Summary1Hornets are omnivorous scavengers: in their visits to the discarded carcases of a Palm Dove (Streptopelia senegalensis aegyptiaca) and of a Pied Kingfisher (Ceryle r. rudis) at Beni Suef, Middle Egypt, they were observed to discriminate in favour of the flesh of the dove. Selective feeding byVespa orientaliswas subsequently used as a method of assessing the relative edibility of the flesh of 38 species of birds in terms of hornet preference.2During the period 1941–1944, 141 hornet experiments were carried out in Egypt and the Lebanon: in each experiment the insects were presented with a choice as between the flesh of two species of birds, or (in control tests) of the same species. The two food elements offered in each experiment were approximately similar in bulk and anatomical content, but the quantity and quality were varied from one experiment to another throughout the series.3The number of insects feeding at each carcase was counted at regular intervals of (usually) five minutes, until one food element had been exhausted, preference for one bird of the pair being established by a differential rate of attack, based on the total number of hornets counted at each food element. Experimental data are tabulated to show: (a) the total number of hornets counted at each test; (b) the ratio of attack at the two food elements; (c) standard error;(d) observed deviation.4The consistency of hornet discrimination and reliability of the experimental method was checked by linking and overlapping experiments (between pairs of species more or less widely separated in the edibility series);and by repetition and control experiments (between similar pairs, or the same species). In general, the experimental data show the hornets to be highly sensitive to differences in the relative palatability of the birds offered.5Thirty‐eight species of birds tested include: members of several Passerine families—Oriolidae, Fringillidae, Ploceidae, Alaudidae, Motacillidae, Paridae, Laniidae, Pycnonotidae, Sylviidae, Turdidae and Hirundinidae; and of other orders— Coraciiformes, Piciformes, Strigiformes, Ciconiiformes and Columbiformes. Arranged in order of acceptability, they fall into the following groups:—6Palatable: Wryneck (Jynx t. torquilla), Crested Lark (Galerida cristata magna, G. c. maculata), White‐vented Bulbul (Pycnonotus barbatus arsinöe), Buff‐backed Heron (Ardeola i. ibis), Greenfinch (Chloris chloris chlorotica), Little Green Bee‐eater (Merops orientalis cleopatra), Palm Dove (Streptopelia senegalensis aegyptiaca), House Sparrow (Passer domesticus niloticus, P. d. biblicus), Spanish Sparrow (Passer h. hispaniolensis), Pied Chat (Oenanthe pleschanka), Spotted Flycatcher (Muscicapa s. striata), Blue‐headed Wagtail (Motacilla f. flava), Lesser Whitethroat (Sylvia c. curruca).7Intermediate: Little Owl (Athene noctua lilith), Cretzschmar's Bunting (Emberiza caesia), Goldfinch (Carduelis c. niediecki), Rufous Warbler (Erythropygia g. galactote.s), White Stork (Ciconia c. ciconia), Garden Warbler (Sylvia borin), Willow Warbler (Phylloscopus t. trochilus), Savi's Warbler (Locustella l. luscinioides), Sprosser (Luscinia luscinia), Great Tit (Parus major terrae‐sanctae), White Wagtail (Motacilla a. alba), Redstart (Phoenicurus p. phoenicurus).8Unpalatable: Blackcap (Sylvia a. atricapilla), Red‐rumped Swallow (Hirundo daurica rufula), Kingfisher (Alcedo atthis atthis), European Swallow (Hirundo r. rustica), Lesser Grey Shrike (Lanius minor), Red‐backed Shrike (Lanius c. collurio), Golden Oriole (Oriolus o. oriolus), Hooded Chat (Oenanthe monacha), Hoopoe (Upupa epops major, U. e. epops), Masked Shrike (Lanius nubicus), Mourning Chat (Oenanthe l. lugens), Pied Kingfisher (Ceryle r. rudis), White‐rumped Black Chat (Oenanthe l. leucopyga).9Consideration of the species investigated from the point of view (a) of their relative edibility to the hornets, and (b) of their relative visibility in the field, demonstrates a general inverse correlation between conspicuousness of the plumage and palatability of the flesh. When classified according to greater or less degree of conspicuousness and edibility respectively, the species fall into four groups as follows: (1) inconspicuous and palatable, 18;(2) conspicuous and unpalatable, 14;(3) inconspicuous and unpalatable, 2;(4) conspicuous and palatable, 4.10When arranged in order of acceptability, it is found that the species rated as most palatable include those with the most effective cryptic coloration and habits—such asJ. t. torquillaandG. cristatawith the lowest visibility—and highest edibility rating. Conversely, those of lowest grade edibility are, without exception, species rendered highly conspicuous in the field by their habits and coloration;and the most conspicuous entirely black and white members of the series—C. r. rudis, O. l. lugensandO. l. leucopyga—are also the most distasteful.11This inverse correlation between visibility status and edibility status appears to hold even within the limits of certain families and genera. For example, members of the families Fringillidae and Ploceidae, taken together, occur in the following order of ascending conspicuousness and descending palatability:—C. chloris, P. domesticus, P. hispaniolensis, E. caesia, C. carduelis. Again,M. albais both less palatable and more conspicuous thanM. flava.12A similar correlation is found with members of the order Coraciiformes (sensu latu), which occur, with one slight deviation, in the following order of increasing conspicuousness and distastefulness:—J. torquilla, M. orientalis, A. noctua, A. atthis, U. epops, G. rudis. On the other hand, no such close correlation has been traced for shrikes of the genusLanius, or for six species of Turdidae tested.13Two outstanding exceptions to the general correlation between conspicuousness and distastefulness are provided byA. ibisandC. ciconia, both of which are highly conspicuous, the former being also of high‐grade and the latter of moderate edibility. It is significant that both birds are much larger than any other species tested, and both are gregarious, with powers of defence rendering them relatively non‐vulnerable to predatory attack.14The distribution of different plumage colours, occurring both as predominant and subordinate elements in the coloration of the birds investigated, has been considered in relation to edibility status. A marked tendency is observed for the cryptic colour‐groups (light, dark, and greyish‐brown, green, olive and yellowish‐green) to predominate among the more palatable species;and for the sematic colours (blue, red and rufous, white and black) to predominate among the more distasteful species.15Forty‐eight further edibility tests have been carried out on 19 species of birds in the Lebanon, Egypt and England, using the domestic cat as experimental taster. Comparison of the available data shows a general agreement in the food preferences of the cat and hornet. In the 15 experiments, where similar pairs of birds were offered to both types of predator, discrimination in favour of the corresponding species was recorded on every occasion except one.16Cats agreed with the hornets in rating more conspicuous species, such asC. rudis, U. epops, and a black and white chat (Oenanthe finschii), as relatively distasteful when compared with more cryptic species such asS. senegalensis, P. domesticusandC. Moris. Further evidence suggests that Titmice (Parus major, P. caeruleus) may be relatively unpalatable to the cat, as to the hornet.17Information has been collected on the relative edibility, to man, of the birds (or of allied species) used in the hornet experiments. The available evidence again suggests a general agreement between the food preferences of hornet and man. For many genera, of whatever edibility grading, the agreement is very close, as inJynx, Galerida, Chloris, Merops, Streptopelia. Passer, Oenanthe(non‐deserticolous species),Carduelis, CiconiaandCeryle.18For other genera, more or less divergence is shown, as inErythropygia, Emberiza, Phoenicurus, Hirundo, Lanius, Parus, andUpupa—apparently rated somewhat highs—higher by the hornet than by man. In only four genera,Motacilla, Sylvia, OriolusandAlcedo, is a marked discrepancy indicated. Of the seven genera covered by data for hornet, cat and man, namely,Chloris, Passer, Merops, Streptopelia, Parus, UpupaandCeryle, in none does the evidence suggest a considerable difference in the edibility rating of these three very different meat‐eaters.19The species used in the hornet experiments have been graded in terms of relative vulnerability to predaceous attack, the assessment being based upon;(a) size;(b) weapons of defence;(c) powers of offence;(d) habitat; (e) habits;(f) sociability. The relation between vulnerability and visibility is discussed. The inverse correlation between visibility and edibility applies essentially to the more vulnerable species;but not to those less liable to attack.20The more vulnerable birds include both cryptic and phanerie types: the former are typically palatable; of the latter, all the more highly conspicuous species are also highly distasteful; the remainder are intermediate in both respects. The less vulnerable birds include both palatable and unpalatable species: all these are highly conspicuous, with the exception ofAthene, noctua, where alone the value of conspicuousness is likely to be outweighed by the need for concealment.21The investigation has been extended, in terms of the palatability of the flesh to man, over a wide range of birds outside the limited series used in the hornet experiments. The evidence suggests that the correlation between vulnerability, palatability and visibility found within the experimental series applies in general to birds as a class.22Birds, of many diverse orders, which are relatively vulnerable and which are also highly prized for the table, are in general cryptically coloured (in both sexes or in the female): for example, various species of Ciconiiformes—Botaurus: Anseriformes—Anas, Aythya; Galliformes—Francolinus, Lagopus, Bonasa. Phasianus, Perdix, Lerwa, Tympanuchus, Tetrastes, Ammoperdix, Coturnix;Ralliformes—Crex, Rallus, Porzana;Gruiformes—Otis, Sypheotis, Turnix; Charadriiformes—Scolopax, Capella, Lymnocryptes, Tringa, Calidris, Philomachus, Crocethia, Pluvialis, Sguatarola, Eudromias, Rhinoptilus, Burhinus;Psittaeiformes—Pezoporus, Neophema, Myiopsitta;Columbiformes—Crocopus, Treron, Pterocles;Caprimulgiformes—Caprimulgus; Passeriformes—Alauda, Sturnella, Emberiza, Anthus.23Further, many species, of whatever order, having the most specialized and effective cryptic coloration are also those which are especially prized for the excellence of their flesh, for example:Anas platyrhyncha, A. discors, A. strepera, Lagopus scoticus, Bonasa umbdlus, Perdix perdix, Coturnix coturnix, Otis tetrax, Turnix suscitator, Scolopax rusticola, Capella gallinago, Pluvialis apricaria, Squatarola squatarola, Burhinus capensis, Treron pompadora, Pterocles gutturalis, Pezoporus wallicus, Caprimulgus europaeus, Alauda arvensis.24Again, within certain natural groups of birds, the general correlation between edibility and visibility obtains. Thus, among the Ralliformes,Crex crex, Rallus longirostrisandPorzana Carolinaare highly palatable and cryptic in coloration and habits;whileFulica atraandGallinula chloropus, though eatable, are hardly accountable as table birds. Similarly, among the Ciconiiformes, bothBotaurus lentiginosusandB. stellarisappear to be more palatable than their more conspicuous congeners. Among plovers, similar relations are found in the seriesPluvialis apricaria, Squatarola squatarola, Vanellus vanellus, Lobivanellus indicus, Pluvianus aegyptius.25Birds which are relatively non‐vulnerable to attack and in which the female or both sexes are conspicuous—unlike vulnerable species—appear to show no general correlation between edibility and visibility. Such species often combine conspicuousness with moderate or high palatability, as seen, for example, in various members of the following genera:Struthio, Dromiceius, Pucheramphus, Pygoscelis, Diomedia, Puffinus, Majaqueus, Oestrelata, Sula, Phalacrocorax, Pelecanus, Ardea, Mycteria, Jaribu, Ciconia, Threskiornis, Inocotis, Plegadis, Hagedashia, Platalea, Phoenicopterus, Cygnus, Branta, Haliaëtus, Ibycter, Rostrhamus, Anthropoïdes, Grus, Balearica, Larus, Pagophila, Catharacta, Ara, Amazona, Pionus, Dacelo, Corvus, Pyrrhocorax.26Birds which are relatively vulnerable (omitting the factor of distaste‐fulness) and conspicuous, appear in general to be more or less highly distasteful —to a degree likely to serve as a deterrent to most predators. Striking examples of this correlation are afforded by members of many different orders, and include such species asColymbus immer, Tadorna tadorna, Mergus albellus, Tachyeres brachyptera, Melanitta nigra, Choëphaga hybrida, Opisthocomus hoazin, Pluvialis aegyptius, Lobivanellus indicus, Fratercula arctica, Uria aalge, U. lomvia, Turacus fischeri, Musophaga rossae, Crotophaga ani, Cuculus solitarius, Chrysococcyx cupreus, Phoeniculus purpureus, P. damarensis, Rhinopomastus cyanomelas, Alcedo cristata, Halcyon albiventris, Chelicutia chelicuti, Ispidina picta, Bucorax caffer, Bycanistes cristatus, Corvus macrorhynchos, Pica pica, Sternus vulgaris, Campephaga flava, Coracina caesia, Dicrurus afer, D. ludwigi, Bradornis ater, Tchitrea perspicillata, Pyrrhula pyrrhula.27The factors causing distastefulness in birds are discussed. The nature of the food plays a part in determining the flavour of the flesh: it is probable that with some distasteful species, whether insectivorous or herbivorous, the deterrent flavour is largely food‐derived. On the other hand, several other factors appear to play a part in determining distastefulness. Palatability may change with growth and age of the bird;and it differs markedly in different parts of the same individual.28There does not appear to be any close correlation between edibility status and feeding‐habits in the birds investigated. Graminivorous or frugivorous species may be of either high‐grade (G. cristata, P. barbatus) or low‐grade edibility (O. oriolus). Similarly, insectivorous species may be either palatable (J. torquilla, M. orientalis, A. ibis) or unpalatable (L. nubicus, U. epops, O. leucopyga). Moreover, members of the same genus with generally similar feeding‐habits may differ widely in their edibility status. No simple explanation of distastefulness, in terms of food‐derived deterrents, is admissible.29The work of Morton Jones on the relative edibility of insects to insectivorous birds and to ants is cited as affording a series of facts parallel to those here considered. His experiments suggest that various unpalatable insects possess food‐derived deterrents;but that distastefulness cannot be attributed widely to this cause. He concludes that relative edibility among insects must depend upon factors other than, or in addition to, the food eaten.30Distastefuhiess is considered in its rôle as a deterrent to attack by predators. The evidence suggests that, despite individual and specific differences in food and feeding‐habits, widely dissimilar animals tend to show similar aversions and preferences for particular foods;and that distastefuhiess operates as a successful deterrent not merely against one or a few types of predator but over a very wide range of potential enemies. Raptorial (as opposed to other) birds appear to be especially sensitive to bitterness, which is a common attribute of the flesh of potential aposematic prey.31The fundamental nature of distastefuhiess as a deterrent is illustrated by various aposematic insects such asPhymateus, Zonocerus, Abraxas, Syntomeidaand others, which are variously rejected as relatively inedible by animals so different in habits and organization as ants, spiders, frogs, lizards, birds, bats, cats, dogs, mongooses, monkeys and man. These facts, taken together with the direct comparison of food preferences of the hornet, cat and man, suggest that the hornet experiments may afford indirect evidence of the relative edibility of the flesh, not merely to the hornets themselves but to meat‐eaters in general, includicuous coloration of distasteful birds subserves an aposematic function is derived from a study of their behaviour in the field. In general, aposematic animals, of whatever group, enhance the advertising effect of their conspicuous colours and bold patterns by specialized behaviour, both in general demeanour and in particular activities in the presence of an enemy. Many birds here considered exhibit the attributes of aposematic animals—freely exposing themselves to view, being apparently fearless or indifferent towards enemies, having slow or laboured flight, or uttering loud warning cries—and, in general, behaving as though to attract attention.33Aspects of the coloration and behaviour of desert birds are discussed. An adaptive interpretation is suggested for the black plumage of deserticolous chats, which is considered to subserve an aposematic function. The evolution of black and of isabelline coloration, and of correlated cryptic and sematic habits, in related species of desert birds, finds a parallel in the coloration and habits of certain desert insects. Evidence is adduced in support of an adaptive interpretation of the concealing and revealing coloration of desert birds, which afford a specialized example of phenomena widespread in other environments and in other groups of animals.34The functions of conspicuous characters in birds are reconsidered in the light of the present experiments and observations, and with special reference to the theory of warning colours, which may be widely extended in its application to birds as a class. It is suggested that insufficient weight has hitherto been allowed to the selective action of predation—as opposed to that of the reproductive environment—as a factor in the evolution of conspicuous plumage.35The inverse correlation, here demonstrated, between visibility of the plumage and edibility of the flesh suggests that the interspecific struggle for existence has played an important part in the evolution, not only of concealing, but also of revealing coloration among birds otherwise liable to attack. Selective pressure by pledators appears to have forced vulnerable species along two divergent lines of specialization; leading, in those which are relatively palatable, towards concealment, and in those which are relatively distasteful, towards advertisement.36Conspicuous species which arenotalso distasteful appear usually to possess other means of defence, and to be little subject to attack. Conspicuousness thus comes to be a general concomitant of relative unacceptability in birds, and is likely to be so associated in the mind of a predator. It is suggested that even where revealing characters subserve functions primarily other than warning, nevertheless their evolution is likely to have been promoted in the interspecific struggle as having a subsidiary aposematic function carrying a biological advantage not only in the sphere of reproduction, but also of s