Six things about sawflies
You’re probably familiar with the less than endearing larvae of sawflies - “spitfire grubs”.
But if you’re like me, that will be about the limit of your knowledge of this group of insects.
Our recent discovery of a new home species has prompted me to learn more about sawflies. Here is some of what I’ve found.
Sawflies are not flies
They’re not even close. Flies belong to the order Diptera, which means two wings.
Sawflies have four wings and differ from flies in many other respects. They belong to the order Hymenoptera, which includes wasps, ants and bees. The German word for these insects - Blattwespe, meaning leaf wasp - is much closer to the mark than sawfly.
But where do they fit in the hymenopteran family tree?
This diagram summarises our current understanding of the evolutionary history of the Hymenoptera - based on an analysis of gene sequences (ref. 1).
The 153,000 species of hymenopterans known globally are the product of around 281 million years of evolution.
Around 240 million years ago, one line evolved with a thin, flexible waist and gave rise to the wasps, ants and bees.
The remaining lines retained a thick waist and became the sawflies.
Australian sawflies evolved from one of these lineages, outlined in red in this diagram.
2. Sawflies are classic Gondwanan insects
Most Australian sawflies (140 of 161 species) belong to the family Pergidae. It is the third largest family of sawflies globally and its members display a great degree of diversity in their morphology and biology.
We have seen 7 pergid species in our forest, representing 4 different subfamilies. I think this gallery of both adults and larvae nicely illustrates the diversity of Pergidae.
Our new home sawfly, which I mentioned at the beginning of this post, is the black and yellow number with the extravagant antennae - Lophyrotoma cyanea.
The Pergidae are almost entirely restricted to Australasia and South America and are absent from South Africa - the hallmark of a Gondwanan group.
A study of the gene sequences of Australian pergid sawflies (ref. 2) showed that this family split from their closest relatives 153 million years (Myr) ago. They have a long history!
At that time Australia was part of the Gondwana supercontinent. It was still joined to South America but Africa had moved away. This could explain why pergid sawflies are shared between Australia and South America but are not present in Africa.
By Fama Clamosa - Own work, CC BY-SA 4.0, https://commons.wikimedia.org/w/index.php?curid=67003306 https://commons.wikimedia.org/w/index.php?curid=67003340
While Australia and South America both have Pergidae sawflies, the species in those two continents differ.
The gene sequence comparison study (ref. 2) dates this taxonomic split to 90-140 Myr ago, coinciding with the known time of formation of seaways between South America, Antarctica and Australia at 90-110 Myr ago.
So although it started with the same stock, pergid sawfly evolution took a different course in South America and Australia.
3. Sawfly larvae are gregarious herbivores
Most hymenopteran larvae eat meat - one way or another. Adult wasps either directly feed their developing young captured insects or they parasitise another insect larva, laying their eggs on or inside its body. The larva hatching from the egg eats the tissues of its host as it develops.
Ant larvae are fed a range of animal and/or plant tissue, depending on the species, while bee larvae are provided with pollen and nectar.
But almost all sawfly larvae are herbivores. Most Australian Pergidae feed on the leaves of eucalypts and other Myrtaceae species, including Leptospermum and Melaleuca. Other known food plants include the leaves of grasses, dock (Rumex), brambles (Rubus) and blueberry ash (Elaeocarpus reticulatus). Some species inhabit the leaf litter, feeding on dead leaves.
The feeding biology of the familiar Steelblue or Spitfire Sawfly (Perga affinis) is well understood (ref. 3).
The larvae are gregarious feeders. During the day they congregate in a tight, quiescent cluster of 20-40 individuals on a Eucalyptus stem. As the larvae become mature, they cluster on the trunk of trees.
Clusters can come together to form mega-colonies of thousands of larvae. In some cases these consist of two or more different species (ref. 3).
Come nightfall, larvae of Perga affinis move in a line higher up the tree to foliage, where individuals separate and feed at different leaves. At dawn they come together in a single cluster.
Our local spitfire species Perga sp. shows a somewhat different feeding behaviour.
At night, the cluster migrates from its roost on a bare stem back into the foliage. At their chosen feeding site, the cluster spreads out over a couple of adjacent leaves so that every individual can access a leaf edge. Having consumed leaves in that patch, the larvae form a cluster again and take a break. The next morning they can be found at their roost on a bare, defoliated stem.
The cluster of about 12 larvae shown in this gallery was seen feeding on the same 1.5m tall Eucalyptus globoidea sapling (shown in last image) for over two weeks, from 4th-20th November.
No sign of the cluster was seen after that time. The larvae may have migrated to the soil to begin pupation, although a search of the area beneath the sapling revealed no trace of them.
Upper branches of a eucalypt sapling showing almost complete defolation by sawfly larvae
As shown in this photo, larvae can eventually almost completely defoliate a small sapling. However the tree probably recovers in most cases.
We often find larvae on epicormic growth of large trees. This foliage is often surplus to requirements as such trees have regrown their crowns. We have never observed extensive defoliation of large trees in our forest by sawfly larvae. It’s likely that the population of these larvae is kept in check in our relatively undisturbed ecosystem by natural forces.
Having defoliated their original tree, the group heads off in search of another. It moves in single file following a leader which is always the same individual.
During their march, the larvae arch their body and touch the back of the animal behind with the end of their abdomen - kind of like a sawfly conga-line.
They regularly tap their follower (4 taps every minute or two) to maintain contact (ref. 4).
If an individual loses contact with the line, it begins tapping more frequently on a leaf or branch. Individual larvae or the whole group respond to its signal and call the lost animal back to the group by tapping. A remarkable system of communication and group cooperation!
The larvae have a very simple visual system - just a pair of eyes, each consisting of a single ocellus (ref. 5). However, it appears to work well enough to locate a new food tree from up to 15 metres away.
Carne (ref. 3) reports that migrating colonies will move towards a human figure standing on their horizon - presumably mistaking them for a tree. To quote Carne, “by moving from side to side, the colony can be made to follow a sinous path”. Now that’s mean!
4. Spitfire sawflies don’t spit
…but they do have various strategies for self-defence.
You’ll be familiar with one if you’ve ever disturbed a cluster of spitfire sawflies (Perga sp.). Kerri played the aggressor here to elicit the behaviour.
Sawfly larvae (Perga sp.) responding to a threat
The sight of several larvae simultaneously raising the front and rear ends of their bodies would deter many potential predators. This makes a strong argument for group living.
The larvae of another type of sawfly (Pseudoperga sp.) make an equally confronting sight as they flick their bodies in unison.
Sawfly larvae (unknown Perginae sp.) simultaneously flick their bodies to deter a predator
If this doesn’t scare off the predator, the larvae have a second line of defence. They discharge a thick, sticky fluid from their mouth with a strong smell of eucalyptus oils. They don’t spit it at the aggressor. The fluid just oozes out of their mouths - perhaps pressed out by their body movements. It often ends up over their body or on their neighbours.
I find the odour quite pleasant - just like eucalyptus oil - but predators undoubtedly associate it with the ill-effects of ingestion.
The discharged fluid is a concentrated form of the oils in the food plant (ref. 4), the main component being a metabolite of the toxic compound cineole.
This begs the question of how the sawfly larvae are able to protect themselves from the same toxic compound when they ingest eucalypt leaves.
The answer lies in part in a specialisation of the sawfly mandibles. Their inner surface houses a brush-like structure, which may help to separate toxic oils such as cineole from nutritive leaf material (ref. 6).
The oils are then passed to diverticula in the pharyx, where they are stored, rather than moving directly into the foregut. When threatened, the larva regurgitates this oil from the diverticula. This also serves to clear the diverticula for further storage of oil.
So the one compound, a potential threat to the sawfly’s health, is used to protect it against another threat - predators. Clever!
While these strategies work effectively against many predators, they provide only limited protection against parasitoids.
Sawfly larvae are targeted by a number of parasitic insects, including tachinid flies and trigonalid and ichneumonid wasps. Their larvae develop inside the body of the sawfly larva, consuming it from within. Mortality levels can reach 80% from such parasitoid infections (ref. 7).
However, it has been reported that sawfly larvae wave their bodies defensively at ichneumonid wasps that are trying to parasitise them and dead wasps can be later found stuck to the sawfly exudate. So they’re clearly not totally defenceless!
UPDATE - Sawfly larvae sometimes do spit!
An article by Philip Weinstein in the magazine Australian Natural History (ref. 8) has alerted me to the fact that in warmer climes in spring and summer, sawfly larvae can spit!
At higher temperatures their regurgitated oil becomes more fluid, so that when they bend their body the liquid can be “shot out almost 20 cms”! We’ve not experienced this in our southern forest - perhaps because it never gets warm enough in spring/early summer. By high summer, the larvae have dug into the soil to pupate.
As Philip points out, the oil spat out by the sawfly has the same composition as eucalypt cough drops, so there’s no reason to worry about being hit by it.
5. How sawflies get their name
Female sawflies lay their eggs on the leaves of their larval food plant. Like many insects they use a special extendible organ - an ovipositor - to deposit the eggs.
But sawflies don’t just place the eggs on the surface of the leaf. They make a cut and deposit the egg under the epidermis through that slit. The ovipositor has sharp teeth, like a saw to make that cut - hence the name for these insects.
The shape of the saw is often diagnostic for the species. This photo from a 1918 publication illustrates the variability in saw shape in different pergid species.
Plate XIII from Morice, F.D. (1918) “Notes on Australian Sawflies, especially the Authors Types and other specimens in the British Museum of Natural History and the Hope Collections of the Oxford University Museum”. Trans. Entomol. Soc. London pp. 247-332
We recently found a female Lophyrotoma interrupta sawfly, which died soon after we captured her. I took the opportunity to dissect out her ovipositor in order to photograph the saw.
The saw lies within a pair of sheaths. After removal of one sheath, the saw with its fine teeth (arrowed) is revealed. The cups on the sheaths presumably help the ovipositor to maintain contact with the leaf surface as the female deposits her eggs.
The females of some groups of pergids stand watch over their freshly laid eggs and the larvae. They will often sit motionless near or over their clutch, ready to ward off anything threatening it.
We observed this female Pteryperga galla doing just that. Kerri spotted her on a blueberry ash leaf and was surprised when she didn’t fly away as she approached. She was able to remove the leaf with the female still attached and bring it inside to get close up photos.
It was only when processing the images of the sawfly for this post - several years after they were taken - that I noticed the row of developing embryos she had been guarding on the underside of the leaf. Each of these is covered by a protective layer of leaf epidermis.
It’s amazing what you see when you know what to look for!
Pseudoperga guerinii shows the same behaviour. A recent iNaturalist observation shows the female guarding her clutch of eggs, which she had laid on the midrib of a eucalyptus leaf. Then two weeks later, after her larvae had hatched, she remained the vigilant mother.
I have recently observed the same thing for myself in the closely related species Pseudoperga ferruginea. You can read about this in the third part of my post Sawfly Galleries.
6. Sawflies make only a fleeting appearance as adults
Like many insects, sawflies spend the great bulk of their lives as larvae and pupae.
The lifecycle of Perga affinis has been well-characterised by Carne (ref. 3). In the ACT, eggs are laid in the autumn by freshly eclosed females. This species reproduces parthenogenetically. Males play no part in reproduction, even though a small number appear at the same time as the females.
The same applies to some, but not all pergid species. Carne notes that mating has been observed in Pseudoperga guerinii. The highly branched antennae of Lophyrotoma males suggest that these sawflies seek out and presumably mate with females. Indeed, this iNaturalist observation shows a coupled male and female Lophyrotoma interrupta. while this one shows mass mating of hundreds of Lophyrotoma zonalis adults on a weeping paper bark (Melaleuca leucadendra) tree.
The eggs of Perga affinis are injected into eucalypt leaves in closely spaced rows perpendicular to the midrib. The larvae hatch out after 30 days, cut their way out of the leaf tissue and gather together on the leaf surface in a rosette pattern.
Larvae pass through six instars over the next 5-6 months, increasing their body weight 2000x until fully grown in October.
At that time the larvae migrate to the base of their food tree and burrow into the soil. A leathery brown cocoon is made, within which the larva moults to a prepupa stage. Cocoons are linked together in a compact mass.
The prepupa enters a period of diapause, remaining at that stage of development until the following autumn, when it becomes a pupa and ecloses to the adult stage.
So development from egg to adult takes a full year. A proportion of the prepupae can remain in diapause for up to 4 years.
After emerging from the cocoon, the adult sawfly climbs to the surface of the ground and attempts to take flight. After landing on a leaf in a suitable eucalypt tree, she lays her eggs and the life cycle is completed.
The adult female does not feed and depends upon her larval fat-body for energy reserves. She lives for only around a week, dying from exhaustion soon after laying her last eggs.
The life cycles of few other pergid sawflies are understood in the same detail as Perga affinis.
I described above how a cluster of feeding Perga dorsalis larvae we had been monitoring disappeared in late November. Adults of this species were observed for the first time in our forest in mid-late February of the following year, suggesting that it has a similar developmental timetable to Perga affinis.
Macdonald and Ohmart (ref. 7) reviewed the available knowledge for other Pergidae sawflies in 1993.
Aggregation of larvae appears to be a common feature across the different pergid subfamilies. Larvae of Pteryperga galla remain in a cluster as they feed rather than separating, as shown by observations on iNaturalist like this one. The same applies to Pseudoperga lewisii.
The group migration seen in Perga affinis is also displayed by Pteryperga galla, although this species moves as a 2-3 deep mass of overlapping individuals rather than in single file (see this iNaturalist observation).
The long 7-8 month larval period of the genus Perga appears to be an exception. Other genera in the subfamily Perginae spend only a few months as larvae, while species in the Pterygophorinae subfamily complete larval development in just a few weeks.
In the Pterygophorinae, pupation occurs on the foliage of the food plant rather than in the soil.
Finally, while adult Perga affinis sawflies do not feed, the same does not apply to all Pergidae. Macdonald and Ohmart (ref. 7) state that female sawflies in the subfamily Pterygophorinae have been observed feeding on nectar. I have photographed this adult Eurys sp. sawfly (subfamily Euryninae) on Burchardia umbellata flowers and there is a report of this genus feeding on Philotheca buxifolia flowers.
So much remains to be learnt about the biology of these insects!
References:
Peters, R.S. et al. (2017) “Evolutionary History of the Hymenoptera.” Current Biology 27: 1013-1018
Schmidt, S. and G.H. Walter (2014) “Young clades in an old family: Major evolutionary transitions and diversification of the eucalypt-feeding pergid sawflies in Australia (Insecta, Hymenoptera, Pergidae).” Molecular Phylogenetics and Evolution 74: 111-121
Carne, P.B. (1962) “The characteristics and behaviour of the saw-fly Perga affinis affinis (Hymenoptera)” Aust. J. Zool. 10: 1-34.
Meyer-Rochow, V.B. (1972) “Verständigungsweisen bei koloniebildenden Blattwespen- und Käfer-Larven.” Zeitschrift für Tierpsychologie 30: 451-455.
Meyer-Rochow, V.B. (1974) “Structure and function of the larval eyes of the sawfly, Perga.” J. Insect. Physiol. 20: 1565-1591.
Schmidt, S., G.H. Walter and C.J. Moore (2000) “Host plant adaptations in myrtaceous-feeding Pergid sawflies: essential oils and the morphology and behaviour of Pergagrapta larvae (Hymenoptera, Symphyta, Pergidae).” Biological Journal of the Linnean Society 70: 15-26.
Macdonald, J. and Ohmart, C.P. (1993) “Life history strategies of Australian pergid sawflies and their interactions with host plants”. pp. 485-502 in Wagner, M. and Raffa, K.F. (eds). Sawfly Life History Adapations to Woody Plants. San Diego: Academic Press. (pdf available here).
Weinstein, P. (1991) “Undermining spitfire defence strategies”. Australian Natural History 23 (11): 849-857.
