Go to National Library of New Zealand Te Puna Mātauranga o Aotearoa
Volume 85, 1957-58
This text is also available in PDF
(570 KB) Opens in new window
– 275 –

The Breeding Habits of the Corduliidae (Odonata) in the Taupo District of New Zealand

[Received by the Editor June 25, 1957.]


Identification of the three Corduliidae found at Taupo, both taxonomically and in the field is given; followed by notes on their breeding in nature, and for the identification of their ova, together with the time required and method of hatching.


The three Corduliidae regularly seen in the Taupo district are:— Procordulia grayi (Selys), P. smithii (White), and Hemicordulia australiac (Rambur). The fourth Corduline on the New Zealand list is Somatochlora braueri (Selys) or Antipodochlora braueri (Fraser, 1939) to give its more recent name; it is a rarity in this district if found at all, and is not considered further in this paper.


Hudson (1904) describes ovipositing by P. smithii and gives a description of the final larval instar; of P. grayi he states: “This interesting and very rare dragonfly has occurred at subalpine localities in the South Island.” The female was first described by Fraser (1939) from specimens sent by Hudson. Salmon (1950) considers that P. smithii and P. grayi are synonymous. Tillyard (1926) says that the species H. australiae “Ranges far and wide over Australia, Tasmania, Norfolk, Lord Howe and Kermadec Islands, and is occasionally taken in New Zealand”. Hudson (1945) states he has only a single specimen. There appears to be no description of the ova or the hatching of the New Zealand Cordulines.

The successful recording of the breeding habits of an insect requires first its correct identification both taxonomically and in the field.


The taxonomic data for separating P. smithii and P. grayi have been prepared for me by Lieut.-Colonel F. C. Fraser, of the British Museum.

Comparison of Procordulia smithii with P. grayt

P. smithu (White). P. grayi (Selys).
Pterostigma dark ochreous. Pterostigma bright ochreous
Subtrigone in hindwing invariably absent. (Fig 1A.) Subtrigone in hindwing present and rarely absent as an aberration. (Fig. 1B.)
Postnodal area of female wings always strongly saffronated. Postnodal area of female wings not tinted with amber.
Frons of male brilliant blue metallic. Frons of male poorly metallic.
Male anal appendages slim, sinuous, elongately fusiform, with apex tapering to an acute point; apices converging or parallel. (Fig. 1C.) Male anal appendages stout and with a strong subapical angulation, apices obtuse and strongly divaricate. (Fig. 1D.)
Female anal appendages markedly curved, forcipate and widely separated throughout Female anal appendages straight converging and meeting at apices.
Gemtalia; lamina of male directed posteriorly, parallel to the axis of abdomen Gemtalia: lamina of male directed perpendicular to axis of abdomen.
– 276 –
Picture icon

Fig. 1. Hemicordulia australiac can be separated by its generic characters, black stigmata, and the internally placed spur-shaped projections on the anal appendages situated about a third of the way from the base to the apex; and its smaller size.

Picture icon

Fig. 2. H. australiae P. grayi P. smithii The terminal segments of the abdomen in males of, from left, H. australiae, P grayi, P. smithu; the shaded areas shown on P. smithii are brown during life, and hardly show against the black

– 277 –

In the Field

With a little practice and a good light it is easy to identify these insects on the wing. P. smithii and P. grayi are very similar in size, but smithii appears to have a narrower abdomen due to its dark, almost totally black colour with a narrow bright yellow stripe at the junctions of the abdominal segments. This narrow yellow line is only complete at the junction of the last two segments, in the case of the other junctions the line is formed by two narrow spots which lie almost at right angles to the axis of the abdomen, and taper oft towards the midline where they meet or almost meet, and also taper towards the lateral margins of the abdomen.

P. grayi has two large creamy-yellow spots on each abdominal segment that are situated on the anterolateral area of the segment; the spots appear to be rounded in outline, with the antero-posterior diameter greater than the lateral; these spots occupy about two-thirds of the width of the abdomen, and give it a broad flat appearance, which is accentuated by the short, broad, blunt anal appendages of the male.

H. australiae can easily be recognised by its smaller size and the narrowing of the abdomen at each end, which is noticeable in the female as well as in the male. The yellow marking on the abdomen takes the form of a narrow based triangle situated at the lateral margin of each segment except the last, with the base at the junction and the apex extending to a point a little over two-thirds down the segment; just before the apex the mark is again expanded into a much smaller wide based triangle with its apex directed medially; this terminal triangular expansion is developed to a greater extent on the middle and anterior segments of the abdomen.

The apparent narrowing of the abdomen is probably accentuated by having no yellow markings on the anal appendages or the terminal segment of the abdomen.

The general effect of these markings is that of an irregular yellow line down the lateral borders of the abdomen. There is also a bright yellow spot on the lateral surface of the thorax just below the anterior wing.

If the insect is flying towards you then the brilliant metalic green frons and green eyes are quite distinctive, being much brighter than the metallic blue of P. smithii.


The males of all three species of cordulines like to hawk up and down quiet glades or roads where they can fly through alternating sunshine and shade seeking the small insects that form their food, or settle on some exposed twig to sun themselves. Any glade of this nature may have males of one or two and occasionally all three species flying at the same time. The females seem to keep nearer to the water where they breed, but even they may be found half a mile or so from water during their first copulatory flight.

P. smithii is the most widely distributed species, and seems to be present on all the rivers and streams in the district, but not in any great numbers; it is most constantly found on the Tokaanu stream in the last mile of its course.

P. grayi is found all round Lakes Taupo and Roto Aira, wherever there are suitable weedbanks and rocks on the shore to facilitate the breeding and hatching out of the adults. Towards the end of the final larval instar the larva may make several expeditions from the waters of the lake in search of a suitable rock, always during the night; and when it has found one will return there for the final emergence; the site chosen may be some feet away from the lake, and is usually on a vertical or, better still, an overhanging face of rock; on one such site in Acacia Bay the overhanging face of a single rock had seventeen cast skins, with only five on the other three vertical faces. The cast skins may hang onto a rock for some days but do not survive the winter.

– 278 –

H. australiae regularly frequents the Waikato River, where it has bred in Richardson's Pool for at least 25 years, and also frequents the Spa Valley, Waitahanui Backwater, and Lake Roto Aira, in all of which places it probably breeds.


All three species have the following points in common, the females deposit their ova by dipping the two terminal segments of the abdomen in the water as they fly over it, frequently making from two to four splashes where they hit the water at about a foot or two apart in the course of each flight. The females are not accompanied by the males during these flights; the males are usually seen hawking near the edge of the stream or lake where the females are depositing their ova; this enables copulation to take place quickly and repeatedly following the deposition of a comparatively small number of ova. Ovipositing takes place during the months of January, February and March in all species; and in P. grayi it has been noted as early as December, and as late as the end of April.

All these species choose a green weed bank over which to lay their ova; and it is usually only done in bright sunlight; the green colour showing up very strongly under these conditions. The sites for ovipositing differ, P. smithii prefers a weed bank in a relatively fast flowing stream or river, and always makes her ovipositing flight upstream, so as to use the current to help wash off the ova, and then returns down stream again for the next flight.

P. grayi likes lake water usually over a weed bank in quite deep water up to ten or fifteen feet, and not less than four feet; the ovipositing flights are made in all directions over the weed. H. australiae oviposits over green weed banks in shallow pools in streams and rivers, and sometimes in the lake, usually two to four feet deep; the site chosen seems to require a pine tree or pine wood to be near it, as this species likes to roost at night and to sun themselves during the daytime high up in a large pine tree; in fact when one or more Cordulines are seen flying round the top of a pine tree, they will almost certainly belong to this species as the other two cordulines seem rarely to fly more than fifteen feet above ground, and prefer to rest on manuka, flax or similar shrub.

Cordulines on the wing can best be observed when stalked from cover, remembering that to see the colour pattern the sun must be behind the stalker and also that one's shadow must move only very slowly on the water, otherwise they will be scared away. Field glasses are not of much use, as unless one has had considerable experience in their use in bird watching or on some similar fast-moving objects, it is very difficult to keep dragonflies in the field of vision and in focus at the same time. Glasses of X8 magnification I find most useful, but are mainly used on dragonflies that have settled on some unapproachable object.

The ova for breeding experiments were obtained from females caught during copuation, as this allows one to be certain of the species of both the male and the female; if the capture is made about noon I always allow some 10 to 20 minutes of copulatory flight or resting if possible before making a capture; but if it is late in the afternoon, say after 3 p.m., this period can be considerably shortened.

Usually my captures have been made between 2 and 3 p.m. The capture of an ovipositing female while ovipositing, which would be ideal, is difficult in the species under review because during this action they are very wary and keep well away from any unusual object, while their actions are so quick that the trout which can be seen shadowing them rarely attempt to capture them.

Polaroid glasses are of great help in watching the movements of the trout in relation to the ovipositing dragonfly, and also for observing the movements of the full grown dragonfly larvae on the lake or river bed; in their younger stages the larvae hide in the weed beds and in the mud below, frequently being camouflaged with algae.

– 279 –

The best way to get ova is to hold the females by the forewings, with these wings held together vertically above the thorax, and allow the hind wings to move freely as the insect does in flying; then dip the two terminal segments of the abdomen into a tube of water from the river or lake; the ova will be discharged in a steady stream or intermittently, in which case it is as well to allow the insect to lift the abdomen out of the water as the striking of the water appears to act as a stimulant to ovipositing.

The eggs are usually laid singly or in bunches up to four at a time, and rapidly sink to the bottom of the tube, where they stick together in a mass that is difficult to separate. If the females do not lay at once they are placed in tubes plugged with cotton wool for taking home, where another attempt is made to get them to lay by the same means as before, only instead of a tube a petri dish is used, so that the ova can be spread over its surface, and so not allowed to adhere together in masses.

If this fails, they are strapped by the wings to a flat piece of cork with a slot cut through it to allow the abdomen to move up and down and so dip into water in a petri-dish placed below. This also allows small flies to be fed to the female, which together with stroking of the dorsal surface of the abdomen with a fine paint brush stimulates egg laying, according to Gardner (1952). All these methods were tried on a single specimen of H. australiae without result, but eggs were obtained by liberating the pair taken in copulation in an insectarium about fifteen inches square, with a petri dish of water on the floor. The insectarium was so placed that the inmates could rest in sunshine or shade; unfortunately I do not know whether a second act of copulation took place before the eggs were laid, as the pair had been in the cage for 24 hours before I saw any eggs in the water. P. grayi and P. smithu have always laid their eggs by the first method.

Picture icon

Fig. 3. New laid. Ten days old. Shell after hatching. Ova of top line, P. smithii, middle, P. grayi, bottom line, H. australiae.

– 280 –

The Ovum

All three species when they pass their ova direct into the water, the ova separate and fall singly to the bottom; if on the other hand they are passed into the air while the insect is flying they adhere together in the form of a blob or tape which, when washed off, remains as an adherent mass, and tends to sink more slowly. The ovum when first passed has several points of similarity in the three species, it has the round shape typical of an exophytic method of laying, and has a well developed pedicel at the anterior pole of the ovum, which takes the form of a rounded point arising from a round flat base, with the height of the point about equal to the width of the base. The shell of the egg shows an irregular hexagonal marking all over it, which Tillyard (1917) states is due to the boundaries of the follicle cells of the ovary by which the chorion or outer covering of the egg is secreted. The ova when they leave the oviduct are longer than they are broad, giving them an oval shape, and are covered with a thin gelatinous coat which is slightly thicker at the anterior pole where it surrounds but does not cover the pedicel, and much more so at the posterior pole, where its external surface seems to be ragged and poorly defined. This gelatinous membrane rapidly expands on coming in contact with water, and assumes different shapes in the three species, but they have in common that the area over the posterior pole of the ovum becomes very sticky so that it can adhere to other ova or objects. As the embryo forms in the ovum, it increases the width, thus giving it a more rounded form.

The ova of the three species differ in the following points:–


P. smithii, when first laid pale cream; after 48 hours dark brown.

P. grayi, when first laid creamy white; after 48 hours golden brown.

H. australiae, when first laid pale green; after 48 hours golden yellow.

Size of Ovum Without Pedicel

New laid. After 10 days.
P. smithii 1.0 × 0.6 mm 1.0 × 0.7 mm
P. grayi 1.1 × 0.7 mm 1.12 × 0.72 mm
H. australiae 0.6 × 0.35 mm 0.6 × 0.4 mm

The ovum of P. grayi appears to have a more prominent pedicel because it is set at the apex of a projecting process of the amniotic cavity. This is most noticeable when the egg is first laid, and the process gradually shortens and is absorbed into the body of the egg as it enlarges, but does not entirely disappear.

No process under the pedicel was seen in either P. smithii or H. australiae.

The gelatinous membrane surrounding the egg of P. smithii swells on contact with water to about twice its original thickness from 0.15 to 0.30 mm at the anterior pole of the egg, and keeps to the contour of the egg; the membrane over the posterior pole is more swollen and irregular and is apparently more sticky, as separated eggs when they come together stick by their posterior poles, and the gelatinous membrane becomes continuous between them, making it hard to separate them again.

As long as the egg mass is not too large this means that the anterior pole is directed out and so facilitates hatching, the same holds good if the egg becomes attached to a piece of weed.

In P. grayi the swelling of the outer gelatinous covering is much greater, taking on a tubular form with the length and width of the tube twice that of the corresponding diameter of the egg.

The egg is at first suspended in the centre with a cone of membrane arising from the perimeter of the pedicel and extending forwards to join the anterior end of the tube; the posterior end of the tubular sheath is more ill-defined and irregular, but has some sheet-like structures extending from the posterior pole of the egg directly backwards or tending to meet at a point; these are probably the supports for the micropile tube. The egg appears to be completely mobile in the tube. If the eggs adhere together in a mass they do so by their posterior poles, and the eggs move.

– 281 –

forwards in the tube until the pedicel projects completely and only a thin film of membrane covers the anterior pole. If on the other hand the egg is kept moving in a large volume of water without contact with anything it tends to migrate to the posterior end of the tube, which becomes more or less closely applied to the posterior pole, whilst the anterior edge of membrane bells out into a cone-shaped structure with the pedicel at its base, and may quite well perform the function of a sea-anchor. (Centre Fig. 3.)

H. australiae has a gelatinous membrane that is intermediate in its swelling between the other two. It still retains some of the lateral curve of the egg but increasing the thickness at the anterior end sufficiently to bury the pedicel at the bottom of a cone with its apex just projecting. The smaller size of the egg and pedicel and greater transparency of the sheath make it more difficult to observe.

Hatching of Ova

This differs from that depicted by Walker (1953) in that the pedicel is often not separated completely, but just around the circumference of the base adjacent to the dorsal surface of the embryo, this allows it to bend over laterally with a small part of the circumference as a hinge, while the chorion or shell, instead of having numerous splits around the circumference of the hole formed by the separation of the pedicel, has a single long split extending almost to the opposite pole of the egg. This allows the pronymph to back out of its shell as well as move forwards; birth takes place very rapidly and is followed in a few seconds by the shedding of the amniotic membrane or pronymphal skin, which splits over the head and down the back of the thorax, for the emergence of the second instar larva.

The shells of the hatched eggs of the three species can easily be recognised. P. smithii appears to have a thicker as well as darker shell, and the split that takes place from the anterior pole towards the posterior is usually sinuous in line, and the shell is so stiff that the margins of the split come together in almost perfect alignment after the pronymph has escaped, making it hard to detect where the split has occurred.

In P. grayi and H. australiae the split occurs in a more or less direct line from pole to pole, and after the pronymph has escaped both shell margins become inverted, leaving a gaping fissure down the shell; the shell of H. australiae is much smaller and of lighter colour than that of P. grayi.

The time required for the hatching of the ova varies considerably even when the ova come from the same batch of eggs and are kept under similar conditions; a batch of ova of P. smithii laid on 21/2/57 hatched from 25 to 33 days after laying, a batch of P. grayi ova laid on 14/12/56 took 25 to 33 days, and a batch of H. australiad ova laid on 6/3/57 took 22 to 29 days, but about half of these ova have not yet hatched, probably due to sharp frosts on the 3rd to the 5th of April.

Fraser (1951) has reported that keeping the ova of a corduline Oxygastra curtisii (Dale) under cool conditions may delay the gestation period very greatly, and points out that under the cooler temperatures prevailing in their natural riverine habitat the ova would normally not hatch until the ensuing spring As the natural habitat of our cordulines depositing their ova in rivers and lakes would be much cooler than in the water used experimentally, it is probable that in nature their ova do not hatch until the following spring.

The separation of the larval instars of these species will be left to another paper, when I have completed the breeding under controlled conditions of their complete life cycles.

– 282 –


My thanks are due to Lieut-Colonel F. C. Fraser, I.M.S. (Ret.), F.R.E.S., for preparing the taxonomic data on P. smithii and P. grayi and for advice and help with Odonata for many years; to the late Mr. G. V. Hudson, F.R.S.E., for supplying a specimen of Antepodochlora braueri for comparison with the local Cordulines; to Mr. A. E. Gardner, F.R.E.S., for advice on breeding Odonata and for the recent literature on this subject; to Mr. Allan Pye for assistance in capturing H. australiae.


Fraser, F. C., 1939. Proc. R. Ent. Soc. Lond. “B” 8; 91–94.

—— 1951. The Entomologist's Month. Mag. 87: 38.

Gardner, A. E., 1952. Breeding Dragonflies. Entomologist's Gazette. 3. 87–88.

Hudson, G. V., 1904. New Zealand Neuroptera. 16–17.

—— 1945. Fragments of New Zealand Entomology: 121–122.

Salmon, J. T., 1950. Trans. R. Soc. N.Z. 78: 1.

Tillyard, R. J., 1917. The Biology of Dragonffies. Cambridge. 229.

—— 1926. The Insects of Australia and N.Z. 85.

Walker, E. M., 1953. The Odonata of Canada and Alaska. Toronto: 1–292.

Dr. J. S. Armstrong

3 Titiraupenga Street,