Great size, as in Aphilotrix radicis, occasionally puts the larva beyond the reach of the parasite, while the very opposite condition protects by insignificance. Another effective protection is found in Andricus curvator, where the inner gall lies in a large hollow chamber, an arrangement which makes the work of the parasite difficult and uncertain.

Outside enemies such as tits, pheasants, and squirrels are as much to be feared as parasites. The larvae are defended from these, sometimes by the nature of the outer gall, which in Aphilotrix fecundatrix consists of closely imbricated scales resembling a hop strobile. In A. Sieboldi the outer gall is hard and stony; in Cynips Kollari and Trigonaspis crustalis, the tannin which is contained in the tissues renders them distasteful. As the galls mature the percentage of tannin becomes less, but the hardening of the epidermic layers which then takes place affords a new line of defence. After the gall has fallen another set of influences secure its safety by the changes they produce in its surface-colouring. The galls of Andricus ostreus, Biorhiza renum, and many others, are supplied, as Beyerinck has pointed out, with certain hydrocarbon compounds, which absorb moisture and undergo molecular changes after they reach the ground; with these chemical changes the growth of the larva and the development of protective coloration in the gall take place.

Cynips Kollari, Dryophanta scutellaris, and a few other gall-larvae and gall-flies, have the power of emitting a disagreeable bug-like odour, which is not sexual, since agamous species possess it, but probably protects the flies to some extent from birds. Certain galls have a fruity and aromatic smell1, the use of which does not 1 Paszlavszky, Wien. Ent. Zeit., 1883. p. 130.

seem quite clear, unless they can be swallowed and voided undigested by a temporary host.

It is remarkable that characters, closely resembling those acquired by fruits, should have been evolved from a totally different cause. In the case of fruits these characters have been of service in securing the distribution of the seeds; in the case of galls, in securing the safety of the larvae; but in both cases it has been their fitness that has brought them into existence.

Darwin and all writers before him held that the force calling out gall formation was due to a chemical secretion injected by the gall-mother. Malpighi considered that it acted as a ferment on juices existing in the plant; and this was the view of Réaumur, but he added to it the thermal effect of the egg, and the nature and character of the wound, which varies according to the shape of the ovipositor of each species. Dr. Derham thought the formation was 'partly due to the act of the plant, and partly to some virulency in the juice or egg, or both, reposited on the vegetable by the parent animal; and just as this virulency is various according to the difference of its animal, so is the form and texture of the gall excited thereby.' Darwin speaks of galls as produced 'by a minute atom of the poison of a gall-insect,' and compares them to the specific local processes of zymotic diseases. Sir James Paget, writing in 1880, said that 'the most reasonable, if not the only reasonable theory, is that each insect infects or inoculates the leaf or other structure of the chosen plant with a poison peculiar to itself.' This may be taken as the view accepted by scientists', until in the following pages Dr. Adler showed conclusively that there was no foundation for supposing

1 See 'Galls,' Encyclop. Britann. ed. 9, where the same view is expressed.

that the gall-mother injected any irritating secretion whatever, and Beyerinck 1 proved that the fluid ejected by the gall-fly is without taste or smell, and absolutely unirritating if injected under the skin. It is probably nothing more than a very mild antiseptic dressing applied to the wound made in the plant. Both these authors show plainly that it is not in the gall-mother, but in the larva, that we must seek for the cause of gall-growth; and that it is the nature of the salivary secretion, and the manner of feeding of the larva, peculiarities inherited by each species, which give the characteristic growth to the gall. The fact that sometimes a blastem has actually begun to form before the egg-shell has ruptured, proves that one of the exciting causes must be a chemical fluid, secreted by the salivary glands, and possessing amylolytic and proteolytic ferments. This fluid is capable of passing through the cell-walls and producing effects at a distance of several mm. beyond actual contact with the larva 2.

The necessity for the continuance of the excitation during the whole period of gall-growth is shown by its cessation when the larva has perished by parasites. In some galls, however, the parasites are evolving the power of prolonging gall-growth beyond the death of the gall-maker, although they have not yet actually acquired the art of initiating it.

The duration of gall-vitality is shortest in succulent galls, such as Spathegaster baccarum, S. albipes, S. tricolor, and S. verrucosus, growing from leaves; or such as S. Taschenbergi, S. similis, and Trigonaspis crustalis, growing from dormant buds. Galls which grow within the leaf substance, like Spathegaster vesicatrix and

1 Beyerinck, Über die ersten Entwickl, einiger Cynipidengallen, p. 179. 2 Hoffmeister.

Andricus curvator, live as long as those leaves of which they form part. Aphilotrix fecundatrix and Cynips Kollari die at the end of the first summer; Dryophanta longiventris, Aphilotrix collaris, A. globuli, A. autumnalis, and Neuroterus ostreus, perish during the first winter; the spangle-galls and Biorhiza renum live till spring; Aphilotrix radicis, A. Sieboldi, and Biorhiza aptera do not die till the second winter; while Andricus inflator may almost be termed perennial, since new oak-buds are developed upon it in the following year.

DESCRIPTION OF PLATES1

PLATE I.

Fig. 1. Galls of Neuroterus lenticularis.

Fig. 1. Galls of Spathegaster baccarum, on the leaf and flowering catkin.

Fig. 2. Galls of Neuroterus laeviusculus.

Fig. 2a. Galls of Spathegaster albipes (×2).

Fig. 3. Galls of Neuroterus numismatis, one gall enlarged.

Fig. 3. Galls of Spathegaster vesicatrix, one gall enlarged.

Fig. 4. Galls of Neuroterus fumipennis.

Fig. 4. Galls of Spathegaster tricolor.

Fig. 5. Galls of Aphilotrix radicis. One in the fresh state, the other in section, after maturity.

Fig. 5. Galls of Andricus noduli. One shoot bears the fresh galls, the other the galls of the previous year.

Fig. 6. Galls of Aphilotrix Sieboldi. fresh, on the other mature and woody.

Fig. 6. Galls of Andricus testaceipes.

On one stem the galls are

Fig. 7. Galls of Aphilotrix corticis. One portion of the bark exhibits the fresh, the other the mature galls.

Fig. 7. Galls of Andricus gemmatus, showing the holes through which the flies have emerged.

Fig. 8. Galls of Aphilotrix globuli. The woody inner gall shown at maturity.

1 The galls were drawn from fresh specimens by Herr O. Peters, Göttingen.