia butterflies (found in South Australia) differ from O. subterrestris mainly by being significantly larger. Flight period and location would also be important differentiation criteria.

Larval Food Host

Theorised to initially be fed by the attendant ants, but later turning myrmecophagous and feeding on the early stages of the ants. Several attempts have been made to set up artificial colonies of the attendant ants in captivity, one of which included a queen ant. All attempts to rear the larvae in these ant colonies have so far resulted in failure, with an early second instar larva being the largest size attained. No artificial ant food, or other vegetable or fungal food was accepted by the first instar larvae, and these young larvae were not interested in the early stages of the ants and actively avoided them.

Larval Attendant Ant

Larvae spend their entire life period inside the nests of the large, sand dwelling sugar-ants Camponotus terebrans. Interestingly, the butterfly associates only with the northern, pale coloured form of the ant.


Large, dark grey-brown (closely matching the colour of the bark on which they are laid), with a white ring dorsally, the micropylar area is white, hemispherical shape, base flattened, and with a very fine reticulated pattern. The reticulations are hexagonal. The eggs are typical for the species-group. They are laid in clusters at or near the entrance of the attendant ant nest. Near the point of larval emergence, the top of the egg, from about the limits of the white ring, collapses inwards. The larva eats out this collapsed portion to emerge. Individual egg clusters usually range up to 20 eggs, but clusters to 50 odd are known, and one female in captivity once laid 120 eggs. The larger clusters in the wild are probably the work of several females. Eggs hatch in about 8-10 days in summer, but up to 19 days in early spring. Larvae only emerge late at night when the ants are active. They immediately move downwards to ground level looking for the entrance to the ant nest. Most are intercepted by the ants, whereby the larvae immediately thrust their rear ends into the face of the ants, presumably to emit pheromones which triggers a passive response from the ants which then pick up the larvae in their giant mandibles and take them below into the ant nest. Hatched eggs are not destroyed by the ants, and remain clustered around the entrance to the ant nest until they are abraded away by the elements or the tree sheds its bark. Some of these old clusters can contain hundreds of egg shells.


Initially pale pinkish or greyish yellow, with a slightly darker dorsal area. Long onisciform shape, with scalloped lateral edges, the subdorsal and peripheral surfaces bearing short bristly hairs. There is a large brownish yellow smooth area above the head (prothoracic plate). There are two pairs of long recurved spinose setae situated dorsally near the rear end (sixth and seventh abdominal segments), and another shorter forward directed pair near the front. These long setae differentiate this species from the very similar Ogyris halmaturia whose immature larvae do not possess these dorso-posteriorly located spinose setae. The second instar loses the long spinose setae.

In captivity, the ants initially show great interest in newly hatched larvae, often picking them up in their jaws and carrying them around, but have not been observed to directly deliver them to the ant's brood chamber or to any other defined area within the nest. The ants later lose interest in the larvae which then wander around, and do not end up in the brood chamber.

The final instar is about 22 mm long, pinkish cream colour with darker, very indistinct dorsal chevron markings, and obvious black spiracles.Long onisciform shaped, with a double row of scalloped lateral edges, the anterior and posterior areas are flattened. Except for the sublateral perimeter, the body is without bristly hairs and secondary setae.The posterior, dorso-lateral organs are not developed. The opening for the posterior, median dorsal sugar secreting organ is present but has never been seen to be functional, and ants have never been seen to investigate this area. The head is small, smooth, yellowish brown, hidden beneath the body. Its mouth parts comprise mandibles and a further tubular-like part behind the mandibles, but the tube could not examined in detail for fear of injuring the larva, so it is not known if that part is a silk spinneret or a piercing mouth part.

In captivity, a mature larva (reared by Lindsay Hunt) was observed to rapidly thump its rear end down on the substrate when agitated, which may be an attentive warning signal to any ants in the vicinity or even other larvae. This thumping sometimes invited a mild reaction from nearby ants, but in general the ants did not pay much attention to the mature larva, presumably the latter is masked by pheromones similar to the ants. Larva movements were usually very slow in the presence of light, but in darkness it tended to be more active and quicker in its movements. A small eversible organ was sometimes seen being extruded at the extreme posterior end of the larva, but its movement was so quick that its origin (possibly from the anus) or nature was not determined by Hunt, although an ant was once seen to briefly grab hold of it!

If the butterfly larvae do feed on the young stages of the ants, then the socio-economic drain on the ant colony must be considerable. It is therefore probably no coincidence that where the butterfly occurs, the ant colonies are enormous. The reason for the latter may be due the ant's practice of farming coccids, which have a symbiotic relationship with the ants in supplying them with 'honey-dew' in return for protection. The coccids shelter in the ant nests during the day, emerging to suck on the tree sap by night. The higher the numbers of coccids, the more vigorous the ant colony. The only downside is the vigour of the tree, which is weakened by the sap sucking of the coccids.


About 15 mm long, brownish-cream, long cylindrical, rounded anteriorly, the abdomen is strongly arched, and there are some very short bristles on the top of the head. A girdle has not been reported about the pupae. Pupation occurs underground in chambers within the ant nest. The pupae are fragile and easily damaged. The empty pupal skins are eventually destroyed by the ants after the adult butterflies emerge. The pupal duration is about 20 days.

Flight Period in South Australia

Late September to late May. There are probably continuous broods over the warmer months, with the main emergence during spring and late summer-early autumn.

flight_subterrestris.gif (3.22 KB)

Within South Australia the butterfly has only been documented from the Riverland at four sites. The butterfly is also recorded from north-west Victoria and from the Broken Hill area in New South Wales. A different subspecies (petrina Field) is found in southwest Western Australia. The butterfly may be restricted to the distribution of its host ant, the northern pale coloured form of Camponotus terebrans, which occurs in the dry, hot inland areas of southern Australia.

map_subterrestris.gif (10.4 KB)

So far, this butterfly has only been found in the dryer mallee and temperate semi-arid inland areas having annual rainfall less than about 300 mm. (The closely related Ogyris halmaturia has evolved its adapation to habitat having annual rainfall greater than about 300 mm, and to its association with the southern dark coloured form of Camponotus terebrans). There may also be a common link between the known colonies along and adjacent to the River Murray system, and colonies often seem to be located near to ephemeral lake systems. The cessation of annual floods along the River Murray since the enormous uptake of water in interstate agricultural areas much further upstream, may be a bonus for the butterfly, allowing permanent colonisation of the riverine and adjacent areas by the host ant. The obligate host ants only associate with sandy soils.

Conservation Status in South Australia

Appears to be in decline in South Australia with populations affected by increasing urbanisation, the effects of climate warming and drought, and butterfly collecting, with the colonies now endangered.Colonies in the nearby Mildura area of north-west Victoria seem to be in a slightly more stable condition, at least for the present.


Known colonies in South Australia are associated with disturbed urban and agricultural areas, and roadside verge, and their safety is therefore extremely precarious. As the ant brood chambers are deep below ground level, the colonies are believed able to survive fires. The effects of toxic spray drift from adjacent vineyards and citrus orchards at two of the colonies has not had any major adverse affect, and this may be because the spray is not applied aerially. Small colonies could be disrupted by the diggings of large burrowing animals, particularly feral rabbits. Known colonies also occur adjacent to the plague locust belt, and this belt continues to be heavily sprayed, and there are no studies of what after-effects the sprays have on the environment. Males require an open area adjacent to the breeding colony in which they can aggregate for mating purposes (lek area). This area is highly vulnerable to any detrimental process. Undocumented colonies may be affected by the new heavy-mineral sand mining ventures in the Murray Mallee.

Conservation Strategy

Farmers living adjacent to the colonies have been advised of the presence of the threatened butterflies. Known colonies need to be continuously monitored for detrimental processes. It will be very important that appropriate Environmental Impact Studies are enforced by the relevant Authorities regarding the new heavy-mineral sand mining ventures in the Murray Mallee, which take into account this and other potentially affected threatened butterflies such as the Anisynta cynone and Herimosa albovenata. It should theoretically be possible to artificially transfer egg clusters to distal colonies of the attendant ant, although some initial experimentation would be required, as the pheromones of the different ant colonies may be different, in which case the eggs may not be accepted.


Some of the biological information for this butterfly was provided by Mike Moore who was the first person to discover this butterfly in South Australia in 1992. A final instar larva and pupa were recently discovered by David Lohman and David Moore, and the same larva reared to pupation by Lindsay Hunt.