The characters of the eggs may be most easily compared if shown in a tabular form :

The greater height of myrica is due to its being a somewhat larger egg than the others and being more often laid singly. Such a specimen was selected for measurement and showed a greater height than could be found in them in an imbricated mass. The heights are given as actually recorded, but my impression is strong that the differences between the several species in this respect is trifling or evanescent.

It is curious that the number of ribs should vary as it does, and especially that the difference should be so great in the most allied species, menyanthidis and myrice having respectively 66 and 50, and rumicis and venosa 54 and 44. This is quite parallel to what obtains between psi and tridens, and probably has some relation to the circumstance that, though these pairs occupy the same areas and emerge at the same times (or nearly so), hybridisation as to none of them has been recorded.

Acronycta (Cuspidia) tridens.-We begin here that section of the genus in which the eggs are laid solitarily, the moth in the wild state probably taking a flight after the deposition of each egg. The egg (Plate VIII., fig. 2) of tridens is nearly colourless, almost glassy when first laid, but acquiring a certain whitish opalescence as the young larva within is developed. Its greater size, and, to a slight degree, the ribbing, render it only slightly less favourable for the observation of the embryonic development than those of Botys hyalinalis, for the opportunity of observing which species I am indebted to Mr. Jeffrey, of Ashford. being solitary takes the very regular form of a portion of a sphere, less than a hemisphere, or roughly, that of a bun. The diameter is 83 mm., and height *38

mm.

The egg

The ribs are 38 in number; in all the Acronyctas this number varies, and the numbers I give are either the average or the actual number of a particular specimen counted. The egg of tridens rarely, if ever, exceeds 44 ribs, that of psi as rarely has less than 45. In colour, or rather want of colour, secondary ribbing, form, etc., they seem to be identical. When about to hatch, the young larva is

very conspicuous inside, the head forining a central black spot.

The

The newly hatched larva (Plate VI., fig. 2) is paler than it shortly becomes, but the head is already black, and the pale and dark segments are clearly pronounced. Each tubercle is a large flat plate, somewhat angulated, so as to fit against and amongst the others; this feature is common to a number of Acronyctas. trapezoidals are thus somewhat pear-shaped, the anterior with the narrow end backwards, the posterior with the narrow end forward, between the anterior trapezoidals and the supra-spiracular. The dorsal tubercles of 3 and 4, being apparently fused trapezoidals, have each two hairs, all the other tubercles have each one long black hair. The plate of the 2nd segment has four hairs on either side. The trapezoidal and supra-spiracular tubercles of 11 are very small, and not so markedly angulated. The trapezoidals of 12 are large, rounded, and the posterior set immediately behind the anterior, again a common arrangement in Acronycta and many other larvæ. The tubercles are fuscous rather than black, the head black. Below the sub-spiracular there is a small ventral or marginal" tubercle in 5, 6, 11, and 12. The colour of the segments, that is the skin of the larva, is reddish brown on the 5th, 8th, 9th, and 12th, pale or whitish on the others. In psi, which is also somewhat larger, the 13th segment belongs to the coloured series.

66

The pupa (Plate III., 2, 2a, 2b, 2d) is of a usual NOCTUÆ type, i.e., of a polished brown corneous texture, more semitransparent than usual, though not so much so as strigosa, still less as compared say with Hadena chenopodii or Cucullia: the abdominal segments tapering, 5th and 6th being as usual free; length 19 mm., width 5 mm., no hairs or bristles, though the double hairs at the antennal bases exist in little more than microscopic form. The sculpturing is in the form of very minute pits, which are most numerous dorsally, and do not exist on the leg and wing cases or thorax, which are finely wrinkled, but not so as to interfere with the shining polished character. The transparency permits, especially on the abdominal segments, certain markings due to the interior structure to be seen, and including a darker dorsal line (dorsal vessel ?) a paler lateral one, with darker and lighter (fat masses?) marbling between, the ventral aspect being paler. The prothoracic spiracle is a very slender slit, almost obsolete, indeed I am inclined to say that no aperture exists, those of the 2nd to 7th abdominal segments each being marked, being raised on a slight conical projection followed by a depression. The 8th abdominal spiracle is visible but obsolete. The anal armature consists of a wrinkled projection of the dorsal half of the extremity, armed dorsally by two central spines, and ventrally by three similar spines on either side. All this group have a similar armature, but vary, especially in the number of ventral spines on either side and in the curvature of the hooks, which they often form or terminate in. In tridens there is a very slight curvature, and the hold taken of the silk of the cocoon is slight. In tridens their number is very usually three, but a considerable portion of pupæ have four, or, not unfrequently, four on one side only. When this occurs, the extra spine is often very slender and close to the outer side of the outer one, as if split off it. It is curious that the name tridens, given no doubt on account of the trident or psi (¥) mark of the imago, should be so applicable to this typical point in the pupal structure. The

curvature of the ventral set is inwards, of the dorsal pair downwards (ventrally) more decidedly than the others (see Plate III., 2a, 2b, 2d). Certain flattenings of the dorsal surfaces of the first four abdominal segments, which are more evident in some other species, are easily observed in this species when carefully looked for. To form its cocoon this species appears to prefer to get behind a piece of loose bark or into a chink of rotten wood, where it hollows out a suitable cavity, which it completes into a cocoon with some white silk and the removed chips, very slightly, if at all, lining the excavated hollow. It will, however, very readily accept an already prepared tubular hollow, or will perform all the work of excavating one for itself in rotten wood or in the pith of a piece of elder twig, and in this case closes the opening with the top of the cocoon. In default of a more suitable nidus it will go down into sawdust or even earth, forming an ordinary cocoon of silk and the surrounding material.

I have already referred to the fact that on one occasion half of a certain brood emerged in August as an autumnal brood, in time enough for a second brood to have occurred, but that on no other occasion among hundreds of moths has an autumnal specimen shown itself. This shows that it is very unsafe in the matter of habits of this sort to regard as invariable in a species, any habit, which we may have found to be so, in even a very large experience.

This consideration prevents my saying that tridens never has four-moult larvæ, so frequent in some species, but I have never detected one.

Tridens occurs here at precisely the same seasons, and in precisely the same places as psi. Wherein they differ in habit, why there is room for the two species, why the one does not displace the other, are matters on which I have still everything to learn. Tridens like psi, will eat almost anything arboreal, but I think it has a closer relation to rosaceous plants than psi, especially fruit trees, and is perhaps commonest here in pear orchards; whilst psi is at least equally at home on forest trees, and may be met with on oak, birch, etc., on which I never happen to have taken tridens. have a suspicion that the fine pink tinge that has characterised some of my broods, and which occurs in several Acronyctas as a variety, is here related to cherry as a food, but I have instituted no special experiments to test the point.

Acronycta (Cuspidia) psi.—Psi is in many respects so like tridens that having fully described those aspects of that species to which I have paid most attention, psi may be most conveniently treated by noting the points of distinction between them, rather than by going into a fully detailed account of each stage. Psi is the only Acronycta of which it has happened to me to meet with the egg as laid naturally by the moth in the wild state. This egg was found on July 4th, 1888, laid on the upper surface of an oak leaf, the diameter was .97 mm., and the height about .33 mm. ; it had 51 ribs, of a pale straw tint or almost colourless. An egg laid in captivity on a glass slide measured 1.03 mm. in diameter and had 50 ribs, other specimens had 54 ribs. It is thus seen that the egg is distinctly larger than that of tridens, and has a larger number of ribs ; in colour (or want of colour) and other characters they are very much the same; in the figures (Plate VIII., fig. 1, psi; 2, tridens), the difference in colouring represents the different method

taken by the artist, at different times, to show the glassy transparency of the eggs, and does not correspond to any actual difference of tint in the eggs themselves. These two eggs exhibit perhaps more distinctly than any others, what is very obvious in all Acronycta eggs, and is common to all eggs of Lepidoptera so far as I have observed them, viz., that the egg contents shrink away from the shell na very early stage of development, leaving a space containing only a clear fluid between, and the flatness of these eggs leaves this space very evident as a margin round the contents, and in the species with coloured egg contents, this has the form of a colourless ring round the coloured internal egg proper. In most species the young larva is very plainly visible through the shell before hatching. In psi and tridens it is perhaps most evident, owing to the transparency and thinness of the egg shell, and the transparency of the larva itself. It lies coiled round the egg, making one complete circle with the head in the centre, and the arrangement of dark and pale segments in psi and tridens is such that the black head in the centre is surrounded by a margin divided into six nearly equal parts which are alternately dark and light tinted.

The hatching may occur in from five to twelve days after laying, according to the temperature prevailing. It is perhaps repeating unnecessarily, as the sculpturing is almost identical in all the species, to point out that the transverse ribs are only represented by a waved outline of the summits of the primary ribs and hollows on their sides, the hollows and projections of the sides of the ribs corresponding to each other on opposite sides of each furrow, and therefore alternating in adjacent furrows, and that the micropylar area is marked by a small circle of slightly raised radiating lines, surrounded by a hardly raised irregular margin in which the ribs terminate; the ribs arise from this to the number of about twenty, and increase in number towards the margin by dividing dichotomously in some instances, in others by arising de novo, in the hollow between two other ribs.

The newly hatched larva (Plate VI., fig. 1) is 2 mm. in length, very distinctly larger than that of tridens, this is unmistakably seen by drawing them under the camera when the head of the larva of psi is decidedly larger than that of tridens, in the proportion of 8 to 7 in diameter. The only other point of difference that I can be sure of is that the 13th segment in psi belongs rather to the dark series, in tridens certainly to the pale. I think I may also say that the tubercles of psi are rather larger and more markedly angulated than those of tridens, and the lateral plates of the pro-legs are nearly colourless in tridens, distinctly dark in psi.

When fully grown in this skin, it has a trace of a broad yellow dorsal line on the pale segments, viz., 3.4, 6.7, 10.11, the 12th segment is already large and dark, with its four tubercles set four-square; the 13th segment seems intermediate in tint between the dark and light series. The hairs (this applies also to tridens) are one to each tubercle, those of the anterior trapezoidals being very long, those on 11 very short, on 5-10 nearly twice the diameter of the larva in length, the others longer; the posterior trapezoidal hairs on 12 have the appearance of belonging to the anterior trapezoidal set, being equally long and merging with them.

Its habits of pupating seem to be identical with those of tridens already

noticed. The pupa (Plate III., figs. 1, 1a, 16) is not to be distinguished with certainty from that of tridens. Psi usually has four spines on each side forming the ventral portion of the anal armature, whilst tridens usually has but three, but just as tridens has not unfrequently four, so psi has at times only three. Tridens is also usually smaller and more delicate and transparent in appearance, and I cannot with certainty say of any individual pupa which it is, but of a score of pupæ said to be all one species, I should take a census of the numbers having 3 and 4 spines to the lateral anal armature, and if 3 predominated, I should say they were tridens, if 4, then they were psi.

On one or two occasions I have fancied this larva missed the 5th moult, but being on occasions when the moults were not being carefully recorded, am in doubt, nor have I reared an autumnal specimen.

The young larvæ of Cuspidia have each their own method of eating and resting. Psi and tridens affect somewhat impartially either side of the leaf, leaving the small ribs and the cuticle of the opposite surface, and when at rest are curled round in a circle.

Notes on Plate VIII.-The ova here delineated are those of the sections Cuspidia and Bisulcia, together with those of Moma orion, Demas coryli, and Diloba cæruleocephala, three species associated by many systematists with the Acronyctas; in my opinion correctly so in the case of M. orion, doubtfully in that of D. cæruleocephala and erroneously in that of D. coryli. I am very well satisfied with the success of the artist in these delineations. As pictures of the eggs they are everything that can be desired, and convey to the mind a most correct idea of the actual objects. As a matter of scientific accuracy they may be criticised on two points:-1st. The glassy transparency of psi, tridens and strigosa is of precisely the same character, and that of ligustri is nearly the same, and it is therefore unfortunate that, the drawings being made at different times, the method of representing this has involved different, instead of identical tints, in each instance. 2nd. In several cases the ribs are represented as all proceeding to the summit of the egg, instead of diminishing largely in number either by coalescing or by certain ribs stopping short as shown in the lateral view of M. orion (fig. 10 a).

Fig. 3.-Leporina is most accurate in this respect, and is indeed a wonderfully successful representation of one of the most beautiful of these beautiful objects. The marginal clear zone is shown in all the Cuspidia eggs, and is widest of all, as shown in megacephala, the largest but also the flattest of the group. The eggs of aceris and alni most resemble those of the Viminia group, auricoma being, at its best colouring, not unlike them; psi and tridens which in the larva state most approach Viminia, both in the arrangement of dark and light segments in the young larva and in the rumicis attitude of the older larva, and in these respects are to some extent intermediate between Viminia and Cuspidia, depart from the types of both groups in being colourless. It may be that they are the more ancient forms and that the colouring of the others has been acquired later.

It may be useful to append a note of the sizes of these eggs and the number of their ribs, both items being subject to variation within small limits.