Other formats

    TEI XML file   ePub eBook file  


    mail icontwitter iconBlogspot iconrss icon

Tuatara: Volume 25, Issue 1, July 1981

An Emergence Study on Uropetala Carovei Carovei (Odonata: Petaluridae) near Wellington, New Zealand, with Notes on the Behaviour of the Subspecies

An Emergence Study on Uropetala Carovei Carovei (Odonata: Petaluridae) near Wellington, New Zealand, with Notes on the Behaviour of the Subspecies.

Collections of final-instar exuviae of Uropetala carovei carovei from six sites near Wellington in January-March 1980 showed that the emergence duration is similar to that in U. c. chiltoni in the South Island but emergence commenced and finished later in the season at the sites studied. The study reveals a slight preponderance of males over females (54%).

As has been described for all petalurids with the exception of U. c. chiltoni, the rectal plates in all exuviae have been found to be open distally. The emergence stance, including novel aspects of the larva's attachment to the emergence support is described and notes are incorporated on oviposition, habitat and larval behaviour.


At the completion of its aquatic larval stages, an odonate larva leaves the water, clamps upon some suitable surface and the winged adult emerges from the larval skin. When the adult departs, a durable record of emergence remains in the form of the final-instar exuviae which can be identified to species and to sex. Thus as Corbet (1962) has demonstrated ably, regular collections of final-instar exuviae can provide information on the duration of the emergence period, the numbers and species of dragonflies and damselflies emerging, and the sex ratios for the population.

Seasonal patterns in the New Zealand dragonflies (Suborder Anisoptera) are not yet well understood. Wolfe (1953) reported briefly on emergence in Uropetala carovei chiltoni Tillyard near Lake Sarah, Cass, (43°03′S 174°46′E, S66 245155) and Deacon (1978, 1979) investigated emergence patterns at Lake Sarah in two damselfly species, Austrolestes colensonis (White) and Xanthocnemis zealandica (McL.), and two endemic corduliid anisopterans, Procordulia grayi (Selys) and P. smithii (White). Except for P. grayi in one year, however, Deacon's anisopteran samples were small.

page 23

There have been no emergence studies published on the Anisoptera in the North Island. I am involved in a study, at Gollans Valley near Wellington, on the biology of a further endemic corduliid, Antipodochlora braueri (Selys), and the lack of information on the other species locally against which to compare its behaviour was a serious hiatus from my point of view. To overcome this, I have made emergence studies on P. grayi, P. smithii, Hemicordulia australiae (Rambur), Aeshna brevistyla Rambur, and Hemianax Papuensis (Burmeister) at Pukepuke Lagoon (40°20′S 175°16′E, N148 782368) near Foxton which will be reported separately, and this present paper incorporates the results of emergence studies on Uropetala carovei carovei (White) at six sites around Wellington.

The Petaluridae comprises a small family in the Anisoptera with only nine known species in five genera. Svihla (1959) reported the distribution of eight species: Petalura gigantea Leach, P. ingentissima Tillyard and P. pulcherrima Tillyard in Australia; Tanypteryx hageni (Selys) in western U.S.A. and British Columbia and T. preyeri (Selys) in Japan; Tachopteryx thoreyi (Selys) in eastern U.S.A.; Phenes raptor Rambur in Chile; and U. carovei in New Zealand. Watson (1958) described the ninth species, Petalura hesperia from Western Australia; and Fernet and Pilon (1968) reported the range of T. thoreyi to include Canada.

Detailed life history investigations have been published for P. gigantea (Tillyard 1909, 1911), U. carovei chiltoni (Wolfe 1953), T. pryeri (Taketo 1960, 1971a) and T. hageni (Svihla 1959, 1960); and a contribution on T. thoreyi is in preparation (S. W. Dunkle, pers. comm.). Our understanding of the ecology and behaviour of the various species in the Petaluridae is, however, still far from complete. Additional information on the behaviour and morphological adaptations of U. c. carovei became available in the course of this study and the opportunity has been taken to include comments on these topics in this paper.

Study Areas

Six study sites were selected all close to Wellington and within 12.5km of each other. Site 1 was chosen for its proximity to Victoria University of Wellington, and the remaining sites are all close to the route I follow into my A. braueri study area in Gollans Valley, 41°19′E 174°45′S, N164 257147. Each site comprises a water seep on sloping ground in the shade of indigenous evergreen forest, the composition of which varies at each site but is basically coastal broadleaf forest at site 1, mixed beech (Nothofagus spp.)-coastal broadleaf forest at site 2, and beech-podocarp-broadleaf forest at sites 3-6. An understorey of shrubs, ferns and saplings is developed at each site. The sites conform to the general descriptions given elsewhere for U. c. carovei habitat (Winstanley and Rowe 1980). The six sites, identified by latitude and longitude, NZMS 1 series map sheet number and grid reference in parenthesis are as follows:

  • Site 1 — Botanical Gardens, Wellington, approximately 100m from the Meterological Office, Kelburn (41°17′S 174°46′E, N164 327221). An area of approximately 4m x 10m containing several hundred burrows was searched, part only of a more extensive colony in a gully facing NNE.

  • Site 2 — 50m north of the upper end of Cheviot Road, Lowry Bay (41°15′S 174°55′E, N164 463261). The area of 5.4m x 9.0m searched contained several hundred burrows on an ESE facing slope.

    page 24
  • Site 3 — on the Kowhai Street-Butterfly Creek Track, Eastbourne, 60m north of the junction with the Muritai Park Track, (41°19′S 174°45′E, N164 447191). A wedge shaped area 4.6m x 3.5m about a slight trickle in a gully facing east. Less than 100 burrows present.

  • Site 4 — on the Kowhai Street-Butterfly Creek Track 49m south of the Muritai Park Track (41°19′S 174°45′E, N 164, 448189). A more or less equilateral wedge 11m long on the slopes of a small stream facing SSE. Several hundred burrows present.

  • Site 5 — on the Kowhai Street- Butterfly Creek Track 22m north of Buttterfly Creek (41°19′S 174°45′E, N164 448187). A steep bank 0.5m high and 2.7m long facing east. Only 11 burrows over 10mm wide counted.

  • Site 6 — on the true right bank of Gollans Stream 102m downstream from its junction with Butterfly Creek and extending 13m further south (41°19′S 174°45′E, N164 447186). An area about 5m wide on an east facing slope. Several hundred burrows.

Duration and Pattern of Emergence

Each of the sites was searched at intervals of a few days to a week from early November to 28 December 1979 (site 1) and to 2 January 1980 (sites 2-6) without final-instar exuviae being found. All sites were searched on 7 January 1980 and final-instar exuviae were found that day at sites 1 and 6. Thereafter site 1 was searched on January 15, 18, 22, 24, 29, February 4, 7, 12, 14, 19, 21, 26 and March 4, 11 and 18, and sites 2-6 on January 9, 11, 14, 21, 23, 25, 28, February 1, 4, 8, 11, 18, 27, March 5, 12, 19. The maximum and minimum emergence period for each site is shown in Table 1. The total numbers of exuviae collected and their distribution between the sites are shown in Table 2. The cumulative emergence figures for each site are graphed separately in Figure 1, and for all sites combined in Figure 2.

Fifty percent of the combined populations had emerged (EM 50) by the 16th day after emergence was detected. Peak emergence occurred on the 18th day, and the observed duration of emergence (EM 100) was 58 days. Emergence commenced and finished much later than expected: I observed my first two U. c. carovei for the season on 21 November 1979, along the south Wellington Coast, (41°21′S 174°41′E, N164 257147). I saw the first adults active at Gollans Stream on 28 January 1980 and the last record I have for the season near Wellington is a sighting by G. W. Gibbs (pers. comm.) on 5 April 1980 at Field Hut, Southern Tararua Range (40°54′S 175°15′S, N157 776682) at approximately 474m altitude. The flight season in some years may be even longer than this: in the Orongorongo Valley (41°21′S 174°59′E, N164 717148), 7.2km east of sites 2-6, the first U. c. carovei for the season was seen on 30 October 1972 and the last on 12 April 1973 (M. J. Meads, pers. comm.).

The duration of emergence is similar to that determined for U. c. chiltoni by Wolfe (1953) at Cass. Emergence commenced there sometime in the first week of December 1948 and the last exuviae was collected on 2 February, 1949, thus emergence spanned a maximum of 62 days. Peak emergence occurred in the second and third weeks of December and only scattered emergences thereafter. Wolfe stated that emergence in North Island localities (he cited North Auckland and Coromandel Peninsula) was page 25
Fig. 1: Graph of the cumulative emergence pattern at sites at 1, 2, 4, 5 and 6. Heavy triangle indicates EM 50 point.

Fig. 1: Graph of the cumulative emergence pattern at sites at 1, 2, 4, 5 and 6. Heavy triangle indicates EM 50 point.

generally two to three weeks in advance of the southern localities. Wolfe found that emergence was delayed at a cold, south facing site near Cass. Presumably the forest shade mediates temperature and delays emergence in the sites I have studied though I have no temperature records for larval burrows. Emergence in U. carovei is protracted when compared with other petalurids. In Japan, Taketo (1960) found that with a natural population of T. pryeri the EM 50 was eight days and EM 100 13 days, but a laboratory population had an EM 50 of 14 days and EM 100 of 60 days. T. hageni was found to emerge from early July 1958 to the end of that month (Svihla 1959) and from the end of July to the end of August 1959 (Svihla 1960) at Tipsoo Lake, Washington. In contrast, Meyer and Clement (1978) reported an emergence period of at least five weeks, commencing in late May 1976, for this species in California. page 26
Fig. 2: Graph of the cumulative emergence pattern at sites 1-6 combined, with male and female emergence shown separately. Heavy triangle indicates EM 50 point.

Fig. 2: Graph of the cumulative emergence pattern at sites 1-6 combined, with male and female emergence shown separately. Heavy triangle indicates EM 50 point.

There is no statistically significant difference in the mean emergence dates for males and females in the combined records of all sites. Nor is there a statistically significant departure from a 1: 1 ratio of males against females but the slight preponderance of males (54%) is of interest. Many studies — reviewed by Corbet (1962) and Lawton (1972) — have shown a page 27
TABLE 1: Maximum and minimum emergence periods at each site
FirstLastEmergence Period
17 Jan14 Feb29/12-14/2487/1-13/238
211 Jan18 Feb8/1-18/24211/1-12/233
327 Feb27 Feb11
49 Jan27 Feb8/1-27/2519/1-19/242
59 Jan28 Jan7/1-28/1229/1-26/118
67 Jan5 Mar3/1-5/3637/1-28/253
TABLE 2: Total exuviae collected at each site and male percentage of each population
SiteMaleFemaleTotalPercentage Males
consistent excess of male larvae over females in the damselflies (S.O. Zygoptera) and an equally consistent excess of females over males in the Anisoptera when large samples, more than 100 specimens, are considered. In small samples, Svihla (1960) collected at one emergence site 11 male and nine female exuviae for T. hageni (55% males) and S. W. Dunkle (pers. comm.) in a study of T. thoreyi emergence found eight male and five female exuviae (61.50% males). In their study of T. hageni, Meyer and Clement (1978) obtained 39 male and 42 female exuviae (48% males). Taketo (1960) found 61 male and 62 female exuviae (49.6% males) in his study of T. pryeri emergence. Male exuviae represented 47% of the total in U. c. chiltoni (Wolfe 1953). The proportion of males at emergence in the Petaluridae is higher than is generally seen in other families in the Anisoptera.

It is interesting to speculate on possible explanations for the high level of emergence at site 5, in relation to the size of the colony, and the low level at site 3. Those from site 5 probably represent a single age cohort and could be the progeny of one female. The solitary emergence from site 3 occurred late — the female exuviae was found on 27 February — and might have resulted from precocious development. If there were an element of fidelity on the part of females to the sites of their own larval development (philopatry), a level of synchrony would persist in each colony for a considerable time in a species such as U. c. carovei in which larval development may span more than four years.

Time of Emergence

Emergence in P. gigantea occurs about dawn (Tillyard 1917) as it does in U. c. chiltoni (Wolfe 1953). T. pryeri also emerges in the early morning (Eda 1959, Taketo 1960) and T. hageni emerges from early morning to early afternoon (Svihla 1960). S. W. Dunkle (pers. comm.) found T. thoreyi to emerge from mid to late morning.

The duration of emergence in individual specimens varies with different odonate species and with different environmental conditions but generally ranges from 40 minutes to four hours (Eda 1959). In the late stages of page 28 emergence the teneral adult hangs with the body fully formed and the wings held together erect above the thorax. The wings are finally extended laterally and within a short time, if conditions are favourable, the dragonfly makes its maiden flight. Emergence takes about 3 1/2 hours in the petalurid T. pryeri (Eda 1959) and about 2 1/2 hours in T. hageni (Svihla 1960); one might expect U. carovei to require a similar length of time for its emergence. Each species usually has a particular time of day at which emergence occurs, dusk and dawn being two phenomena with which it is often correlated. Inclement weather may interfere with the typical pattern so that some members of an emergence group will complete emergence at the normal time whilst others will not do so until later in the day when conditions are more favourable: this phenomenon has been described as “divided emergence” (Corbet 1962). The time of emergence was not determined in this study but Wolfe (1953) stated that emergence in the genus Uropetala occurs about dawn. At site five on 21 January, 1980, I found one male specimen close to its exuviae with its wings still in the vertical position six hours 45 minutes after dawn and 25 minutes before the solar noon. If a typical dawn emergence holds good for Uropetala throughout its range, this record represents an example of divided emergence.

The teneral adult observed with its wings erect on 21 January 1980 had its full adult colouring — blackish-brown and a vivid buttercup yellow. P. gigantea also achieves full colour before flight (Tillyard 1917) in contrast with T. hageni which requires a further post-flight period for adult colour maturation (Svihla 1960).

Emergence Supports

Of the 105 exuviae recovered, I am aware that my searching activities displaced five exuviae from the vegetation before I noticed them, and 36 others were also found displaced. All 64 attached exuviae were on vegetation: exposed roots, tree trunks, tree fern caudices, fern stripes, rachides and laminae, sapling stems, small twigs and branches, petioles, leaf surfaces and stout lianes. All other petalurids studied have been shown to select a wide variety of vegetation for their emergence supports (Tillyard 1909, 1911, 1917; Svihla 1960; Watson 1958; S. W. Dunkle pers. comm.), and T. pryeri shares a penchant for leaf surfaces (Taketo 1960) with U. c. carovei.

All exuvae were found within an estimated horizontal distance of 0.5m from a burrow but the burrows from which emergence occurred were not determined. Two exuviae were found 0.7m above ground level but most were below 0.3m and some almost touching the ground. Other petalurids vary in the height at which they emerge: T. hageni will emerge whilst the distal segments of the abdomen are still in water (Svihla 1960); T. thoreyi has a mean emergence height of 0.6m and a range of 0.2-1.4m (S.W. Dunkle pers. comm.); P. hesperia may climb as much as 15 feet (4.5m) (Watson 1958).

U. carovei larvae at emergence are much larger than those of other species in New Zealand. Green (1977) reported a mean weight of 2.68g for final-instar larvae and I obtained a weight of 2.99g for one surface-dried female cleaned of mud. A striking feature in this study was the very frail supports such a large larva would use for emergence which is presumably related to its sheltered habitat within the forest. page 29
Fig. 3: The emergence stance in Anisoptera: (a) — “upright” type (Tanypteryx pryeri); (b) — “hanging” type (Procordulia grayi); (c) — “upright” emergence angle 0° to 90°; (d) — “Hanging” emergence angle 90° to 180°. 3(a), (c), (d) after Eda (1959).

Fig. 3: The emergence stance in Anisoptera: (a) — “upright” type (Tanypteryx pryeri); (b) — “hanging” type (Procordulia grayi); (c) — “upright” emergence angle 0° to 90°; (d) — “Hanging” emergence angle 90° to 180°. 3(a), (c), (d) after Eda (1959).

Emergence Stance

Four anisopteran families are represented in New Zealand: Aeshnidae, Corduliidae, Libellulidae and Petaluridae. The emergence stance in the first three families is similar and certain generalisations can be made about it: the larva chooses a site where it can support itself at an angle between 90° and 180° (Figures 3b and 3d); a firm grip must be obtained by the tarsi on each side of the body, that is, the grip must have radial or lateral integrity otherwise the contortions of the emerging adult will lead to displacement and failure; on a slender support, the limbs may overlap behind the support; and finally, if any tarsi are not attached to the substrate at emergence, it is usually those of the forelimbs. There are subtle differences between the Aeshnidae-Corduliidae-Libellulidae pattern and that of U. c. carovei.

In the 64 attached U. c carovei exuviae examined the stance was always vertical irrespective of the orientation of the emergence support. The tip of the abdomen was usually curved dorsally. The limbs never crossed behind the support and were more often than not held spreadeagled. Sixteen exuviae (25%) were found to be attached by the tarsi on one side of the body only. A closer examination showed that the tarsi and tibial spurs are used antagonistically to achieve a vice-like grip, the mechanical principle being similar to that applied in a cant-hook (Figure 4d). Any pressure applied to the tibiae will increase the tenacity of the grip. This type of grip has not been reported in other petalurids but each species has well-developed tibial spurs which would make it possible. The illustrations of page 30
Fig. 4: Tibiae and tarsi of the left limbs in Uropetala carovei carovei final-instar larvae. (a) — prothoracic; (b) — mesothoracic and (c) — metathoracic legs. Posterior lateral view shown on the left and ventral view on the right. There is a notable similarity between these structures and a cant-hook (d).

Fig. 4: Tibiae and tarsi of the left limbs in Uropetala carovei carovei final-instar larvae. (a) — prothoracic; (b) — mesothoracic and (c) — metathoracic legs. Posterior lateral view shown on the left and ventral view on the right. There is a notable similarity between these structures and a cant-hook (d).

page 31
TABLE 3: Tarsal attachment to the emergence support in a sample of 32 exuviae
Eda (1959, 1964) and Taketo (1971b) appear to support its occurence in T. pryeri. The tibio-tarsal joint also has elastic properties which would strengthen the grip of the tarsal claws: the elastic properties of the joint are demonstrated in dead larvae and those feigning death in each of which muscular relaxation causes the tarsi to flex away from the substrate.

For 32 exuviae, detailed records were kept of which tarsi were attached to the substrate. The results are summarised in Table 3. The rear tarsi were the least frequently attached. In 7 exuviae (22%) the hind tarsi were attached each side, in 11 (34%) they were attached on one side only and in 14 (44%) they were free from the substrate. A chi-squared test confirms a significant difference between the attachment of the hind and other limbs (chi-square equals 8.799 p less than .05%). Taketo (1971 b) has illustrated a T. pryeri exuviae with the hind tarsi below and free from the emergence support. The anterior tibial spurs are more developed than the posterior spurs on the prothoracic and mesothoracic legs which are held prograde during emergence. In contrast, the interior and posterior spurs are equally developed on the metatibiae which are usually held laterigrade or retrograde (Figure 4).

In all 105 exuviae examined, the tips of the wing cases had been widely separated by the withdrawal of the adult wings. Wolfe (1953) stated that, other than in points he enumerated, emergence in U. carovei did not differ from that described for P. gigantea by Tillyard (1917). Tillyard, however, illustrates the exuvial wing cases being held parallel throughout emergence, which is exceptional in a petalurid. In his earlier description of the exuviae in this species (Tillyard 1909), he also described the wing cases as lying parallel along the back of the abdomen but remarked that in some exuviae they are turned in all directions. In various collections I have examined exuviae of P. gigantea, P. hesperia, T. thoreyi, T. hageni, T. pryeri, P. raptor and both U. carovei subspecies: invariably, the wing cases were widely separated distally. J.A.L. Watson (pers. comm.) has confirmed that all the petalurid exuviae in the C.S.I.R.O. collection at Canberra — including P. ingentissima, the larva of which has not yet been described—have the wing cases separated also.

Eda (1959, 1963, 1964) observed that two different emergence patterns occur in the Odonata, the “upright” type and the “hanging” type (Figure 3). In the upright type, the head and thorax are withdrawn from the exuviae and, during the recess or resting period which follows, the head and thorax are held more or less parallel to the exuvial head and thorax. In the hanging type, the head and thorax are held at an obtuse angle from the exuvial head and thorax during the resting period. The descriptions of Tillyard (1917) for P. gigantea, Wolfe (1953) for U. c. chiltoni, and Svihla (1960) for T. hageni suggest that emergence in these species is of the hanging type. On the other hand, Eda (1959) has shown with a series of photographs that P. pryeri has an upright emergence stance and S. W. Dunkle (pers. comm.) has found that T. thoreyi also follows this pattern. It is otherwise unknown for species in the same genus, or for those in the same family, to have such radically different emergence behaviours and I page 32 share with Professor Eda and S. W. Dunkle (pers. comm.) the view that emergence in P. gigantea, T. hageni and U. carovei warrants re-examination.

Emergence Mortality

Emergence had not been completed in only 4 (3.8%) of the 105 exuviae collected. One of the four had been displaced, which might explain the failure, otherwise the cause of mortality was not established. Two of the attached specimens had ants feeding on them but there is no evidence, such as Kiauta (1971) found in Aeshna juncea (L.), to implicate the ants as predators.

The wings of a predated teneral specimen were found at site 2 on 8 February, 1980, but the predator was not identified. Removing wings from the prey is a trait of small passerines and Corbet (1957) found blackbirds (Turdus merula L.) a common predator on Anax imperator Leach. Blackbirds were heard frequently at each of my study sites and one was observed searching on the ground at site 4. Wolfe (1953) and Winstanley and Rowe (1980) have listed some of the birds which prey on U. carovei adults. Kingfishers (Halcyon sancta vagans (Lesson)) may be the main avian predators of U. carovei adults in the forest (B. M. Fitzgerald, pers. comm.). Prey remains collected from three kingfisher nests in the Orongorongo Valley (41°21′S 174°59′E, N164 717148) by M. J. Meads and B. M. Fitzgerald contained remains of many Uropetala. However movie films taken by G. J. H. Moon (pers. comm.) show that a kingfisher swallows aeshnid and corduliid adults without removing the wings hence it could not be implicated as the predator at site 2.

The morepork (Ninox novaeseelandiae (Gmelin)) is another potential predator: Moon (1957, 1967, pers. comm.) has shown that it may catch corduliid dragonflies in the forest at night. Fitzgerald and Karl (1979) established that feral cats (Felis catus L.) catch and eat U. carovei adults, and three other observers have presented me with information about cat predation on the species. Cannings (1978) has reported cat predation on T. hageni.

Rectal Air Breathing in Larvae

Dr F. Ris observed that the rectal plates in P. gigantea exuviae were open leading him to suspect that the larvae were not aquatic, but air breathers (Tillyard 1911). Tillyard confirmed aquatic respiration in the larva but suggested that final-instar larvae were able to breathe air through the rectal chamber when above water. Wolfe (1953) after examining 2000 exuviae of U. c. chiltoni in which he found the rectal plates closed, stated that air breathing could only be spiracular in the final-instar larva near emergence. Svihla (1960) determined that the rectal plates are open in exuviae of T. hageni, T. pryeri and T. thoreyi. Taketo (1971a) has shown that the rectal plates are in fact held together in T. pryeri exuviae but their truncated tips form a tube which leaves the rectal chamber permanently open. J. A. L. Watson (pers. comm.) has intimated that the rectal plates are open in this sense in P. hesperia and P. ingentissima and two exuviae of P. raptor which I examined in the British Museum (Natural History) had a similar opening at the tip. In a sample of 46 U. c. carovei exuviae in my collection, 8 (17%) had the rectal plates apart and all were open at the tip (Figure 5). All final-instar exuviae I have examined of each U. carovel subspecies have had an opening at the tip of the rectal plates and I must conclude that Wolfe (1953) was mistaken in his pronouncement. page 33
Fig. 5: The terminal abdominal segments of male (left) and female (right) exuviae. In the male the three rectal plates — the ventral paraprocts (P) and the dorsal epiproct (M) — are widely separated. The shape and size of the cerci (C) provide a simple means of sexing exuviae.

Fig. 5: The terminal abdominal segments of male (left) and female (right) exuviae. In the male the three rectal plates — the ventral paraprocts (P) and the dorsal epiproct (M) — are widely separated. The shape and size of the cerci (C) provide a simple means of sexing exuviae.

Svihla (1959) described rectal air breathing in T. hageni larvae and Taketo (1971a) has done the same for T. pryeri. Final-instar U. c. carovei placed in shallow water bend the tip of the abdomen dorsally placing the rectal opening at the surface as does P. gigantea (Tillyard 1911).


Wolfe (1953) described the oviposition behaviour of U. c. chiltoni. The female alights on Schoenus tussock, struggles down amongst the tangled stems to the bog surface, lays a few eggs in one place and then flutters to another tussock to repeat the process. Abrasion against the tussock leaves leads to chronic wing wear.

I have seen oviposition in U. c. carovei on only one occasion but the behaviour differed in detail from that which Wolfe recorded. The female stood directly on the ground, arched the abdomen and moved backwards over the wet ground probing repeatedly with her ovipositor. Throughout oviposition the wings were gently vibrated so that the hind wings touched the ground. Presumably, wing vibration would accelerate abrasion on contact with any rough surface. The oviposition stance is similar to that which Asahina and Eda (1957) illustrate for T. pryeri where the wings also contact the ground. T. hageni lands directly on the ground to oviposit (Svihla 1959) as does T. thoreyi (Williamson 1901, S. W. Dunkle, pers. comm.).

General Observations

Once the location of a colony is known, exuviae are easily found. Mud-coated final instar exuviae which have had time to dry are light coloured and easily detected against vegetaton. Not all exuviae are mud-coated however, which may mean that not all larvae sit covered in silt in their burrows as Wolfe (1953, p.265) has described. Perhaps the mud-covered exuviae are those of larvae which have recently repaired the burrow. Clean page 34 exuviae have been collected after dry weather hence have not simply had the coating sluiced away by rain. Whether the exuviae is clean or mud-covered, the eyes are always clean, bright and shining. In an excavated burrow, the larva is usually detected by the torchlight gleaming on its eyes.

I have examined several hundred larval colonies of U. carovei and a general description of the larval habitat has appeared elsewhere (Winstanley and Rowe 1980). In many colonies, a few larval burrows, perhaps ten per cent, have the burrow opening occluded in some way, usually by tailings ejected from the burrow. Burrow obstructions may be less ephemeral however. I have found burrows with tree roots across the mouth and, in colonies on level ground, leaf and other plant debris may form a deep cover over the burrows. R. J. Rowe (pers. comm.) also observed blockages of the latter type in the U. c. carovei colony at the Dome, near Warkworth (36°21′S 174°38′E). Clearly the larvae cannot leave the burrows to hunt at the surface under such circumstances and it may be that the larva is not only pit-dwelling (as Taketo (1958) described T. pryeri) but also pit-trapping. Perhaps the larva has to rely on its prey tunnelling through the ground or litter to penetrate the burrow. Remarks made by Wolfe (1953, p.265) on the excavation of a winter feeding chamber by U. c. chiltoni larvae appear to support this hypothesis. When spiders, harvestmen or beetles are disturbed at the surface, they will frequently take shelter in an open burrow where they may be preyed upon.

The material ejected from a burrow can accumulate to form a durable, steep-sided cone as much as 10cm or more above the level of the surrounding ground. Emerging from such a structure to feed would pose problems for the larva: the apex rarely provides a large enough perching surface on which the larva can sit, and one imagines that the cone would reduce the numbers of prey which might simply stumble into the burrow opening. Those burrows which open onto wet, vertical faces must also impose problems for surface feeding.

Larvae maintained in the laboratory have proved to be negatively phototactic and positively thigmotactic. Larvae kept in an inclined aquarium with the choice of resting on soil either above or below water level have been observed at night to assume a stance submerged in a corner of the aquarium where they have stood almost vertically with the head at the water surface and the antennae and foretarsi just penetrating the surface film. They also leave the water freely and eventually will construct a burrow, commencing their excavations above water level. Larvae kept in deep circular containers on wet leaf mould usually assume a more or less vertical stance against the side of the container at night also.

Harding (1977) found that light of all the wavelengths she tested modified behaviour in U. c. carovei larvae in some way. I have used a red light for observations at night. The reaction to white light at night is either immobility or a rapid backward retreat with the abdomen thrashing from side to side. This behaviour may be related to the normal mode of locomotion within the burrow. Tillyard (1921) has described the hair tufts on the dorsolateral surfaces of the abdomen in U. carovei larvae and most other petalurid larvae have similar tufts on the abdomen. A side to side movement of the abdomen inside the burrow would cause these projections to contact the burrow walls and would facilitate movement, particularly in the vertical sections. Analogous abdominal structures abound in burrowing or wood boring insect larvae; the Tipulidae, Cicindellidae and Cerambycidae provide good examples.

page 35

Wolfe (1953, p.265) observed that the burrows in Uropetala are of constant diameter throughout their length. I have found the burrow may be oval in cross section near its mouth with the wide axis-corresponding to the horizontal and this applies particularly to burrows opening onto steep faces. The burrows of final-instar larvae may be much shorter than those which Wolfe (1953, p.266) illustrates. In an extreme case, I found one final-instar larva in a simple vertical burrow, slightly expanded at its base, which was only 15 cm long. This particular burrow was not water filled and was no damper than would normally be expected in the root zone. I have found one other intermediate stage larva in a burrow which lacked free water. S. W. Dunkle (pers. comm.) has shown that T. thoreyi is one species which does not burrow and he located final-instar larvae in the wild hiding under leaves in seepage areas either in the water or just above the water-line, thus demonstrating in them a measure of independence from constant immersion.

This paper has refuted several minor points which have appeared to emphasise heterogeneity of behaviour and morphological divergence in the Petaluridae. The likelihood is that further research will reinforce the essential homogeneity of behaviour and adaptation in this ancient group of dragonflies.


I am grateful to all those people acknowledged in the text who have provided me with their personal observations. In particular I am indebted to Professor Shigeo Eda, Matsumoto Dental College, Nagano, Japan, and to Dr Akira Taketo, Kanazawa University, Ishikawa, Japan, for information on T. pryeri and for papers published in Japan on that species. Dr J. A. L. Watson, C.S.I.R.O., Canberra, Australia, provided me with information on Australian petalurids and Dr S. W. Dunkle, University of Florida, Gainsville, U.S.A., made his unpublished observations on T. thoreyi and other petalurids freely available to me. Dr Dunkle also commented on an early version of this paper as did Dr G. W. Gibbs, Victoria University of Wellington, New Zealand and my wife Christine. Mr and Mrs B. Marshall, Auckland, New Zealand and Mr and Mrs R. N. Greenslade of London, England were my generous hosts in their respective cities. I am grateful to the staff of the Systematics Section, Entomology Division, D.S.I.R., Auckland, New Zealand, and to Mr S. J. Brooks, British Museum (Natural History), London, England, for access to specimens in their care. This research was completed during the tenure of a Junior Lectureship at Victoria University of Wellington. I am grateful to my family for financing my travel. My thanks too to Nesta Black for her cheerful typing services.


ASAHINA, S. and EDA, S. 1957: Oviposition of Tanypteryx pryeri (Odonata, Petaluridae). Kontyu, Tokyo 25: 1.

CANNINGS, R. A. 1978: The distribution of Tanypteryx hageni (Odonata: Petaluridae) in British Columbia. J. Entomol. Soc. B. C. 75: 18-19.

CORBET, P. S. 1957: The life history of the Emperor Dragonfly Anax imperator Leach (Odonata: Aeshnidae). J. Anim. Ecol. 26: 1-69.

—— 1962: A biology of dragonflies. Witherby, London: 247 pp.

DEACON, K. J. G. 1978: Seasonality in New Zealand dragonflies. N.Z. Entomol. 6: 359-360.

—— 1979: The seasonality of four Odonata species from mid-Canterbury, South Island, New Zealand. Unpubl. Ph D. Thesis, University of Canterbury, New Zealand: 209 pp.

EDA, S. 1959. Observations on the emergence of Tanypteryx pryeri Selys. Tombo, Tokyo 2: 18-24.

—— 1963: Emergence of Epiophlebia superstes Selys. Ibid 6: 2-7.

—— 1964: -The comparative studies on the emergence and oviposition behaviour of dragonflies. Insect Journal (Tokyo) 1: 91-103.

FERNET, L. and PILON, J-G. 1968: Une premiere mention oubliée pour le Québec et pour le Canada. Nat. Canad. 95: 1033.

FITZGERALD, B. M. and KARL, B. J. 1979: Foods of feral house cats (Felis catus L.) in forest of the Orongorongo Valley, Wellington. N.Z.J. Zool 6: 107-126.

GREEN, L. F. B. 1977: Aspects of the respiratory and excretory physiology of the nymph of Uropetala carovei (Odonata: Petaluridae). Ibid. 4: 39-43.

HARDING, D. 1977: A consideration of some aspects of the feeding behaviour, along with a morphological investigation of the mouthparts and alimentary canal of Uropetala carovei White (Odonata). Unpubl. B. Sc. Hons. Thesis, Massey University.

KIAUTA, B. 1971: Predation by ants, Formica fusca L. and F. rufa polyctena Bondr, on the emerging dragonfly Aeshna juncea (L.) and its teratological consequences. Tombo, Tokyo 14: 2-5.

LAWTON, J. H. 1972: Sex ratios in Odonata larvae, with particular reference to the Zygoptera, Odonatologica 1: 209-219.

MEYER, R. P. and CLEMENT, S. L. 1978: Studies on the biology of Tanypteryx hageni in California. Ann. ent. Soc. Am. 71: 667-669.

MOON, G. J. H. 1957: Focus on New Zealand birds. Cameo Press, Warkworth: 132 pp.

MOON, G. J. H. 1967: Refocus on New Zealand birds. Reed, Wellington: 143 pp.

SVIHLA, A. 1959: The life history of Tanypteryx hageni (Odonata). Trans. Amer. ent, Soc. 85: 219-232.

—— 1960: Emergence and transformation of Tanypteryx hageni Selys (Odonata). Ent. News. 71: 131-135.

TAKETO, A. 1958: Discovery of the pit-dwelling larva of Tanypteryx pryeri Tombo, Tokyo 1: 21.

—— 1960: Studies on the life history of Tanypteryx pryeri Selys (Odonata: Petaluridae). 1. Observations of adult dragonflies. Kontyu, Tokyo 28: 97-109.

—— 1971(a): Studies on the life history of Tanypteryx pryeriSelys. II. Habitat and habit of the nymph. Ibid 39: 299-310.

—— 1971(b): The odonate-fauna of Ishikawa Prefecture. Rep. Ishikawa Ent. Soc. 3: 1-21.

TILLYARD, R. J. 1909: Studies in the life-histories of Australian Odonata. 1. The life-history of Petalura gigantea Leach. Proc. Linn, Soc. N.S.W. 34: 256-267.

—— 1911: Studies in the life-histories of Australian Odonata. No 4. Further notes on the life-history of Petalura gigantea Leach. Proc. Linn. Soc. N.S.W. 36: 86-96.

—— 1917: The biology of dragonflies. Cambridge University Press: 396 pp.

—— 1921: Description of a new dragonfly belonging to the genus Uropetala Selys. Trans. Proc. N.Z. Inst. 53: 343-346.

WATSON, J. A. L. 1958: A new species of Petalura Leach (Odonata) from Western Australia Proc. R. ent. Soc. London (B) 27: 116-120.

WILLIAMSON, E. B. 1901: On the manner of ovipositon and on the nymph of Tachopteryx thoreyi(Order Odonata). Ent. News 12: 1-3.

WINSTANLEY, W. J. and ROWE, R. J. 1980: The larval habitat of Uropetala carovei carovei (Odonata: Petaluridae) in the North Island of New Zealand, and the geographical limits of the subspecies. N.Z.J. Zool. 7: 127-134.

WOLFE, L. S. 1953: A study of the genus Uropetala Selys (Order Odonata) from New Zealand. Trans. R. Soc. N.Z. 80: 245-275.