A Study of the Genus Uropetala Selys (Order Odonata)
from New Zealand

By L. S. Wolfe, M. Sc., Department of Zoology, Canterbury University College, Christchurch, New Zealand*

The Canterbury University College Mountain Biological Station situated in the Cass-Arthur's Pass area of the Southern Alps of the South Island, New Zealand, provided a convenient base for a detailed study of the life history and ecology of Uropetala chiltoni Tillyard. The present paper is the result of one year's study of the southern species and nine months' general survey of the genus throughout New Zealand and comparison of the southern species with the northern species, Uropetala carovei White. Complete series of larval and imaginal stages were examined from the Cass-Arthur's Pass area and the descriptions of the egg and early larval stages were made from collections in this area in January and February, 1949. Imaginal and nymphal material was collected or sent from many localities in both the North and South Islands. Museum and private collections throughout the country were examined.

I wish to express my thanks to Professor E. Percival for suggesting the present research and for helpful criticism; to Dr. M. A. Lieftinck for the loan of literature and for suggestions; to Mr. R. Forster, Canterbury Museum, for the identification of the Arachnids and to the Cawthron Institute and the Canterbury College Library for assistance in obtaining literature.

Life History and Ecology

Collection records for Uropetala indicate that it has a longer emergence period and flying season than has been recorded for any petalurid. This difference may be due to the incomplete field records so far available for the Petaluridae. It is concluded that Uropetala is a mid-summer dragonfly. The length of imaginal life varies from locality to locality, depending on the degree of exposure and climatic conditions.

The Sex Ratio

Throughout the flying season at Cass males were far more often seen and collected than females. It might be concluded that males are in excess of females, but this field observation is more apparent than real An attempt was made to show statistically the results of collections throughout the season. It is a common phenomenon among many groups of insects, especially the Diptera, that males emerge before females, and it was thought desirable to obtain quantitative evidence for Uropetala For some time there has been argument about the sex ratios in the Odonata (Tillyard, 1905).

The method adopted was exuviae collecting over marked areas commencing mid-December and continuing until the end of January. Every tussock in the marked bog area was examined throughout the period. Winter snow, spring rain and bacterial dissolution removed almost completely the exuviae of the previous season. The collected exuviae were sexed and the enumeration data statistically analysed using the chi-squared test. The individual locality records are not shown here, but only the locality totals. In all cases many more males were collected at the beginning of the season. Table 1 shows the locality totals and the

x²[5] = 5.46 (Probability 35–40 per cent. : not significant)
x²[1] = 5.06 (Probability of 2 per cent. : significant)

calculation of chi-squared for five different areas. Chi-squared for five degrees of freedom is not significant. This means that all the localities are homogenous and can be considered together. Chi-squared for one degree of freedom gives a significant figure meaning that the overall sex ration is uniform and only significantly different at two per cent. from 1:1.

Assuming that there are no differences in different localities and that there is no loss during periods with no collection, the cumulative sex ratios have been worked out for each date and plotted in Figure 1. The percentage of males in December was always greater than in January. When collections were commenced in January instead of December the sex ratios were approximately 1:1. Five per cent. and twenty per cent. confidence limits have been calculated from Mainland's tables (1948) and plotted along with the percentage males on Figure 1. The upper and lower confidence limit lines delimit the confidence belt and it is unlikely that the population percentages will fall outside the five per cent. confidence limits. The graph shows clearly the continuously changing sex ratio over the period exuviae were collected. If collections had been taken earlier than December 17, this drift, it is assumed, would have been more marked.

Fig. 1—Graph of the results of exuviae collections over marked areas during emergence The 5 per cent. and 20 per cent. confidence limits are also plotted.

The sex ratio is 1:1. However it is commonly observed in the field that there are more males than females. Three factors at least contribute to this observation. Firstly, the emergence of males before the females. This factor will be operating at the beginning of the season Tillyard (1917) concluded that the females emerge a few days before the males. Secondly, after copulation the females disappear to their oviposition grounds and are generally well hidden. Thus again in the mid-season the males will more frequently be encountered than the females Thirdly, after oviposition the females are exhausted and badly battered and soon die Because of this, towards the end of the season the males are still predommant and are usually the last to die.

Field Observations on the Copulatory Position and the Function of the Oreillets

Male Petaluridae have characteristic forms of anal appendages. The superior appendages of Uropetala chiltom are broad and leaf-like. The internal margin of the superior appendage has a moderately sharp spur near the base and a row of small spines more distally. The inferior appendage, though greater in length than width, is very robust and widens considerably towards the base. There are two dorsal, highly sclerotized, slightly recurved spurs at the base. (Fig. 2.)

Pairing was observed to take place any time between late December and late January when the females were mature. The ovaries in newly emerged females were very immature, but rapidly matured in five to seven days. Males in the Cass area were generally found basking in the sun on the shingle and rocks bordering most of the bogs. They could be approached slowly from behind and from in front, but never from above, without being disturbed. When shadows crossed them they immediately took to flight If a Uropetala flow above them they commenced rapid pursuit, approaching first from below and then rising until just overtaking from above If a male, there was a quick flutter of wings and the two separated; but if a female, the superior appendages of the male firmly grasped the female by the constriction between the head and prothorax. The broad superior appendages were firmly pressed dorso-laterally on the anterior margin of the pronotum and extended over the mesepisterna. The small spines on the basal internal margin of the superior appendages gripped in a shallow groove between the orbit and the occipital triangle at the rear of the head. The inferior appendage was bent forward at a sharp angle to the superior appendages and gripped on the occiput just posterior to the ocelli and completely covered the occipital triangle. The spurs gripped on to the epicranial region between the eyes just in front of the occipital triangle Tillyard (1909) figured the position of the inferior appendage of Petalura gigantea in cop. In Uropetala the inferior appendage was placed further forward than shown in Tillyard's figure and was very closely apposed to the epicranium. The spurs on the inferior appendage gripped round the anterior face of the yellow occipital prominence which was conspicuously flattened by their pressure In virgin females this occipital triangle is in a rounded and undented condition. Once the male and female were united the pair continued flight in tandem. Attachment other than during flight was never observed.

Two different procedures were observed that were unquestionably concerned with the transference of sperm capsules to the seminal vesicle on the ventral surface of the second abdominal segment of the male before coition. A male was observed on a rock ventrally flexing its abdomen, drawing the ninth sternite to the second sternite Again, a pair in tandem was observed flying very low; the male forwardly flexing its abdomen and attempting to appose its ninth sternite to the second In doing this, the head of the female was pulled forward and raised till it touched the thorax ventrally. When observing this, I was impressed by the strength of the sternal and tergal muscles which must be required for this method of charging the intertergital fossa.

Oviposition

The Petaluridae because they posses well-developed aeschnine-like ovipositors have been classified among the dragonflies that have endophytic oviposition No direct observations have so far been recorded. Tillyard (1909) stated for Petalura gigantea, “the probability is that the eggs are actually inserted in decaying tissues.” Williamson (1901) noted observations on the manner of oviposition of Tachopteryx thoreyi but did not locate the eggs. The oviposition habits of Uropetala chiltoni were carefully observed and the eggs located among the moss. liverwort and other bog vegetation just below the water-level.

After copulation the pair separated and the females moved to the egg-laying grounds, which could be any boggy or swampy situation. There was no migration from selected feeding grounds where they matured and copulated to the egg-laying areas as was observed in the cordulines, e.g. Procordulia smithii White Both males and females had a very restricted habitat; mating and ovipositing in the same bog area they developed in During January the females were found ovipositing in all localities. They alighted on to the Schoenus tussocks. slowly struggled down among the tangled stems to the bog surface, and inserted the last three to four segments of their abdomens into the surface vegetation which usually consisted of matted mosses and sphagna, encrusting liverworts or other water-loving herbaceous plants. The females remained in one position up to two minutes They then withdrew and fluttered clumsily among the tussocks to another locality and repeated the process By the time the females had exhausted their egg supply their wings were extremely battered and in many cases only the anal portion of the wing remained In this condition they could not fly efficiently, but fluttered among the tussocks and soon died. On several occasions females were observed ovipositing in localities suitable for oviposition but where no nymphal holes could be found It was noticed that these localities were always unsuitable for the long nymphal period. They were liable to either flooding or complete drying out, e.g. swamps extending from the slopes on to the flat shingle beds of the Waimakariri River.

Small areas where egg laying had been observed were marked and the vegetation, roots and subterranean stems were carefully dug out in a block and examined in the laboratory. Every blade, stem and decayed piece of vegetation was examined and the earth sieved through graded meshes to fine bolting silk. Batches of six to ten brown, elongated eggs were found, each attached separately to the roots and subterranean stems at approximately one inch below the water-level. No eggs were found inserted in the plant tissues, but they were always found in groups attached to the plant just below the water-level (Figure 3). The curved form of the female ovipositor was admirably suited to placing the eggs in the position found (Figure 4). During the egg-laying the terebra extended and retracted alternately, guiding the egg down between the valvulair and terebral blades and out against the decaying stems or plant roots to which they were attached by a small mass of secretion which coagulated when in contact with water.

Uropetala and probably all the Petaluridae do not have true endophytic oviposition. The Petaluridae probably represent a stage in the evolution of endophytic oviposition. A complete odonate ovipositor is an ancient structure and it is probable that in Carboniferous times when bogs were very extensive the Proto-odonates possessed well-developed ovipositors, but laid their eggs among the decaying stems and leaves in essentially a similar manner to Uropetala. The bog habitat is primitive and is retained by modern Petaluridae The Aeschnidae retained the ovipositor and increased its efficiency for true endophytic egg-laying, whereas in the Gomphidae and Libellulidae reduction took place and the eggs were now dropped or deposited in various fashions directly into water If it is held that the bogs were the primary habitat of the Odonata these divergences are associated with a change from the primary habitat.

The Egg and Pronymph

Uropetala has an elongate egg with a small, rounded prominence at the anterior pole and a slight thickening at the posterior pole.

Average length, 1.38 mm; range, 1.34–1.42 mm.

Average width, 0.54 mm; range, 0.52–0.55 mm.

An extremely fine canal (micropyle) runs into the egg slightly to one side of the cap of the anterior pole A very thin transparent envelope extends over the whole egg. The chorion is sculptured completely with a polygonal and hexagonal pattern. The eggs when first laid are white and opaque, but within three hours are completely brown. The egg production per female ranges from 400 to 600 eggs (Figure 5)

Tillyard (1917) regarded the pronymphal stage as the first instar, the pronymphal sheath representing the first larval instar cuticle. In Uropetala the pronymphal sheath closely invests the body, limbs, mouthparts and anal appendages. Tillyard stated that the pronymph possessed a spine at the anal end that caught against the broken egg shell and anchored the nymph. No spine was observed on the pronymphal sheath of Uropetala (Figure 7). The sheath covered

Fig. 7—The pronymph. The limbs are shown on one side only.

the anal region rather loosely and was projected into three blunt unsclerotized spurs which were attached to the inside of the egg case. After the hatching of the second instar the pronymphal sheath was still attached halfway down the inside of the egg shell. The pronymphal period lasted approximately thirty seconds. The time elapsing between the cutting of the chorion and the rupture of the pronymphal sheath varied between 15 and 60 seconds. The pronymph possessed a conspicuous hatching spine as a sharp sclerotized ridge on the head.

It would be impossible for the pronymphal period to be anything but transitory, since it cannot feed, walk or respire by its branchial gills. A possible function of this stage is to enable unhindered freeing of the nymphs from the eggs when they are deposited in plant tissues, among detritus or surrounded by a thick gelatinous mass.

The Hatching of the Second Instar Nymph and the “Cephalic Heart”

As the time for emergence from the egg approached, movements were observed in the egg. A clear space was seen in the head region between the embryo and the surrounding egg membranes. This is the head vesicle. Waves of contractions passed up the abdomen causing a distension of the thorax and head. Irregular pulsations were seen in the dorsal head region between the eyes and also contractions in the dorsal blood sinus passing up the thorax to the head from the abdomen. The posterior abdominal pulsatile hearts were observed. It was difficult to observe the pulsations in the head region and abdomen simultaneously because the posterior region of the abdomen is sharply flexed ventrally towards the head region at the posterior end of the egg. The course of the blood during circulation was observed by following the movement of the blood corpuscles. The pulsations in the head were observed to be associated with the pulsations of the main aortic heart. There was always a brief time lag between the contraction of the aortic heart and the distension of the head pulsatile organ This organ was called by Tillyard the “cephalic heart”. Besides peristaltic movements in the embryo, slight movements of the mouth parts and limbs were observed. These various movements were all directed to the expansion of the head region, which was soon pressed hard up against the vitelline membrane and chorion. A slight dorsoventral arching movement of the head and thorax resulted in the hatching spine on the head of the embryo cutting through the egg membranes. Just before this the peristaltic contractions and head pulsations increased markedly. The head bulged through the fissure cut by the spine. The fissure extended one third of the length of the egg and then at right angles around the egg. This resulted in a flap of the egg shell being pushed aside as the emergence proceeded. Once the limbs were free from the egg case violent arching movements commenced; the divided labium and limbs arched away from the thorax and abdomen. The nymphal spines and hairs gripped and cut through the pronymphal sheath dorsally in a line from between the eyes to the beginning of the abdomen. The writhing, arching and contractions continued until the nymph had extricated its legs from the pronymphal sheath. The nymph pushed itself free of the sheath by the tibial spines. No pulsations were visible in the head of the nymph as seen in the embryo just before eclosion. Very rapid contractions of the branchial basket commenced and within thirty seconds of the time of release a small bubble of gas appeared in the dorsal tracheal trunks in the mid-gut region and extended posteriorly and anteriorly filling the whole tracheal system within

five seconds. Fine tracheal branches filled simultaneously as the gas extended in the main trunks. The mid-gut of newly emerged nymphs is packed with yolk globules and feeding does not commence for some time. In the laboratory nymphs must be given a rough substratum on to which the tibial spines can grip to free themselves from the pronymphal sheath.

Tillyard (1914, 1916, 1917) described a distinct two-chambered “cephalic heart” in the embryo of Anax papuensis. He regarded this pulsatile organ as having a transitory appearance in the head between the larval mouth and aorta just before hatching. The presence of this “heart” as special organ could not be demonstrated in sections. He described also in the embryo before hatching a distension in the anterior oesophageal region containing blood with, however, “no visible corpuscles”. The head vesicle disappears just before emergence and Tillyard concluded that the liquid in this vesicle was liquid blood minus corpuscles and was withdrawn through the larval mouth into the oesophagus by the action of the “cephalic heart”. Lieftinck (1933) described a pulsating “cephalic heart” in the late embryo of Procordulia artemia.

The following observations have been made on the pulsations in the embryo of Uropetala. The dorsal aorta is attached to the anterior region of the oesophagus and from it blood is forced into the head sinuses. The blood flows to the bases of the antennae and part is directed ventrally around the oesophagus to the labium and backwards to the ventral thoracic spaces and the coxae. Blood is also directed up around the optic ganglia, brain, mushroom gland and back into the neck and laterally into the thorax. Blood flowing into the head haemocoele from the dorsal aorta in the embryo is forced directly against the bases of the antennae, which because of the head flexure are ventral in position. Most of the blood is reflected dorsally into the dorsal sinuses and around the oesophagus back into the thorax. Just before emergence the mouth-parts move and the head swells till it closely apposes the egg membranes. Pulsations can be seen in the head region and suddenly the egg case splits. During the mouth movements the embryo, it is concluded, is swallowing the amniotic fluid in the head vesicle (the space between the embryo and the pronymphal sheath) The fluid accumulates in the oesophagus and is prevented from entering the mid-gut region by the yolky food plug and also by the abdominal peristalses forcing blood into the head region.

The combined effects of abdominal peristalses, increased blood pressure in the head, modified circulation of the blood in the head resulting from the head flexure, swallowing of the amniotic fluid and consequent swelling of the oesophagus and muscular arching movements of the head, create sufficient pressure on the hatching spine for it to cut the vitelline membrane and chorion. No cephalic heart has been seen and it is concluded that the “cephalic heart” is not a definite structure. The pulsations in the head are partially produced by the blood forced into the head sinuses from the dorsal aorta and augmented by the swallowing movements and swelling of the oesophagus. Wigglesworth and Sikes (1931) showed that in all cases investigated the force employed to break open the egg is muscular and that the swallowing of amniotic fluid before hatching is very common among insects.

The Second Instar (Figure 8)

Colour, pale brown. Length, 1 5–1 9 mm. Width abdomen, 0 3 mm.

from the gut separately. The abdominal and thoracic ganglia are very large; those in the thorax being two-thirds of the width of the nymph.

Ontogenetic Changes at the Various Ecdyses

The changes at the various moults in Uropetala do not differ basically from those of other dragonfly nymphs. Apart from increase in total length and width of the head, thorax and abdomen, the more important changes taking place throughout the nymphal period are tabulated below.

1.	The colour changes progressively from a very light brown in the earlier instars to dark brown in the last instars. Nymphs from burrows in light coloured soils are light brown, while those from dark soils are dark brown or almost black.
2.	Antennae are three-segmented in instars 2–3, four-segmented instars 4–6, five-segmented instars 7–11, and six-segmented instars 12–15.
3.	The tarsi are undivided until the third instar and then are divided first into two segments and then three.
4.	A small spine is present on the outer border of the lateral lobe of the labium in all instars and overlaps the large movable spine. (Figures 9 and 10.) The labium of the last instars is slightly concave. (Plate 54.)
5.	Wing cases in the form of triangular flaps develop from the pleural ridges at the eighth to ninth instars. At first, widely separated mesially, the hind Fig. 9—The labium of the second instar nymph. Fig. 10—The labium of the last instar nymph.

Growth of the Nymph

Random samples of nymphs collected monthly from the field and measured did not yield results that could be analysed statistically to determine the growth

Fig. 11—The anal appendages of male last instar nymph.
Fig. 12—The anal appendages of female last instar nymph, showing the dorsal hair tufts. The pattern on segment 7 marks the areas of muscle attachments.

Kennedy (1928) and Walker (1912, 1925) in studies on the Somatochloras and Aeschnids have shown that the higher the latitude the longer is the nymphal period. Somatochlora kennedyi is thought to have a nymphal period of at least four years in its most northern Canadian habitats, which are cool shallow bog and swamp pools. In fact all members of the arctica group have long nymphal periods. Odonates with long nymphal periods generally have more nymphal instars than those that complete their life cycle in a year or less. For the genus Aeschna, Walker regards three years as the probable nymphal period for southern Canadian and northern United States localities and adds at least an additional year for northern Canadian habitats.

The Emergence of the Imago

Tillyard (1917) has described the emergence of Petalura gigantea. Only additional notes on the process in Uropetala will be given here; otherwise it did not differ from Petalura. The alimentary canals in nymphs several months before emergence were completely empty. The most noticeable changes taking place during this period were the withdrawal of the musculature, blood vessels and nerves of the mask, formation of the thoracic wing musculature, swelling of the wing buds, general darkening of colour, extension of the abdomen which became turgid and the appearance of the imaginal colour pattern through the larval cuticle. Before emergence the nymphs were found above the water-level close to the entrance of their burrows. Respiration at this time was spiracular. The branchial gills were sloughed off in the hind gut. Emergence takes place in the early morning around dawn. The nymphs moved some distance from the entrance to their burrows and ascended the bog vegetation, chiefly the Schoenus tussocks. Weather conditions did not affect emergence except that during south-west storms the fully expanded imago did not take to flight until fine weather returned. The pulsations of the meso- and metathoracic pulsatile hearts visible during emergènce rapidly increased from 40 to 60 beats per minute within 20 minutes during the extension and expansion of the wings. It has been shown by Brocher (1917) that these pulsatile hearts are essential to the complete expansion of the wings. Tillyard (1911) stated that the caudal plates of Petalura gigantea were separated in the exuviae and suggested that the larva when about to emerge keeps its rectum open for air breathing. On the examination of two thousand exuviae of Uropetala it was found that in every case the caudal plates were firmly closed. This suggests that when about to emerge respiration is probably entirely spiracular and not rectal.

The Habitat of the Nymphs

The nymphs of Uropetala are found in spring-fed boggy or swampy areas where there is a permanent and consistent flow of water. The tussock-covered faces of the foothills and the main divide of the Southern Alps of the South

	Fig. 14—Plots of summer-winter bog temperatures. ▴—▴ Dec.-Feb. •—• June-Aug. 1 and 3 are the extremes; 2 represents the most frequent gradients. the winter months they do not ascend to the moss surface. The temperature under a thick clump of matted moss was always found to be above freezing.
5.	In calm weather the air layer just above the bog surface between Schoenus pauciflorus tussocks is 3–4°C below the temperature above the tussocks and 100 per cent. saturated. For January, 1949:

Max. temperature in the shade was 32°C.

Max. temperature in the sun was 39°C.

Min. temperature was 1°C.

Small valleys were found often to have temperatures 3–4 degrees above the open country. Many imagos were found flying over the bogs in these areas and the nymphal population was dense.

The Structure of the Nymphal Burrows

Tillyard (1911) described the nymphal burrows of Petalura gigantea These unique burrows are probably constructed by the nymphs of all petalurids. The burrows of Uropetala differed in several details from those described by Tillyard for Petalura. The entrances to the burrows are found most commonly at the base of Schoenus tussocks, in Sphagnum and other mosses, in liverwort, in peaty or clay banks through which water is flowing, among shingle in the bogs or opening at the edge of the bog streams. The burrows occur both in bogs in open country and in bogs running through bush. Near Warkworth, North

manner of oviposition. In these patches the nymphal populations can be very large. (See Table 2.)

A very interesting feature of the burrows is the presence of several very distinct side chambers of different diameters from the main burrow. The diameters of these side burrows is smaller in the upper parts than in the lower. It is concluded that the burrow of the ultimate instar is produced by the progressive enlarging (excavation and cleaning) of the burrows of the earlier instars. The construction of the burrow of the ultimate larval instar is the product of the whole nymphal period.

The Behaviour of the Nymphs

The nymphs of Uropetala move sluggishly. If they are pricked, there is a quick recoiling movement, but if continuously disturbed they contract their limbs and become immobile, feigning death. They react negatively to light and become immobile if prevented from moving into a darker situation. Nymphs from 10 mm.–45 mm. total length have been observed to show a diurnal activity. At night they ascend their burrows and proceed to clean the canals, using their broad, flat masks as shovels. The excavations are scooped out of the burrow and left in a pile at the entrance. The mandibles and lateral lobes of the labium cut any dead tussocks stems or roots. These observations were made on moonlight nights. The nymphs ascend the burrows in a spiral fashion, cleaning and shovelling as they do so. Once at the surface of the burrow above the water-level the nymphs remain still, waiting for nocturnal insects to come within catching distance. Carabid beetles were observed captured by the labium on several occasions. The nymphs never migrate far from the burrow entrances and are quite capable of living long periods out of water in a water-saturated atmosphere. Many aeschnine larvae have been observed to migrate at nights from the lakes on to damp rocks or grass. In Uropetala nymphs the meso- and metastigmata are open in all the later instars, and it is probable that they are functional during the nocturanal periods out of the water.

Enemies and Parasites of the Nymph and Imago

The chief enemy of the nymphs is themselves. Many of the smaller nymphs are eaten by the larger. For the imago the most vulnerable period is at the time of emergence Black capped terns (Sterna albostriata) and seagulls (Larus dominicanus) patrol the bogs, diving and capturing imagos either during emergence or when drying their wings. Imagos are not often captured by trout. The cordulines are the chief dragonflies that the trout capture. The only parasite found was a large gregarine from the midgut of nymphs in the last three instars. They were not found in early instars or in images. (Table 4.)

[The section below cannot be correctly rendered as it contains complex formatting. See the image of the page for a more accurate rendering.]

The Food of the Nymphs and Imagos

The food habits of the earlier instars were found after examination of gut contents to be different from the last instars. The second instar does not feed for a considerable time after hatching. Food for the very young nymphs consisted of ciliates, flagellates, Cladocera, Ostracods, Copepods, small Amphipods, small oligochaete worms, free-living nematodes, rotifers and microtrichopteran larvae. The food of later instars, however, consisted entirely of ground-living Arthropods that could easily have been captured by the nymphs during the nocturnal period out of their burrows. This type of food for dragonfly nymphs is unique. Table 5 shows the results of gut examinations on nymphs of the last seven instars. Many particles of food could not be identified. The table also shows the type of food of the imagos.

[The section below cannot be correctly rendered as it contains complex formatting. See the image of the page for a more accurate rendering.]

Geographical Variations

Tillyard (1921) created a new species of Uropetala, Uropetala chiltoni Tillyard, separate from Uropetala carovei White. He based his argument for the establishment of this new species on six differences. I have re-examined a complete series of Uropetala imagos from both the North and South Islands in museums and private collections. The colour of the labrum was found to be yellow in all specimens examined from the South Island except for several specimens collected in the Nelson and northern Marlborough areas which had the labrum partly yellow but merging into brown and black ventrally. The labrum was completely black in all North Island imagos. The degree of blackening of the frons was very variable, but in extent was slightly greater in North Island imagos (Figure 23). The extent of the dorsal bands on the thorax and the yellow spots on the abdomen varied irregularly even in specimens from one area.

Fig. 23—The frons and labrum of imagos from North and South Islands.
1—Rimutaka R., ♂ (N.I.).
2—Palmerston North, ♂ (N.I.)
3—Hunua, Auckland, ♀ (N.I)
4—Aniseed Valley, Nelson, ♂ (S.I.)
5—Maitai Valley, Nelson, ♂ (S.I.)
6—Awatere Valley, Marlborough. ♂ (S.I.)
7—West Coast, ♂ (S.I.)
8—Cass, Canterbury, ♂ (S.I.)
9—Porter's Pass, Canterbury, ♂ (S.I.)

No difference was found in the colour of the femora of mature imagos in the two species. The differences in the truncation of the tip of the inferior appendages was extremely slight. In all North Island forms, the superior appendages were broader. From scale drawings the superior appendage of U. carovei is 4 sq. mm. greater in area than that of U. chiltoni. In the Nelson district, however, there were intermediates. The average size of the North Island imagos is a little greater than the South Island imagos, but this is not an invariable rule. No difference was observed either in the behaviour or structure of the nymphs.

The differences between the two species is not as distinct as Tillyard thought. The colour of the labrum, size of the superior appendages and average size are the most common differences. These have intermediates, however, in the Nelson area. This intergradation of the North and South Island imagos suggests geographical variation rather than specific difference. I suggest naming the North Island form Uropetala carovei carovei and the southern imagos Uropetala carovei chiltoni. Tillyard's suggestion of an invasion of chiltoni from the west through Arthur's Pass was based on inadequate collection records.

Some Systematic Aspects

(a) History and Synonymy of the Genus Uropetala and Related Forms

The first mention of Uropetala from New Zealand was made by Adam White in “The Zoology of the Voyage of H.M.S. Erebus and Terror … during the years 1839–1843”, published in sections between 1846 and 1874. Pages 1–24 and Plates 1–6 of the section on Insects by White appeared in 1846. Pages 25 on and Plates 8 on appeared in 1874. Plate 6 was White's (1846) original figure of Petalura carovei, collected near Auckland. The corresponding text appeared in 1874. The first description, however, of Petalura carovei was that of J. E. Gray in Dieffenbach's travels, published in 1843, in advance of White's plate. Because Gray expressly states, “Petalura carovei White, n.s.”, it is unquestionable that White is the author of the species. Gray compared this dragonfly with the type of the genus Petalura, viz. gigantea, created by Leach (1815) in the family Aeschnidae. Petalura carovei was included in the Libellulidae by White.

Selys (1854) placed Petalura carovei in his 6th Legion, Petalura, which contained two genera, Petalura Leach and Phenes Rambur (1842). In his monograph (1858) a new sub-genus, Uropetala Selys, appeared containing Uropetala carovei White and Uropetala thoreyi Hagen. The sub-genus Petalura contained Petalura gigantea Leach, and the sub-genus Phenes, Phenes raptor Rambur (figured by Gay, 1854). The sub-genus Uropetala was characterized by 3-segmented antennal stylus, 3-celled discoidal triangle of forewing, and the form of the male anal appendages. In the additions (1859) Uropetala thoreyi Hagen was moved to a new sub-genus Tachopteryx Uhler (Tachopteryx obscura Uhler changed to Tachopteryx thoreyi Selys and Hagen). Tachopteryx thoreyi Hagen was synonynious with Petalura thoreyi Hagen, 1861. In 1879 a new species, Tachopteryx hageni Selys was added to the sub-genus Tachopteryx. Gompus pryeri Selys, discovered in 1883 in Japan, was named Tachopteryx pryeri Selys in further additions to the Gomphines (1889).

Calvert (1893) suggested that Selys' Legion Petalura should represent a distinct subfamily Petalurinae rather than be incorporated in the Gomphinae. Hutton (1899) redescribed Uropetala carovei very briefly. Hudson (1904), in a popular account, unsatisfactorily describes the nymph of Uropetala carovei. A Tachopteryx thoreyi exuvia was described and figured by Williamson in 1901. Tillyard (1907) reviewed the genus Petalura and described a new species, Pelalura ingentissima Tillyard, from Queensland, the largest living dragonfly. He later (1912) added Petalura pulcherrima Tillyard to this genus, also from Queensland He figured (1909) the nymph of Petalura gigantea and in 1911 gave some life history notes Ris (1904, 1910) recognised the Petalurinae as a distinct subfamily of the Aeschnidae. This is the classification held by Tillyard (1917), where the Petalurinae are placed between the Chlorogomphinae and Cordulegastrinae.

B. Median labial lobe denticulate, not triangular, but projecting forward.

Tachopteryx thoreyi

Tanypteryx pryeri

(a) Abdominal hair tufts on tubercles. Labium as wide as long. 7-segmented antennae.

Tachopteryx thoreyi

(b) Abdominal hair tufts not on tubercles Labium wider than long. 6-segmented antennae.

Tanypteryx pryeri

Petalura gigantea

Phenes raptor

A. Median lobe of labium triangular; mentum longer than wide; prothorax not ridged; no hairy tubercles on abdomen; superior appendages as rounded prominences.

Petalura gigantea

B. Median lobe of labium not triangular, but projecting forward; mentum wider than long; ridged prothorax with tubercles, hairy tubercles on abdomen; superior anal appendages with forked appearance.

Phenes raptor

Until the nymph of Tanypferyx hageni is discovered and described little can be said of the relationships of the various species. The nymphal characters, however, do not indicate that Tanypteryx pryeri should be placed in a different subfamily. (Fraser, 1933.)

(c) Imaginal Characters

Adequate descriptions of the imagos of Uropetala have been recorded by Fraser, 1933, and Tillyard, 1921.

The Spermatocyte Chromosomes

Testes of newly emerged males and last instar nymphs were fixed in Fleming's solution and stained either with Heidenhain's haematoxylin or with crystal violet. Spermatogonial divisions were found in last instar nymphs and spermatocyte divisions and spermateleosis stages in teneral males. The primary spermatocyte chromosome number of Uropetala carovei is nine made up of seven large autosomes, one medium-sized sex chromosome, and one very small microsome. Kichijo (1939) found exactly the same number in Tanypteryx pryeri. This is the lowest so far recorded for any dragonfly. Asahina and Makino (1935) showed that the haploid number for Libellulidae is generally 13 and for the Aeschnidae and Gomphidae it varied from 10–13. It is probable that the haploid number of 7 + X + m as found in petalurids is the basic odonate chromosome number, the numbers in other species being derived from this by fragmentation of the autosomes and/or the fusion of the small microsomes with the autosomes.

References

Asahina. S., and Okumura, T., 1949. Nymph of Tanypteryx pryeri. Mushi, 19, (7) 37, 1 fig.

Asana, J. J., and Makino, S., 1935. Chromosomes of Indian dragonflies. J. Fac. Sc. Hokk. Imp. Univ., Ser 6, 4 (2), 67–86.

Brocher, F., 1916. Circulation in Dytiscus. Arch. Zool. 55, 347–373.

Brocher, F., 1917. Ibid, 56, 347–358.

Brocher, F., 1917. Circulation in Odonata. Ibid., 56, 445–490.

Calvert, P. P., 1893. Introduction to the study of Odonata. Trans, N.Y. Ent. Soc., 2, 219–272.

Calvert, P. P, 1895 Odonata of New York State. J. N.Y. Ent. Soc, 3

Claude-Joseph, H., 1929. Observaciones sobre el Phenes raptor. Rev. Chile Hist. Nat., 32, 8–10.

Dieffenbach, E., 1843. Travels in New Zealand, vol. 2. 281–282.

Fox, M., and Wingfield, 1938. Micro Winkler method. J.E.B., 15, 437.