Birds of the Thera Wall Paintings

Identifications are based on plumages, proportions, behaviour and modern ranges, with speculation as to late Bronze Age ranges. Six species and one additional genus are found; thirty-three of the birds are identified to the species level and two to genus. The historical and contemporary occurrence of these species on Santorini is discussed. The swallow Hirundo r. rustica is the commonest bird in the paintings, with fourteen individuals known, all portrayed in flight, and including two pairs. Possible explanations for the small racquets at the tips of the swallows' outer tail feathers are given, and the behaviour of the swallow pairs is examined. Finally, several artistic and iconographic observations are made.

INTRODUCTION 

The seven swallows of the 'Spring' fresco from Akrotiri are a magnificent achievement of Aegean Bronze Age art, unequalled for perhaps three millennia in their elegant, dynamic portrayal of bird flight. But they are not the only avian masterpieces in the Thera wall paintings. Swallow fragments exhibit the master's hand; a dove is caught in flight; ducks fly through reeds or along a river bank. What species are being portrayed? What can we learn from studying these birds that will provide insight into Theran life, or the second millennium BC ecology of Santorini, or historical avian distributions?

Morgan (1985, 6-7) has questioned whether modern zoological taxonomy adequately reflects the system of animal classification used by the Minoans, or any other culture. Instead, she suggested that iconographic interpretation, leading to insights into how ancient peoples viewed their environment, might be more valuable. Following this method, Vanschoonwinkel (1990) compiled a catalogue of animal representations in Theran art, without consideration of zoological identification, relying instead on iconographic analysis. But Morgan's reservations notwithstanding, the association of ancient images with modern taxonomic forms provides a context and a setting from which associations can be made and inferences drawn, around the world and back through time. Thus classification, while not an end in itself, is the starting point of this study of the Thera wall paintings from an ornithological perspective.

One example of the successful use of this method is Houlihan's catalogue of seventy-two species (or, in a few cases, genera) identified from the rich and detailed corpus of ancient Egyptian art (Houlihan 1986, 1-137). Another example is Ruuskanen's study of the birds depicted on Aegean Bronze Age seals, where many families are represented but where identification to species is usually impossible (Ruuskanen 1992, 52-63). The present work is more limited in scope, being confined to the birds depicted in the Theran wall paintings which, as we will see, comprise only seven species or genera.

In analysing the avian content of ancient art, we must keep in mind two limitations that the artist faced. First, no optical instruments were available to obtain a close view of distant birds. Killed or captive specimens would have been the only means to observe details of structure or plumage. Second, limitations of the artist's palette have to be considered. At Akrotiri the absence of green (Doumas 1992, 19) meant that other colours, such as black or yellow, had to be substituted. Shades of grey were apparently difficult to prepare, so blue was often used instead. Such a substitution led Evans (1928, 454, col. pl. XI) to misidentify as a roller (Coracias garrulus) the rock dove (Columba livia) portrayed in the 'Blue Bird' fresco at Knossos.

In Table 1 the published birds found in the Thera wall paintings are listed, with a code number for ease of discussion, exact location if known, avian type or family, a few words on the bird's attitude in the depiction, and a reference. Not including nestlings, there are thirty-five avian representations. The only surprise here should be W4, which has been called at various times a swallow or a dove, and X1-X6, which have usually been labelled as ducks.

Logically W4, W5-W12 and X1-X6 should not be included, for they are not paintings of birds but paintings of representations of birds. Nevertheless, we will treat these 'bird paintings once removed' in the same manner as the others, keeping in mind that they may be further from any naturalistic roots.

Still further from nature are creatures of the artist's imagination, which are found throughout avian art. Aside from the griffins, which we will not discuss here, there appear to be few flights of fancy in the Theran wall paintings. One exception to this is W2, with its strange blue and yellow coloration. Moreever, even this is probably the exaggerated depiction of a real bird.

IDENTIFICATION

Athough, as we will find, some of the birds depicted in the Thera wall paintings are not found on Santorini today, we will start here from the hypothesis that the birds depicted were local Theran species, rather than African birds either seen in Libya or Egypt by travelling Therans or painted on Thera by visiting Africans. Avian ranges have not been constant over 3600 years. Even in this century there have been major range changes, for example the north-westward expansion of the collared dove (Streptopelia decaocto) from the Balkans to southern Scandinavia and the British Isles over a forty-year period (Cramp 1985, 341-343). Either climatic changes or the activities of man usually drive such changes, including extinctions.

On a much longer time scale, the fossil record from islands in the Mediterranean from the middle and late Pleistocene shows relatively few extinctions, mainly large raptors, a large flightless swan on Malta, and some Corvidae (Alcover et al. 1992,280-282). However, a number of species, which now breed in northern Europe only, bred in the Mediterranean during the Pleistocene; for example, the northern gannet (Morus bassanus) and red-breasted goose (Branta ruficollis) have been found on Crete (Acover et al. 1992, 280; Weesie 1987, 76-80).

In attempting to match known taxonomic forms with the birds depicted in the Thera wall paintings we will attempt to take these considerations into account. The reader is warned, however, that identifying birds in ancient art is not an exact science.

WATERFOWL

       X7. From the proportions of bill, head, neck and wings, the bird (Pl. 16) is obviously a duck, and was identified as such when unearthed by Spyridon Marinatos (1976, 27, col. pl. L) in a fresco he called 'Thicket of Reeds'.

The white neck-ring immediately narrows the choice to the male mallard (Anas platyrhynchos) (see Fig. 1) or the Australian shelduck (Tadorna tadornoides). The shelduck, while resembling X7 in several other details (black head, conspicuous white wing patches) is not a possibility, since it is endemic to southern Australia (Madge and Burn 1988, 164 and col. pl. 12). The mallard, on the other hand, is distributed throughout the northern hemisphere, but does not occur on Santorini except possibly as a scarce migrant. However, it winters in northern Crete (Handrinos and Akriotis 1997, 118), and may well have occurred on seventeenth century BC Thera, at least as a winter visitor or migrant. For the mallard to have bred on Santorini, the climate would have had to have been considerably colder and wetter than it is now (see discussion below).

While X7 is clearly identifiable as a mallard, there are several discrepancies in the coloration. First, the male mallard's bill is yellow, not dark brown; this is an inexplicable error. Second, the male mallard's head is green, not black. This can be explained by the fact that, as noted above, the Theran artists did not know how to mix a green pigment. It can also be explained by noting that at a distance a mallard's head and neck can look black.

Continuing restoration of the 'Thicket and Reeds' (or 'Reed bed') fresco is revealing more ducks, perhaps seven altogether (see Vlachopoulos this volume). The most complete bird of this new lot is another male mallard, but in this mallard yellow-ochre instead of black is substituted for the green of the head above the clearly visible white neck-ring.⁽³⁾ Another, partly reconstructed bird appears to be a female mallard, flying right but looking left.

       A2. Although incomplete, A2 is obviously the same species as X7, with white neck-ring, black head, brown belly and extensive white on the wings, all characteristic of the Theran mallard. These mallards may be the earliest known depiction of this species, which was not in Houlihan's catalogue of ancient Egyptian birds (Houlihan 1986, 67-73), but does appear on the inlaid dagger blade from the roughly contemporary Shaft Grave V at Mycenae (Evans 1930, 113-117, col pl. XX).

       W1. This bird (Fig. 2a) has received much attention, being first identified as a duck (Marinatos 1974, 42), then a greylag goose (Anser anser) or possibly bean goose (Anser fabalis) (Morgan 1988, 63). The bird's proportions, however - apparently small bill, relatively short wings -and the strong impression of rapid, headlong escape flight, suggest a duck. Furthermore, all the geese that might have occurred in this region are dark under the wings, whereas W1 shows extensive white. The Egyptian goose (Alopochen aegyptiacus) (Madge and Burn 1988, 170-171, col. pl. 13), not a true goose but a sheld-goose, has white under the wings, but it is confined to the inner forewing. One other mark to note in W1 is the white underside of the tail, separated by a dark area from the lighter coloured belly. But which species of duck did the artist intend? One clue is the white eye, another the extensive white in the wing, together suggesting the genus Aythya. Possible species are ferruginous duck (A. nyroca) (Madge and Burn 1988, 252-253, col. pl. 37), tufted duck (A. fuligula) (Madge and Burn 1988, 255-256, col. pl. 39) and greater scaup (A. marila) (Madge and Burn 1988, 256-257, col. pl. 39).

The male ferruginous duck (Fig. 2b), has a bright white eye, extensive white on both surfaces of the wing, a rust-coloured (ferruginous) head and neck, white under-tail coverts, and a small whitish patch on an otherwise brownish belly. Its present breeding range includes parts of Greece, and it has bred on Crete in this century (Bauer et al. 1969, 39). Although the head and neck colour in W1 appears to be wrong for this species, at a distance the ferruginous colour may not be visible, and the head and neck will look dark.

The male tufted duck has a drooping tuft of feathers over the back of the head which a careful observer would not miss, but the female has an inconspicuous tuft, a yellow eye, dusky brown head and neck, a brownish belly and white under-wing. The female tufted duck is normally dark under the tail, but occasionally shows white in the autumn (Madge and Burn 1988, 255). It breeds in northern Europe, but its present wintering range includes the northern Aegean and possibly the Cyclades (Bauer et al. 1969, 39).

As for the final possibility, greater scaup, the male has a bright white belly, while the female shows white at the base of the bill, neither of which is exhibited by W1. However, the male goes into an eclipse plumage in the fall, in which the belly becomes brownish; it has a dark tail. Its present breeding range is extreme northern Europe, and it does not currently winter in the Aegean except in the extreme north (Handrinos and Akriotis 1997, 123).

Because of its dark tail, northern range and the need to assume an eclipsed male, the greater scaup can be safely eliminated. Therefore W1 is either a male ferruginous duck or a female tufted duck. Because identification as a tufted duck requires that the bird's plumage would have been atypical (white under-tail), and because of the ferruginous duck's more prominent white wing stripe and much closer modern breeding range, we identify W1 as a ferruginous duck. As far as I am aware, this is the oldest representation anywhere of this species.

       W2. Goose-like in form but sporting some bizarre features, W2 (Fig. 3) preens unaware of the approaching feline. S. Marinatos, upon discovering the frieze with this scene, speculated that the bird was an ibis or flamingo (Marinatos 1974, 42, col. pl. 8). But as Morgan (1988, 63-66) recognised, its proportions are goose-like. The long neck, long dark bill and thin white band around the base of the bill suggest instead swan goose (Anser cygnoides) (Madge and Burn 1988, 134, col. pl. 4). However, this is a bird of north-east Asia, not believed to have been brought west to become an ancestor of the domestic Chinese goose until a millennium later.

Nevertheless, the heavy body of W2 strongly suggests a domesticated or semi-wild goose. The strange left foot (Morgan 1988, 64) adds to this impression, as does the neck-ring. The latter might actually be a collar, to which a lead was once attached. More likely than swan goose is greylag goose, the commonest European Anser, a fairly long-necked goose with a distinct wing pattern and also a domestic goose ancestor. Other possible 'components' of the bird depicted in this painting are bean goose (Jonsson 1992, 78) and white-fronted goose (Anser albifrons). In summary, the most that can be said with a high degree of certainty is that W2 is a not completely wild Anser goose. Although not enough of W3 remains to be certain, it appears to be the same species.

CORMORANTS 

       X1-X6. Forming a necklace on the 'Mistress of the Animals', just above another necklace adorned with five dragonflies, are six ornaments in the form of birds (Pl. 13). Depicted in decorative rather than naturalistic colours of alternating deep orange, blue and pale orange, they appear to be swimming, have slightly raised wings, fairly long necks and thin, rather long bills. The bill is best seen on X5, while the body, wings and feet are best preserved on X2. Note the forward placement of the visible feet, suggesting somewhat long-legged birds and ruling out grebes (Podicipedidae) as a possibility. The thin bills are completely wrong for waterfowl (Anatidae), so what are these birds? All the features noted above, plus upturned head on X3, and even the head-to-tail linear arrangement, all point strongly to cormorants (phalacrocoracidae) (see Fig. 4).

Of the thirty or so species worldwide, by far the most probable is European shag (Phalacrocorax aristotelis) (Jonsson 1992, 57). First, the bill of the European shag is thinner and less noticeably hooked than that of other possible species, such as the great cormorant (Phalacrocorax carbo), consistent with the birds on the necklace. Second, the European shag, a bird of high cliffs and rocky coasts, is the only cormorant currently inhabiting the Cyclades or Crete, where it is widespread and fairly common. Finally, fossil remains of the European shag from the late Pleistocene have been found on Crete (Weesie 1987, 12), further strengthening the case that this species was present in the southern Aegean during the Late Bronze Age.

The reason for the unnatural orange and blue colours is clear: in its dark brown to black plumage, the shag is an ugly bird, unfit for the necklace of a goddess. Add bright colours, and we have a miniature masterpiece.

Some comments on the dragonflies (see Pl. 13): the shape of the wing tips is completely wrong - they should be rounded, not pointed. If we ignore that problem, then general proportions, particularly the enlarged abdomen, indicate that they belong to the family Libellulidae. One guess, based on wing and abdomen markings and general coloration, is Libellula fulva (d'Aguilar et al. 1986, 269-270, col. pl. 19). If the necklace ornaments were not imports, and the dragonflies were indigenous, there is an implied presence of standing water, which dragonflies require to breed. This is consistent with the finding by Rackham (1990), based on the methods of historical ecology, that Santorini was somewhat less arid 3500 years ago than it is today.

FALCON

W4. Identified first as a swallow (Marinatos 1974, 47-49), then as a probable dove (Morgan 1988, 66), this prow emblem (Fig. 5a) in fact resembles neither. The bill appears to be hawk-like and open. The wings are narrow, pointed and bent at the carpal joint. The tail is fairly long and thin and without adornment. These features, together with the bird's overall proportions, strongly suggest a small to medium-sized falcon. In the absence of any clear plumage markings, which one is it?

The possibilities are lesser kestrel (Falco naumanni), Eurasian kestrel (F. tinnunculus), red-footed falcon (F. vespertinus), merlin (F. columbarius), hobby (F. subbuteo), Eleonora's falcon (F. eleonorae), sooty falcon (F. concolor), lanner (F. biarmicus) and peregrine (F. peregrinus) (Cramp 1980, 277ft.). Other falcons (there are forty-five species worldwide) were not considered because they are either completely out of range or too massively built to be W4. Of these nine species, we can immediately eliminate F. biarmicus and F. peregrinus because their wings are much too broad. F. concalor, a bird of north-east Africa and the Middle East, does not occur in Greece, and probably never did (Handrinos and Akriotis 1997, 309). F. columbarius has rather short, broad wings; F. subbuteo has too short a tail; F. vespertinus is an inland, not a coastal species.

Thus we are left with lesser kestrel, Eurasian kestrel and Eleonora's falcon from which to choose. Of these, two things point to Eleonora's falcon (Fig. 5b) as the most likely model for W4: first, its wings are longer and thinner than those of either kestrel (the wing tips of both kestrels are slightly rounded). Second, its habitat - wild Aegean island cliffs - would seem more in keeping with its use as a ship's emblem than the cultivated fields and pastures that are the preferred habitat of the lesser kestrel and to some extent the Eurasian kestrel (Jonsson 1992, 156-158), although I have seen Eurasian kestrels with Eleonora's falcons over the near-shore waters at Phira. Eleonora's falcon is today locally common in the Cyclades and Crete (Handrinos and Akriotis 1997, 145-146), breeding on many Aegean islets, and was also found in the fossil record from the late Pleistocene on Crete (Weesie 1987, 25). It is therefore highly probable that this species was familiar to the inhabitants of Late Bronze Age Thera.

We conclude that the prow emblem W4 is a representation of an Eleonora's falcon. It is possible, however, that close inspection of the original painting could reveal some new detail that could alter this conclusion. I believe that this emblem is the earliest known representation of Eleonora's falcon.

DOVES

       A3. S. Marinatos (1969, col. pl. B2) speculated that the fragment he illustrated from Arvaniti 1 (Sector Alpha) might be part of an eagle. However, further reconstruction (Fig. 6a) reveals the small head, chunky body, relatively short wings, fanned tail and blue colouring of an Aegean-style dove. Except for the outer left corner, the tail is intact and shows a brown terminal band and sub-terminal white outer tail feathers. Although the white outer tail feathers are not usually seen, this pattern is that of the rock dove (Cramp 1985, 285ff., col. pl. 29), viewed from above (Fig. 6b:1). Further corroborating the identification as a rock dove is the white rump, visible on either side of the tail-body crack in Fig.6a.

But what are we to make of the extensive white in the wings, since the upper surface of a rock dove's primaries is dark? The answer is that the artist has chosen to depict the bird in a twisted view, showing the largely white undersides of the wings (Fig. 6b:2) but the upper side of the tail. Immerwahr (1990, 141, pl. 81) noted this same perspective for the 'Bluebird' frieze at Pylos, several centuries after Akrotiri.

The rock dove is commonly found today throughout the Cyclades and Crete (Handrinos and Akriotis 1997, 198), and has also been found in the Cretan fossil record from the late Pleistocene (Weesie 1987, 32). Thus there is little doubt it was resident on Late Bronze Age Thera.

       W5-W12. Decorating the starboard side of the only ship under full sail in the West House flotilla are at least five and perhaps eight pigeon-like birds in full flight. The narrow, dark terminal tail band indicates rock dove, but what about the white in the primaries and secondaries? Again, we have a twisted view, showing the undersides of the wings but the upper side of the tail.

SWALLOWS

       A1. The bird A1 (Fig. 7), which has been variously attributed to Beta 6 (Marinatos 1970, 64, col. pl. B1) and Sector Alpha (Doumas 1992, 184), is obviously a swallow, with small bill, long thin wings and long outer tail feather (the left outer tail feather is missing). The extreme length of the tail feather points to the genus Hirundo, of which two members might be expected in the Mediterranean region: barn swallow (H. rustica) (Turner and Rose 1989, 164-169, col. pl. 15) and red-rumped swallow (H. daurica) (Turner and Rose 1989, 201-204, col. pl. 20). The white spot(s) on the tail is characteristic of H. rustica but absent in H. daurica, so we conclude that this is indeed a barn swallow. Confirming this identification is a red patch on the throat.

Two features illustrated on this barn swallow are not quite accurate. First, the brownish colour of the head, back and wings is that of a juvenile swallow; adults show much more blue. Yet the long outer tail feather, or streamer, does not grow in until late in the post-juvenile moult (Cramp 1988, 275-276). One possibility is that the artist had observed juvenile barn swallows at close range, perhaps shortly after their fledging, but having no binoculars was uncertain of the colour of the adults' upper parts. The other anomaly is the appendage at the tip of each streamer, a signature of Theran swallow artists which will be discussed in detail below.

       D1-D7. The seven swallows of the 'Spring' fresco (Marinatos 1971, 49-52, pls. 121-126, col. pls. A-C) have captured not only the hearts of viewers of the paintings but also the attention of archaeologists and art historians, who have attempted to find the behavioural basis of the aerial acrobatics so masterfully depicted (Pl. 17). In this section we will discuss some points on form, colour and distribution, saving discussion on behaviour for below. In contrast to A1, all seven birds are depicted as if viewed from below, which is just how they would be seen if painted from life, in flight. Markings on all seven birds are nearly the same: they are largely white below, with white under the wings reaching nearly to the wing tips, red on throat and face, white spots or lines near the tips of the inner tail feathers, and long outer tail feathers forming streamers.

Based upon white markings on swallows on Akrotiri pottery, S. Marinatos (1969, 14) believed that the Theran swallows were an artistic hybrid between the barn swallow and the house martin (Delichon urbica). However, it was clear to him that the basis of the 'Spring' fresco swallows was Hirundo rustica (Marinatos 1971, 52), and this remains undisputed.

These barn swallows belong to the west Eurasian and central north-African race H. r. rustica. The non-migratory Levant race, transitiva, and Egyptian race, savignii, have darker underparts, and the latter has dark under-wing coverts (Cramp 1988, 262). In H. r. rustica, the outer tail feathers are much longer in the male than the female, and shortest of all in the juvenile (Figs. 9a and 9b). The most useful measurement for determining sex and age is the depth of the tail fork. Mean values are 62 mm. in the male, 44 mm. in the female, and only 23 mm. in the juvenile (Cramp 1988, 276). Examination of D1-D7 suggests that D1-D6 are adult males, and D7 an adult female (Pl. 18). It is a tribute to the skill and accuracy of the artist that it is even possible to speculate as to the age and sex of these birds.

One deviation from strict, naturalistic representation, however, is the extent of white in the wings, which runs nearly to the wing tips in D1-D7. Compare Pl. 17, which shows D4 and D5, with Fig. 8. I believe the artist intentionally extended the white wing linings because he thought it looked more striking, and indeed it does. An even less natural flourish, mentioned above on A1, is the small loop, or racquet, at the tip of each streamer, which will be discussed below.

The question of the occurrence of the barn swallow on Thera has received some attention, beginning with the widely quoted statement of S. Marinatos (1971, 51-52) that swallows are "non-existent on the island today". The barn swallow is a widespread and common breeding bird throughout Greece (Handrinos and Akriotis 1997, 223). Several ornithologists have visited Santorini and published their findings: Watson (1964, 107) paid a brief visit on 9 May 1954, as part of a study of the passerine birds of the Aegean, and observed barn swallows, which he included in his list of breeding birds of Santorini, but noted that a visit later in the season might result in this species being subtracted from the list, along with several other suspected passage migrants. However, he concedes that 9 May is a very late date for a migrant barn swallow.⁽⁴⁾ From 2-5 May 1892, Douglass (1892, 454) observed the fauna of Santorini and recorded barn swallow. Earlier still, Erhard (1858, 54), who spent many years in the Cyclades, listed barn swallow as a summer resident; unfortunately, he failed to note on which islands he found it.

Thus, although it is possible that the barn swallow does not breed anywhere on Santorini today, it almost certainly has in the recent past. Finding mud to use in nest building is probably the limiting factor, although barn swallows have been known to use marine algae as a substitute (Duffin 1973, 237-238). In any event, there can be little doubt that the barn swallow bred on ancient Thera, since the swallows D1-D7 give every appearance of having been painted from life. Consistent with this conclusion is the finding by Rackham (1990) that 3500 years ago Santorini was wetter than it is today.

       X8-X13. These fragments from Xeste 3, Room 4 (Doumas 1992, 128), but described incorrectly by S. Marinatos (1976, 27, pl. 39b) as being found in Xeste 3, Room 2, are again barn swallows. Swallow X9 is particularly interesting, in that it is shown feeding oddly coloured nestlings (yellow-ochre instead of black) in a bowl-like nest perched on a small rock ledge.

SUMMARY 

Of the thirty-five avian forms depicted in the published Theran wall paintings (not including nestlings), thirty-three have been identified to species and two to genus. Six species are represented, plus one additional genus, belonging to five families, as summarised in Table 2. The mallards, rock dove A3 and all of the barn swallows are positive identifications. The remaining assignments of species are presented here with a high degree of certainty, but study of the original wall paintings could result in changes. The occurrence of all of the identified species on Late Bronze Age Thera is highly probable.

DISCUSSION 

       Swallows

The racquets at the ends of the swallow tail streamers appear to be unique in avian art and may be considered to be the defining feature of the Theran swallow. Polinger Foster (1995, 413) noted that the streamers in the wall paintings terminate in open loops, whereas in the swallows depicted on Theran pottery the loops are solid. However, close examination of the streamer tip on A1 reveals a split tip, with an indistinct loop drawn around the split ends. This suggests the possibility of a natural explanation for the streamer racquet on at least one swallow, which might have been copied on the others. Looking for deformed streamer tips, I examined all 113 specimens of Hirundo r. rustica at the Harvard Museum of Comparative Zoology^,(5) and found none, aside from slight damage to a few that had obviously occurred in handling or storage. Extending the search to other subspecies, I examined the 374 remaining specimens of Hirundo rustica, and found one, an H. r. erythrogaster from Brazil, with the last 5 mm. of each outer tail feather bent outward about 30 degrees (Fig. 9b, right). It would appear, then, that damage or congenital deformation is a rare event, unlikely to have been the inspiration for the tail racquets in the Theran swallow paintings.

Is it possible that swallows on ancient Thera were a species now extinct that actually had such tail ornaments? Ernst Mayr, a leading authority on avian evolution, does not believe that such a bird could have existed 3600 years ago. Only twenty-nine species worldwide (out of about nine thousand) have such appendages: nine species in the parrot genus (Prioniturus) from the Philippines and Sulawesi, three South American hummingbirds (Trochilidae), five species of paradise kingfisher (Tanysiptera) from New Guinea, a roller from south-eastern Africa (Coracias spatulata), two bee-eaters (Merops), one from south-eastern Africa the other from Australia, five species of motmot (Momotidae) from the neotropics, two species of drongo (Dicrurus) from south Asia and two birds of paradise (Paradisaeidae) from New Guinea. There are no swallows so adorned, and it is not plausible that there ever were. Mayr (who is familiar with the wall paintings) thinks that the tail racquets were painted there because the artist felt that the swallow should have them, that they just looked better than bare outer tail feathers.⁽⁶⁾

I propose that the Thera swallows have racquets on their tail tips because one artist, the creator of the 'Spring' fresco, thought that the racquets improved upon nature, adding balance and grace to the birds' flight. The racquets were his 'signature'. However, subsequent artists were not close observers of swallows in flight, and simply copied the master.

The behaviour of the swallows depicted in the 'Spring' fresco, particularly the pairs D1-D2 and D4-D5, has been an ongoing source of controversy. S. Marinatos (1971, 52) thought they were mating, but there are two problems with that hypothesis. First, as discussed above, all four of these birds are probably males. Second, the face-to-face positions depicted are not characteristic of swallow mating behaviour. In sexual display ('courting'), the male fans out his tail in a series of swoops, and the female, if interested, follows him to a potential nest site (Møller 1994, 77-78). Copulation, which occurs only after formation of the pair bond, takes place with the birds perched, not in flight (Møller 1994, 78-79).

Hollinshead (1989, 341-343) recognised that bill-to-bill contact was not mating behaviour, and suggested instead that the pairs were adults feeding young. But both birds of each pair are adults: post-juvenile moult may start on the breeding grounds, but is completed in the winter quarters in Africa, and the tail feathers are the last to be fully grown (Møller 1994, 75-76). Nor is there any sign of food in the paintings, or of a gaping bill on one bird only of each pair, which would be the case if young were being fed.

Polinger Foster (1995, 413-414) rejected both of these interpretations of the behaviour of the Delta 2 swallows, and proposed instead that the birds in each pair are engaged in aggressive display. Citing the signs of high-intensity swallow display - extended head and neck, gaping bill, extended legs, flexed claws, lunging at the opponent's throat - she makes a strong case for this behaviour. I had reached the same conclusion independently. Such display has been studied by many observers,⁽⁷⁾ most thoroughly by Møller (1994, 78-79), who observed many entangling encounters (Fig. 10), sometimes ending with both birds falling to the ground with feet interlocked, where they may fall prey to cats.

One question remains: what are these birds fighting over? Polinger Foster (1995, 414ff.), citing some recent studies by Winkler (1993a, 29; 1993b, 30-34) of feather-fighting in the tree swallow (Tachycineta bicolor), extrapolates this behaviour to the barn swallow and concludes that the Delta 2 swallow pairs are engaged in aerial battle over feathers, which the barn swallow uses for nest lining. First, from this conclusion, she identifies as feathers several objects in the airspace around swallow-like birds in Minoan and Mycenaean gold rings (Polinger Foster 1995, 418-419). Second, she suggests that the holes in the west and north walls of Delta 2 once held pegs from which were hung hooks, and from those hooks near the swallows D1-D5 were suspended feathers.

Being sceptical that feather-fighting is as prevalent in the barn swallow as Polinger Foster hypothesised, I asked David Winkler if he has observed it in the North American barn swallovv, H. rustica erythrogaster. He has seen this behaviour, not as persistently as in the tree swallovv,⁽⁸⁾ but recommended I contact Anders Møller, undoubtedly the world's leading expert on H. r. rustica. Møller replied that⁽⁹⁾

The most common reason for fights is males fighting at the territorial boundaries, mainly over females...Fighting over feathers is uncommon. Barn swallows do take a lot of interest in individuals carrying feathers, but it is not clear why. It is mainly females that carry feathers, and males often show a great interest in such females because they are fertile. More than 25% of all offspring are fathered by males other than the mate in this socially monogamous species.

Thus there is no ornithological basis for assuming that feathers were involved in the D1-D2 or D4-D5 swallow pairs: either one of the birds in these is an intruder, or they were fighting at the common border of their territories.

ART AND ICONOGRAPHY

How many artists might there have been painting birds on the walls of Akrotiri? The swallows D1-D7 and at least four of the Xeste 3 swallows (X8-X11) appear to have been drawn by the same hand, with fluid  acrobatic flight and exaggerated white wing linings. The swallow A1, on the other hand, has long, stiff wings shown in top view, a too small head, and is dearly the work of a second artist. Based upon wing and tail feather delineation and overall style, a third artist was probably responsible for the ducks A2 and W1 and the dove A3. The artist who depicted the cormorants X1-X6 appears to me to be the same one who drew the geese W2 and W3, having in common fanciful colours, sweeping curves and similar nearby griffins. I suspect he also painted the duck X7. Yet another artist appears to have done the doves W5-W12, with their crude draftsmanship, and presumably the falcon W4. Thus the birds of the Thera wall paintings were the work of perhaps four or five different artists.

Do the bird paintings tell us anything about where these artists and their subjects might have come from? Since all species depicted (except possibly for the Anser geese, which I do not believe are wild birds) were probably indigenous to seventeenth century BC Thera, there is no reason to doubt that both the artists and the birds were Theran.

Several peculiarities of the Thera avian art should be noted. First, all but seven of the birds are portrayed in flight. Standing birds predominate in ancient Egyptian art, although there are many examples of flying ducks, vultures, kingfishers and other birds. Why did the Thera wall painting artists take on this more difficult task? Could it be that birds' flight was held in great awe by the Therans, that flying was considered the essence of avian existence? The birds that are not flying are either swimming or preening, not simply standing. Compare, for example, the Delta 2 swallows with the pair of Egyptian barn swallows from Deir el-Bahri, of the Twenty-first Dynasty (Houlihan 1986, 122-124).

Second, some avian proportions are distorted in ways that appear to be deliberate. In the flying mallard X7, the wings and body are foreshortened relative to the head and neck. I suggest that the artist wanted to give the impression of a head-on view but still draw the bird in profile. The opposite disrortion can be seen in the barn swallow A1, a unique top view of a barn swallow. The wings are too long and the head is too small, reflecting the artist's impression that a flying swallow is mostly wing.

Third, there is a consistent exaggeration of the extent of white in depicting the wing linings of birds. Whether the mallards A2 and X7, the ferruginous duck W1, the rock doves A3 and W5-W12, or the barn swallows D1-D7, X8, X10 and X11, there is too much white on the underside of the wings. Perhaps this has some particular significance in the Theran view of bird flight, perhaps it was a convention observed by the artists of Thera, or perhaps it was simply how the wings looked in bright sunlight.

Finally, with lions, wild cats, even griffins pursuing their prey, why is there such a paucity of raptorial birds? Were eagles, hawks and owls scarce on ancient Thera? In the Cretan fossil record, birds of prey abound (Weesie 1987, 72ff.). Are there paintings of other avian raptors yet to be discovered, or is the prow falcon W4 the only one? Perhaps the Therans rejected the concept of avian aggressor, and as prow emblem the falcon might have represented speed and purposefulness more than aggression. The swallow pairs D1-D2 and D4-D5 exhibit controlled, programmed, ritualistic aggression, much like the 'Boxing Boys' and 'Antelopes' of Beta 1 (Marinatos 1984, 106-112).

SUMMARY

An attempt was made to identify the birds depicted in the Theran wall paintings. The study was done with reference to ancient avian art, earlier Theran studies and contemporary knowledge of plumages, distribution and behaviour. Species identified were European shag, mallard, ferruginous duck, Eleonora's falcon, rock dove and barn swallow, plus an Anser goose. The unusual swallow tail racquets were concluded to be the product of one artist's aesthetic sense and imagination. The swallow pairs in Delta 2 are males engaged in aerial combat in defence of their territories, a motif echoed in the 'Antelopes' and 'Boxing Boys' of Beta 1. With the possible exception of the goose, these species were probably indigenous to Thera. This supports two hypotheses: that the artists and the birds were Theran, and that the climate on Thera was somewhat wetter than it is now. Three of these species (two ducks and falcon) may be their earliest recognisable depiction anywhere. While meagre in quantity when compared to the vast Egyptian corpus of avian art, there was a very high level of artistry at work here, and we must thank the Therans for leaving us such a valuable and inspiring record.

(1).      Bird paintings listed from north to south as found at Akrotiri. A = Sector Alpha (Arvaniti 1), W = West House, D = Building Complex Delta, X = Xeste 3.

(2).      Preferred reference given here is Doumas 1992 because of the high quality and scale of there productions.

(3).      I thank Prof. C. Doumas for the opportunity to view this restoration.

(4).      George E. Watson, private communication.

(5).      Harvard University, Cambridge, Massachusetts, USA.

(6).      Ernst Mayr, private communication.

(7).      For summaries see: Glutz von Blotzheim and Bauer 1985, 438-440; Cramp 1988, 270; Turner and Rose 1989, 20-22, 167.

(8).      David W. Winkler, private communication.

(9).      Anders Pape Møller, private communication.

---------------------------------------------------

---------------------------------------------------