As02-019.pdf

Migratory song thrushes Turdus philomelos
hunted in Europe: survival rates and other
demographic parameters
Vladimir A. Payevsky1 and Vadim G. Vysotsky
Long-distance ringing recoveries of song thrushes ringed in the Eastern Baltic between 1957and 1999 were analysed to determine the circumstances of recovery and annual survivalrate. Of 358 recoveries, 86 % referred to birds killed by hunters and bird-catchers. The high-est proportions of recoveries due to hunting were recorded in France, Italy, Spain, and Por-tugal (73–93 %), whereas in other countries only 28 % were recovered in this way. Survivalprobabilities of the song thrush estimated for different time periods varied between 0.368 and0.420 for first-year birds and between 0.561 and 0.633 in adults. In the British Isles, wherethe song thrush is mainly a year-round resident and is not a quarry bird, adult survival ratesdid not differ from the values obtained for the Baltic populations, whereas first-year survivalwas somewhat higher than in Baltic birds.
Key words: Song thrush, Turdus philomelos, first-year survival rate, annual adult survival rate,age distribution, hunting bag.
Zoological Institute, Russian Academy of Sciences, 199034 St. Petersburg, Russia.1Corre-sponding author: e-mail: [email protected] Turdus thrushes are important quarry species in West- Material and methods
ern Europe (McCulloch et al. 1992), with annual bagscomprising many tens of millions of birds (Aebischer Birds have been trapped and ringed annually since et al. 1999, Ricci 2001). Hunting is evidently a signifi- 1957 by the staff of the Biological Station Rybachy of cant cause of mortality to song thrushes Turdus philo- the Russian Academy of Sciences Zoological Institute.
From April until November birds are caught at two The migratory route of several thrush species from sites on the Courish Spit: at the permanent ‘Fringilla’ the populations of north-western Russia and Finland field station (55º 05’ N, 20º 44’ E) and at the Rybachy passes through the Eastern Baltic (Payevsky 1973).
field site on Rossitten Cape (55º 09’ N, 20º 51’ E). At One of segments of this route is the Courish (=Curon- ‘Fringilla’ birds are trapped in stationary Rybachy-type ian) Spit in the Kaliningrad Region in the south-eastern traps (for a detailed description of the traps see Payev- corner of the Baltic Sea, where massive bird trapping sky 2000), and at the Rybachy site birds have been provides a large data set on migrating birds and allows an analysis of their demographic parameters. The aims Lists of long-distance recoveries of birds on the of this paper are: (1) to compare recovery circumstan- Courish Spit in 1956–1999 have been published by ces of song thrushes in different countries of Europe; Payevsky (1973) and Bolshakov et al. (1999, 2000, (2) to determine the age distribution of song thrushes 2001). The latter three publications gave all recoveries in hunting bags; and (3) to estimate age-specific sur- available in the end of 2000. We have analysed a total vival rates and to discuss the effects of hunting on song of 358 recoveries of song thrush from 22 907 birds ringed. Most thrushes were ringed during migration.
V. Payevsky & V. Vysotsky: Demography of song thrushes hunted in Europe These migratory birds belong to populations from probability that a bird alive on 10 June survives until 9 Finland, north-western Russia and the eastern Baltic June of the next year and therefore covers 12 months.
The data were organised in accordance with these peri- All trapped birds were aged using plumage charac- ods to make our results comparable with the published teristics, moult and wear, following Svensson (1970, data of the British trust for Ornithology (Thomson et al.
1992) with our additions (Vinogradova et al. 1976). The main age character when ringing passage populations We used 178 dead recoveries of full-grown (first of song thrushes was the incomplete replacement of the year) thrushes during 1958–1978 to estimate survival rate. We modelled survival and reporting rates using To calculate the age distribution of song thrushes in program MARK v. 2.1 (‘recoveries only’ option; Cooch the hunting bag, we considered seasons of the avian & White 2001, White & Burnham 1999). To estimate annual cycle, not calendar seasons. According to data survival rate we applied age-dependent models (Free- on breeding biology (Haartman 1969, Malchevsky & man & Morgan 1992; Catchpole & Morgan 1996) due Pukinsky 1983) the average hatching date of these to young-only data being available. Age-dependent sur- populations is 10 June. As all birds hatch between May vival models were fitted by MARK using the parame- and July, we have denoted the immature age group ter index matrix and sine link function.
as birds after independence from their parents until The final model was selected by starting from a ‘glo- 10 June of the next calendar year. The average age bal’ model in which first-year survival depends on the of our immature thrushes at ringing is 3 months 20 year of ringing and adult annual survival probability varies with the age of bird, and then selecting between To estimate survival rate, we used passage thrushes a variety of a set of alternative submodels. We believe ringed only between 10 September and 21 October (au- that for song thrushes aged two years or older, age-de- tumn peak in Fig. 1). Accordingly, the midpoint of the pendence of annual survival probability (S) was not ne- ringing period falls on 30 September. The formal con- cessary as it is generally considered that survival stabi- dition that the sampling period must be short as com- lises by the age of two years. Thus we chose the ‘sub- pared to the period between sampling occasions (Brow- global’ model (using the terminology of Freeman & nie et al. 1985) was thus observed. Survival probabili- Morgan 1992) in which first-year survival probability ty (S) was defined as the probability that a bird alive in varies with year of ringing, and age-dependence of sur- year i survives until year i+1. First-year survival pro- vival for adult birds aged ≤ 2 years and reporting pro- bability was defined as the probability that a bird alive bability (r) varies with year of recovery. We assume that at the midpoint of the ringing period (30 September) all birds aged > 2 years have the same annual survival survives until 10 June (the middle of their first breeding probability. The limitation of the age-dependent models season) and therefore covers 8 months and 20 days.
is that the age-specific reporting rate cannot always be Adult annual survival probability was defined as the Figure 1. Relationship
It is well known that first-year reporting rate may be Results and discussion
higher than the adult reporting rate, which may poten-tially cause a serious bias in the estimate of first-year survival. Freeman & Morgan (1992), however, foundthat this bias is small and may be ignored in practice.
The overall recovery rate of song thrushes is very low For British song thrushes, Thomson et al. (1997, 1999) (no more than 1.5–2 %), as in other passerines. On the found no evidence that reporting probability varies be- basis of 1958–1999 data there is a correlation between tween first-year and adult birds. Even though it was the numbers of birds ringed and number of recoveries later shown that reporting rate differs between age clas- (rs = 0.72, n = 41, P < 0.001). Comparison of ringing ses (W. Peach, pers. comm.), we believe that the use of dates on the Courish Spit and recovery dates indicates models based on age-independent reporting rate remain seasons of most intensive contacts of thrushes with hu- mans (Fig. 1). The number of recoveries peaks during To describe the various Freeman-Morgan submodels winter months, corresponding to the hunting season in we use following notation: SFY denoted first-year sur- west European countries. Song thrushes are not a quar- vival probability, SAD1 denoted constant annual survi- val probability of adult birds aged 1 year, SAD2 denotedconstant annual survival probability of adult birds aged Finding circumstances of ringed song thrushes in 2 years or older, SAD denoted constant annual survival probability of adult birds since age of 1 year. Symbol(t) refers to general dependence on time, and (.) refers Of 358 recoveries of song thrush available to us, to no dependence on time. For reporting rates symbol 255 (71 %) were reported as ‘shot’. Cases reported as ‘caught, further fate unknown’ (29 recoveries), and A simulation approach was used to calculate the ‘manner of recovery unknown’ (23 recoveries) are also goodness-of-fit between our data and the global model.
likely to belong to this category. This assumption is We used c (variance inflation factor; Cooch & White based on the existence of a popular tradition of netting 2001) as a measure of overdispersion within each data- migratory songbirds for food, especially in Italy (Ghigi set, and to adjust sampling variances. We calculated c 1958). It is probably no exaggeration to assume that as the observed deviance from the global model divided of all recovered thrushes with rings, 307 (86 %) were by the mean deviance from 500 parametric bootstrap si- mulations of the global model (Cooch & White 2001).
Differential hunting pressure on thrushes between A distribution of expected c for the global model was countries is clearly shown by the proportion of all find- generated from 500 simulations. Observed c was then ing circumstances that comprise hunting recoveries.
compared to the distribution of simulated values to de- The highest proportion is found in Italy (92.9 ± 2.8 %), termine whether the global model was an adequate fit France (92.3 ± 2.1 %), Spain (81.1 ± 4.1 %), and Por- to the data. Goodness-of-fit tests based on simulations tugal (72.7 ± 13.4 %), where thrushes are evidently sub- indicated that the global model was a satisfactory start- ject to considerable hunting pressure. The difference ing point for all datasets (P-values > 0.05). We pro- from other European countries, where the proportion of ceeded to improve model fit by fitting nested models birds killed is only 27.7 %, is highly significant (χ2 = with reduced numbers of parameters. Model selection was based on the difference in QAICc (quasi-Akaike’sInformation Criterion; Cooch & White 2001) values Age distribution of song thrushes in the hunting between models (∆QAICc). The best fitted model had a ∆QAICc of zero. We chose a model with constant sur-vival rate for estimating survival probability for all da- Among nocturnal migrants ringed on the Courish Spit tasets. Other statistical tests followed Sokal & Rohlf the vast majority of individuals captured in autumn are immatures hatched in the current year (Payevsky 1985,1998). This phenomenon of a very low proportion ofadults in captures has been termed the ‘coastal effect’, V. Payevsky & V. Vysotsky: Demography of song thrushes hunted in Europe Table 1. Age distribution of song thrushes in the hunting
ted separately (Table 1). As the proportion of adult song thrushes in autumn does not exceed 7 % on average, asafe assumption is that of all birds aged as ‘full grown’, c. 93 % were immatures. We can therefore assume that over 60 % of song thrushes hunted are immatures, and that the immature/adult ratio in the hunting bag is 1.6 Despite these calculations, some doubt remains whether the age distribution in the hunting bag indeed reflects the true proportion of age classes in nature. Al- though immatures and adults are shot in equal propor- tions, we have no information on adults just because they have not been ringed. It may be assumed, however, that immatures are more likely to be shot due to their Some published data sets on breeding performance of thrushes (Haartman 1969, Malchevsky & Pukinsky1983, Payevsky 1985) suggest that the mean producti- * the EURING codes 1, 3, and 5 in the process of ring- vity of the song thrushes studied in north-western Rus- sia and Finland does not exceed 3 young fledged per * * the EURING codes 2, 4, and 6 in the process of ring- breeding pair per breeding attempt (estimate based on clutch size of 4.3–5.2 eggs, breeding success of 42–61 % and 10–20 % of pairs which have a second breed- from the similar example of passerine birds trapped at ing attempt after a successful first brood). Thus, the im- coastal sites in North America (Ralph 1978). Disregard- mature/adult ratio just after breeding season should be ing the causes of such an effect, it should be emphasised c. 1.5. This figure is very close to that obtained from the that it concerns only nocturnal migrants as opposed to species that migrate by day. Such a high proportion ofimmature birds is not consistent with the potential pro- ductivity of the populations involved. The impact of the‘coastal effect’ on demographic studies of migrants is In a number of European countries songbird hunting apparent, in that we cannot use the trapping data of noc- was legally restricted in 1979 (McCulloch et al. 1992).
turnal migrants in studies of age structure and its annu- We have therefore analysed separately the survival rates of thrushes before 1979. Furthermore, because an un- The exact age of passage song thrushes has been re- derstanding of the impact of hunting on survival rates corded during ringing on the Courish Spit since the may be best achieved by comparing the rates between early 1970s. It became immediately apparent that the hunted and non-hunted populations (Aebischer et al.
bulk of birds captured by us are immatures. Trapping 1999), we have compared the survival rates of song data for 1972–1995 have shown that in some years not thrushes that were ringed in the Eastern Baltic and a single adult song thrush was captured. Of 11 544 song spend their winter in continental Europe and the Medi- thrushes trapped in those years, adults comprised on terranean regions with the published data on survival of average 6.5 % during autumn and 19.2 % during spring song thrushes on the British Isles over the same time pe- riods (Thomson et al., 1997, Siriwardena et al., 1998).
Although most song thrushes ringed on the Courish Most British thrushes remain on the islands all year Spit are immatures, the exact age was not known in all round, and only some birds from southern England mi- cases. Therefore when we calculated the age distribu- grate to France and Spain (Ashmole 1962, Cramp tion of thrushes in hunting bags, birds with known and 1988). The main difference in the life cycles of these unknown age at ringing (307 birds in total) were trea- birds is that within British Isles they are not hunted.
Table 2. Selection of Freeman-Morgan age-specific models for the song thrush data for 1958–1978. Model selection
based on c=1.485.
The results of fitting Freeman-Morgan age-specific old. Our data refer to the period between 3 months 20 models are shown in Table 2 and a comparison of sur- days old and 12 months, and the difference of 1 month vival rates of hunted and non-hunted populations of 20 days between studies must be taken into account in song thrushes in Table 3. The use of age-dependent any comparison. We used the published data on British models made it possible to estimate the survival rates song thrushes (Thomson et al. 1997, 1999) to make this of immature and adult thrushes on the basis of ringing correction. In Britain from 1962 to 1975 the survival data on immatures only. The first-year survival rate of rate was 0.376 from fledging until 2 months old, and Baltic birds wintering on the continent appeared to be 0.484 between 2 and 12 months old. Survivorship until slightly below the first-year survival rate of British song 12 months after fledging is therefore 0.376*0.484 = thrushes. The values of first-year survival rate obtained 0.182. We are interested in the relative frequency of by us (Table 3) are in the lower part of the range of first- birds that survive until 3 months 20 days. We used lin- year survival rates (0.371–0.528) reported for British ear interpolation between 0.376 and 0.182 and estima- song thrushes for 1962–1995 (Siriwardena et al. 1998).
ted the survivorship to 3 months 20 days as 0.344. Then To compare our data on first-year survival rate with the survival rate between 3 months 20 days and 12 the published British data, it is necessary to use com- months will be 0.182/0.344 = 0.529 in the British birds.
parable periods. In Britain, the first-year survival rate This value is directly comparable with estimate of 0.411 was estimated for the period between 2 and 12 months Table 3. Survival rates of song thrushes in intensively hunted populations (birds wintering on the European continent)
and in non-hunted populations (British Isles).
V. Payevsky & V. Vysotsky: Demography of song thrushes hunted in Europe Our data on first-year survival rate (Table 3) are sub- weather in winter (Baillie 1990, Thomson et al. 1997, stantially lower than figures reported from Britain. In Siriwardena et al. 1998, Wernham et al. 1998). Other contrast, adult survival rates in the Baltic area and in studies, especially of waterfowl, have shown that hunt- Britain are similar. At the same time, the values of adult ing mortality is compensated for by other forms of mor- annual survival rate obtained by us (Table 3) are con- tality (Burnham & Anderson 1984, Mihelsons et al.
siderably higher than those reported from the adjacent 1985). In other words, mortality due to hunting is not additive to the natural mortality, as long as the formerdoes not exceed a certain threshold. Productivity and Compensatory reproduction and population mortality are compensated reciprocally, presumably on the basis of density-dependent regulation mechanisms.
Legislative restrictions on hunting songbirds estab- It is not immediately clear to what extent the survival lished in 1979 seem to have had a very limited effect on rate is governed by hunting pressure. Along with hunt- the real hunting pressure. Some decrease in the report- ing, a number of other factors affect survival, many of ing rate of rings is believed by many authors to be due which are density-dependent, and it is virtually impos- not to a smaller hunting pressure but to the fear of sanc- sible to isolate the importance of hunting pressure tions (McCulloch et al. 1992). In our data, however, alone. The problem is best approached by a comparing 65 % of recoveries after 1979 were still reported as estimates of survival rates from hunted and non-hunted ‘shot’. We therefore believe that our conclusions con- populations, as in our example with Baltic and British cerning the impact of hunting on the survival of song thrushes. In this case, however, different living thrushes will also be relevant for the period after the conditions may influence the annual survival rate at least as profoundly as hunting pressure. For instance, The results of our comparison of survival rates in dif- these two populations differ in migratory habits. Win- ferent regions and years imply that if hunting does in- ter movements, if any, of British thrushes are an order duce some additive mortality of song thrushes, it occurs of magnitude shorter than the movements of their con- in immatures only. Compared with adults, first-year specifics from Finland and Eastern Baltic. It has been birds seem to be more vulnerable to an adverse envi- suggested that ‘relatively low survival rates may be an ronment. It has been shown for British song thrushes intrinsic feature of largely migratory song thrush popu- that ‘first-year survival was lower during years with lations, and the apparently high index of hunting loss cold or dry winters and adult survival was lower during may simply be a reflection of the passage of the birds years with cold winters’ (Thomson et al. 1997). What through the Mediterranean’ (Aebischer et al. 1999). Ac- could be the reason under these age-related differences? cordingly, seasonal migrations and/or hunting pressure We suggested above that immatures are more vulnera- may be the factors which reduce the survival of song ble because they migrate for the first time in their lives.
thrushes from the Eastern Baltic. That migration cancause additive mortality has been shown in robins Erithacus rubecula by Adriaensen & Dhondt (1990).
On the other hand, a study of the adaptive significance of seasonal migration in various bird species showed that mortality during migration is no higher than during In some studies, e.g. on willow grouse Lagopus la- gopus, a significant decrease has been shown in survi- val rates in areas with a strong hunting pressure (Smith & Willebrand 1999). Decreases in annual survival rates of British blackbirds Turdus merula, song thrushes and Figure 2. Fluctuations of song thrush numbers on the
mistle thrushes T. viscivorus were found only in years Courish Spit in 1960–1978 (total n = 5622). Birds were when populations declined. In the song thrush this ef- trapped by standard Rybachy-type traps, located in the fect occurred in juveniles only and was due to severe same sites throughout (Payevsky 1985).
On the other hand, there is no reason to believe that Bal- bird populations. Ringing & Migration 19 (Suppl.): tic thrush populations are endangered due to hunting.
During the period 1958–1978 numbers of passage song Adriaensen, F. & Dhondt, A. A. 1990. Population dy- thrushes in the Eastern Baltic fluctuated without show- namics and partial migration of the european robin ing significant trends (Fig. 2), and similar patterns for (Erithacus rubecula) in different habitats. J. Anim.
song thrushes during these years were also observed in Finland and Sweden (Järvinen & Väisänen 1978a,b, Ashmole, M. J. 1962. The migration of European thrus- hes: a comparative study based on ringing recove- In conclusion, we have shown that in the Baltic area ries. Ibis 104: 314–346 and 522–559.
adult survival rates did not differ from the values ob- Baillie, S. R. 1990. Integrated population monitoring of tained for the British populations, whereas first-year breeding birds in Britain and Ireland. Ibis 132: 151– survival was lower than in British birds. This implies the occurrence of compensatory hunting mortality for Bolshakov, C. V., Shapoval, A. P. & Zelenova, N. P.
adult song thrushes. The absence of any indication that 1999. Results of bird trapping and ringing by the Baltic populations have shown a decline suggests that Biological Station ‘Rybachy’ on the Courish Spit in hunting mortality of first-year thrushes may be bal- 1998. Avian Ecol. Behav. 2: 105–150.
anced by a higher productivity. We can only guess Bolshakov, C. V., Shapoval, A. P. & Zelenova, N. P.
which population mechanisms cause such higher pro- 2000. Results of bird trapping and ringing by the ductivity. The published breeding data (Haartman Biological Station ‘Rybachy’ on the Courish Spit in 1969, Malchevsky & Pukinsky 1983, Payevsky 1985, 1999. Avian Ecol. Behav. 4: 85–145.
Cramp 1988) suggest that neither clutch size nor brood Bolshakov, C. V., Shapoval, A. P. & Zelenova, N. P.
size show a marked difference between these popula- 2001. Results of bird trapping and ringing by the tions. The parameter most difficult to measure in field Biological Station ‘Rybachy’ on the Courish Spit: studies of any bird species is the proportion of normal long-distance recoveries of birds ringed in second clutches laid after a successful rearing of the 1956–1997. Avian Ecol. Behav. Suppl 1: 1–126.
first brood. The reason is that they look very similar to Brownie, C., Anderson, D. R., Burnham, K. P. & Rob- the repeat clutches made after nest loss. The only plau- son, D. S. 1985. Statistical inference from band re- sible assumption is that in the Baltic populations which covery data: a handbook. U. S. Fish and Wildlife are subject to a hunting pressure, more pairs are double- Service. Resource Publ. 156, Washington.
brooded annually than in other populations. Burnham, K. P. & Anderson, D. R. 1984. Tests of com- pensatory vs. additive hypotheses of mortality in Acknowledgements. The authors are grateful to their
colleagues who have participated in the long-term trap- Catchpole, E. A. & Morgan, B. J. T. 1996. Model se- ping and ringing of thrushes on the Courish Spit. This lection in ring-recovery models using score tests.
study was supported by a grant from OMPO (France) on the project ‘Long-term monitoring of the thrushes’ Cooch, E. & White, G. 2001. Program MARK: Analy- and by a grant N 02-04-48617a from Russian Founda- sis of data from marked individuals. 2nd edn.
tion of Basic Research. The authors are specially grate- ful to Dr. Will Peach for his constructive criticism and Cramp, S. (ed.). 1988. The birds of the Western Pale- a number of valuable comments and recommendations arctic. Vol. V. Tyrant flycatchers to thrushes. Oxford which helped us the improve the manuscript consider- Freeman, S. N. & Morgan, B. J. T. 1992. A modelling strategy for recovery data from birds ringed as nest-lings. Biometrics 48: 217–235.
Ghigi, A. 1958. L’uccellagione in Italia. Vie Italia 64: Aebischer, N. J., Potts, G. R. & Rehfisch, M. 1999.
Haartman, von, L. 1969. The nesting habits of Finnish Using ringing data to study the effect of hunting on birds. I. Passeriformes. Comment. Biologic. Societ.
V. Payevsky & V. Vysotsky: Demography of song thrushes hunted in Europe Ricci, J.-C. 2001. The thrushes hunted in Europe: con- Hjort, C., Lindholm, C.-G. & Pettersson, J. 1981. Ring- servation status and number estimates. OMPO märkningssiffror vid Ottenby fågelstation 1946– 1980. Special report from Ottenby Bird Observa- Robinson, R. A., Baillie, S. R. & Peach, W. J. 2002.
How important is post-fledging survival? 23 Int.
Järvinen, O. & Väisänen, R. A. 1978a. Recent changes Ornithol. Congr. Beijing. Abstract Volume: 163– in forest bird populations in northern Finland. Ann.
Siriwardena, G. M., Baillie, S. R. & Wilson, J. D. 1998.
Järvinen, O. & Väisänen, R. A. 1978b. Long-term Variation in the survival rates of some British pas- population changes of the most abundant south serines with respect to their population trends on Finnish forest birds during the past 50 years. J. Or- Smith, A. & Willebrand, T. 1999. Mortality causes Malchevsky, A. S. & Pukinsky, Yu. B. 1983. The birds and survival rates of hunted and unhunted Willow of Leningrad region and adjacent areas. Vol. 2. Pas- Grouse. J. Wildl. Manage. 63: 722–730.
serines. Leningrad University Press, Leningrad. [In Sokal, R. R. & Rohlf, F. J. 1998. Biometry. 3rd edn. McCulloch, M. N., Tucker, G. M. & Baillie, S. R. 1992.
Svensson, L. 1970. Identification guide to European The hunting of migratory birds in Europe: a ringing passerines. Naturhist. Riksmuseet, Stockholm.
recovery analysis. Ibis 134. Suppl. 1: 55–65.
Svensson, L. 1992. Identification guide to European Mihelsons, H., Mednis, A. & Blums, P. 1985. Regula- passerines. 4th ed. Naturhist. Riksmuseet, Stock- tory mechanisms of numbers in breeding popula- tions of migratory ducks. Acta XVIII Congr. Int. Or- Thomson, D. L., Baillie, S. R. & Peach, W. J. 1997. The demography and age-specific annual survival of Payevsky, V. A. 1973. Atlas of bird migration according song thrushes during periods of population stability to banding data at the Courland Spit. Pp 1–124 in: and decline. J. Anim. Ecol. 66: 414–424.
Bykhovsky, B. E. (ed.). Bird migration – ecological Thomson, D. L., Baillie, S. R. & Peach, W. J. 1999. A and physiological factors. Halstead Press. John method for studying post-fledging survival rates using data from ringing recoveries. Bird Study 46 Payevsky, V. A. 1985. Demography of birds. Nauka Vinogradova, N. V., Dolnik, V. R., Yefremov, V. D. & Payevsky, V. A. 1990. Population dynamics of birds ac- Payevsky, V. A. 1976. Identification of sex and age cording to trapping data on the Courish Spit of the in passerine birds of the USSR. Nauka Press, Mos- Baltic sea during twenty seven years. Zool. Zhurnal 69: 80–93. [In Russian with English summary].
Wernham, C. V., Baillie, S. R., Clark, J. A., Crick, H.
Payevsky, V. A. 1998. Age structure of passerine mi- Q. P. & Siriwardena, G. M. 1998. The BTO inte- grants at the eastern Baltic coast: the analysis of the grated population monitoring programme: the con- ‘coastal effect’. Ornis Svecica 8: 171–178.
tributions of ringing. Acta Zool. Lithuanica, Orni- Payevsky, V. A. 1999. Adaptive significance of sea- sonal migrations: are annual migrations dangerous White, G. C. & Burnham, K. P. 1999. Program MARK: for birds? Russian J. Zool. 3: 78–84 [translated from survival estimation from populations of marked ani- mals. Bird Study 46 (Suppl.): 120–139.
Payevsky, V. A. 2000. Rybachy-type trap. Pp 20–24 in: Busse, P. (ed.). Bird station manual. Gdansk Uni-versity, Gdansk.
Ralph, C. J. 1978. Disorientation and possible fate of young passerine coastal migrants. Bird-Banding

Source: http://www.huntinginmalta.org.mt/uploads/90/903/Migratory_Song_Thrushes.pdf