Upper Pleistocene and Holocene fossil avifauna from Moleta Cave (Mallorca, Balearic Islands) Bartomeu SEGuí, Cécile MOURER-CHAUVIRÉ & Josep Antoni ALCOVER SHNB Seguí, B., Mourer-Chauviré, C. & Aleover, J.A. 1997. Upper Pleistoeene and Holoeene lossil avilauna Irom Moleta Cave (Mallorca, Balearie Islands). BolI. Soe. Hist. Nat. Balears, 40: 223-252. ISSN 0210-260X. Palma de Mallorca. The assemblage 01 lossil birds Irom Moleta Cave (Muleta Cave) is deseribed. The site was diseovered in 1962 and exeavated by Waldren (1982). Paleontologieal material s eovered a range Irom 32,000 to 2,000 years b.p. At the time 01 the exeavations, bird remains were not studied. Between 41 and 50 avian taxa have now been reeognized. Seventeen are new lor the gymnesie lossil record (Brantal Anser, Aleetoris el. rufa, Alauda arvensis, Anthus pratensis, Phoenieurus oehruros, Saxieola rubetra, S. torquata, Oenanthe oenanthe, Sylvia el. sarda, Sylvia sp. el. S. melanoeephalaleurruea, Fieedula hypoleuea, Passer sp. el. domestieus/ SOCIETAT O'HISTORIA hispaniolensis, Petronia petronia, Carduelis ehloris, Tadorna sp., Calandrella sp., NATURAL DE LES BALEARS Phylloseopus sp. and Sturnus sp.). Six 01 them are al so new lor the insular Mediterranean. In eontrast with Upper Pleistocene avilaunas 01 other Mediterranean islands, passerines are extremely abundant while proeellariiforms and birds 01 prey are underepresented. Unlortunately, available stratigraphie data do not allow strong eonclusions on the eolonization 01 the island by anthropie launa in the Holoeene. Keywords: Moleta Cave, Mallorca, Upper Pleistoeene, Holoeene, paleornithology. AVIFAUNA FOSSIL DEL PLlSTOCE SUPERIOR I HOLOCE DE LA COVA DE MOLETA (MALLORCA, ILLES BALEARS). L'objeetiu d'aquest treball és donar a eoneixer el registre aviari lóssil de la Cova de Moleta. El jaeiment lou deeobert el 1962 per Waldren (1982). Els materials paleontológies eontinguts enregistren un període de temps compres entre els 32.000 i 2.000 anys A.P. Rere I'exeavaeió, els materials paleornitológies romangueren sense estudiar. En el moment actual, entre 41 i 50 taxons aviaris han estat determinats. 17 són nous per a les Gimnesies (BrantalAnser, Aleetoris el. rufa, Alauda arvensis, Anthus pratensis, Phoenieurus oehruros, Saxieola rubetra, S. torquata, Oenanthe oenanthe, Sylvia el. sarda, Sylvia sp. el. S. melanoeephalaleurruea, Fieedula hypoleuea, Passer sp. el. domestieus/hispaniolensis, Petronia petronia, Carduelis ehloris, Tadorna sp., Calandre/la sp., Phylloseopus sp. and Sturnus sp.). D'aquests, sis són també nous per a la Mediterrania insular. A dilereneia del que sueeeeix en altres iIIes del nostre mare geogralie, els passerilormes són extremadament abundants, mentre que els proeeHarilormes i els rapinyaires estan poe representats. Malauradament, tot i que el jaeiment enregistre un període de temps estens, les dades estratigraliques no permeten extreure massa eonelusions sobre la eolonitzaeió de I'illa per la launa antrópiea durant I'Holoee. Para u/es e/au: Cova de Moleta, Mallorca, Plistoee Superior, Holoee, paleornitologia. Bartomeu SEGUí, Departament de Cieneies de la Terra, Universitat de les /l/es Balears, Carretera de Valldemossa km 7,5, E-07071 Ciutat de Mallorca. Céeile MOURER-CHAUVIRÉ, Centre de Paléontologie Stratigraphique et Paléoéeologique. CNRS, UMR 5565, Université Claude-Bernard, Lyon-I. 27-43, boulevard du 11- 224 Bo/l. SOCo Hist. Nat. Balears, 40 (1997) Novembre- 1918, 69622 Vifleurbanne Cedex. France. Josep Antoni ALCOVER, Institut Mediterrani d'Estudis Avan9ats-CS/C-UIB. Carretera de Valldemossa km 7,5, E-07071 Ciutat de Mallorca. Recepció del manuscrit: 18-nov-97; revisió acceptada: 1-des·97. Introduction Moleta Cave was discovered in of Sóller (Fig. 1). The cavity. is a karst 1962 and intensively studied, along with dissolution system which has developed Abric de Son Matge, as part of a doc in Jurassic materials. The galleries are toral dissertation (Waldren, 1982). The disposed on two superposed levels con latter discusses the excavation process, nected by a vertical pitfall (Fig. 2). Most and presents and interprets the nume of the excavated sectors are in the rous paleontological and specially lower part of the cave. The upper part, archeological data collected. Neverthe where one finds the only two openings, less, most of the fossil microfauna was both of which are small, consists of a ignored. The author was aware of the horizontal gallery. 7 m from the main relevance of this remains and at the entrance is the chimney which has a time considered their possible future vertical drop of 3 m. As discussed fur analysis (Waldren, 1982). ther below, this structure is important The present contribution is based for understanding the taphonomic pro- on avian material found amongst the vertebrate microfauna. This was 40° 2"40' 3"20' 3°40' partly examined by Dra. Cécile Mourer-Chauviré and some determina- tions were published by Alcover et al. (1981). Most of the material, however, remained un published. We herein report on the fossil or nithofauna from Moleta Cave. 39°20' o 10 20km Localization and description 2"20' ~40' 3° 3°20' Moleta Cave is in the Serra de Tramunta Fig. 1. Location of Moleta Cave (Mallorca, Balearic Islands). na, in the municipality Fig. 1. Localització de la Cova de Moleta (Mallorca, Balears). B. Seguí et al., Fossil avifauna from Moleta Cave 225 CAVE UF MULETA' PLAN MULETA· SECTlON.LOWER CAVE. E.W. Fig. 2. Topography 01 Moleta Cave (Irom Waldren, 1982). Fig. 2. Topografia de la Cova de Moleta (tret de Waldren, 1982). 226 Bol/. SOCo Hist. Nat. Balears, 40 (1997) esses which gave rise to the fossi X 200-300 and Z 200-300, are those liferous deposit. which register the arrival of anthropic fauna. The lower ones of the latter two levels appear to be prehuman and ca. Chronostratigraphy and age 15,000 years old. of the materials For a parallel discussion on chrono logical and stratigraphic problems of The' Moleta Cave displays a Moleta Cave, see Marcus (in press). cotinuous stratigraphic series of ca. 105 years (Waldren, 1982). Various periods have been characterized on the basis of Sediment deposition stratigraphic, paleontological and archeo and taphonomy logical data. The Presettlement Period (32,000-7,000 b.p. sensu lato, up to There are useful data on the level 7; Waldren, 1982: 43) includes taphonomy and sedimentology of the most of the avian material, although fossiliferous deposit. On the basis of some belongs to the oldest part of the geochemical data and paleomagnetic Early Settlement Period (7,000-2,000 b. datation, 0 4 and racemization of aspartic p. sensu lato, level 7 and 6), where acid, it is known that sediment deposi human presence in the cave is first re tion started during the Riss Glaciation, corded. The material studied herein is 230,000 years b.p. Nevertheless, the however problematic, as the correlation main accumulation took place ca. 80,000 between the artificial levels determined years b.p. The fossil remains of by fossil extraction with the natural ones Myotragus balearicus were deposited described by Waldren (1982) is not mainly during the last 32,000 years clear. The artificial levels contain sedi (Waldren, 1982). mentary layers up to 1 m thick, and The physical structure of the cave there are no data on the dip of the seems to be of major importance for strata. Hence, the great caution which explaining the recovery of a large the valuable information gleaned from number of specimens of M. balearicus. 7 datations at specific points must be in m from the opening believed to be the terpreted. An approximated age has main entrance, the gallery connects with been estimated for the excaved materi a natural well 3 m deep. The small dia als (Table 1). We consider O 100-150, meter of the opening and of the gallery CD 100-150 and E 100-150 coetaneous interfere with the access of light and with Man, while Z 700 and Z 750 ap probably animals fe" in it once they pear to be prehuman, with an antiquity had entered the cave (Waldren, 1982). close to 30,000 years b.p. E 150-250, X This sort of natural trap has al so been 200, X 200-300 and Z 200-300, theoreti recognized at other sites (e.g. Cueva del cally prehuman, contain anthropic fauna. Viento, Tenerife; Rando & López, 1996). Remains of Apodemus sylvaticus have One may see a parallel situation in vol been recovered, and other workers al so canic tu bes, where vertebrate remains describe anthropic remains in other sec tend to accumulate, and which in fact tors at ca. 200 cm (F 200, O 200; are key sites for knowledge on oceanic Adrover, 1966). We suggest from this paleocomunities worldwide (e. g. Olson that the lower layers of E 150-250, & James, 1991; James & Olson, 1991; which correspond to the upper ones of Rando, 1995; Rando & López, 1996). B. Seguí et al., Fossíl avifauna from Moleta Cave 227 Here light conditions vary on a longitudi archipelago (data from GOB 1996, nal axis and visibility drops with the di s 1997), as well as its fossil record in the tance to the opening. A critical point is Western Mediterranean islands. The recognized, from which visibility is nil minimum number of individuals (MNI) (Fiol, 1996). From here onwards, one present at the fossil site is given for finds most of the fossil remains, be long each species, corresponding to the maxi ing to individuals which have lost their mum number of skeletal elements from orientation. Thanks to its tri-dimensional the same side of the body registered. structure, the Moleta Cave reproduces, at small scale, the conditions in volcanic tubes, which take place in one direction Systematic paleontology over hundreds or thousands of meters. Order ANSERIFORMES Family ANATIDAE Material and methods Branta/Anser The fossil material referred to MATERIAL: O 100-150: MNCM 2658, herein is kept at the vertebrate collection medial fragment of radius. MNI= 1. "Museu de la Naturalesa de les IIles It is very difficult to difterentiate Balears" (acronym MNCM). between the radii of Anser and Branta For its determination, material has (Miller, 1937; Woolfenden, 1961; Olson been matched with present-day skeletons & James, 1991). The bone belonged to in the same collection. Specimens from an adult animal, and its shaft diameter the United States National Museum (WD=3,60) fits well within the range (Natural History; USNM) have also be en found in B. bernicla, while it is rather compared. smaller in comparison with B. leucopsis In the case of non-passeriforms, all (Bacher, 1967). Nevertheless, important skeletal elements have been studied similarities were stated with material we except ribs and vertebrae. With regard were able to examine belonging to a to passeriforms, usually only the cra possible new Anserini from Eivissa nium, humerus, coracoid and tarsometa (Ibiza). The description of this material tarsus have been studied. The rest of (McMinn, in prep.) is not yet finished, the postcranial generally lacks characters thus strong conclusions on the material of enough diagnostic value (Weesie, discussed may not be stated. 1988). The anatomical terminology fo lIows Baumel (1979). Terms used by Genus Tadorna Howard (1929) are sometimes used. Tadorna sp. The measurements referred to herein were taken with a caliper graduated to MATERIAL: X 200-300: MNCM 47050, 0.05 mm, as in Seguí (1997, see this fragmented synsacrum. MNI= 1. volume). This fossil belongs to a middle-sized From a geographic point of view, Anatidae. Important similarities stand out Gymnesic Islands refer to Mallorca and both with Tadorna tadorna and Anas Menorca, while Balearic Islands includes plathyrhynchos, the biggest species in both Gymnesic and Pityusic Islands. the genus. Despite of the synsacrum After determination of each taxon, being fragmented, it is clearly seen that there is a brief reference to the present the ridge that originates in the Extremitas status of the species in the Balearic cranialis synsacri and extends ventrally 228 Bo/l. SoCo Hist. Nat. Balears, 40 (1997) through the Corpus vertebra e of the data is correct, the material proves its toracic vertebrae reaches at least the presence in the island long befo re his second one. This situation is also present torical introductions. in all T. tadorna examined. Against this, It cannot be excluded that the re in A. platyrhynchos the ridge reaches just mains belong to other circummedi the first vertebra. The size of the fossil is terranean species of the genus Alectoris, slightly bigger than the latter species, fit for example A. barbara (Bonnaterre), A. ting better with T. tadorna. It has been graeca (Meisner) and A. chukar (J. E. impossible to examine any skeleton of T. Gray), with whom comparison was not ferruginea, and thus the material is attrib possible. In fact, any of them could have uted to Tadorna sp. been introduced in the Balearic Islands This is the first record of the taxon in the past (Blondel, 1987). for the Gymnesic Islands. T. tadorna has also been found in the Pleistocene of Genus Coturnix Corsica (Alcover et al., 1992). Nowa Coturnix coturnix (L.) days, both T. tadorna and T. ferruginea are rare in the Balearic Islands. MATERIAL: Sector CD 100-150: MNCM 41005, right tarsometarsus. MNI= 1. Order GALLIFORMES These remains were obtained from Family PHASIANIDAE the first human occupation levels. A Genus Alectoris close form was found in the Gymnesic Alectoris cf. rufa. Islands from Plio-Pleistocene and Lower Pleistocene sites (Alcover et al., 1981). MATERIAL: CD 100-150: MNCM 2922, Nowadays C. coturnix is a summer visi distal fragment of right radius; O 100- tor as well as migrator in the Balearics, 150: MNCM 2651, distal fragment of beeing partially sedentary in Mallorca. right radius; MNCM 2652, distal fragment of right tarsometatarsus. MNI= 2. Order GRUIFORMES A. rufa seems not to be autoch Family RALLlDAE tonous in the Balearic Islands, beeing Genus Porzana absent from all prehuman insular Medi Porzana porzana (L.) terranean faunas (Alcover et al., 1992). Its introduction in Mallorca is first docu MATERIAL: AB: MNCM 2951, distal frag mented for the XIV century (Ferrer et ment of right humerus; MNCM 2482, al., 1986). distal fragment of left metacarpus; Z 200- The MNCM 2922 fragment shows 300: MNCM 47056, proximal fragment of some incisions which seem to have right metacarpus; MNCM 47057, proximal been made with a sharp tool. AII mate fragment of left coracoid. MNI= 1. rials from CD 100-150 and O 100-150 In spite of the fragmentary state of sectors seem to belong to the level 7 the remains, they share all features with ("Early Settlement Period Horizons", Porzana porzana. A close form was Waldren, 1982) from which some found in Pedrera de S'Onix (Alcover et datations are available: SM54 (0175)= al., 1981), from the Pliocene-Pleistocene 3985 b.c.+/- 109 years; SM56 (0150)= boundary. The species was also recov 5185 b.c.+/- 80 years and SM57 ered in Crete (Weesie, 1988) and in (0150)= 6620 b.c.+/- 350 years Corsica (Alcover et al., 1992). The ac (Waldren, 1980). If these stratigraphic tual status of the species in the Balearic B. Seguí et al., Fossil avítauna trom Moleta. Cave 229 Islands is unknown, being established species were exhumed from the first populations at least in winter and during human occupation levels (Waldren, migration. 1982). There is so far no preholocenic evidence of the species in the archi Order COLUMBIFORMES pelago. Family COLUMBIDAE Genus Columba Family STRIGIDAE Columba sp. cf. C. lív/a Gmelin/oenas L. Genus Otus Otus scops (Linnaeus) MATERIAL: X 200-300: MNCM 47052, fragmented left tibiotarsus; MNCM MATERIAL: E 150-250: right radius. 47053, fragmented radius; MNCM 47054, MIN= 1. medial fragment of coracoid. MNI=1. This complete radius is identical to The fossil material, despite being those of Otus scops. The species has fragmented, is clearly attributed to the also been recorded in two other fossil genus Columba. Its size is smaller than sites in Mallorca (Ballman & Adrover, C. palumbus. It is difficult to differentiate 1970): Pedrera de S'Onix (Pliocene between C. lívía and C. oenas (Fick, Pleistocene boundary) and Cova de Son 1974; Weesie, 1988) and almost impos Bauga (Middle Pleistocene). It seems to sible with this poorly preserved fossil have been a common element of material. C. lívía was recorded in three Pliocene and Pleistocene Mediterranean Upper Pleistocene sites in Mallorca insular avifaunas (Alcover et al., 1992). (Florit & Alcover, 1987a; Alcover et al., Nowadays it is widespread in the Bale 1992) and, as well as C. oenas, in other arics. Mediterranean islands (Alcover et .al., 1992). Nowadays C. lívía is commom in Genus Athene the Gymnesic Islands, C. oenas being Athene noctua (Scopoli) rather rareo MATERIAL: Z 200-300: MNCM 2667, Order STRIGIFORMES distal fragment of right coracoid; MNCM Family TYTONIDAE 2666, fragmented left coracoid, MNCM Genus Tyto 2668, proximal fragment of scapula; AB: Tyto alba (Scopoli) MNCM 2969, medial fragment of left humerus. MNI= 1. o. MATERIAL: O 100-150: MNCM 2669, The coracoid of A. noctua and fragmented right tarsometatarsus; CD scops, the smaller Western Paleartic 100-150: MNCM 47055, left tibiotarsus; Z Strigidae can be differenciated due to 200-300: MNCM 47051, proximal frag the bigger Processus procoracoídeus in ment of right coracoid. MNI= 1. the former species, as well as for its The fossil material shares all osteo wider and more angular Facíes artícularí logical diagnostic features with the living clavícularís in the lalter. The coracoid of species. The tibiotarsus MNCM 47055 A. noctua is generally more strongly built o. shows juvenile characters. Tyto alba and rather bigger than that of scops. o. was found in three other Holocene sites The scapula of scops is more in the Gymnesic Islands (Mourer slightly built and its Corpus scapulae, Chauviré et al., 1977; McMinn & Alcover, especially at the Extremítas caudalís, is 1992). In Moleta Cave, remains of this more rounded. Humeri of both species 230 BoII. Soco Hist. Nat. Balears, 40 (1997) differ in size, A. noctua being bigger. On found in Eivissa and Karpathos (Alcover this basis, the medial fragment, has et al., 1992; Sondaar et al., 1995). been attributed to this species. Nowadays A. apus is a commom sum A. noctua is not present nowadays mer visitor in the Balearics. A. pallidus in the Balearic Islands The remains from is a rather rare summer visitor and mi Moleta Cave are the first evidence of grant. the species in the Gymnesic Islands, the material cited by Alcover et al. (1992) Order PICIFORMES being in fact the one described herein. Family PICIDAE The species has also been found in Genus Jynx other insular Mediterranean fossil sites Jynx torquilla L. (Alcover et al., 1992). In Crete (Weesie, 1988) the genus Athene evolved into an MATERIAL: O 100-150: MNCM 2622, endemic form, A. cretensis, developing a Distal fragment of right tarsometatarsus. lager size, longer legs and shortened MNI= 1. wings. On the other hand, remains from The tarsometatarsus of Jynx the Moleta Cave are identical to modern torquilla is easily recognized due to the A. noctua. large, well developed Trae/ea metatarsi IV, the remaining woodpeckers beeing Order APODIFORMES excluded due to the larger size. Family APODIDAE Blondel & Frochot (1976) and Genus Apus Blondel (1982) consider Jynx torquilla as Apus sp. cf. A. apus L. /pallidus a typical element of some modern Medi Shelley terranean insular avifaunas. It is present nowadays in the Balearic Islands, but MATERIAL: Z 200-300: MNCM 2653, left has never been found as a fossil spe ulna. MNI= 1 cies in any other island on the Medite Although comparison with Apus rranean. melba was not possible, this species can be excluded due to its larger size Order PASSERIFORMES (Florit & Alcover, 1987a). A. affínis is Family ALAUDIDAE al so excluded because of its smaller Genus Calandrella size. Osteology per se is close to that Calandrella sp. of the remaining congeneric species. Features of the fossil ulna are similar to MATERIAL: MNCM 47006, MNCM those of A. apus, but the bone is rather 47007, distal fragment of rostrum. MNI= stouter. A. pallidus cannot be excluded 2. due to the lack of a large comparative The rostrum of Alaudidae pos series. Morphology of both species sesses a Crista tomialis that narrows at proves to be similar (Weesie, 1988) and the anterior end of the narial openings thus the fossil ulna is attributed to Apus (Moreno, 1985). Both in Calandrella and apus/pallidus. in Melanocorypha, narial openings are The cited material is the first record round and smaller in proportion to the for the Gymnesic Islands, the citation of size of the beak, and the Processus Alcover et al. (1992) being the bone dorsonarialis and the whole rostrum is described in this work. Some other stouter than in the remaining genera in Pleistocene remains of the taxon were the family. Comparison with a single B. Seguí et al., Fossil avifauna trom Moleta Cave 231 specimen of M. calandra was enough to Alauda arvensis has never been discard this species, much larger than found in a Gymnesic site before. It was fossil remains. Coincidence of features recorded in Eivissa (Alcover et al., 1992; and size with Calandrella species was Sondaar et al., 1995). Nowadays the high, but not fui!. A more developed species is a common winter visitor and bony ridge ventrally to the Processus migrant in the Balearic Islands. dorsonarialis and the general thorough ness of the fossil material are the main Alaudidae sp. differences. In spite of this, attribution to this genus is attempted. MATERIAL: E 150-250: MNCM 2624, No other record of this taxon is right humerus; MNCM 2626, proximal known so far for the Balearic Islands. fragment of humerus. Sorne poorly preserved material has Genus Alauda been attributed to the family Alaudidae . Alauda arvensis Linnaeus Family MOTACILLlDAE MATERIAL: Z 200-300: MNCM 2623, Genus Anthus right humerus; MNCM 2631, MNCM Anthus pratensis (Linnaeus) 2630, left humerus; MNCM 2633, proxi mal fragment of left coracoid; MNCM MATERIAL: O 100-150: MNCM 2659, 2634, right coracoid. MNI= 1. right humerus. MNI= 1. Jánossy (1983) characterized the AII features described by Jimossy humerus of Alaudidae by having a (1983) to define the family Motacillidae unique, pneumatized depression in the are shared by the fossil humerus. Com proximal epiphysis, the Fossa pneu parison with Anthus and Motacilla spe moanconaea. In the distal epiphysis, the cies proves that the material should be Processus supracondylaris dorsalis is assigned to the first taxon due to the weakly developed. AII these features are development and position of the spine shared by the exhumed bones. Within located between the diaphysis and the the group, species distinction can be Processus supracondylaris dorsalis. In done merely on size (Jánossy, 1983). Motacilla the spine is sharp and long, The coracoid of Alaudidae is also and is located in the middle of the angle well defined due to the Foramen between the diaphysis and the pro pneumaticum, located at the ventral sur cessus. In Anthus it is a rather rounded face of the proximal epiphysis. The tuberculum (in some cases slightly elon Spina lateralis is lacking and the gated) placed on the processus. Inside Processus lateralis is heavily developed the genus, the humerus was compared (Moreno, 1985). Lullula arborea and both with A. campestris, A. spinoletta, A. Calandrella species are smaller than trivialis, A. berthelotii and A. pratensis, fossil coracoids while Galerida cristata being atributed to the latter mainly on and G. tecklae differ from the fossil biometric criteria (Moreno, 1986). material in their larger size, as well as The cited material is the first of the in the different shape of the proximal species tor both the Gymnesic archi epiphysis, in which the weakly developed pelago and tor the rest of the Mediterra Facies articularis clavicularis is a main nean islands. Nowadays A. pratensis is feature. Coincidence of characters with common in winter and during migration Alauda arvensis is fui!. in Mallorca. 232 Bol/. Soco Hist. Nat. Balears, 40 (1997) Motacillidae sp. Family PRUNELLlDAE Genus Prunella MATERIAL: O 100-150: MNCM 2660, Prunella modularis (Linnaeus) proximal fragment of left humerus; Z 200-300: MNCM 2684, MNCM 47099, MATERIAL: AB: MNCM 47009, distal proximal fragments of right humeri; fragment of rostrum. MNI= 1. MNCM 2685, proximal fragment of left The premaxilla in Prunel/a is char humerus. MNI= 2. acterized by a Processus dorsonarialis These four fragmented humeri lack which bends upwards at the distal end features for a specific attribution, and of narial openings. This feature is al so are referred to just family level. shared by so me Anthus species as well as by some Muscicapidae (Moreno, Family HIRUNDIDAE 1986), but in both cases the bill is more cf. Hirundo sp./ Ptyonoprogne rupestris slender. Although comparison with (Scopoli) Prunella collaris was not possible, this species may also be excluded due to its MATERIAL: Z 200-300: MNCM 2942, larger size (Moreno, 1987). proximal fragment of left ulna. MNI= 1. P. cOllaris was cited from two other The fact that osteology of the gymnesic fossil sites: the Cova Nova Hirundinidae is well defined (Moreno, (McMinn & Alcover, 1992) and the Pe 1986) enabled the determination of this drera de S'Onix (Mourer-Chauviré et al. ulna, while for the rest of the small 1977). There are no fossil records of the Passeriformes this skeletic element was species on the remaining Mediterranean not studied. islands. Nowadays the species is a win Riparia riparia and Delichon urbica ter visitor and migrant in the Balearics. were excluded because of their smaller size. Comparison with H. daurica was Family TURDIDAE not possible, and distinction between P. Genus Erithacus rupestris and H. rustica very difficult Erithacus rubecula (Linnaeus) (Moreno, 1986). The fossil bone may belong to either of these species. MATERIAL: E 150-250: MNCM 2615, left Only P. rupestris has been reco humerus; Z 200-300, MNCM 2739, ve red in the Gymnesic Islands (Florit & MNCM 2740, MNCM 2741, MNCM 2742, Alcover, 1987a) but almost all living spe left humeri; MNCM 2743, proximal frag cies have been found on other Mediter ment of left humerus; CD 100-150: ranean islands (Alcover et al., 1992). MNCM 2939, proximal fragment of right Nowadays there are both sedentary humerus; O 100-150: MNCM 2725, pro and migratory populations of P. rupestris ximal fragment of right humerus; MNCM in the Balearics, while H. rustica is a 2726, proximal fragment of left humerus; summer visitor. H. daurica is also seen AB: MNCM 2962, left coracoid. MNI= 7. in summer, but it seems to be rather In the humerus of Erithacus and scarce. Luscinia, the Fossa tricipitalis is less developed than in the remaining small sized Turdidae and Muscicapidae (Jánossy, 1983). In Muscicapa striata, the proximal epiphysis is stouter and