branch of the media, vein iv2, becomes cubital; whereas in the Pieridæ it becomes radial, as in the Hesperiades generally, with the exception of Leptidia, an aberrant pierine form in which this branch also becomes cubital on the hind wings. In the LycænidæHesperiadæ it remains central, while it becomes radial in the Nemeobiidæ, as in the Pieri-Nymphalidæ. The upper branch of the media, vein iv,, ascends the radius in the specialized forms of the Papilionides, as in the Pieridæ, and does not remain permanently attached to the cross-vein, from the upper corner of the cell, as in the Nymphalids.

The second direction in which specialization shows itself lies in the suppression of the branches of the radius on the fore wings. The five-branched radius, exhibited in a generalized state in the Papilionidæ, becomes four-branched in the most specialized butterflies of the group I have yet examined, in Parnassius apollo and its very close ally, Doritis mnemosyne.

For the rest, the specializations of the neuration generally show themselves in absorption, so that I have laid it down as a principle that the amount of the specialization is measured by the extent of the absorption or disintegration.

The so-called "tails" to the hind wings in this group are prolongations of vein iv,. They are probably to be regarded as characters of specialization, and they possibly had their origin as secondary sexual ornaments of the male sex, although now most of the females have followed suit. In certain Papilios in which the female is mimetic, the "tail" in this sex may have been abandoned after having been originally acquired.

NOMENCLATURE AND HOMOLOGY OF THE VEINS.

The ancestors of the Papilionides must have exhibited vein ix of the primaries, since this is evidently a retained and not an acquired character. I follow Comstock in numbering the loop at the base of vii as viii in the Hesperiades and other groups. This vein viii is absent in the Papilionides, where there is no place for it. It may have originated in a splitting of vii at base, and not be a relic of a longitudinal separate vein. In this case the number assigned to ix would be incorrect, but the numbering having been introduced, to change it would make confusion, although the vein

itself would be homologous with vein viii of hind wings and not with vein ix of the same pair.

The cross-vein between cubitus and vii very gradually fades out in the more specialized forms of the Papilionides and finally disappears. It fades from its base, where it joins on to vii, upwardly, becoming a mere remnant in the Teinopalpidæ, extending below the cubitus. Mr. Quail has discovered a similar slight blotch in Anosia, and I believe correctly homologizes it with the crossvein of Papilio. I have found it still more extended in Heliconius, where it reaches, running a little obliquely downward, to about the place where vi would be had this latter vein not faded completely out. In my preparation and the original photograph of Heliconius this fragment of the cross-vein is with difficulty to be seen, and I overlooked it at the time. I also failed to notice that Heliconius shows a trace of vein viii.

The presence of relics of a cross-vein below cubitus in the Limnadide and Heliconiidæ, homologous with that in Papilio, does not necessarily imply consanguinity between the groups. The hypothesis has suggested itself to me that the lepidopterous wing may have originally shown a series of longitudinal and independent veins, connected by a system of cross-veins, and without the present furcations of the branches of the media and radius. The disappearance of the cross-veins would allow of the contact of the longitudinal veins, and probably assist the shifting movements we now perceive in action (PROC. Aм. PHIL. Soc., Jan., 1898). The cubital cross-vein would be a relic of these.

GENERAL CLASSIFICATION.

If we assume vein viii of the fore wing in the Hesperiades to be a splitting of the vein vii at base, it might not be held to be homologous with the vein viii, occupying the same position, in the Sphingides and other groups of moths. This supposition seems to me untenable. In the lycaenid genera Aurotis, Zephyrus and Feniseca I do not perceive any difference in this vein from its appearance in the Hesperiada. It becomes simply more prominent and somewhat less strap-like and rigid in the moths. In any case, this vein viii, while allying the Hesperiades to the higher groups of moths, is absent in the Papilionides and is replaced by

out.

what I consider to be the undoubted relic of a true longitudinal vein, taking an opposite direction to vein viii, and running outwardly and downwardly to the inner margin of the wing, and which I have called vein ix. If the view is accepted that vein viii originates in a splitting of vein vii at base, then it might be held. that this splitting occurred after vein ix had been gotten rid of, and that consequently butterflies with vein viii represent a succeeding stage in this respect. I believe to have shown that, in certain of the more specialized of the Hesperiades, vein viii gradually fades In the Nymphalids I find very faint traces of it only in certain Argynninæ. I have found no trace of it in any Satyrid. It has vanished in Leptidia, is present in Pseudopontia, and, while strong in Colias, is fainter in some of the other Pierinæ. Certain Lycænidæ appear to have lost it, or it is very faint, and this may be the case also with some Hesperians, though in this latter group it is usually quite legible. Conceding its variability, no better evidence perhaps of its power of extinguishment can be offered than that it is strongly marked in Libythea and Limnas, and that it is incomplete in the related Heliconius.

Let our theories as to viii be as they may, one thing seems clear : that no traces of vein ix of the Papilionides have been found in the Hesperiades, and that this latter group is held together, as opposed to the former, by the negative character of its absence.

My attempt therefore, to connect the Papilionides with the rest of the diurnals, has failed. In my first draft of the genealogical tree (1896–1897), and which is still pinned above my desk, I supposed that the point of contact might be with primitive forms, less specialized than the Hesperians. These would have all exhibited vein ix, which the Papilionides had retained, while the other ascending branch had lost it. I reproduce here this sketch, the readier since Dr. Chapman, in letters to me, has queried whether an analogous scheme might not work.

The point x is supposititious and, in my original sketch, marked with a query. I had endeavored also to connect the Papilionides with the Nymphalids. Aside from the major difficulty, the only point of similarity I could then find was the five-branched radius, which in both the Papilionidæ and in the Nymphalids was in a generalized condition. To this must be added traces of the cubital cross-vein in the Limnadide and Heliconiidae. While too much importance should not be paid to a suppression of branches of the radius now in a fluid state, as in the Pieridæ and Saturniadæ, stress must be laid on the fact that throughout the Nymphalids the five-branched condition is retained, while in the Parnassians we have also a four-branched type, in which the more generalized five-branched condition has been very clearly abandoned. The only movement I have found in the Nymphalid radius consists of a transference of iii, to beyond the cell in the Heliconiidæ and certain long-winged forms, or in forms perhaps tending in that direction, such as Thalerope, Araschnia, Melitaea and Euptoieta; while in the long-winged forms, Agraulis, Dione, the vein iii, has followed suit. A trace of this movement is seen in Argynnis, but not in Issoria lathonia. I was also impressed by the fact that in the Papilionidæ vein iii, did not attain the apex of the primary wing,

but this condition is not constant in the Nymphalidæ, and is abandoned in the Satyrids and Libythea. On the other hand, the differences between the Papilionides and Nymphalidæ are numerous. Vein iv, the middle branch of the media, becomes radial in the latter, in the former cubital, in specialization. Vein iv, also leaves the lower outer corner of the cell in the Papilionidæ, and, although this position is abandoned in the higher genera of the Parnassiidæ, still it may have been a primitive one, since it occurs now with the more generalized forms. The peculiarities of the papilionid wing are very strong, and notwithstanding the discovery of a remnant of the cubital cross-vein in Anosia (Danaus) by Mr. Quail, and by myself in Heliconius, I cannot find evidence sufficiently weighty to connect the groups from the neuration. But while the coincidences allow of some comparison of the Papilionides with the Pieri-Nymphalidæ, although an affinity appears to me to be illusory, it is impossible to consider them as representing in any nearer way an ancestral form of the Lycæni-Hesperiadæ.

Upon the generalized condition of Hesperia too much stress has, perhaps, been laid in literature. All the forms of which I have examined the neuration seem relatively specialized upon their peculiar plan of venation, of which the more modern and advanced outgrowths are to be found in the Lycanidæ, culminating in Thecla. But all the butterflies belonging to the Lycanid-Hesperid phylum seem relatively too specialized as to represent adequately the primitive form of the diurnals. Whatever the primitive form was like, the only character in which it may have resembled Hesperia, or the primitive form of the Charaxinæ, is that of the separation of the longitudinal veins. The primitive butterfly may have had separated veins, together with cubital cross-veins, of which we find a trace in the Heliconians and Limnads, and an anal vein, like the Papilionides, on the primaries. And, perhaps, by conjuring up a creature rejoicing in apparently residual features, we might attain to a picture in somewhat like manner as Gabriel Max has painted Haeckel's Pithecanthropus alalus europæus. But the muse of morphology, as I am now able to understand her, abandons me at this juncture, with the unconnected threads of the groups A and B, the Papilionides and Hesperiades, dangling downwards into the abyss of Time, kept apart by the presence of vein ix of primaries in the one and its absence in the other.

The object of my communications upon the wings of butterflies,