External and Internal Anatomy Last Nymphal Instar.

first contributes one branch to the mesothoracic leg, the second one to this appendage and a posterior branch to the metathoracic leg; whilst the last branch of this latter is derived from the trunk to the first tracheal gill. The posterior stem of the first spiracular trunk gives off a branch to the forewing-pad, the costo-radial, which supplies the following veins:—C. Sc. R. and M. together with their cross-veins.

The anterior stem of the second spiracular trunk gives off the cubito-anal trachea to the forewing-pad, and supplies the Cu and A. veins of this structure. In this way the wing-pad receives tracheae from two sources, each in direct communication with a spiracle, and hence is assured of a rich oxygen supply.

In a like manner the hindwing-pad is supplied from two sources, the costo-radial arising from the posterior stem of the second spiracular trunk, and the cubito-anal from the posterior leg stem which arises from the trunk to the first tracheal gill.

As has been stated by Comstock (1918), etc., there is no connection between the radio-costal and cubito-anal tracheae, and the median is still in the primitive condition attached to the former. There is no fusion of the two wing tracheae as found by Tillyard (1917) in Dragonflies, where he states that this condition is the combined result of three interacting forces, one of which being that of relying almost wholly on an oxygen supply coming from the anal end of the alar trunk, owing to the larval gills being situated in this region. Another authority has disputed this point, and it must be stated here that, on the one hand, no support for Tillyard's view can be deduced from the condition of the alar trunks, whilst, on the other, there is the reduction of the costal vein which Tillyard uses to support his theory. This latter vein is lost in the adult, resembling the condition Tillyard found in Dragonflies.

The Wing-Pads.

These are about 0.3 cm. in length, lie with their dorsal surfaces uppermost, and project postero-laterally on the tergum.

The Forewing Pad (Pl. 38, Fig. 26) has a thick margin well provided with large, stout macrotrichia, and shows quite clearly the fusion of the veins at their bases to form the costo-radial and cubitoanal tracheae, which according to Tillyard (1921) are characteristic of the order as a whole.

The costo-radial trachea on entering the wing-pad splits into three, the anterior branch forming Sc. which runs two-thirds of the length of the pad where it bifurcates into Sc.₁ and Sc.₂, the latter of which bends down and almost touches R, or else fuses with it for a short distance, afterwards curving back to the margin of the wing. Just after Sc. leaves the costo-radial a very small costa, which ends before the margin is reached, may or may not be given off. The humeral veinlet is the next branch and is constant in position; following the humeral are a variable number, usually 3 to 4, of cross-veins across the costal space between Sc. and the margin. These are provided with macrotrichia, and there is always a large

space between the first of these and the humeral veinlet. The second branch constitutes R, which runs some distance before branching into a straight, unbranched R₁, and into Rs. This latter branches into R₂ + ₃ and R₄ + ₅, the former again bifurcating into R₂ and R₃. The third branch of the costo-radial is M, which runs some distance before bifurcating into M₁ + ₂ and M₃ + ₄.

A small cross-vein between M and Cu₁ is regarded by Tillyard (1919), as M₅ a basal concave vein connecting M with Cu₁, based on the study of the fossil Belmontia. It is a characteristic feature of the orders forming the Panorpoid complex. In 1925, after further study, he decided that it was merely a convex strut which has secondarily developed providing rigidity to the base of Cu₁.

In 1921, this author figures the tracheation of Stenoperla prasina, but shows no trachea between M and Cu₁. In 1923, he figures the venation, naming M₅ as a small cross-vein between M and Cu₁, but gives no reason for doing so. Now this author (1919) also holds that only in special cases do cross-veins bear macrotrichia and that these are more characteristic of main-veins. An examination of the wing-pads and venation by the present author revealed the fact that macrotrichia occur on many cross-veins, especially those between M and Cu₁, including Tillyard's M₅, and between Cu₁ and Cu₂. A similar condition was found to exist in the venation of Austroperla cyrene and two species of Zealandoperla.

Now, main-veins are preceded by tracheae except in orders where the venation is highly specialised as in Hymenoptera, Mecoptera, Diptera, and Trichoptera, so that if these veins possess tracheae in the wing-pads we might expect them to be rudiments of the archedictyon and so explain the anomaly. If not, then they must be merely cross-veins, and the presence or absence of macrotrichia is not an infallible criterion for main-veins.

The present author examined a large number of wing-pads of S. prasina and found that not one of these veins, including Tillyard's M₅, is preceded by tracheae (Pl. 35, Fig. 18). The wing-pads of Zealandoperla sp. and Austroperla cyrene were also examined, and again no tracheae were found. From these facts it is concluded that in stone-flies M is primitively four-branched.

Again, Comstock's M is a concave vein, and hence must be homologous with Lameere's (1922) MP, which is primitively four-branched. MA likewise is four-branched and, being the anterior branch of M, is convex. According to both these authors then, MP is four-branched, and according to Lameere so also is MA, and as stone-flies are primitive, one would expect a similar condition to exist in this order.

Sections of the wings of S. prasina and Megaleptoperla grandis were cut, revealing the fact that M was convex, hence it must be homologous with Lameere's MA and not with Comstock's M. After studying the Proto-perlaria, Tillyard (1928) found that MP was already becoming obsolete basally, and that in true Perlaria has entirely disappeared. This it is seen is in agreement with the condition in recent forms and resembles the condition in Odonata.

so that the lumen is almost obliterated. This region is surrounded by a large amount of fat.

The Crop is normal in position, but differs in that it is greatly distended with a gas the composition of which is unknown.

The Gizzard in the imago has increased in length and extends backwards as far as the fourth abdominal segment. Like the crop, it is distended with gas, this being prevented from escape into the mesenteron by the oesophageal valve.

The Mesenteron is shortened correspondingly with the elongation of the gizzard, and now lies in the fifth and sixth abdominal segments. It is much reduced in diameter and is pigmented posteriorly. The parasites referred to in the section on bionomics are extremely common.

The Hindgut is very much reduced in diameter, but displays the same structure as in the nymph. Posteriorly there is a constriction dividing it into ileum and colon-rectum, in the latter of which a trace of faece was detected.

Secretory and Excretory Glands of the Gut.

The Salivary Gland.—The salivary glands resemble those of the nymph in every respect.

The Malpighian Tubules.—These structures resemble those of the nymph except that they are of a darker yellowish brown suggesting a greater deposition of solid urates. Similarly the tubules pass forwards and are attached to the gizzard. They are all well supplied with tracheae, lie in close relation to the gonads, and arise from the mesenteron in two bands, each tubule opening into the latter by a separate opening situated on a small papilla.

The Nervous System.

The nervous system is identical with the nymphal.

The Circulatory System.

This is identical with what has already been described.

The Respiratory System.

The general nymphal respiratory plan is adhered to (Pl. 37, Fig. 24), although slight modifications have taken place to adapt it to a terrestrial mode of life. It has been shown that in the last larval instar the thoracic spiracles were present but closed; now, however, they are functional, and eight functional abdominal spiracles have been added. The main longitudinal trunk and its branches are still present, as are also the typical Y-shaped tracheae to the leg. In each of the first five abdominal segments, branches still pass downwards and outwards to the vestigial gills, a discussion of which has already been given. Each branch again sends a trachea to the gut, the muscles, and fat body, but with this difference that the large branch to the former is now connected with its fellow of the opposite side, to form a ventral commissure not found in the nymph. This condition occurs in the second to fifth abdominal segment, whilst the

first segment has a slightly different arrangement described below. Intersegmentally the longitudinal trunk gives off eight short, spiracular, unbranched tracheae to the first eight abdominal spiracles, except that the spiracular trachea of the first arises by the bifurcation of the tracheal trunk to the first gill, one branch of which passes to this latter, the second to the spiracle, whilst the third forms the posterior stem to the leg. This latter shortly after its origin from the spiracular trunk gives off a branch internally to join a similar branch from the opposite side, forming the first ventral commissure. This commissure is connected to the second on either side by a small longitudinal trachea running parallel and ventral to the main trunk. This again is not found in the nymph, and may or may not be a variation. In the specimen dissected the trunks to the wings were severed, since dissection was performed from the dorsal surface, but the cubito-anal branch to the forewing could be traced, and was found to be normal in position. From this there seems no reason to doubt that the basal tracheae to the legs and wings are the same as for the last nymphal instar. The atria of the thoracic spiracles are now much more marked than in this latter, and it is from these that the basal tracheae to the wings arise, the trachea having migrated along the leg stems to these regions.

In the mesothorax the main longitudinal trunk gives off a fairly large branch to the muscles, represented only by a single small branch in the nymph. Together with this are the smaller branches to the muscles, the increased tracheation being correlated with the increased activity of the wing-bearing segments.

The atrium of the first spiracle is deviated forwards and constitutes that portion of the main trunk lying in the pro-thorax. This by bifurcation gives rise to the dorsal and ventral head trunks, the latter of which gives off a large branch to the legs as it passes forwards. The former resembles that of the nymph, except that at the point where it curves outwards in the head, a large trachea is given off. This curves forwards and inwards to meet its fellow from the opposite side, forming a dorsal, anterior commissure between the two longitudinal trunks, a condition which is absent in the nymph.

The development of commissures is no doubt correlated with flight, this rendering it necessary to have direct communication between the trunks, thus making it possible for all parts to be oxygenated more rapidly.

The Reproductive System.

The Male Reproductive Organs.—These consist of a pair of long, convoluted tubes, the testes, extending forwards to the metathorax, where each joins its fellow from the opposite side (Pl. 40, Fig. 35). In a freshly dissected specimen each tube is seen to have an inner, opaque, central portion, and an outer transparent. Each tubule lies more or less dorso-laterally to the gut, and extends backwards. passing imperceptibly into the vas deferens. In the eighth segment

each vas deferens descends and runs backwards to the posterior region of the ninth, where it doubles back on itself, running forwards into a large vesicula seminalis (Pl. 40, Fig. 36). Each of these is joined to its fellow anteriorly, is pear-shaped, convex above and below, and opens into the ejaculatory duct by an opening of its own. There are a pair of accessory glands lying immediately beneath the vesicles, each of the former opening into the corresponding vesicle of its side. In shape they are irregular cylinders communicating with one another anteriorly. The accessory gland is revealed to be of the mesodenium type of Escherich.

The ejaculatory sac lies immediately above the hypandrium, and appears to have a portion of the median unpaired aedeagus together with the ductus ejaculatorius retracted within it.

The Female Reproductive System.—The ovaries are of the panoistic type in which no nutritive cells are present; they are two in number and extend backwards from the first abdominal segment to the eighth, lying dorso-laterally to the gut, which is much reduced in size. Each ovary has a long, straight tube, the oviduct, from which on either side are given off a large number of ovarioles (Pl. 40. Fig. 37), which increase in length from before backwards until a maximum is reached. Anteriorly on the mid-dorsal line each ovary joins its fellow. Each ovariole may be divided into three regions:—

• (1) the terminal filament;

• (2) the germarium;

• (3) the vitellarium.

The Terminal Filament is a minute, slender prolongation of the peritoneum which invests each ovariole. All the filaments of an ovary are bound together, and anteriorly where the ovaries join they all meet and continue forwards as the median or dorsal ligament, which finds attachment on the pericardial diaphragm.

The Germarium (Pl. 40, Fig. 38) in the youngest ovarioles-appears as an undifferentiated portion containing large nuclei with nucleoli. From this region the ova and follicular epithelium are developed. In older ovarioles the germarium shows the ova differentiated in linear series; this condition appears to be present in the ovarioles nearer the gonopore, the more anterior ones seemingly being of a younger nature.

The Vitellarium (Pl. 40, Fig. 37) forms the major portion of the ovariole and has the ova disposed in a linear series, each ovum being enclosed in a definite follicle and surrounded by a layer of follicular cells, which, as the ova pass down the ovariole, secrete the chorion. Meanwhile, before this occurs, the ova increase in size by the accumulation of oil droplets. Each ovum has a large nucleus.

Each oviduct extends backwards so that the ovarioles lie on top, and on the sides of the gut, and in the intersegmental region between the seventh and eighth segments, each duct descends to

At first it was thought that they might be of a different species, but a careful examination has revealed no grounds for regarding them as such. On the contrary, it is contended that they are brachy-pterous males of Stenoperla prasina, supported by the fact that external and internal anatomy conform to the same general plan as found in the nymph and other imagines, and that the external genitalia show no difference whatsoever from the normal male. Moreover, these males were sexually mature and endeavoured to copulate with the female, which, however, very unfortunately died, together with two males, during a very severe frost. Where possible a comparison of their internal anatomy was made with that of normal imagines, and was found to be identical.

The Wings.

The wings (Pl. 42, Fig. 40) are one-quarter of the normal size and lie flat on the terga of their segments, the forewings slightly overlapping the hind; the anal fan of the latter is doubled under the rest of the wing so that its dorsal surface lies ventrally on the tergum. It is slightly shorter than the fore. Both are the same colour, the wings of three specimens are olive green, whilst those of a fourth are purple in colour. When the insect is disturbed, no attempt is made to use these, but escape is effected by their very rapid running. The anterior margin of both is thickened around to the apex, and both are clothed with microtrichia. The venation is extremely variable, the wings of a single individual showing great variations.

The Forewing.—In the forewing C is absent unless the anterior thickening of the wing can be regarded as such, whilst Sc is two-branched as in the normal wing. The humeral veinlet is constant in position, and a variable number of costal veinlets pass from Sc across the costal space. Sc₂ in all cases curves downwards and is fused with R₁ for a short distance. R + M pass outwards some distance and then bifurcates into R and M. It is here that the greatest variation takes place. In the more typical forms R continues towards the margin for a short distance, and then divides into R₁ and Rs; the former is a straight, unbranched vein terminating just anterior to the apex. Rs bifurcates distally into R₂ + ₃ and R₄ + ₅, but in some cases is undivided and passes out as Rs to the wing margin. M divides distally into M₁ + ₂ and M₃ + ₄, but M₁ + ₂ may or may not be fused at its distal extremity with R₄ + ₅. Cu runs some distance along the wing membrane and then splits into Cu₁ and Cu₂ (both of which are usually unbranched). A strong set of cross-veins in all cases is developed between Cu₁ and Cu₂. 1A is straight and unbranched, whilst 2A is two-branched.

The Hindwing.—The hindwing is lacking in C, and Sc is twobranched, Sc₂ fusing with R₁ as before. The humeral veinlet is constant, and there are a variable number of veinlets across the

Newport, G., 1851. Anatomy and Affinities of Pteronarcys regalis, Trans. Linn. Soc., Zoo., pp. 425–452.

Snodgrass, R. E., 1909. The Thorax of Insects and the Articulation of the Wings, Proc. U.S.A. Nat. Mus., Vol. 36, pp. 511–595.

Tillyard, R. J., 1917. Biology of Dragonflies, Camb. Univ. Press, Cambridge.

——, 1919. Panorpoid Complex, Proc. Linn. Soc. N.S.W., Vol. 11, Pt. 3.

——, 1921. Revision of Family Eustheniidae, Proc. Linn. Soc. N.S.W., Vol. 1, Pt. 2.

——, 1923. The Stone-flies of N.Z., Trans. N.Z. Inst., Vol. 54, pp. 197–217.

——, 1925. Amer. Journ. Sci., 9, p. 328.

——, 1926. Insects of Australia and N.Z., Angus and Robertson, Sydney.

——, 1928. Kansas Permian Insects, Part II, Order Protoperlaria, Family Lemnatophoridae (contd.), Amer. Journ. Sci., Vol. 16, pp. 313–348.

Walker, W., 1852. Neuroptera from N.Z., Cat. Brit. Mus., Pt. 2, pp. 206–210.

Wigglesworth, V. B., 1931. The Physiology of Excretion in a Blood-sucking Insect, Journ. Exp. Biol., 8.

Fig. 1 —Sketch of Pitot tube Fig. 2.—Side view basal portion of Pitot tube.

Fig. 3 —Gregarine from mesenteron last nymphal instar X 73

Fig 4.—Tentorium side view Fig. 5.—Pleurites of nymph.

• po.a—posterior arm of tentorium

• d.a.—dorsal arm of tentorium

• a.a.—anterior arm of tentorium

• b.—body of tentorium

• ep—eplmeron

• epi—episternum

• tr.—trochantin

• t.—trochanter

• c.—coxa

• a.c.—antecoxal plate

Fig. 9—Nymphal labrum Fig. 10.—Nymphal mandible. Fig. 11.—Nymphal maxilla. Fig. 12 —Nymphal labium. Fig. 13.—Nymphal hypopharynx.

• ad a.—adductor apodeme

• s.d —salivary duct

• 1 —lacinia

• di g—distigalea

• b.g.—basigalea

• ea.—eustipes

• di.c.—disticardo

• b c.—basi-cardo

Fig 14—Thoracic sternites and tentorium of nymph dorsal aspect. Fig. 15.—Thorax of adult ventral aspect.

Fig. 16—Tibio-tarsus of nymph showing aquatic adaptation. Fig. 17.—Basal leg and wing tracheae of last nymphal instar.

Fig. 18.—R + M and Cu_{1 + 2} showing their tracheae and the absence of a trachea in the cross-vem between.

Fig. 19.—Tibio-tarsus of imago showing the absence of nymphal adaptation and the development of an empodium.

• hw.—hmdwing cu a.—cubito-anal trachea co.r.—costo-radial trachea st.—sternum f.—furca sp.—spinasternite f in.—furcal invagmation v.g.—vestigeal gills

Fig. 20.—Brain and sympathetic nerves ventral aspect. Fig. 21—Dorsal aspect ditto.

• fr g.—frontal ganglion

• p.c.—paroesophageal connective

• a.n.—antennary nerve

• tri c—triterocerebrum

• de.c—deuterocerebrum

• o g.—optic ganglion

• sym.n.—sympathetic nerve

• pro.c.—proterocerebrum

• st g.—stomachic ganglion

Fig. 28—S. prasina imago. X 4. Fig 29.—Male cersus showing bristles. Fig. 30—Male gemtalia ventral aspect. Fig. 31.—Ditto, side view Fig. 32—Female genitalia ventral aspect.

• hy.—hypandrum i.a—inferior appendage sup a.—superior appendage ae.—aedeagus

Fig 33—Alimentary canal of nymph. Fig. 34—Brain and suboesophageal ganglion in situ, side view. Fig 35—Anterior region testes Fig 36—Male accessory organs, ventral aspect. Fig. 37.—An ovariole. Fig. 38.—The germarium.

• oe.—oesophagus

• s.g—salivary gland

• cr.—crop

• r.n.—recurrent nerve

• g—gizzard

• m.—mesenteron

• m.t—malpighian tubules

• i.—ileum

• c.r.—colon-rectum

• ac.g—accessory gland

• v.s—vesicle seminalis

• v.d—vas deferens

• ej c.—ejaculatory sac

• d.e.—ductus ejaculatoris

• d.n.—anteunary nerve

• de.c.—deuterocerebrum

• tri.c.—triterocerebrum

• pro.c.—proterocerebrum

• o.g.—optic ganglion

• svmn.n.—svmpathetic nerve

• st g—stomachic ganglion

• oe.—oesophagus

• su.oe.g.—sub-oesophageal ganglion