In the Polydorus-group two species, P. Antiphus and P. Diphilus, inhabiting India and the Indian region, are tailed, while the two which take their place in the Moluccas, New Guinea, and Australia, P. Polydorus and P. Leodamas, are destitute of tail, the species furthest east having lost this ornament the most completely.

Western species, tailed. Eastern species (closely allied), less tailed.
Papilio Pammon (India) tailed. P. Thesus (islands) very short tail.
P. Agamemnon, var. (India) tailed. P. Agamemnon, var. (islands) not tailed.
P. Antiphus (India, Java) tailed. P. Polydorus (Moluccas) not tailed.
P. Diphilus (India, Java) tailed. P. Leodamas (New Guinea) not tailed.

The most conspicuous instance of local modification of form, however, is exhibited in the island of Celebes, which in this respect, as in some others, stands alone and isolated in the whole archipelago. Almost every species of Papilio inhabiting Celebes has the wings of a peculiar shape, which distinguishes them at a glance from the allied species of every other island. This peculiarity consists, first, in the upper wings being generally more elongate and falcate; and secondly, in the costa or anterior margin being much more curved, and in most instances exhibiting near the base an abrupt bend or elbow, which in some species is very conspicuous. This peculiarity is visible, not only when the Celebesian species are compared with their small-sized allies of Java and Borneo, but also, and in an almost equal degree, when the large forms of Amboyna and the Moluccas are the objects of comparison, showing that this is quite a distinct phenomenon from the difference of size which has just been pointed out.

In the following Table I have arranged the chief Papilios of Celebes in the order in which they exhibit this characteristic form most prominently. (See Plate VIII.)

Papilios of Celebes, having the wings falcate or with abruptly curved costa. Closely allied Papilios of the surrounding islands, with less falcate wings and slightly curved costa.
1. P. Gigon, n. s. P. Demolion (Java).
2. P. Telephus, n. s. P. Jason (Sumatra).
3. P. Miletus, n. s. P. Sarpedon (Moluccas, Java).
4. P. Agamemnon, var. P. Agamemnon, var. (Borneo).
5. P. Macedon, n. s. P. Peranthus (Java).
6. P. Ascalaphus. P. Deiphontes, n. s. (Gilolo).
7. P. Hecuba, n. s. P. Helenus (Java).
8. P. Blumei. P. Brama (Sumatra).
9. P. Androcles. P. Antiphates (Borneo).
10. P. Rhesus. P. Aristæus (Moluccas).
11. P. Theseus, var., ♂. P. Thesus, ♂ (Java).
12. P. Codrus, var. P. Codrus (Moluccas).
13. P. Encelades. P. Leucothoë (Malacca).

It thus appears that every species of Papilio exhibits this peculiar form in a greater or less degree, except one, P. Polyphontes, Bd., allied to P. Diphilus of India and P. Polydorus of the Moluccas. This fact I shall recur to again, as I think it helps us to understand something of the causes that may have brought about the phenomenon we are considering. Neither do the genera Ornithoptera and Leptocircus exhibit any traces of this peculiar form. In several other families of Butterflies this characteristic form reappears in a few species. In the Pieridæ the following species exhibit it distinctly:—

1. Eronia tritæa compared with Eronia Valeria (Java).
2. Iphias Glaucippe, var. „ „ Iphias Glaucippe (Java).
3. Pieris Zebuda „ „ Pieris Descombesi (India).
4. P. Zarinda „ „ P. Nero (Malacca).
5. P., n. s. „ „ P. Hyparete (Java).
6. P. Hombronii have the same form, but are isolated species.
7. P. Ithome
8. P. Eperia, Bd. compared with P. Coronis (Java).
9. P. Polisma „ „ P., n. s. (Malacca).
10. Terias, n. s. „ „ P. Tilaha (Java).

The other species of Terias, one or two Pieris, and the genus Callidryas do not exhibit any perceptible change of form.

In the other families there are but few similar examples. The following are all that I can find in my collection:—

Cethosia Æolecompared withCethosia Biblis (Java).
Junonia, n. s.„ „Junonia Polynice (Borneo).
Limenitis Limire„ „Limenitis Procris (Java).
Cynthia Arsinoë, var.„ „Cynthia Arsinoë (Java, Sum., Born.).