1. The purpose of this paper is to give an account of certain aspects of the natural history of East African tree-frogs belonging to the genera Hyperolius, Megalixalus, and Leptopelis. The material upon which it is based was obtained by the writer in the palm-forest region of the lower Zambesi Valley during the Zoological Society's Expedition of 1927.

2. In particular, a detailed investigation of the food-habits of tree-frogs, as indicated by the examination of food eaten under natural conditions, has been made (1) in order to test the efficiency of procryptic and warning colours, and mimicry, in defending insects from predatory attack; and (2) to indicate the part played by these batrachians as a factor in the production of adaptive coloration in insects.

3. The general habits, habitat, and abundance of the frogs are described. It is suggested that the colour-scheme and resting posture of M. fornasinii have a concealing and aggressive function which enables the species to exploit an abundant source of food available in the active, alert, and strongly-flying Odonata, muscid Diptera, and Lepidoptera—groups which are relatively inaccessible to, and little eaten by, the other tree-frogs whose habits were investigated.

4. In H. argus certain differences have been observed (1) in the feeding-habits of the sexes, and (2) in the choice of habitat. These have been correlated respectively with the difference in the size of the sexes and with the marked sexual dichromatism in the species. The conspicuous colouring of the female, associated as it is with aposematic habits, may have a “warning” significance.

5. Frogs are preyed upon by numerous enemies. They are hunted by small carnivorous mammals, by monitors, by fish, and by members of their own order; they are eaten by a wide range of birds, including crows, shrikes, kingfishers, owls, eagles, harriers, herons, bitterns, ibises, hammer-heads, and spoonbills; and they are relentlessly persecuted by snakes, especially tree-snakes, which probably rank first in importance as enemies of tree-frogs in East Africa.

6. The stomach-contents of 798 tree-frogs belonging to the following species are tabulated (the figures refer to the number of frogs examined and to the number of food-animals recovered):—Hyperolius marmoratus, 40 (2675); H. bayoni, 110 (3688); H. argus, 254 (3300); Megalixalus fornasinii, 360 (1119); M. brachycnemis, 11 (31); Leptopelis johnstoni, 8 (13); Phrynobatrachus acridoides, 15 (602).

7. 11,428 insects and other food-animals were obtained from this source, representing 11 orders and at least 153 species. Taken collectively, this material comprises the following groups of animals:—Hymenoptera, 87.16 per cent.; Hemiptera, 5.84; Diptera, 3.40; Coleoptera, 1.88; Lepidoptera, .70; Orthoptera, .46; Araneida, .34; Odonata, .15; Isopoda, 04; Isoptera, 02; Batrachia, .01.

8. There is a close general similarity in the food-habits of the three species of Hyperolius examined. Ants (Pheidole megacephala) in each case comprise more than 90 per cent, of the total; Hemiptera, Coleoptera, and Diptera, in varying proportions, make up the bulk of the remaining food. In contrast to the above, Megalixalus fornasinii is essentially a fly-catching frog—Diptera comprising the bulk of the food eaten.

9. The frogs are considered in relation to their biological environment, with special reference to enemies and prey, and an attempt has been made to indicate certain “food-chains” in which they form an essential link. There is evidence that these forms may be reckoned among the main predatory enemies of small insects in the tropics, and it is suggested that their depredations are not without economic value to man.

10. Ants, which are usually regarded as well-defended insects, are widely used as food by frogs. Collected records of stomach examinations made by thirty-five investigators (embracing data relating to 153 species of Anura and to some thousands of individuals) shows that about 60 per cent, of the species examined include ants in their diet. This figure is only slightly exceeded in the case of one other kind of prey, namely, beetles.

11. In the present material ants are represented by 9937 out of 11,428 food-animals. The seven species examined each preys upon ants. These insects comprise the main food of the following species:—H. marmoratus, H. bayoni, H. argus, P. acridoides, and M. brachycnemis, the percentage of ants in terms of total food-content being 98, 96, 93, 92, and 77 respectively.

12. In three species over 90 per cent, of the frogs containing recognizable food had been feeding upon ants. The figures for each species are as follows:—P. acridoides, 100 per cent.; H. bayoni 97; H. marmoratus, 92; H. argus, 79; M. brachycnemis 78; L. johnstoni, 33; M. fornasinii, 27.

13. One is forced to conclude from these facts that the various phenomena of ant-resemblance, in so far as frogs are concerned, can have little, if any, adaptive significance.

14. Hymenoptera (other than ants) comprise a significantly low percentage of the food, being represented by 20 specimens (i. e., .18 per cent, of food-animals) obtained from the stomachs of 18 frogs (i. e., less than 2–3 per cent, of those examined).

15. The food-animals of the frogs have been classified according to colour, and the results are graphically shown for each of the following groups:—Orthoptera, Hemiptera, Coleoptera, Diptera, Hymenoptera, and Araneida. Of 10,968 specimens sufficiently complete for analysis, not more than 48 insects are conspicuously coloured, and of these only 14 specimens (.13 per cent.) belong to the typically aposematic group.

16. These observations lend strong additional support to the classes of facts upon which the theory of warning colours rests: (1) that conspicuous pattern and colour in insects are typically associated with disagreeable or dangerous attributes which render their owner (relatively) unpalatable; (2) that insectivorous enemies learn by experience in nature to recognize and to avoid unpalatable prey; (3) that in leading to immediate recognition by enemies, and by checking the fatalities and injuries caused by experimental tasting, aposematic habits and colour are of vital benefit to their possessor.

17. Tree-frogs are serious enemies of small insects in the tropics. They depend mainly upon sight in hunting and capturing prey. There are grounds for believing that they exercise discrimination in the choice of food, and that they learn to recognize and avoid unpalatable prey. There is strong presumptive evidence that aposematic colour and habit assist frogs in the recognition of distasteful species. The food-habits and mental equipment of tree-frogs is such as to suggest their claim to a not unimportant share with birds and lizards in the production, through natural selection, of procryptic and aposematic coloration in insects.