Spring migratory routes of eight trans-Saharan passerines through the central and western Mediterranean; results from a network of insular and coastal ringing sites zyx ANDREA PILASTRO*, STEFAN0 MACCHIO, ALBERT0 MASSI, ALESSANDRO MONTEMAGGIORI zyxwv & FERNANDO SPINA lstituto Nazionale per la Fauna Selvatica, Via Ca’ Fornacetta 9, 40064 Ozzano Emilia zyxw (BO), Italy Detailed information on spring migration routes of songbirds across the Mediterranean is still scanty. Results are presented here from a study on the occurrence of eight Palaearctic- African migrants across the western and central Mediterranean based on ringing data collected during the Progetto Piccole Isole, a co-ordinated project based on standardized mist netting at 21 islands and coastal stations. The species were Melodious Warbler Hip- polais polyglotta, Icterine Warbler Hippolais icterina, Bonelli’s Warbler Phylloscopus bonelli, Wood Warbler Phylloscopus sibilatrix, Pied Flycatcher Ficedula hypoleuca, Collared Flycatcher Ficedula albicollis, Red-backed Shrike Lanius collurio and Woodchat Shrike Lanius senator Their capture frequencies were expressed as the frequency of each species relative to the total number of migrants ringed at each station in order to correct for differences in capture effort between stations. For most species, the detailed, quantitative results con- firmed the scanty information available so far, whereas in the case of Melodious Warbler and Bonelli’s Warbler, evidence of undescribed and unexpected migration routes was found. These results suggest that standardized mist netting provides an important method for studying the migratory routes of small birds, in particular for those species which are not frequently ringed in their breeding and wintering grounds. Numbers of nocturnal songbird migrants may be particu- larly difficult to estimate visually at stopover sites, where birds stop for short periods during migration and their pres- ence even during daytime may be hard to quantify (Karr 1976). Visual estimates are heavily influenced by the skill of the observers (see Bibby zyxwvutsr et al. 1992), and caution is need- ed when comparing the frequencies of different species across localities. This method may therefore be difficult to apply in large-scale studies aimed at collecting information on the relative frequency of a given species and not simply its presence/absence. Because of the difficulty of visually estimating the direction of their movements, as is usually done with large, soaring migrants, the knowledge of song- bird migratory routes largely derives from recoveries of ringed birds (Zink 1973). On the other hand, estimations of the migratory routes of a species based on ringing recoveries are influenced not only by the number of individuals of that species staging in a given area but also by the fraction of individuals that were ringed elsewhere and by the probability of recovery, which may be dependent on factors such as hunting or ringing * Present address: Dipartimento di Biologia, Universita di Padova, via Trieste 75 1-35121 Padova, Italy. effort in that area as well as on local reporting rates (Le- breton & North 1993). A migration route can be completely overlooked if passing migrants belong to populations which are not ringed or when the route involves geographic areas lacking ringing effort. Coordinated ringing activities in southern European countries are relatively recent developments, and data on songbird migration movements within the Mediterranean area, and in particular on those species breeding in this region, are scanty and anecdotal or even absent (Cramp 1992, Cramp & Perrins 1993). From a more general perspective, spring migration routes of Palaearctic-African songbirds are less well known than their autumn movements (Zink 1973, Dowsett et zyx aI. 1988); this may be a consequence of the large effort traditionally devoted to ringing during the postnuptial movements, es- pecially in northern Europe. The routes followed by the same birds in autumn and spring may be different (“loop migra- tion”; Zink 1973, Alerstam 1990, Berthold 1993), and for the reasons listed above, such routes are not always easily studied. Ringing stations distributed over a wide geographic area, where contemporary ringing activities are based on mist netting and standardized methods of data collection, may provide a unique opportunity to get an overview of the mi- 591